EPA-650/2-74-126-b
DECEMBER 1974
Environmental Protection Technology Series
?:!?:;:
-------
EPA-650/2-74-126-B
PROCEEDINGS:
SYMPOSIUM
ON FLUE GAS DESULFURIZATION
ATLANTA, NOVEMBER 1974
VOLUME II
ROAP NO. 21ACX-AA
Program Element No. 1AB013
Chairman: E.L. Plyler
Vice-Chairman: W.H. Ponder
Sponsored by
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, N. C. 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON D.C. 20460
December 1974
-------
EPA REVIEW NOTICE
This report has been reviewed by the National Environmental Research
Center - Research Triangle Park, Office of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency, have been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:
1. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2 . ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5 . SOCIOECONOMIC ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to
develop and demonstrate instrumentation, equipment and methodology
to repair or prevent environmental degradation from point and non-
point sources of pollution. This work provides the new or improved
technology required for the control and treatment of pollution sources
to meet environmental quality standards.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
Publication No. EPA-650/2-74-126-b
11
-------
PREFACE
The development of technologies to control sulfur dioxide emissions is
an issue of national importance. Indications that the use of sulfur-containing
fossil fuels to generate electricity will increase by 50 percent by 1985,
recognition that more than half of all sulfur dioxide ($02) emissions are caused
by electric power generation, and documentation of the deleterious effects
caused by such emissions make the acceleration of development to commercial
application of SC>2 control technology one of the most important goals of the
U.S. Environmental Protection Agency (EPA). The Control Systems Laboratory
(CSL), as part of EPA's Office of Research and Development, has accelerated
the development of flue gas desulfurization (FGD) technology so that it is now
in the process of commercialization. Each year, CSL sponsors a Symposium which
brings together users and developers of the technology for open exchange and
discussion of their experience and results. The 1974 Symposium was held on
November 4-7 at the Sheraton-Biltmore Hotel in Atlanta.
Commercial application of full-scale FGD processes was the primary topic
of the Symposium which provided a means to transfer the latest information
on the status, development and assessment of current FGD processes to developers,
vendors, potential users, and those concerned with regulatory deadlines and
enforcement. The Symposium was attended by more than 600 national and inter-
national representatives of utilities, vendors, government, and universities.
One-third of the presentation was assessment and status reports from utilities
which are utilizing full-scale FGD systems. Information from small- and
large-scale pilot testing of alternate FGD systems, developments in FGD
by-product disposal and utilization, and cost studies of FGD application
completed the extensive overview of FGD technology.
In concluding remarks, it was noted by an international FGD technology
consultant that the Symposium is recognized as the world's most significant
conference on S02 pollution control by FGD technology. The growth of the
FGD Symposium to this latest, widely attended, in-depth information exchange
forum supports the conclusion that FGD is the only viable near-term alternative,
other than the burning of clean fuels, that will permit compliance with current
regulatory requirements.
The contents of these Proceedings are comprised of copies of the
participating authors' papers as received. Although the papers have been
reviewed and approved for publication by the Environmental Protection Agency,
approval does not indicate that the contents necessarily reflect the views
and/or policies of the Agency. The mention of trade names or commercial
products does not constitute endorsement or recommendation for use.
As supplies permit, copies of the Proceedings are available free of
charge and may be obtained by contacting the Air Pollution Technical
Information Center, Environmental Protection Agency, Research Triangle
Park, North Carolina 27711.
iii
-------
CONTENTS
TITLE PAGE
VOLUME I
OPENING SESSION
Keynote Address - THE ROLE OF ENVIRONMENTAL HEALTH ASSESSMENT
IN THE CONTROL OF AIR POLLUTION
John Finklea, Environmental Protection Agency,
National Environmental Research Center,
Research Triangle Park, North Carolina 1
FLUE GAS DESULFURIZATION AND OTHER ALTERNATIVES FOR PRODUCING
ELECTRICITY FROM COAL
S.J. Gage, Environmental Protection Agency,
Washington, D. C
STATUS OF FLUE GAS DESULFURIZATION SYSTEMS IN THE
UNITED STATES
T.W. Devitt and F. Zada, PEDCo-Environmental, Inc.,
Cincinnati, Ohio 17
STATUS OF FLUE GAS DESULFURIZATION TECHNOLOGY IN JAPAN
J. Ando, Chuo University, Tokyo, Japan 125
COST COMPARISONS OF FLUE GAS DESULFURIZATION SYSTEMS
G.G. McGlamery and R.L. Torstrick,
Tennessee Valley Authority,
Muscle Shoals, Alabama 149
NON-REGENERABLE PROCESSES SESSION
EPA/RTP PILOT STUDIES RELATED TO UNSATURATED OPERATION OF
LIME AND LIMESTONE SCRUBBERS
R.H. Borgwardt, Environmental Protection Agency,
Research Triangle Park, North Carolina 225
LIMESTONE AND LIME TEST RESULTS AT THE EPA ALKALI
SCRUBBING TEST FACILITY AT THE TVA SHAWNEE POWER PLANT
M. Epstein, Bechtel Corporation,
San Francisco, California 241
OPERATIONAL STATUS AND PERFORMANCE OF THE ARIZONA PUBLIC
SERVICE COMPANY FLUE GAS DESULFURIZATION SYSTEM AT THE
CHOLLA STATION
L.K. Mundth, Arizona Public Service,
Phoenix, Arizona 307
WET SCRUBBER OPERATING EXPERIENCE AT LA CYGNE
STATION UNIT NO. 1
C.F. McDaniel, Kansas City Power and Light,
Kansas City, Missouri 319
-------
TITLE PAGE
DUQUESNE LIGHT COMPANY, PHILLIPS POWER STATION,
LIME SCRUBBING FACILITY
S.L. Pernick, Jr. and R.G. Knight,
Duquesne Light Company,
Pittsburgh, Pennsylvania 329
THE HORIZONTAL CROSS FLOW SCRUBBER
A. Weir, Jr., J.M. Johnson, D.G. Jones,
and S.T. Carlisle, Southern California Edison,
Rosemead, California 357
OPERATIONAL STATUS AND PERFORMANCE OF THE LOUISVILLE
FGD SYSTEM AT THE PADDY'S RUN STATION
R.P. Van Ness, Louisville Gas and Electric
Company, Louisville, Kentucky 389
DISPOSAL OF BY-PRODUCTS FROM NON-REGENERABLE FLUE
GAS DESULFURIZATION SYSTEMS
J. Rossoff, R.C. Rossi, L.J. BornsCein,
The Aerospace Corporation,
El Segundo, California
J.W. Jones, Environmental Protection Agency,
Research Triangle Park, North Carolina 399
AN OVERVIEW OF DOUBLE ALKALI PROCESSES FOR
FLUE GAS DESULFURIZATION
N. Kaplan, Environmental Protection Agency,
Research Triangle Park, North Carolina 445
INITIAL OPERATING EXPERIENCES WITH A DUAL-ALKALI
S02 REMOVAL SYSTEM
PART I - PROCESS PERFORMANCE WITH A COMMERCIAL
DUAL-ALKALI S02 REMOVAL SYSTEM
T.T. Dingo, General Motors Corporation, Parma, Ohio .... 517
PART II - EQUIPMENT PERFORMANCE WITH A COMMERCIAL
DUAL-ALKALI S02 REMOVAL SYSTEM
E.J. Piasecki, General Motors Corporation,
Parma, Ohio 539
EPA-ADL DUAL ALKALI PROGRAM—INTERIM RESULTS
C.R. LaMantia, R.R. Lunt, J.E. Oberholtzer, E.L. Field,
Arthur D. Little, Inc., Cambridge, Massachusetts
N. Kaplan, Environmental Protection Agency,
Research Triangle Park, North Carolina 549
-------
TITLE PAGE
VOLUME II
REGENERABLE PROCESSES SESSION
SEW ENGLAND SO, CONTROL PROJECT FINAL RESULTS
G.R. Koehler, E.J. Dober, Chemical Construction
Corporation, New York, New York 671
ASSESSMENT OF PROTOTYPE OPERATION AND FUTURE EXPANSION
STUDY - MAGNESIA SCRUBBING MYSTIC GENERATING STATION
C.P. Quigley, J.A.. Burns, Boston Edison Company,
Boston, Massachusetts 709
MAG-OX SCRUBBING EXPERIENCE AT THE COAL-FIRED DICKERSON
STATION, POTOMAC ELECTRIC POWER COMPANY
D.A. Erdman, Potomac Electric Power Company,
Washington, D.C 729
POWER PLANT FLUE GAS DESULFURIZATION BY THE
WELLMAN-LORD S02 PROCESS
PART I - THE DEAN H. MITCHELL STATION
E.L. Mann, Northern Indiana Public Service Company,
Michigan City, Indiana 739
PART II - CONTINUING PROGRESS FOR THE WELLMAN-LORD
S02 PROCESS
E.E. Bailey, Davy Powergas, Inc.,
Lakeland, Florida 745
THE CAT-OX DEMONSTRATION PROGRAM
E.M. Jamgochian, The Mitre Corporation,
McLean, Virginia
W.E. Miller, Illinois Power Company,
Decator, Illinois , 761
THE SHELL FLUE GAS DESULFURIZATION PROCESS
J.B. Pohlenz, Universal Oil Products Company,
Des Plaines, Illinois S07
STATUS REPORT ON CHIYODA THOROUGHBRED 101 PROCESS
M. Noguchi, Chiyoda International Corporation,
Seattle, Washington 837
FLUE GAS DESULFURIZATION BY-PRODUCT
DISPOSAL/UTILIZATION PANEL
FLUE GAS DESULFURIZATION BY-PRODUCT DISPOSAL/UTILIZATION
- REVIEW AND STATUS
H.W. Elder, Tennessee Valley Authority,
Muscle Shoals, Alabama 851
vii
-------
TITLE PAGE
LIME/LIMESTONE SLUDGE DISPOSAL - TRENDS IN THE
UTILITY INDUSTRY
C.N. Ifeadi and H.S. Rosenberg,
Battelle, Columbus, Ohio 865
ENVIRONMENTALLY ACCEPTABLE DISPOSAL OF FLUE GAS
DESULFURIZATION SLUDGES: THE EPA RESEARCH AND
DEVELOPMENT PROGRAM
J.W. Jones, Environmental Protection Agency,
Research Triangle Park, North Carolina 887
FGD SLUDGE FIXATION AND DISPOSAL
W.H. Lord, Dravo Corporation,
Pittsburgh, Pennsylvania 929
UTILIZING AND DISPOSING OF SULFUR PRODUCTS
FROM THE GAS DESULFURIZATION PROCESSES IN JAPAN
J. Ando, Chuo University,
Tokyo, Japan 955
TVA-EPA STUDY OF THE MARKETABILITY OF ABATEMENT
SULFUR PRODUCTS
J.I. Bucy and P.A. Corrigan,
Tennessee Valley Authority,
Muscle Shoals, Alabama 969
THE PRODUCTION AND MARKETING OF SULFURIC ACID FROM
THE MAGNESIUM OXIDE FLUE GAS DESULFURIZATION PROCESS
I.S. Zonis, F. Olmsted, K.A. Hoist, D.M. Cunningham,
Essex Chemical Corporation, Clifton, New Jersey 1003
SECOND GENERATION PROCESSES SESSION
SECOND GENERATION PROCESSES FOR FLUE GAS
DESULFURIZATION - INTRODUCTION AND OVERVIEW
A.V. Slack, SAS Corporation, Sheffield, Alabama 1029
PILOT PLANT TESTING OF THE CITRATE PROCESS FOR S02
EMISSION CONTROL
W.A. McKinney, W.I. Nissen, D.A. Elkins and
J.B. Rosenbaum, Bureau of Mines,
Salt Lake City, Utah 1049
TVA-EPA PILOT-PLANT STUDY OF THE AMMONIA ABSORPTION -
AMMONIUM BISULFATE REGENERATION PROCESS
C.E. Breed, Tennessee Valley Authority
Muscle Shoals, Alabama
G.A. Hollinden, Tennessee Valley Authority,
Chattanooga, Tennessee 1069
viii
-------
TITLE PAGE
DESCRIPTION AND OPERATION OF THE STONE & WEBSTER/IONICS
S02 REMOVAL AND RECOVERY PILOT PLANT AT THE WISCONSIN
ELECTRIC POWER COMPANY VALLEY STATION IN MILWAUKEE
K.A. Meliere and R.J. Gartside, Stone & Webster
Engineering Corporation, Boston, Massachusetts
W.A. McRae and T.F. Seamans, Ionics, Inc.,
Waltham, Massachusetts 1109
CALSOX SYSTEM DEVELOPMENT PROGRAM PRESENTED AT
THE EPA FLUE GAS DESULFURIZATION SYMPOSIUM
R.E. Barnard and R.K. Teague, Monsanto
Enviro-Chem Systems, Inc., St. Louis, Missouri
G.C. Vansickle, Indianapolis Power & Light Company,
Indianapolis, Indiana 1127
WESTVACO ACTIVATED CARBON PROCESS FOR SOX
RECOVERY AS ELEMENTAL SULFUR
F.J. Ball, G.N. Brown, A.J. Repik, S.L. Torrence,
Westvaco Corporation, North Charleston,
South Carolina 1151
IX
-------
NEW ENGLAND S02 CONTROL PROJECT
FINAL RESULTS
BY
George R. Koehler
Edward J. Dober
Chemical Construction Corporation
Air Pollution Control Company
1 Penn Plaza
New York, New York
Prepared For Presentation At
EPA Flue Gas Desulfurization Symposium
Atlanta, Georgia
November U - 7, 1974
Regenerable Process Session
671
-------
NOTE
This project has been funded in part with Federal Funds
from the Environmental Protection Agency under Contract
No. CPA 70-11>4. The content of this publication does not
necessarily reflect the views or policies of the
Environmental Protection Agency, nor does mention of
trade names, commercial products, or organizations imply
endorsement from the United States Government.
672
-------
NEW ENGLAND S02 CONTROL PROJECT
FINAL KESULTS
George R. Koehler*
Edward J. Dober
INTRODUCTION
Operational testing of the Magnesia Slurry System for S0? control
has been completed in a full size, prototype plant employing a
single stage venturi absorber for S02 and particulate removal.
This Chemico-Basic System treated all the flue gas from an oil
fired boiler powering a 150 MW generator at Boston Edison Company's
Mystic Station.
The program spanned a 21 month period for the construction of the
plants and an additional 27 month period of operation in which
over 4,000 hours of running time on the absorption system were
logged. During the operating period, it was demonstrated that:
1) The process could remove SO,, and particulate from
a high sulfur fuel oil flue gas.
2) The basic concepts of the system were sound.
3) Process guarantees of 90% removal of the inlet S0?
could be met, and particulate removal could be
achieved.
4) Magnesia could be regenerated and recycled. During
the course of the program over 6,800 tons of MgSO,
•3
was processed, and more than 3,000 tons of magnesia
regenerated and recycled.
5) 98% sulfuric acid of high quality could be recovered
from the S0~ removed in the pollution abatement
plant. Over 5,000 tons of 98% sulfuric acid was
produced from MgSO. and marketed in the conventional
manner.
" To whom correspondence should be addressed.
673
-------
INTRODUCTION continued
In addition, operation of the full size plant yielded valuable
information for the design of future plants:
1) A correlation was developed allowing prediction and
process control of S0« removal for other design
requirements.
2) A correlation was developed allowing prediction and
process control of the regeneration plant to produce
magnesia suitable for recycle.
3) Studies of the complex physio-chemistry of the process
were undertaken yielding data, not previously reported on:
a) The kinetics of the MgS03 hydrate formation
and transformation.
b) The activation phenomena of regenerated
MgO slaking.
c) The formation and influence of MgSO^ in
the system.
4) Measurements of the physical properties of the chemical
components of the system were made.
5) New analytical methods and techniques were developed to
assure process and quality control at the plants.
6) Observation and testing of the suitability of the
equipment and materials of construction yielded new
recommendations for the design of improved plants.
Finally, the accumulation of operating experience and a large
data bank for reference allowed initiation of the second phase
of the overall application of the magnesia slurry system for
SO control at a coal fired generating station.
674
-------
REVIEW OF INSTALLED FACILITIES
A Chemico-Basic magnesia S0? absorption system (Fig. 1) was installed
on a fuel oil fired boiler at Boston Edison Company's Mystic Station.
The absorber and related equipment was "retrofitted" to the 156 MW
unit No. 6 of the station, which has a Combustion Engineering controlled
circulation;tangentially fired boiler. This boiler is rated to produce
1,000,000#/hr. of steam at 1800 psig, and 1000°F with reheat to
1000°F firing 97CO GPH of #6 fuel oil.
The equipment "retrofitted" on the Mystic #6 boiler was:
1) Two new 800 H.P. booster fans.
2) Single stage venturi absorber and slurry recycle pumps
and lines.
3) Centrifuge system including centrate tank.
4) Dryer.
5) MgO storage silo (an existing ash silo was modified
for MgSO, storage).
6) MgO make-up system, including weigh feeder, slurry
tank, and pumps.
7) Interconnecting materials handling equipment.
The venturi absorber is a co-current gas contactor, 31' in diameter
by 50 ft. overall height with a double annular throat. The absorber
was designed for optimum dispersed absorbing surface at a gas flow
of 425,000 A.C.F.M. In actual operation, however, the absorber
had to treat as much as 650,000 A.C.F.M. because of a high leakage
in the existing boiler ductwork.
The regeneration section of the Chemico-Basic process was added to
a sulfuric acid plant with a nominal output of 50 TPD of 98%
sulfuric acid. The Essex Chemical acid plant is located in
Rumford, R.I., and had been supplying sulfuric acid for industrial
uses since 1928 when it was built by Chemico. Some modifications
to the acid plant were necessary to allow it to take SO- produced
during the regeneration of magnesia as its feed.
675
-------
REVIEW OF INSTALLED FACILITIES continued
The regeneration facilities are comprised of feed MgSOQ and product
O
MgO storage silos (again, an existing soda ash silo was used for
bulk MgS03 storage), coke storage, metering equipment for coke
and MgSO- feed, a direct fired, refractory lined, rotary calciner
7 ' 6" ID x 120' long, and interconnecting materials handling equipment. A gas
cleaning and conditioning system consisting of a hot cyclone,
followed by a venturi scrubber for particulate removal and a
separator - cooler and related heat exchange equipment to reduce
the gas temperature to 100°F, is also included in the system.
Acid plant modifications were limited to: a new main blower to
provide both suction for the regeneration plant and pressure for
the acid plant, a "cold heat exchanger", and small gas stripping
towers.
Recently, operation of a similar S0« and particulate control system
has been started at Potomac Electric Power Company's Dickerson
Station. Again, a Chemico Basic magnesia system was retrofitted on
their unit No. 3. This is a coal fired boiler, and the S0?
absorption system was installed so that it could treat the flue gas
taken off either before or after the electrostatic precipitators.
The system installed on PEPCO's unit No. 3 is similar to the Boston
design, differing in these important areas:
Flue gas is first treated for particulate removal in a 1st stage
venturi scrubber. This unit, while it has been installed in the
same shell as the venturi absorber is a separate fly ash removal
system with its own thickener, slurry pu.mps and piping for fly
ash slurry handling.
676
-------
REVIEW OF INSTALLED FACILITIES continued
After particulate removal, the gas passes to the S0» removal
stage. A single 3500 HP fan, designed for 290,000 ACFM, is
placed after the absorber and provides the pressure drop for
the system. This is a "wet" fan design, and is followed by a
mist eliminator vessel before the gas returns to the stack.
The same regeneration plant, at Rumford, R.I., is used for this
installation.
677
-------
PROGRESS
(The progress of the work which was accomplished to mid 1973 has
been reported in several previous papers (1~"6). The object of
this paper is to describe the work undertaken during the last
year of the project, to establish the significant findings and to
describe the correlations which have been developed from the data) .
By June 1973, most system modifications had been made. One remaining
problem area, the formation of hard deposits in the dryer conveyor
where the cyclone underflow was reintroduced, was eliminated by
installing a pneumatic conveying system to take the dry dust directly
to the product silo. The spill back hopper was also serviced by this
system eliminating an MgO loss point. The series of improvements
made to the dryer converted it to a partial granulator which allowed
it to handle either the magnesium sulfite hexa or trihydrate.
One further attempt at improvement was made when one of the four
dryer cyclones in the multiple array was blocked off to increase the
unit's efficiency at the Boston Edison Plant.
The result was so successful that the increase in collected dust
overloaded the conveying system. The baffle was removed and the unit
returned to its original condition.
One other modification was also made at Rumford in August 1973. The
MgSO elevator was converted from a centrifugal to a continuous
O
discharge type. This immediately reduced the "recycling" that had
caused elevator "trip outs" and plant shut downs in previous attempts
to operate with the dusty feed sometimes introduced.
678
-------
PROGRESS continued
Attempts to achieve longer operational periods during the remainder
of the year were frustrated by a series of mechanical, material
handling problems interspersed with shut downs due to boiler tube
leaks. However, the first of several performance series runs was
made. All of the tests confirmed the results of the
previously reported continuous monitoring in that 90% S0~ removal
was attained by the system. The first tests in October also showed
that the system was being called on to treat a gas flow 50% in
excess of design.
Corrosion - erosion of the unprotected steel piping, valves, and
pumps also caused more delays and outages were taken in August and
November for repair work. Reliability was not achieved until
extensive replacement of valves and pumps was undertaken in December.
Also, centrifuge inspections in August, September, and February
resulted in additional delays of almost a month, as new hard facings
were applied to determine which of them resulted in longer service life
In all, for the period June 1973 thru February 1974, 1630 additional
hours of absorber running time were logged with 700 of those hours
in the June - July period.
By mid-February, it appeared that the magnesia system and operational
staff were in their best state of readiness. Preparations were made
to run performance tests and also to attempt sustained operations
at the design gas flow. Permission was obtained to run the system
with a partial by-pass of the excess flow for most of the following
test period.
679
-------
PROGRESS continued
Fig. 2 shows, in graphical form, the operations during 1971,
including the four months of Operations Testing. During the
period from March 8th to the end of the program on June 26th,
1350 hours of additional running time were logged, despite the
large number of interruptions caused by boiler tube failures.
These numerous interruptions did demonstrate, unequivocally, that
the absorption system could cycle with an operating boiler. In
the 35 day period from 4/12 thru 5/10, 623 hours were logged at
essentially 100% availability, despite nine interruptions caused
by boiler related problems.
Performance Test Results
A condensed summary of the performance test results obtained in
February 1974 are given in Tables 1 and 2.
680
-------
S02 REMOVAL
The absorption of gas in a venturi device can be conveniently
described in terms of conventional mass transfer principles. In
the venturi absorber used for this process, flue gas containing
S0« enters the converging section of the vessel and is accelerated
towards the throat area, passing over surfaces which are irrigated
by the absorbant slurry.
The moving gas creates a wave motion on the liquid surface until, at
a critical velocity, the energy resulting from the frequency and
amplitude of the waves exceeds the cohesive (surface tension, etc.)
forces of the liquid. When this occurs, some portion of the wave is
detached and dispersed into the gas stream.
The liquid, dispersed as droplets into the gas stream, provides the
media for absorption. The equation for point efficiency,
SO
2f
(1 - E) =
- Pe
' Pe
exp
-Kg a Z
(1)
681
-------
S02 REMOVAL continued
( E = Efficiency
( P = Partial Pressure of the Absorbed Gas
where ( Kg = Overall Mass Transfer Coefficient
( a = Area ft2
( Z = Axial distance ft.
( G = Mass Velocity Ib mole/hr. ft.
is useful in deriving relations for removal efficiency.
The surface area can be determined from the liquid to
gas ratio in the absorber and the mean liquid drop size
S = 1.83 x 106 vi
Do Vg (2)
S = Specific Surface Area ft.2/ ft.3
VI = Volume of liquid ft .
Vg = Volume of gas ft.
Do = Mean drop diameter, ML.
An estimate of Do can be obtained from the Nukiyama and
Tanasawa equation (7)
DO =
<3)
Ug = Velocity of the gas, meters/sec.
•f - liquid density, gm/cc
-*< = liquid viscosity, poise
or- = liquid surface tension, dynes/cm
For the case of absorption with gas phase resistance
controlling, it is possible to use an analogous form of
682
-------
S02 REMOVAL continued
equation 1 to correlate the data,
1 - E = exp
-A
A.P = differential pressure (in. HO)
A 6 B are constants
For a particular installation (8 ) this has been used for a
successful correlation as the vessel geometry and most operating
conditions are set in the real system and dispersed area and gas
flow are functions of /\P.
For cases other than gas film controlling, an over-all mass
transfer coefficient is defined as:
Kga kga kla ksa kra
(5)
where the subscripts refer to the gas film,liquid film, solubility,
and reaction resistances respectively.
The reactions involved are acid-base (fast) reactions, therefore,
the final term in 5 can be considered zero. Recent investigations
( 9 ) have determined that the solids dissolution resistance is also
zero.
In order to incorporate the contribution of liquid film resistance,
we have correlated the data using the form
1 - E = exp
-A A P
(10^
B
(S02)
(6)
683
-------
S02_REMOVAL continued
This correlation is shown graphically in Fig. 3 , which presents
the results obtained using the final form of the prediction equation.
F = 1 - exp (2.666( A P)'1'01U(S02I) ~3'?5 + '271 ln S°2I
1
(10)6 -'°31 #) -3
F = Fraction SO- Removal
£\ P = Pressure Drop, inches H20
S02I = Inlet Concentration S02 (PPM)
This prediction equation has been developed from both the data
obtained during the operations at Mystic Station in Boston, and
from some subsequent additional data obtained during initial
operations at an installation on PEPCO's coal fired boiler. This
latter operation provided data for extension of the equation to
higher pressure drops.
An estimate of the accuracy of the correlation for predicting S0?
removal efficiency is given in the following table:
No. Of Standard
Source of Data Points Deviation
Boston Edison 600 H.O %
PEPCO 5 2.9 %
Boston Test Runs
Used to Check
Correlation 5 3.1 %
This correlation also illustrates the relatively small change in
SO- removal efficiency (Fig. 4 ) experienced over wide turn down
ratios of the plant in the normal operating range of the venturi
absorber. An explanation of this potential to maintain high
684
m
-------
S0? REMOVAL continued
efficiency over the ranges of power output of a cycling generating
station, is the relative invariance of the surface area available
for mass transfer over that range. Surface area calculations based
on uniform drop size, and a constant liquid rate, show only a 33%
reduction in surface area for a 4 to 1 turn down equivalent to
operations between HO MW and 150 MW.
685
-------
CENTRIFUGE OPERATION
The centrifuge has operated satisfactorily for the greater part
of the program. Problem areas have been wear of some parts of
the centrifuge and four isolated periods when centrifuging did
not separate solids at a rate sufficient for solids level control
in the recycle slurry. It was demonstrated that proper selection
of materials of construction would reduce the first problem to an
acceptable level and that the second was related to a combination
of chemical concentrations in the slurry which could be controlled.
Wear in the centrifuge was limited to the bowl plows, head plows,
and the conveyor. No wear was observed in the centrifuge bowl itself.
The conveyor, was removed on three occasions to build up and re-hardface
the flights. The plows and wear beads were also replaced. In the
final repair, a stellite hard facing was applied to the flights..
Later examination of the machine and comparison of wear rates after
the stellite was applied indicated that this resulted in increasing
the operating duration to one year before repair of the conveyor is
necessary.
Wash-out connections were also added to the case and an internal
wash pipe was installed to insure that solids deposits did not
build up inside the unit.
Fig. 5 , shows the monthly average centrifuge performance from
January thru June 1974 and is typical of the results obtained
previously. In general, the centrifuge removes HO to 60% of the
solids in the stream fed to it. This level of separation has been
satisfactory for control of recycle solids concentration in the
desired 8 to 10% range.
686
-------
CENTRIFUGE OPERATION continued
The low level of separation experienced in January had also occurred
on three previous occasions. At these times the solids level in the
recycle slurry rose uncontrollably forcing a system shut down. At
first this was thought to have resulted from wear in the centrifuge
(two of the centrifuge inspections were made as a result of these
problems). Analyses of the data file revealed, however, a
relationship to other chemical species in the system and this is
shown in Fig. 6 .
While the factors affecting centrifuge performance are complex, it
appears that the combination of increasing magnesium sulfate
concentration and excess magnesia in the slurry result in an increase
in viscosity coupled with hindered settling, which makes the separation
more difficult.
Both of these concentrations can be controlled in the system.
Actually, high magnesium sulfate levels had been induced during the
problem periods by a practice of fuel switching to low sulfur fuel
when a maintenance problem arose. This change in the S0?/0? ratio
resulted in a higher oxidation rate, more MgSO. in solution, and a
worsening of the problem. This was easily remedied by eliminating
the practice.
687
-------
REGENERATION OF MAGNESIA
The principle operation in the regeneration plant is the calcination
of the dryer product from the absorption system. This can be most
simply described as a thermal decomposition:
MgS03 *- MgO + S02
heat
The rate of the decomposition is temperature dependent, and some
decomposition has been observed at temperatures as low as 300°C.
For the production of useful product in this program, the rotary
calciner has been operated at mid kiln temperatures above 1000°F.
In addition to the simple decomposition of MgSO., the calcination
step also serves the purpose of reducing the MgSCK , a side product
in the absorption reaction, which is also present in the feed.
This reaction can be represented by either of the following
equations:
2MgS04 + C >- 2MgO + 2S02
F1100°F = ~31 KCal
MgSO., + C ^ MgO + SO, + CO
F1100°F = -14 KCal
688
-------
REGENERATION OF MAGNESIA continued
Other reactions can also take place in the kiln environment
1/2 MgS
J. / L i -gu.
MgSO^
2 CO +
1/1 /"\
1/2 02
>UH ^-
V.
so2 — •>-
+ f~l f\ \
SO- ^>-
2
Mg02
2 CO
O rs
so3
S°
All of these reactions are both time and temperature dependent
with the calciner equivalent to a multi-zone linear reactor. In
the course of the program, in addition to the production of
thousands of tons of acceptable regenerated MgO, upset conditions
have resulted in the formation of high concentrations of S0_ and
O
at other times in the formation of elemental sulfur in the
calcination step.
Attempts have been made to study these upset conditions in the
laboratory, and thermal decomposition studies have been carried
out there using infrared spectroscopic techniques to investigate
the phenomena of sulfur formation by detection of its precursors
appearance. These studies have not as yet yielded any specific
information on the formation of elemental sulfur. However, they
have yielded information indicating that the decomposition process
for MgSO- obtained from the trihydrate form and from the hexahydrate
form may be different.
Retention time in the calciner can be determined from the following
equation:
Q = 0.19 L
N D S (8)
where -8 = time (min.)
L = kiln length (ft.)
N = Rotational Speed (RPM)
D = Dryer diameter
S = Slope (ft./ft.)
689
-------
REGENERATION OF MAGNESIA continued
and for most operations undertaken during this project this time
was approximately 1 hour. Because of the use of tube coolers on
the calciner, an additional 1 hour was required before final
discharge to the product conveyors.
An additional factor in the calcining operation is the formation
of periclase, an unreactive form of MgO. This is generally
accompanied by an increase in the density of the magnesia, and the
following table shows the results of experiments to determine the
effect of calcining temperature on specific gravity.
T°C
600
710
850
1000
1200
moo
T°F
1112
1310
1562
1832
2192
2552
Specific
Gravity
2.94
3.04
3.22
3.39
3. 48
3.52
The data resulting from the test program phase of this project was
analyzed by regression methods to provide correlations for both the
percent of MgSO in the product and the bulk density of the product
These are given in the following equations and provide a means of
determining the processing condition necessary for the production
of an active magnesia. The effect on bulk density is shown
graphically in Fig.7 .
690
-------
CALCINER OPERATIONS
CORRELATION FOR PREDICTION AND CONTROL
OF BULK DENSITY OF REGENERATED MAGNESIA
BD= 169.2+ (T'*A) + (C*BH 0.741 (% MgSO in Feed) - 0.744(% MgSO, in
4 d Feed)
- 4.9 (% 0, in Acid Gas) -0.4 (Feed Rate, LB./MIN.) - 104 (Furnace
^ Draft)
(9 )
"l/o
Where T' = (1700 - °F Mid Kiln Temperature)
A = -0.891 + 0.166 (C)
B =-28.4 + 3.44 (C) + 71.2 (Furnace Draft) + 1.95 (% 0, in Acid
* Gas)
- 1.24 (% MgSOH in Feed)
C = % Carbon in Feed
BD = Product Bulk Density, LB./FT.3 (Operation @ 1.82 RPM
Add 5 to BD for 1.56 RPM)
Statistics: Standard Deviation = 7.2
Multiple Correlation Coefficient = 0.79
Confidence Level of F Ratio = 99.9% +
No. of Data Points = 456
691
-------
CALCINER OPERATIONS
CORRELATION FOR PREDICTION AND CONTROL
OF % MgSOH IN REGENERATED MAGNESIA
% MgSOu = -90.2 + (T'*A)+(C*B) -2.H6 (%MgSO in feed)+0.989(% MgSO in
(% 02 in Acid Gas) +0.28 (Feed Rate, Lb./Min.)
(10)
Where T'
A
B
(1700 - °F Mid Kiln Temperature )
-0.870 + 0.185 (% MgSO^ in Feed)
-1/7
x/
= 15.7 - 23.
(Furnace Draft) - 2.19 (% 0 in Acid Gas)
-0.61 (% MgSO^ in Feed)
C = % Carbon in Feed
Statistics:
Standard Deviation = 5.U
Multiple Correlation Coefficient = 0.77
Confidence Level of F Ratio = 99.9% +
No. of Data Points = 207
692
-------
MgO LOSS AND REGENERATION CYCLES
A) Loss
In early operation of the system, losses of MgO were high.
These losses resulted from the necessity for frequent clean
outs, discarding both oversize material and any spilled solids ,
as well as loss to the stack, overflows, and absorber draining
during the frequent shutdowns .
The installation, by June 1973, of the various lump crushers,
as well as rerouting the dryer off gas to the absorber reduced
this loss substantially in the following period but did not
eliminate it. Spills were still discarded and the vessel was
still drained on shut down.
Another contributing factor was the higher than design gas flow,
which may have resulted in entrainment losses.
In the final period of operation, controls were imposed to
eliminate many of the previously occurring gross losses, and
during part of this period a careful measurement of system
losses were made in order to identify these sources for future
design improvements.
This loss history is shown graphically in Fig. 8, which accounts
for the total system MgO loss by operating periods. The first,
of about 1127 hours duration accounting for the greatest loss of
606 tons, corresponds to the pre-start, start-up, and shake-down
period prior to June 1973.
The second period for test and development had a loss of 338 tons
of MgO in 1630 hours of operation, and the last period for
operational testing showed a loss of 151 tons in 1350 hours of
operation.
693
-------
MgO LOSS AND REGENERATION CYCLES continued
A) Loss continued.
In the final period starting in March 1974, gas flow to the
absorber was maintained at the design rate. In addition,
the system was run as a closed loop with no vessel drainage,
returning the small spills back to the process. Measurements
were made at the following m loss points at Boston for a
continuous operations period and the silos were emptied and
the contents weighed before and after the run.
Potential Sources of MgQ Loss at Mystic Station
1) Stack
2) Centrifuge Washing
3) Centrifuge Case Leaks
4) Pump Packing Gland Leaks
5) Absorber Overflow
6) MgO Slurry Tank Blow-Down
7) MgO Slurry Tank Overflow
8) Centrate Tank Overflow
9) Solids Loss at Dryer Feed End
10) Dust Loss at Dryer I.D. Fan
11) Dust Loss at Expansion Joints
12) Spillage at MgO Feeder
13) Spillage in MgSO,, Belt Gallery
14) Spillage at TrucR Loading Point
The tests, conducted over 13 days, in which 336,470 Ibs. of
regenerated material were fed (an additional 2,504 Ibs of
MgO added with the fuel oil was also accounted for) , showed
a loss of 0.37 tons/operating day at the absorber system
distributed as follows:
Loss to Stack 0.13 Ton/Day
Absorber Overflow 0.14 "
Misc. Measured Loss 0.07 "
Unmeasured Loss (By
Difference) 0.03 "
0.37 "
694
-------
MgO LO_SS_AND REGENERATION_CYCLES continued
A) Loss continued
With an average MgO consumption of 10.61 tons/day during the
period, this amounts to a 3.5% loss as compared to a 5% design.
The greatest losses were found to occur at the regeneration
plant. Here 1.5 tons/day of equivalent MgO is lost from the
neutralizer system overflow and another 0.5 tons per day is
scalped off, for future reclamation, before being pulverized.
Both of these losses would be virtually eliminated in a full
size regeneration plant.
B) Regeneration Cycles
Fig. 8 can also be used to evaluate the number of regenerations
of magnesia in the various operational periods.
In the start-up and shake-down period, losses were 606 tons of
MgO, with 559 tons of regenerated MgO recycled to the Mystic
#6 System. This rate limited the number of recycles from 2 to
3 before the material was lost from the system.
During the test and development period, 1717 tons of regenerated
MgO were returned to the system, while 338 tons of MgO equivalent
were lost. This corresponds to 5.1 cycles for the magnesia
before it is lost from the system.
During the final period, 151 tons of magnesia was lost, while
875 tons of regenerated alkali was shipped back to Boston. This
would correspond to 5.8 cycles; however, during this same period
there was an inventory build-up with 80 tons of equivalent MgO
695
-------
MgO LOSS AND REGENERATION CYCLES continued
B) Regeneration Cycles
inventory at the beginning, and 214 tons remaining at the close
of the program. The actual number of cycles reached a maximum
of 5 at the beginning of May.
The negligible effect of regeneration of magnesia on the process
is also seen in the monthly average analysis of excess MgO in
the centrifuge cake, which only increased from 1.7% in March to
3.6% in April.
696
-------
CONCLUSIONS
Problems encountered by previous workers attempting to use magnesia
as the chemical agent for flue gas desulfurization have been
overcome in the Chemico-Basic Process. Operation of the prototype
plant over the two year period covered by this series of reports
has demonstrated that:
1) Slurry solids separation has been maintained. Use
of a centrifugal separator and control of the
chemical components in the slurry, which influence
the seperational methods, both contribute to the
solution of this problem area.
2) Magnesium sulfate concentration can be controlled
and the magnesium sulfate formed can be regenerated.
Controlled addition of carbon to alter the calciner
atmosphere and modify the reducing conditions has
proven successful in eliminating the problem of
eventual complete conversion of the alkali to MgSO .
3) Plugging and scaling of the absorption equipment
has not been a problem. High circulation rates,
controlled slurry composition, high crystallization
nucleii concentration, and limited residence time
are all contributers to the solution of this problem
area.
Several new problems were encountered and overcome during
operation of the prototype system, and their description has been
reported in this paper and references 5, 6. It now would appear
that incorporation of the developed modifications into the next
plants, coupled with a mechanical redesign to incorporate simpler
and more robust material handling equipment and the use of rubber
lined slurry pipe, pumps, and valves, would result in a reliable
system for flue gas desulfurization.
697
-------
Bibliography
1. "Magnesium Base SC>2 Recovery Process, A Prototype Installation",
I. S. Shah, C. P. Quigley, 70th AIChE National Meeting,
August 1971.
2. "Magnesium Base SC>2 Recovery Scrubbing Systems", P. M. Wechselblatt,
R. H. Quig, 71st AIChE National Meeting, February 1972.
3. "The Magnesia Slurry SC>2 Recovery Process, Operating Experience
With A Large Prototype System", M. A. Maxwell, G. R. Koehler,
65th AIChE Annual Meeting, November 1972.
4. "Progress Report - Magnesium Oxide System At Boston Edison
Company's Mystic Station", C. P. Quigley, Electrical Worlds
Technical Conference, Chicago, October 1972.
5. "Operational Performance of the Chemico Basic Magnesium Oxide
System at the Boston Edison Company", Part I - G. R. Koehler,
Part II - C. P. Quigley, Flue Gas Desulfurization Symposium,
New Orleans, May 1973. (EPA-650/2-7 3-038, December 1973).
6. "New England S02 Recovery Project - System Performance",
G. R. Koehler, 66th AIChE Annual Meeting, Philadelphia,
November 1973.
7. S. Nukiyama, Y. Tanasawa, Trans. Soc. Mech, Eng. (Japan) 4,
No. 14, 86 (1938).
8. "E.P.A. Alkali Scrubbing Test Facility: Sodium Carbonate
and Limestone Tests Results", M. Epstein, EPA-65072-73-013,
August 1973.
9. "Sulfur Dioxide Removal in Venturi Scrubbers", C. P. Kerr,
Ind. Eng. Chem., Process Design, 13, No. 3, 222 (1974).
698
-------
TABLE 1
S02 REMOVAL - TEST RESULTS
BOILER
TEST LOAD: INLET GAS S02 IN S02 OUT % S02 SOg OUT
NO. MW RATE: ACFM PPM - VOL. PPM - VOL. REMOVAL LB/106 BTU
1
146 446,953
926.1
71.1
92.3
0.125
486,991
1004 .5
89.0
91.1
0.199
151 658,207
983.9
63.3
93.6
0.201
148 503,233
833.3
86.6
89.6
0.243
699
-------
TABLE 2
PARTICULATE REMOVAL - TEST RESULTS
BOILER
TEST LOAD: INLET GAS PARTICULATES: LB/HR % PARTICULATES
NO. MW RATE: ACFM TT7 OUT REMOVAL LB/106 BTU
1 146 446,953 380 116 69.5 0.072
2 144 486,991 232 115 50.4 0.084
151 658,207 399 150 62.4 0.111
4 148 503,233 151 82 45.7 0.068
700
-------
MgO ADDITIVE
SCRUBBER SYSTEM FOR S02 RECOVERY
OIL FIRED BOILER
SCHEMATIC PROCESS FLOW SHEET
M«O FROM ACID PLANT
REGENERATION SYSTEM
MgO RECYCLE PROCESS, FOR PRODUCTION OF 987. SUIFURIC ACID
SCHEMATIC PROCESS FLOW SHEET
M,SO, TO ACID PIAN1
SO2 GAS CLEANING
CONVENOR
CONCENTRATED S02 GAS
TO SULFURIC ACID PLANT
REGENERATED
MgO
SILO
POWER PLANTS
701
-------
Outage Code
o
NJ
0
1/7-
pj
0
o_
rv
Q
cr
5°
— 'i/i_
T— *
2
5T
Sg
'
f 3
3 n
•9
•*
J J
t
V-j
t
i.
S 33
3
3 .
* 3
.?.* n
^
3 3 a) Operating Period of 6 hrs. or less
fa a £ b) Interruptions of 3 hrs. or less
~i
3
*3
3
8
10
BOSTON ED. CPERRTI0NS JRN- -JUNE 1974
Fig. 2
-------
D
O.
O.
CO
u_
u_
UJ
_J
CEO.
>^
CD
z:
UJ
cc
CO
O
U)
DP*12 IN.
DP-6 IN.
DP-H IN.
DP-3 IN.
OP-2 IN.
ioo 1200 \oo ieoo
INLET S02 - PPM
1800
2000
SG2 REMOVRL....EFFICIENCY
EFFECT OF QELTR P flND INLET S02 CONG
Fig. 3
-------
O-i
O)
D.
JN; to
u_
O-
cr
CD
2:
UJ
C
L
OJ
CO
•o.
to
o.
l/J
INLET 502=1000 PPM
INLET S02=700 PPM
INLET S02=UOO PPM
8 10 12
OELTfl P - 1N.H20
502 REMOVRL EFFICIENCY
EFFECT OF DELTfl P RND INLET 502 CONG
16
is
Fig. U
-------
CENTRIFUGE
SOUD3 BEMOVAL EFFICIENCY
3H20 Case
FEB.
MAD.
APR.
MAY
JUNE 1974
I ig . 5
705
-------
o
o
to.
O
O
to.
coo
o
CO
UJo
—I
0
o
UJ
cc
o
o
o
o
MgS03 . 3H20 Case
*5.SO 6.00 6.50 7.00 7.SO 8.00 8.SO 9-00 9-50 10-00
EFFECT OF SYSTEM PH RND MGS04 LEVEL
ON RECY. SOLIDS CONTENT -DftTfl TO 3/12/7U
Fig. 6
-------
ID-
o
"vl
Q;
Zi
en
o.
XC-0.5
XC-l.O
~"?00
800
900
1000
UOO
1200
1300
CRLCINER
MIDKILN TEMP
M1DK1LN TEMP-DEG F
moo
1500
RflTIGN - EFFECT
S, 7.C. ON BULK DENSITY
OF
1600
Fig. 7
1700
-------
o
O
I\J.
O
O
O,
o
00
CO
_i
D
I O.
U3
CO
CO
o
og.
o
o_
t\J
PRE-STARTUP
AND
SHAKEDOWN
AVERAGE LOSS PER PERIOD
TEST
AND
DEVELOPMENT
/ / /
/ /
OPERATING
H/72
6/73
3/7U
6/7H
MYSTIC STATION
REGENERATION
t> 500 1000 1500 2000 2500 3000
OPERRTING HOURS
PROCESS MGO CONSUMPTION
BY OPERRTING PERIODS
3500
4000
4500
Fig. 8
-------
ASSESSMENT OF PROTOTYPE OPERATION
AND FUTURE EXPANSION STUDY - MAGNESIA SCRUBBING
MYSTIC GENERATING STATION
BY
Christopher P. Quigley
James A. Burns
Boston Edison Company
Boston, Massachusetts
Prepared For Presentation At
EPA Flue Gas Desulfurization Symposium
Atlanta, Georgia
November 4 - 1, 1974
Regenerable Process Session 709
-------
Acknowledgment
Funding for this project was provided in part by the Environmental
Protection Agency under Contract No. CPA 70-114. In addition,
financial support was provided to Boston Edison Company by New
England Gas and Electric Association and Eastern Utilities Associates,
710
-------
Kt^'r_ OF 1 > HOT UT YPK OPE HAT ION
i-'U'i'ui
-------
Some of the early operational problems requiring additional man-
power included inoperable flue gas dampers, inability to keep rotary
dryer burner operating along with cleanup associated with hardened
deposits, dust carryover, and formation of agglomerates in the rotary
dryer, unreliable Ph instrumentation which forced constant manual
sampling and analysis, insufficient capacity of dryer feed screw,
and general cleanup of dust and slurry on equipment and ground.
As equipment modifications and additions were made to improve
scrubber operation a more complex system resulted. Also, with an
increasing knowledge of existing equipment limitations and complex-
ities, a requirement for anticipating and adjusting for system
excursions evolved. Under these circumstances it became necessary
to consider a more technically oriented operator. In late 1973, one
Central Control Operator and one Supervisor on each shift were as-
signed to the project replacing the less experienced auxiliary
operators. An additional CCO and cleanup crews were available on any
shift as required. This reassignment of manpower and the extensive
maintenance program on pumps and centrifuge in December 1973 and
January 1974 was in large measure responsible for the generally more
successful operation experienced through the end of the program.
SCRUBBER IIAIHTEHANCE
Scrubber maintenance throughout most of the test program was exces-
sive both in terms of manpower requirements and spare parts. This
was especially true when considering the very low quantity of scrubber
operating hours attained during this period. This condition existed
until the Spring of 1974 when most of the major problems were resolved,
712
-------
Liu-cause we wore making concentrated efforts to demonstrate contin-
uous scrubber operation we could not in a short test program take
the time to make major changes. This resulted in recurring
breakdowns of equipment which was not really suited for the applica-
tion. While many examples could be cited to illustrate the problems,
three areas will be discussed.
Dryer Feed Screw Conveyor
The dryer feed screw conveyor carrying wet cake from the centrifuge
to the rotary dryer caused repeated scrubber outages. Initially,
dryer dust captured in the mechanical collector on the dryer off-gas
was reinjected into the wet cake just prior to passage into the dryer,
Premature drying resulted in the buildup of hardened deposits in the
conveyor, both inside and outside the rotary dryer. These deposits
caused overloads, damaged gear boxes and motors, and resulted in
erosion and twisting of the conveyor housing.
Changes to the conveyor housing, and transport of the dryer dust to
storage rather than to the wet cake conveyor eliminated some of the
problems but buildups in the conveyor within the rotary dryer re-
mained a problem.
Centrifuge
Centrifuge problems basically involved buildup of hardened air-dried
slurry within the centrifuge. Maintenance work included a range of
tasks from repetitious replacement of shear pins (pins provide
mechanical overload protection to internals in the event of binding),
to the complete dismantling of the centrifuge for cleaning and repair
to worn solids conveyor. Assignment of 3 or 4 maintenance men for
713
-------
dinmanllimj and cleaning, LJhipmont of internal conveyor to manu-
facturers shop Cor rebuilding and hardface replacement, and rein-
stallation of centrifuge was required on several occasions. Each
of these maintenance experiences required approximately 1 week.
Erosion/Corrosion of Pumps, Valves, etc.
The major area of maintenance committment in terms of both manpower
and spare parts was due to erosion/corrosion of slurry pumps, large
isolation valves and expansion joints around recycle pumps, and
various piping. Original design had included cast iron slurry
pumps, carbon steel unlined piping, carbon steel valves with rubber
lined plugs, and rubber expansion joints around the large recycle
pumps. Because of the limited time available in the test program
very few attempts at optimizing equipment could be made. Stainless
steel impellers, for example were installed with little or no im-
provement observed. As the test program progressed we considered
ourselves fortunate to obtain the necessary cast iron parts to keep
pumps running. Erosion/corrosion in the 6-18 inch valves on the
discharge and suction of the main recycle pumps eventually prevented
isolation of the pumps for maintenance. Six new valves were installed
in the same materials, that is, carbon steel bodies with rubber-lined
plugs. New valves with the entire body rubber-lined could not be
made available on time. Several of the large rubber expansion joints
installed on the recycle pumps showed erosion and splitting of the
rubber and were replaced. Maintenance welding of piping especially
at changes of flow direction in small piping was constantly required.
Erosion/corrosion of piping in the tangential spray header piping to
the absorber was a particular problem.
714
-------
From Lilurlup in Api.il i'J72 to February L'JIA the requirement for
maintenance was continuous. Assignment of manpower varied from 1
man for the smallest job to 12 men during the heaviest periods of
maintenance for overhaul. After the major maintenance program on
pumps, valves, centrifuge, etc. was completed in February 1974
maintenance requirements fell off considerably through the end of
the program (June 1974) as would be expected with the increase in
scrubber operating hours in this period.
SCRUBBER AVAILABILITY
From startup in April, 1972 through May, 1973 the scrubber operated
with a very low availability of 17% of Unit j)6 operating hours.
Availability is defined as the hours of scrubber operation divided
by the hours of boiler operation. Through June and July 1973 scrubber
availability improved to 68 and 61% respectively. During the re-
mainder of 1973 availability fell off due to the heavy erosion-
corrosion experienced in pumps and centrifuge. Maintenance work was
completed in February, 1974 with all efforts directed to improving
the status of equipment for the final stages of the 2 year test
program. Scrubber availability improved from March through June
1974 as evidenced by the following data:
Unit #6 Operated Scrubber Operated Availability
June 1973 592 402 68%
July 575 351 61%
August Unit $6 and scrubber out for overhaul
September 637 243 38%
October 627 377 60%
715
-------
Unit _ |G _Ope ra tod S c r ubbor Operated Availability
November 629 162 26%
December 658 86 13%
January 1974 555 152 28%
February 541 138 25%
March 408 353 87%
April 585 471 81%
May 488 280 57%
June 359 288 80%
Total 6654 3303
Scrubber availability from June 1973 through June 1974 was 50%.
The longest continuous periods of scrubber operation occurred for
approximately 7 days on 3 separate occasions. Two of these contin-
uous runs were in April, 1974. From March, 1974 through the end of
the program in June, 1974 the inability to demonstrate longer runs
was due primarily to Unit #6 boiler problems which were not related
to scrubber operation.
COST REVIEW - SCRUBBER PROTOTYPE
The capital investment in scrubber plant at Mystic Station which
includes the centrifuge and drying facilities amounts to $3,635,000.
This figure does not include $1,382,447 for general and administrative
expenses and allowance for funds used during construction.
Scrubber operation and maintenance costs at Mystic Station over the
2 year test program are shown as follows:
716
-------
Operating Labor $390,000
Mechanical & Electrical Maintenance
including Spare Parts 380,000
Extraordinary Cleanup Maintenance 65,000
Instrument Maintenance 32,000
Operating Materials 132,000
Total $999,000
The cost to operate and maintain this prototype system was exces-
sive due to the unusual operational and maintenance problems
experienced. It is not advisable we feel, to apply the above cost
during a 2 year test program to a future design. Experience through
a 2 year program has, however, enabled us to prepare a reasonably
accurate assessment of future capital, operating, and maintenance
cost.
Estimates of these costs for scrubbers installed on Mystic Station
Units with 1050 MW of electrical capacity are presented later in
this paper.
MYSTIC STATION PROTOTYPE SUMMARY
Boston Edison has appraised the scrubber project at Mystic Station
as reasonably successful under the circumstances of an erratic test
program due to very intermittent operation. We prefer, however, to
continue to detail areas of success and failure in order for others
to individually assess the project.
Areas of Success
1. S02 removal efficiency obtained throughout the program was
generally in excess of 80%. Independent performance testing
using EPA test methods during late February and early March 1974
717
-------
showed -SO2 removal efficiencies averaging 1)1.1% with the use
of regenerated McjO.
2. Minor scaling in the absorber was observed but never inter-
fered with operation. No plugging of mist eliminators occurred
at any time. Once past early problems of slaking the regenerated
MgO which caused plugging in MgO slurry lines, no plugging oc-
curred in the pumps and piping systems.
3. Late in the scrubber test program tube failures in Unit #6 boiler
caused several scrubber shutdowns. The scrubber system repeatedly
showed its ability to cycle with the boiler by returning to ser-
vice immediately after boiler repairs.
4. The project resulted in the production and sale of approximately
5000 tons of a commercial grade of sulfuric acid.
5. As a research prototype the major goal was to demonstrate the
chemical reactions on a large scale and uncover and solve tech-
nical problems introduced in the scale up from pilot plant
operation. Although the 2 year test program was conducted under
very erratic conditions it was clear that basic concepts of SC>2
removal, MgO regeneration, and production of sulfuric acid were
feasible.
Areas of Failure
1. Poor availability over a 2 year test program.
2. fkjO losses throughout the system (scrubber and calciner) were
excessive and amounted to approximately 10% at the end of the
program.
718
-------
J. The ability oC the regenerated MgO to bo recycled continuously
while maintaining its capability to react with SO2 in a tight
system with minimal MgO makeup was not demonstrated. In addi-
tion, with high MgO system losses an evaluation of buildup of
ash, vanadium, etc. was inconclusive.
EXPANSION STUDIES
Conccptual Arrangement
Expansion of scrubbing at Mystic Station using the MgO process has
been studied in detail. In the expanded system the flue gases of
Unit Nos. 4, 5, 6 and 7 would be scrubbed and a central calciner
facility would be located on site to process the magnesium salts
produced. The S02 gas would be fed to a sulfuric acid plant on an
"across the fence" basis.
The system was sized to scrub 1050 MW of oil fired boilers burning
a residual fuel of up to 3.5£ sulfur content. Operating costs were
evaluated for the system on the basis of firing 9,738,000 barrels
per year of residual fuel oil with a 2.5% average sulfur content
(plant load factor of 71.2%).
Figure 1 shows a simplified process flow diagram. The scrubbers
and slurry recirculating pump systems would be located in the
vicinity of the individual unit stacks similarly arranged to the
existing Unit ?,6 system. Unit Nos. 4 and 5, each rated at 150 MW,
will be virtual duplicates of the Unit 86 scrubber arrangement.
Unit #7, a new 600 MW unit, would be equipped witli two scrubbers.
The expanded scrubbing system would supply magnesium sulfite slurry
to a central regeneration plant located either on the property or
on property adjacent to the power plant site.
719
-------
Tho rocienc-rntioii plant would conLriTutio, dry, and .store the
rn.iqnc.'jium sulfite received from the boiler plant. From
storage, the sulfitc salts would be processed through a cal-
cincr, driving off the 502 gas to an adjacent acid plant and
regenerating the MgO for reuse in the scrubbers. MgO would
be slurried with water at the regeneration plant and pumped
back to the scrubbers.
Built into the process would be points for buffering against
equipment breakdowns, planned equipment shutdowns and variances
in equipment process loadings. Each scrubber would be equipped
with a by-pass duct to enable the boiler to continue to operate
in the event scrubber shutdown is required. The MgS03 silos
and the MgO silo would be sized to provide buffer storage. This
excess storage would provide for the varying flow of materials to
and from the power plant while supplying an essentially base
loaded acid plant. In addition, this storage permits the
continuation of scrubbing operations during a normal acid plant
annual overhaul. The calciner and acid plant would be sized to
provide "catch-up" calcination ability.
This arrangement differs from the Unit 6 experience in that no
"off-site" transportation is involved. Further, the centrifuging,
drying, dry material handling and MgO slurry preparation steps
are divorced from the power plant area and centralized with the
calciner operation. This would enable us to adequately control
material losses. The handling of dry dusty materials is localized
and minimized. The losses from cleaning dryer and calciner off
gases could be virtually eliminated by using the effluent from
720
-------
.'•.cruhbei:.1; on UH-:;I' oil'
-------
U t i 1 ity Requirements
Utility requirements at the scrubber plant are estimated as follows
Power; 13,800 kW at 4160 kV
230 kW at 480V
Water: 960 gpm
Steam: 500 #/hr. max. at 50 psig
Mother Liquor (From Regeneration Plant): 2350 gpm
MgO Slurry (From Regeneration Plant): 262 gpm
Major Components
The following lists the major components to be provided at the
scrubber plant:
Unit #4 Unit #5
Unit #6
Unit j}7
1
2
2
2
1
2
2
2
1
3
2
(exist . )
(exist. )
2
(exist. )
2
4
2
2
Item
Fixed throat venturi
absorber
Scrubber Recycle Pumps
Sump Pumps
Induced Draft Booster Fans
Instrument & Control
Boards
Scrubber Holding Tank
Instrument Air System
4 kV Switchgcar
480 V Switchgear
300 kVA Transformer
33' x 70' Control Dldg.
REGENERATION PLANT DETAILS
Production Capability
Operating logistics have been estimated based upon power plant
generation of 1050 MW at an average annual plant load factor of
722
111
1 - Common to all units
-I __ II I! II II
1 - '
T II 11 II |1
I _ II II II ft
•1 tl It II II
-------
71.2'i burning a 2.51 sulfur oil with a 90% SC>2 removal efficiency.
On this basis, the plant would produce approximately 72,000 Ton/year
of S02 supplying the feed stock for a 350 Ton per day H2SC>4 plant,
Utility Requirements
Utility requirements at the regeneration plant are estimated as
follows:
Power: 2800 kW at 4160 kV
1360 kW at 480 V
Water: 392 gpm
Steam: 29,765 #/hr. at 50 psig
#6 Fuel Oil: 28,840 gal./day
Coke: 3.3 tons/day
MgO Makeup: 7.7 tons/day
Major Components
Major equipment provided at the regeneration plant consists of:
Item Quantity
Centrifuges 10
Rotary Dryers 2
Dryer Scrubber 1
Calciner 1
Calciner Scrubber 1
Weak Acid Coolers 2
Cooling Tower 1
Mechanical Dust Collectors 12
Miscellaneous Tanks 12
Miscellaneous Conveyors 14
Fans and Blowers 12
Process Pumps 22
723
-------
f^ajor Commoner)ts (continued)
I_te m Q u an tity
Fuel Oil System including storage &
2 pump and htr. sets 1
Instrument Air System 1
Instruments & Control Boards
4 kV Swgr. Units 2
480 V Mtr. Control Centers 2
1000 kVA 4 kV/480 V Transformers 2
Centrifuge Building
(48' x 75' x 55' high) 1
Office/Warehouse Bldg.
{401 x 80' x 12' high) 1
Calciner Dldg.
(35' x 60' x 48' high) 1
MgSO3 Storage Silo
(40' dia. x 130' high - reinf. cone.) 2
MgO Storage Silo
(40* dia. x 130' high - reinf. cone.) 1
Cost Summary
Capital cost for such a conceptual plant is currently estimated at
$51/kW. This cost does not include the cost of a sulfuric acid
plant which would be owned and operated by others. Total capital
and operating costs for the scrubber/regeneration facility based
on the plant loading used in this study are estimated to be
2.44 mils/kWh.
Assessment
The Boston Edison Company believes we have demonstrated the tech-
nical feasibility of magnesia scrubbing. However, we were not able
724
-------
Lo ck.'muM.'jL.LMLi.: i\ J.OIHJ :;u:;La i nod run eommcJKJuriJ Lo with power plant
needs. We arc reasonably certain we can build a second generation
largo scale system based on the knowledge gained /ith the Unit §6
prototype. Some risks would be involved. For example, we believe
the erosion/corrosion problems associated with pumps and piping
would be solved by rubber coating critical components. We have,
however, not had the opportunity to test this modification. The
absorber size for Unit #7 will escalate to 900,000 acfm from the
450,000 acfm design of Unit #6. Although there is some risk in
this extrapolation, it certainly does not match the original risk
in extrapolating from the 1500 acfm pilot to the 450,000 acfm
Unit #6 absorber. Centrifuge operation and wet cake material handling,
although working well in the late stages of the program, were not
demonstrated long enough to develop a high level of confidence that
all major problems with this equipment have been resolved.
We would reiterate that our confidence level in building a second
generation system is high.
Future Plans
The Company has been evaluating the potential expansion of scrubbing
at Mystic Station for some time. There are factors in this evaluation
beyond the strictly technical which must be seriously considered.
The economic attraction for scrubbing is favorable based on the
present price differentials between low and high sulfur fuel oils.
The fuel oil energy picture is so fluid as to make it extremely
difficult to assess the future price differentials. These differ-
entials can be adjusted for reasons other than solely cost and are
therefore not readily predicted.
725
-------
In Uiis period uL dire uhortucjo oi' capital, the Company is looking
very hard at all capital committments. Our ability at this time to
raise the capital for this project in competition with other com-
pany capital requirements is subject to question and concern.
Possibilities of funding this project by leasing and/or by revenue
bond financing are being explored.
Allowing utilities to burn higher sulfur fuels than is now permitted
in Massachusetts is under study. Higher sulfur fuel use will re-
duce the economic incentive to install scrubbers.
A selective burning program using fuel switching based on meteorolog-
ical conditions if allowed, would be considerably less costly than
stack gas SC>2 control. The rising consumer and energy crisis pres-
sures may redirect SC>2 control to this approach.
The President lias recently issued a policy statement that the burning
of fuel oil in utility boilers be greatly reduced. The impact of
this policy on the Mystic Station fuel situation both from a fuel
use and a fuel pricing point of view is uncertain.
We do not rule out but we cannot yet commit to scrubber installations
at Mystic Station until many of these uncertainties are resolved.
726
-------
FIGURE i
Expansion Study - Magnesia Scr
Unit Nos. 4, 5, 6, 7
Mystic Generatir.c: Station
Boston Edisor. "Cc-pany
T~ * X '-> - -» —
-------
MAG-OX SCRUBBING
EXPERIENCE AT THE COAL-FIRED DICKERSON STATION
POTOMAC ELECTRIC POWER COMPANY
Donald A. Erdman
Project Engineer
Potomac Electric Power Company
1900 Pennsylvania Avenue, N.W.
Washington, D.C. 20006
ABSTRACT
The paper presents a description of the Cheraico-Basic Mag-Ox scrub-
bing system installed at the Dickerson Generating Station. Operating
experiences for the first year of operation are reviewed. Projected
costs for future installations are discussed.
729
-------
MAG-OX SCRUBBING
EXPERIENCE AT THE COAL-FIRED DICKERSON STATION
SYSTEM DESCRIPTION
The Potomac Electric Power Company - PEPCO - in late 1970 in con-
tinuation of their search for a stack gas desulfurization system,
selected a Chemico-Basic magnesium oxide process for installation at the
Dickerson Generating Station. The selected system consists of a two-
stage scrubber system sized to process half the flue gas, 295,000 acfm,
from the 190 Mw Unit #3. The system is designed as a parallel system
with its own wet I.D. fan. With the parallel arrangement loss of the
scrubber does not cause any loss of load. The scrubber is capable of
taking flue gas from either ahead of or after the existing electro-
static precipitator.
Flue gas enters the first stage, the particulate removal scrubber,
where the gas is cooled and saturated from 250°F to 120°F. The flue gas
passes through an adjustable throat venturi where the fly ash is removed.
The ash-laden water is recycled through the scrubber with a bleed stream
carrying ash to the thickeners. The thickener underflow is discharged
to a dilution tank where water is added to pump the ash to a settling
pond. The overflow from this pond cascades through four ponds. Water
from the lowest pond is pumped back to the dilution tank. The thickener
overflow is pumped back into the first stage scrubber.
The flue gas leaves the first stage passing through mist elimina-
tors then to the fixed throat venturi where it contacts the recycling
MgO slurry for absorption of the S02- As the MgO absorbs S02 magnesium
sulfite crystals build up in the slurry. A bleed stream is sent to the
centrifuge where the sulfite crystals are stripped and sent to a rotary
oil-fired dryer. The dried MgS03 is sent to a silo for storage and
ultimate shipping to the calciner. The dryer off gas is returned to the
inlet of the second stage scrubber.
The mother liquor from the centrifuge is returned to the scrubber.
A bleed from the mother liquor is used to slake the MgO. The MgO is fed
on demand to the scrubber to maintain a neutral pH.
The flue gas leaves the second stage through mist eliminators to
the wet I.D. fan through the mist eliminator tank to the stack where
730
-------
It is mixed with the unscrubbed flue gas from Unit #3 giving a mixed gas
temperature of about 175°F.
OPERATIONS
The S02 removal system was placed in operation on September 13,
1973. Operation since that time can be divided into four phases as
follows:
Phase I September 13, 1973 to January 14, 1974 -
Initial operation and debugging.
Phase II January 14 to April 15, 1974 - Maintenance
and modification.
Phase III April 1 to July 1, 1974 - Modification
verification.
Phase IV July 1 to December 31, 1974 - Performance
testing, optimization and reliability.
The salient features of each phase will be reviewed. Attached to
this report is a summary of operations (Appendix A) listing each run,
the duration and the reason for shutdown.
Phase I
Initial start-up and operation was reasonably smooth. There were
two shutdowns caused by failures of stainless steel expansion bellows
allowing first stage slurry to leak into the second stage. Examination
verified that the bellows were not made of the specified 316 stainless.
The major problem was with the MgO feed system. Continual plugging
occurred in the MgO mix tank and suction lines to the MgO make-up pumps.
The longest continuous run during this phase was for 271 hours.
Approximately midway through this run the boiler was forced out for 24
hours with a tube leak. All liquid flows and levels were maintained and
flue gas returned to the scrubber as soon as the boiler returned to ser-
vice. This phase concluded when the boiler shut down January 14 for
annual maintenance.
Phase II
Inspection of the scrubber system was made after about 700 hours
operation. The system was basically in good condition with absolutely
731
-------
no sign of scaling or build-up. However, in the first stage where the
operating pH is less than two, there was corrosion of nuts, bolts,
hanger rods, spray nozzles, bellows and the vessel itself. Examination
determined that all the corroded parts were of non-specified material.
There was some very minor corrosion on 316 stainless. The corrosion of
the vessel occurred where the protective flake glass lining was pene-
trated. The problem here was in some part due to improper application
and in part due to construction damage after the lining was installed.
This represented a very small percentage of the total flake glass
lining.
The major moditication was the addition of a pre-mix tank in the
MgO feed system. Steam sparging was added to the MgO mix tank for
future use with recycle MgO. Other minor modifications were made in
piping.
Phase III
Operation resumed with start-up on April 15, 1974. The intent was
to operate to verify the modifications and then shut down to set up and
operate for performance testing. The pre-mix tank improved slaking but
not to an acceptable standard for long-term operation. We decided the
end of April that the system could be operated for performance testing.
We shut down, checked our inventory of MgO and storage space in the
magnesium sulfite silo and found we did not have enough MgO to run the
test nor storage space for sulfite. At this time we had 130 tons of
sulfite at Rumford waiting to be calcined. Boston Edison was using the
calciner and it appeared there was no chance of PEPCO's material being
calcined before July. Additional virgin MgO had not been ordered as we
were expecting to be able to test on recycle MgO. We did operate in May
to empty the MgO silo.
Chemico decided to replace the pre-raix tank with a "solids liquid
mixing eductor" to improve slaking.
Phase IV
PEPCO received permission to use the EPA owned calciner at Rumford,
Rhode Island July 1, 1974. Virgin MgO was on order and we were expecting
recycle MgO. Virgin MgO arrived first near the end of July. The first
start-up was August 1. The mixing eductor proved totally unsatisfac-
tory - plugging continually. After 10 hours we shut down and re-installed
the pre-mix tank which had been modified. Preliminary tests operating on
virgin MgO indicated an S02 removal in the 707» range. Chemico felt the
pressure drop across the absorber throat would have to be increased to
732
-------
obtain the 90% design removal. These modifications were made and the
indicated removal efficiency on virgin MgO is in excess of 90%.
Our first recycle MgO was received and introduced into the system
on August 16. The dryer feed material became sticky and caused caking
in the dryer. This was believed to be caused by unreacted MgO in the
centrifuge cake. During the next run steam sparging was used to raise
the temperature in the MgO mix tank to correct this problem. It has
also been necessary to change the dryer operating temperature on recycle
MgO. Slaking with the modified pre-mix tank has been satisfactory on
both virgin and recycle MgO.
In conjunction, with Chemico, Basic, Essex and EPA, we are currently
into a 6-month program to test, optimize operating conditions and gain
reliability experience. The system looks encouraging - there has been
no problem to rule out the technical feasibility of the Mag-Ox process
for S02 removal. There are still some problems in the sulfite handling
equipment. The equipment is apparently sized for steady state operation.
The centrifuge hopper and the dryer tend to hold up material and then
release it in a slug that overloads the sulfite conveyors.
A performance test program has been completed by York Research and
while formal results are not available, the indicated S02 removal effi-
ciency is in the 88% to 96% range as gas flow varies from 150,000 to
300,000 acfm. Particulate removal is in excess of 99% when taking flue
gas from either before or after the precipitators.
Operating Summary
We still continue to be plagued with minor problems causing shut-
downs. Examples of these are - corrosion leaks in rubber lined pipe,
erosion leaks in second stage piping, pump seal problems, bearing failure
in sulfite bucket elevator, etc. Operating availability continues to
improve - for the month of August it was 43.5% and since August 13 the
availability is running 55 to 60%. In contrast to these figures, the
availability of the Dickerson units has been around 90% since their
installation. A scrubber system should be capable of matching boiler
availability. We would therefore recommend the installation of a scrubber
by-pass system rather than spare scrubber capacity until more experience
is obtained on long-term reliability.
Economics
When PEPCO started investigating scrubbers in 1969 the estimated
capital cost of a scrubber system was $12 to $20 per kilowatt. In the
733
-------
January 1974 EPA report on the October 1973 scrubber hearings the capital
cost was given as $50 to $65 per kilowatt. PEPCO's current in-house
estimate in 1974 dollars is in excess of $100 per kilowatt.
Operating costs are difficult to estimate as we do not have suffi-
cient information on MgO make-up required or a real estimate of the
maintenance costs. Based on our present experience, it will require two
additional operators per shift.
The estimated increase in power cost for the Dickerson Station
operating at an 857. capacity factor would be about 5 mills per kilowatt
hour of which over 3 mills is the fixed charge on investment. These
costs assume a recovery system through on-site production of SO? gas.
The assumption is made that the sale of elemental sulfur or sulruric
acid will cover all fixed and operating costs associated with an on-site
sulfur or acid plant. We realize that this is a debatable assumption,
but at this time we do not have a good estimate for this cost. For
instance, a chemical company operating an acid plant would have to be
guaranteed a continuous feed for the plant. What will be the availa-
bility of a scrubber system? What will be the variation in sulfur con-
tent of coal? These are the kind of questions that will have to be
resolved to arrive at definite costs.
CONCLUSION
PEPCO entered into the development of stack gas desulfurization
based on an analysis of low sulfur fuel availability and the projected
costs of these systems. Throw-a-way systems were rejected as at our
particular plant sites, the sludge disposal would cause more environ-
mental problems than the removal of sulfur dioxide would solve. We
believe these were correct decisions based on the then available data.
Where do we stand after six years and expenditures approaching
$9,000,000? The Mag-Ox process shows promise as a means of removing
sulfur dioxide from flue gas. The SO? can be absorbed, the sulfite
crystals can be removed and a marketable sulfuric acid can be produced.
There are still some problems to be solved. These problems are:
1. MgO - is there a limit to the number of times it can
be recycled? To date, we have been through only one
cycle.
2. Material balance - it must be determined how much
make-up MgO is required and whether any MgO is being
lost in an environmentally unacceptable manner.
734
-------
3. Long term reliability and availability must be demonstrated.
This is a material selection problem to select the best
material to stand up under corrosion and erosion attack.
4. Economics.
The economic situation has changed so drastically that decisions on
future installations must be based on a valid cost-benefit analysis for
each specific location. For example, we are required by the State of
Maryland to maintain at each plant a seven day supply of fuel containing
less than 1% sulfur. This fuel is to be used on order from the State
when there is an air pollution problem in the area. This fuel has been
on-site for almost three years and we have never been ordered to use it.
This indicates to me that sulfur dioxide has not been a health problem
in any of the pollution alerts.
PEPCO continues to analyze the status of various stack gas desul-
furization systems, the availability of low sulfur fuels, coal gasifica-
tion, solvent refined coal and any other method for meeting our obligation
to produce reliable power with a minimal adverse impact on the environment.
735
-------
APPENDIX A
SUMMARY OF OPERATIONS
CHEMICO-BASIC MgO SULFUR REMOVAL SYSTEM
DICKERSON UNIT 3
POTOMAC ELECTRIC POWER COMPANY
1. Run #1 - September 13, 1973 for 36 hours
Shutdown - failed stainless bellows in first
stage flow to recycle pump suction.
2. Run #2 - September 24 for 10 hours
Shutdown - MgO feed problems - could not control pH
3. Run #3 - October 1 for 60 hours
Shutdown - plugged MgO tank and lines
4. Run #4 - October 12 for 30 hours
Shutdown - failed stainless bellows in first stage flow to
recycle pump suction (Not same bellows as #1)
5. Run #5 - November 6 for 13 hours
Shutdown - bearing failure MgSO-j bucket elevator
6. Run #6 - November 8 for 51 hours
Shutdown - boiler outage - blown tube
7. Run #7 - November 12 for 54 hours
Shutdown - erosion corrosion leak - first stage slurry line
to thickeners
8. Run #8 - November 30 for 24 hours
Shutdown - lost fuel oil to dryer
9. Run #9 - December 1 for 2 hours
Shutdown - lost fuel oil to dryer - underground fuel oil
line failure
10. Run #10 - December 9 for 138 hours
Shutdown - boiler outage - tube leak (maintained all scrubber
flows and levels during 24-hour boiler outage)
11. Run #11 - December 16 for 109 hours
Shutdown - partial pluggage of MgO feed lines was restricting
gas flow to scrubber
736
-------
12. Run #12 - January 6, 1974 for 192 hours
Shutdown - boiler off for scheduled overhaul and partial
plugging of MgO
Summary
September 13, 1973 to January 14, 1974 - 719 hours operation
and scrubber available for 70 hours that boiler was unavailable.
Scrubber Availability = 789 = 26.7%
2952
December 9, 1973 to January 14, 1974
Scrubber Availability = — = 53.67.
13. Run #13 - April 15, 1974 for 17 hours
Shutdown - flop gate in MgSC>3 screw conveyor
14. Run #14 - April 16, 1974 for 13-1/2 hours
Shutdown - Solidified MgO in pre-mix tank, K-Tron MgO weigh
feeder
15. Run #15 - April 22, 1974 for 10 hours
Shutdown - centrifuge hopper filled - bucket elevator tripped -
leak in 1st stage pinch type control valve
16. Run #16 - April 25, 1974 for 89-1/2 hours
Shutdown - bucket elevator tripped - mother liquor line to
pre-mix tank plugged
17. Run #17 - April 29, 1974 for 43 hours
Shutdown - not forced - to make minor modifications -
restart will be for test program
Availability April 15 to May 1 = rrr = 40.5%
18. Run #18 - 9:40 am, May 13 to about noon on May 17 - 100 hours
Shutdown - Out of MgO @ 6:45 am on May 17 - continued to operate
to empty MgO tank and strip crystals - bucket elevator
shaft broke
19. Run #19 - August 1 for 10 hours
Shutdown - Test solids -liquid mixing. Educ tor -plugged -down to
replace with pre-mix tank.
737
-------
20. Run #20 - August 6 for 54 hours
Shutdown to install throat restrictor
21. Run #21 - August 13 for 24 hours
Shutdown - offloading overweight MgSOj truck plugged conveyor
dryer to silo
22. Run #22 - August 15 for 64 hours
Shutdown recycle 1^0 caused caking in dryer
23. Run #23 - August 20 for 12 hours
Shutdown - bucket elevator would not handle flow of
24. Run #24 - August 26 for 84 hours
Shutdown - leak in 14" elbow
August summary operation 248 hours 33.2%
Available 323 hours 43.5%
25. Run #25 - September 3 for 28 hours
Shutdown - leak in 14" elbow
26. Run #26 - September 6 for 28 hours
Shutdown - 8" flex valve failed
738
-------
POWER PLANT FLUE GAS DESULFURIZATION
by the
WELLMAN-LORD SO PROCESS
PART I
THE DEAN H. MITCHELL STATION
(Northern Indiana Public Service Company)
by
E. L. Mann
Plant Engineering
Northern Indiana Public Service Company
Michigan City, Indiana
PART I I
CONTINUING PROGRESS FOR THE
WELLMAN-LORD SO PROCESS
by
E. E. Bailey
Senior Process Engineer
Davy Powergas Inc.
Lakeland, Florida
Prepared for Presentation at
Flue Gas Desulfurizatfon Symposium
Sponsored by the Environmental Protection Agency
Atlanta, Georgia
November k-7, 197**
739
-------
THE
DEAN H. MITCHELL STATION
(NORTHERN INDIANA PUBLIC SERVICE COMPANY)
WELLMAN-LOKD (DAVY POWERGAS INC.)
ALLIED CHEMICAL CORPORATION
S02 EMISSION CONTROL FACILITY
BY
E. L. Mann
Plant Engineering
Northern Indiana Public Service Company
Michigan City, Indiana
ABSTRACT
In a jointly funded EPA-NIPSCo. project, the Northern Indiana Public
Service Company will combine two technologies to provide a flue-gas-
desulfurization system in a 115MW pulverized coal-fired boiler at the
Dean H. Mitchell Station in Gary, Indiana. The S0£ controls will be
back-fitted onto a unit which went in service in 1970. These systems,
scheduled to go into service in late 1975, are The Wellman-Lord S02
Recovery Process which will produce a concentrated S0£ gas and the
Allied Chemical S02 Reduction Process which will produce high quality
elemental sulfur having a minimum assay of 99.5%. This quality is
suitable for sulfuric acid production.
Following acceptance tests at the conclusion of construction, a one-
year period of operation by Allied Chemical will be funded entirely by
NIPSCo.. A comprehensive emission-testing program will be conducted
by EPA during this period. In this final phase, severe test conditions
and refinements will be imposed on the system consistent with safe
operation of Unit 11 and evaluated by NIPSCo. and the EPA, each evaluat-
ing in his respective area of concern. The EPA and NIPSCo. expect to
use the data obtained to extrapolate to obtain costs and other data for
larger facilities.
Prepared for Presentation at
Flue Gas Desulfurization Symposium
Sponsored by the Environmental Protection Agency
Atlanta, Georgia
November 4-7, 1974
740
-------
THE
DEAN H. MITCHELL STATION
(NORTHERN INDIANA PUBLIC SERVICE COMPANY)
WELLMAN-LORD (DAVY POWERGAS INC.)
ALLIED CHEMICAL CORPORATION
SO, EMISSION CONTROL FACILITY
Northern Indiana Public Service Company is a combination gas and electric
utility operating in the Northern third of the State of Indiana. The
Company has 3 coal fired stations in operation, each station being
between 600 and 700MW in capacity. Each is located in environmentally
sensitive areas. The stations have used Mid West coal of 3 to 4% sulfur,
10% ash and 11,000 plus BTU/lb..
In seeking a suitable method for cleaning flue gases of S02, NIPSCo.
concentrated on finding a system that would be totally acceptable on
all environmental fronts. Water and land use in the area indicated that
a "throw-away product" removal system would not be acceptable. It also
appeared that recovery of the sulfur in the form of S02> sulfuric acid,
or any other sulfur compound might prove to be a problem at some future
date when it came to marketing or disposing of the product.
Combining two proven processes appeared the best assurance of meeting all
environmental requirements. The Wellman-Lord S02 Recovery Process is
meeting, or exceeding, required emission standards in treating tail gases
from several sulfuric acid and sulfur recovery plants in the U.S. and
Japan and at a Japanese 70MW oil fired generating plant and a 220MW
oil fired peaking unit in Japan. Allied Chemical's process has been
demonstrated in Canada, where it operated a system recovering elemental
sulfur from the Falconbridge nickel mine's 13% S02 smelter gas.
Furthermore, the information which we had indicated that retrofitting
some units with a throw-away system could be about the same capital cost
as a recovery system.
Initial discussions involving Davy Powergas, Allied Chemical Company,
Northern Indiana Public Service Company and the EPA took place early in
1972. The EPA and NIPSCo. reached agreement on a contract in June, 1972.
The EPA-NIPSCo. contract and, in turn, the NIPSCo.-Davy Powergas contract
has requirements for an emission control system which will operate at a
minimum of 90% S02 removal from the flue gases of the 115MW boiler while
feeding coal with a sulfur content of 3.5%. The process also is
guaranteed not to allow any emission in excess of 200 ppm by volume of
S02 in the exit gas with lower sulfur fuel. An existing electrostatic
precipitator is expected to reduce the dust loading to 0.044 grains per
actual cubic foot at the scrubber inlet. The Davy Powergas system is
741
-------
being designed with a pre-scrubber to remove 0.2 grains per ACF with
the capability of handling considerably greater fly ash loadings for
short periods. Ash content of the Midwest Coal is 10.6%, the BTU
content 11,008/lb..
The ductwork between the existing I.D. Fans and the 236 foot tall
chimney will be revised so as to divert the flue gases, 320,000 acfm
at 300°F, to the new installation. A booster fan will be installed
between the I.D. Fans and the SC>2 system. The flue gases, after
treatment in the scrubber system, will be reheated to 180°F by a
direct fired gas reheater. The release of the treated flue gases will
be at a point about 165 feet above ground level through a stack on
the top of the absorber. A by-pass back to the original stack is being
installed. The flue gases enter the 862 absorber relatively free of
particulate matter. The Wellman-Lord system employs a sodium-sulfite
rich solution of sulfite and bisulfite to remove the SC>2 from the
stack gas. The resulting bisulfite rich solution is pumped to a re-
generation unit in which the reaction is reversed in an evaporator/
crystallizer. The application of heat from a 50 psig steam system
generates water saturated S02 and sodium sulfite crystals which are
separated, redissolved and recycled as clean absorber feed. Caustic
soda or soda ash is added to replenish the system.
Some sodium sulphate is formed in the Wellman-Lord system. A small
purge stream is continuously withdrawn to prevent a sulfate buildup
in the system. Sulfates and a small amount of sulfites in the purge
amount to approximately 8 to 10% of the sulfur coming into the system.
The output of the Wellman-Lord plant is a gas stream of about 85% S02,
the remainder mostly water vapor. The S02 feed stream is received by
the Allied reduction plant. The gas is compressed by a bloxrer and
mixed with natural gas which serves as a reductant.
The stream then enters the primary reaction system consisting of two
packed-bed regenerative heat exchangers and a catalyst-packed reduction
reactor. The two heat exchangers operate cyclically, one giving up
heat to the entering gas mixture while the other is being reheated by
exit gas from the exothermic reaction. Periodic reversal of the gas
flow through the heat exchangers provides essentially stable operating
conditions in the reactor.
After exiting from the regenerative heat exchanger, gas from the primary
reactor passes first through a sulfur condenser, which removes the
elemental sulfur for storage in molten form. The remaining gas mixture
then enters a two-stage Claus unit containing an interstage condenser,
which also removes elemental sulfur to storage. A final sulfur condenser
removes the last portion of the elemental product, and the unrecovered
sulfur values in the tail gas are oxidized to SC>2 and finally returned
to the absorber of the Wellman-Lord S02 Recovery Process.
The contracts between NIPSCo. and EPA and the other principals, Davy
742
-------
Powergas and Allied Chemical, contain penalty-assessable guarantees
for the following: SC>2 emission levels - The process, according to
the contracts, will be at least 90% efficient firing coal with a
sulfur content of up to 3.5% (approximately 2,300 ppm by volume of
SC>2 in the stack gas). As an override, no more than 200 ppm by
volume of S02 will be present in the exit gas.
Mechanical reliability - The process plants will be mechanically sound.
Utilities consumption - The aggregate cost of steam, electric power,
and natural gas required for operation of the complex will not exceed
a specified cost; i.e., $56.00 per hour based on:
Electric Power - $0.007 per KWH
Steam - $0.50 per 1000 Ib. at 550 psig, 750°F
Natural Gas - $0.55 per million BTU.
Chemicals requirements - The quantity of make-up chemicals (caustic
soda or soda ash) will not exceed a fixed daily amount. The average
chemical make-up over a twelve (12) day operating period at an average
of 92MW shall be no greater than 6.6 tons per day of NA2CC>3. The value
of antioxidant used during the 12-day period shall not exceed an average
of $400 per day at an average of 80% load factor.
Product quality - Sulfur produced is to have a quality suitable for
use in the manufacture of sulfuric acid by the contact process.
It is recognized that the demonstration system is not intended or
designed to reduce emissions other than SC>2 and particulates. Reduction
of other pollutants by the demonstration system is incidental to the
demonstration objectives. Therefore, the required compliance with
current emission regulations shall apply only to particulates and SC>2.
If the project were completely financed by NIPSCo. , we calculate the
following data.
Capitalization, including top charges, escalated to 1975- $13,441,434
This is $117/KW installed capacity or $12,000 per million BTU/Hr.input.
Operation & Maintenance, including cost of capital * - $4,977,186
(based on 72% load factor)
This is 6.3 mils/KWH gross generated or 65 cents per million
BTU coal burned.
We, at NIPSCo., are looking forward to the results of this demonstration
project. We hope the data we and the EPA obtain will be of value in the
decision making of others.
743
-------
* Includes management fee, Allied process license (Davy Powergas Inc.
process license does not apply to this project).
Based on 10 year plant life.
744
-------
CONTINUING PROGRESS
FOR
WELLMAN-LORD SO, PROCESS
by
E. E. Bailey
Senior Process Engineer
Davy Powergas Inc.
Lakeland, Florida
Flue gas desulfurization for two large utility boilers, will be the
second application of the Wellman-Lord S02 process on a coal-fired
power plant. In Japan there are presently four successfully operating
Wellman-Lord S02 plants on oil fired, power and steam generating
plants.
Capital and operating costs for the Wellman-Lord flue gas desulfuriza-
tion plant can be minimized by the proper selection of design criteria.
This paper will present a brief outline of the above new project.
Coverage will also be given to some of the Wellman-Lord installations
in Japan. Finally a section on critical design criteria will be
presented to emphasize savings in capital and operating costs.
745
-------
The client has contracted to build two coal-fired power plants and
although the coal that would be burned in both of these power plants
was low in sulfur content -- .8% to 1.3% ~- it was still necessary
for each plant to be suitably equipped with a sulfur dioxide and
particulate removal facility. Burning the low sulfur coal and in-
stalling electrostatic precipitators for fly ash collection would
bring these power plants - 330 MW and 340 MW - very close to meeting
the Federal Emission Standards. State standards, however, require
that pollution control facilities be installed.
TABLE 1 - Comparison of NonControl1ed Emissions With
Existing Federal and State Air Pollution Regulations
Sulfur Dioxide Particulate
Stack Concentration Stack Concentration
ppm Ibs/MM BTU Grains/acf Ibs/MM BTU
UNIT NO. 1
Avg. Grade Coal 700 1.98 0.026 0.104
Low Grade Coal 1150 3-25 0.044 0.175
UNIT NO. 2
Avg. Grade Coal 700 1.96 0.039 0.154
Low Grade Coal 1150 3-23 0.062 0.245
FEDERAL STANDARDS
(As of June 1974) 1 .2 0.10
STATE STANDARDS
(As of Sept. 1973) -34 O.OS(Max)
0.02( 2 )
746
-------
The task of selecting a suitable sulfur dioxide-particulate removal
process was performed by the client and their engineer/architect
for the two power plants. Preliminary inquiries were solicited and
studies were made of existing power plant pollution control tech-
nology. Following completion of these studies, a detailed bid
specification was prepared. After evaluation of all bids by the
Client, the sulfur dioxide-particulate removal project was awarded
to Davy Powergas for installation of its Wellman-Lord sodium sulfite/
bisulfite process in combination with the Allied Chemical Corporation
SC>2 reduction process for the production of elemental sulfur.
DESCRIPTION OF PROCESS
The entire sulfur dioxide-particulate removal facility will consist
of a gas handling area for each power plant, an absorbing solution
regeneration plant, an area for the reduction and concentration of a
by-product purge stream and finally an elemental sulfur plant using
Allied Chemical Corporation's technology.
Each gas handling area is designed to process 1,588,000 acfm of gas
containing between 700 ppmv - 1200 ppmv sulfur dioxide. The fly ash
loading of the incoming gas will range between 0.06 and 0.0^» grains/acf.
The flue gas from each power plant is picked up at the stack entrance
by three hot side fans, operating in parallel. THe discharge pressure
from each fan is sufficient to force the flue gas through a wet venturi
type prescrubber and a tray type sulfur dioxide absorption tower. Each
booster fan is directly connected to its own venturi and absorber. The
cleaned gases from each of the absorbers are combined for reheat before
re-entering the power plant stack for discharge to the atmosphere. Each
power plant flue gas handling area has one spare fan, venturi and ab-
sorption tower at the request of the Client.
The sulfur dioxide rich absorbing solution from both power plant gas
handling trains is collected at a central point before being processed
in the regeneration area. This central collection point provides
storage of rich absorber solution. It also provides a supply of lean
absorber solution so that the regeneration area may be shut down for
seven (7) to twenty-four (24) hours for maintenance and operational
clean out requirements without affecting the plant's S02 removal
capabi1i ty.
The chemical regeneration area consists of two parallel, double effect,
forced circulation evaporators. The rich absorbing solution is therm-
ally regenerated by driving off sulfur dioxide and water vapor. This
vapor passes through a series of partial condensers where the sulfur
747
-------
dioxide vapor is concentrated to 85 vol. %, compressed and forwarded
to the Allied sulfur plant. The regenerated absorbing solution leav-
ing the evaporators is combined with stripped condensate from the
partial condensers and recycled back to the absorbing solution storage
area.
A slip stream of absorbing solution is taken from the absorbers on
both power plants for processing in a purge treatment area. Through
a series of unit operations the inactive sodium salts are concentra-
ted and removed from the remaining absorbing solution. The final
purge liquor will then be dried and the resulting product - 85 wt %
Na^SO, ~ will be marketed. As a result of this purge treatment area,
the chemical make up requirement can be reduced by a factor of k.
Therefore the chemical make-up cost for a plant of this size would be
reduced by some $2.7 million per year with the installation of a
purge treatment area.
The Allied sulfur plant consists of two identical, parallel trains,
each capable of producing 60 LTPD of sulfur.
The sulfur dioxide-water vapor which enters the sulfur plant is reacted
with natural gas over a catalyst, generating a mixture of hydrogen sul-
fide and sulfur dioxide. This resulting gas then passes through a
series of Claus reactors, generating sulfur and an off gas which is
incinerated and recycled to the gas handling areas. The sulfur
produced will have a purity of 99-9% and will be marketed.
PROJECT SCHEDULE
The compliance date for this power plant is July 1, 1977- By this
date both of the gas handling, sulfur dioxide-particulate removal
facilities must be operating successfully to meet both the State and
Federal air pollution abatement regulations. Since the project was
awarded in the Spring of 197^+ and expected start up is early 1977,
the total length of this project through engineering, design, pro-
curement and construction is planned for thirty (30) to thirty-three
(33) months.
748
-------
PURGE REDUCTION PROGRAM
The Wellman-Lord process is, like all other alkali scrubbing processes,
plagued by oxidation and disproportionation of the sodium sulfite/bi-
sulfite absorbing solution. Whatever the mechanism is for breaking
down the active sodium salts, it is difficult to reverse this action
to rejuvenate the inactive sodium. Therefore, all inactive sodium
salts presently must be purged from the system and their sodium ions
replaced by active sodium ions in the form of soda ash or caustic
soda.
Due to the waste purge streams, Davy Powergas has been attempting for
two years to find or develop a process for reactivating the sodium ion.
A number of processes have been investigated, especially those related
to the pulp and paper industry where the chemistry is very similar.
Preliminary laboratory scale tests and even one full scale plant test
have been run on the different available processes. Since this devel-
opment program is of a highly confidential nature, the specific pro-
cesses and the results of the tests conducted will not be discussed.
However, a few processes do look promising. Engineering for a demon-
stration unit will probably begin early in 1975. Additional tests are
being conducted this Fall to determine the most applicable process and
the necessary scale up factors which will be required for the demon-
stration unit. Location of the planned demonstration unit has not
been chosen, but it will probably be in the United States at one of
the existing, or possibly new, Wellman-Lord sulfur dioxide recovery
pi ants.
In searching for this purge reduction process, Davy Powergas is not
looking for a mechanism to reactivate the sodium ion at the expense of
a "More Easily Disposed of Product" or by-product that "can be sold"
but that nobody wents to buy, but instead we hope to close the loop
by generating sodium carbonate. In closing the loop we intend to re-
activate as much of the inactive sodium ion as economically feasible.
But, even if this is 100% successful, there still must be some waste
purge stream due to the build up of make up material contaminants.
Details of the specific demonstration plant will be the subject of a
later paper.
749
-------
FLUE GAS DESULFURIZATION IN JAPAN
The Wellman-Lord process has been installed on an oil-fired, 220 MW
power plant for Chubu Electric Power Company in Nagoya, Japan. This
installation was designed and constructed by one of our Japanese
licensees, Mitsubishi Kakoki Kaisha, and came on stream in late Spring
of 1973. This particular power plant is a peaking station; therefore,
the gas handling-SCL absorption equipment must be capable of handling
load variations of from 35% to 105% of design flow a number of times
each day. Also to be considered is the suspension of power plant op-
eration for the weekend.
This weekend suspension is handled by the installation of ample ab-
sorbing liquor surge capacity so that the chemical plant area and the
sulfuric acid plant (end product) can operate at low rates. When the
power plant is brought back on line after the weekend suspension of
operation, the flue gas sulfur dioxide absorption area is started up
automatically. This automatic start up and any subsequent suspension
of operation of the entire desulfurization plant can be handled from
the control room. As a result of this highly automated system the
entire operation of the desulfurization plant, including start up and
shut down, can be handled by two operators.
Although the Wellman-Lord sulfur dioxide recovery system installed at
Japanese Synthetic Rubber Company in Chiba, Japan is not on an electric
power plant it is still worthy of mention because of its highly success-
ful operating history. This unit is installed on two oil-fired boilers,
each generating 185,000 Ibs/hr of steam, and it handles about 12^,000
scfm of flue gas. The plant started up in June 1971 and maintained
a 97% stream time availability in its first year and a 100% stream
time availability its second and third years of operation. This
sulfur dioxide recovery plant is removing more than 90% of the S0_
from a flue gas containing approximately 2100 ppmv SO.,.
At the present time there are ten (10) active projects for the in-
stallation of Wellman-Lord flue gas desulfurization plants on oil
fired boilers in Japan. These ten plants are being designed and
constructed by the two Japanese Wellman-Lord licensees. The total
boiler exhaust gas to be desulfurized is in excess of k million
actual cubic feet per minute.
750
-------
Overall View of Well man-Lord SOg Installation
At Chubu Electric Power Plant, Nagoya, Japan
751
-------
DESIGN CRITERIA AND HOW THEY AFFECT COSTS
Although many design criteria must be considered when designing any
sulfur dioxide-particulate removal facility, there are also several
process criteria that should be considered carefully by both the
client and by Davy Powergas when designing a Wellman-Lord SO.. Recovery
System. In this section we will highlight the most influential
criteria and point cut ways to minimize their affect on the plant de-
sign. The criteria covered will include: SO content of the gas,
water content of the gas, oxygen and sulfur trioxide concentrations of
the gas, fly ash loading and expected removal, and total flue gas
quanti ty.
SULFUR DIOXIDE CONCENTRATIONS
Since sulfur dioxide removal is the primary function of this process,
SO concentrations in the feed gas have no maximum or minimum limita-
tions. The normal minimum S0? concentration in the processed exhaust
gas is 200 ppmv; however, lower concentrations can be achieved if nec-
essary.
The size of the S0« absorption area is mainly a function of the flue
gas flow rate whicn is governed by the size of the power generating
unit; however, the sulfur dioxide concentrations, in and out, do affect
the number of mass transfer stages required in the absorber. Generally
speaking, a flue gas with a sulfur dioxide concentration of 1200 ppm
requiring 90% removal will take four (k) to five (5) stages. Flue gases
with higher concentrations - say 3000 ppm, requiring 95% removal -
may need only three (3) stages. The difference is due to the higher
driving force of the higher SO^ concentration gas. Each trayed stage
will require a minimum of three (3) inches water pressure drop. Using
an airfoil booster fan, this may mean as much as $22,120'per year
additional operating cost for each additional mass transfer tray in-
stalled on a 100 MW power plant.
The chemical plant and any subsequent final product plant, elemental
sulfur or sulfuric acid, will be directly affected by S0? concentra-
tions in the flue gas. Each 1000 ppmv of SO removed from the exhaust
gas of a 100 MW power plant represents approximately 15 LTPD of elemen-
tal sulfur or 50 TPD of 100% sulfuric acid. Concentrations of less
than 2000 ppmv SO in the flue gas increase the size of the chemical
plant on a per pound of S0_ absorbed basis. This is due to the
higher absorbing solution circulation rate and subsequent higher
pounds of water vapor per pound of sulfur dioxide evaporator boil up
requi red.
752
-------
WATER CONTENT OF FLUE GAS
The water content of the flue gas for a given dry bulb temperature
will have a direct affect on the number of stages required to satisfy
the sulfur dioxide removal requirements. The absorber section of the
gas handling train will operate at a temperature *+°F to 12°F higher
than the adiabatic saturation temperature of the flue gas. Since
this operating temperature of the absorbing section will affect the
SCL equilibrium, a 2.0 volume % change in water content of the flue
gas may require one additional mass transfer unit.
OXYGEN AND SULFUR TRIOXIDE CONTENT OF FLUE GAS
The formation of inactive sodium salts caused by the oxidation of
sodium sulfite is directly proportional to the amount of oxygen and
sulfur trioxide contained in the flue gas. Therefore, the area most
affected by these component concentrations is the Purge Treatment
Area. The size of this area is almost directly proportional to the
oxygen content of the flue gas. The amount of by-product sodium
sulfate formed for each one percent oxygen in the flue gas from a
100 MW power plant is approximately 1.3 TPD.
FLY ASH LOADING AND PERCENT REHOVAL
The most economic removal of fly ash from a power plant flue gas is
accomplished by combining an electrostatic precipitator with a wet
scrubber. Since a wet scrubbing type device (Prescrubber) is used as
a gas saturation unit on the Wellman-Lord process, its use to effect-
ively and efficiently remove fly ash particles must be investigated.
When sizing the prescrubbing equipment careful attention must be paid
to the actual fly ash loading, the percent removal required and the
particle size distribution of the fly ash in the gas to the pre-
scrubber.
The more critical the prescrubber design criteria becomes (lower grain
loading, higher removal efficiencies, smaller particle size) the
higher the energy required to meet the demands. Each additional inch
of water pressure drop across the prescrubbing unit will cost approx-
imately $9130 in operating cost per year for a 100 MW power plant when
using a radial blade booster fan.
Another design parameter which must be considered is the disposal of
the captured fly ash. This captured fly ash is usually in the form
of a 1 to 10% by weight slurry in very acidic water. Depending on the
overall fly ash removal equipment, this fly ash stream may be blended
with the dry electrostatic precipitator fly ash or it may require
neutralization and solids concentration before disposal.
753
-------
FLUE GAS QUANTITY
Although the sulfur dioxide concentration and the fly ash loading of
the power plant flue gas play an important role in the sizing of the
gas handling area, the quantity of flue gas has by far the largest
affect. Since our operating history to date has shown better than
96% on stream time for our absorption areas we do not recommend spare
units, but instead suggest that the design be based on a realistic
maximum gas flow. To date it has been our recommendation that a
modular concenpt be used in the gas handling area. The present maxi-
mum size blower, prescrubber and absorber module would have a gas
handling capacity equivalent to a 120 - 150 MW coal-fired power plant.
WELLMAN-LORD SO2 RECOVERY PROCESS
(UTILITY INSTALLATIONS)
754
-------
WELLMAN-LORD S02 RECOVERY PROCESS
(UTILITY INSTALLATIONS)
-------
APPENDIX A
DETAILED PROCESS DESCRIPTION
The Wellman-Lord Sulfur Dioxide Recovery Process is a chemical ab-
sorption, thermal regeneration process using an active sodium based
absorbing solution. The total processing plant consists of a gas
handling area, a chemical regeneration area and a purge treatment
area. The description which follows will emphasize the removal and
recovery of sulfur dioxide from power plant flue gas.
ABSORPTION AREA
The primary function of the Absorption Area is to remove the sulfur
dioxide from the flue gas. Secondary functions of gas saturation,
particulate removal and exhaust gas reheat are also accomplished.
The flue gas leaving the power plant air preheaters (or electrostatic
precipitators) at 250°F to 300°F and essentially atmospheric pressure
is picked up by large booster fans. These fans impart sufficient
static pressure to force the gas through the entire gas handling train
and out the stack. The gas discharging from the blower will first
enter a gas-saturation, prescrubbing unit. The amount of fly ash
contained in the gas, the size distr;bution and the required percent
removal will dictate the unit type and size and pressure drop required.
The prescrubber will remove fly ash by recirculation of an acidic
water-fly ash slurry through a venturi type or a valve tray type unit.
Saturation of the flue gas also occurs in the prescrubber, adiabatically
reducing the gas temperature to 120-130°F. At the exit of the pre-
scrubber, chevron demisting devices minimize fly ash carry over into
the absorbing section. Either all or a portion of the prescrubber
is housed in the lower one-third of the absorption tower.
The relatively particulate-free, saturated flue gas now passes into
the sulfur dioxide absorption section of the tower. The removal of
sulfur dioxide from the flue gas is achieved by a chemical absorp-
tion process. A solution which is rich in sodium sulfite enters the
top of the column and is converted to a solution rich in sodium bi-
sulfite as sulfur dioxide is absorbed. The mass transfer medium used
in the absorption tower will be a series of sieve or valve trays. Due
to the extremely low liquid to gas ratio required for absorption, each
set of trays requires recircu1 at ion of the absorbing solution. This
recirculation provides the necessary amount of liquid required to
hydrau1ical1y load the valve trays.
The absorbing solution rich in sodium bisulfite leaves the bottom of
the absorbing section for processing in the chemical plant area. The
degree of conversion, amount of sodium sulfite converted to sodium
756
-------
bisulfite, of the absorbing solution is important because it governs
the size of the chemical plant and the amount of steam required to re-
generate the sodium bisulfite to sodium sulfite.
Clean flue gas passes through either a mesh pad or chevron type de-
misting sectFon when leaving the absorbing section. This demister
minimizes absorbing solution carryover into the power plant stack.
The final section of the gas handling area is a stack gas reheater.
This reheater will eliminate the plume, provide bouyancy to the ex-
haust gas and minimize corrosion of the stack by supplying superheat
to the flue gas. The amount of reheat can vary, depending on stack
height, area climatic conditions, and materials of construction, but
usually requires a 50°F increase in gas temperature.
CHEMICAL PLANT AREA
Thermal regeneration of the bisulfite rich absorbing solution is
accomplished in a forced circulation, vacuum evaporator. The sodium
bisulfite in the feed solution is thermally converted to sodium sul-
fite by driving off sulfur dioxide overhead. Water vapor is also re-
moved overhead in an amount which is necessary to satisfy the water
balance on the absorbing solution. The sodium salts remaining in the
evaporator circulating liquid now form a crystalline slurry. The
major component of this crystal \s sodium sulfite. The concentration
of crystals in this evaporator circulation slurry is kept at an optimum
level in order to reduce heat transfer area fouling and decrease
evaporator size.
The overhead sulfur dioxide-water vapor mixture is concentrated
through a series of partial condensers. Depending upon the size of
the chemical plant, which is based on the amount of sulfur dioxide
absorbed, the first partial condenser for a portion of the sulfur
dioxide-water vapor may be the heat exchanger for a second stage
evaporator. As the size of the chemical plant increases somewhere
beyond the "production" of 5,000 1bs of sulfur dioxide the operating
cost economics begin favoring a double effect evaporator. The con-
densers following the evaporators may be air cooled or water cooled
depending on the process absolute pressure, product purity required
and cooling medium available. Since a vacuum pump is required, it is
proposed to cool the SO product gas down as far as possible with the
cooling medium and method available. The product SCL, at least in the
case of a power plant, will be used to make elemental sulfur or sulfuric
acid and it is therefore advantageous to have its concentration as high
as possible. Normally, this product gas will be approximately 85 volume
% sor
Water vapor condensed in the partial condensers contains varying small
amounts of dissolved sulfur dioxide. All of this "sour" condensate is
757
-------
Single Effect Evaporator Chemical Plant at the
SOCAL, El Segundo, California Installation
of the WeiIman-Lord Process
758
-------
The purge streams from the evaporator and purge treatment area are then
combined to be further processed. The next processing step can be one
of drying or neutralization. Drying of the purge will result in a
marketable sodium sulfate by-product while neutralization with
sulfuric acid will give a soluable waste product with a very low COD
value.
759
-------
T?
U
ILLINOIS POWER COMPANY
THE MITRE CORPORATION
M74-106
761
NOVEMBER 1974
-------
M74-106
E.M.JAMGOCHIAN
The MITRE Corporation
McLean, Virginia
W.E. MILLER
Illinois Power Company
Decatur, Illinois
Presented to
THE FLUE GAS DESULFURIZATION SYMPOSIUM
NOVEMBER 1974
ATLANTA, GEORGIA
Sponsored By
CONTROL SYSTEMS LABORATORY
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA
NOVEMBER 1974
762
-------
CONTENTS
Page
-MM&bM»
Part I—The "Cat-Ox11 Project at Illinois Power 765
1.0 Introduction 765
2.0 Capital and Operating Costs 766
3.0 Theoretical Description 767
4.0 Project Timetable 769
5. 0 Guarantee and Operating Program 770
6.0 Summary 771
Part n--Mitre Test Support for the "Cat-Ox" Project 774
1.0 Background 774
2.0 Program Summary 774
3.0 One-Year Test Program Objectives 776
4. 0 Test Program Design 800
763
-------
ABSTRACT
THE "CAT-OX" PROCESS DEMONSTRATION PROGRAM
This paper consists of two parts. Part I by W. E. Miller of
Illinois Power describes the installation and step-by-step operation
of the "Cat-Ox"1 process. Part II by E. M. Jamgochian of The MITRE
Corporation discusses the one-year test program which will be conducted
when the Cat-Ox process is again operating.
Part I - The catalytic oxidation method developed by Monsanto Enviro-
Chem Systems, Inc., for removing sulfur dioxide from flue gas of fossil
fuel generating stations was installed as a prototype installation with a
15 megawatt capacity at the Portland Station of Metropolitan Edison
Company in 1967. The first commercially-sized installation has been
installed on the 103 megawatt Wood River #4 unit of Illinois Power Company.
Although start-up and operational problems have kept the system from
operating reliably, the principle of operation has been proven. The Wood
River project was financed jointly by the Control Systems Laboratory of
the Office of Research and Development of Federal EPA and by the Illinois
Power Company.
Part II - MITRE test support for the "Cat-Ox" Demonstration Program is
sponsored by the Control Systems Laboratory, Environmental Protection
Agency. Status of the major task areas, and the accomplishments to date
are reviewed. The areas of investigation for the one-year demonstration
program are identified and the specific test objectives are discussed.
The test program will evaluate process design performance, process main-
tainability, process availability, and process operating costs, and is
designed on the basis of statistical considerations.
764
-------
PART I
THE "CAT-OX" PROJECT AT ILLINOIS POWER
1.0 INTRODUCTION
In 1969, Illinois Power Company employed Battelle Memorial Insti-
tute to make a survey of research, and development programs being con-
ducted on S02 removal systems. After an intensive study, it appeared
that the Cat-Ox System, following many years of research by Monsanto,
was a feasible method and the system most nearly ready for a demonstra-
tion installation. Therefore, early in 1970, Illinois Power decided to
install the first commercially-sized demonstration installation of the
Cat-Ox System on the 103 megawatt Unit /M at Wood River. This project
has been jointly funded by the Federal EPA and by Illinois Power Company
in an effort to advance the science of sulfur dioxide removal by develop-
ing a system which will produce a usable by-product in the form of sul-
furic acid. Unit #4 at Wood River normally burns approximately 275,000
tons of coal per year with an average sulfur content of 3.1%. Based on
these figures, the Cat-Ox System should produce about 25,000 tons per
year of 78% concentration sulfuric acid.
Upon completion of contract negotiations, construction of the Cat-Ox
System started in January, 1971, and the associated Research-Cottrell
precipitator designed for Cat-Ox was completed and placed in service in
February, 1972. Due to construction delays, initial start-up of the
sulfur removal equipment did not occur until September 4, 1972, using
natural gas for the in-line reheat burners. The System was operated for
444 hours during this period. When natural gas became unavailable, in
October, 1972, it became necessary to operate the inline reheat burners
on #2 oil. The reheat burners were tested with fuel oil and modified
until operable during the period November, 1972 to June, 1973. The
Performance Guarantee Test was satisfactorily completed using #2 oil in
July, 1973. Operation in July, 1973 brought the total operating time for
the "Cat-Ox" system to 602 hours. However, it was apparent to everyone
involved, including Federal EPA personnel, that the continuous long-term
operation of the in-line heaters on #2 oil would cause excessive con-
tamination of the catalyst, thereby causing frequent shutdowns for catalyst
cleaning. Since this would not be acceptable, it was agreed to construct
an external reheat burner using #2 oil with the required heat being ducted
into the System at the present location of the in-line burners. The
Federal EPA, Monsanto, and Illinois Power are sharing the additional costs
which now exceed $300,000 for the modification. The installation of the
external burner was completed in April, 1974, and the unit was placed in
operation May 7, 1974. During start-up, a leak in the lead lining of the
absorbing tower was discovered. In addition, the water intake screens
were plugging with debris from the Mississippi River which requires a
change in design of the intake structure. It was further discovered that
the impellers in the acid circulation pumps have been thermally shocked
765
-------
and must be replaced. These repairs and modifications have now been
made and the System was placed in operation on August 14, 1974. How-
ever j additional mechanical and structural problems have since developed
and the operation of the system has been sporadic.
2.0 CAPITAL AND OPERATING COSTS
The original capital costs involved in the design and construction
of the Cat-Ox System were estimated at $7,300,000, including the cost of
providing off-battery services. Of this amount, Illinois Power contracted
to pay $3,800,000, and the Office of Research and Monitoring of the
Federal EPA contracted to pay $3,500,000. Illinois Power agreed to pro-
vide the required auxiliary services, including natural gas or fuel oil
for the reheat burners, cooling water for acid coolers, electric services
for the induced draft fans, electrostatic precipitator, pumps, and
facilities for ash disposal. Final costs are now approaching $8,500,000
resulting in an installed cost of about $84.00 per kilowatt based on the
103 megawatt net capacity of the generating unit. Of these final costs,
Federal EPA will pay $3,700,000 and Illinois Power approximately
$4,800,000.
Operating and maintenance costs are expected to run approximately
$962,000 per year, as shown in the following tabulation:
Operating & Maintenance Cost Estimate
Operating & Maintenance Labor $ 36,000
General Maintenance 2% x 8,000,000 $160,000
Make-up Catalyst 200,000 liters x .025 x 4
screenings per year $ 30,000
Reheat burner fuel (burner rating 50 MM BTUH) $534,000
Electric Power (4% of generator output)
Capacity cost 4200 Kw x $200/Kw x .15
carrying charge $126,000
Energy cost 4200 Kw x 6000 hrs. x $.003
per Kwh equals $75,600. Use $76,000. $ 76,000 $202,000
$962,000
Possible credit for sale of acid $200,000*
It should be emphasized that these figures do not include the cost of
money or the cost of any major maintenance.
* Assuming a $10 per ton net return.
766
-------
3.0 TECHNICAL DESCRIPTION
The operation of the Cat-Ox System consists basically of the following
six separate phases:
1. Fly Ash Collection
2. Conversion of S02 to 303
3. Heat Recovery
4. Removal of sulfuric acid
5. Acid mist elimination
6. Acid storage and loading
These basic steps are shown diagrammatically in Figure 1 and are
described below.
1. Fly Ash Collection
The existing mechanical collector remains in service on Unit #4
to remove most of the fly ash from the flue gas. A new Research-Cottrell
electrostatic precipitator with a design efficiency of 99.6% has been
installed to work in series with the mechanical collector to remove
essentially all the particulate matter from the flue gas. After leaving
the electrostatic precipitator, the cleaned flue gas is heated and passes
into the converter of the Cat-Ox System, or, during start-up or unusual
operation, can be bypassed directly to the stack. The fly ash collected
by the precipitators is conveyed pneumatically to the existing ash pit
area. The electrostatic precipitator installation was completed in
February, 1972 and has been operating with Unit #4 since that time.
2. Conversion of S02 to SO^
The temperature of the flue gas leaving the electrostatic pre-
cipitator is 310°F. and is reheated to 850°F. to allow a 90% conversion
of S02 to 803. It was proposed that this be done by two in-line reheat
burners using natural gas or No. 2 fuel oil and by recovery of sensible
heat from the treated flue gas. The reheat burners were designed to
maintain the 850°F. conversion temperature regardless of boiler load.
Following reheat to conversion temperature, the flue gas enters the con-
verter where the Cat-Ox © catalyst (a vanadium pentoxide catalyst) reacts
with the S02 to form 803. The converter is designed so that the catalyst
bed can be emptied onto a conveyor system for transport to a screening
process after which the cleaned catalyst is conveyed back to the con-
verter. About 2.5% of the catalyst mass is lost during each cleaning
process which is anticipated to occur about four times per year. About
forty-eight hours is required for each catalyst cleaning.
767
-------
FLUE GAS
FROM EXISTING
ID FAN
DAMPER
STACK
REHEAT1
BURNERn
CONVERTER
GAS HEAT
EXCHANGER
f
SULFURIC
ACID
CAT-OX
MIST
ELIMINATOR
ABSORBING
TOWER
ACID
COOLER
RECYCLE
STORAGE
FIGURE 1
THE REHEAT CAT-OX SYSTEM
-------
3. Heat Recovery
The treated flue gas, now containing 803, passes to a Ljungstrom-
type heat exchanger where about 400°F. sensible heat is recovered to heat
the incoming untreated flue gas. As a result of heat recovery in this
exchanger, the overall need for fuel usage is to add 150°F. of sensible
heat in the Cat-Ox process. The temperature of the gas is maintained
well above the dew point. Normal flue gas leakage in a regenerative
heat exchanger of this type will allow about 5% of the flue gas to by-
pass the converter, thereby reducing the overall efficiency of S(>2
removal to approximately 85%.
4. Removal of Sulfuric Acid
The flue gas is further cooled in a packed-bed absorbing tower
which operates in conjunction with an external shell and tube heat
exchanger. During cooling, the H20 and 863 in the flue gas combine to
form sulfuric acid which condenses in the absorbing tower. The tower
brings cooled sulfuric acid into direct contact with the rising hot flue
gas. Exit gas leaves the packed section at about 250°F. while hot acid
is constantly being removed from the bottom of the tower and cooled in
the external heat exchanger and either recycled or sent to storage.
5. Acid Mist Elimination
Very fine mist particles of sulfuric acid are formed in the gas
as it is cooled in the absorbing tower. These mist particles in the flue
gas are removed along with some entrained droplets of circulating acid
from the tower by the Cat-Ox mist eliminator system. The packed section
of the absorbing tower and the mist eliminators are contained within one
vessel. The flue gas leaving the mist eliminator to enter the exit stack
contains less 803 than the amount normally emitted from the combustion
process.
6. Acid Storage and Loading
The cooled acid amounting at full load to 12 gallons per minute
of 78% H2S04 is collected in two 400,000 gallon steel storage tanks. An
acid loading pump and tank car loading facilities are provided adjacent
to the storage tanks. Tank trucks may be loaded from this station if
desired.
4.0 PROJECT TIMETABLE
The time schedule for the Cat-Ox System includes:
Design and Capital Cost Estimates - Initiated June, 1970
Detailed Engineering & Equipment Procurement - Initiated June,
1970
769
-------
On-Site Construction - Initiated January, 1971
Electrostatic Precipitator - Placed in Operation February, 1972
Performance Guarantee Test - July, 1973
Data Phase, including 15 months operational testing by the
MITRE Corporation under Federal EPA contract - estimated
to be completed late 1975 or early 1976.
5.0 GUARANTEE AND OPERATING PROGRAM
The 24-hour test by Enviro-Chem was conducted to assure that the
Cat-Ox System can meet the following guarantees:
1. The system is capable of operating with a gas flow of 1,120,000
#/hour entering the system at 310°F.
2. The system is capable of producing 60° Baume (77.7% H2S04)
sulfuric acid.
3. The exit gas emitted to the stack does not contain, on the
average, more than 1.0 milligram of 100% sulfuric acid mist
per actual cubic foot of gas when the system is operated at
rated capacity.
4. The conversion of S02 to SO^j of the gas entering the converter
shall be at least 90% of rated capacity.
5. The system shall operate so that over 99% of the fly ash in the
flue gas leaving the boiler is removed when operating at rated
capacity.
6. The system shall remove 85% of the S02 in the flue gas entering
the system.
Results indicate the removal of S02 entering the system to be 92.1%
during the period of the performance guarantee tests and 91.4% over a
monitoring period exceeding 24 hours. The sulfuric acid produced was in
excess of 78% concentration, the exit gas to the stack contained 0.25 -
0.53 milligrams 100% sulfuric acid mist per actual cubic foot of gas,
and the fly ash removal across the Cat-Ox System alone was 99.5%.
Following repairs and modifications, the unit was started on August
14, 1974 and Illinois Power Company is attempting to operate the Cat-Ox
System for a minimum of 15 months as mentioned previously. During this
period, and for a subsequent period of four years, if Illinois Power
Company decides to continue to operate the System, data will be obtained
to evaluate the following items :
770
-------
1. Operating characteristics and plant performance (relative to
SC>2 and fly ash removals and to H.2S04 recovery).
2. Maintenance procedures, requirements, and costs.
3. Total process operating costs.
The photographs in Figures 2 and 3 show the Cat-Ox installation.
6.0 SUMMARY
In an attempt to advance the frontiers of knowledge in the science
of removing sulfur dioxide from flue gas, Illinois Power, with Battelle
Institute, conducted a comprehensive survey of possible sulfur removal
systems. The possibility of using low sulfur Western coal was considered,
but, in addition to increased costs, it was found to be less efficient
than local coal (lower Btu, higher ash and moisture), and precipitator
efficiencies declined with its use, thereby causing a violation of
particulate regulations. As a result of these studies, Illinois Power
Company decided that the Cat-Ox System, after both a pilot installation
and prototype installation, was most nearly ready for commercial demon-
stration. In 1970, the capital costs being considered for Cat-Ox were
much higher than those proposed for other S02 control systems. However,
as more actual experience has been gained since 1970, it seems that the
capital costs originally estimated by Enviro-Chem for Cat-Ox were more
realistic than those estimated for other systems, primarily because of
the decade of research and the advanced stage of development of the
Cat-Ox System.
From our limited experience in operating the Cat-Ox, it is evident
that the System can remove at least 85% of the sulfur dioxide from the
flue gas as 78% sulfuric acid and can remove essentially all of the fly
ash. Sale of the sulfuric acid may offset some of the operating costs.
Illinois Power Company, by removing the pollutants from the environment
and conserving natural resources by recovering a product which is pres-
ently being thrown away, is not solving one pollution problem and
creating another one. While numerous start-up difficulties have been
encountered with the Cat-Ox System, we are still hopeful it will prove
to be a feasible method.
ACKNOWLEDGMENT
We hereby express grateful acknowledgment to the Control Systems
Laboratory of the Office of Research and Development of the Federal
Environmental Protection Agency for their financial assistance and to
Monsanto Enviro-Chem Systems, Inc. for their technical competence.
771
-------
FIGURE 2
REHEAT CAT-OX PROCESS, I LLINOIS POWER,
WOOD RIVER UNIT NO. 4
772
-------
FIGURE 3
REHEAT CAT OX INSTALLATION SHOWING CONVERTER, HEAT EXCHANGER,
ABSORBING TOWER, AND MIST ELIMINATOR
773
-------
PART II
MITRE TEST SUPPORT FOR THE "CAT-OX" PROJECT
1.0 BACKGROUND
MITRE was contracted by the Environmental Protection Agency in
April, 1971, to provide technical support for the Cat-Ox Demonstration
Program. In general, support consisted of the following: (1) program
management assistance, (2) test plan development, (3) design, development,
and installation of measurement systems, (4) service related to the per-
formance' of test programs, and (5) evaluation of test results.
A management plan was developed which blueprints the overall pro-
gram for support of the Cat-Ox demonstration. The details of the plan
are described in Management Plan for Test Support for the Cat-Ox
Demonstration Program, MTR 6054, July, 1971. The major tasks defined
for the program are shown in Figure 4.
2.0 PROGRAM SUMMARY
The first task area involved a definition of test requirements in
which MITRE evaluated the needs of potential users and defined a set of
test conditions which were reflected into the baseline measurement pro-
gram. The test conditions were effectively limited in actual practice
by the limiting capabilities of the Unit 4 steam generator with which
the Cat-Ox process was integrated.
The second task area involved a characterization of the Unit 4
steam generator prior to installation of the Cat-Ox process with regard
to operating conditions for the steam generator, flue gas properties at
the steam generator/Cat-Ox process interface, pollutant emissions includ-
ing S02 and particulates, and the evaluation of instrumentation and
measurement procedures to be used during the one-year demonstration pro-
gram.
The third task involves a one-year demonstration test during, which
a measurement program will be conducted to evaluate the Cat-Ox process
control performance, process maintainability, process availability/reli-
ability, and process operating costs. The start of this task has been
delayed for approximately two years while the reheat burners of the
Cat-Ox were being tested and redesigned.
In the interim, a special test program on the electrostatic pre-
cipitator (ESP) was performed to determine overall ESP efficiency and
efficiency as a function of particle size. In this particular test
series precipitator operating conditions were varied as well as steam
generator operating conditions. In addition, specialized measurements
and analyses such as in-situ resistivity, fly ash composition and gaseous
803 concentrations were performed to determine their effects on ESP per-
formance. The results of these tests have been incorporated into a draft
report which is currently being reviewed by EPA.
774
-------
TASK 1
TASK 2
TASK 3
TASK 4
DEFINE TEST REQUIREMENTS
BASELINE MEASUREMENT PROGRAM
CAT-OX DEMONSTRATION
-PRECIPITATOR TEST PROGRAM
-TEST PREPARATION
-ONE-YEAR MEASUREMENT PROGRAM
EVALUATION OF RESULTS
FIGURE 4
PROGRAM SUMMARY
COMPLETED 1971
COMPLETED 1971
INTERIM TESTS 1973
COMPLETED 1974
SEPTEMBER 1974 -
AUGUST 1975
SEPTEMBER 1975 -
LATE 1975 OR
EARLY 1976
775
-------
At this time the Cat-Ox process is still experiencing difficulties
with the redesigned reheat burners. MITRE has, therefore, initiated its
test program with a block of tests on the ESP to determine if its per-
formance has changed significantly in the past year. These tests have
been completed and the results are being analyzed. It is now expected
that start up of the Cat-Ox process will be initiated in late October, 1974.
The final task will be to evaluate the results of the one-year test
program in terms of performance, maintainability, availability/reliability
and operating costs.
3.0 ONE-YEAR TEST PROGRAM OBJECTIVES
Major areas which will be evaluated during the one-year test pro-
gram are:
o Process performance
o Process maintainability
o Process availability
o Operating costs
Specific tests will be designed and performed to evaluate process per-
formance and maintainability, whereas process availability and operating
costs will be determined mostly from accurate record keeping.
Performance tests will be performed at particular periods of time
to determine if the process is performing as specified. Maintainability
tests will be conducted to measure changes in performance with time; to
measure the effects of specified maintenance procedures and operating
conditions on process performance; and to identify additional or modified
maintenance procedures required to achieve design performance. Process
availability will be determined by measuring the operating hours of the
steam generator and the Cat-Ox process and calculating percentage avail-
abilities relative to scheduled operation. Operating costs will take
into consideration normal operational support, required maintenance,
utilities for process operation, steam generator power loss due to process
failure, capitalization, and by-product marketability.
Particular areas of concern are as follows:
o Operating characteristics and process performance
relative to S02 and fly ash removal, H^SO^ recovery,
and mist/gaseous emissions
o Response of process to fuels (varying sulfur content),
load, excess air, and soot blowing
o Precipitator efficiency and outlet grain loading
776
-------
o Thermal efficiency of regenerative heat exchanger
and seal leakage with time
o Converter efficiency, longevity of catalyst and
catalyst make-up
o Mist eliminator performance and required frequency
of washing
o Corrosion rates of process materials and components
The program by which these results will be achieved is discussed in
greater detail in subsequent paragraphs.
3.1 Cat-Ox Process and Measurement Locations
Figure 5 shows the flow diagram of the steam generator and the
reheat Cat-Ox process. The dashed line shown in the Figure separates
the steam generator from the process. The Cat-Ox process has been
installed between the existing I.D. fans and the stack. The electro-
static precipitator is shown as part of the process because it is
essential to the process in order to minimize contamination of the cat-
alyst in the converter.
Flue gas from the boiler passes through the economizer and the air
heater, and then enters the mechanical collector where particulates are
initially removed. Particulates are further reduced to a very low level
by the electrostatic precipitator. During the course of this flow the gas
temperature has dropped approximately 400°F. and therefore must be re-
heated prior to entering the converter. Reheating of the flue gas is
accomplished by the external reheat burner and the heat exchanger. The
reheat burner injects heated flue gas into the main flue gas stream by
means of ducts A and B at approximately 25% and 75% of the total burner
output respectively. A portion of the main flue gas stream is recycled
to temper the heated gas being emitted directly from the burner. This
new design using an external burner replaces the original in-line burners
which were not suitable for long-term operation using no. 2 fuel oil
because of possible excessive catalyst contamination.
Flue gas is then ducted to the converter where the S02 is catalytically
oxidized to SO^. The flue gas with the 803 constituent is redirected to
the heat exchanger so that it is cooled prior to entering the absorbing
tower where the 803 combines with the water vapor and is absorbed into
the recirculating acid. The mist eliminator removes the sulfuric acid
mist which escapes from the absorbing tower with the flue gas. A second
induced draft (I.D.) fan is required to make up pressure drop losses and
to restore flow to the stack. Sulfuric acid from the absorbing tower and
mist eliminator are pumped, cooled and recirculated through the absorbing
tower. A small fraction of the total acid is periodically drawn off,
cooled and pumped to the acid storage tanks.
777
-------
A'" *=£> ^ADN
J
bXISIINU ,
in eiMC >
Ik
r MECHANICAL S* A|R HEATER /T^j ECONOMIZER (
^ COLLECTOR \^ V^
DUCT A
«. • \^
t
(Q> BOILER
K
1
p;
(R) ^ BURNER
^ . DUCT B ^AIR
\S 1 1
~ * K. . * \/ K. or*fci\/co-rco
\ BY-PASS
V TOST/
^CK
ABSORBING "N. MIST
TOWER _-/ ELIMINATOR
•'v^ vT ^ ^
^ ^ ^\\\\\\SS\\\\v
^ s
•w^^^^\^^^^\^^^^^^\\^^\\^^ ^
1
, ... x
^x:^^^^^^^^^^NN^s^s^^^^^^
^^
§ ACID CIRCULATION srjwA. rmrui ATION <7?S<& PRODUCT ACID TjrrA PRODUCT ACID <^rrr
$; PUMPS iii^ CIRC ULATION ^^ COOLER ^^ COOLER S$£>
^\ r U(Jv/Lttt I^Cl
§ ^
^x^N^^c<^\\\\\\^\\\\\^\\\\^\^^^
FIGURES
STEAM GENERATOR & CAT-OX PROCESS
-------
The numbers shown in the flow diagram identify the locations at
which measurements will be made. Points I1, 2', and 14 are locations
where measurements were made during the baseline program. Measurements
will be repeated at these three locations to control test conditions, to
permit correlation with the baseline results and to reevaluate improve-
ments in the operational efficiency of the mechanical collector. The
other measurement points identified are pertinent to evaluation of the
overall Cat-Ox process and major subsystems of the process. The circular
dots shown in the flow diagram identify the locations where corrosion
coupons will be placed.
3. 2 Discussion of Major Areas of Investigation
The Cat-Ox process has been designed to satisfy specified levels of
performance for both the overall system and the major subsystems. In
addition, maintenance procedures have been identified for some of the
critical subsystems such as the converter. The test program which has
been designed will measure the actual process performance and compare
the results to the manufacturers performance specifications.
3.2.1 Design Performance Testing
Design performance testing will consist of two general areas,
testing of the overall Cat-Ox process and testing of the major subsystems.
The major subsystems which will be evaluated are the electrostatic precipitator,
the gas heat exchanger, the converter, the absorbing tower and the mist
eliminator. Where practical, testing of the overall process and various
subsystems will be combined.
Figure 6 shows the test objectives of the overall Cat-Ox
process with the corresponding measurement parameters and the particular
measurement locations as identified previously in the process flow dia-
gram. The measurement methods and instrumentation were reported in the
Proceedings of the Flue Gas Desulfurization Symposium held in May, 1973, and
will not be discussed in this paper.
Of particular interest is the overall S02 conversion
efficiency which will be obtained by measuring the mass flow of S02 at
the input and output of the overall process, points 1 and 14 respectively.
In addition, the S02 contribution from the external burner will be deter-
mined by measuring the net flow, i.e., the output of the burner less the
amount which is recycled. The S02 contribution of the external burner
is extremely small. The overall process conversion efficiency is con-
siderably influenced by the leakage of the heat exchanger which will be
discussed further in this section.
The H2SO^ mist concentration and mass flow will be measured
in the stack because this is an important pollutant, and because the
process by its nature can act as an emission source. In addition, the
rate of formation of sulfuric acid in the absorbing tower will be measured
directly by observation of acid level as a function of time. Based
779
-------
OBJECTIVE
OVERALL S02 CONVERSION EFFICIENCY
H2S04 HIST CONCENTRATION
BY-PRODUCT GENERATION
SULFUR BALANCE
PARTICULATE REMOVAL EFFICIENCY,
GRAIN LOADING 4 MASS FLOW RATE
CASEOUS EMISSIONS
MEASURED PARAMETERS
MASS FLOW S02: GAS CONC.;
GAS VOLUME FLOW
MASS FLOW H2S04 MIST/S03: HIST CONC.;
GAS VOL. FLOW
RATE OF H2S04 FORMATION: SPECIFIC GRAV., TEMP.;
ACID LEVEL
MASS FLOW SO?: GAS CONC.; CAS VOL. FLOW
MASS FLOW S03: GAS CONC.; GAS VOL. FLOW
RATE OF H2SOi, FORMATION: SPEC. GRAV., TEMP.; ACID LEVEL
MASS FLOW S02, H2S04 MIST/S03: S02, MIST CONC.; CAS VOL. FLOW
MASS FLOW PART.: MASS DENSITY; CAS
VOL. FLOW
MASS FLOW N02, NOX, THC, C02: GAS CONC.;
CAS VOL. FLOW
LOCATIONS
INPUT: PTS. 1, 9, R
OUTPUT: PT. 14
OUTPUT: PT. 14
ABSORBING TOWER (STORAGE TANK)
INPUT: PTS. 1, 9, R
OUTPUT: ABSORBING TOWER
(STORAGE TANK)
INPUT: PT. 1
OUTPUT: PT. 14
OUTPUT: PT. 14
FIGURE 6
OVERALL CAT-OX PROCESS PERFORMANCE
-------
on these series of measurements which will be 'conducted under steady
state conditions within a normal test day, a sulfur balance will be
performed to check the overall accuracy of the results.
Particulate is removed from the flue gas stream prior to
entering the Cat-Ox process primarily to minimize contamination of the
catalyst and, therefore, to reduce the number of times the catalyst other-
wise would have to be cleaned. Also, particulate collects in the heat
exchanger and the mist eliminator requiring that both be periodically
cleaned, and, furthermore, contaminates the sulfuric acid. Therefore,
the particulate removal efficiency of the ESP is significantly important
for proper functioning of the process. The particulate removal efficiency
for the overall process will be higher than that for the precipitator
itself because of the additional filtering inadvertently provided by the
several subsystems just identified. The overall particulate removal
efficiency will be measured by comparing the particulate loading in the
stack to the loading at the input to the precipitator.
Finally, all the polluting gaseous emissions being emitted
at the stack other than those for which the process has been designed to
control will be measured. These are nitrogen oxide, nitrogen dioxide,
total hydrocarbons, and carbon dioxide. In addition, oxygen and water
vapor will be measured for purposes of performing necessary mass flow
calculations.
Figure 7 shows the test objectives of the major subsystems.
As regards the electrostatic precipitator, measurements of mass loading
will be performed at the inlet (point 1) and outlet (point 3) to deter-
mine efficiency, grain loading and mass flow rate.
The heat exchanger transfers a substantial quantity of energy
stored in the hot gas exiting from the converter to the lower temperature
flue gas from the electrostatic precipitator. However, due to the nature of
the heat exchanger design, a rotary Ljungstrom, approximately 5.2% of the
flue gas leaks across the heat exchanger and is, therefore, not processed
by the converter. This leakage reduces the overall process 862 con-
version efficiency by slightly less than the amount of the leakage to
approximately 85.3%. The gas leakage of the heat exchanger is specified
not to exceed 5.2% provided the pressure drop from the input cold side
to the output hot side does not exceed a specified level. The gas leak-
age will be determined by measuring the mass flow of S02 at the inputs
and outputs of the heat exchanger. The static pressure drop will be
measured at the same time to assure that it complies with specification.
In addition, the mass flow of 62 and C02 will be measured at the inputs
and outputs to determine if there are any air leaks in the heat exchanger.
A heat balance will be performed for the heat exchanger by
measuring the gas concentrations, the gas volume flow, and the temper-
atures of the gases at the inlets and outlets. The specific heats of
the various gases will, of course, be required to perform the calculation.
By applying the data on gas leakage, the thermal efficiency of the heat
exchanger will be computed.
781
-------
OBJECTIVE
MEASURED PARAMETERS
LOCATIONS
ELECTROSTATIC PRECIPITATOR
PRECIPITATOR EFFICIENCY,
GRAIN LOADING &
MASS FLOW RATE
MASS FLOW OF PART.: GRAIN LOADING;
GAS VOL. FLOW
INPUT: PT. 1
OUTPUT: PT 3
GAS HEAT EXCHANGER
GAS LEAKAGE
00
IvJ
HEAT BALANCE &
THERMAL EFFICIENCY
COHVERTER
CONVERSION EFFICIENCY
MASS FLOW OF S02: GAS CONCS.;
GAS VOL. FLOW
MASS FLOW OF 02, C02: GAS CONCS;
GAS VOL. FLOW
STATIC PRESSURE DROP
MASS FLOW OF GAS & ENERGY: GAS CONCS. (INCL.S03);
GAS VOL. FLOW;
TEMPERATURES
MASS FLOW OF S02-' GAS CONC.;
GAS VOL. FLOW
MASS FLOW OF S03: GAS CONC.;
GAS VOL. FLOW
GAS TEMPERATURE
STATIC PRESSURE DROP
INPUT: PT 4, 8
OUTPUT: PT 5 + R, 10
INPUT: PT 4, 8
OUTPUT: PT 5 + R, 10
PT 4 - PT 10
INPUT: PT 4, 8
OUTPUT: PT 5+R, 10
INPUT: PT. 3, 4, 5, 9
OUTPUT: PT. 8
INPUT: PT 5
OUTPUT: PT 8
INPUT: PT 7
OUTPUT: PT 8
PT7 - PT8
FIGURE 7
PERFORMANCE OF MAJOR SUBSYSTEMS
-------
OBJECTIVE
ABSORBING TOWER & MIST ELIMINATOR
ACID FORMATION RATE &
SULFUR BALANCE
MIST ELIMINATOR PERFORMANCE
MEASURED PARAMETERS
MASS FLOW OF H2s°i> MIST/GASEOUS S03, S02:
GAS CONC.; GAS VOL. FLOW
ACID LEVEL VS. TIME
SPECIFIC GRAVITY AND TEMPERATURE OF ACID
MIST DENSITY
DIFFERENTIAL PRESSURE
LOCATIONS
INPUT(ABSORB.): PT 10
OUTPUT (MIST ELIM.): PT 11
ABSORBING TOWER (STORAGE TANK)
PRODUCT ACID DISCHARGE VALVE
OUTPUT (MIST ELIM.): PT 11
FIGURE 7 (CONT'D)
PERFORMANCE OF MAJOR SUBSYSTEMS
-------
The SC>2 conversion efficiency of the converter will be
measured by determining the mass flow of SC>2 at the input and output of
the converter. The mass flow of SC>2 into the converter is equal to the
sum of the flows at location 5 and the contribution from the reheat
burner through duct B. Because of limited number of access ports, the
contribution from duct B is indirectly obtained by measuring the total
S02 mass flow of the burner at point 9 and subtracting the fraction of
the flow to duct A. The flow in duct A is obtained by measuring the
difference of the flows at locations 3 and 4.
As a check on the S0£ measurements, the mass flow of 803 at
the input and output of the converter will also be measured. The 803
contribution from the external burner is expected to be negligible and
will not be measured.
The temperature of the flue gas is significant as regards
the S02 conversion efficiency and is recorded as part of the gas volume
flow measurement. In addition the static pressure across the converter
will be measured as an indication of fly ash accumulation and possible
catalyst contamination. The design specification of the converter states
that the converter efficiency is independent of fly ash accumulation.
Finally the acid formation rate of the absorbing tower will
be determined by measuring acid level as a function of time. The acid
concentration will be determined by measurement of specific gravity and
acid temperature. In addition, the acid formation rate will be deter-
mined by measurement of the mass flow of gaseous 303 and S02 at the input
to the absorbing tower and mass flow of I^SO^ mist/gaseous 863 and SC>2
at the output of the mist eliminator. The acid formation rate is deter-
mined by assuming that the difference between input to the absorbing
tower and the output of the mist eliminator goes into the formation of
sulfuric acid.
The mist eliminator performance will be based on the mea-
surement of sulfuric acid mist at the output of the mist eliminator.
The differential pressure of the mist eliminator will be measured
simultaneously as an indication of mist eliminator operating condition.
3.2.2 Maintainability Testing
As the Cat-Ox process operates over an extended period of
time, performance characteristics will change causing lower operating
efficiencies and higher emission levels for the major subsystems and the
overall process. The degradation in performance will be corrected at
certain points in time when maintenance procedures now planned are applied
to the various subsystems. Because of the lack of experience with the
Cat-Ox process, the existing maintenance procedures and the frequency of
application of these procedures can be expected to be modified during the
course of the one-year test program. In fact, it is possible that
additional maintenance procedures will be identified and applied which
will permit the process to operate at best expected performance levels.
784
-------
Modifications to maintenance procedures will be identified
as operational experience with the process is obtained and as data is
collected and evaluated.
The general objectives of maintainability testing can be
stated as follows:
(1) Measure performance changes of the major subsystems
and the overall Cat-Ox process with time based on an
optimum statistical test design. The statistical
design utilizes blocking replication and randomization
to determine the major inter-relationships of boiler
and process parameters while performing a minimum
number of tests, thereby, minimizing test costs and
interference with normal steam generator operation.
(2) Measure the effect of specified maintenance procedures
and operating conditions on process performance by
observing process performance before and after the
application of maintenance procedures and changes in
operating conditions. Measurements will be an integral
part of the statistical test design and will be observed
as part of the time dependent performance changes.
(3) Identify additional or modified maintenance procedures
required to achieve design performance or best level of
performance. Particular areas of maintenance have not
been specified in the Cat-Ox design or operating manuals.
Some of these areas will become better defined during
the course of the test program.
Specific Objectives
The specific objectives of maintainability testing are stated in
Figure 8 for each of the major subsystems. The test approach consists
of several types of measurements and observations:
(1) Changes of design performance parameters with time
using statistical testing.
(2) Observation of changes in operational levels of
particular parameters with time.
(3) Observation of changes in mechanical and electro-
mechanical properties of components and subsystems
with time.
785
-------
OBJECTIVE
MEASURED PARAMETERS
MONITORING
SUBSYSTEM OPERATING
ELECTROSTATIC PRECIPITATOR
4 OBSERVATIONS
FREQUENCY
STATUS
CHANCE OF EFFICIENCY &
OUTPUT GRAIN LOADING VS. TIME
MASS FLOW OF PART.:
GRAIN LOADING; GAS VOL. FLOW
FULL FACTORIAL
TEST BLOCKS
SCHEDULED MAINTENANCE-
BEFORE/AFTER
CHANGE OF ELECTRICAL
CHARACTERISTICS VS. TIME
PRIMARY VOLTAGE
PRIMARY CURRENT
SECONDARY CURRENT
RECORDED
PERIODICALLY
ADJUSTED AS REQUIRED
CHANGE OF ELECTRO-MECHANICAL
CHARACTERISTICS VS. TIME
INTEGRITY OF DISCHARGE
ELECTRODES
OPERATION ELECTRODE VIBRATORS,
FREQ. & INT.
OPERATION PLATE RAPPERS,
FREQ. & INT.
ASH HOPPER LEVEL: INDICATORS;
VIBRATORS; HEATERS
OBSERVED
PERIODICALLY
ON-LINE MAINTENANCE
WHEN POSSIBLE
CORROSION
MATERIALS, PT3: C-1008 C.S.,
316 SS, COR-TEN
TEMPERATURE, PT 3
MEASURED
PERIODICALLY
RECORDED PERIODICALLY
FIGURE 8
MAINTAINABILITY TESTING
-------
OBJECTIVE
EXTERNAL REHEAT BURNER
CHANGE OF FLUE GAS
TEMPERATURES VS. TIME
MEASURED PARAMETERS
S OBSERVATIONS
INPUT A: OUTLET TEMP., PT 3, A
INPUT B: OUTLET TEMP., PT 5, 8
HEATER OUTLET TEMP.
MONITORING
FREQUENCY
RECORDED PERIODICALLY
SUBSYSTEM OPERATING
STATUS
MAINTENANCE OF BURNER-
BEFORE/AFTER
CHANGE OF BURNER
INPUTS WITH TIME
CAS HEAT EXCHANGER
CHANGE or HEAT BALANCE &
THERMAL EFFICIENCY
WITH TIME
NATURAL GAS FLOW
NO. 2 OIL FLOW
COMBUSTION AIR FLOW
AMBIENT TEMPERATURE
MASS FLOW OF GAS & ENERGY: GAS CONCS.
(S03 INCL.) GAS VOL. FLOW;
TEMPERATURES
RECORDED PERIODICALLY
FRACTIONAL
FACTORIAL
TEST BLOCKS
SOOT BLOWING-
BEFORE/AFTER
WASHING -
BEFORE/AFTER
CHANGE OF LEAKAGE WITH
TIME
MASS FLOW OF S02: GAS CONCS.; GAS VOL. FLOW FRACTIONAL
FACTORIAL
MASS FLOW OF 02, C02: GAS CONCS.; GAS VOL. FLOW TEST BLOCKS
STATIC PRESSURE DROP
FIGURE 8 (CONT.)
MAINTAINABILITY TESTING
-------
OBJECTIVE
GAS HEAT EXCHANGER (CONT'D)
CORROSION
CONVERTER
CONVERSION EFFICIENCY VS. TIME
STATIC PRESSURE DROP VS. TIME
CORROSION
MEASURED PARAMETERS
& OBSERVATIONS
MATERIALS, PT 4: C-1008 C.S., 316 SS,
COR-TEN
TEMPERATURE, PT 4
STATIC PRESSURE, PT 4 - PT 5
MATERIALS, PT 5; C-1008 C.S., 316 SS,
COR-TEN
TEMPERATURE, PT 5
MASS FLOW OF S02, S03: GAS CONC. ;GAS VOL. FLOW
GAS TEMPERATURE
STATIC PRESSURE DROP
CATALYST CLEANING FREQUENCY
CATALYST LOSS
DOWN TIME & MANLOADING
STATIC PRESSURE DROP
FUEL CONSUMPTION; FUEL TYPE
MATERIALS, PT 8: C-1008 C.S., 316 SS,
COR-TEN
TEMPERATURE, PT 8
MONITORING
FREQUENCY
MEASURED PERIODICALLY
RECORDED PERIODICALLY
RECORDED PERIODICALLY
MEASURED PERIODICALLY
RECORDED PERIODICALLY
FRACTIONAL FACTORIAL
TEST BLOCKS
RECORDED CONTINUOUSLY
RECORDED CONTINUOUSLY
RECORDED AS REQUIRED
RECORDED AS REQUIRED
RECORDED AS REQUIRED
RECORDED PERIODICALLY
RECORDED PERIODICALLY
MEASURED PERIODICALLY
RECORDED PERIODICALLY
SUBSYSTEM OPERATING
STATUS
CATALYST CLEANING
BEFORE/AFTER
FIGURE 8 (CONT'D)
MAINTAINABILITY TESTING
-------
OBJECTIVE
ABSORBING TOWER AND ACID LOOP
PRODUCTION OF ACID OVER
EXTENDED PERIOD OF TIME
MEASURED PARAMETERS
4 OBSERVATIONS
CONDITION OF ACID CIRCULATION
PUMPS VS. TIME
CONDITION OF ACID CIRCULATION
AND PRODUCT ACID COOLERS VS.
TIME
CONDITION OF ACID PRODUCT PUMPS
VS. TIME
MASS FLOW OF H2SOil VAPOR, S02 GAS-PT 10
ACID LEVEL OF ABSORBING TOWER
ACID LEVEL OF STORAGE TANK
SPECIFIC GRAVITY AND TEMP. OF ACID
PRESSURE DROP ACROSS ABSORBING TOWER
FUEL CONSUMPTION RATE, FUEL TYPE, LOAD
LEVEL
ACID FLOW
TEMP. ACID TO CIR. COOLERS
TEMP. ACID TO PRODUCT COOLER
TEMP. ACID FROM PRODUCT COOLER
TEMP. WATER TO ACID COOLERS
TEMP. COOLING WATER RETURN
PH COOLING WATER RETURN
TEMP. FLUE GAS FROM ABSORBING TOWER
DISCHARGE VALVE TO STORAGE
ACID LEVEL IN STORAGE TANK
MONITORING
FREQUENCY
RANDOMIZED COMPLETE
BLOCKS
PERIODICALLY RECORDED
PERIODICALLY RECORDED
PERIODICALLY RECORDED
PERIODICALLY RECORDED
PERIODICALLY RECORDED
PERIODICALLY RECORDED
RECORDED PERIODICALLY
SUBSYSTEM OPERATING
STATUS
CONDITION OF PACKING;
LEVEL OF PACKING;
CONDITION ACID DISTRIBUTOR
RECORDED PERIODICALLY
FIGURE 8 (CONT'D)
MAINTAINABILITY TESTING
-------
OBJECTIVE
ABSORBING TOWER AMD ACID LOOP (CONT'D)
CORROSION
MIST ELIMINATOR
HjSO,; MIST DENSITY VS. TIME
MEASURED PARAMETERS
& OBSERVATIONS
MATERIALS, PT 10; C-1008 C.S.,
316 SS, COR-TEN, CARP. 20,
ARMCO 22-13-5
TEMPERATURE, PT 10
MATERIALS, INSIDE ABSORBING TOWER:C-1008C.S.,
316 SS. COR-TEN, CARP. 20, ARMSCO 22-13-5,
CHEMICAL LEAD, USS T-l
MATERIALS, AGIO FROM ABSORB. TOWER:
C-1008 C.S., 316 SS, COR-TEN, CARP. 20,
ARMCO 22-13-5, CHEMICAL LEAD
MATERIALS, ACID FROM PRODUCT COOLER:
C-1008 C.S., 316 SS, COR-TEN, CARP.20,
ARMCO 22-13-5, CHEMICAL LEAD
H2S04 MIST DENSITY - PT 11
MONITORING
FREQUENCY
MEASURED PERIODICALLY
PERIODICALLY RECORDED
AT PROCESS
SHUTDOWN
AT PROCESS
SHUTDOWN
MEASURED
PERIODICALLY
RANDOM! ZtD COMPLETE
BLOCKS
SUBSYSTEM OPERATING
STATUS
REPLACEMENT FIBER PACKED
ELEMENTS - BEFORE/AFTER
WASHING - BEFORE/AFTER
FIGURE 8 (CONT.'D)
MAINTAINABILITY TESTING
-------
OBJECTIVE
HIST ELIMINATOR (CONT'D)
MIST ELIMINATOR WASHING
MEASURED PARAMETERS
S OBSERVATIONS
WASHING FREQUENCY:
DIFFERENTIAL
PRESSURE
SUBSYSTEM OPERATING
STATUS
RECORDED PERIODICALLY
CORROSION
I. D. FAN
GAS VOL. FLOW AND STATIC
PRESSURE VS. TIME
MATERIALS, PT. 11 - C-1008 C.S.
316 SS, COR-TEN, CARP. 20,
ARMCO 22-13-5. USS T-l
TEMPERATURE, PT. 11
GAS VOLUME FLOW: DIFF. PRES.,
STATIC PRES., AND TEMP. AT
AVAILABLE LOCATIONS
STATIC PRESSURE, PT 13
ID FAN MOTOR "A" CURRENT
ID FAN MOTOR "B" CURRENT
ID FAN SPEED
OIL COOLER OUTLET TEMP.
INBOARD BEARING TEMP.
OUTBOARD BEARING TEMP.
OIL SUMP TEMP.
MEASURED PERIODICALLY
RECORDED PERIODICALLY
AS AVAILABLE
DURING OTHER TESTING
RECORDED PERIODICALLY
REQUIRED MAINTENANCE WILL
BE RECORDED
FIGURE 8 (CONT'D)
MAINTAINABILITY TESTING
-------
Electrostatic Precipitator
The major subsystems are discussed in sequence as they appear in
the process flow. The electrostatic precipitator is considered part of
the process because the output grain loading will determine maintenance
requirements for most of the subsystems which follow.
The most significant parameter of the precipitator is the efficiency,
99.67%, which determines the output grain loading of .005 Grains/SCF at
32°F. for an assumed input grain loading of 1.5 Grains/SCF. The
efficiency and output grain loading will change with time because the
electrical and electro-mechanical characteristics of the precipitator
change with time. Therefore, in addition to monitoring mass flow at the
input and output of the precipitator, electrical operating settings for
each of the transformer test sets will be monitored for primary voltage,
primary current and secondary current to observe unusual changes or
gradual drifting of these parameters. These changes, if significant in
magnitude, will be correlated with changes in efficiency and grain loading.
In addition, any mechanical and electro-mechanical changes and failures
will be noted for those cases where diagnosis is feasible. Of particular
interest are the mechanical integrity of the discharge electrodes,
changes in the frequency and intensity of the electrode vibrators, changes
in the frequency and intensity of the plate rappers, and changes in the
operation of hopper vibrators, heaters and indicators.
Electrical parameters will be manually recorded periodically and
electro-mechanical/mechanical characteristics will be observed during
each test.
Corrosion rates at the output of the precipitator will be measured
for materials used in the construction of the process. The temperature
of the flue gas will be monitored periodically for correlation with
corrosion.
At the present time no regular maintenance is scheduled for the ESP.
It is possible that the ESP will degrade below design performance levels
and operate at a degraded level for a considerable period of time before
the process is shutdown for maintenance. Degraded operation will
accelerate plugging of the converter bed and may force more frequent
cleaning of the catalyst. Of course, many causes of ESP degradation will
be reparable on line and these will be fixed as quickly as is possible
by power plant personnel.
External Reheat Burner
The temperatures of the flue gas at insertion and before/after
insertion of hot gases into the main stream are important parameters for
evaluating the reheat burner performance as a function of time. At the
insertion point from duct A, the burner raises the temperature from 310°F.
to 350°F. to prevent cold corrosion at the input to the heat exchanger.
At the insertion point from duct B, after the heat exchanger, the flue
792
-------
gas temperature is raised from 750°F. to 850°F. at which temperature
S02 conversion occurs efficiently. These temperatures will be recorded
periodically as an indication of reheat burner performance.
In addition to flue gas temperatures, the fuel flows and combustion
air flows will be monitored to determine the relationship of flue gas
temperatures to energy consumption. Energy consumption will vary
depending on ambient air temperatures which, therefore, will also be
recorded.
The process is designed to operate over the ambient temperature
range from -10°F. to 110°F. The control system of the burner will
normally compensate for seasonal changes by increasing or decreasing
fuel flows. These measurements will demonstrate the capability of the
burner to operate over the ambient temperature range of the process.
Gas Heat Exchanger
Observations will be made of the changes in heat balance, thermal
efficiency and gas leakage as a function of time using fractional
factorial testing. Also a special test will be performed to check the
effects of soot blowing by measuring thermal efficiency and gas leakage
before and after soot blowing is performed. Soot blowing is now specified
to occur at a frequency of once per week.
Washing frequency of the heat exchanger has not been specified.
Observations of thermal efficiency and leakage over the period of
factorial testing may indicate a desirable washing frequency. However,
to assure a measure of the effect of washing, the heat exchanger will
be washed during the test series and the effect on thermal efficiency and
gas leakage will be observed. Washing of the heat exchanger will be
selected to coincide with a time period when the process is shutdown for
catalyst cleaning.
Corrosion rates will be measured at the input to the cold side of
the heat exchanger and will provide a measure of the actual corrosion
of the heat exchanger. Corrosion rates will also be measured at the
output of the cold side of the heat exchanger where gas temperatures
are higher and corrosion is expected to be lower. Temperature and static
pressure will be recorded periodically at both of these locations.
Converter
Changes of conversion efficiency with time will be determined
periodically on the basis of fractional factorial testing. It will be
of particular interest to observe the effect of catalyst cleaning on
efficiency. Therefore, fractional testing will be designed to include
a measure of efficiency over one full time period including catalyst cleaning.
The temperature of the flue gas and the pressure drop across the
converter will also be measured and related as conditions of converter
efficiency.
793
-------
Fly ash will accumulate in the converter as time progresses
until the pressure drop across the converter reaches a level at which
it has been specified that no further pressure build-up be allowed and
that the catalyst bed be cleaned. Therefore, the pressure build-up
across the catalyst will be monitored over time as a measure of the
fly ash accumulation. The conversion efficiency will be plotted as a
function of pressure build-up in order to verify as is specified,
that conversion efficiency is independent of fly ash accumulation.
The fly ash input to the converter will also be determined by keeping
a continuous record of the quantity of fuel consumed by the boiler, the
type of fuel consumed, the loads at which the boiler is operated, and the
operating condition of the electrostatic precipitator. The pressure drop
will be plotted as a function of the integrated fly ash input to provide
an indication of the rate of fly ash bxiild-up, and to determine the approx-
imate quantity of fuel which must be consumed before the critical pressure
drop is achieved. This information could then be used to predict when
the next cleaning will be required based on assumptions of fuel con-
sumption and precipitator performance.
It is specified that cleaning of the catalyst will be required no
more than once every three months on the average, and that cleaning will
require 3 maintenance men for 4 shifts or a total of 96 mn-hrs. A 2.5%
loss of catalyst is expected during the cleaning operation. These three
conditions of catalyst cleaning, i.e., frequency, time/manhours, and
catalyst loss, will be determined during each cleaning operation of the
one year demonstration program. As the catalyst is delivered in 50 liter
containers, the makeup loss will be determined by counting the number of
50 liter containers required to fill the catalyst bed.
The corrosion rate of several materials will be determined at the
output of the converter. Despite the high concentrations of 803 at the
output of the converter, high corrosion rates are not anticipated because
of the high temperature of the flue gas.
Absorbing Tower and Acid Loop
The product acid at full load will flow at an average rate of 12
GPM and the acid level in the bottom of the absorbing tower will be
maintained at approximately 24 inches. Acid flow will be measured
periodically using randomized complete blocks to determine changes in
flow rate with time for particular operating conditions. The rate of
formation of acid will be determined indirectly by measuring the sulfur
flow into the absorbing tower less the sulfur flow out of the mist
eliminator as discussed previously.
794
-------
Acid formation in the absorbing tower will also be measured
directly using an acid level indicator. This will be accomplished by
shutting off the product acid flow and allowing the acid to build up in
the absorbing tower to near the alarm level. During the build-up, the
acid level will be measured as a function of time. The time period of
measurement will be a fraction of a test day. In addition, the acid
formation in the storage tank will be measured over a longer period of
time. Approximately one full test day at full load should make a
measurable change in the acid level. Finally the acid level in the
storage tank will be measured over an extended period of operation and
will be correlated with fuel consumption, fuel type, and load.
The acid concentration will be measured once each shift on weekdays
and once each day on weekends by Illinois Power personnel. The acid
concentration will be determined by measurement of specific gravity and
temperature.
The condition of the absorbing tower internals will be observed
when feasible to ascertain the condition of the packed bed, the
level of the packed bed, and the condition of the acid distributor.
Inspections will be made during process shutdown, and degradations or
corrections of defects will be noted for correlation with acid forma-
tion rate.
The flue gas pressure drop in the absorbing tower will be measured
periodically as an indication of the status of absorbing tower internals.
The condition of the acid circulation pumps will be checked by
monitoring the acid flow from the acid coolers to the absorbing tower
periodically. Similarly the condition of the product acid pumps will
be checked by determining acid formation rate in the storage tank as
previously discussed.
The condition of the acid circulation system and product acid coolers
will be monitored by periodically recording acid temperatures to the
coolers and from the coolers. In addition, the temperature of the water
to and from the coolers will be monitored as will the PH of the cooling
water return as an indication of acid leakage. The flue gas temperature
at the output of the absorbing tower will be monitored as an indirect
indicator. The temperature of the flue gas is dependent on operation
of the acid coolers and absorbing tower.
Corrosion rates will be determined at the flue gas input to the
absorbing tower, a location where considerable corrosion has already
occurred. Also corrosion rates will be determined at several locations
in the acid stream, i.e., inside the absorbing tower, after the absorbing
tower, to the product cooler, and from the product cooler.
795
-------
Mist Eliminator
The sulfuric acid mist density at the output of the mist eliminator
will be measured over an extended period of time by means of randomized
complete blocks. Deterioration and replacement of fiber packed elements
will be observed when possible, and noted for correlation with test data.
The pressure drop across the mist eliminator will be recorded
periodically and mist eliminator washing requirements will be deter-
mined. Sulfuric acid mist density will be measured before and
after washing to determine the effect of washing on mist eliminator
performance.
Corrosion rates of several materials will be determined at the out-
put of the mist eliminator.
Induced Draft Fan
The gas volume flow and the static pressures will be measured
throughout the Cat-Ox process at various points during the one year test
program. These measurements will indicate the performance of the induced
draft (ID) fan as a function of time. In particular, the pressure at
the outlet of the ID fan will be continuously monitored. In addition,
motor current, fan speed, oil cooler outlet temperature, inboard bearing
temperature, outboard bearing temperature and oil sump temperature will
all be monitored to determine the operating condition of the I.D. fan.
Corrosion at the outlet of the ID fan will be measured and related
to materials used in the fan.
3.2.3 Process Availability and Operating Costs
As yet, little is known of the operational reliability of
the Cat-Ox process over an extended period of time. To obtain this
information, accurate and detailed logs on the operation of both the
steam generator and the Cat-Ox process will be maintained during the
year long demonstration program.
The operational status of the steam generator and process
and the resulting interactions will be recorded. Both the steam generator
and the Cat-Ox process will operate in one of three states, that is
normally, degraded or completely shutdown. Therefore, nine operating
states for both the boiler and process are possible. These operating
states are shown in Figure 9. Only six of the states are significant
from the viewpoint of processing S02 in the flue gas. The other three
cases are trivial because, when the boiler is out, no gas is available
for processing. These six combinations resolve into three conditions,
of process operation, full process availability, partial process
availability and no process availability. Determination of these avail-
ability statistics are of major importance for determining the practicality
of applying the Cat-Ox process.
796
-------
BOILER
X
CAl-OX.
NORMAL
DEGRADED
OUT
NORMAL
Full S02
Processing
Full
S02 Processing
DEGRADED
Partial
S02 Processing
Partial
S02 Processing
OUT
No S02
Processing
No S02
Processing
FIGURE 9
STEAM GENERATOR/PROCESS OPERATING MODES
797
-------
Data Collection and Analysis
The data which will be collected to determine process availability
is shown in Figure 10. The time period of steam generator or process
degradation will be recorded. The faulty subsystem/component which
caused degradation or shutdown will be identified and the cause of failure
will be diagnosed. Also, the recommended maintenance procedure will be
recorded.
The following information will be recorded in connection with the
maintenance performed:
(1) Types and number of Illinois Power personnel
(2) Subcontractor labor
(3) Man-hours by labor category
(4) Materials required
(5) Materials and labor costs and
(6) Total time required for repairs taking into account
delays in assigning personnel or obtaining materials
Most of this data will be obtained from logs and records covering
a full year of operation on a 24 hour/day basis. A majority of the data
will, therefore, be recorded by Illinois Power in logs and records as
follows:
(1) Operating Log
(2) Mechanical Maintenance Log
(3) Electrical Maintenance Log
(4) Shift Supervisors' Logs
(5) Cost Accumulation Record and
(6) Monthly Plant Managers Report
MITRE will also maintain, during its normal working day, a Test Director's
Log and the recorded maintainability data discussed previously.
The data which is collected by both Illinois Power and MITRE will
be analyzed to provide the following information:
(1) Total hours of steam generator/process operation in
the three states of operation, normal, degraded and
shutdown
798
-------
RECORD TIME
Degradation
Shutdown
Startup
IDENTIFY FAULTY SUBSYSTEMS/COMPONENTS
Cause of failure
Recommended fix
RECORD MAINTENANCE PERFORMED
Types and number of IP personnel
Subcontractors
Man-hours by category
Materials required
Materials and labor costs
Time required for repairs
Personnel assignment
Parts/materials availability
FIGURE 10
STEAM GENERATOR AND CAT-OX PROCESS AVAILABILITY
799
-------
(2) Percentage availability of the steam generator and
process
(3) Mean-time-to-failure of process subsystems and components
for cases where sufficient data is collected.
(4) Labor categories and -man-hours by category for the
Cat-Ox process
(5) Significant causes of process failure
(6) Spare parts requirements for the process and
(7) Redesign requirements for the process
Evaluation of Operating Costs
The factors which will be evaluated to determine Cat-Ox process
operating costs are summarized in Figure 11. Estimates of most of these
factors have been made in Part I of this paper.
Normal Cat-Ox process operation will include costs of process
operating personnel and the cost of borrowed money. The cost of borrowed
money will reflect actual costs which would be passed on in higher rates
to the consumer. Maintenance costs will include both scheduled and
unscheduled maintenance, and for the most part will be collected as
described previously for determination of process availability. Most of
the costs for utilities will result from fuel requirements for the
external reheat burner and electrical power for the process subsystems.
Water treatment, cooling water, fly ash handling, and pressurized air/steam
costs are expected to be small by comparison.
If the steam generator is forced to shutdown as a result of process
failure, the cost of power loss will be included in the operating cost.
Although the possibility of such an occurrence is not great, there is
finite probability that it could occur.
Finally the quantity of acid which is generated and sold will be
determined and the income received will be applied as a credit toward
reducing operating costs.
4.0 TEST PROGRAM DESIGN
The design performance and maintainability tests will be performed
over a range of boiler operating conditions representative of normal
boiler operation. These conditions are shown in-Figure 12. The minimum
load was established as the lower limit achievable while consistently
satisfying Illinois Power load demands. The 80 MW level was selected as
the mid-point between 60 MW and 100 MW. The low load tests will
generally be conducted at night when load demand is down.
800
-------
NORMAL OPERATION
Operating personnel
Cost of borrowed money
MAINTENANCE (Scheduled/Unscheduled)
Overall maintenance costs
Personnel costs by category - mechanical, electrical
Subcontractor costs
Parts and Materials costs (make-up catalyst, mist eliminator
wash solution)
Overhead charges
UTILITIES
Fuels - external reheat burner
Electrical power and capacity charge - all process subsystems
Water treatment - mist eliminator wash
Cooling water chlorination
Fly ash handling
Pressurized air/steam
POWER GENERATION LOSS DUE TO PROCESS FAILURE
MARKETABILITY OF ACID
Quantity generated/sold
Application/markets/transportation
FIGURE 11
EVALUATION OF OPERATING COSTS
801
-------
OPERATING CONDITION LEVEL
Load 100 MW
80 MW
60 MW
Fuel 3.6% wght. S Coal
1.8% S equiv. COAL/GAS mixture
Excess Air 4.5% 02
4.0% 02
3.5% 02
Soot Blowing Normal
None
FIGURE 12
STEAM GENERATOR OPERATING CONDITIONS
802
-------
The fuel normally burned in the Unit 4 steam generator is a
Southern Illinois coal with sulfur content of approximately 3.6% by
weight. A lower level of sulfur equivalent to 1.8% S wght. will be
obtained by burning a mixture of the 3.6% sulfur coal with natural gas.
Natural gas is usually available during the late spring, summer, and
early fall months. Low sulfur tests, therefore, will only be conducted
during the early part of the program which was initiated in September
and then again toward the end of the program starting around the end of
April.
The majority of the tests will be conducted without soot blowing.
When soot blowing is employed, the normal soot blowing cycle will be
conducted including full cycling of both the wall and retractable blowers.
The range of excess air from 3.5% to 4.5% C>2 is the maximum
practical for the loads which will be employed. Therefore, these con-
ditions of excess air will be used, the two extreme levels and one at
mid-point.
Because of the number of operating conditions and the several
levels for each operating condition, a large number of test condition
combinations are possible. It is not practical to test each subsystem
for all of these combinations. Therefore, the number of operating con-
ditions used to evaluate a particular process subsystem have been
restricted to those which will significantly effect subsystem performance.
Furthermore, the overall test design has been based on statistical con-
siderations permitting the significant results to be obtained while
conducting a small number of tests.
Figure 13 outlines the test program design. The program was initiated
on 11 September 1974 with a block of 12 tests on the electrostatic pre-
cipitator. Precipitator performance was evaluated as a function of load,
fuel and soot blowing. This first block of tests has now been completed
and the data is being evaluated.
Testing of the precipitator was performed first so as to get the
test program started even though the Cat-Ox process was experiencing
start-up difficulties and was not yet on line. It was hoped that by
the time the precipitator tests were completed the Cat-Ox process would
be fully operating and the remainder of the test program could be carried
out.
If gas is still available after the process does start up a special
block of 4 tests will be conducted to evaluate the converter performance
as a function of fuel. After which a series of 9 tests divided into 3
blocks of 3 tests each will be performed at approximately 3 week intervals
to evaluate the converter and heat exchanger as a function of load and
excess air, and to evaluate time dependent changes resulting from
catalyst contamination and heat exchanger seal leakage. A special one
day test will be performed to evaluate heat exchanger performance
resulting from soot blowing of the heat exchanger. This test will be
repeated again approximately 6 weeks later to verify results of the first
test.
803
-------
SUBSYSTEM
Electrostatic Precipitator
No testing (Cat-Ox Process
Start-up)
Converter
Converter, Heat Exchanger
Heat Exchanger Soot Blowing
No Testing
Absorbing Tower
Heat Exchanger Soot Blowing
Converter, Heat Exchanger
No Testing
Mist Eliminator Wash
Absorbing Tower
Converter, Heat Exchanger
No Testing -
Steam Generator and
Cat-Ox Process
Maintenance (Catalyst
cleaning, H.E. Wash)
Converter, Heat Exchanger
Mist Eliminator Wash
Absorbing Tower
No Testing
Mechanical Collector,
Precipitator and Overall
System (Part.)
TIME PERIOD
(WKS)
2 1/2
2
2
2
2
2
1
1
5
2
2
1
NO. OF
TESTS
12
4
3
1
3
1
3
1
3
3
3
1
3
3
DESIGN
Full Factorial
Special
Frac.Fact.I
Special
Rand . Compl .
Block
Special
Frac.Fact.II
Special
Rand.Conipl.
Block
Frac. Fact. Ill
Frac.Fact.I
Special
Rand. Compl.
Block
Special
STEAM GENERATOR
VARIABLES
Load, fuel,
soot blowing
Fuel
Load, Excess Air
Load
Load, Excess Air
Load
Load, Excess Air
Load, Excess Air
Load
Load
FIGURE 13
SUMMARY OF TEST PROGRAM DESIGN
804
-------
oo
o
SUBSYSTEM
Converter, Heat Exchanger
No Testing
Converter, Heat Exchanger
No Testing
Overall System
No Testing
Overall System
No Testing
Overall System
TIME PERIOD
(WKS)
1
3
1
2
1 1/2
2 1/2
2
2
1 1/2
NO. OF
TESTS
3
3
6
6
6
DESIGN
Frac.Fact. II
Frac. Fact. Ill
Frac.Fact. I
Frac. Fact. II
Frac. Fact. Ill
STEAM GENERATOR
VARIABLES
Load, Excess Air
Load, Excess Air
Load, Fuel, Excess
Air
Load, Fuel, Excess
Air
Load, Fuel, Excess
Air
FIGURE 13
SUMMARY OF TEST PROGRAM DESIGN (CONT.*D)
-------
The absorbing tower performance will be evaluated as a function of
load only. Subsequently a second block of the same tests will be
repeated three weeks later to observe time dependent changes which may
have occurred and provide needed replicates. A special one day test will
be performed to evaluate the effects of mist eliminator washing.
A scheduled outage of the steam generator and Cat-Ox process is
planned during January, 1975. During this outage, maintenance will be
performed on both the steam generator and the Cat-Ox process. In partic-
ular, the catalyst in the converter will be cleaned and the heat
exchanger will be washed. The effect of these two maintenance procedures
will be measured by performing a second series of tests on the converter
and heat exchanger consisting of a series of 9 tests in 3 blocks of 3
tests each. In addition, the absorbing tower tests will be repeated a
third time to observe any improvement which may have resulted from
maintenance performed on the absorbing tower.
A special block of tests will be performed to determine the efficiency
of the mechanical collector, any changes in performance of the precipi-
tator since the start of the program, and the overall Cat-Ox process
particulate removal efficiency.
The final series of tests will involve the overall system performance
as a function of load, fuel and excess air. This series will consist of
18 tests divided into 3 blocks of 6 tests each separated by approximately
2 weeks to obtain time dependent changes on the overall process per-
formance.
ACKNOWLEDGEMENT
The MITRE test support program is sponsored by the Control Systems
Laboratory, Office of Research and Monitoring, U. S. Environmental
Protection Agency. The MITRE team working on various aspects of the pro-
gram consists of Dr. B. Baratz, E. M. Jamgochian, R. Reale, J. Verhoeff,
and A. Wallo. Additional support has been provided by C. R. Simcox
and P. J. Schneider of Consultants and Designers. J. Verhoeff has been
responsible for the statistical design considerations of the test program.
Earlier support in the formulation and direction of the project was pro-
vided by G. Erskine of MITRE.
REFERENCES
1. Proprietary term and registered trademark of Monsanto Enviro-Chem
Systems, Inc.
806
-------
The Shell Flue Gas Desulfurization Process
by
J. B. Pohlenz
UOP Process Division
Universal Oil Products Company
Des Plaines, Illinois
For presentation at the
Environmental Protection Agency
Flue Gas Desulfurization Symposium
Atlanta, Georgia, November 4-7, 1974
807
-------
THE SHELL FLUE GAS DESULFURIZATION PROCESS
by J. B. Pohlenz
UOP Process Division
Universal Oil Products Company
Des Plaines, Illinois
MECHANISM OF S02 CAPTURE AND RELEASE
In the early ]960's Shell (Shell International Petroleum)
initiated a development program to separate SO2 from flue gas
containing, along with the usual products of combustion, oxygen
and particulate matter. It was the objective that the process
should utilize a dry, selective adsorbent to avoid the complications
often characteristic of "-wet" systems and that the energy require-
ments to achieve isolation would represent but a modest fraction
of the energy transfer being effected by the flue gas generator
being serviced. A number of metal oxides have the desired
thermodynamic properties but, to be used effectively, must be
displayed with porous supports and, since the properties of
supported systems cannot be predicted from those of pure compounds,
a number of supported metal oxides were experimentally investigated
to determine which, if any, would show a high reactivity with
SO2 to the metal sulfate and which could be regenerated at the
same temperature to release the sulfur as S02•
It was found that cupric oxide had outstanding properties
for such an application in that CuO readily reacts with S02 in
the presence of oxygen at temperatures around 400°C to yield
809
-------
CuSC>4, which can be easily reduced at the same temperature to
yield the sulfur as S02; further, CuSO4 is stable to
thermal dissociation at the acceptance/regeneration temperature.
In the selection of the acceptor support, the requirements
to be met are adequate acceptance and regeneration rates coupled
•with chemical and physical stability. Activated alumina is the
preferred carrier but is attacked by 803, forming £12(804)3,
which creates strains in and -weakens the alumina structure.
But by applying stabilizing techniques to the alumina during
manufacture of the catalyst, avoiding excessive
temperatures in the processing cycle and maintaining good regen-
eration procedures, the acceptor support does not disintegrate in
use.
Effective exposure of the acceptor material to the flue
gas can be achieved with conventional reactor designs employing
moving packed beds, fixed packed beds, or fluid systems.
Reactor designs with moving solids have the advantage of providing
continuous processing but have the disadvantage of acceptor
attrition, particulate matter separation and relatively complex
solids-handling operations. This leaves a fixed, packed bed as
possessing the most suitable characteristics. However, it is not
applicable in its usual form since pressure drop would be excessive
and plugging with particulate matter would be unavoidable. A
810
-------
new fixed-bed reactor design -was developed in which the flue
gas flows through open channels alongside and in contact with
the acceptor material. With this design, pressure drop is low,
soot and fly ash pass through the channels without causing
plugging, and the CuO is so effectively presented to the 502
in the flowing gas that capture is completed in less than- one-
half second.
Summarizing, a process has been developed employing a
dry acceptor, CuO on alumina in a fixed, packed bed, to
extract sulfur oxides from flue gas containing oxygen ai
particulate matter. As the acceptor becomes loaded
with sulfur, the desulfurization efficiency decreases, and the
material is regenerated in-situ and at the same temperature
to release the sulfur as sulfur dioxide in the regeneration
off-gas. Two or more identical reactors are applied in swing
operation to provide for continuous processing of flue gas.
The 862 in the off-gas from regeneration, free from oxygen and
particulate matter, can be processed to yield liquid SC>2 or
elemental sulfur.
Sto ich iometry
The stoichiometry of the three reaction steps - oxidation,
acceptance, regeneration - serves not only as a convenient sum-
marization but can also provide a quantitative measure for
reaction efficiency. The copper in the regenerated acceptor
811
-------
exists primarily as elemental copper with small quantities as
cuprous sulfide (Cu2S) . Upon contact with oxygen in the flue
gas, the copper and CU2S react rapidly to form CuO and CuSO4 :
Cu + 1/2 02 - — CuO
CU2S + 5/2 02 - «^CuO + CuSO4
Thus, the formation of CU2S is undesirable since, on subsequent
oxidation, one-half of the copper in this form is converted to
and is unavailable to participate in the acceptance reaction:
CuO + 1/2 O + SO - *
The SO2 is reacted until the conversion of the CuO to
has proceeded to such an extent that the unconverted CuO has been
reduced to a low concentration level; the amount of S©2 which
escapes capture reaches the maximum acceptable and the partially
loaded acceptor is regenerated. This acceptance reaction is shown
graphically in Fig. ]'"' , expressing the SO2 content of the
treated gas as a function of elapsed time after flue gas is contacted
with regenerated acceptor. The SO2 "spike" at the beginning
of the cycle represents the SO2 slippage which occurs during the
first few minutes while the copper is converted to cupric oxide.
If the acceptance is continued past 120 minutes, the concavity of
the curve reverses and SO2 content of the treated gas approaches
that of the feed asymptotically.
812
-------
Thus, at practical levels of desulfurization, a portion
of the copper exists as CuO when the acceptor is regenerated -
CuO + H2 - — Cu = H2O
CuSO^j + 2 H2 - •- Cu + SC>2 + 2 H2O
The efficiency of copper utilization can r>e expressed in
terms of two parameters which, in turn, can appear in the
stoichiometric coefficients of the reaction cycle:
Basis: 1 Mole Cu
Let a =
Cu
Cu + 2 CU2S
CuSO4 _
CuSCXg + CuO
, after regeneration
, after acceptance
OXIDATION:
a Cu -
Cu2S +
5"3
\ 2
CuO
)
2 /
ACCEPTANCE:
CuO
S02 + 1/2
REGENERATION:
/-, _ a \
/3CuS04 + (1 - $ ) CuO + ( fi- a + 2) H2 - ^ a Cu + I- - j
-, _ a
Cu2S +
13-
~ a
S02 + (p-a+ 2) H20
The hydrogen required per unit of SO2 removed and released can
be expressed in the mole ratio -
813
-------
52 = (/8-c+ 2)
S02 (a- 1 -a
\ 2
The minimum requirement is that for which all the copper is present
as CuS04 going into regeneration and all the copper is present
as elemental after regeneration. In this event ,8 = a = 1 and the
H2/S02 is 2.0. The characteristics of acceptor material produced
in commercial quantities is such that values of /3 commonly range
from 0.7 - 1.0 depending on operating temperature, gas velocity
and sulfur removal required. A combination of acceptor improvements
and operating techniques almost totally suppress cuprous sulfide
formation so that real hydrogen requirements exceed theoretical
by only 10-25%.
Copper sulfate releases the accepted sulfur in the form
of S02 upon regeneration with reducing agents such as V^^i CO,
and light hydrocarbons at temperatures of approximately 400°C.
The rate of regeneration with various regenerants has been
investigated and reported (1); from this work it was determined
that hydrogen has economic advantages due to its high reactivity
with copper sulfater in fact, at temperatures above 300 C, the
reduction of CuSO^ is limited by the supply rate of hydrogen.
The regeneration reaction is illustrated graphically in Fig. 2(5)
814
-------
which shows the sulfur release rate against cumulative time after
regeneration commences.
DEMONSTRATION UNIT - SHELL'S REFINERY IN PERNIS
After extensive bench scale testing to define reaction
mechanism, kinetics and acceptor performance, a demonstration
unit was erected in 1967 in the Shell Refinery near Rotterdam.
The design of the unit was based on a kinetic model of the
parallel passage system and the concept of unit cells, a basic
module for containing the acceptor was first introduced. The nature
of the process permits such an approach, since any size of
reactor can be based on a unit cell of internals, each containing
a series of acceptor layers and gas channels. The cross-sectional
area of the cell fixes the volumetric rate of gas processed per
cell; the height (number of cells stacked in series) fixes the
extent of desulfurization and sulfur-holding capacity. An
adiabatic reactor containing a single stack of unit cells operates
under the same restraints of heat flow, temperature and composition
profiles, space time and contact time as if it were in parallel
with many other identical cell-stacks. Thus, scale-up to commercial
size can be accomplished in a single step.
In the Pernis unit, 400-600 scf/minute (600-1000 Nm3/h)
of flue gas containing 0.3- 0.3 vol. % SC>2 was isokinetically
815
-------
sampled from the main flue gas duct of a process heater fired
with heavy, high-sulfur fuel oil. The flue gas was drawn through
the reactor at 350-450°C by means of a steam ejector, and the
treated gas was returned to the main flue gas duct of the furnace.
The reactor normally contained 3.5 cu. ft. (0..1 m-3) of acceptor.
The acceptance time was varied between 0.5 and 2 hours in order to
arrive at an BC>2 removal efficiency of about 90% for gases with
various contents of SO . The regeneration of the acceptor was
carried out in a time equal to or shorter than the acceptance
time. The unit is fully automated and runs unattended.
During the approximately 20,000 operating hours of the
Pernis unit a great number of tests were run with various regen-
eration agents, with different types of copper-on-alumina acceptor
of varying copper content, and with a number of different
construction materials for the reactor internals.
The results of these tests are summarized below (6):
• Diluted hydrogen-containing gases are preferred as regeneration gas.
• An acceptor based on a stabilized alumina support will have a
life in excess of 8000 cycles (i.e., in excess of 1.5 years'
service life). The acceptor is resistant to chemical and
mechanical attrition; no fines will be formed or copper lost.
• The many chemical species present in the flue gases originating
from firing a variety of fuels (gas up to and including heavy
816
-------
asphalt) do not affect the physical and chemical stability
of the acceptor.
• The pressure drop over the reactor was low and remained constant
during all life-testing runs. A typical value for the particulate
content of the flue gas passing through the reactor is 0.12
grains/scf (300 mg/Nm ) when firing long residue. This value
increases sharply during soot-blowing.
• Corrosion rates fo± a number of metals under SFGD reactor
conditions have been established. With the proper materials of
construction service life of the reactor internals has been
estimated to exceed ]5 years.
COMMERCIAL UNIT AT SHOWA YOKKAICHI SEKIYU (SYS)
The SFGD unit at SYS was designed to effect 90% desulfuri-
zation on 125,000 Nm^/hr. of flue gas containing 2500 ppm of SC>2;
this was the combined flow of flue gas from an oil-fired boiler and
incinerated tail gas from a Glaus unit. Between design and
start-up, the Glaus unit had been permanently shut down so that
the SFGD unit processed only flue gas from the boiler. The S02
in the regeneration off-gas is reduced to elemental sulfur in a
second Glaus unit located approximately 1 kilometer away and its
tail gas is processed elsewhere. Design and actual sulfur loads
are shown in Fig. 3. A simplified process flow diagram is shown
817
-------
in Fig. 4; a photograph of the reactors is in Fig. 5.
The reactor section consists of two identical, parallel-
passage reactors, each containing 68 cell-stacks. The regeneration
of the spent acceptor is carried out with a diluted hydrogen stream
producing off-gas at an intermittent rate and containing SC>2, water vapor,
unconverted regeneration gas and whatever inerts are present with
the hydrogen. The SC>2 is co-processed in a Claus unit with
normal refinery acid gas streams and, in order not to suppress
the Claus conversion by flow fluctuations, water vapor and inerts,
the SO2 present in the regeneration off-gas is isolated in an
absorber-stripper system. The flow fluctuations of the stripper
overhead are damped by incorporating surge capacity in the bottom
of the absorber.
Two features of the process received special attention:
The flue gas valves which serve to isolate the reactors from
flue gas during regeneration and the unit cells in which the
acceptor is held in gauze envelopes in such a way as to meet the
requirements of the parallel passage design. The valves selected
are a special flapper-type design with good sealing characteristics
manufactured by Gebr. Adams, Bochum, Germany. After some
modifications, this valve has been subjected in a special test
818
-------
rig at Shell's Pernis Refinery to over 40,000 cycles (approx.
8 years' operation) at 450°C using steam as the test gas. It
was found that this design would meet the specifications.
Commercialization of the manufacture of unit cells was achieved
by Nihon Mesh Co. (Tokyo) with the assistance of Japan Gasoline
Co. (Tokyo), after a number of trials to meet the rather stringent
tolerances. Both cells and valves are available from fabricators
in the U.S.A.
The acceptor was developed by Shell nt its Amsterdam labora-
tories and the successful translation to commercial production
was completed by Ketjen Catalyst, AKZO Chemie B.V., Amsterdam,
who supplied the acceptor loading for the SYS unit.
Operating Experience at SYS (5)
The SYS unit was successfully started in August 1973
and. except for acceptor life, had successfully completed its
test runs by Sept. 10, 1973. The main operating data are
summarized in Table I and show that the performance of the reactor
section was as expected: an overall SC>2 removal efficiency of
90% and a hydrogen consumption of about 0.2 wt-% H2/wt-% S
recovered.
A close match was obtained between the predicted and
measured SC<2 concentrations in the treated flue gas (Fig. 1)
819
-------
and the predicted and measured SO2 release during regeneration
(Fig. 2) These operating results indicate that:
• The commercial acceptor has the required and expected SC>2
removal activity and capacity.
• The computer programs developed to simulate the performance
of commercial units provide a useful means for reactor design
and design optimization.
• Engineering scale-up from the Pernis demonstration unit size
(600 Nm3/h) to the SYS reactor size (325,000 Nm3A) or any
other size consists, apart from paying attention to flow distri-
bution and mechanical requirements for -which the scale-up
principles are known, simply of the parallel installation of
standardized unit cells. The Pernis unit has one unit cell
in a horizontal reactor cross-section, whereas the reactors
in the SYS unit have 68 identical (larger) unit cells in a
cross-section.
The experience gained with the special features incorporated
in the reactor section is as follows:
• The unit cells were easy to handle and easy to load into the
reactors.
• The automatic sequence controller operates the unit by setting
the duration of the acceptance and regeneration cycles and by
820
-------
continuous checking of all important valve positions, flows
and temperatures. The required attention of operators, after
the start-up and the initial operation period, was reduced
to almost nil. No extra manpower is required to operate this
unit.
• The special, large flue gas valves are working satisfactorily,
as are all other sequence-controlled valves (wedge-in-wedge
type) .
• The pressure drop over the reactors is low (8.6 inches WG)
but somewhat higher than expected (6 inches WG).
• The "open bypass" (continuous recycle of a small quantity
of treated flue gas) effectively prevents any pressure surges
caused by reactor switching from influencing the boiler operation.
After some initial problems in the work-up section had
been remedied, the absorber efficiency is 99.9% and the total
sulfur content of the stripped water is about 20 wt.-ppni S,
of which 75% is present as sulfate. The co-processing of the
SC>2 in the Glaus unit has no effect on the efficiency of this
unit.
During the first year, the only significant operating
problem requiring adjustment in processing conditions resulted
from partial condensation of the steam used to purge the reactor
821
-------
before and after regeneration. Since this purge steam flow is
intermittent, somewhat more than the usual attention to engineering
detail is required to avoid excessive loss of superheat temperature
and subsequent formation of liquid water.
SFGD for Coal-Fired Boilers
Flue gas from oil-fired boilers normally contains 0.02-0.12
gram/SCF (50-300 mg/Nm^) of solid material. Coal-fired boilers,
however, produce off-gas with solids contents which may be higher
by a factor of 50 to 100. Moreover, the solids compositions
differ considerably, and components might be present which have
an unfavorable effect on the performance of the copper acceptor.
Thus, a sound program on coal-fired systems must demonstrate
(J) the ability of the cell design to operated in a stable manner
with very high loadings of particulate matter and (2) the
chemical and physical stability of the acceptor material.
Accordingly, a two—phase program has been devised, the first
phase comprising continuous operation of a relatively simple
dummy reactor set-up with no acceptance/regeneration cycle.
The second phase is the demonstration of the actual performance
of the acceptance/regeneration system.
Test - Coal-Fired Utility Boiler in Rotterdam
For the first phase a dummy parallel-passage reactor was
822
-------
erected next to a coal-fired 57 MW B & W utility boiler of the
"Galileistraat" Power Station of the Gemeente Energiebedrijf,
Rotterdam, The Netherlands.
In these tests a slip stream of flue gas upstream of the
air preheater and the precipitator -was isokinetically sampled
from the main duct and drawn through a parallel-passage reactor
by means of a steam ejector. The unit was further provided with
a heater to ensure a constant reactor inlet temperature of
400 C, a cyclone and filter assembly to measure actual particulate
content of the flue gas at the inlet and the outlet of the
reactor, and facilities for measuring erosion/corrosion of
various materials at different gas velocities.
With particulate matter loadings of 2-4 gr/SCF (6-9 g/Nm-)
and peak concentrations of 8 gr/SCF during soot blowing, the dummy
reactor set-up was first tested for 120 hours under typical
SFGD reactor conditions to determine the resistance of the
parallel-passage system to fouling by the unfiltered flue gases.
Over that period no increase in pressure drop was observed, nor
did visual inspection show any signs of deposits interfering
with the process.
In a second test, the unit was subjected to continuous
operation for 703 hours at a flue gas velocity of twice the
823
-------
velocity normally experienced in the SFGD process, so as to
provide erosion/corrosion rate data for engineering design aspects.
In this second 703-hour test, the reactor had been filled
with copper-on-alumina acceptor which, therefore, has been subjected
to the rather severe conditions of continued acceptance during
703 hours. Measurement of the performance of the acceptor
in laboratory apparatus after this test showed that the acceptor
stability was equivalent to that of acceptor from life tests in
the oil-fired Pernis demonstration unit.
This work demonstrated the effectiveness of the parallel
passage principle to cope with high particulate loadings from
a coal-fired utility boiler and suggested that the system could
function successfully through oxidation-reduction cycles.
Demonstration Unit at Tampa Electric Company
The SFGD unit is fitted to process a slip stream from units
1 or 2 of TECO's Big Bend Station. These units are fired with
predominantly Western Kentucky coal, have a design rating of
446 MW, and were designed and constructed by Stone and Webster.
Flue gas is withdrawn upstream or downstream of the cold
precipitators at a temperature of approximately 275°F.
A simplified process flow of the SFGD system is shown in
Fig. 6.
824
-------
The single reactor contains a stack of up to 5 unit-cells
or modules which are identical with those in the SYS unit. The
flue gas capacity per stack of unit cells is approximately the
flue gas output of 0.6 megawatt, and the desulfurization capacity
per stack is that which will provide 45-60 minutes of operating
time and achieve 90% desulfurization of a flue gas containing
approximately 3000 ppm of SC>2 • The hydrogen used to regenerate
the acceptor is drawn from a portable hydrogen trailer and is diluted
before use. The reactor is purged before and after regeneration
with superheated steam; the regeneration off-gas is returned
to the main flue gas duct.
An SFGD unit may be applied to a boiler in two ways:
(1) integration between the economizer and air preheater to
reduce the temperature adjustment required for the acceptance
reaction and (2) add-on downstream of the air preheater, requiring
additional fuel to increase the gas temperature to 700°P and
heat exchange to recover some portion of the added thermal
energy. As a practical matter, the net re-heat requirement for
the add-on application is approximately 35°F of semsible heat on
the flue gas, equivalent to 1% of the fuel input to the boiler.
Since the SFGD process removes SO^ as well as SCu,
integration between the economizer and air preheater permits
greater heat recovery in the air preheater due to the decrease
825
-------
in the sulfric acid dew point. When burning high sulfur fuels,
this depression of acid dew point will allow a reduction of flue
gas temperature as much as 70°F, equivalent to 2% of the fuel
input to the boiler.
The demonstration unit has been designed as an add-on
retrofit. A regenerative heat exchanger reduces the fuel requirement
to the in-line heater to raise the flue gas to the temperature
required for good acceptor utilization, 700-800°F. The blower
is oversized so that the gas rate to the SFGD unit is held constant
with the excess by-passing the reactor. The valves and sequence
controller are similar to those used in the SYS design. SC>2
content of the flue gas in and out of the reactor, as well as in
the regeneration off-gas, is monitored automatically with
infrared analyzers.
The test program originally planned for the SFGD unit at
TECO called for approximately 90% desulfurization on 1200-1400
scfm of flue gas carried out over several thousands of cycles
to clearly establish physical and chemical stability. However,
since the plant is well instrumented, the test program has been
modified to allow for a process variable study while the acceptor
life-test is in progress. This will permit a direct comparison
of the performance of the same acceptor material at SYS and TECO.
826
-------
The acceptor activity/stability test, which is in progress,
span 5000 cycles (equivalent to 20,000hrs. operation at SYS)
of -which 500 cycles -will be on flue gas taken up-stream of the
precipitator. During this life-testing, a series of tests of
short duration is being performed in order to observe the following
(1) The effect of gas velocity on the desulfurization
efficiency and reactor pressure drop.
(2) The effect of acceptance time on slip curves at
75-90% desulfurization.
(3) The effect of flue gas inlet temperature on removal
efficiency at various gas velocities and on
regeneration efficiency.
(4) The effect of reactor length on desulfurization
efficiency and pressure drop.
The reactor will be opened after 3000 and 5000 cycles to
obtain samples of acceptor for laboratory evaluation and to
determine erosion and corrosion rates of the unit cells. Also
at suitable intervals, the gas streams will be analyzed for
solids, SO- and H2S contents.
Near the end of the life-test the systems will be operated
to demonstrate NOX reduction.
827
-------
LITERATURE
1. Dautzenberg, P.M., Naber, J. E., van Ginneken, A. J. J.,
"The Shell Flue Gas Desulfurization Process," A I Ch E,
Sixty-eighth Annual Meeting, Feb. 28-Mar. 4, 1971, Paper 3.1'd
2. Dautzenberg, F. M. , Naber, J. E., van Ginneken, A. J. J.,
"Shell's Flue Gas Desulfurization Process," CEP, Vol. 67,
Aug. 1971, pp. 86-91
3. Conser, R. E., Anderson, R. F. , "New Tool Combats SC>2 Emissions",
Oil and Gas Journal, Oct. 29, 1973
4. Naber, J. E. , Wesseling, J. A., and Groenendaal, W. ,
"New Shell Process Treats Claus Off-Gas," CEP, Vol. 69, Dec. 3973,
*
pp. 29-34
5. Ploeg, J. E. G., Akagi, E., and Kishi, K.,
"Shell's Flue Gas Desulfurization Unit at Showa Yokkaichi
Sekiyu K.K," Petroleum International, Vol. 34, No. A, July 1974
6. Groenendaal, W. , Naber, J. E. , Pohlenz, J. B.,
"The SFGD Process - Demonstration on Oil-And Coal-
Fired Boilers", A I Ch E National Meeting, Mar. ]0-13, 3974, Tulsa
828
-------
Table
COMPARISON OF DESIGN AND ACTUAL PERFORMANCE DATA
SO2 Removal Efficiency, %
Hydrogen Consumption,
wt H2/wt S
Absorber Efficiency, %
SO2 Concentration of
Stripped Water, ppm wt S
Total S Concentration of
Stripped Water, ppm wt S
Influence on Clous Efficiency
Design
90
0.20
99.5
10
Actual
90
0.19
99.9
None
20
None
829
-------
CD
OJ
O
300
0.
Q.
OC
UJ
O
O
U
CM
O
200
100
30
60 90
ACCEPTANCE TIME .MINUTES
FIGURE 1
120
SO2 CONCENTRATIONS IN TREATED FLUE GAS
-------
00
I
\
o
UJ
h-
oc
z
O
h-
o
Q
O
cc
Q.
CM
O
500
400
300
20O
100
PREDICTED
— - ACTUAL
30 60 90 120
REGENERATION TIME,MINUTES
FIGURE 2
SO2 CONCENTRATIONS IN REGENERATION
OFF - GAS
-------
TREATED FROM
FLUE-GAS OFF-GAS UN
I).99 0.05
fl.07
SFGD 9.38 SFGD 9.33
ADIP
ITS
r
50.0
CLAUS 56.37 .^ C111 BU1 ,„
SECTION SECTION
10.45 ^ r
t i INCINERATED
fe 2-96 4 TflM-HAS
7.41
FLUE -GAS
ORIGINAL DESIGN
TREATED FROW
FLUE-GAS OFF-GAS UN
0.44 0.00
0.07 0.51
SFGD 4-19 SFGD "-19
ADIP INCINERATED
TS TAIL-GAS
130.0 4.02*
_ ^ CLAUS 13qi17^cillou,,p
ir SECTION SECTION
14.70
1 4.63 • AFTER INSTALLATION OF SCOT UNIT: 0.20
FLUE-GAS
ACTUAL OPERATION
FIGURE 3
SULPHUR LOADS OF SFGD UNIT AT SYS,
IN TONS/DAY OF SULPHUR
832
-------
00
u>
u>
ME GENERATION GAS
ACCEPTANCE TIME: 120 MIN.
BOILER
FEED WATER
FIGURE 4
SIMPLIFIED PROCESS FLOW SCHEME OF SFGD UNIT AT SYS
-------
FIGURE 5
SFGD UNIT AT SYS'S YOKKAICHI REFINERY
834
-------
oo
UJ
Ul
REGENERATION
GAS
FLUE GAS
FROM DUCT
UP/DOWNSTREAM
PRECIPITATOR
AIR
FIGURE 6
SIMPLIFIED FLOW SCHEME OF SFGD DEMONSTRATION UNIT FOR
COAL FIRED UTILITY BOILER AT TAMPA ELECTRIC, FLORIDA
-------
STATUS REPORT
ON
CHIYODA THOROUGHBRED 101 PROCESS
Masaaki Noguchi
Engineering Manager
Chiyoda International Corp., Seattle, Wash.
ABSTRACT
This process uses water to absorb S02S which is oxidized by air to
yield sulfuric acid. The resulting weak acid is actually circulated as
an absorbent of S02> and a part is neutralized with limestone to produce
the by-product gypsum. The gypsum is recoverable as wallboard or as a
cement retarder, depending on local conditions.
Ten commercial plants have been already commissioned in Japan
including a 250MW power plant application, which will be followed by
several plants now under construction. After slight modification all
plants are running quite satisfactorily with over 90% removal of S02-
All applications on boilers in Japan are treating oil burnt gas.
Therefore, a demonstration plant for coal burnt gas has been constructed
in Florida to likewise prove the process.
Numerous studies in a laboratory for catalyst activity with impur-
ities from coal have been conducted. Comprehensive study is also being
done to optimize and economize the plant for large applications and to
improve it for better air pollution abatement processing.
837
-------
Status Report on
Chiyoda THOROUGHBRED 101 Process
1. INTRODUCTION
Several scrubbing processes in Japan and in the United States are
currently reputed as highly developed for removing S02 from waste gases.
The Chiyoda THOROUGHBRED 101 (CT-101) Process is also highly appraised
in Japan with its dry, marketable by-product and simple operation. Ten
plants are operating smoothly and five others are under construction in
Japan and in the United States (Table 1), three years after the process
was announced by Chiyoda Chemical Engineering & Construction Company.
The largest plant now operating is for Hokuriku Electric Power Company
at its 250MW boiler in Toyama, and a 350MW plant is now under construc-
tion for the same utility conpany. All CT-101 applications for boilers
in Japan are treating oil-fired gases, because of lack of coal resources
there. Recently Chiyoda International Corporation (CIC) was established
in Seattle, Washington to introduce the process to the United States.
CIC is constructing a demonstration plant at Scholz Plant of Gulf Power
Company in Sneads, Florida for coal-fired application.
The purpose of this paper will cover the CT-101 process, plants and
operations.
2. PROCESS DESCRIPTION
A simplified flow diagram is given in Figure 1. Flue gas generated
in a boiler or an incinerator, is introduced into a venturi type prescrub-
ber to remove particulates and to be cooled down. For gases, heavily
loaded with particulates, as from coal-burning power plant boilers, it
is desirable to have an electrostatic precipitator before the Chiyoda
system to remove most of the particulates.
The flue gas is then fed into the absorber where S02 is absorbed by
water. The absorption is countercurrent, that is, the gas enters at the
bottom and the absorbent at the top. The absorber is a tower with pack-
ing supported on grids. The gas leaving the absorber is reheated to
restore buoyancy and to reduce the incidence of a visible steam plume.
The scrubber effluent solution flows to the oxidizer tower where
air bubbled through the liquor converts all the absorbed S02 to H2S04,
with the help of ferric ion dissolved as a catalyst. Part of the liquor
leaving the oxidizer passes to gypsum production and the remainder is
recycled to the absorber. Thus, the recycling absorbent is dilute sul-
furic acid solution with a concentration of about 2 to 5%. In the
absorber, oxygen in the recycling liquor and in the flue gas oxidize the
absorbed S02> thus improving the abosrption capacity of the liquor.
The sidestream from the oxidizer is neutralized with lime or lime-
stone to crystallize gypsum. Chiyoda has developed a crystallizer
838
-------
capable of producing crystals suitable to be dehydrated by conventional
centrifuges. The mother liquor leaving the crystal is recycled to the
adsorber. Final chemistry is as follows:
S02 + H20 =-• H2S03
H2S03 + 1/2 02 ="- H2S04
H2SC>4 + CaC03 + H20 *> CaSO^ . 2H20 + C02
3. FEATURES
3-1. Simple Process Flow
The CT-101 process has quite a simple plant structure. That is,
1) The process itself is simple, using few vessels.
2) The plant is composed of simple, already known but valuable
apparatus of each unit operation, except the gypsum crystal-
lizer.
3) No slurry is used in the absorber where flue gas passes
through, securring no clogging problems.
Therefore, the process promises easy operation, wide range of
operational flexibility and stable operation.
3-2. High Efficiency Desulfurization
S02 in the cleaned gas can be held to SOppm or lower. In other
words, more than 95% desulfurization is easily attainable, because of
the combination of the absorber and the oxidizer and the usage of Fe~t"++
catalyst.
3-3. Dry By-product Gypsum
Gypsum is a highly stable chemical compound of sulfur, that is
harmless when stored. The by-product of CT-101 contains only about 10%
of free water, therefore, it can be readily stock-piled without the slump
problems which normally occur when excessive moisture is present. Trans-
portation is a simple matter which can be accomplished by utilizing dump
trucks. Furthermore, the by-product gypsum, can be recoverable as a
wallboard and cement constituent as has been done in Japan.
4. HISTORY OF DEVELOPMENT
Investigation and analysis of referential literature covering all
proposed processes for stack gas cleaning, including dry and wet processes
was performed and a suggestion for the process written by Johnstone in
1931 (Reference 1) was found promising. Starting by tracing the original
paper, the process was developed in the following manner:
4-1. Laboratory Work
A series of experiments was conducted to locate a possible catalyst.
Earlier attempts by Copson et al (Reference 2, 3) were focused primarily
on manganous sulfate as catalyst because of its high activity, but not
industrialized of poisoned catalyst by contaminants carried in the gas
and others. Ferric ion was chosen because of its harmlessness, availa-
bility and stability to the possible contaminants as on Table 2,
839
-------
is active enough comparing with Mn^" catalyst, as shown on Figure 2, at
the temperature that usual flue gases reach when directly contacted by
water.
4-2. Bench Test (December 1970 - June 1971)
Bench scale tests were conducted using flue gas of 50 Ntrr/hr
generated by burning fuel oil. During this period the main effort was
concentrated on collecting absorption and oxidation process data. S(>2
and other components were added to the gas when needed.
4-3. Pilot Test (July 1971 - )
A pilot plant with a capacity of 1000 Nnr/hr was constructed to
collect data necessary to design equipment for commercial application,
including gypsum production. Bunker oil, sometimes mixed with asphalt
was burned in the incinerator. After three months operation proved the
reliability of the process, the process was introduced to Japanese
industry in October 1971. Laboratory work is still being continued to
collect catalyst stability data and to analyze reaction mechanism.
5. PRIMARY COMMERCIAL PLANTS
After the introduction of CT-101, four plants were designed and
constructed almost simultaneously. These plants encountered several
problems. However they have already been overcome as follows and all
plants are now running quite satisfactorily.
1) Too slow flowing velocity in horizontal pipings caused gypsum
deposits. Pipings were inclined and smaller sized pipes were adopted.
2) Gypsum erroded rubber lined glove valves which were changed to
stainless steel valves of a different style.
3) Air spargers for oxidation were plugged by gypsum crystals. To
eliminate this trouble, a clarifier was installed to minimize carry-
over of gypsum crystals in recycling absorbent from the centrifuge.
Sparger holes were enlarges and air was cooled down before entering
the oxidizer.
4) By-product gypsum contained more than 157o of free water that would
decrease dryer capacity of wallboard manufacturer and become diffi-
cult of bulk handling with dump trucks. The crystallizers were
redesigned with longer holding time of crystals to obtain a larger
crystal size for easier dewatering in the centrifuge.
5) Pitting corrosions on some parts of stainless steel plates caused
by the accumulated chlorine ions in the absorbent when burning low
grade oil. The limitation of chlorine content was reduced to ZOOppro
by adjusting a purge rate.
6. LARGE SCALE PLANTS
Several large scale plants have been under construction, and two of
840
-------
them were commissioned in June of this year, one for Hokuriku Electric
Power Co. and the other for Mitsubishi Chemical Co.
6-1. Plant Outline
For Hokuriku the design was started in November 1972 and concrete
was poured July 1973 for foundation work. Hot gas was introduced on June
4th of this year. With the exception of one month intermission in July
for inspection of both boiler and turbine which were constructed at the
same time, the plant has been running very satisfactorily.
A general layout of this desulfurization unit is given in Figure
3. The whole SCL removal process is located in 5,000 m^. The crystal-
lization process is done in a separate area from the absorption section
leaving an area for future expansion.
The main feature of the Hokuriku plant is in the absorption and
the oxidation apparatus. They are combined in one cylindrical column.
That is, the column consists of two concentrical shells with a center
area for the oxidation and an annuli for the absorption. This configur-
ation is good for the internal supports in mechanical strength and also
helpful for liquid distribution at the top of the column. Table 3,
shows the specification of the main apparatus. Stainless steel, FRP,
rubber lining, etc. can be applied for construction materials without
difficulties. Chiyoda International Corporation estimates roughly the
CT-101 plant cost in the U. S. will be in the range of $80 - 100/kw for
a 250MW power plant.
6-2. Operation
Operational data obtained in both plants are described in Table 4.
The Hokuriku Plant is extracting half of the flue gas from a'SOOMW boiler.
Instrumentation allows gas fluctuation according to boiler operation in
the range of 25 to 100% of gas treating capacity. The plant is treating
gas by burning 1% sulfur oil, however, it can treat gas of 37o sulfur oil
in the future when the facility will be doubled to treat all the gas
generated in the boiler.
The plant is operated by two people per shift; except when solids
are handled such as, unloading limestone and loading gypsum into trucks.
Almost all plant apparatus can be operated and their performances can be
observed at the control panel room. The plant performances demonstrates
that the process serves for pollution abatement, that is, for removal of
S02 and fine particulates. The S02 level is reduced to below 30ppm at
the stack outlet. Particulates, are also removed to 16.5 mg/Nm3 from
280 mg/Nm3.
Gypsum is being sold to wallboard manufacturers and cement compan-
ies at good prices. Bleed stream from the system to maintain the concen-
tration of cholrine and others is disposed after recovery of catalyst
dissolved, being controlled of pH and removed of suspended solids. The
final properties of water meet the waste water qualification regulation.
841
-------
Generally the properties are as follows:
pH 6.5 — 7.5
COD less than I5ppm
Suspended Solid less than 30ppm
7. APPLICATION FOR COAL-BURNING
CT-101 units are being applied successfully for oil-burnt flue
gas in Japan. However, there is no application for coal gas. The main
difference between oil and coal-burnt gases would be in particulates
loading and compositions of fly ash and gas.
Grain removal is recommended by installing some device before the
process, such as an electrostatic precipitator to reduce it to be below
2000 mg/Nm^. After using such a dust collector in the process, particu-
lates will be removed by prescrubbing first and next by simultaneous
S02 absorption.
Several kinds of fly ash collected in the U. S. utility boilers
were added in the absorbent of test units in Kawasaki Laboratory of
Chiyoda and the results obtained indicates CT-101 application of coal
gas without trouble. Also, individual components of the fly ash and
organic constituents in the coal gas were investigated in the same
manner. Some results are tabulated in Table 2.
Meanwhile, Gulf Power Co. has a testing project for S02 removal
using several kinds of U. S. coal. The decision was made to join the
project and to demonstrate the process in the U. S. The plant has a
capacity equivalent to 23MW. The gas flow rate will follow the fluc-
tuation of the boiler operation. Before the process, an electrostatic
precipitator has been installed of which particulate removal efficiency
will be adjusted accordingly with test schedules to obtain data on a
wide range grain loading into the stack gas cleaning process. The
plant will be completed within this year.
8. FUTURE
We anticipate successful operation of CT-101 process at the Florida
plant which will demonstrate process applicability to coal burning
boilers. Besides} studies are being continued for better construction
materials, processing apparatus, economics and for a better abatement
process. One of such achievements is the process for simultaneous
removal of S02 and NOX, basically using the CT-101 process with addi-
tion of an 03 generator. (Reference 4)
842
-------
REFERENCES
1.
2.
3.
4.
TABLES
1.
2.
3.
4.
Figures
1.
2.
3.
Johnstone, H.F., Ind. Eng. Ghent. 23 (5), 559 (1931)
Copson, R.J., and J.W. Payne, Ind. Eng. Chem. _25 (3), 909
(1933)
Tarbutton, G., J.C. Driskell, T.M. Jones, F.J. Gray, and
C.M. Smith, Ind. Eng. Chem, 49 (3), 392 (1957)
Yamamoto, 0., S. Fukui, H. Hishino, Y. Kameoka and J.
Miyazaki, 74-259, 67th APCA annual meeting, Denver, 1974
LIST OF COMMERCIAL PLANTS
PARTIAL LIST OF CHEMICALS TESTED IN LAB FOR CATALYST
ACTIVITY
EQUIPMENT LIST
TYPICAL OPERATIONAL DATA
Process flow diagram
Catalyst activity
General layout of Hokuriku Electric Co., plant
843
-------
Table 1, LIST OF COMMERCIAL PLANTS
00
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Owner of plant
Nippon Minning Co.
Fuji Kosan Co.
Mitsubishi Rayon Co.
Tohoku Oil Co.
Daicel Co.
Amagasaki Coke Co.
Hokuriku Elect. Co.
Mitsubishi Chem. Co.
Mitsubishi Pet. Chem. Co.
Mitsubishi Pet. Chem. Co.
Gulf Power Co.
Denki Kagaku Co.
Hokuriku Elect. Co.
Elect. Co.
Elect. Co.
Location
Mizushima
Kainan
Otake
Sendai
Aboshi
Kakogawa
Toyama
Yokkaichi
Yokkaichi
Yokkaichi
Florida
Chiba
Fukui
Nm3/hr
33,500
157,200
90,000
14,000
100,000
36,200
750,000
400,000
150,000
700,000
90,000
120,000
1,050,000
750,000
750,000
Gas source
Glaus
Boiler & Glaus
Boiler
Glaus
Boiler
Gas incinerator
Power boiler (250MW)
Boiler
Boiler
Boiler
Power boiler (23MW)
Boiler
Power boiler (350MW)
Power boiler (250MW)
Power boiler (250MW)
Completion
Nov.
Nov.
Jan.
Feb.
Nov.
Feb.
June
June
Sept.
Oct.
Nov.
Feb.
May
Oct.
Spring
1972
1972
1973
1973
1973
1974
1974
1974
1974
1974
1974
1975
1975
1975
1976
-------
Table 2, PARTIAL LIST OF CHEMICALS TESTED IN LAB FOR CATALYST STABILITY
1. All chemicals listed do not have negative influence on catalyst
activity in CT-101 process.
2. Test Condition
Catalyst
Absorbent
Fe^ 2000ppm
Free H2S04 2.2wt%
3. Concentration in numerals gives only the maximum
not the limitation.
Component
Formaldehyde
Acetaldehyde
m-Tolualdehyde
Phenol
Resorcinol
Phlorogrucinol
Acetaldehyde
+ Resorcinol
Acetaldehyde
+ Phenol
Fluoranthene
+ Perylene
+ Benzopyrene
+ Pyrene
Ditto
+ Ether
ppm
300
800
100
100
100
300
300
100
800
200
30
30
30
30
ditto
+1%
Component
V205
v2o5
-fNiO
NiS04
Metal Cr
PbO+PbS04
H3AS04
Metal Sn
Na2S04
MgS04
(NH4)2S04
HC1
HN03
valve tested and
ppm as ion
100
500
100
1,000
500
5
10
100
10,000
10,000
15% as (NH4)2S04
200
1,000
845
-------
Table 3, EQUIPMENT LIST
Hokuriku Electric Power Company
250MH CT-101 Unit
No.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Name
Precooler
Abs-Oxidizer
Absorbent Tank
Limestone Silo
Limestone Slurry Tank
Crystallizer
Clarifier
Centrifuge
Flue Gas Fan
Air Blower
Absorbent Pump
Mist Eliminator
Type
Venturi
Double shells
Cone roof
with bag filter
Agitated
Agitated
with rake
Basket
Turbo
Turbo
Double suction
Chevron
Dimensions
Diameter 21m
Height 28m
5,000 m3
500 m3
40 m3
500 m3
300 m3
750,000 Nm3/hr
7,500 Nm3/hr
Quantity
2
1
1
1
1
1
1
3
1
2
3
1
846
-------
Table 4, TYPICAL OPERATION DATA
Owner of plant
Gas source
Hokuriku
Electric Power Co.
Oil burnt boiler
Mitsubishi
Chemical Co.
Oil burnt boiler
Operational data
Flue gas rate
m3/hr
Boiler capacity equiv. MW
S02 In
Out
Removal
HoSOA in Absorbent
Limestone
Elect, power
Water
ppm
ppm
%
ton/ day
KW
ton/hr
750,000
250
436
21
94 7
2.2
34.8
4,500
45
324,000*
110
780
25
97.
1.
27.
3,400
20
3
8
1
* Boiler is loaded to 80% of capacity in summer.
847
-------
Cleaned gas
Reheater Absorber Oxidizer
00
-IS
00
Crystallizer
Limestone
Centrifuge
Prescrubber Filter
Sludge
Air
Purge
to treatment
Figure 1, Process flow diagram
-------
o
c
01
c
o
•H
•a
•^
x
o
100
20 _
Temperature (°c)
Figure 2, Catalyst activity
849
-------
EP
00
m
o
Gas ducts
O
Stack
0 10 20m
Scale
Absorption
Section
Gas mixer
eheater
Pipe rack
Crystallization
Section
Gypsum
store
8 10
n°©
O
Numbers; refer to Table 3
Figure 3, General layout of Hokuriku Electric Go,, plant
-------
FLUE GAS DESULFURIZATION BYPRODUCT DISPOSAL/UTILIZATION
REVIEW AND STATUS
H. W. Elder
Tennessee Valley Authority
Muscle Shoals, Alabama
Prepared for Presentation at
Flue Gas Desulfurization Symposium
Sponsored by the Environmental Protection Agency
Atlanta, Georgia
November 4-7, 1974
851
-------
FLUE GAS DESULFURIZATION BYPRODUCT DISPOSAL/UTILIZATION
REVIEW AND STATUS
H. W. Elder
Tennessee Valley Authority
Muscle Shoals, Alabama
In the symposium last year, several excellent papers intro-
duced the subject of this session and the general considerations will
not be covered in this paper. For those who did not attend the last
meeting in New Orleans, I recommend that a copy of the proceedings be
obtained. The topics that we will cover in this discussion focus on
progress in understanding and solving problems associated with disposal
and use of byproducts from flue gas desulfurization processes.
During the initial period, the discussion will center on dis-
posal of waste materials and later in the program, use of recovered
products will be considered. The first paper will review the status of
waste disposal from full-scale lime-limestone scrubbing systems currently
installed. Then the EPA program to develop better alternatives will be
reviewed, followed by a discussion of the experience with one of the
methods for chemical fixation. One of the highlights of the session will
be a review of the status in Japan where the emphasis is on utilization
of recovered products even from lime-limestone systems. During the por-
tion of the program directed toward work in this country on utilization,
the preliminary results of market evaluations for sulfuric acid, sulfur,
and gypsum will be presented. The experience with production and market-
ing of acid from the Boston Edison magnesium oxide scrubbing system will
conclude the formal presentations after which a discussion with the
audience and among the panelists is scheduled. My presentation will
attempt to identify the important considerations from the utility view-
point.
Removal of sulfur oxides from stack gas has received far more
attention than the question of what to do with it after it is caught.
The technology for removing S02 is gradually evolving to the point that
within a year or two, alternative methods probably will be available for
use where they are needed. A very real concern in considering use of the
technology is disposal of products. Recurring questions include:
* How much is produced?
• What are the choices?
« For throwaway processes, how do I throw it and where is away?
853
-------
• For recovery processes, is the market reliable for continuous
production of varying amounts and what happens if the pipeline
clogs?
• Are the rules going to change?
• What is the cheapest way out?
Of course, no one has all the answers at present but reliable
responses are needed to provide the guidance required for intelligent
process selection. The following discussion will help to give perspec-
tive to some of the concerns.
HOW MUCH IS PRODUCED?
The quantity produced for a single installation depends on
several factors including process used, the sulfur content of the fuel,
the extent of control required, and the load factor for the plant; the
quantity on a national basis depends on how many sources will use de-
sulfurization systems. There probably is no such thing as a typical
installation, but for example, a 500-MW power plant with an average load
factor of 60$ and burning 3-5$ s coal would produce about one-half
million tons (dry basis) of waste over a 25-year period when equipped
with a limestone scrubbing system to meet the new source performance
standard; the volume of solids (50$ moisture) produced would occupy
approximately 6000 acre-feet. Installation of a recovery process to
produce sulfuric acid at the same plant would result in about 85,000
tons per year of acid.
The number of plants that will be equipped with stack gas
cleaning systems is uncertain. In spite of a general preference for
use of low-sulfur fuels to meet emission requirements, several utilities
are proceeding with plans for installation of desulfurization systems
either because of insufficient supplies of low-sulfur fuel or because
of potential economic advantages.
One trend is apparent--nuclear generation capacity has not
reached the level of earlier forecasts (Figure l)1'2 and with the
emphasis for use of gas and oil elsewhere, coal is likely to be the
primary fuel for power generation for many years to come. On the other
hand, energy conservation measures may decrease the growth rate for
power demand. This has been partially responsible for deferment of
some new power plant construction in recent months.
854
-------
200
NOTE:
DIFFERENCE IN 1985 EQUIVALENT
TO ITS MILLION TONS OF COAL
( | ESTIMATED IN 1974*
ESTIMATED IN 1973*
150-
o
g
M
X
Z
o:
ui
u
o
(X
UJ
o
Q.
UJ
-I
o
100
50
1975
1980
1985
Figure 1. Projected Nuclear Power Generation
855
-------
Future energy requirements were estimated in a recent report
prepared for the Office of Coal Research, U.S. Department of Interior,
by the Hudson Institute. The coal needed to meet the direct com-
bustion energy demands through 1985 is shown in Figure 2. To develop
the capability to produce and deliver the estimated amounts will re-
quire heroic effort. About j6%> of the states have S02 regulations that
limit sulfur content of coal to 1% or less, but in 1970 only 1J$ of the
coal used by utilities had less than 1$> sulfur. Incentives to produce
low-sulfur coal will probably improve the ratio of low to high-sulfur
coal and an arbitrary assumption of 30$ low S in 1985 will allow some
estimates of desulfurization product quantities. Further assumptions
are average sulfur content of k$> (high sulfur fraction only), 80$
removal, and 50$ of the high-sulfur coal is used in installations
equipped with flue gas desulfurization systems. On this basis, the
total sulfur removed would be 20 million tons annually.
Recovery of this amount in calcium-based wet scrubbing
systems would require approximately 200,000 acre-feet per year of
storage for waste disposal. Conversion to sulfuric acid would result
in production of 60 million annual tons of acid, an amount that ex-
ceeds the total projected market. No one suggests that a single method
will be used universally. Instead, a variety of processes will probably
result in a mix of throwaway and recovery products.
WHAT ARE THE CHOICES?
Several different processes are being tested in full-scale
facilities and others are planned. Depending on successful completion
of continuing development work, the list of candidate processes for
sulfur oxide control could include the methods listed below.
Processes that Produce Waste Products
Limestone scrubbing
Lime scrubbing
Double alkali (sodium or ammonia with lime or limestone)
Dilute acid neutralization with lime or limestone
Sodium salt scrubbing
856
-------
2000
V)
o
2
2E
z*
o
I
13
I50O
o
o
a
UJ
UJ
o
UJ
b
UJ
o
-------
Limestone and lime scrubbing systems account for most of the large-scale
installations primarily because development work has been focused on
these methods. Use of limestone generally produces more waste solids
than use of lime because a larger amount of excess absorbent is required
for comparable removal. The sludge storage characteristics are poor
except when oxidation of sulfite to sulfate is essentially complete.
High oxidation can sometimes be achieved directly in the scrubber when
the S02 concentration in the gas is low. A separate oxidation step is
used in some processes. Interest in the double alkali method stems from
operating problems with slurry scrubbing. The problems can be avoided
by scrubbing with a soluble salt and precipitating the absorbed sulfur
oxide as a waste material by addition of lime or limestone; the proper-
ties of the waste solids are similar to those produced directly in the
scrubber. Dilute acid neutralization with lime or limestone has the
advantage of producing an essentially completely oxidized product.
Dilute acid may be produced by absorption of S02 in a solution contain-
ing an oxidation catalyst or by carbon adsorption and subsequent washing.
Sodium salt scrubbing to produce a waste product is uniquely suited to
applications where abundant natural sodium carbonate deposits are
available and where solar evaporation can be used to dewater the waste
stream.
Although calcium reaction products are normally considered
waste materials, gypsum (calcium sulfate) has potential for use as a
material for wallboard construction or for use as a cement additive.
TVA has carried out a preliminary study3 for EPA on the market for by-
product gypsum in the wallboard industry.
Processes that Recover Sulfur in a Useful Form
Magnesium oxide scrubbing with thermal regeneration
Sodium salt scrubbing with thermal regeneration
Catalytic oxidation
Copper oxide - dry absorption-reduction
Except for the catalytic oxidation process, these recovery methods pro-
duce an intermediate, relatively concentrated stream of sulfur dioxide
that can be converted in conventional equipment into sulfuric acid or
can be reduced to elemental sulfur. Catalytic oxidation converts the
dilute sulfur dioxide in the total gas to S03 and produces sulfuric acid
directly. The yield of products (sulfur equivalent) is similar for all
of the methods. The magnesium oxide method requires scrubbing with a
slurry and involves solids handling for regeneration and recycle. Sodium
salt scrubbing permits use of a solution in the scrubber, but heat re-
quirements for regeneration are relatively high and oxidation compli-
cates the process. Catalytic oxidation requires good particulate removal
858
-------
ahead of the process, operates at a high temperature (800°F), and pro-
duces acid that is less concentrated than standard grades. Like the
catalytic oxidation method, the copper oxide dry absorption method
requires no reheat of the exit gas, but since the absorption beds are
fixed, gas switching is necessary.
The^most logical recovery product is sulfuric acid because
most of the sulfur used is in this form. However, tank storage is re-
quired to allow for distribution upsets. The most practical form for
storage and shipping is elemental sulfur, but the technology to reduce
S02 complicates the processes and most of it would have to be reoxidized
for eventual use. Moreover, natural gas is the conventional source of
reductant and the limited supply could affect both the availability and
economics. Technology is needed to use coal as the source of reducing
gas or solid for the conversion to sulfur; development work is in pro-
gress.
In addition to the basic sulfur products, acid and elemental
sulfur, processes that produce fertilizers and intermediates for the
pulp and paper industry are also being studied. These markets will be
limited, but they may add to the alternatives for specialized market
situations.
FOR THROWAWAY PROCESSES--HOW DO I THROW IT AND WHERE IS AWAY?
It is almost certain that discharge of significant quantities
of waste solids to inland streams will not be acceptable.
An important question that has not been adequately answered
is the extent of clarified liquor discharge or seepage that will be
acceptable from the standpoint of water quality. The simplest method
for disposal of the waste solids and the best approach for enhancing
scrubbing system reliability is to discharge a side stream from the
scrubber loop to a settling pond and to overflow the clarified liquor to
a receiving stream. Because of the low solubility of calcium sulfite
and sulfate, the total dissolved solids concentration is low. The
volume can be reduced by use of a thickener to concentrate the solids in
the stream from the scrubber. This technique is being used in some of
the installations. However, the objective for development work on lime-
limestone scrubbing systems has been to operate with no liquor discharge.
For closed-loop operation, pond management becomes a problem.
The goal is to minimize seepage and to return the total outflow to the
scrubber system. Is the local soil sufficiently impermeable so that a
liner is not required? Does the pond system need to have sufficient
storage volume to accommodate short-term precipitation? How is the
water balance maintained where long-term precipitation exceeds evapor-
ration? The answers to these questions may vary for each specific
installation, but some guidelines are needed. The pond management
859
-------
problem can be avoided by mechanical dewaCering of the scrubber solids
so that all of the recoverable liquor is retained in the system. How-
ever, filtered or centrifuged sludge generally contains about 50$
moisture and is structurally unstable; it is not suitable for landfill.
Compaction may be a feasible means for improving the stability, but the
solids revert to the characteristic sludge properties when exposed to
water. Also, soluble components may be leached out.
The most sophisticated (and expensive) technology being
developed for disposal of waste solids involves chemical fixation to
improve both the physical properties and the resistance to leaching.
Several companies now offer processes based on this approach and the
technology is being tested on operating systems. Chemical fixation is
apparently somewhat empirical and depends on sludge composition; the
composition is a variable not only among different applications of lime
or limestone scrubbing, but also in any specific system where ash con-
tent, excess absorbent, oxidation, and moisture can change significantly
over short periods. The treatment process will either have to have
rapid response to composition changes or will have to be very forgiving
regarding excursions.
The question of site selection for storage of wastes is de-
pendent on a variety of factors, particularly geography. For rural
areas where real estate is available, use of ponds (with impermeable
liners if necessary) should be acceptable. Dewatering, compaction, or
fixation might be justified if the trade-off in cost for transportation
to an alternate nearby site offsets the value of adjacent land. For
urban areas, nearby sites simply may not be available. In these cases
dewatering may be essential to avoid excessive transportation costs and
fixation may be required for permission to dump, particularly in con-
junction with sanitary landfill practice. It has been suggested that
waste materials be hauled back to the quarry from which the limestone
came or to the hole where the coal used to be. In theory, this concept
sounds good but in practice, the logistics may be virtually impossible.
It may be more practical to improve land surface features by systematic
land reclamation or development, provided that stabilization is the only
alternative.
FOR RECOVERY PROCESSES--IS THE MARKET RELIABLE FOR CONTINUOUS PRODUCTION
OF VARYING AMDUNTS OF A BYPRODUCT AND WHAT HAPPENS IF THE PIPELINE CLOGS?
It probably is not fair to jump 10 to 15 years into the future
and conclude that the supply of sulfur products exceeds the demand.
During a period of technology adoption, when the supply of recovered
products is small compared to the total market, the incremental produc-
tion can probably be integrated into growth markets with minimum impact,
particularly if an established distributor serves as a broker. Vari-
ations in production volume from the recovery operation could be com-
pensated by the flexibility of conventional sources. TVA has recently
860
-------
completed a hypothetical sulfuric acid marketing study for EPA based on
the TVA system.4 As a greater share of the total market is supplied
from variable sources, the ability of the shrinking basic production to
absorb the swings will diminish. On the other hand, multiplicity of
recovery sources should stabilize the short-range supply and seasonal
swings would cause the biggest problem. The net effect would be to
downgrade the value of recovery products so that reserve capacity from
conventional sources could be maintained to supply the peak demands; in
other words, the cost of turndown of backup facilities would have to be
deducted from the value of recovered products in order to stabilize
price. This is obviously an oversimplification because where different
companies in different industries are involved, coordination is diffi-
cult; industry-wide distribution strategies would be required. The
situation where the supply from abatement sources approaches the annual
demand would probably require an impractical amount of standby capacity
and the only alternative would be unprecedented reserve storage, a
costly option that needs definition. By the time that production of
recovered sulfur products exceed the total demand, the value will reach
a level consistent with the least cost for disposal. As the value de-
creases, new markets may develop and the total demand is likely to grow.
A very real concern is the need to continue production even if
there is no market, at any price. An obvious answer is to bypass the
desulfurization system and the circumstances for exercising this option
should be defined. Flexibility to convert a recovery system to a
throwaway type is a desirable feature, although the rational that led
to use of recovery in the first place would probably complicate this
consideration. However, for short-term interruptions waste disposal
might be feasible if not practical. For processes that make sulfuric
acid, the acid could be neutralized with limestone and discard gypsum.
Solution scrubbing systems could be converted to a double alkali pro-
cess. Processes producing elemental sulfur would fare best because the
product could probably be stored indefinitely. These manipulations
would play havoc with sinking fund projections, but they would provide
for continued power generation.
ARE THE RULES GOING TO CHANGE?
The regulations that apply to air quality are defined more
clearly than water and solid waste considerations and even they are
still being debated. It is impossible to intelligently evaluate alter-
native processes when reliable guidelines for all effluent streams are
not available. For example, selection of a throwaway system as the
least cost method on the basis that waste disposal by ponding is accept-
able might be a poor choice if subsequent solid waste regulations re-
quired stabilization. It would be similarly unfortunate if a process
to produce elemental sulfur were chosen to minimize the impact of market
uncertainties and later regulations prohibited open-pile storage.
861
-------
Regulatory organizations need to consider the interactions of
air, water, and solid waste control so that efforts to solve one problem
do not create another; economics of compliance must take into account
the total regulatory requirement.
WHAT IS THE CHEAPEST WAY OUT?
Separation of the cost elements associated with waste disposal
and recovered product distribution is difficult and direct comparison
of results could be misleading. The most effective way to evaluate the
overall process economics is to compare the lifetime operating costs in-
cluding amortization of the investment.
An earlier paper5 in this symposium provided detailed invest-
ment and operating economics for five processes including both throwaway
and recovery methods. As an example of the effect of disposal alterna-
tives on cost, use of an onsite, unlined pond (no thickener) with
clarified liquor recycle contributed approximately 7% of the total
operating cost; including an impervious liner increased the cost to
about 12$ of the total. Offsite disposal costs (including facilities
for filtration) were estimated to be approximately 10$ higher than
costs for onsite ponding. Chemical fixation (at an assumed cost of
$10/ton of dry solids) and disposal offsite would be approximately
higher than onsite ponding.
Credit for sale of recovered products does not constitute a
major reduction in operating costs; revenues at a price equivalent to
$25 per short ton S, would reduce the lifetime operating costs less
than 10$.
862
-------
REFERENCES
1. Armbruster, F., et al. (Hudson Institute). "Policy Analysis for
Coal Development at a Wartime Urgency Level to Meet the Goals of
Project Independence/1 prepared for Office of Coal Research, U.S.
Department of Interior (February k,
2. Davis, W. K. (fiechtel Power Corporation). Chem. Eng. Progr. 69,
No. 6, 48-53 (June 1973)- ~~
J. Corrigan, P. A. (Tennessee Valley Authority). "Preliminary Feasi-
bility Study of Calcium-Sulfur Sludge Utilization in the Wallboard
Industry." prepared for the Environmental Protection Agency (June
21, 197^).
k- Waitzman, D. A., et al. (Tennessee Valley Authority). "Marketing
H2S04 from S02 Abatement Sources—The TVA Hypothesis/' prepared
for the Environmental Protection Agency (December 1973 )•
5. McGlamery, G. G., and R. L. Torstrick (Tennessee Valley Authority).
"Cost Comparisons of Flue Gas Desulfurization Systems," prepared
for EPA Flue Gas Desulfurization Symposium, Atlanta, Georgia,
November h-J
863
-------
LIME/LIMESTONE SLUDGE DISPOSAL - TRENDS IN THE UTILITY INDUSTRY
by
C. N. Ifeadi and H. S. Rosenberg
Battelle's Columbus Laboratories
Columbus, Ohio
ABSTRACT
This paper is an evaluation and summary of the present state of
the art and future trends in the ultimate disposal of calcium-sulfur
sludges from utility flue gas desulfurization processes. Twelve
full-scale utility plants currently have the potential to gen-
erate calcium-sulfur sludges at the annual rate of about 2.5 x 10
metric tons (50 percent moisture). It is estimated that by 1980,,
lime/limestone scrubbing systems may generate as much as 120 x 10
metric tons of sludge (50 percent moisture) annually. Thus, the major
problem in the application of lime/limestone scrubbing technology is
the disposal of the large quantities of sludge. Other attendant
problems are associated with the physical and chemical characteristics
of the sludge. Physical characteristics influence markedly the land
use of the disposal site. Chemical characteristics are strongly
related to water pollution problems. Before final disposal, the
scrubber sludges may be subjected to various treatments which will
change their physical and chemical characteristics in order to effect
an environmentally acceptable disposition. Depending on economics and
local conditions, utilities may practice ponding only, or a combination
of ponding and/or other mechanical pretreatment processes. Ponding
without pretreatment is practiced in regions where there is relatively
cheap land and/or abundant solar energy to aid evaporation. Clari-
fication and vacuum filtration are common in regions with a high land
cost and a wetter and colder climate. Current popular final disposal
methods adopted by utility companies are ponding and landfill.
Disposal through utilization is not yet practiced on full-scale
installations, but the technical feasibility of various uses has been
demonstrated by some research groups. However, most have so far
proven economically unattractive.
865
-------
LIME/LIMESTONE SLUDGE DISPOSAL - TRENDS IN THE UTILITY INDUSTRY
BACKGROUND
In response to the need for up-to-date, factual information on
the treatment of stack gas for pollution control, Battelle-Columbus
began a study focused on this subject in January, 1974. The purpose
of this effort was to (1) establish a mechanism to collect installation
and operating data, (2) analyze these data, and (3) distribute the
resulting reports to management personnel concerned with the problem
of S02 in stack gas.
The original program was underwritten by the Electric Power
Research Institute. Additional work is being supported by the
American Petroleum Institute, and a number of individual companies
interested in both utility and industrial boilers. Battelle-Columbus
is actively seeking additional support from other companies for this
very significant group program.
The program comprises seven tasks. In brief, these are:
1. Survey installations in North America now using SO^
control devices
2. Establish an information and analysis center for information
on stack gas control
3. Prepare topical reports on subjects of importance to
members of the group
4. Survey overseas installations of SO control devices
5. Conduct laboratory studies on problems related to
stack gas pollution control
6. Contact vendors of S0? control devices to obtain
information on operating experience
7. Conduct review meetings.
867
-------
One of the topical reports prepared is on the disposal of sludge
from lime/limestone scrubbing processes. This effort will form the
basis for the current presentation.
INTRODUCTION
The United States Clean Air Act of 1970 requires stationary
combustion sources to control SO- emissions. The shortage of low-
sulfur fuels has increased the importance of stack-gas cleaning to
meet S0~ emission regulations. A major division in S0« removal
technologies is between recovery of S0_ in useful form and formation
of a solid waste. Most throwaway processes utilize limestone or
-------
Table 1 SELECTED OPERATING PARAMETERS FOR SLUDGE-PRODUCING S02 SCRUBBERS
co
o\
VO
Facility
Kansas Power & Light -
Lawrence 6
Kansas Power & Light-
Lawrence 5
Kansas City Power & Light-
Hawthorn 3
Kansas City Power 6 Light-
Hawthorn 4
Commonwealth £dl eon-
Will County 1
City of Key West-
Kansas City Power & Light-
La Cygne
Arisone Public Seirvlce-
Cholla
Louisville Cas & Electric-
Paddy's Run
Dusquesne Light-
PhiHips )
(0
820
125
70
180
160Cd)
32(e)
Nominal
SuLCut Content
of Coal,
3.8
3.8
3.0
3.0
3.5
2 Q
5.3
0.5
3.7
0.4
Nominal
Ash Content
of Coal,
12
12
13
13
15
0 W
22
10
14
10
Primary Method
of PartlcuUte
Marble bed
Marble bed
Marble bed
Marble bed
Venturi
Mechanical
Venturi
Flooded disc
scrubber
Electrostatic
precipitator
Electrostatic
precipitator
Method of
2
Limestone injection-
wet scrubbing
Limestone injection-
wet arrubblng
Limestone Injcctlon-
wet scrubbing
Tall -end limestone
scrubbing
Tall-end limestone
scrubbing
Tall-end limestone
Tall -end limestone
scrubbing
Tail-end limestone
scrubbing
Line scrubbing**'
Lime scrubbing
lirae scrubbing
Double alkali process
Stoichromctrlc
Ratio,
moles Ca/mole S02
1.15
1.15
1.2
1.5
1.5
5 0
1.9
1.0
1.0
1.0
SO2 Removal
Efficiency,
person
60
60
70
70
85
70
eo
90
90
90
pK Range
In Scrubber
9.0-5.3
9.0-5.3
9.0-5.0
5.5-4.5
5.9-5.7
7.5-6.5
6.0-5.6
6.5-5.2
9.0-5.3
9-5
Approximate
Mole Ratio of
Oxygen to S02
In Caa to Scrubber
20
20
20
20
40
30
30
100
30
300
(a) Hot yet visited.
(b) Cyclone boiler.
(c) Boiler fired with Bunker C fuel oil.
(d) 20 percent of gas flow from 790 >W unit.
(e) Four stoker-fired boilers.
(f) Carbide sludge.
-------
The double alkali process attempts to circumvent plugging and
scaling problems in the scrubber. By scrubbing the SCL with soluble
sodium salts and reacting the scrubber effluent with lime, calcium-
sulfur compounds are produced for waste disposal, while yielding a
solution of NaOH and Na?SCL for recycle back to the scrubber.
Although this process has not been applied to a utility boiler, it
is currently in operation on an industrial boiler.
The factors influencing sludge characteristics include size of
power plant, type of boiler, type of fuel burned, sulfur and ash
content of fuel, method of fly ash removal, method of SCL control, stoi-
chiometric ratio of calcium to SO- removed, and SCL removal efficiency.
It is speculated that the pH range in the scrubber and the mole ratio
of oxygen to SCL in the gas entering the scrubber may also affect
sludge characteristics. The size of the power plant affects the
total quantity of sludge produced. The type of fuel burned, sulfur
content of the fuel, and SCL removal efficiency affect the quantity
of calcium-sulfur compounds in the sludge. The type of boiler, ash
content of the fuel, and the method of fly ash removal affect the
fly ash content of the sludge. In some cases, most of the fly ash is
removed in an electrostatic precipitator ahead of the scrubber so
that it can be kept separate from the sludge. In other cases, the
fly ash is removed in either the same scrubber as the SCL or in a
Venturi-type scrubber ahead of the SCL scrubber. The quantity of
fly ash that is removed by wet scrubbing in conjunction with SCL
removal ends up in the sludge. The stoichiometric ratio of calcium
to SCL removed affects the CaCCL content of the sludge. A mole ratio
in excess of one means that unreacted limestone leaves with the
scrubber effluent. If the reactant is lime, any excess presumably
reacts with CCL in the flue gas to form CaCCL.
The calcium-sulfur compounds present in the sludge are calcium
sulfite and calcium sulfate. The relative amount of each depends
870
-------
upon the level of oxidation in the scrubber system. The factors that
appear to affect oxidation include the amount of fly ash removed in
the scrubber, the pH range in the scrubber, and the mole ratio of
oxygen to SC>2 in the gas entering the scrubber. The method of SO
control affects the pH range in the scrubber (limestone slurries are
lower in pH than lime slurries). Also, in the double alkali process,
some sodium salts can appear in the sludge.
Table 2 summarizes the dry sludge production rate on an hourly
basis and the approximate composition of the sludge for the full-
scale plants currently in operation. The data listed in this table
are best estimate calculations based on information obtained during
plant visits. The amount of water in the sludge influences the total
tonnage. The scrubber effluents contain a solids content in the
range of 5 to 15 percent by weight and must be dewatered to a point
where sludge disposal can be economically feasible and ecologically
acceptable. The sludge production rate on a yearly basis depends
upon plant capacity factor and availability of the SO control system.
Assuming a 65 percent capacity factor for each plant and 90 percent
availability for each S02 control system, the plants listed in
Table 2 have the potential to generate sludges at the rate of about
2.5 x 10 metric tons (50 percent moisture) annually. It is estimated
that by 1980 lime/lime stone scrubbing systems may generate as much as
120 x 10 metric tons (50 percent moisture) annually.
The limestone injection systems produce a sludge consisting mainly
of fly ash and calcium sulfate with some calcium sulfite and a little
calcium carbonate. This composition occurs because the fly ash is
removed in the same marble bed scrubber as the S09, and the presence
of fly ash appears to promote oxidation. The calcium carbonate in
the sludge results from a stoichiometric ratio of about 1.2. The
sludge composition for the tail-end limestone scrubbing systems varies
greatly with process parameters. The fly ash content for the Stock
871
-------
Table 2
CHACTERISTICS OF SLUDGE FROM OPERATING S02 SCRUBBERS
Oo
-vl
Facility
Lawrence 4
Lawrence 5
Hawthorn 3
Hawthorn 4
Will County 1
Stock Island
La Cygne
Cholla
Paddy's Run 6
Phillips ^
Mohave 2
(a)
Parma
Rate (dry basis),
metric tons/hr
10.7
34.4
12.4
15.4
17.5
-------
Island sludge is very low because it is an oil-fired boiler, while
calcium carbonate content is high because of the very high stoichio-
metric ratio of calcium to SCL removed. The Will County and La Cygne
sludges contain less fly ash than might be expected because cyclone
boilers produce less fly ash than other types of boilers. The Cholla
sludge does not contain any calcium carbonate because the stoichio-
metric ratio is one. The concentration of calcium sulfate is greater
than calcium sulfite in this sludge because of the presence of fly ash
and the high ratio of oxygen to SCL entering the scrubber. This
latter effect is a direct result of burning low-sulfur coal.
The two lime scrubbing installations visited (see Table 1) present
an interesting contrast. Both installations remove fly ash in
electrostatic precipitators ahead of the S0_ scrubbers and both
scrubbers probably operate in the same pH range. The Paddy's Run
scrubber is a two-stage marble bed and the Mohave scrubber is a
horizontal spray chamber with four spray headers in series. The
sludge at Paddy's Run is practically all calcium sulfite, presumably
because the relatively high pH level in the scrubber prevents oxi-
dation by keeping sulfite out of solution. However, the sludge at
Mohave is practically all calcium sulfate. The low-sulfur coal and
high excess air in this instance cause a very high ratio of oxygen
to SCL in the inlet gas to the scrubber.
WATER POLLUTION AND LAND USE PROBLEMS
The tendency of sludge to pollute surface and groundwater is
largely dependent on its chemical composition. Sludges from lime/
limestone S0_ control processes generally contain CaSO-j * y H^O,
CaSO,* 2H 0, CaCO-, and fly ash. Other minor compounds (e.g.,
sodium, ammonium, and magnesium salts) may be present because they
occur in the lime or limestone, in makeup water, or are purposely
introduced into the system as in the double alkali process. The
873
-------
composition of fly ash varies greatly with coal type, but generally
consists of silica, alumina, and iron oxides with lesser amounts of
alkaline earth and alkali metal oxides and trace amounts of heavy
metal oxides. However, the fly ash content of the sludge would not
be expected to present a greater disposal problem than that of fly
ash alone which has been ponded and landfilled for many years. On
the other hand, the CaSO^' j H?0 content of the sludge could pose
problems because of the chemical oxygen demand of sulfite ions.
The physical and chemical characteristics of the sludge influence
the amount of land required for disposal and future use of the ponded
area. The pond volume requirement for lime/limestone sludge is larger
than that required for fly ash disposal because of the larger packing
volumes of the former. Most fly ash settles compactly requiring only
3 3
about 0.6 m per metric ton of dry solids (20 ft per ton) while
3
scrubber sludges appear to have packing volumes between 1.4 and 2.3 m
3
per metric ton of dry solids (45 and 75 ft per ton). The ratio of
calcium sulfite to calcium sulfate in the sludge is significant to
the packing volume. Sulfite presents a more significant land use
problem than sulfate. Sulfites tend to crystallize in small, thin
platelets which settle to a loose bulky structure that may occlude
a relatively large amount of water. The net result is to increase
the land required for disposal. For example, the Paddy's Run sludge
which consists mainly of calcium sulfite contains only 40 percent
solids after vacuum filtration, while the Mohave sludge which consists
mainly of calcium sulfate contains 65 percent solids after settling
in a clarifier. For a typical 1,000 MW coal-fired station, the
disposal of lime/limestone sludge will require between 12.2 to 24.3
hectares (30 to 60 acres) per year ponded to a 3-meter (10-foot)
depth. The lower figure represents the area required for sulfate
sludge, while the higher figure represents the area required for
sulfite sludge.
874
-------
Currently, ponding and landfilling are the only methods practiced
by utilities for the disposal of lime/limestone sludges. However,
the problems associated with these disposal methods are becoming
increasingly difficult. These problems are related to finding suitable
disposal sites close to the plant and the high cost of removal to a
disposal site away from the plant. The magnitude of the problems tend
to be site specific. In the case of Key West or Will County, there is
little or no land available adjacent to the plant while at Lawrence or
La Cygne, there is abundant available land adjacent to the plant.
The land used for waste disposal may be structurally unstable and
aesthetically objectionable. If the disposal site is not available
for future land use, the land loses its production capacity. There-
fore, additional sludge treatment by chemical fixation and alternative
disposal processes through commercial utilization are being developed.
Unfortunately, chemical fixation can significantly increase the cost
of waste disposal, and limited utilization options are available.
Thus far, the sludge by-products are too expensive to compete with
natural sources of the same or similar material.
CURRENT SLUDGE PROCESSING TECHNIQUES
The present status of sludge processing and disposal by utilities
is in the development stage, with wide variations in technique being
practiced at different locations. In most cases, economics have been
the controlling factor so that the easiest and cheapest method of
sludge treatment and disposal has usually been selected. Therefore,
the disposal method adopted has always been a throwaway method--
ponding or landfilling. Table 3 provides a summary of the sludge
disposal systems currently in use at plants employing lime/limestone
scrubbing. The particular processes selected by a utility are
dictated as stated earlier by site location and, to a lesser extent,
environmental pressures, in addition to economics.
875
-------
Table 3 CURRENT SLUDGE DISPOSAL PRACTICES IN THE UTILITY INDUSTRY
CO
Facility
Lawrence 4
Lawrence 5
Hawthorn 3
Hawthorn 4
Will
County 1
Stock
Island
La Cygne
Cholla
Paddy ' s Run
Phillips
Mohave 2
Parma
Location
Lawrence ,
Kansas
Lawrence ,
Kansas
Kansas City,
Missouri
Kansas City,
Missouri
Lockport,
Illinois
Key West,
Florida
La Cygne,
Kansas
Joseph City,
Arizona
Louisville ,
Kentucky
South
Heights,
Pa.
South Point,
Nevada
Parma, Ohio
Pretreatraent Method
Effluent Chemical Ultimate Disposal Method
Management Oxidation Clarifier Vacuum Filter Centifuge Pond Fixation Ponding Landfill Utilization
Closed loop X X
Closed loop X X
(a)
Open loopv ' X XX
(a )
Open loop X XX
.
Closed loop X X( ' X
Open loop X X
Closed loop X X
(b)
Open loop XX
Closed loop Xx X
Closed loop X X X X
Closed loop(c^ X X X(6^ X X
Closed loop XX X
(a) Closed loop with respect to clarifier and open loop with respect to pond.
(b) Solar evaporation.
(c) Aided by solar evaporation.
(d) Chicago Fly Ash method.
(e) Dravo method.
-------
PRETREATMENT
Pretreatment refers to those processes to which the sludge is
subjected prior to ultimate disposition. There are three broad
approaches to pretreatment: (1) oxidation of calcium sulfite to
calcium sulfate, (2) reduction of moisture content by various
dewatering schemes, and (3) chemical/physical fixation processes. It
is possible for a power plant to employ one or more approaches
depending upon the desired method of ultimate disposal and the
economics involved; however, dewatering is almost always a necessity.
Oxidation improves sludge compaction and can produce a salable by-
product (gypsum); dewatering provides reduced sludge volumes to handle,
manage, or transport to the disposal site; fixation processes convert
the sludge into a relatively nonleachable stabilized material.
Oxidation. The oxidation of the small gel-like sulfite crystals
in the sludge to the larger sulfate crystals improves agglomeration
and compaction and, thus, dewaterability of the sludge. No operating
lime/limestone system in the United States practices intentional
2
oxidation. This is in contrast with the Ando (1973) account of the
Japanese systems where oxidation is practiced extensively in order to
produce gypsum for sale. However, the supply and demand situation
for gypsum is much different in Japan than in the United States.
Incidental oxidation occurs in lime/limestone scrubber systems
because of oxygen present in the flue gas and because of contact
with air in the recirculation tank, clarifier, and pond. Oxidation
appears to be enhanced by low pH, the presence of fly ash in the
scrubber, and a high ratio of oxygen to S02 in the flue gas entering
the scrubber.
Dewatering. The scrubber effluent must be processed to settle
the thixotropic suspension and dewater the solids prior to ultimate
disposition. The solids are recovered and concentrated, while the
clarified liquid effluent is recirculated (closed-loop operation),
877
-------
or discharged to a receiving body of water (open-loop operation). In
some cases, the scrubber effluent is evaporated in the sun (open-loop
operation). Open-loop operation increases the dissolved solids
content of the receiving body of water, except in the case of solar
evaporation, but it reduces plugging and scaling problems in the
scrubber system by permitting more fresh water makeup for washing
critical components such as the mist eliminator. However, open-loop
operation is not permitted in most locations. The plants currently
operating in the open-loop mode are Hawthorn, Stock Island, and
Cholla. Hawthorn is actually operating in what may be described as
semi-open loop since the clarifier overflow is recycled but the pond
overflow is discharged to the Missouri River. The pond overflow at
Stock Island is discharged to the Atlantic Ocean while the pond water
at Cholla is subjected to solar evaporation. Mohave operates in the
closed-loop mode but there are times when no recycle of pond water is
necessary because of solar evaporation.
There are many possible schemes available for sludge dewatering.
However, only three schemes are currently used at full-scale lime/
limestone scrubber installations. These include ponding (or clarifi-
cation) only, clarification followed by ponding, and clarification
followed by a vacuum filtration. Dewatering by ponding only is shown
schematically in Figure 1 and is practiced at Lawrence, Stock Island,
La Cygne, and Cholla. At Lawrence and La Cygne, the relatively high
solids content of the recirculating slurry (9 to 10 percent at
Lawrence and 15 to 20 percent at La Cygne) is kept in suspension by
high power mixers and baffles. A bleed stream from the recycle tank
is sent to the sludge pond and the pond overflow is recycled to the
scrubber system. At Stock Island, a slip stream from the scrubber
hoppers goes to two settling ponds and clarified effluent is discharged
to the sea through overflow pipes; sea water is used in the scrubber
system because fresh water is very expensive at this location. At
878
-------
Cholla, a bleed stream from the scrubber recirculation tank is sent
to two sludge storage tanks; the tanks are emptied to an existing ash
pond about once per shift and the water inflow to the pond is lost by
evaporation. Ponding is not a problem in Kansas and Arizona where
land is available, but it is a very serious problem in Key West where
the entire island is only five square miles in area and the sludge
ponds have a useful life of only 21 days.
The dewatering system of a clarifier followed by a pond is shown
schematically in Figure 2. The use of a clarifier ahead of the pond
tends to decrease the pond area required. The clarifier surface area
2 2
varies from 1 to 10 m /metric ton/day (10 to 100 ft /ton/day) depending
on the settling characteristics of the sludge. This method of
dewatering is practiced at Hawthorn, Phillips, and Mohave. The
clarifier used at Hawthorn is shown in Figure 3.
The dewatering system of a clarifier followed by a vacuum filter
is shown schematically in Figure 4 and is practiced at Paddy's Run
and Parma. The addition of a vacuum filter is a great aid in reducing
the moisture content of the sludge. At Paddy's Run, the clarifier
bottoms contain 23 percent solids and are filtered to produce a cake
containing 40 percent solids for landfill. This filter is shown in
Figure 5. The filtrate is returned to the clarifier.
Alternatively, a centrifuge can be used instead of a vacuum
filter. Although none of the plants listed in Table 3 use a centri-
fuge, it is being used at the Shawnee pilot plant and will be tried
at the limestone scrubbing installation on Mohave Unit 1. The use of a
centrifuge can effect a high degree of dewatering at the expense of a
879
-------
From scrubber
Recycle
tank ~
I
oo
\ Pond /
To scrubber
Figure 1 Sludge dewatering by ponding,
From scrubber
Recycle
tank
do
Clarif ier
To scrubber
\ Pond /
Figure 2 Sludge dewatering by clarification and ponding.
Figure 3 Hawthorn clarifier.
880
-------
From
scrubber
CO
Recycle
tank
To scrubber
To Landfill
Figure 4 Sludge dewatering by clarification and filtration.
'
Figure 5 Paddy's Run filter,
881
-------
relatively high power consumption. Also, the centrifuge internals
may be subjected to undue wear because of erosion.
Chemical Fixation. Chemical fixation stabilizes the sludge by
changing its physical and chemical characteristics so that it can be
landfilled or utilized as a by-product. Currently, Will County,
Phillips, and Mohave are using chemical fixation processes. At Will
County, the fixation is handled by the Chicago Fly Ash Company. The
clarifier bottoms, which are about 35 percent solids, are treated with
10 percent lime and 20 percent fly ash based on dry solids in a redi-
mix truck. The sludge is mixed while it is being hauled one mile for
dumping in a clay-lined basin where it sets up in a few days.
3
Phillips and Jones (1973) have reported the Chicago Fly Ash process
to cost $16.50/metric ton ($15/ton) on a dry solids basis.
Sludge stabilization is going to be studied in an advanced
program at the Shawnee test facility under the direction of the
Aerospace Corporation. Three different proprietary processes will
be studied--Dravo, IU Conversion Systems, and Chemfix. There will
be a large pond for untreated sludge and three small ponds for
treated sludge (one for each process). Each pond will have a leachate
well and a groundwater well to monitor any groundwater contamination.
ULTIMATE DISPOSAL
Utilities can finally dispose of the scrubber sludge by ponding,
landfilling, or utilization. Currently, ultimate disposal is about
evenly divided between ponding and landfilling with no utility producing
a by-product for utilization. Ponding can be used to dispose of un-
treated scrubber effluent, dewatered sludge, or chemically fixed sludge.
Untreated scrubber effluent is ponded at Key West, Lawrence, La Cygne,
and Cholla; dewatered sludge is ponded at Hawthorn; and chemically
882
-------
fixed sludge is ponded at Mohave. Typical pond disposal sites at Key
West and Lawrence are shown in Figures 6 and 7, respectively. The
significance of these figures is the wasted land area occupied by
the ponds and the cracking of the pond surface at Key West. Thus far,
there has not been any study of pond seepage or leaching at full-
scale scrubber installations.
Landfilling of scrubber sludges requires dewatering and
stabilization. Dewatering improves handling characteristics and
stabilization improves the load-bearing strength of the landfilled
site. At Paddy's Run, a filter cake of calcium sulfite containing
40 percent solids is trucked to a borrow pit where it is landfilled
together with fly ash. Now and then the sludge and fly ash are mixed
with a bulldozer. Presumably the fly ash aids in stabilizing the
sludge. At Will County, stabilized sludge is dumped in a seven acre,
clay-lined basin, as shown in Figure 8. As far as is known, there
have not been any leachate studies or load-bearing studies at either
site. Thus far, Commonwealth Edison has not found a private
landfill operator willing to take the waste material.
Utilization of scrubber sludges as building material has been
demonstrated as technically feasible. However, economic factors
have discouraged utilities in the United States from practicing any
of the utilization approaches in a full-scale application. In 1972,
sludge from the Lawrence site was transported to Dulles Airport and
mixed with fly ash for use as a parking lot pavement. This test
resulted in the sponsoring of further research by the Federal
Highway Department. The first large-scale application of sludge
utilization will be at the limestone scrubber installation at Mohave
Unit 1 where IU Conversion Systems will convert the sludge into a
building material.
383
-------
Figure 6 Key West
pond.
Figure 7 Lawrence
pond.
Figure
Will County
sludge basin.
834
-------
CONCLUSIONS AND RECOMMENDATIONS
The disposal of sludge is probably the major problem in the
application of lime/limestone scrubbing technology. The dozen plants
that are employing this technology have tried to dispose of the
sludge in the most expedient manner. Thus far, the success of the
sludge disposal operation has largely depended on plant location.
In areas where abundant land adjacent to the plant is readily
available, disposal by ponding appears to be an adequate solution.
However, in other locations the sludge disposal problem remains
unresolved. Chemical fixation and landfill may provide the answer;
or perhaps by-product utilization is the way to proceed. The advanced
program at the Shawnee test facility to study sludge stabilization,
pond leaching, and groundwater contamination should prove to be
extremely valuable. However, more extensive work at existing full-
scale installations practicing either ponding or landfilling would
appear worthwhile. It would also be highly desirable to encourage
further development of sludge by-products.
REFERENCES
(1) Sulfur Oxide Control Technology Assessment Panel (SOCTAP),
"Projected Utilization of Stack Gas Cleaning Systems by Steam
Electric Plants", Final Report, prepared for Federal Interagency
Committee for Evaluation of State Air Implementation Plans,
Washington, D. C., PB-221 356 (April, 1973).
(2) Ando, J., "Utilizing and Disposing of Sulfur Products from Flue
Gas Desulfurization Processes in Japan", presented at the Flue
Gas Desulfurization Symposium, New Orleans (May 14-17, 1973).
(3) Phillips, N. P., and Jones, D. C., "Evaluation of Lime/Limestone
Sludge Disposal Options", prepared for the Environmental
Protection Agency by Radian Corporation, Austin, Texas
(November 19, 1973).
(4) Lord, W. H., "Transportation of Sludges and Off-Site Disposal",
Dravo Corporation, proceedings from Electrical World Conference
on the Problem Beyond Removal - Waste Disposal in Utility
Environmental Systems, Chicago (October 30-31, 1973).
885
-------
ENVIRONMENTALLY ACCEPTABLE DISPOSAL
OF FLUE GAS DESULFURIZATION SLUDGES:
THE EPA RESEARCH
AND DEVELOPMENT PROGRAM
BY
Julian W. Jones
Control Systems Laboratory
Environmental Protection Agency
Research Triangle Park, North Carolina
Presented at
EPA Control Systems Laboratory
Symposium on Flue Gas Desulfurization
Atlanta, Georgia
November 4-7, 1974
887
-------
This paper has been reviewed by the Environmental Protection
Agency and approved for presentation. Approval does not signify
that the contents necessarily reflect the views and policies of
the Agency, nor does mention of trade names, commercial products,
or commercial processes constitute endorsement or recommendation
for use.
CONTENTS
Page
1.0 Introduction 891
2.0 Summary 891
3.0 Definition of the Problem 896
4. 0 Current Approaches to Disposing of or
Utilizing Scrubber Sludge Materials 907
5. 0 Current EPA R and D Programs 912
6.0 References 926
888
-------
ABSTRACT
A recent assessment has been made concerning the environmental
and economic factors associated with disposal of sludge from non-
regenerable flue gas desulfurization (FGD) processes. Lime/limestone
scrubbing systems are expected to comprise the majority of FGD process
installations on power plants through 1980, producing approximately
35,000,000 short tons (dry, excluding coal ash) of sludge annually by
1980. This compares to an expected coal ash production of about
83,000,000 tons (dry) annually by 1980. Since uncontrolled disposal of
raw (untreated) sludge presents potential water pollution and land
reclamation problems, environmentally sound disposal techniques (e.g.,
chemical fixation) need to be employed. Most of the chemical
constituents of concern originate in the coal, indicating that FGD
processes provide a multi-pollutant control capability. Because of
considerable variation in FGD system applications, FGD sludge properties,
and electric utility sludge disposal approaches, EPA has undertaken
additional research and development efforts to increase the environmental
and cost effectiveness of current sludge disposal techniques. Several
programs are underway at National Environmental Research Centers (NERCs)
in Research Triangle Park (RTP), N.C.; Cincinnati, Ohio; and Corvallis,
Oregon. The most broadly-based program is a NERC-RTP contract with
The Aerospace Corporation. Initiated during late 1972, this effort
includes: quantification of potential water pollution and land reclamation
problems; technical and economic evaluation of currently available
treatment/disposal techniques; and support of an EPA field study of
sludge disposal at TVA's Shawnee Steam Plant. Results to date have
verified the need for environmentally sound disposal techniques; the
most practical are landfilling of chemically fixed sludge and disposal
of untreated sludge in ponds lined with an impervious material such
as clay, plastic, or rubber. Disposal cost estimates range from $2
to $9 per ton (wet, 50 percent solids). Although less expensive than
fixation, ponding may cause subsequent land reclamation costs to be
incurred. The Shawnee field study, which includes evaluation of the
Chemfix, Dravo, and IU Conversion Systems sludge fixation processes,
was initiated in September 1974. In mid-1974, NERC-Cincinnati initiated
an interagency agreement with the U.S. Army Corps of Engineers to study
the leachability and durability of raw and chemically fixed hazardous
industrial wastes and FGD sludges. Also in mid-1974, NERC-Cincinnati
initiated an interagency agreement with the U.S. Army Materiel Command
to study migration of chemical constituents from industrial and FGD
sludges through soils. NERC-Corvallis has issued a grant to Aerospace
to ascertain the water pollution and reuse potential of treated and
untreated scrubber liquors; the study will be completed in January
1975.
889
-------
ACKNOWLEDGEMENTS
The author is especially indebted to several individuals for their
technical assistance in preparing this paper, including the following:
Frank T. Princiotta
Richard D. Stern
EPA Control Systems Laboratory
Norbert Schomaker
George Huffman
Robert Landreth
Michael Roulier
} EPA Solid & Hazardous Waste
Research Laboratory
Dennis Cannon
EPA Pacific Northwest Environmental
Research Laboratory
Jerome Rossoff
} The Aerospace Corporation
The author is also appreciative of the patience, helpfulness and
skill of Mrs. Carolyn Fowler in typing several drafts and the "final
edition".
890
-------
1.0 INTRODUCTION
A major problem inherent in any flue gas desulfurization (FGD)
system is the necessity for disposing of or utilizing large
quantities of a sulfur product. Application of FGD systems in
the United States is accelerating; the majority of these are
lime/limestone wet scrubbing systems producing a sludge by-product.
Because of environmental, economic, and other concerns related to
the disposition of this FGD sludge, there have been considerable
research, development, and demonstration activities, both in the
public and private sector. In this paper, the FGD sludge disposition
problem is defined and quantified, and efforts to solve the problem
are discussed. In the discussion of these efforts, emphasis is placed
on U.S. Environmental Protection Agency (EPA) research and development
programs.
2.0 SUMMARY
Based on a review of sulfur oxides (SOX) control technologies,
FGD systems installed on units burning high sulfur coal is the major
alternative to scarce clean fuels between now and 1980. Installation
of FGD systems is presently demand-limited; regulatory pressures are
expected to change this to a supply-limited situation sometime between
1975 and 1977. This situation is expected to continue through
about 1980. Under these conditions, it is estimated that FGD
control will most likely be installed on 90,000 Mw or about
35 percent of total estimated coal-fired utility generating capacity
by 1980. Most of the 90,000 Mw capacity is expected to be controlled
by lime/limestone wet scrubbing systems producing a throwaway
sludge. If all of the installations are lime/ limestone systems,
up to 118,000,000 metric tons/year (130,000,000 tons/year) of
wet sludge including ash will be produced by 1980 (assuming
limestone scrubbing, combined sludge + ash disposal, 50 percent solids)
In placing sludge production in quantitative perspective, it
has been determined that approximately 346,000 metric tons (381,000
tons) of scrubber sludge and 307,000 metric tons (338,000 tons)
of coal ash, on a dry basis, will be produced annually by a
typical 1000 Mw coal-fired plant. This represents an increase of
about 110 percent over the solid waste produced by a plant with
ash particulate controls only. However, when compared to the land
usage associated with a typical 1000 Mw coal-fired plant, the waste
disposal area for sludge and ash is only about one percent of the
total. Nevertheless the environmental impact of each individual
land usage varies considerably, requiring consideration of many
factors other than the relative areas involved. When compared
to other solid wastes, the quantity of dry sludge (excluding ash)
projected to be produced in 1980, although larger than the quantity
891
-------
of some wastes, is considerably less than that of others. For
example, the dry sludge production rate is anticipated to be
about 40 percent of the production rate of all coal ash from
power plants.
In placing scrubber sludge in qualitative perspective,
data have been assembled on physical and chemical properties of
various scrubber sludge materials. These data show a wide variation
in properties of sludges from different units. However, in general,
the data indicate that raw (untreated) scrubber sludges, depending
on the amount of calcium sulfite present, are difficult to dewater
and have little or no compressive strength as produced. These properties
present potential disposal area reclamation problems. In addition,
the raw sludge liquors contain dissolved chemical species
in concentrations which considerably exceed water quality criteria.
These include mercury, selenium, boron, chloride, sulfate, and
total dissolved solids. Most of the constituents of concern
originate in the coal, indicating that FGD processes provide a multi-
pollutant control capability. These problems must be addressed
by the lime/limestone process user. Several approaches have been
or are being investigated, including commercial utilization of
the sludge and a variety of currently available sludge disposal
techniques.
Technology exists for utilization of scrubber sludge in
products such as mineral wool, bricks, gypsum, road base materials,
artificial aggregate, and aerated concrete. However, because of
major marketing inhibitions, appreciable commercial utilization
of sludge is unlikely. The only important near term alternatives
appear to be disposal by ponding and disposal by landfill.
Based on current and planned lime/limestone FGD system
installations, utilities are using ponding and landfill disposal
techniques, with and without fixation processes. The choice is
usually based on environmental/economic considerations peculiar to
each plant site. For this reason EPA has undertaken several
programs to develop the additional information and data needed to
assure more general application of sludge disposal technology and
to further minimize environmental effects, at reasonable costs.
The most broadly-based EPA program relating to scrubber sludges
is the National Environmental Research Center - Research Triangle
Park, N.C. (NERC-RTP) study with The Aerospace Corporation. This
study is designed to identify environmental problems associated
with scrubber sludge disposal, to assess current sludge disposal
technologies, and to make recommendations regarding alternate
disposal approaches. The effort includes technical support of an
EPA/TVA sludge disposal field evaluation currently underway at the
TVA Shawnee Steam Plant.
892
-------
In the Aerospace Corporation study, sludge samples have been
analyzed from three scrubbers involving lime/Eastern coal, limestone/
Eastern coal and limestone/Western coal. (Samples from two additional
scrubbers were analyzed under a NERC-Corva.llts study. Results of
these analyses have also been assessed as part of the NERC-RTP study.)
A preliminary assessment was made with regard to any potential
environmental problem that might be posed if the sludge liquors
entered water supplies by comparing the liquor analyses with the
EPA Proposed Public Water Supply Intake Criteria (October 1973).
Based on the concentrations of several trace metals and major dissolved
species (previously described) the assessment strongly indicated
that environmentally sound techniques, are needed for scrubber
sludge disposal.
Detailed physical characterization of various samples has
also been performed under the Aerospace Corporation study. Results
of the characterization have verified the water retentive nature of
calcium sulfite and have quantified the effect of moisture on sludge
physical properties. For example, at 65 percent solids content,
sludges will support personnel, and at solids content greater than
70 percent, sludges will support heavy equipment. However, the
cost of dewatering the sludge and maintaining the disposal site at
the necessary degree of dryness may or may not compare favorably
with the cost of an alternative disposal technique such as chemical
fixation. Data from chemical fixation processors indicate that
treated sludge quickly attains a compressive strength which can
make the disposal site reclaimable for either structural or recreational
use. Tests are currently underway in the Aerospace study to independently
verify the performance of these chemical fixation techniques.
Based on available information, the best environmentally sound methods
currently available for sludge disposal are landfilling of chemically fixed
sludge and disposal of untreated sludge in ponds lined with an
impervious material, such as clay, plastic, or rubber. Estimates
from various sources indicate disposal costs of from $2 to $9/(short)
ton for chemical fixation and from $2.50 to $4.50/ton for ponding,
excluding possible subsequent pond reclamation costs. In a typical power
plant application, a $5/ton disposal cost would be equivalent to 1.12
mills/Kwhr.
Initial results of the Aerospace contract were reported in
May 1974. A second report should be issued in mid-1975. The
final report of the effort is expected to be released in late 1976.
893
-------
In the EPA field study, sludges will be obtained from 10 Mw
lime/limestone pilot scrubbers at the TVA Shawnee Steam Plant near
Paducah, Kentucky, and will be placed into five ponds dug in the clay
soil nearby. One pond will receive raw lime sludge; one pond will
receive raw limestone sludge; one pond will receive lime sludge
chemically conditioned by IU Conversion Systems, Inc. (IUCS); one
pond will receive limestone sludge chemically conditioned by
Dravo Corporation; and one pond will receive limestone sludge
chemically conditioned by Chemfix (Division of Environmental
Sciences, Inc.). All fixation will be performed on-site during
the period October - December 1974. Each pond will have a
leachate well and a ground water well. Tests will be performed
to evaluate the environmental acceptability of disposal of both
untreated and treated sludge in clay-lined disposal areas.
Preliminary results are expected in mid-1975; a final report
will be issued in late 1976.
Two programs were initiated in mid-1974 by NERC-Cincinnati
to evaluate the environmental effects of FGD sludge disposal. One
of these is an interagency agreement with the U.S. Army Corps of
Engineers' Waterways Experiment Station in Vicksburg, Mississippi.
Under this agreement, the leachability and durability of raw and
chemically fixed hazardous industrial wastes and FGD sludges are
being studied. Five industrial sludges and up to six FGD sludges
are being obtained for the study. Results of laboratory studies
are expected to be reported in mid-1975. Interim results of field
studies will be reported in mid-1976; final field study results will
be reported in mid-1977.
The second program is also an interagency agreement, with the
U.S. Army Materiel Command's Dugway Proving Ground, Dugway, Utah.
Under this agreement research is being conducted to determine the
extent to which heavy metals and other chemical constituents from
thirteen industrial and three FGD sludges could migrate through the
soil in land disposal sites. After initial screening tests with a
variety of U.S. soils, leachate column studies will be performed with
two selected (best and worst) soils. Long-term permeability tests
with selected clays are also planned for the FGD sludges. Results
of the screening tests will be reported in mid-1975. Preliminary
results of the column studies are expected to be reported in late
1975; the final results are expected to be reported in late 1976.
894
-------
NERC-Cincinnati is also currently considering a full-scale
FGD sludge disposal demonstration program with a utility which uses
high sulfur coal.
NERC-Corvallis (Oregon) has issued a grant to Aerospace
Corporation to determine the implications of open-loop or partially
open-loop operation of lime/limestone FGD systems. Analyses of
various sludge liquors will be performed and technologies for
liquor treatment will be evaluated. These data will be used to
ascertain the water pollution and reuse potential, for various
plant uses, of treated and untreated scrubber liquors. The final
report is expected to be issued in mid-1975.
The EPA organizations are closely coordinating the sludge
disposal research and development efforts just described. These
efforts represent a multi-disciplinary approach, since personnel
with expertise in air pollution control (NERC-RTP), solid waste/
residuals management (NERC-Cincinnati) and water treatment/water
quality (NERC-Corvallis) are involved.
895
-------
3.0 DEFINITION OF THB PROBLEM
3.1 Availability of Alternative SOX Cont ;ol Options
United States Department of the Interior data indicate that net
electric generation by fossil-fueled power plants will increase from
1310 billion Kwhrs in 1971 to 1950 billion Kwhrs in 1980.l
To meet this rapid increase in demand, the electric utility industry
will have to consume large additional quantities of fossil fuels. The
utilities will have to do this, however, without violating air pollution
emission restrictions on sulfur oxides, nitrogen oxides, and particulates.
Sulfur oxide restrictions require that the utilities make a choice
from several alternatives. These include low-sulfur fuels,
fuel cleaning and conversion, and flue gas desulfurization (FGD).
The amount of natural gas and low-sulfur fuel oil available to
electric utilities will be supply-limited at least until 1980.1
In addition, although abundant low-sulfur coal reserves exist (primarily
in the Western states), availability in the near term will be hampered
by the mining industry's difficulty in expanding rapidly and the high
transportation costs in delivering coal to Eastern and Mid-Western
regions where the greatest demand exists. Differences between
Western and Eastern coal characteristics could also cause operating
problems in Eastern plants.
Since low-sulfur fuel availability is inadequate, other alternatives
for meeting emission restrictions must be considered. Technological
developments in fuel cleaning, advanced combustion, and fuel conversion
areas have been rapid in the past few years. However, since none of
these schemes has advanced past the pilot plant stage, it is unlikely
that any of these processes will provide a significant percentage of the
low-sulfur fuel needed in this country by 1980. Therefore, the only major
alternative to low-sulfur fuel between now and 1980 is flue gas
desulfurization (FGD).2 The leading FGD processes include
lime/limestone scrubbing, sodium scrubbing with thermal regeneration
(Wellman-Lord), magnesium oxide scrubbing, and catalytic oxidation
(Cat-Ox). The lime/limestone scrubbing processes produce a
throwaway product (non-regenerable processes), whereas the
other three processes produce a saleable product (regenerable
processes). The potential demand for these processes is examined
below.
3.2 Potential Demand for Lime/Limestone Scrubbing
Installation of flue gas desulfurization systems in the utility
industry is presently demand-limited and Is expected to remain
896
-------
demand-limited through 1975. The total generating capacity
controlled by the end of 1975 is projected to be no more than
10,000 Mw. (Current commitments through 1977 total about 22,000 Mw.)
Because of regulatory pressures, the installation of FGD systems
should become supply-limited sometime between 1975 and 1977. From
1977 to 1980 the installation of flue gas desulfurization systems
will almost certainly be supply-limited. Based on regulatory
pressures and expanding generating capacity coupled with a clean
fuels deficit, it has been forecast that a maximum of over 130,000 Mw
of installed coal-fired generating capacity will need to be controlled
by FGD systems by 1980. However, the most likely demand figure by
1980 is estimated at 90,000 Mw of FGD control, or about 35% of
total estimated coal-fired generating capacity. In the post-1980
period, depending on the commercial availability and viability of
alternate clean fuel technologies, FGD systems could be installed to
the extent that they approach demand requirements. Recent projections by
EPA for the need for coal-fired utility FGD systems are shown in
Figure I.3
In the late 1970*s the projected annual rate of application
of regenerable FGD systems is expected to exceed that for the
non-regenerable FGD systems. These systems produce a variety
of sulfur-containing by-products which represent alternatives
to the non-regenerable sludge, including elemental sulfur,
sulfuric acid, gypsum, sodium sulfate, ammonium sulfate, and
liquid S02.However, based on current emphasis, technology availability,
and lead time considerations, most of the FGD systems installed
by 1980 are expected to be lime/limestone. Other applications
of lime/limestone scrubbing (e.g., oil-fired utility boilers,
coal-fired industrial boilers) could make the 90,000 Mw projected
for control by these processes a realistic figure.
3.3 Quantification of the Problem and Comparison with Analogous
Environmental Problems
One of the major problems inherent in any FGD system is the
necessity to dispose of or utilize large quantities of a
sulfur product. Lime/limestone (and double alkali) scrubbing
systems generate throwaway sludge products with little commercial
value projected at the present time.
897
-------
125
00
UD
00
CO
o
rH
X
CO
I
B
H
§
U
100 --
75 --
50
25 --
1975
1976
Notes;
Curves include new and existing plants requiring controls
to achieve either primary standards or new source perform-
ance standards.
*5
Based on pessimistic projections for new low sulfur coal
supplies and minimal redistribution of existing supplies.
Based on optimistic projections for low sulfur coal
supplies and maximum redistribution of existing supplies.
1977
1 1978
TIME,YEAR
1979
1980
1981
Figure 1. CUMULATIVE NEED: FGD FOR COAL FIRED POWER PLANTS
-------
A power plant SC^ scrubbing system can be designed with the
alternatives of collecting flyash simultaneously with the flue gas
scrubbing operation or of collecting fly-ash upstream of the scrubbing
operation by precipitators and/or mechanical collectors. Additionally,
the ash may be disposed of separately or with the scrubber sludge.
As yet, no consistent approach has been taken by the utility
industry. For coal-fired installations where efficient particulate
removal does not take place upstream of the wet lime/limestone absorber,
scrubber sludges can contain large quantities of coal ash.
The amount of sludge generated by a given plant is a function of
the sulfur and ash content of the coal, the coal usage, the load
factor (on-stream hours per year), the mole ratio of additive to
SO^, the SCL removal efficiency of the scrubbing system, the
composition of the sludge (e.g., sulfite/sulfate ratio), and the
moisture content of the sludge. Table 1 shows typical quantities
of ash and sludge produced by a plant burning the national average
(3 percent S, 12 percent ash) coal projected for 1980, and using
a lime/limestone scrubbing system. Quantities for combined
and separate disposal of sludge and ash are given to show the
effects of these options.
Assuming the projected 90,000 Mw of FGD control is accomplished
entirely by lime/limestone installations and using the national average
annual sludge production rate per 1000 Mw of controlled generating
capacity, the amount of wet sludge and ash (limestone sludge, combined
disposal, 50 percent moisture) that will have to be disposed of
annually by 1980 could be as high as 118,000,000 metric tons/year
(130,000,000 tons/year). (It is unlikely that all coal-fired utility
FGD installations will be lime/limestone systems. However, the
majority are expected to be and other applications, e.g., oil-
fired utility boilers, coal-fired industrial boilers, could make
the projected sludge production figure quite realistic). Depending
on the viability and commercial availability of alternate clean
fuel technologies, sludge production rates could substantially
increase in the post-1980 period.
Although the quantities of sludge produced by a large power plant
are considerable, they should be put in perspective by the comparisons
discussed below. Table 2 shows the annual land and solid waste impact
of a 1000 Mw coal-fired electric energy system equipped with flue gas
desulfurization (FGD) for SOX and particulate removal.4 The
coal mining operations appear to have the greatest impact in terms
899
-------
TABLE 1. TYPICAL AiNwUAL PRODUCTION OF ASH AND SLUDGE
BY A 1000 MW COAL-FIRED GENERATING STATION CONTROLLED
WITH LIME/LIMESTONE FLUE GAS DESULFURIZATION SYSTEM
(Quantities In Short Tons)3
Coal Ash, dry
Coal Ash, wet (80% solids)
338,000
422,000
Limestone Sludge,
CaS03-l/2H20
CaS04-2H20
CaCOn (unreacted)
TOTAL
Limestone Sludge, wet
(50% solids)
Limestone Sludge 4- Ash, wet
(separate disposal)
Limestone Sludge + Ash, wet
(combined disposal, 50% solids)
260,000
29,000
92,400
381,400
762,800
1,184,800
1,438,800
Lime Sludge,
CaS03-l/2H20
CaS04-2H20
CaO (unreacted)
TOTAL
Lime Sludge, wet
(50% solids)
Lime Sludge + Ash, wet
(separate disposal)
Lime Sludge + Ash, wet
(combined disposal, 50% solids)
260,000
29,000
22.200
311,200
622,400
1,044,400
1,298,400
Assumptions:
Coal:
Plant:
3.0%S; 12% ash
6400 iS, 0.88 lb coaA
yr kw-hr
Scrubber:
85% S02 removal;
1.0 CaO/S02 mole ratio;
1.2 CaC03/S02 mole ratio
1 short ton = 0.907 metric ton.
Sulfite/sulfate ratio based on performance of Chemico scrubbing
unit at Mitsui Aluminum Company, Japan.
900
-------
TABLE 2. COMPARATIVE ANNUAL LAMP AHD SOLID WASTE IMPACT OF 1,000 MW ELECTRIC ENERGY SYSTEM (0.75 LOAD FACTOR)
(Lou Levels of Environmental Controls Except for Installation of a Limestone FGD System for SOX and Partlculate Removal)
Land Affected,
acres6
Solid Waste
Produced ,
short tonsc
Environmental
Impact
\D
o
Typical Tech-
nique^) Avail-
able to
Minimize
Impact
MininK (Coal)4
Deep
9,120
97,141 (wet,
97% solids)
with acid
drainage
sludee)
1) potential
land degra-
dation due
to subsi-
dence; 2)aci<
mine drainage
water pollu-
tion problem!
l)no well
developed
cost-effec-
tive tech-
nology to
control sub-
sidence;
2)neutraliza-
tion of mine
drainage
with line
Surface
lU ,010
5,7*5,000 (wet,
98% solids)
(2,762,328
with acid
drainage
sludge )
1) mined lane
made barren
precluding
wildlife
habitat ,
recreation
and most
other uses;
2)acid mine
drainage
water pollu-
tion problems
Dintensive
land recla-
mation can
restore most
strip-mined
land;2) neu-
tralization
of mine drair
with lime
Processing*
161
1*5^,092 (wet,
$9% solids)
1) culm piles;
2) water pollu-
tion: a) acid
drainage ;
b)siltation-,
3)air pollu-
tion: a)nts-
charges SOj, CO
&HgS; bjpotentia
spontaneous com-
bustion
compacting in
holes, mines,
quarries , etc .
age
Transport
2,213
0
use of land for
railroad beds
1
N/A
Conversion^
(plant site)
350
0
use of land for
power plant site
M/A
aSee Table 1 for assumptions (also includes ash)
Land affected is expressed as a time average of Che amount of land in use over 30 years.
Fixed land is taken at its full amount; average variable use (waste storage) is IS times
the annual Incremental damage.
Limestone FCD
System8
Untreated
Ponded Sludge
367
(30 ft. depth)
l.liliO.OOO (wet,
50% solids)
1) potential
groundwater
pollution
problems;
2)land poten-
tially made
useless if
sludge not
treated or
permanently
dewatered
1) although
reclamation is
feasible, no
well developed,
cost-effective
U
Transmission
17 ,188
0
use of land foi
transmission
line right of
way
N/A
2) sound pond manage-
ment, use of imperme-
able pond
liner and operation
of FCD system in
closed-loop mode can
minimize water pollution. (As an
ponding, chemical fixation with
to have potential for solving bo
and land reclamation problems . )
Totals
Deep
29,399
1,991,233
N/A
N/A
Surface
34,289
lt,656,092
N/A
N/A
alternative to
landfill appears
th water IpbUution
1 short ton - 0.907 metric ton.
-------
of land use and environmental effects. Although the right-of-way
required for transmission lines actually consumes more land than coal
mining, this land is still available for some other uses and, aside
from aesthetics, the environmental effects are minimal.
The area required for the site of a plant equipped with a lime/
limestone FGD system and using ponding for combined sludge
and ash disposal would be just over 2 times as much total area
as a plant site without the SOX and particulate control system.
For comparison, this same plant site area would be about 1.5
times the total area of a plant with particulate control only
and ash disposal by ponding.
Table 2 also shows that large quantities of wastes are involved
in coal mining and processing operations. It can be seen that the
FGD system will produce about 3 times as much waste as deep mining,
but only about half that produced by strip mining.
Table 3 presents a semi-quantitative comparison of major U.S.
solid wastes on a dry basis. In addition to quantities of waste,
disposal methods and potential environmental problems are shown.
Quantities are not directly comparable in every case since they
are based on many different sources and time periods.
The quantities of dry sludge (excluding ash) projected to be
produced in 1980 by plants using lime/limestone FGD systems are
about 40 percent of projected 1980 production rates of coal
ash from all power plants. They are about 34 percent of recent
production of ore wastes. However, the quantities are greater
than those for municipal sewage sludge, phosphate rock slime,
gypsum from fertilizer manufacture, or acid mine drainage sludge.
As shown in Table 3, ponding and landfilling provide the
major mechanisms of disposal of most waste products. In terms
of environmental effects, these disposal mechanisms have many
points of similarity for the various industries. In some cases,
land use for waste disposal has destroyed wildlife habitat and is
aesthetically objectionable. In addition, all wastes have the
potential for varying degrees of surface and groundwater pollution
depending on their chemical compositions and solubilities, and the
location, design, and operation of the disposal site. To reclaim
the disposal site, most of the stable wastes require only a cover
material to support growth of vegetation and to prevent eventual
erosion of the wastes by run-off water. However, some wastes
(e.g., phosphate rock slime, sewage sludge) are resistant to
dewatering. In these cases, the disposal sites could become only
902
-------
TABLE 3. COMPARISON OF MAJOR SOLID WASTE DISPOSAL PROBLEMS
Waste Material
Municipal and Industrial
Refuse 5
Culm Piles6
Mineral Ore Wastes
Coal Ash (including
coal ash from lime/limestone
scrubbing installations)
Lime/limestone Scrubber
Sludge (excluding coal ash)
o
Taconite Tailings
Gypsum from Fertilizer
Manufacture 9
Quantity Disposed
Annually In Referenced Year,
metric tons, drv
a
270,000,000 (1973)
> 100, 000, 000 b (1969)
89,000,000 7 (1970)
75,500,000d(l980)
32,000,000 6 (1980)
55,000,000 (1971)
25,000,000 f (1973)
Method of
Disposal
Landfill ing,
incineration
Surface piles,
landfill ing
Ponding,
surface piles,
landfill ing
Ponding,
landfill ing
Ponding,
landfill ing
Ponding; lake
dumping (Reserve
Mining Company
only)
Ponding,
surface piles
Land Use or
Reclamation
Considerations
Cover material neede
to support vegetatio
Cover material requi
for plant growth.
Provision for col lee
of drainage.
Needs cover material
Needs cover material
Untreated sludge
difficult to dewater
Fertilization, mulch
etc. required for
reclamation of ponds
Needs cover material
to support vegetatio
and mskp aocthoti rsl
VO
o
acceptable.
-------
TABLE 3. (Continued) COMPARISON OF MAJOR SOLID WASTE DISPOSAL PROBLEMS
Waste Material
Quantity Disposed
Annually In Referenced Year
metric tons, drv
Method of
Disposal,,
Land Use or
Reclamation
Considerations
Municipal Sewage Sludge
10
Phosphate Rock Slime
n
Acid Mine Drainage Sludge
12
11,000,000 (1980)
1,800,000 f (1967)
Ponding,
landfill ing
Ponding
410,000 9
Ponding
Hard to dewater.
Difficult to develop.
Hard to dewater
(settles to only 30%
solids after years).
Not established whether
dried solids will
support vegetative
growth.
Hard to dewater.
aExcludes agriculture & mining wastes
^Bituminous coal only
CRock wastes from metal ores only; no processing wastes included
Ash, 6400 hrs/yr, 245,000 Mw installed coal fired generating capacity (1980), 0.88 Ib coal/Kwhr
S, 12% Ash, 6400 hrs/yr, 85% SO? Removal, 0.88 Ib coal/Kwhr, 90,000 Mw controlled generating
capacity, 1.2 CaC03/S02 (inlet) mote ratio, 10% oxidation
^80% disposed of in Florida.
9Most acid mine drainage comes from abandoned mines and these are not treated.
-------
a temporary storage site. Although land reclamation is feasible,
no well-developed technology, other than fixation, appears to be
established. However, with proper site selection and design (including
a permanent, impermeable liner) and sound operating practices,
surface water and groundwater pollution can be avoided.
3.4 Nature of the Material
Scrubber sludge is normally a moist, grey material containing
varying amounts of flyash from the combustion process. (If no
flyash were present, the material would be white.) The physical
behavior of the sludge is primarily dependent on the percentage
of retained moisture. However, the degree to which a sludge can
be dewatered is dependent on the relative quantities of the major
solid constituents present, i.e., calcium sulfite, calcium sulfate,
calcium carbonate (and/or calcium hydroxide), and flyash. The
moisture-retaining tendency has been attributed to the thin, platelet-
like crystal structure of calcium sulfite hemihydrate (CaS03*l/2 H^O) .
The result of this moisture is thixotropic behavior, with the sludge
exhibiting little or no compressive strength. This behavior can
be substantially modified by several techniques, which are discussed
in Section 5.1.
Table 4 compares a sludge with high calcium sulfite content
(typical of high-sulfur Eastern coal-fired plants) to a sludge with
high calcium sulfate content (typical of low-sulfur Western coal-fired
plants).
TABLE 4. COMPARISON OF TYPICAL SLUDGE DEWATERING PROPERTIES13'14
Sludge Type
High CaS03-l/2H20
(low flyash)
High CaS03-l/2H20
(high flyash)
High CaS04-2H20
(low flyash)
Approximate Degree
of Dewatering, percent
solids
Settling
30-35
35-40
60-65
Filtration
50
55-60
80
Approximate Percent
Solids for Optimum Compaction
80
80
90
905
-------
It should be noted that the presence of significant amounts of
flyash will slightly alter the dewatering characteristics by
increasing the solids content of sulfite sludges but slightly
decreasing the solids content of sulfate sludges (the percent solids
for optimum compaction of flyash is about 80). Another significant
point to'note is the wide gap between the percent solids for filtered
calcium sulfite sludge and the percent solids required for optimum
compaction of that material. This is an important consideration in
examining disposal options. A final point is that since sludges vary
from plant to plant, the physical behavior will vary between the
figures shown in Table 4.
The major constituents of scrubber sludge range from very slightly
water soluble (e.g., calcium carbonate, calcium sulfate) to water
insoluble (flyash). Although some of the constituents are soluble
in acid, the presence of alkaline materials makes their dissolution
unlikely. The sludge liquors contain dissolved species in concentrations
which vary with individual solubilities and the rate at which the
species enter the scrubbing system. The prevalent ions include calcium,
sulfate, chloride, and magnesium. The liquors and solids also
contain trace elements, primarily from the flyash and coal. From
an environmental standpoint, the constituents of the sludge liquor
are the most significant, since they represent the potential ground
water pollution problem for untreated sludge disposal on land.
Based' on data from sludge liquors analyzed thus far,13,14 the following
constituents are of concern:
(1) Mercury
(2) Selenium
(3) Boron
(4) Chloride
(5) Sulfate
(6) Total Dissolved Solids (primarily calcium, chloride, sulfate)
All of these constituents have appeared substantially in excess of water
quality criteria. These results strongly indicate that environmentally
sound scrubber sludge disposal techniques (such as chemical fixation or
ponds lined with impermeable materials) need to be employed.
In summary, the nature of the sludge indicates problems which
must be addressed by the lime/limestone process user. Various
approaches have been or are being investigated, including commercial
utilization of the sludge and a variety of sludge disposal techniques.
906
-------
4.0 CURRENT APPROACHES TO DISPOSING OF OR UTILIZING SCRUBBER SLUDGE
MATERIALS
4.1 Commercial Utilization
Investigations into the potential commercial utilization of power
plant desulfurization sludges have been made by numerous government
and private organizations. These organizations include Federal agencies;
research centers; universities; commercial research, processing and
sales corporations; national trade associations; and private researchers.
The results of their efforts have been disseminated through symposia,
technical reports, newspapers and periodicals. A review of these
references indicates that the general consensus is that, although
some commercial usage is feasible from a technical and economic stand-
point, the potential outlet is so small that the vast majority of the
sludges will not be marketed.
Attempts have been made to develop technology to apply sludges
to the existing ash product market or to develop new applications in
which the sludges might be used. Such developments and investigations
have been reported by research centers including the Coal Research
Bureau - West Virginia University, Combustion Enginering, IU Conversion
Systems, Inc. and Michigan Institute of Technology. These developments
include mineral wool, bricks, sintered concrete products, soil
amendment, sulfur recovery, gypsum, mineral recovery, road base
materials, parking lot materials, artificial aggregate,
lightweight aggregate, and aerated concrete. It was eventually
recognized that despite the fact there were many potential flyash
products with quality equal to or superior to other existing materials,
the use of flyash was limited; it was also recognized that the
situation would be even worse for sludge. Major inhibitions to the
use of sludge include highly variable chemical and physical
properties, high transportation costs, requirement for dewatering
for many applications, and inability to economically compete with
other materials.
This is in contrast with the trend in Japan, which since 1972 has
been toward the conversion of scrubber sludge to gypsum. All
of the gypsum has been used so far for wallboard production
and as a retarder of cement setting. It should be noted that
the presence of flyash in appreciable quantities reduces the
market value of the wallboard gypsum in Japan^ but, since most
Japanese boilers are oil-fired, this does not present a major
problem. This could present a marketing problem for applying
this approach to the coal-fired systems in the U.S. unless most
of the flyash is removed ahead of the scrubber.
907
-------
Although the supply of gypsum in Japan is expected to exceed
the demand in the next few years, and despite the fact that land
there is at a premium, the trend toward byproduction of gypsum
continues because gypsum is considered more acceptable for disposal?-^
This is due primarily to the fact that gypsum is much more easily
dewatered than sludge containing large quantities of calcium sulfite.
Consequently, the volume of sludge for disposal is substantially
reduced. Conversion to gypsum also eliminates any chemical oxygen
demand which the calcium sulfite might present.
Since gypsum is a relatively plentiful commodity in the United
States (although much of it is imported), conversion of scrubber
sludge to gypsum is not expected to be a major utilization activity.
However, it does present a potential alternative to other disposal
schemes.
From the above discussion, it is concluded that for the immediate
future, disposal of scrubber sludges will be the major alternative
selected in the United States. The discussion below summarizes
current utility programs for sludge disposal.
4.2 Present and Planned Utility Industry Disposal Programs
The major options for sludge disposal are ponding and landfill.
Table 5 summarizes dewatering techniques and ultimate disposal
modes for fifteen lime and limestone FGD systems at utility sites.
It can be observed that utilities are selecting both ponding and
landfill as the disposal mode. For those sites selecting the
landfill mode, dewatering techniques (such as filtration or
centrifugation) and/or sludge fixation processes have been or
will be used to attempt to produce an acceptable landfill material.
The wide variety of approaches indicated may be based on factors
such as: non-uniformity of local regulations, disposal site location
and ownership, disposal site proximity to ground or surface waters,
soil permeability, variations in sludge chemical and physical properties,
and variations in scrubber processes and types of ash collection
and disposal.
Detailed examples of some of the utility sludge disposal programs
are described below:
Commonwealth Edison (Will County) - Treated sludge material
will be stored in clay-lined basins with groundwater wells. The
material will cure for approximately one month and will be
908
-------
TABLE 5. SLUDGE TREATMENT/DISPOSAL TECHNIQUES FOR SELECTED UTILITY LIME/LIMESTONE FGD SYSTEMS
(X = Current; P = Possible Additions)
Facility
(Operating
Status)
TVA-Shawnee
(Current)
EPA Test
Facility
City of Key
West-Stock
Island
(Current)
Coniinonwca 1 th
Edison Co. -Will
County
(Current)
vc
o
0 Southern
California
Cdison-Mohave
Lime: Current
Limestone :0ct .]
Kansas City
Power & Light-
1 law thorn
(Current)
Kansas Power ft
Light-
Lawrence
(Current)
Louisville Gas
G Electric-
Paddy's Run
(Current)
Sorbent ^___^-- — '~~~~^~^
pi i p i
^^^*~~~^ r uc i
Limestone ^^^
& lime ^^*^
^^^ Eastern
^^^^ coal
Limestone ^^*
(coral marl) .^^
^^^^^
.^'^ Residual
^^ oil
Limestone ^^^
-s^^ Eastern
,^ coal
Limestone ^^- — '
6 lime ^^^^
^_^^"^
^^^^ Western
yWT coal
Limestone - boiler s'
injected (Unit 3Y/
and "tail-end"^
(Unit 4) ./
/^ Coal
/^ (possible H&W
^ blend)
Boiler ^^
injected ^^^
limestone ^^
^^^
^^^ Eastern
^^ coal
Carbide ^-^"
sludge ^^--^^^
(Ca(OHK)^-^"^
^^^ Eastern
^^ coal
Scale
Proto-
type
(Three
LO Mw
Jnits)
Full
Pull
Full
Full
Full
Ku 1 1
Clari-
f i er
•
X
X
X
X
X
Dewatering Technique
Fi 1 ter
X
P
X
X
Centri-
fuge
X
Dryer
P
Pond
X
X
(clay-
lined
well
points)
X
(well
points)
X
Final Disposition
Ponding
X
(unlined)
X
(fixed)
X
(unlined)
X
(unlined)
Landfill
x
(unfixed)
X
(fixed)
X
(unfixed)
-------
TABLE 5. (Continued) SLUDGE TREATMENT/DISPOSAL TECHNIQUES FOR SELECTED UTILITY LIME/LIMESTONE FGD SYSTEMS
(X = Current; P = Possible Additions)
Facility
(Operating
Status]
Sorbent
Fuel
Scale
Clari-
fier
Dewatering Technique
Filter
Centri-
fuge
Dryer
Pond
Final Disposition
Ponding
Landfill
Northern
States Power-
Black Dog
(Current)
Limestone
Western
coal
Proto-
type
(3-5 Mw)
(unlined)
Kansas City
Power (,
Light-
LaCysne
(Current)
Limestone
Eastern
coal
Full
(unlined)
Ari zona
Public
Service-
Choi la
(Current)
Limestone
Full
Kestern
coal
X
(solar
evap)
pnlined)
vo
}—*
o
Uuquesne
Light-
Phillips
(Current)
Lime
Full
Eastern
coal
X
curing)
(un-
lined)
X
(fixed)
Detroit
Cdison-
St. Clair
(Jan. 1975)
Limestone
Full
Eastern
coal
(unfixed)
TVA-Widows
Creek (1976)
Limestone
Full
Eastern
coal
X
(unlined)
Mansfield
(1975/1976)
Line
Full
Eastern
coal
X
(fixed)
Northern
States Power-
Sherburne
(1976/1977)
Limestone-
Flyash
Full
Western
coal
(clay lined)
-------
inspected by local authorities to obtain permission for off-site
disposal. Activities to determine the technical quality of
the fixed material and attendant costs are still underway.
Kansas City Power & Light (Hawthorn) - Fourteen wells are
located around the unlined on-site pond and are sampled periodically.
No definite data are available but it is believed that results to
date are inconclusive because the groundwater in the general area
may be heavily contaminated by leachates from on-site flyash
ponds.
Duquesne Light Company (Phillips) - After curing for about
30 days in clay-lined basins, sludge treated with Dravo's "Calcilox"
is being dredged out and hauled to a disposal demonstration site
about one mile away. The site includes two ponds lined with
Hypalon. Each pond has underdrainage and overdrainage piping
provisions to collect water for testing. These demonstration
tests are still underway.
Southern California Edison (Mohave) - Two commercial sludge
fixation processes are currently being evaluated. The Dravo
process is being demonstrated through production of a landfill
(soil-like) material; the IUCS process is being demonstrated
through production of a synthetic aggregate. Tests are being
conducted to determine the environmental acceptability of these
materials.
Through activities of the type described above, utilities, FGD
vendors, and sludge handling technology vendors can be expected
to identify environmental problems and solutions, and to optimize
costs of sludge disposal, as necessary for their specific applications.
However, it is felt that more information and data are necessary
to assure more general application of sludge disposal technology
and to further minimize environmental effects, at reasonable
cost. In an attempt to fulfill those needs, EPA has initiated
several programs, which are discussed below.
911
-------
5.0 CURRENT EPA R&D PROGRAMS
5.1 Aerospace Study
The most broadly-based EPA program relating to scrubber sludges
is the National Environmental Research Center-Research Triangle Park,
N.C. (NERC-RTP) contract with The Aerospace Corporation (El Segundo,
California) entitled "Study of Disposal of By-Products From Non-Regenerable
Flue Gas Desulfurization Systems." This study was initiated
during late 1972 and has the following major elements:
(1) An inventory of sludge constituents in both the solid and
liquid phases. Sludges produced from the following sorbent/
fuel combinations are being studied: limestone/Eastern
and Western coals, lime/Eastern coal, and double alkali/
Eastern and Western coal.
(2) An evaluation of the potential water pollution and solid
waste problems including consideration of existing or
proposed water effluent, water quality and solid waste
standards or guidelines.
(3) An evaluation of treatment/disposal techniques with
emphasis on ponding and treated and untreated landfill.
In particular, sludges treated by two or more commercially
offered processes will be evaluated in the laboratory for
mechanical properties, permeability, leachability, etc.
(4) A recommendation of the best available technology for
sludge treatment/disposal based on the elements delineated
above.
(5) Support of an EPA field study of FGD sludge disposal at
TVA's Shawnee Steam Plant, which will include test
planning, program coordination, analyses of liquid
and solid samples, and reports.
Initial results of the Aerospace contract were reported in
May 1974 J-6 A second report should be issued in mid-1975. The
final report of the effort is expected to be released in late 1976.
Sections 5.1.1 and 5.1.2 summarize recent information generated
under this contract; this information was reported in detail in
two previous papers.•",14
912
-------
5.1.1. Sludge Chemical and Physical Characterization
In the Aerospace Corporation study, sludge samples have been
analyzed from three scrubbers involving lime/Eastern coal, limestone/
Eastern coal and limestone/Western coal. (Samples from two additional
scrubbers were analyzed under a NERC-Corvallis study. Results of
these analyses have also been assessed as part of the NERC-RTP study.)
Chemical characterizations of the sludge liquors were performed to
identify the concentrations of nine trace metals of interest and
seven major soluble species, as well as the pH and total dissolved
solids (TDS) for each of the samples collected. (Input materials
such as coal, flyash and make-up water were also characterized to
identify sources of constituents found in the sludges.) By comparing
the analytical data with the EPA Proposed Public Water Supply Intake
Criteria (October 1973),-^ a preliminary assessment was made with
regard to any potential environmental problem that might be posed
if these sludge liquors entered water supplies.
In assessing the potential impact of trace metals, it was
found that in each of the sludge liquors at least one of the
following trace metals exceeded the EPA proposed criteria: arsenic,
cadmium, chromium, lead, mercury, and selenium. Except for mercury
and selenium, these trace metals exceeded the criteria by not more
than a factor of 5. Mercury and selenium exceeded the criteria
in each of the sludges analyzed by more than an order of magnitude.
Comparing the concentrations of the major soluble species with
the criteria, it was found that particular excesses exist for
chloride, sulfate and TDS for each of the samples analyzed. To
date, only one sample has been analyzed for boron, which was also
found to be in excess. It should be noted that most of the
constituents of concern originate in the coal, indicating that FGD
processes provide a multi-pollutant control capability.
The assessment indicated that in all of the power plant sludge
liquors analyzed, water quality criteria are appreciably exceeded
for mercury, selenium, boron, chloride, sulfate and TDS (See Section 3.4)
Attenuation of chemical species by soil is widely accepted in many
land disposal practices; however, attenuation of three of the species
discussed above, namely, selenium, boron and chloride, is known to
be ineffective. Whether the concentrations of the other constituents
are acceptable or not, the concentrations of these three species
strongly indicate that environmentally sound techniques such as
lined ponds or chemical fixation are needed for scrubber sludge
disposal.
913
-------
Detailed physical characterization of various samples have
also been performed under the Aerospace Corporation study to
determine properties such as: ease of dewatering/settling, bulk
density, viscosity, load bearing strength and permeability. Results
of the characterization have verified the water retentive nature of
calcium sulfite and have quantified the effect of moisture on sludge
physical properties. For example, at 65 percent solids content,
sludges will support personnel, and at solids content greater than
70 percent, sludges will support heavy equipment. However, the
cost of dewatering the sludge and maintaining the disposal site at
the necessary degree of dryness is strongly dependent on the raw
sludge properties, and may or may not compare favorably with the
cost of an alternative disposal technique such as chemical fixation.
5.1.2. Current Disposal Techniques
Ponding - Available information indicates that water pollution
problems can be prevented by proper pond engineering, installation
of an impervious pond liner, and by operating the FGD system in a
closed-loop mode. Pond liners may consist of flexible materials
such as polyethylene, polyvinyl chloride, du Pont Hypalon films,
or non-flexible materials such as asphalt concrete or clay. Costs
of ponding are primarily dependent on construction costs and the
type of liner employed, with a minor dependency on the cost of
land. Estimated costs for ponding of sludge range from about $2.50
to $4.50 per ton of wet (50% solids) sludge, exclusive of possible pond
reclamation costs.
Ponding of scrubber sludge containing considerable amounts of
calcium sulfite has the disadvantage—in areas of the country where
precipitation exceeds evaporation (this is the case in most of the
Eastern U.S.)—of excessive moisture being retained in the sludge
(see Section 3.4). Under these conditions the sludge has little
compressive strength, and presents a land reclamation problem which
can be avoided by chemical treatment (fixation). Sludge consisting
mostly of calcium sulfate (gypsum) may not present this problem,
but this has not yet been demonstrated. Even if land reclamation
difficulties are avoided, eventual deterioration of pond linings
would still be of concern. Available data indicate that chemical
treatment could avoid the need for'-a pond lining.
914
-------
Landfill - The problems in landfilling untreated scrubber
sludge are identical to those of ponding the material, except
that additional dewatering is normally necessary. As discussed
previously, this presents the most difficulty with sludges
containing considerable calcium sulfite. Commercial processes
for chemical treatment (fixation) of scrubber sludge are currently
offered by at least three companies—Chemfix (Division of.Environmental
Sciences, Inc.)> Dravo Corporation, and IU Conversion Systems,
Inc. (IUCS). Although distinctly different chemicals and
operations are employed, all three processes are designed to produce
a material with sufficient compressive strength to be suitable
for landfill and to chemically and/or physically bind up the
soluble constituents of the sludges.
Costs of sludge chemcial treatment (fixation) processes are
still developing, and depend to a great extent on the local plant
conditions, e.g., land availability. Estimates from various
sources indicate chemical fixation costs of from $2 to $9 per
(short) ton of sludge (wet, 50 percent solids). All relevant
factors pertaining to the cost estimates are not available, but
based on recent vendor contacts, a total cost for treated scrubber
sludge disposal of about $5/ton appears reasonable. (In a typical
power plant application, $5/ton would be equivalent to 1.2
mills/Kwhr.) These costs are for disposal in natural soil disposal
sites, which means sludge fixation costs are more than the costs
of disposal in a lined pond. However, the potential for future
environmental difficulties is substantially reduced.
As indicated in Section 3.4, scrubber sludge consisting mostly
of gypsum is much more easily dewatered than other scrubber sludges.
If most of the solubles could also be removed by techniques such as
cake washing, this material might be suitable for direct disposal
on land.
The FGD processes applied to utilities using low-sulfur Western
coal normally produce high gypsum sludge. Technology also exists for
conversion of high calcium sulfite sludges to gypsum sludges. If
landfilling of untreated gypsum sludge proves feasible, this would
provide a direct competitor to chemical fixation.
915
-------
5.2 Shawnee Field Study
5.2.1 Overview
A field study of sludge disposal has been initiated by NERC-RTP
at the Tennessee Valley Authority (TVA) Shawnee Steam Plant near
Paducah, Kentucky. The purpose of the study is to evaluate current
sludge disposal technology regarding the ponding of raw sludges
and simulated landfilling using chemical-treated sludges. The study,
which began in September 1974, will last approximately 18 months.
The disposal test site contains clay-like soils such that all
disposal areas will simulate clay lined ponds or basins. Five
different disposal ponds, each approximately 405 sq m (0.1 acre) in
size, will be filled to a depth of about 0.9 m (3 ft). One will
contain untreated filter cake from the venturi-spray tower (lime)
scrubber and one will contain untreated clarifier underflow from
the mobile bed (limestone) scrubber. Each of the three other
ponds will be filled by sludge fixation process contractors as
follows:
(1) Chemfix-treated limestone clarifier underflow,
averaging about 40 percent solids which contain
about 45 percent flyash (dry basis).
(2) IUCS, Inc.-treated lime filter cake averaging
about 50 percent solids which contain
45 percent flyash (dry basis).
(3) Dravo-treated limestone filter cake averaging
about 55 percent solids which contain
45 percent flyash (dry basis).
All fixation will be performed on-site during the period November -
December 1974. No additional materials, such as flyash or soil
will be used for further dewatering of the sludge. Technical and
economic data attendant to each of the pond sludge operations will
be collected to determine an estimate of fixation costs for disposal
of a similar material at a full-scale plant.
Each pond will have a leachate well and ground water wells, all
of which will be sampled periodically. Soil cores from the pond
bottoms will be taken prior to filling and periodically throughout
the program; fixed sludge cores will also be taken periodically.
916
-------
In addition, sludge samples from the scrubber system discharge
(clarifies underflow or filter cake) will be taken as well as
periodic samples of disposal pond surface liquors. Selected
analyses of all samples will .be conducted by TVA and Aerospace
to determine the following: 1) the nature of the bottom .soil
of each pond; 2) the quality of the water from all wells;
3) the seepage through the bottom of all ponds; 4) the interaction
between the sludges and the bottom soil of each pond; 5) the
quality of the chemically conditioned sludges as to strength,
permeability, and leaching effects.
All construction, maintenance, sampling and selected testing
will be performed by the TVA. The Aerospace Corporation is responsible
for developing and maintaining the test plan, coordinating the program,
performing selected tests, analyzing the test data and writing the
final report. Bechtel Corporation will coordinate the scrubber test
program with the disposal study as necessary. A status report will
be released in mid-1975 and the final report will be issued in late
1976.
5.2.2 Current Status
The first untreated sludge pond was filled in late September-
early October 1974. Designated Pond "A", this pond was filled
with filter cake from the venturi-spray tower lime scrubbing system.
Figure 2 shows the filter cake discharge; the filter cake drops
to a horizontal conveyor, then is transferred to an inclined
portable conveyor (Figure 3), which deposits the filter cake into
a cement truck. During filling of the truck, the cement mixer
is rotated, keeping the sludge in a fluid state.
Figure 4 shows Pond "A" prior to filling. Figure 5 is a
close-up of the leachate well and access platform. Figure 6
shows the cement truck in position for discharge of the sludge.
Figure 7 is a close-up of the sludge in Pond "A" when the pond
was approximately half full. In the foreground the surface of
the sludge has dried and has begun to crack.
After Pond "A" was filled, the filling of the second untreated
sludge pond was initiated. Designated Pond "D", this pond was
filled during October 1974 with clarifier underflow from the
mobile bed limestone scrubbing system. In early November, the
917
-------
Figure 2. Discharge of Sludge Filter
Cake
Figure 3. Portable Sludge Conveyor
918
-------
Figure 4. Pond "A" Prior to Filling
Figure 5, Pond "A" Leachate Well &
Access Platform
919
-------
Cement Truck in Position to
Discharge Sludge in Pond "A"
Figure 7. Pond "A11 Half-Filled with Sludge
920
-------
sludge from this pond will be treated by the Chemfix process and
placed in an adjacent pond, designated Pond "E". Pond "D" will be
refilled with untreated limestone sludge, according to current
plans, beginning in mid-December. The remaining ponds (Pond "B"
and Pond "C") are expected to be filled with treated sludge
during the period early November - early December.
5.3 Corps of Engineers Fixation Studies
5.3.1 Overview
In mid-1974, NERC-Cincinnati initiated an interagency agreement
with the U.S. Army Corps of Engineers' Waterways Experiment Station
in Vicksburg, Mississippi. Under this agreement, the leachability
and durability of raw and chemically fixed hazardous industrial
wastes (Type "A" sludges) and FGD scrubber sludges (Type "B"
sludges) will be studied. Five Type "A" sludges and up to six
Type "B" sludges will be obtained for the study. In addition,
five fixation processors will be selected to treat each of the
two categories of sludges (not necessarily the same five
processors for each sludge category).
The program will be conducted in three phases:
(a) Phase I will consist of sludge characterization and
analysis, preliminary leaching tests, and the design
of the testing program for Phase II.
(b) Phase II will include laboratory studies of leachability,
permeability, durability, and stability tests. The
leachability tests will consist of column studies
designed to simulate field conditions associated with
land disposal.
(c) Phase III will be a field study during which selected
sludges will be disposited directly on land surface
plots (simulated landfill).
Formal progress reports on the effort are expected to be
issued as follows:
(a) Laboratory Report on Leachability and Durability
Studies (Phases II and III) - mid-1975.
921
-------
(b) Interim Report on Field Study (Phase III) -mid-1976.
(c) Final Report on Field Study - mid-1977.
The leachability and durability studies are expected to
supplement and enhance the efforts being conducted under the
Aerospace study (Section 5.1). The field studies should supplement
the information generated in the Shawnee field study (Section 5.2).
The EPA project officer for this study is Mr. Robert Landreth.
5.3.2 Current Status
The scrubber sludges obtained by the Corps of Engineers so far
in this study include the following:
Eastern (high-sulfur) coal - lime
- limestone
- double alkali
Western (low-sulfur) coal - limestone
- double alkali
As of September 30, 1974, samples of most of the sludges had been
sent to the appropriate fixation processors. A total of seven
processors were selected for this study. Three of the seven will
be treating both the (Type "A") industrial sludges and the (Type "B")
scrubber sludges.
Preliminary qualitative and quantitative analyses have been
conducted of the scrubber sludges and the industrial sludges. In
addition, two sets of elutriate (batch-wise leachate) samples
have been obtained and analyzed. The first set were obtained by
contacting the sludges with deionized water, the second set by
contacting the sludges with deionized water adjusted to pH 4.0
with HC£. Since these data have not been completely analyzed,
they are not reported here.
The leachate column studies were expected to be underway in
late October 1974.
922
-------
5.4 Dugway Proving Ground Study
In January 1974, NERC-Cincinnati initiated an interagency
agreement with the U.S. Army Materiel Command's Dugway Proving
Ground, Dugway, Utah, to determine the extent to which heavy metals
and other chemical constituents from five industrial wastes could
migrate through the soil in land disposal sites. In mid-1974,
this study was expanded to include eight additional industrial
wastes and three untreated FGD scrubber sludges.
The program will be conducted in three phases:
(a) Phase I will consist of physical and chemical
characterization of the sludges and preliminary
screening tests with a variety of U.S. soils.
(b) Phase II will consist of leachate studies in
columns with the sludges applied to two selected
(best and worst) soils, long term permeability
tests with selected clays are also planned for
the FGD scrubber sludges.
(c) Phase III will consist of data interpretation of
the Phase II tests to identify soil attenuation
mechanisms and to develop empirical "attenuation
coefficients" for specific chemical substances.
The Phase I results will supplement the Aerospace and
Corps of Engineers' efforts. The Phase II results are expected
to assist considerably in interpretation of the results of the
Shawnee field study, particularly the permeability tests with
clays.
Results of the Phase I screening tests with scrubber sludges
are expected to be reported by mid-1975. Preliminary results of the
column studies are expected to be reported in late 1975; the final
results are expected to be reported in late 1976. As of October 22,
1974 Phase I with the scrubber sludges was almost ready to be
initiated. The EPA project officer for this study is Dr. Michael Roulier.
5.5 Full-Scale Sludge Demonstration Program
NERC-Cincinnati is currently considering a full-scale sludge
disposal demonstration program with a utility which uses high sulfur
923
-------
coal. This program would serve to supplement the Shawnee field
demonstration. The program, if it is undertaken, would include
laboratory tests as well as small and large scale field tests of
landfill disposal sludge. Preliminary negotiations are still under-
way 'as of October 10, 1974. The EPA project officer for this effort
will be Mr. George Huffman.
5.6 Aerospace Water Treatment/Reuse Study
NERC-Corvallis has issued a grant to Aerospace Corporation to
determine the implications of open-loop or partially open-loop
lime/limestone FGD systems. Under the study, samples of scrubber
liquors from three different plants have been obtained, including
the following system types:
Eastern (high-sulfur) coal - limestone
- lime
Western (low-sulfur) coal - limestone
The study has the following objectives:
(a) Characterize the constituents in scrubber liquor
under various process conditions.
(b) Identify constituents in the liquor that may
potentially require control or treatment.
(c) Assess potential liquor treatment alternatives
on the basis of treatment efficiency and costs.
(d) Recommend practicable approaches for control and
treatment of scrubber liquor for potential reuse
or discharge.
The assessment of water treatment systems will cover new
technology developments as well as currently operational treatment
and control systems. This will include information on the types,
sizes, investment costs and operating costs, as well as an evaluation
of the technical and economic feasibility of the wastewater treatment
and control alternatives. Recommendations will be made for the most
practicable approaches to control and treatment of scrubber liquor
for water reuse and disposal.
924
-------
Results of the study could indicate alternate approaches to the
sludge disposal problem. One very likely prospect is treatment
of filter cake wash from FGD processes producing a mostly gypsum
sludge. The EPA project officer for this study is Mr. Dennis Cannon.
5.7 Remarks on the EPA Program
As Sections 5.1 - 5.6 indicate, there are several significant
EPA research and development efforts underway at three different
research centers. The interrelationship between the various efforts
is also described. A substantial amount of close coordination is
required for these separate efforts to accomplish the overall task
which must be done. This close coordination exists through frequent
communication and a spirit of cooperation. The real significance
of these efforts, however, is the multi-disciplinary approach to
the overall task. Expertise in air pollution control (NERC-RTP),
solid waste/residuals management (NERC-Cincinnati) and water
treatment/water quality (NERC-Corvallis) have all been brought
together in this research and development program. This approach
is the only effective way to address environmental concerns.
925
-------
6.0 REFERENCES
1. Dupree, Walter G., Jr. and James A. West, U.S. Energy Through
the Year 2000, U.S. Department of the Interior, 1972.
2. Gage, S.J., "Technological Alternatives to Flue Gas Desulfur-
ization," Presented at the Flue Gas Desulfurization Symposium,
New Orleans, Louisiana, May 14-17, 1973.
3. EPA, Report of the Hearing Panel, National Public Hearings
on Power Plant Compliance with Sulfur Oxide Air Pollution
Regulations, January 1974.
4. Council on Environmental Quality, Energy and the Environment;
Electric Power, August 1973.
5. Personal Communication, Norbert Schomaker, EPA/Solid and
Hazardous Waste Research Laboratory.
6. U.S. Bureau of Mines, Information Circular 8572.
7. U.S. Bureau of Mines, Minerals Yearbook, 1970, Vol. 1.
8. Weston Environmental Scientists and Engineers, Concept
Evaluation Report, Taconite Tailings Disposal, Reserve
Mining Company, Silver Bay, Minnesota, West Chester,
Pa., 1971.
9. Personal Communication, Mr. Stowalzer, U.S. Bureau of Mines
Phosphate Commodity Office.
10. Dean, R. and J. Smith, Jr., "The Properties and Composition of
Sludges." Presented at the Seminar of Methodology for Monitoring
the Marine Environment, University of Washington, Seattle,
Washington, October 1973.
11. Battelle Memorial Institute, Inorganic Fertilizer and Phosphate
Mining Industries—Water Pollution and Control, Columbus, Ohio, 1971.
12. Personal Communication, R.D. Hill, EPA/Industrial Waste Treatment
Research Laboratory.
926
-------
13. Jones, J.W., et. al., "Disposal of By-Products from Non-
Regenerable Flue Gas Desulfurization Systems", Presented
at the American Society of Civil Engineers Annual and
National Environmental Engineering Convention, Kansas
City, Missouri, October 21-25, 1974.
14. Rossoff, J., et. al., "Disposal of By-Products from Non-
Regenerable Flue Gas Desulfurization Systems: A Status
Report", Presented at the EPA Control Systems Laboratory
Symposium on Flue Gas Desulfurization, Atlanta, Georgia,
November 4-7, 1974.
15. Ando, Jumpei, "Utilizing and Disposing of Sulfur Products from Flue
Gas Desulfurization in Japan." Presented at the EPA Control
Systems Laboratory Symposium on Flue Gas Desulfurization,
Atlanta, Georgia, November 4-7, 1974.
16. Rossoff, J. and R. C. Rossi, Aerospace Corporation, "Disposal of
By-Products from Non-Regenerable Flue Gas Desulfurization
Systems: Initial Report," EPA-650/2-74-037-a, May 1974.
17. Proposed Criteria for Water Quality, U.S. Environmental
Protection Agency, Washington, D.C., 20460, October 1973.
927
-------
FGD SLUDGE FIXATION AND DISPOSAL
by
William H. .Lord, P.E.
Projects Director
Special Projects - Waste Disposal Group
Eastern Construction Division
Dravo Corporation
Pittsburgh, Pennsylvania
The fixation and disposal of by-products produced by Flue
Gas Desulfurization (FGD) processes has increased the operational
and environmental problems confronting the electric utility
industry. As a result of laboratory research and development efforts
apd pilot plant testing, significant data has been compiled
concerning the physical and chemical properties of the sludge.
Work to date indicates that stabilization by chemical additives is
required and feasible. Technical analysis of sludges from FGD
systems stabilized with CALCILOX additives indicate these sludges
can be used in a land fill type disposal operation.
Market studies and technical analysis indicate that land
fill disposal is the only currently feasible and acceptable disposal
method. Land fills constructed with stabilized FGD sludges can meet
all engineering, environmental and ecological requirements. In
addition, land fill disposal is the only disposal method which
can handle the large volumes expected from FGD systems. Full scale
disposal operations conducted by Dravo Corporation are described.
929
-------
TABLE OF CONTENTS
Page
List of Tables and Figures
Introduction and Background 933
On Going Disposal Research 93^
Stabilization and Soil Mechanics 937
Laboratory Testing of Stabilized Sludges 940
General Index Properties 940
Strength Tests 940
Consolidation Characteristics Q, _
Disposal System Planning 947
Disposal Site Selection o/7
Current Disposal Studies 950
Current Projects
Cost 952
Summary
931
-------
LIST OF TABLES AND FIGURES
Table Number
1
Title Page
Chemical Constituents 935
Desulfurization System Sludge
Grain Size Distribution 938
Desulfurization System Sludge
Chemical Constituents of Stabilized 946
Desulfurization System Sludge
Leachate
Figure Number
1
2
3
4
5
Title Page
Solid Waste Products from a 1200 MW 934
Station
Settling Test Curve 939
Effect of Solid Content and 941
Percent Additive on Strength
Grain Size Curve 942
Effect of Various Additives on 944
Shearing Strength
932
-------
FGD SLUDGE FIXATION AND DISPOSAL
INTRODUCTION AND BACKGROUND
After the enactment of the 1970 Amendments to the Clean Air
Act of 1967, Dravo Corporation's Research and Development Department
began extensive "state of the art" studies of Flue Gas Desulfurization
(FGD) systems. One important conclusion of these studies was that
although much effort had been directed toward the mechanics of S02
removal from flue gases, very little effort had been directed toward
the disposal of the FGD systems by-products. Aside from a general
knowledge of the chemical constituents of these wastes, no practical
data were available concerning the physical and chemical behaviors of
these materials and the magnitude of the disposal problems they would
create at fossil-fuel-burning generating stations. Realizing the need
for additional data to evaluate the alternatives for waste disposal,
Dravo Corporation began an extensive study of FGD systems wastes. The
major parameters which required identification included:
1) Quantities of wastes to be produced.
2) Chemical constituents of the FGD wastes.
3) Physical properties of the FGD wastes.
This study lead to the realization that very large quantities of
FGD systems wastes would be generated. Figure 1 shows the estimated
waste quantities generated by a 1200 MW power station burning high
sulfur coal. Although useable products such as light-weight aggregate
or road sub-base material could be made from the wastes, marketability
studies, existing technology and economic .considerations indicated
that wastes of this quantity must be placed in land fills.
Evaluation of the chemical constituents of the wastes
indicated that the waste solids were generally inert and non-toxic
in a controlled environment. Problems could result, however, from the
chemical qualitv of the large quantities of water required in the wet
scrubbing method\)f desulfurization. Reutilization of all the super-
natant and leachate from FGD systems wastes was not feasible as the
water chemistry after S02 removal did not meet the standards required
for FGD system makeup water. The water which could not be used had to
be treated for reuse, discarded or included in a disposal scheme.
A chemical analysis of an S02 scrubber sludge from a lime scrubber
is shown in Table 1.
Study of the physical properties of FGD systems wastes showed
that, due to the particle size, shape and gradation, these materials
were thixotropic, i.e., subject to liquefaction.
933
-------
CC
O
X
a:
UJ
0_
t-
x
ts
-1 30 -
O
CO
40 -
cc
a
SOLID WASTE PRODUCTS (IN DRY TONS PER HOUR)
FROM A 1200 MW STATION AS A FUNCTION OF
COAL FIRING RATE •
FLY ASH
SOLIDS
SO SCRUBBER
SLUDGE SOLIDS
BOTTOM ASH
SOLIDS
20 •
10 •
50
100
150
DATE 10-3-74
DR ALP CH
REF.
2767-2-1
200 250 300
COAL FIRING RATE (TONS PER HOUR)
FIGURE I
Based on I 6 . 8% Ash and 3.5% Sulfur
content of cool with Z0%/80% split for
bottom ash/fly osh. 100% ash removal
and 85% S02 recovered.
-------
TABLE 1
Chemical Constituents
Desulfurization System Sludge
Fly Ash 17-20%
CaS03.l/2 H20 65%
CaS04-2H20 5-7%
CaC03 3%
Others 3-5%
935
-------
Results of the preliminary study indicated that utilization
of the FGD systems wastes in stabilized land fill was the only
acceptable and economical method of handling the waste problem.
Further study of the materials generated in FGD systems determined
that, without the aid of a stabilizing agent, the material could not be
satisfactorily placed in a land fill. As a result, the series of
stabilizing additives known as CALCILOX was developed by Dravo's Research
and Development Department. These additives sufficiently alter the
properties of the wastes to permit placement in a safe land fill.
ON GOING DISPOSAL RESEARCH
The handling and disposal of large quantities of waste
materials generated by FGD systems are directly affected by the in-
herent physical and chemical properties of the wastes. These proper-
ties are a function of the: 1) characteristics of the coal burned;
2) type of boiler; 3) type of scrubber; 4) scrubbing medium; and
5) plant operating procedure. As these variables are unique to each
power station, study of the wastes from each station is required to
evaluate their properties. Each study should include a pilot FGD
system of the type selected for the plant and should be operated under
conditions which simulate anticipated conditions during power gen-
eration.
Laboratory analyses of the sludges generated by the pilot
plant in each study will yield data for evaluation of various disposal
methods. A typical laboratory evaluation might include:
1. Bulk density of sludge settled from water slurry
2. Specific gravity of dry solids
3. Particle size, shape and gradation
a. Wet sieve analysis
b. Sub-sieve analysis
c. Scanning electron microscope evaluation
d. Elaine fineness number
4. Chemical analyses
a. Sludge solids
b. Supernatant liquid
c. Leachate
5. X-Ray diffraction studies
6. Settling tests
936
-------
7. Filter tests
8. Rheological and pumping tests
Although the sludge generated in each study will be
different, test results from several sludges indicate most will have
properties comparable to those listed below:
1. Bulk density of settled sludge -
85 to 95 pounds per cubic foot
2. Specific gravity of dry solids - 2.48 to 2.55
3. Particle size - see Table 2
4. Chemical analyses - see Table 1
5. Settling tests - see Figure 2
To date, Dravo has gathered rheological data on numerous
samples of FGD sludges from four different power stations in Penn-
sylvania, Ohio and West Virginia and has conducted three large-scale
(6 and 8 inch pipe diameter) pump loop test programs on FGD sludges.
As a result of extensive rheological testing in the
laboratory and in pump loop tests, it has been concluded that FGD
systems waste products are suitable for pipeline transport. Although
each FGD system sludge has different flow characteristics, they are
all non-corrosive, exhibit relatively low abrasion and can be trans-
ported over a wide range of tonnages within any given pipe size.
More detailed discussions of the properties of FGD system
sludges have been presented by Selmeczi and Knight(1) and by
Selmeczi and Elnaggar(2).
STABILIZATION AND SOIL MECHANICS
Once ranges of probable physical and chemical properties of
sludges to be produced at a plant have been determined, evaluation
of the stabilization characteristics of the sludges is required.
Sludges generated in the pilot scrubber should be treated with
stabilizing agents such as the CALCILOX series of additives. Testing
of the various admixtures of sludges and stabilizing agents will
indicate the most economically feasible stabilization procedures
for each power station.
Samples should be prepared for continuous monitoring to
establish the degree and rate of stabilization. A measure of the
degree and rate of stabilization is the resistance to penetration of
937
-------
TABLE 2
Grain Size Distribution
Desulfurization System Sludge
Wet Screen Analysis (Dravo)
+ 50 mesh
+100 mesh 1.1%
+200 mesh 4.9%
+325 mesh 9.3%
+400 mesh 14.3%
-400 mesh 85.7%
Sub-Sieve Analysis (Micromeritics Instrument Corp.)
Equivalent Spherical Diameter Cumulative Wt.
+18 micron 10%
-18 micron 90%
-12 micron 80%
- 9 micron 70%
- 8 micron 60%
- 6.4 micron 50%
- 5 micron 40%
- 4 micron 30%
- 3 micron 20%
- 1.5 micron 10%
Elaine Fineness Number 7600
938
-------
14
DRUM 30 - 35 CMS/LITER
SETTLING TEST
FIGURE 2
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 ISO 190 200
MINUTES
-------
the treated sludges by a penetrometer. Continuous testing will yield
strength-time relationships for samples treated with the various
additives. The stabilization test data fot sludges treated with
CALCILOX H35 are shown on Figure 3. These curves show that strength
gain was rapid in some of the samples. Some of the stabilized
sludge samples reached the maximum limit of the penetrometer
(4.5 TSF) in a few days while others required about four weeks. Based
on results similar to those shown in Figure 3, judgment must be
exercised in selecting the most promising samples for an extensive
testing program. It must be recognized, however, that variables
other than sludge properties will affect the degree and rate of stab-
ilization. In addition to the type of additives, consideration must
be given to: 1) combinations of additives; 2) quantity of additives;
3) solids content of the sludge leaving the thickener; 4) curing
temperature; and 5) slurry pH.
LABORATORY TESTING OF STABILIZED SLUDGES
As the stabilized sludges are new materials and little,
if anything, is known about their behavior in a land fill, Dravo
Corporation has conducted extensive soil mechanics laboratory tests.
The purpose of these testing programs was to develop soil mechanics
parameters for correlating the properties of the stabilized sludges
with those of known soil types for use in the design of land fills.
General Index Properties
All of the cured, stabilized sludges tested were composed of
silt-sized particles; most fall into the ML classification of the USCS
classification system. Figure 4 shows a typical particle size dis-
tribution of an unstabilized sludge containing 55-60% fly ash. The
average particle size of a sludge containing less fly ash would be
slightly smaller as calcium sulfite and sulfate crystals generally do
not exceed 10 microns in size. Some of the samples exhibited
plasticity and were classified as MH.
Natural water contents ranged from 106% (48% solids) to 140%
(42% solids). Dry densities of the undisturbed stabilized sludges
ranged from a low of 33.1 pounds per cubic foot to a high of 43.7
pounds per cubic foot. Permeability ranged from 1 X 10~^ centimeters
per second for remolded material to 1 X 10~° centimeters per second
for the undisturbed material.
Strength Tests
Undisturbed and remolded stabilized sludges were subjected to
direct shear and triaxial shear testing to develop strength parameters
for embankment design and to evaluate bearing capacity for a land fill
to be used for building sites. A sample of unstabilized sludge was
940
-------
I-'
u_
*
o
\
CO
o
\-
o
LJ
cr
o:
UJ
h-
Ui
S
O
(T
H
LU
2 -
I -
LIMIT OF PENETROMETER
ADDITIVE
38% SOLID
50% SOLID
25
30
5 /O 15 20
TIME, DAYS
FIGURE 3. EFFECT OF SOLID CONTENT AND PERCENT ADDITIVE
ON THE STRENGTH
-------
SIEVE ANALYSIS
HYDROMETER ANALYSIS
•JD
JN
ho
CLEAR SO. .. . _Tn op-p.pQ
OPENINGS U.S. STD. SERIES
3" \]/9 3/J' 3/ft" *4 *IO *20 *40 *80 *200 1 4 15 60 180 1440
IAA ' 1 1 | 1 1 111
IOO
go
Qf\
O\J
H-
I
UJ
ft A
>- OU
CD
cr
z
u.
z
Ul
o
cc
uj 30 -
a.
2O
i f\
l U
i
i
i
i
i
1
i
1
i
1
•4
1
i
l
i
l
l
i
l
I
X
-v
\,
^
100 10 1.0 0. 1
PARTICLE DIAMETER IN
COBBLES
GRAVEL
CRSE. | FINE
SAND
I CRSE.j MED. | FINE
1
S
s
\
V
>
\
^
I
1
0.01
MM
\
\
S
S
\
^
0.001
SILT AND CLAY
ICLAY
GRAIN SIZE CURVES
STARfi I7FD
-------
subjected to a direct shear test to evaluate the relative effects of
stabilization on the strength of the material.
The angle of internal friction for both the undisturbed
and remolded treated materials varied from 37° to 51° with a typical
value for a thirty-day cured sludge being on the order of 39°. Most
of the samples tested showed very little cohesion. This is to be
expected as the low plasticity indicates that the stabilized sludges
should behave as granular soils. The untreated sludge exhibited an
angle of internal friction between 27° to 30°.
The strength of the stabilized sludges is comparable to
that of a dense sand and gravel under static loading conditions while
the strength of the unstabilized sludges is similar to that of
medium dense sand. It is interesting to note that the stabilized
sludges show no significant reduction in strength upon remolding and
only a slight reduction upon continued shearing. For example, the
peak strength for a thirty-day cured sample (Figure 5) exhibits an
angle of internal friction of 40° which was reduced, upon continued
shearing, to a residual angle of internal friction of 35.5°. Con-
sidering an embankment slope of two horizontal to one vertical under
drained conditions, use of the residual strength would provide a
factor of safety of 1.4 against slope failure.
Samples of stabilized sludges consolidated to dry densities
greater than 45 pound per cubic foot showed significant increases in
strength with increased confining pressures. This may be due to the
particles adjacent the failure plane slipping over one another during
shear at low confining pressures, while at higher confining pressures
the failure plane is forced through individual particles, actually
breaking the crystal structure.
Based on these tests, several conclusions can be drawn:
1. The strength of the remolded material is dependent
upon the unit weight. For the strength data presented
here to be valid, the stabilized sludges must be
placed in the disposal area within the range of
unit weights that were tested in the laboratory
(35 to 50 pounds per cubic foot dry density).
2. Increased solids content of the undisturbed sludges
results in increased strength. It will be desirable
to stabilize the sludges at the maximum solids content
that is practical under operating conditions.
3. Increased curing time results in higher strengths
and reduction of natural water content.
943
-------
if
* UNTREATED, CONSOLIDATED UNDER A VERY
HIGH STRESS BEFORE SHEARING
40 50
NORMAL STRESS; P.S.I
EFFECT OF VARIOUS ADDITIVES ON SHEARING STRENGTH
FIGURE 5
-------
4. Increased confining pressures result in a significant
increase in strength with stabilized sludges placed
at dry densities greater than 45 pounds per cubic foot.
It must be kept in mind that all of these strength data
are based oh static loading. However, based on the information
available to date, there is no significant reduction of strength for
samples subjected to vibrating loads in the range of 2 to 5 cycles
per second.
Consolidation Characteristics
One-dimensional consolidation tests were run on both undis-
turbed and remolded samples of stabilized sludges by Dr. Hameed A.
Elnaggar of the University of Pittsburgh to determine the consolidation
characteristics of the stabilized sludges and to develop the necessary
parameters for evaluating potential settlement. The compression index
Cc varied from 1.20 to 1.35 for the undisturbed material and from
0.69 to 1.15 for the remolded material. The magnitude of potential
virgin settlement is directly proportional to the value of the com-
pression index.
These values to the compression index are relatively high;
however, due to the chemical hardening of the sludges, the undisturbed
material appears to have a preconsolidation stress of 8 to 10 tons
per square foot, the equivalent of the weight of 120 to 160 feet of
compacted fill soil. This indicates that relatively high foundation
loadings, such as those resulting from multi-story structures, could be
placed on the cured sludge with negligible settlement. As the time
rate of consolidation is very high, any expected settlement would
occur during construction.
If the sludges, after hardening, were to be excavated and
then placed as compacted fill, the expected settlement would depend
upon the dry densities, and moisture contents of the compacted material.
Laboratory tests on remolded samples have shown that the magnitude
of settlement of structures supported on compacted sludges would be
rather high. It is not known if chemical hardening continues after
stabilized sludges have been remolded, but if additional hardening
does occur, the compressibility of the material may be reduced.
More detailed discussions of the properties of stabilized FGD system
sludges have been presented by Selmeczi and Elnaggar.(2)
Leachate tests and analyses have also been conducted on
stabilized sludges. Below is a typical leachate analysis from a sludge
cured for four weeks. Five-hundred (500) grams of the material was
slurried with 2000 ml of distilled water and agitated for 48 hours
using a Phipps-Bird stirrer. This slurry was allowed to settle for
a period of 48 hours and the supernatant decanted for analysis. The
results are shown in Table 3.
945
-------
TABLE 3
Chemical Constituents of Stabilized
Desulfurization System Sludge
Leachate
pH
Dissolved salts
Dissolved Si02
Hardness, CaC03
Fe-H-
Total iron
Ca-H-
Mg-H-
Mn++
Na+
A1-H-+
Alkalinity as CaC03
S03
PO/.
11.6
590 mg/1
Not detectable
430 mg/1
Not detectable
Not detectable
172 mg/1
.05 mg/1
.03 mg/1
4 mg/1
4 mg/1
140 mg/1
66 mg/1
100 mg/1
92 mg/1
Not detectable
It should be noted in this leachate analysis that the pH is high but
all other criteria are within the range of regulatory requirements.
946
-------
DISPOSAL SYSTEM PLANNING
In the past, long-term (30-40 years) disposal of power plant
wastes had not been a part of the planning associated with designing
and constructing a fossil fuel power generating station as most
utilities considered the disposal of fly ash to be an operating pro-
blem. This resulted in the development of numerous short-term
disposal solutions which required continuous planning. With the
advent of new regulations and public concern on environmental matters,
utilities must now develop long-range plans to control their waste
by-products, both gaseous and solid.
The planning for FGD waste disposal systems begins at the
outfall of the FGD system arid progresses through several steps
designed to change the material into either a useful product or one
which can be disposed of satisfactorily in a land recovery operation.
In view of the current state of technology of air quality control
systems and the large quantities of material involved, land fill
disposal is the best solution. Test data indicate that only stab-
ilized sludges possess adequate strength for a structurally sound
land fill.
Rising costs for land, materials, and labor make short-term
disposal systems economically unattractive. The costs associated
with obtaining local and state agency approval are high and the process
is generally time-consuming. It will be less expensive in terms of
delay time and public antagonism, as well as money, to have one large,
well designed and well maintained disposal operation, especially if
some beneficial use can be made of the land following completion of
the disposal operation. Planning at the outset for a long-term
disposal system is the key to long-range disposal economics.
DISPOSAL SITE SELECTION.
The largest and most complex task associated with the
disposal of 862 scrubber wastes in a land recovery operation is the
investigation, evaluation and design of the disposal site. To carry
a disposal site project from conception to completion requires an
interdisciplinary effort utilizing the talents of professionals
from design engineers to environmentalists. Studies must be made to
evaluate the physical, environmental, sociological, economical and
political ramifications of the proposed disposal site.
The disposal area must be designed and constructed to meet
all environmental and regulatory requirements. Existing utilities
must be relocated or removed as required. The area must be cleared
of vegetation. Existing surface drainage must be collected and
carried out of the disposal area to prevent disequilibrium of the
947
-------
system water balance or reduction of sludge storage capacity.
Numerous other important details must be considered, many of which
will vary from site to site.
The relocation of existing utilities such as roads, power
lines, pipeline, etc. within the disposal area should be done only as
required by the progress of construction and disposal. In this
manner negative sociological and economical impacts can be minimized.
Clearing of vegetation only as required will help to prevent ex-
cessive erosion within the watershed and to maintain the natural
aesthetic characteristics which occur within the disposal area.
As most disposal sites will lie in valleys, an embankment
will be required to impound the stabilized FGD system sludges. The
embankment will retain the sludge during settling, consolidation,
curing and stabilization, and will provide a stable and permanent
downstream face for the land fill. It should be designed for a full
hydrostatic load of unstabilized sludge. Emergency overflow pro-
tection through gated or non-gated spillways, decant towers, etc.,
which meet the requirements of regulatory agencies and good
engineering practice during construction and operation of the disposal
area should be provided. Normal drainage protection for earth and
rock-fill embankments should be provided to control seepage or drain-
age in the embankment. Standard monitoring equipment designed to
evaluate the behavior of the embankment should be installed.
The embankment should be constructed using native materials
found in the immediate area. This is the most economical approach,
and if the borrow area is located within the impoundment, construction
of the embankment will result in increased storage capacity.
Embankment construction can proceed in several stages as required by
the disposal operation. The final embankment design, however, should
be predetermined on the basis of existing foundation conditions
and locally available materials.
The natural drainage existing in the disposal area must be
controlled by collection or by-passing it through the embankment.
This includes both permanent and ephemeral springs and creeks which
occur throughout the disposal area. The main conduit through the
embankment should be designed to carry runoff from normal rainfall.
The runoff from severe storms beyond the capacity of the by-pass
system and the precipitation which falls directly on the reservoir
surface can be collected in the sludge impoundment area and circulated
back to the power plant for make-up water.
A comprehensive evaluation of all factors involved in the
disposal area is required to construct and operate a structurally
sound, economically feasible and environmentally acceptable disposal
site. Items which must be considered during the initial phases of
developing a disposal site include:
948
-------
A. Logistics
1. Proximity to power station
2. Availability of property (acquisition of property
options)
B. Physical Site Considerations
1. Reservoir evaluation
a. Geologic evaluation of site suitability
(stratigraphy, structural geology,
seismicity, etc.)
b. Hydrologic evaluation of site suitability
(ground and surface water inventory, seepage
characteristics of soils and bedrock, water
quality, etc.)
2. Environmental evaluations (environmental impact
of disposal on surrounding flora and fauna, land
use, etc.)
3. Archaeological and historical investigations
(destruction of some archaeological site)
4. Etc.
C. Political Considerations
I. Local government regulations
2. State government regulations
3. Federal government regulations
A. Economic impact of disposal site
5. Social impact of disposal site
Once the site has been tentatively selected, detailed
geologic investigations should be conducted to provide feasibility
and design data. This work should include:
A. Detailed Site Investigation
1. Detailed field reconnaissance to locate borrow
areas, seepage areas, abandoned mines, quarries,
wells, etc.
949
-------
2. Drilling program to evaluate foundation conditions
in the embankment area, pressure testing for
permeability determinations of in-place bedrock,
rock mechanics testing for embankment stability,
evaluation of rock borrow materials, etc.
B. Preliminary Land Fill Design
1. Design of stable embankment with the materials
available.
2. Water balance for disposal area.
3. Relocation and/or removal of utilities.
4. Federal, State and local approval.
a. Completion of all permits
b. Submittal and review
c. Incorporation of regulatory requirements into
disposal system design
C. Establishment of Estimated Construction Costs
CURRENT DISPOSAL STUDIES
Dravo is presently studying alternate methods for disposal of
FGD wastes. Current research is directed toward more sophisticated
dewatering devices and methods in an effort to reduce the quantities
and combinations of CALCILOX additives required for stabilization.
The results of these studies will include not only better disposal
economics, but also better utilization of existing and planned
disposal areas.
Studies are also under way to determine the technical
parameters involved in the utilization of fossil fuel power plant
wastes from existing stations for the construction and operation of
the future FGD waste disposal systems.
Research is being conducted to develop more reactive
CALCILOX additives in an effort to reduce the quantities of additives
required and to improve the characteristics of the stabilized sludge.
This research is also aimed at improving the chemical quality of the
water in the disposal system.
950
-------
CURRENT PROJECTS
Dravo Corporation has five FGD disposal projects under
contract. Dravo is providing engineering services from conceptual
design to final design and construction of an FGD disposal system
for a 1760 MW power station located in Pennsylvania. This system
is designed to handle 21,000 wet tons per day (FGD system sludge and
fly ash) from the power station's FGD system. The disposal system
will mix the wastes with the stabilizing additive, transport the
slurry approximately seven miles by pipeline, and deposit it behind
a 400-foot high embankment. The slurry and additive will cure
behind the embankment and form a stabilized land fill mass. At the
completion of this long-term disposal project (30 years) the disposal
area can be used for a light industrial development, a recreational
area or a man-made lake. The ultimate utilization of the land can
be left to the local municipalities as the various options remain
open to completion of the land fill operation.
Dravo recently completed a stabilization demonstration
program at another Pennsylvania power station. The FGD system at
this station handled an output of approximately 125 MW. For the same
utility, Dravo is conducting a study to develop a disposal system
with the capacity to handle the wastes from FGD systems for a total
output of 900 MW, AOO MW from one station and 500 MW from another.
The expected output from these two stations is 7200 wet tons per day.
This study will develop a conceptual design, locate potential
disposal sites and develop basic economics of the disposal system.
Dravo is also assisting a California utility with on-going
tests of a stabilization demonstration project. The disposal system
was designed by Dravo and testing began early in 1974. The system
handles the FGD system waste from a proto-type S0£ scrubber with a
stack gas flow through equivalent to a 150 MW generating unit
burning low sulfur coal. This project is demonstrating the importance
of make-up water and sludge chemistry in the FGD disposal system.
Dravo is currently under contract to provide detail design
of a disposal system for a 1236 MW power station located in West
Virginia. The system will provide disposal for the wastes produced
by a total plant FGD system. The disposal system will be designed
to handle 5500 wet tons per day of waste solids in a slurry form
which will be mixed with the stabilizing additive CALCILOX and
transported via pipeline to the land fill disposal area. This system
will also utilize 1900 dry tons per day of other power station waste
materials in the construction of the disposal area.
Another Dravo project currently underway is the detail
design of an FGD waste disposal system for an Ohio utility. The
951
-------
power station capacity is 820 MW and the expected maximum waste
quantity is 7700 wet tons per day. The wastes will be received in
slurry form, mixed with the stabilizing additive, and transported by
pipeline to the land fill disposal area. A unique feature of this
project is that the disposal sites under consideration are in strip
and deep mined areas.
COST
The cost for FGD waste disposal systems vary considerably
and depend on such factors as waste quantities, design life of the
disposal area, site logistics, power plant size, location, population
density and area development. It is clear, however, that the best
economics for disposal systems are developed by early planning of
long-term disposal systems where capital investments can be spread
over large disposal tonnages.
Present estimates indicate that long-term disposal system
capital investments will be in the range of 30 million dollars but
can be as high as 60 million. When these long-term investments are
reduced to unit costs, economic studies indicate that S02 sludge
disposal systems will cost in the range of 2 to 4 dollars per ton of
coal burned. These are estimates of the total cost for a long-term
disposal system designed to handle fly ash and bottom ash as well as
FGD system sludges. They include costs for conceptual development,
site investigations and evaluations, land acquisition, capital
costs of the system hardware and operating costs. No attempt has
been made to reduce these estimates by the ultimate salvage value
of the reclaimed land upon completion of the disposal operation.
952
-------
SUMMARY
The current public awareness of the environment and new
federal regulations dealing with air quality have forced the power
industry to increase their efforts to control objectionable emissions.
Although their efforts have been successful in reducing air pollution,
the power industry is now faced with a waste disposal dilemma. With
the introduction of flue gas SC^ removal, the quantities of waste
have increased several orders of magnitude, but the technology for
handling these wastes is almost non-existant. In spite of this,
Federal, State, and local regulatory agencies demand an immediate
and satisfactory solution.
Until recently, very little was known about the physical
and chemical characteristics of the SO. wastes. Realizing this
deficiency, Dravo Corporation began a series of studies to identify
the magnitude of the waste quantities as well as the physio-chemical
parameters which would best describe them. Although it is technically
feasible to turn these wastes into useful manufactured products, the
enormous quantities of these wastes make disposal rather than utiliza-
tion the only economically feasible solution.
Dravo Corporation has conducted an intensive research
effort designed to identify the physical and chemical characteristics
of these new wastes with regard to placing them in a land fill.
In order to make this type of operation meet the stringent environ-
mental regulations which now exist and those which will no doubt be
enacted in the future, Dravo developed a series of additives called
CALCILOX which turn the thixotropic, calcium sulfite-sulfate slurries
into reasonably high strength, low permeability materials. The
feasibility of this solution has been tested on a proto-type install-
ation and will soon be used in a large, 21,000 wet tons per day,
disposal operation. Studies for other disposal operations of a
similar magnitude are in various phases of completion.
All currently available data indicate that costs for a
long-term disposal system, from conception to final design and
construction, will be on the order of 2 to 4 dollars per ton of coal
burned. Dravo Corporation is continuing its research in an effort
to find more economically attractive solutions to the problems of
FGD wastes disposal.
953
-------
References
1. Selmeczi, J. G., and Knight, R. G.
Properties of Power Plant Waste Sludges
Paper No. B-7; Third International Ash Utilization Symposium
March 13-14, 1973; Hilton Hotel; Pittsburgh, Pa.
2. Selmeczi, J. G., and Elnaggar, H. A.
Properties and Stabilization of S02 Scrubbing Sludges
Coal and the Environment Meeting, National Coal Association
October 22-24, 1974; Louisville, Ky.
954
-------
UTILIZING AND DISPOSING OP SULFUR PRODUCTS FROM
FLUE GAS DESULFURIZATION PROCESSES IS JAPAH
Jumpei Ando
Faculty of Science and Engineering
Chuo University
Kasuga, Bunkyo-ku, Tokyo
Flue gas desulfurization in Japan has so far been oriented toward processes
which by-produce salable products* The by-product sodium sulfite has
reached 350,000 tons/year and has already filled the demand. The capacity
of the by-production of sulfuric acid will reach 1,000 tons/day in 1975*
Gypsum is considered to be the most rational by-product because of the
increasing demand for wallboard and cement; the production capacity
will reach 6,000 tons/day in 1975. Sulfur has been by-produced in relatively
small plants. The production of ammonium sulfate is under consideration.
However, since so many FGD plants are to be built, it is likely that in the
future supply of the by-products will far exceed the demand. New uses
of gypsum has been studied extensively.
955
-------
UTILIZING ABD DISPOSING OP SULPDR PRODUCTS FROM
FLUE GAS DESULFORIZATION PROCESSES IS JAPAN
I Introduction
Recently in Japan about 6 million tons yearly of S02 have been eaitted
mainly by the burning of heavy fuel oil. Desulfurization efforts have
been made in earnest since 1966. Among various desulfurization processes,
those which first became popular were hydrodesulfurization of heavy oil
by-producing elemental sulfur and sodium scrubbing of waste gases by-
producing sodium sulfite (Figure l). A wet lime process plant by-
producing salable gypsum has been in operation since 1964 but it was not
until 1972 that construction was started on many plants by-producing
gypsum. Processes that give by-product sulfuric acid, elemental sulfur
and calcium sulfite have been developed since 1971* The discarding of
calcium sulfite sludge is not as widespread in Japan as it is in the
United States, because of the limitations in land space available for
disposal.
However, since many desulfurization plants are to be built, it is likely
that in the future the supply of the by-products will far exceed the
demand, necessitating the discarding of a substantial portion of theft*
Gypsum is generally considered to be the most reasonable by-product because
of the increasing demand for it and also because of the ease with which
it can be discarded. The various by-products from waste gas desulfurization
processes are discussed below.
II Sodium sulfite
The sodium scrubbing processes by-producing sodium salts were reported on
by the author at the EPA FGD symposium last year1). The reasons for the
rapid development of the processes are the simpleness of the processes and
the usefulness of the by-product sodium sulfite for paper mills. More
than one hundred plants have been installed mainly for relatively small gas
sources, industrial boilers, chemical plants, etc. Yearly production of
sodium sulfite has reached 350,000 tons and has already filled the demand
resulting in a decrease of the selling price (Figure 2). Nevertheless, the
S02 recovered as sulfite is only 4$ of the total emission. Asahi Glass Co.
recovers 802 from a glass furnace to produce sodium sulfite, which is
oxidized into sulfate and returned to the furnace. There is not much demand
for sodium sulfate, either. Several smaller plants produce waste sodium
sulfite or sulfate solution. Rot many additional sodium scrubbing plants
are expected to be built in the future.
956
-------
10,000
(X,
1976
Figure 1 Production capacity of waste gas desulfurization
0
1968
1970
1972
1974
Figure 2 Price of by-products
957
-------
Ill SuLfurio acid
The supply and use of sulfuric acid in Japan are listed in Table 1.
Pyrite, which was the major source of the acid, has been gradually replaced
by smelter gas and sulfur. The production of the acid by flue gas
desulfurization was begun in 1971* The production capacity will reach
1,000 tons/day in 1975. But continuing rapid development cannot be
expected because the increase in the demand is only 200,000-300,000 tons
yearly.
Table 1 Supply and use of suxfuric acid in Japan (1,000 tons)
1971 1972 1973
Supply
Smelter gas 3,470 4,203 4,789
Pyrite 3,620 2,600 1,881
Sulfur 146 442 950
Others 72 140 164
Total 7,308 7,585 7,784
Use
Fertilizer 2,418 2,579 2,560
Others 4,921 4,984 5,623
Total 7,539 7,565 7,823
The Wellman-Lord process has been the major process for the production
of sulfuric acid. At present, there are five plants in operation that use
the process and seven plants under construction* The sulfuric acid
production capacity of those plants will reach 840 tons/day by the end
of 1975* The main problem concerning the process is vastewater treatment.
At the Nishinagoya plant of Chubu Electric (220MV), a portion of absorbing
liquor is cooled to 0°C to crystallize out the sodium sulfate, which is
separated by a centrifuge. The sulfate, which contains some sulfite, is
dissolved in a purge stream from the prescrubber, treated with sulfuric
acid to eliminate any sulfite, aerated, oxidized with.ozone, settled to
precipitate solids, neutralized with.sodium hydroxide, and then discharged
to sea. The chemical oxygen demand of the discarded water is kept below
lOppm, which is required by regulations. A portion of the separated eodium
sulfate is sold to the chemical industry.
958
-------
An S02 recovery plant by the Sumitomo activated carbon process has been
in operation at the Sakai Po^er Station of Kansai Electric, treating
90e(XX)scfa of flue gas from an oil-fired utilityboiler. One advantage in
producing sulfuric acid from the recovered S02 by the Wellman-Lord and
Sumitomo processes is the small size of vessels used for the acid production
due to the high S02 concentration of the gas to be fed into the acid plant—
nearly 100$ for ¥ellman and 10-20$ for Sumitomo as compared with 6-10$ for
the gas by pyrite or sulfur burning. There is no further plan, however, to
install a larger carbon process plant.
Two magnesium scrubbing plants respectively using the Mitsui Mining process
and Onahama-Tsxikish1ma process have been in operation. Both are installed
in copper smelters with sulfuric acid plants. The former treats 44,000scfm
of tail gas from a sulfuric acid plant and the latter treats 49,000scfm
of converter gas containing 2$> S02. The recovered S02 is sent to sulfuric
acid plants.
IV Elemental sulfur
Recently in Japan the supply of elemental sulfur has been depending
mainly on the by-product from the hydrodesulfurization of heavy oil
(Table 2).
Table 2 Supply of and demand for sulfur (in 1,000 tons)
1972 1973 1974 (estimate)
Supply Mined 12 0 0
Recovered 622 76! 880
Imported 1J 65 66
Total 647 826 946
Demand Use 606 733 836
Export 53 62 63
There are currently four processes being used to by-produce sulfur
by waste-gas desulfurization: (l) the Wellman-Lord process with a Glaus
furnace, (2) the Shell process with a Glaus furnace, (3) ammonia scrubbing
with an UP reactor, and (4) magnesia scrubbing with a Glaus furnace. A
plant with the process (l) has been in operation since 1971 at the Kawasaki
plant of Toa Fenryo, treating tail gas (39,000scfm) from a Clause furnace
containing 6,000ppm S02. Two plants respectively using the processes (2)
959
-------
and (?) have started operation recently; the former at the Yokkaichi plant
of Showa Yokkaichi Oil, treating flue gas (64,000scfm) from an oil-fired
industrial boiler and the latter at the Shimozu plant of Maruzen Oil,
treating tail gaa (24,000scfm) from a Glaus furnace. A plant with the
process (4) will start operation soon at the Chiba plant of Idemitsu
Kosan, to treat 275,000scfm gas from a Glaus furnace and an industrial
boiler. The recovery of S02 from tail gas of the Glaus furnace to return
to the furnace may be made fairly economically, but the by-production
of sulfur at power plants would be costly. The sulfur by-producing
processes will be further developed when the oversupply of other by-
products becomes obvious.
Elemental sulfur has also been recovered by the hydrogen sulfide
recovery processes. By the Takahax process a fine powder of sulfur is
obtained which is intended for use in agricultural chemicals. Not many
new uses for sulfur have been studied in Japan as they have been in
the U.S.A.
T Ammonium sulfate
Ammonium sulfate was produced until recently at the Yokkaichi Plant of
Chubu Electric using the Mitsubishi manganese process. The plant
discontinued operation recently because of the difficulty of removing
more than 9C$> of the 802 in flue gas and the economical disadvantage.
Until a few years ago there were several ammonia scrubbing plants treating
tail gas from sulfuric acid plants to by-produce ammonium sulfate. All
of the plants were shut down because of the oversupply of ammonium sulfate.
Now there are a few small ammonia scrubbing plants treating flue gas to
produce a dilute ammonium sulfate solution to be discarded. The discarding
of aasftonium sulfate solution is to be restricted because it can cause a
eutrophication problem. Due to the present shortage of nitrogen fertilizers
in developing countries* Nippon Eokan is now considering building an
ammonium scrubbing plant at its Fukuyama Works to by-produce ammonium
sulfate for export. Nippon Eokan has already developed a process by
which S02 in waste gas and ammonia in coke oven gas are both recovered
to produce ammonium sulfate?). There is some doubt, however, as to^whether
a market for ammonium sulfate can be secured until several years from now.
Another problem in ammonia scrubbing is the plume formation. Nippon Eokan
is considering the use of an electrostatic precipitator for plume prevention.
The plume can be prevented by the use of an acidic solution as an absorbing
liquor, as is being done with Kurabo's acidic ammonium sulfate-lime
process4).
960
-------
VI Calcium sulfite
Not much calcium sulfite has been produced in Japan because of the
limitations on the use and on the landspace available for discarding it.
Mitsui Aluminum Co., which has produced calcium sulfite sludge since 1972,
is going to use the gypsum production process for new installations because
of the poor nature of the sludge. There is no focus on sludge stabilization,
as there is in the TJ.S.A.
The calcium sulfite obtained by the wet lime processes usually consists of
very small crystals about 0.1 micron in thickness and about 1 micron in
length and is not easy to filter. The calcium sulfite from the sodium-
limestone processes of Showa Denko and Kureha-Zawasaki grows imto much
larger crystals about 1 micron in thickness and 10 to 30 microns In
length. At the Saganoseki Smelter of Nippon Mining Co., the sulfite
obtained by the Showa Denko process is filtered by a vacuum filter, mixed
with copper ore, And is fed into a smalter to recover 802* Also for discarding,
the sulfite by the sodium-limestone process might be better than that by
the wet lime process.
A synthetic paper from fairly pure calcium sulfite and polyethylene, at a
weight ratio of about 70:50, has been produced recently by Lion Eat and
Oil Co. jointly with Idemitsu Kbsan. This new product has some defects
and is now under improvement.
VII Gypsum
Demand for and supply of gypsum Most of the big S02 recovery plants now
under construction or being planned are oriented toward the by-production
of gypsum for the following reasons: (l) Japan has plenty of limestone.
(2) Other by-products such as sodium salts, sulfuric acid, and ammonium
sulfate will not increase much since they are already in oversupply.
(3) Production of elemental sulfur from S02 in waste gases is not very
easy. (4) Japan does not have much available land On which to dump calcium
sulfite sludge. (5) Demand for gypsum has been increasing considerably.
(6) Gypsum is suitable for discarding in the case of oversupply.
The demand for and supply of gypsum in Japan is illustrated in Figure 3.
All of the by-product gypsum has been used so far for wallboard production
and as a retarder for cement setting because there has been a slight
shortage of gypsum in Japan since 1971• But since so Jhany desulfurization
plants by-producing gypsum are to be installed, oversupply of gypsum is
considered likely to occur in the future. The processes and installations
by-producing gypsum are described in the author's other paper for the
present symposium4).
961
-------
Quantity( millions of tons)
O ro -F- ON oo
-
Demand
Supply
on
B
C
OS
R
P
Demand
OU
B
C
>>
r-t
P<
P.
•3
W5
OS
R
P
OU
B
C
OS
R
P
1970
1973
1976
Demand C:Cement B:Board OU:Other uses
Supply P:Phosphogypsum RrRecovered OSrOther sources
Figure 3 Demand for and supply of gypsum in Japan
Use of gypsum for wallboard and cement For wallboard production^ an
appropriate crystal size ^larger than about JO microns in length and more
than about 10 microns in thickness) and low impurity are favored. Gypsum
obtained from oil-fired flue gas usually meets these requirements.
(Photographs A-F).
Th© by-product gypsum from oil-fired boiler flue gas ia nearly white or light
brown in color. Most of the wet lime-limestone process plants in Japan
have a cooler or a prescrubber of flue gas where the gas is sprayed with
water for cooling as well as for humidifying and removing most of the dust
which was not caught by an electrostatic precipitator.
Dark colored gypsum* since it contains a considerable amount of carbon
dust, has less commercial value. The gypsum obtained at a new plant of
Mitsui Aluminum Co. from coal-fired boiler flue gas passing through an
electrostatic precipitator by the Mitsui Mike limestone process (which
962
-------
Photomicrograms of by-product gypsum ( x 100 )
(A) Mitsubishi-JECCO
( CaCOj scrubbing)
(B) Mitsubishi-JECCO
( Ca(OH)2 scrubbing)
(C) Babcock-Hitachi
( CaCO; scrubbing)
(D) Chiyoda
-------
Photomicrograms of by-product gypsum ( x 100 )
(E) Kureha-Kawasaki)
( Na2SOj-CaCO, )
(F) Nippon Kokan
( (NHv)2SO,-Ca(OH)2
Scanning type electron photomicrograms of the broken
surface of a piece of gypsum(G) and GPC(H) ( x 1,500 )
(G) Gypsum
(H) GPC
-------
has no prescrubber) contains about 10$ fly ash and some carbon and is gray
in color. The gypsum has been tested by the sister company, MtLtsui Toatsu
Chemicals, for wallboard production. Similar gypsum to be obtained at the
Takasago Power Station of Electric Power Development Co., will possibly be
used for cement.
For use as a retarder in cement setting, gypsum should contain less than about
10$ moisture because wet gypsum tends to form a "bridge" in the hopper and
cannot be charged smoothly to the cement mill. Normally the by-product
gypsum contains less than about 10^ moisture after being centrifuged. Well-
grown gypsum produced by the sodium-limestone process (100-500 microns,
Photograph E) contains only 5 to T% moisture after being centrifuged.
Presence of sodium in gypsum can adversely affect the property of cement
but sodium in gypsum produced by the sodium-limestone process is in a
negligible amount because of the simplicity of washing due to the large
crystal size. Fly ash, about 10$ in gypsum, has no bad effects. Calcium
sulfite can be used also for a retarder in cement setting replacing a portion
of gypsum.
The KasMma Station of Tokyo Electric Power has produced powdery gypsum at
a 150MW plant by the reaction of powdered limestone with dilute sulfuric
acid (15-20$) obtained by a water wash of carbon which has absorbed S02*
The filtrate from the gypsum centrifuge has to be discarded because its cir-
culation in the washing produces solid deposits of the carbon decreasing its
activity.
As an alternative, the station has recently operated a 15MW test unit where
the acid is concentrated to 50-70$ by contact with the hot flue gas from
an electrostatic precipitator and then treated with powdered limestone*
By the heat from the reaction most of the water is eliminated to produce
nearly dry gypsum, which is extruded to form pellets about 20mm in diameter
and 30mm long. The pellets are suitable for charging into cement mills and
also might be suitable for landfill or disposal.
Gypsum for building material As a considerable oversupply of gypsum may
occur in the future, new uses of gypsum have been recently studied by many
organizations. The most promising new use is as a building material. The
usual type of calcium sulfate hemihydrate ( £ type) has a lower strength
than concrete (Figure 4)* The hemihydrate of cL type has a much larger
crystal size and higher strength than jS type but is fairly expensive. The
form II anhydrite which is obtained by heating gypsum at 950-1,000°C
hydrates fairly rapidly when a small amount (l-2$) of potassium sulfate is
added and increases the strength. Recent tests by Onoda Cement Co. have
shown that an anhydrite of good quality can be obtained with by-product
gypsum from S02 recovery if the fly ash content is less than about 5$.
A larger amount of fly ash tends to decrease the strength.
965
-------
•H
CO
PH
a
Q>
s-<
-4-J
CO
•H
CO
CO
CD
s
o
o
6,000
if, 000
« 2,000
0
II-anhydride W/G=0.35
ol -hemihydrate W/G=0.36
-hemihydrate W/G=0.6
-hemihydrate W/G=0.?
0
23
Aging ( week )
Figure 4 Strength of various types of gypsum and concrete
( W/G and W/C mean weight ratio of water against gypsum
and cement, respectively)
Technology for reinforcement of gypsum with glass fiber has been developed
recently in England?/. The reinforced gypsum from d- type hemihydrate has
an equal compression, bending, and tensile strength and much higher impact
strength when compared with asbestos-reinforced concrete.
Gypsum plastic composite An important defect of gypsum as a building
material is its lack of resistance to water. To eliminate this weakness,
gypsum plastic composite (GPC) has been recently developed in Japan
through the cooperation of Mitsui Toatsu Chemical and Taisei Construction
Co. (Figure 5)» Usually a resin monomer such as methyl methacrylate (MHA.)
or styrene is used for impregnation. The monomer is polymerized by a
thermal catalytic means. Some results of tests with MMA. are shown in
Tables 3 and 4* G£C has superior qualities in its strength, resistance to
water, acid and base, and also in its good workability and semi-incombustible
property. It may, therefore, be used as a Mgh-grade building material.
966
-------
Gypsum
Additive
Water
Resin monomer
Catalyst
Mixing
Molding
Impregnation
drying
Evacuation
Polymerization
Figure 5 Process for GPC production
Finishing
I
Product
Table 3 Blending of materials
Ho. ft -hemihydrate
A 100
A' 100
B 100
B' 100
Glass
fiber
0
0
3
3
Water
57
57
63
63
MM
0
35-5
0
38.1
Table 4 Properties of gypsum and GPC
Ho.
A
A«
B
B'
Specific
gravity
(G/ml)
1.274
1.699
1.214
1.663
Conpressive
strength
(i>si)
1,870
10,200
1,410
11,220
Bending
strength
(nsi)
740
2,780
966
3,890
Wearing
(mils/1,000
revolutions)
360
32
720
44
967
-------
Observations of the broken surface of pieces of ordinary gypsum and GPC
by ft scanning type electron microscope have shown that fox gypsum, crystals
were not broken but came apart from each other by stress, while for GPC,
each crystal was broken requiring a great stress (Photograph H). This
explains the high strength of GPC.
References
l) Jumpei Ando, Proceedings of Flue Gas Desulfurization
Symposium 1973* PP 875-890
2) Jumpei Ando, ibid page 96
3) M.A. Ali, J. Material Science, 4 (5), p 398 (1969)
4) Jumpei Ando, Status of Flue Gas Desulfurization Technology
in Japan. EPA EDO Symposium, Kov. 1974
968
-------
TVA-EPA STUDY OF THE MARKETABILITY OF ABATEMENT SULFU1 PRODUCTS
J. I. Bucy and P. A. Corrigan
Tennessee Valley Authority
Muscle Shoals, Alabama
Prepared for Presentation at
Flue Gas Desulfurization Symposium
Sponsored "by the Environmental Protection Agency
Atlanta, Georgia
November U-7,
969
-------
CONTENTS
Page
Introduction 972
Results of the Phase I Study 973
Preliminary Feasibility Study of Calcium-Sulfur
Sludge Utilization in the Wallboard Industry 998
Phase 2 of the TVA-EPA Study of the Marketability
of Abatement Products 1000
970
-------
TVA-EPA STUDY OF THE MARKETABILITY OF ABATEMENT SULFUR PRODUCTS1
J. I. Bucy and P. A. Corrigan
Tennessee Valley Authority
Muscle Shoals, Alabama
ABSTRACT
During the past year a hypothetical study of the poten-
tial for marketing abatement sulfuric acid produced from SOg which
is emitted by the power plant stack gases from seven TVA coal-
burning steam plants was made in cooperation with EPA. A production-
distribution model was developed on a time-sharing computer to
determine maximum net sales revenue to the utility from sales of
sulfuric acid to the relatively inefficient, high-cost sulfuric acid
producer. The results of the study indicate that for the approxi-
mately 2 million tons of sulfuric acid that could be produced annually
by TVA, the net sales revenue above distribution costs only would be
approximately $8.75 per ton. If TVA were to use 10 percent of the
acid at Muscle Shoals for the production of wet-process phosphoric
acid or other fertilizer, the net sales revenue would be increased
approximately 50 cents per ton of sulfuric acid.
Based on the above results an expanded study has been
initiated to determine for specific power plant installations the
potential net sales revenue for byproduct elemental sulfur and
sulfuric acid which can be realized from marketing strategies cover-
ing the existing acid market, the existing elemental sulfur market,
and growth markets for such commodities. This study should have
general application for the geographic area located east of the
Rocky Mountains in the United States. The effect of marketing
abatement production of sulfuric acid and/or elemental sulfur on
existing production- shipment- consumption-pricing patterns will be
simulated using the product ion- distribution model developed in Phase
I. The expanded model will be developed on a time-sharing computer
which facilitates access for any specific company.
At the same time a second study was conducted on the po-
tential marketability of abatement calcium sulfate production. That
is, instead of discarding calcium- sulfur sludges, if the sulfur
dioxide is recovered as gypsum, it may be possible to sell it to
the wallboard industry.
Depending on the power plant location and processing costs,
this advantage over a throwaway process could range from nothing to
over $l6/ton of gypsum disposed. For a 500-MW power unit, this
could be as much as a $3»8-million-per-year reduction (^5$) in the
cost of operating a lime-scrubbing, sol ids -throwaway system.
To be presented at the Flue Gas Desulfurization Symposium, Atlanta,
Georgia, November ^-T ,
971
-------
TVA-EPA STUDY OF THE MARKETABILITY OF ABATEMENT SULFUR PRODUCTS
INTRODUCTION
Our Nation faces a dilemma in implementing the Clean Air
Act as amended in 1970, which established strict requirements and
timetables for cleaning the air. The dilemma concerns both the
direction to take and the achievement of objectives once a direction
is determined. Time is critical; the target date of mid-1975 for
implementing the SOX regulations is approaching rapidly.
The solution is far from simple. One major complicating
factor is the energy crisis. Supplies of low-sulfur fuels are in-
adequate for them to be the sole means of achieving compliance. In
addition, the design of many electric power plants is not conducive
to use of low-sulfur fuel. Other alternatives--such as dispersion
techniques, use of products from coal gasification of liquefaction,
or addition of flue gas desulfurization systems—must be considered
for many plants. Then there is the matter of side effects of SOX
removal systems--e.g., disruption of supply channels for raw materials
normally used elsewhere, excessive demand on fabricators of needed
equipment, disposal of waste products, and market impact of saleable
byproducts.
For the air quality control areas where flue gas desulfur-
ization systems will become the major control technique, numerous
alternatives are available as to the type of system to install and
the byproducts produced. Byproducts are both waste and saleable
material, such as calcium sludge, gypsum, liquid S02, ammonium
sulfate, elemental sulfur, and various concentrations of eulfuric
acid. The large tonnages of these byproducts vhich could be produced
from the S02 emissions projected for 1980 could create an over-
whelming disposal problem. If all the S02 emissions projected for
1980 were converted to saleable materials, the amount of sulfur in-
volved would be about 1.7 times the expected 1980 U.S. consumption of
sulfur. On the other hand, if all utilities installed a calcium-
based scrubbing throwaway process the impact on the agricultural lime
market could be severe. The limestone requirements for this use
would be 2.6 times the projected need for agricultural lime. To
further complicate this alternative, the resulting calcium sulfate
sludge piles in total could exceed the size of the pyramids of Egypt.
In recognition of this dilemma EPA has initiated a research
project with TVA to study the possible impacts of saleable abatement
materials on existing and future markets. The overall research effort
has been arbitrarily divided into five phases.
972
-------
The proposed fifth phase involves a more detailed market
study of the potential utilization of calcium-sulfur sludge by the
wallboard manufacturing industry to derive more accurate cost data
for comparison of throvaway alternatives. A thorough economic
evaluation would be made of the more promising processes—such as
the Chiyoda and the carbon absorption plus the oxidation step for
conversion of calcium sulfite to sulfate. A preliminary study was
conducted early in 197^- At that time a report was prepared by TVA
for EPA entitled "Preliminary Feasibility Study of Calcium-Sulfur
Sludge Utilization in the Wallboard Industry" by P. A. Corr,igan.
This report will be discussed later in the paper.
RESULTS OF THE PHASE I STUDY
This study was sponsored by the Office of Research and
Development, US Environmental Protection Agency, Research Triangle
Park, North Carolina in cooperation with TVA.
The objective was to create a model for estimating the net
sales revenue to TVA for marketing the abatement acid which could
be produced. The cost of removing sulfur dioxide and producing the
sulfuric acid is considered independent from this evaluation.
The study assumes that an acceptable flue gas desulfur-
ization system equipped with a sulfuric acid production process is
commercially available and could be installed at TVA steam plants.
Sulfur Dioxide Removal Processes
The sulfur dioxide removal processes being developed with
financial assistance from EPA include several which could produce
sulfuric acid as a marketable product—the Chemico-basic magnesia
scrubbing process, Davy Powergas - Sulfite scrubbing process, and
the Monsanto catalytic oxidation process. Philadelphia Electric is
installing u privately funded magnesia scrubbing process at its
Eddystone plant. The demonstration sized plants in the United States
using technology from these processes are listed in Table 1 below:
Table 1.
Process
MgO scrubbing
Sodium sulfite
scrubbing
Catalytic oxidation
REGENERABLE PROCESS DEMONSTRATION
Demonstration Utility Company Product
155 MW oil (1972)
100 MW coal (19710
120 MW coal (1975)
115 MW coal (1975)
110 MW coal (1974)
Boston Edison
Potomac Electric
Power
Philadelphia
Electric
Northern Indiana
Public Service
Illinois Power
Sulfur
973
-------
gulf uric Acid Market
About 90 percent of the elemental sulfur consumed in
the United States is used to make sulfuric acid. The following
industries use elemental sulfur for non-acid purposes: Agricultural
Chemicals, pulp and paper, carbon disulfide, rubber, sugar, starch,
malt, and dye stuffs. It is apparent that elemental sulfur and
sulfuric acid offer the greatest potential to the electric utilities
for abatement production from the 302 emissions.
Sulfur dioxide (802) is oxidized by air in the presence
of a catalyst to form sulfur trioxide (SO^) which combines spon-
taneously and irreversibly with water vapor to form sulfuric acid.
There are alternatives as to the source of SOo as well as the
method of conversion to SCb.
There are two principal methods for conversion of SOo
to SOo which are known as the chamber and contact processes. The
older chamber process, which was introduced in the 18th century,
uses nitrogen oxides as an oxygen- carry ing catalyst for the
conversion of SC>2 to SOo. The reactions which produce the SOo and
sulfuric acid take place either in huge lead chambers or packed
towers .
The more modern contact process converts SOg to 803 by
use of a metal or metal-oxide catalyst. The SOo is then passed
through an absorption tower where it is absorbed in recirculating
concentrated acid. The major advantages of the contact process are
that concentrated acid of high purity can be produced directly
and compact plants of high capacity are feasible.
Alternatives sources of SOg for manufacture of sulfuric
acid include (l) elemental sulfur, (2) pyrites (sulfite ores or
iron, copper, lead, or zinc), (3) waste gases from metallurgical
refining operations, (U) hydrogen sulfide from sour gas or
petroleum, (5) sulfur-bearing ores of volcanic origin, (6) sulfate
suits such as gypsum or anhydrite, und (7) waste gases from
combustion of sulfur- containing fuels. Only elemental sulfur,
pyrites, and sulfates (gypsum) are considered to be true basic
raw materials for the production of sulfuric acid since other
sources yield S02 as a by-product. This study focuses on the
potential production of HgSO^ from waste gases and combustion of
sulfur containing fuels used in electrical generating steam plants.
974
-------
Current Production
United States production of sulfuric acid in 1972 totaled
over 31.1 million short tons (100$ f^SOlJ or about 1.2 percent above
1971' s production. This represented approximately 80 percent of the
total production capacity in the United States of about kO million
tons. The spacial distribution of the production is approximated
in Figure 1. About 60 percent of the production capacity was
committed for captive use. Only about 12.5 million short tons was
externally marketed from the 1971 production of 29. ^ million short
tons. The 1970 capacity by states (short tons /day) is outlined in
Table 2. The five states having the most capacity for acid manu-
facture include Florida, Louisiana, Texas, New Jersey, and Illinois.
The size of individual acid plants has increased over the
years. Plants of 1,000 tons per day capacity have now become common-
place, and capacity of up to 2,000 tons per day have been recently
constructed. Such plants are usually a part of fertilizer complexes.
A few of the old chamber process plants are still in operation, but
the majority of the plants use the more modern and efficient contact
process. Most existing sulfuric acid plants do not have adequate
pollution control facilities. This is discussed further in the
production-distribution model.
Sulfuric acid is made and used in a variety of concentra-
tions which are usually indicated as follows :
or °Baume - The simplest description of
sulfuric acid concentration is $ HgSO^. However,
because of the distinct relationship between specific
gravity and strength (up to 93$) a"-d the simplicity
of measuring specific gravity by hydrometer, most
acid concentrations up to 93 percent are expressed
as degrees Baume. From 93 to 100 percent acids
are referred to by % concentration of
Current Consumption
The major end uses of sulfuric acid in the United States
in 1970 are shown in Table 3« Fertilizer consumption represented
5^ percent of the sulfuric acid consumed. The long-range growth
in acid consumption is estimated to be about ^ to 6 percent per
year which is closely tied to the fertilizer growth pattern.
Although most sulfuric acid consumed in fertilizer manu-
facture is a concentrated, high-quality material, off-grade acid
could be used as well. For other end uses of sulfuric acid, high
purity and high concentration are almost mandatory.
975
-------
oooooooOO
O. 901. «<». i»d- tOOU t*Ot- 1OOI' UO»- 4OQU
SOO WOO )*OO fOOO ItOO MOD MOO 4OOO tOOO
ooooOO
Figure 1. Sxilfuric acid manufacturing capacity (1970)
-------
Table 2. SULFURIC ACID PLANT CAPACITY (1970)
(short tons/day)
State
Alabama
Arizona
Arkansas
California
Colorado
Delaware
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Capacity
1,610
2,627
737
6,771*
1,1*83
1,050
23,661
1,369
3,1*70
6,9****
2,066
1,877
7l*7
550
12,600
223
2,260
330
1,301
State
Mississippi
Missouri
New Jersey
New Mexico
New York
North Carolina
Ohio
Oklahoma
Pennsylvania
Rhode Island
South Carolina
Tennessee
Texas
Utah
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Capacity
1,067
3,303
6,913
1*1*6
583
3,**8o
3,180
630
2,177
50
321*
1*, U21
9,855
2,133
1,983
333
1*70
.T
67
360
Grand Total
113,
977
-------
Table 3. SULFUR1C ACID END USE PATTERN (l9?0)
Thousand
short tons
End Uses (100% basis)
Fertilizer ~~~~
Phosphoric acid products 13,750
Normal superphosphate 1,2MQ
Cellulesics
Rayon 520
Cellophane 170
Pulp and paper 600
Petroleum alkylation 2,1*00
Iron and steel pickling 800
Nonferrous metallurgy
Uranium ore processing 300
Copper leaching 350
Chemicals
Ammonium sulfate
Coke oven 500
Synthetic kQO
Chemical by-product 190
Chlorine drying 150
Alum 600
Caprolactam 260
Dyes and intermediates 370
Detergents, synthetic 2+00
Chrome chemicals 100
HC1 150
HF 880
TiOa 1, MO
Alcohols 1,800
Other chemicals 380
Industrial vater treatment 200
Storage batteries
Other processing
Total'"2o,6UO
978
-------
It is apparent from Table 3 that sulfuric acid has a wide
variety of uses. Some uses are based on excellent physical qualities,
but cost is also important. Sulfuric acid is often preferred over
other mineral acids, chemicals, or different process technology
because it is the least expensive. For example, in phosphate rock
acidulation and phosphoric acid manufacture, it is the lowest cost
acidularvt available. Also, the use of sulfuric acid for leaching
low-grade copper oxide ores has been feasible only because of the
availability of low-cost byproduct sulfuric acid from western copper
smelting operations.
Current Transportation
Location of power plants equipped with sulfur dioxide
removal and sulfuric acid production facilities will have a major
influence on the marketing costs of abatement acid production.
That is, power plants that enjoy the use of water transportation
will be in a position to reach more distant markets due to the
relatively lower transportation costs as compared to rail or truck.
Figure 2 outlines the relative costs of shipping sulfuric acid by
barge, rail, and truck. (P. *A, Sulfur Markets for Ohio Utilities,
by J. F. Foster, et. al., EPA ^50-3-7^-026).
Sulfuric Acid Production Capacity of TVA
TVA is a corporate agency of the United States created by
the Tennessee Valley Authority Act of 1933- In addition to various
other programs, TVA operates a system supplying the pover require-
ments for an area of approximately 80,000 square miles containing
about 6 million people. Except for direct service by OTA to certain
industrial customers and Federal installations with large or unusual
power requirements, WA pover is supplied to the ultimate consumer
by 160 municipalities and rural electric cooperatives which purchase
their power from TVA. TVA is interconnected at 26 points with
neighboring utility systems.
Power loads on the TVA system have doubled in the past 10
years and are expected to continue to increase in the future. To
keep pace with the growing demand it has been necessary to add
substantial capacity to the generating and transmission system on a
regulur basic. Current plane are based on meeting future additional
requirements with nuclear power stations.
The following tabulation breaks down the TVA power genera-
ting system capacity into several categories:
979
-------
12
o
•tie-
O
o
£L
Q.
RANGE OF
BARGE COSTS
100
Figure 2.
200 300
DISTANCE, MILES
Sulfurie ueid. shipping, cost.
4OO
500
980
-------
Capacity in service Under Construction
June 30, 1972 or scheduled
No. of No. of
plants Megawatts plants Megavatts
Coal-fired steam plants 11 15,509 1 2,600
Hydroelectric plants 29 3,185
Nuclear plants 4 11,101
Gas-or-oil-fired turbines 2 688
Figure 3 shows the location of TVA's present generating
facilities and those under construction.
The total of 18,109 megawatts of coal-fired capacity is of
primary interest because it represents the potential for sulfuric
acid production. However, only a portion of this capacity is used
as "base load"—that is, operated continuously except for maintenance.
These are the newer, larger and more efficient plants. The other
portion is used as "swing load," that is, intermittently, or at
times of peak demand. These are the older, smaller and less efficient
plants.
The TVA plants which would have the greatest potential for
the installation of sulfuric acid production facilities would be the
base-load coal-fired plants, except Bull Run which burns low-sulfur-
content coal, 1.5 percent. This is "based on the indication that SC^
recovery and sulfuric acid-producing facilities would be less
competitive in intermittent service for TVA than limestone scrubbing
or other "throwaway processes" facilities. Also, SOg recovery and
acid-producing facilities operate more efficiently under continuous
duty with constant operating conditions.
One of the relatively new and large units is being
equipped with a limestone scrubbing S02 removal system. This plant
is the Widows Creek Unit 8 and is not considered a potential
sulfuric acid producer. The swing load plants -- Colbert Units 1-4,
John Sevier, Johnsonville 1-6, and Kingston -- generally would
have limited potential for acid production.
Therefore, of the total 18,109 megawatts of coal-fired
capacity, 9,806 megawatts would be considered for sulfuric acid
production.
To determine the amount of acid which could be produced by
the 9,806 megawatts, it was assumed that about 90 percent of the
sulfur in the coal would end up as S02 in the stack gas. The re-
maining sulfur is rejected in the coal mills as pyrites, left in
981
-------
vo
00
to
TENNESSEE VALLEY REGION
S
Strain Plants
Coat-Fired
Nude*
Ojnn
Aluminum O>. of America Dam
Under Construction
Approximate Area* Served
by Municipal & Cooperative
Dirtributor* of TVA Power—
*O«I tCHJPCXJ^ !.«>
**rrs i
lNOXVtLl£
ao7
_
PAOUCAM
MH_£ 22
PROFILE OF THE TENNESSEE RIVER (ALL MAINSTREAM DAMS HAVE NAVIGATION LOCKS)
Figure 3- Location of TVA power plants,
-------
the ash, or unaccounted for. For every pound of sulfur oxidized, 2
pounds of SC>2 are produced and for every pound of SC>2 that is
recovered, 1.53 pounds of sulfuric acid can be produced. Furthermore,
it was also assumed that the SOg removed for each plant would be
based on the EPA emission standard for new coal-fired steam plants—
1.2 pounds of SC>2 per million Btu heat input.
Using these assumptions, plus projections for sulfur in
coal and the amount of generation expected at each plant (supplied
by TVA Division of Power Resource Planning), estimates of potential
culfuric acid production from TVA's plants from 1975 through 1985
were made. Consideration was given to the oncoming new plants --
coal-fired and nuclear -- and the effect of time, age, and mainte-
nance on operating schedules for existing plants. Coal analyses were
based on 1972 data. It would, of course, be impossible for acid
facilities to be installed by 1975> however, this is the base year
assumed for startup. The forecast is as follows:
Forecasted Theoretical Production of Sulfuric Acid
Steam plant and unit
Colbert 5
Cumberland 1-2
Gallatin 1-k
Johnsonville 7-10
Paradise 1-3
Shawnee 1-10
Widows Creek 7
TOTAL
(M
1975
121.9
578.7
165.3
135-9
617.3
270.0
92.6
1,981.7
Tons)
1977
121.9
578.7
159-8
135-9
617.3
253-5
92.6
1,959.7
1980
121.9
578.7
137.8
111.9
608. h
253-5
86.0
1,898.2
1983
84. k
520.8
99-2
55-9
573.2
137.8
66.1
1,537.^
1985
Qk.k
1*71.2
71.6
Uo.o
5^6.7
99-2
52.9
1,366.0
Market Approach
The major sulfuric acid manufacturing-marketing schemes
which prevail in the existing market are:
1. Production of acid neur the point of use from
purchased sulfur.
983
-------
2. Production of acid near the source of sulfur by
the basic sulfur producer.
3. Marketing of spent or regenerated, acid.
JK Marketing of acid recovered from pollution
abatement processes.
The first (production from purchased sulfur) is the most
vulnerable because the producer is dependent on an external source
of sulfur. The acid producer who is basic in sulfur in No. 2, above,
would consider the investment in mining facilities as a sunken in-
vestment. This means that he would take into account only his out-
of-pocket costs in meeting market price pressures from abatement
sulfur products. The arrangements for utilization of spent acids
(No. 3) above) are highly specialized. It would, therefore, be
difficult to place abatement acid in the regenerative acid market.
The most orderly way to incorporate the abatement acid
into the existing market would be to replace the capacity of sulfur-
burning sulfuric acid plants. Such plants purchase sulfur from
external sources. Therefore, the marketing strategy assumed for
this study is to substitute abatement acid for purchased sulfur.
Market Potential for abatement acid
TVA's National Fertilizer Development Center maintains a
computerized file of world-wide manufacturers of fertilizers and
related products. A list of sulfur-burning acid plants currently
in production or planned thru 1975 v&s developed from the TVA file.
The potential market was limited to a 10-state area on the inland-
waterway system in the central United States. All the TVA power
plants are located with access to this waterway. The states
selected included: Alabama, Arkansas, Illinois, Kentucky, Louisiana,
Mississippi, Missouri, Ohio, Tennessee, and Texas. Florida was
included as un alternate marketing area if required.
Information from the TVA file offered the following data
for sulfuric acid plants:
1. Company
2. Location
3- Annual capacity
U. Type of process
Dates of construction and major capital improvements were
obtained from other sources. A total of 6l sulfuric acid plants
were selected as potential points for acid sales. These points can
be grouped into seven metropolitan market areas -- Memphis, Houston,
Chicago, New Orleans, Cincinnati, Columbus, and Tampa.
984
-------
The production from the 6l plants represents the market
potential for abatement acid. The market demand from these plants
is dependent upon price incentive. The primary incentive to purchase
acid vill be the cost reduction enjoyed as compared to manufacturing
the acid from purchased sulfur. This assumes that price, quality,
and convenience are the major factors that influence product or
process substitution. It is further assumed that the acid plant vill
buy abatement acid if the delivered cost is lower than its production
cost. The more inefficient plants become the prime consumer of
abatement acid. In order to move the total production, some of the
more efficient plants would have to be shut dovn; therefore, the
price vill be influenced by the volume of abatement production.
Avoidable Costs
Avoidable costs are defined as those costs which a
producer would not incur if he discontinued the operation of his
plant. They are:
Cost Category Cost Breakdown
Raw material Sulfur
Utilities Electric power, cooling
water, process water,
boiler-feed water.
Operating Expenses Labor, supervision,
payroll overhead.
Capital Costs Amortized costs for
maintenance of existing
facilities plus amortized
costs of new capital
investment at the end of
useful plant life.
An adjustment for loss of steam generation in the acid
plant is required.
Sulfuric Acid Production - Distribution Model
In the derivation of a model to maximize the net sales
revenue from the sale of abatement acid the following factors were
considered:
1. Trade off between avoidable costs at 6l acid
plants and shipping distances from seven power
-------
plants.
2. Effect of sulfur price.
3- Effect of volume on net sales revenue.
The combinations of these factors contribute to the
complexity of the evaluation; therefore, use of a computer is almost
essential to establish maximum revenues. A production-distribution
model (similar to a transportation linear program model) was
developed to handle the several variables. The objective of the
model is to minimize acid costs to the existing sulfuric acid plant
locations while maximizing net sales revenue to TVA.
The programming model was designed so that the key
technical and economic parameters can be varied. Table k lists the
major parameters and shows typical values.
The first three parameters relate to sulfur conversion
efficiency as a function of plant design; the data are based on a
report by the Chemical Construction Corporation. Plants built
prior to 1960 average 95•5 percent conversion. The newer plants are
more efficient, 97 percent. Parameters k thru 9 are used to calculate
the manufacturing costs of sulfuric acid. (An example is shown in
Table 5). The values for the investment parameters k thru 6 are
estimates based on the initial capital estimates shown in Table 5•
The utility costs (parameter 7) are fixed per ton of sulfuric acid
and the operating expenses (parameter 8) are annualized; taxes and
insurance (parameter 9) are proportional to initial capital in-
vestment.
The annual costs are calculated in perpetuity utilizing
the discounted cash flow analysis method. The outlay streams are
then amortized, or averaged over all years in the firms planning
horizon. The cost streams are composed of (l) constant annual
expenditures for sulfur, utilities, and operating expenses; (2)
periodic expenditures for new plants; and (3) maintenance of exist-
ing facilities which is assumed to grow at a compound rate. The
impact of inflation is not included in the analysis. These cost
streams are presented in Figure h. It is noted that the average
capital costs decline as useful life increases.
The optimium useful life is identified as the minimum
point on the average total cost curve. At this point the added
capital cost savings enjoyed by increasing useful life one year
equals the added maintenance saving from shortening useful life
one year. It is noted that the average total cost curve in
Figure 4 is very flat over a wide range of years. The average
capital charge of 1^.9 percent, which is identified in Table $,covers
a range from 29 to Ul years. Possibly random effects, such as
586
-------
Table 4. MAJOR PARAMETERS IN MODEL
No.
1
2
3
k
5
6
7
8
9
10
11
12
15
Ik
15
16
17
18
19
20
Description of variable
Tons of sulfur per ton HgS04 (before YEAR60)
Tons of sulfur per ton H2S04 (after YEAR6o)
Year of technology change
Sulfurlc acid plant investment ($/ton-year)
Capacity for this plant (M tons/year)
Scale factor for determining investment for
other sited plants
Fixed conversion cost per ton ($/ton)
Fixed annual conversion cost ($/year)
Taxes and insurance rate
Time preference rate for money
Compound maintenance rate
Economic useful life
Percent H2S04 concentration
Port Sulphur price ($/short Con)
TVA H2S04 price ($/ton HeS04)
Proportion of 330 TPD capacity estimate
Number of steam plants
Number of acid plants
Number of years considered
Years considered
Example
value
•3053
.3006
60.
27.285
21+7.5
.734054
^7
116.620
.015
.08
.01*
54-
98.
22.32
0.
1.
7-
61.
l.
75-
Fortran
name
PRE60
Post6o
YEAR60
EXPENDO
SIZED
FACTOR
AVC
AFC
TI8
RATEI
RATEM
USEUFE
ACDCON
PS
PA
DEMAND
NPLANTS
JNUM
NY EARS
YEAR(I)
Table 5. PRODUCTION COST ESTIMATES FOR SULFURIC ACID
Acid plant
Tons per day
Tons per year, at 330 days/yr
Initial capital, $
Unit capital, $/ton-yr
Operating costs, $
Utility costs
Electric power
Cooling water
Process water
Boiler feed water
Steam (credit)
Labor
Operating
Supervision
Overhead at 70^ above
Capital costs, $>
Amortized value of maintenance
plus capital outlays at
optimal useful life
(29-'»l yr), II*. 9^
Taxes and insurance, 1-1/2$
Annual operating cost, $
(excluding eulfur)
Unit cost, $/ton (excluding sulfur)
Capacity
50
16,500
909,000
55.09
11,570
6,oi»o
70
980
-10,870
1*7,500
21,100
1*8,020
135,^1
13,635
273^%
16.57
250
82,500
3,090,000
37-1*5
57,800
30,200
350
it, 910
-51*, 1*00
1*7,500
21,100
U8,020
1(60,1*10
1*6,350
662,21+0
750
2lt7,500
6,907,000
27.91
172,700
90,300
1,020
14, 730
-163,000
1*7,500
21,100
1*8,020
1,029,11*3
103,605
1,365,118
8.03
5-52
I
1,500
1*95,000
10,905,000
22.03
31*6,600
181,200
2,100
29,440
-326,000
47,500
21,100
1*8,020
1,624, &»5
165,575
2,138,380
4.32
987
-------
OO
00
CO
O
o
51
UJ
O
o:
ui
o.
OPTIMAL USEFUL
LIFE
Figure 1+. Amortized, value of maintenance and capital outlays for new plants.
(Assuming 8% interest and k% compound maintenance).
-------
abrupt physical, economic, technological, or environmental changes
pluy the dominant role during this period with regards to the timing
of plant replacement.
The unique concept incorporated into this model relates to
the method used for handling existing plants as compared to the
traditional static analysis used to justify the investment in a new
plant. For existing plants, the initial capital expenditures
represent a "sunken investment," and, therefore, do not enter
directly into the firm's decision-making process for making the
decision to discontinue present production and buy abatement sulfuric
acid. Only avoidable costs are considered in making this decision.
It can be demonstrated that only when avoidable costs are
considered for a one-year-old plant the level of costs decrease
from 1^.9 percent of the initial investment to 7-1 percent. At
this point, the added savings from postponing the building of a
new plant is just offset by added maintenance costs in the 3^th
year which is exactly equal to the optimal useful life for a new
plant.
When the model is focused on a thirty-year-old plant it
is noted that the level of costs have risen to lk.6 percent of the
initial capital expenditure and the optimal useful life is still
31*- years. It is recognized that management of a new plant is not
concerned with replacement alternatives, but the management of an
old plant is faced with impending replacement alternatives.
The managers handling the older plants should be receptive
to exploring the alternative of purchasing abatement acid because
their maintenance costs are high and within a few years they will
be faced with the imminent decision of plant modernization.
The computer program calculates the above mentioned costs
based on (l) an interest rate of 8 percent of total investment
(parameter 10, Table k); (2) a maintenance rate of k percent of the
initial investment compounded annually at the rate of k percent
(parameter 11, Table k); and (3) plant age. The user is given the
freedom of selecting useful life (parameter 12, Table k). The
program can be modified to calculate and use the optimum value for
this parameter.
Parameters 13 thru 20, Table 4, relate to the logistical
portion of the model. It is assumed that the competitive pricing
structure for sulfur in the United States is based on a Gulf Coast
price plus transportation costs to a given sulfur-burning sulfuric
acid plant. This assumption seems reasonable since firms buying
imported sulfur continually bargain against Gulf Coast producers.
989
-------
The model allows each acid plant three choices: (l)
continue to operate existing plant; (2) rebuild and operate a new
plant; and (3) shut down existing plant and buy acid. The abate-
ment acid would be expected to enter the market at a price no
higher than the cost which could be avoided by shutting down the
most inefficient plant. The actual price of the abatement acid is
influenced by the volume offered in the market. The model selects
the most profitable choice for the acid plant by minimizing acid
costs at existing sulfuric acid plant locations.
The solution to the model ranks all acid producers in
terms of production costs, age, size, and relative location from
their sulfur source as well as the abatement acid source. The
plants with the highest production costs become the prime consumers
of the abatement acid. This group has the oldest plants with the
lowest production capacity located in the more remote areas from
the Gulf Coast sulfur producer.
The FOB abatement acid price (initially zero) plus fixed
handling cost per ton associated with each steam plant are added to
the transportation costs in order to determine the delivered price
to each sulfuric acid plant. The maximum net sales revenue is
derived by adjusting the FOB price of the acid until the total
volume is sold. It is maximized by realizing the trade-off between
avoidable production costs at 6l acid plants and the shipping
distances from seven power plants.
Another important economic factor is the cost of pollution
abatement facilities that must be added to existing sulfur-burning
acid plants. This cost could be expected to vary considerably from
one plant to another due to age of plant and process used. Based on
Chemical Construction Corporation data, it is assumed that the
average would be about $3 per ton. This factor is not included in
the program; and in many cases, net revenue results shown later in
the report could be increased by this amount.
The program is written so that one or more years can be
considered simultaneously for a given run. The model examines each
plant to determine if that firm would be better off continuing
production or buying abatement acid. It also determines the optimum
distribution pattern from each steam plant to each acid plant. This
optimization is done in such manner as to result in the lowest
possible industry cost. The model can determine the quantity of
acid sold at a given price or the highest price which will just move
the required amount for each steam plant.
The model is written for Control Data Corporation Kronos
time-sharing and can run from most any location through a standard
telephone. Furthermore, the program can be made available to
990
-------
anyone interested in its use. The heart of the model is a conver-
sational linear programming package called APEX. The present
program calculates cost of each acid plant-steam plant combination
(presently over itOO) and then generates the required input data
file. APEX is run to optimize the model and a second program
interprets solutions as printed reports. An interactive system is
also available which can display any or all of the standard linear
programming solution values.
Freight Bates and Handling Charges
Freight rates used in the model were obtained from TVA's
Navigation Economics Branch which is located in Knoxville, Tennessee,
These rates can be divided into two categories:
1. Those used for shipping sulfur from Port Sulfur,
Louisiana to various plant locations. These
rates were used as a factor in determining the
costs of sulfuric acid production of each plant
location.
2. Those rates for shipping sulfuric acid from the
seven TVA steam plants to each of the sulfuric
acid production locations. These rates are a
factor in determining the net back to TVA.
An estimated cost of 20 cents per ton has been programmed
into the model to cover acid storage at the existing acid plants.
This would provide 30-day storage. The storage cost is based on an
estimated capital cost for the tanks and the auxiliary facilities
at $20 per ton.
Results of the Study
The base case market pattern is shown in Table 6 as the
most economical market pattern for TVA abatement acid given the
assumptions outlined in this study. The maximum net sales revenue,
excluding production cost, is identified as $8.76 per ton which is
the lowest of the marginal costs shown for each of the seven steam
plants. In this case, acid would be shipped by barge from seven
production points (steam plants) to 20 marketing points (sulfuric
acid plants). If a credit is added for the estimated increased
cost for installation and operation of tail gas clean-up system on
existing acid plants the net sales revenue would increase by
approximately $3 per ton of acid or a total of $11.76.
The effect that a change in TVA's net sales revenue or
"price" has on acid sales is shown in Figure j? for the base case.
991
-------
Table 6. BASE CASE MARKET PATTERN FOR TVA H2SOil
SULFUR PRICE » $22.32
PLANT
LOCATION
2. N.LITTLE ROCK.AR
23. E.ST.LOUIS,ILL.
29. MnNSANTO,ILL
30. E.ST.LOUIS,ILL.
32. CALUMET CITY,ILL
33. JOLIET,ILLINOIS
35. JOLIET,ILLINOIS
36. STREATOR.ILL.
37. E.CHICAGO,IND.
38. LASALLE,ILLINOIS
40. JOLIET,ILLINOIS
41. CALUMET CITY,ILL
42. CHICAGO HTS,ILL
46. BATON ROUGE,LA.
47. NEW ORLEANS,LA.
54. HAMILTON,OHIO
55. CINCINNATI,OHIO
56. CINCINNATI,OHIO
60. COLUMBUS,OHIO
61. COLUMBUS,OHIO
PLANT CAPACITY
PLANT PRODUCTION
MARGINAL ACID COST
-------
25
I
UJ
V)
QUANTITY H2S04 (MILLION TONS)
Figure 5. Demand for TVA sulfuric acid.
993
-------
As expected, acid movement declines as the "price" of TVA acid
increases. In order to move all of its acid, TVA could charge no
more than $8.76 per ton plus freight. It could expect to move
only about one-half of its production for $10. At $20 per ton of
acid, no acid could be sold externally.
By adjusting the assumptions in the model from 100 percent
barge transportation to a mixed rail and barge coupled by a
reduction in market demand equivalent to 75 percent of on-streara
time for existing acid plants the net sales revenue is reduced to
$5.99 per ton vithout credit for tail gas clean-up at the acid
plant. The results are outlined in Table 7.
The revenue from the sale of abatement acid given the
assumptions used in this study is directly proportional to the
sulfur price. That is, the increase of $5 P6*1 long ton of sulfur is
equivalent to approximately $1.42 increase in potential net sales
revenue for abatement acid. (Figure 6) The shipment of 80 percent
acid instead of 98 percent acid increased the transportation and
handling costs by $1.00 per ton of acid which reduces the net
revenue in the same amount.
Another consideration in the study which has wide
implications involves the use of the abatement acid internally to
produce a more valuable phosphoric acid for fertilizer production.
Table 8 outlines the production costs for a phosphoric acid plant
vhich would utilize the abatement sulfuric acid production vhich
could be produced from both Colbert #5 (550 MW) and Widow's Creek
#7 (575 MW) or a total of 221,000 tons of sulfuric acid. A
phosphoric acid plant sized to use this amount of sulfuric acid
would produce about 7^250 tons of ^2^5 P61" year* This would allow
TVA to enjoy the savings incurred by producing its own ^2^5 in lieu
of purchasing PpO^ and at the same time enjoy an increase in net
sales revenue in the amount of 51 cents a ton for abatement sulfuric
acid.
The study indicates that the most orderly way to incorporate
the abatement acid into the market would be to replace sulfuric acid
currently produced from purchased sulfur. The potential sulfuric
acid production from the TVA system could be incorporated gradually
into the market as long as there was no significant competition from
other abatement sources. The resulting net sales revenue could
reduce the cost of operating the sulfur dioxide control system by an
estimated 10 to 20 percent.
Perhaps the most important result from the study is the
development of a practical versatile computer program which can be
used to extend the market investigation of abatement acid production
to the entire United States. Also, the initiation of a data file on
994
-------
VO
VO
Ul
SULFUR PRICE « $22.32
Table 7- MOST LIKELY MARKET PATTERN FOR TVA H2S0lf
(M TONS)
ACID CONCENTRATION = 98% CAPACITY =
MAXIMUM TVA ACID PRICE WOULD BE S 5.99
75X
BARGE * 80%
PLANT
LOCATION
I . HELENA,ARK.
2. N.LITTLE ROCK,AR
23. E.ST.LOUIS,ILL.
29. M3NSANTO,ILL
30. E.ST.LOUIS,ILL.
31. MARSEILLES,ILL.
32. CALUMET CITY,ILL
33. JOLIET,ILLINOIS
35. JOLIET,ILLINOIS
36. STREATOR,ILL.
37. E.CHICAGO,IND.
38. LASALLE,ILLINOIS
40. JOLIET,ILLINOIS
41. CALUMET CITY,ILL
42. CHICAGO HTS,ILL
46. BATON ROUGE,LA.
47. NErt ORLEANS,LA.
52. GEISMAR,LA.
54. HAMILTON,OHIO
55. CINCINNATI,OHIO
56. CINCINNATI,OHIO
57. COLUMBUS,OHIO
53. COLUMBUS,OHIO
59. COLUMBUS,OHIO
60. COLUMBUS,OHIO
61. COLUMBUS,OHIO
PRODUCTION ACTUAL
CAPACITY P
101
64
115
104
179
157
83
27
192
26
250
26
224
22
22
67
22
58
47
22
12
48
40
40
13
18
PLANT CAPACITY
PLANT PRODUCTION
MARGINAL ACID COST <$)
TOTAL PRODUCTION = 1982
AL
>'N
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
4
0
0
0
0
YEAR
BUILT
67
46
37
67
54
62
56
54
45
51
37
37
42
47
60
53
65
68
48
46
38
65
49
55
37
37
SULFUR
REDUC'N
(S)
7.52
15.19
14.64
8.37
9.13
1.62
7.91
23.92
4,53
22.74
0.
29.41
3.88
30.81
24.91
7.79
19.49
3.42
21 .22
39.16
59.82
0.
9.35
0.
37.49
28.09
COLB
101
0
0
0
0
0
0
0
0
0
0
0
0
0
0
21
0
0
0
0
0
0
0
0
0
0
122
122
STEAM PLANT SALES
CUMB GALL PARA SHArt
0
0
0
0
0
0
0
0
192
0
190
0
1 17
22
22
34
0
0
0
0
0
0
0
0
0
0
579
579
0
0
10
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
44
40
40
13
18
165
165
0
0
35
104
179
0
83
27
0
0
0
0
107
0
0
0
0
0
47
22
12
0
0
0
0
0
617
617
0
0
0
0
0
157
0
0
0
26
60
26
0
0
0
0
0
0
0
0
0
0
0
0
0
0
270
270
WIDC
0
0
70
0
0
0
0
0
0
0
0
0
0
0
0
0
22
0
0
0
0
0
0
0
0
0
93
93
JOHN
0
64
0
0
0
0
0
0
0
0
0
0
0
0
0
13
0
58
0
0
0
0
0
0
0
0
136
136
6.65 6.97 6.62 6.84 7.57 5.99 7.01
TOTAL NET SALES REVENUE « S 11872170
-------
35
VO
o:
UJ
o.
to
LJ
u
cr
a.
a:
u.
14 16 18 20 22 24
Figure 6. Effect of sulfur price on TVA net sales revenue.
-------
Table 8. PRODUCTION COSTS FOR PHOSPHORIC ACID PLANT
(225 tons/day)
Annual Operating Costs $/Ton P 0
Direct cost
Phosphate rock, 31.1$ PpOc (68$ BPL) 51.00
3.58 tons at $1^.25/ton
Sulfuric acid, transportation cost from 2.70
Colbert and Widows Creek, 2,7 tons at
$l/ton (truck rate)
Labor, 0.83 man-hr at $6.50 5.1*0
Maintenance, 6$ of plant cost 7.20
Electricity, 330 kWh, $0.006/kWh 1.98
Cooling water, 5.5 M gal at $0.02/M gal 0.11
Supplies, analysis, a.nd handling 2.20
Total direct cost 70.59
Indirect cost
Insurance and taxes, 2$ of plant cost 2.^0
Depreciation, 12 yr 10.00
Overhead, 100$ labor 5.^0
Interest, 7-1/2$ U.50
Total indirect cost 20.22
^^J3
Total production cost90.81
This IE equivalent to $0.91/unit of ^^ (unit = 20 Ib).
The net savings would be about $1.25 minus $0.91 equals
$0.3Vunit of P2°5 or
$34 per ton of PgO.
$2,524,000 per year
997
-------
sulfuric acid, sulfur sources, and end use patterns is in the
development stage. Both the data file and the computer program can
"be made available to other interested users.
PRELIMINARY FEASIBILITY STUDY OF CALCIUM- SULFUR SLUDGE UTILIZATION
IN THE WALLBOARD INDUSTRY
Several sulfur dioxide removal systems under development
are based on discarding the sulfur in the form of waste calcium
solids; however, there are many steam-electric generating plants,
especially in the East, which do not have the necessary land area
for such disposal. Therefore, an interest has developed toward
conversion of these "throwaway" materials to useful products.
One such possibility is the use of the calcium- sulfur
sludge to replace mined, native gypsum in wallboard manufacture.
In 1972, U.S. gypsum producers mined 12,328,000 tons while 7,718,000
tons were imported. Because of the low cost of water transportation
most imported material was shipped by water to users on the Atlantic,
Pacific, and Gulf Coasts. In the same year, 20,076,865 tons of
gypsum was used in the United States and 1^,205,069 tons of this, or
71 percent, was used in wallboard manufacture. Altogether, gypsum's
value as a material in agricultural and construction uses has
resulted in a plus $^00 million per year industry.
During the initial phases of investigation for this study,
it became apparent that the most likely markets for byproduct gypsum
are in the eastern United States. Most midwestern wallboard markets
can be supplied with native gypsum from nearby mines; however, a
large percentage of the eastern markets use imported gypsum from
Nova Scotia. Therefore, the study was directed toward this market
displacement opportunity.
In this feasibility study, six types of processes were
considered for byproduct gypsum production. Of the six types, those
which produce gypsum by oxidation of calcium sulfite sludge from
lime and limestone scrubbing or by neutralizing a weak sulfuric acid
scrubbing solution with lime or limestone appear to be the best
suited. In Japan, some of these processes are already being
utilized to supply gypsum requirements . One of these, the Japanese
Chiyoda "Thoroughbred 101" process, is now undergoing prototype
plant tests at a Gulf Power Company plant in Florida, and the gypsum
produced will be made into wallboard. Definitive test results are
expected by the latter part of
The lack of firm cost data makes it impossible to directly
compare the economics of gypsum-producing processes with other S0?
removal processes such as the lime or limestone throwaway schemes.
998
-------
However, a gypsum process has two major advantages over a lime or
limestone throwaway process.
1. Disposal costs for the calcium solids produced
in the throwaway processes are eliminated.
2. Sales revenue most likely can be obtained for
the byproduct gypsum.
In addition, in the eastern U.S. where the largest markets
are, byproduct gypsum could be shipped to many wallboard plant
locations cheaper than imported material. Domestic gypsum is not
competitive in this area because of its higher cost ($3-93/ton f.o.b.
vs. $2.38/ton f.o.b. in 1972) and the extremely high rail transporta-
tion costs from the Midwest mines to the East Coast ($22-$23/ton vs.
about $^/ton for water transport of imported material).
Producing byproduct gypsum instead of disposing of calcium
solids from the lime throwaway process results in a disposal cost
savings of approximately $i(-.50-$7• 00/ton of gypsum. It is probable
that the byproduct gypsum can be sold for at least $2.00/ton, $0.38/
ton cheaper than imported and $1.93/ton cheaper than domestic gypsum.
Byproduct gypsum can also be shipped from steam plants on the inland
waterways and along the East Coast to wallboard plants in the same
areas from $0.65 to $7-^5/ton cheaper than imported gypsum.
The potential economic advantage that byproduct gypsum
processes might have in comparison to throwaway systems would,
therefore, be the sum of the disposal co?t savings, sales revenue,
and any transportation credits less any differential in process
cost. Depending on the power plant location and processing costs,
this advantage over a throwaway process could range from nothing to
over $l6/ton of gypsum disposed. For a 500-MW power unit, this could
be as much as a $3-8 million per year reduction (^5$) in the cost of
operating a lime scrubbing - solids throwaway system.
This preliminary investigation indicates that production
of wallboard - quality byproduct gypsum from SCvj removal systems may
be an economically attractive route to waste solids disposal. It is
recommended that a more detailed market study of calcium-sulfur
sludge for wallboard be performed. To derive more accurate cost
data for comparison with throwaway alternatives, a thorough economic
evaluation of the more promising processes such as Chiyoda and
carbon absorption plus the oxidation step for conversion of calcium
sulfite to sulfate should be made.
999
-------
PHASE 2 OF THE TVA-EPA STUDY OF THE MARKETABILITY OF ABATEMENT
PRODUCTS
Phase 2 is titled, "The Potential Abatement Production
Marketing of Elemental Sulfur and/or Sulfuric Acid by the Electrical
Industry in the Eastern United States." The study started June 2k,
197^4, and will run through June 23, 1975- The basic objective is
to determine for specific power plant installations using flue gas
desulfurization system the potential net sales revenue for byproduct
elemental sulfur and sulfuric acid which can be realized from
marketing strategies covering the existing acid market, the existing
elemental sulfur market, and the growth markets for such commodities.
The study is designed to determine the quantities of by-
product sulfuric acid and/or elemental sulfur which could be produced
by air pollution abatement at installations for power plants located
in the states bordering the Mississippi River, its navigational
tributaries, the Great Lakes, and the East Coast. The study should
have general application for the geographic area located east of the
Rocky Mountains in the United States.
The depth of data gathering will be increased so that
important acid costing characteristics such as age, efficiency,
necessity for added pollution control equipment, sulfur receiving,
and storage facilities will be refined. Acid transportation methods
for each power plant and acid plant will be investigated and
accurate freight costs will be determined.
Logistical model requires the generation of hundreds of
thousands of freight rates for rail, truck, and barge transportation
from (l) existing sulfur sources through terminals to acid plants
and (2) from steam plants using flue gas desulfurization systems to
acid plants. The strategy is (l) to have TVA Navigation Economics
personnel generate a large sample of rates by hand as a check
reference (2) a computer system will then be developed to generate
those rates as a check and (3) to generate the remaining majority
within an acceptable margin of error. The system will be designed
such that rate generation for future studies, such as potential
relocation of the fertilizer industry adjacent to steam plant areas
can readily be made. The study will determine the incremental
production tonnages of sulfuric acid or elemental sulfur which can
be produced with expected variation during the year in order to meet
requirements such as AQCR regulations for specific plant abatement
production options. Abatement calculations will be based on the
tonnages and sulfur content of coal and/or oil which the electrical
industry would plan to use through I960 at the coal and oil-fired
power plants where sulfuric acid or elemental sulfur abatement
production is considered. A generalized approach will be used to
meet current local emission standards plus applicable case
1000
-------
variations based on screening criteria such as age of plant, amount
of fuel burned, sulfur content of fuel, plant factor, and site
limitations.
One of the more important concepts which will be encompassed in this
analysis will be that of consumer surplus. In looking, back at the
demand curve in Figure 5, "the consumer surplus would be measured
by integrating the entire area under the demand curve which lies
above the net sales revenue prices for TVA abatement acid. Each of
the acid producers which purchased TVA acid with the exception of
the "swing plant" identified in Table 6 would have paid a higher
price for the abatement acid. The model encourages the higher cost
inefficient producer to close down his acid plant and buy abatement
acid. The relatively more efficient producer continues to produce
acid from externally purchased sulfur. Thus, the model promotes
and enhances efficient production in the sulfuric acid market.
It is the opinion of the author that there is no one
abatement product which will have broad application to all power
plant installations in the country. Each utility may enjoy a unique
location advantage for marketing a specific abatement product.
Perhaps the utilities located in the high calcareous soil areas of
the Great Plains should be looking at ammonium sulfate as an abate-
ment product. Most of the soils in this area show moderate to
extreme sulfur deficiency. Recycling the abatement sulfur products
into the fertilizer industry through an established market (back
to the land) offers the greatest imminent social benefit to the
American consumer. Agriculture is by far the most competitive
sector of the American economy.
1001
-------
Unif;-fl \/.//:'v / 'nvirnnrrwntal Protection Agency
I'hjc f.V/v nt'fdjlfuri/ntion Symposium
Ai/Li:ii
-------
THE PRODUCTION AND MARKETING OF SULFURIC
ACID FROM THE MAGNESIUM OXIDE FLUE GAS
DESULFURIZATION PROCESS
BY
Irwin S. Zonis
Francis Olmsted
Dr. Karl A. Hoist
David M. Cunningham
Presented at the United States Environmental Protection
Agency Flue Gas Desulfurization Symposium
Atlanta, Georgia
November 4-7, 1974
Essex Chemical Corporation
1401 Broad Street
Clifton, New Jersey 07015
-------
TABLE OF CONTENTS
Page
Introduction 1009
Sulfuric Acid Operations with Feed
from an MgO Regeneration Process
1. Calciner Off-Gas 1011
2. Gas Pre-Treatment 1012
3. Acid Plant Changes - Mechanical 1014
4. Acid Plant Changes - Operational
and End Product 1018
5. Costs 1020
Economics of a Sulfuric Acid Plant Related
to an MgO F. G.D. System 102G
Example Economics 1021
Summary and Conclusion 1025
1007
-------
Introduction:
For several years within the chemical industry, many processes
have been under development designed to remove pollutant materials
from waste gas streams. One example is the Chemico-Basic Magnesium
Oxide Flue Gas Desulfurization Process (MgO F. G. D.), particularly
as it has been applied to fossil fuel burning electric power plants.
This process involves, not only the removal of sulfur oxide pollutants,
but also the regeneration of the pollution removing agent, together
with the production of a saleable by-product; commercial grade
sulfuric acid.
The process is capable of removing 90% of the sulfur dioxide
present in the flue-gases of boilers fueled with high-sulfur oil or
coal. As has been previously described, it consists of scrubbing
the flue-gases on their way to the stack with a slurry of magnesium
oxide which reacts with the sulfur dioxide to yield magnesium sulfite.
The magnesium sulfite slurry leaving the scrubbers is then partially
de-watered by centrifuging; the mother liquor being recycled to the
scrubbers while the solids are dried in a direct fired rotary drier to
yield dry, crystalline magnesium sulfite. This is stored in silos
pending shipment to a regeneration plant located in the immediate
vicinity of an existing or new sulfuric acid plant.
The regeneration plant consists essentially of a direct fired
calciner in which the magnesium sulfite is decomposed into magne-
sium oxide and sulfur dioxide gas. The magnesium oxide is returned
1009
-------
to storage silos at the boiler plant for reuse, while the sulfur
dioxide rich gas from the calciner is piped to the sulfuric acid
(1)
plant for use as its feed-stock in place of elemental sulfur.
Sulfuric acid has been manufactured at Rumford, Rhode Island since
1929. Essex Chemical Corporation has been manufacturing com-
mercial grade sulfuric acid at this facility since 1966 and has been
doing so from magnesium oxide regeneration offgas for the past two
years. Until July 1, 1974, the regeneration plant was operated
using magnesium sulfite from an MgO F. G. D. demonstration unit
located at Boston Edison Company's 150 MW oil fired Mystic Unit
No. 6. Following successful completion of this demonstration
project, the plant has been operating on magnesium sulfite from a
similar MgO scrubbing system on 100 MW of Potomac Electric
Company's coal fired Dickerson Unit No. 3.
It will be the purpose of this paper to discuss the differences in
sulfuric acid manufacture when fed from an MgO F. G. D. process
regeneration facility and when fed by an ordinary sulfur furnace,
based on the Essex experience. The discussion will begin with the
gas that leaves the regeneration facility and continues through to the
point in the sulfuric acid plant where there is virtually no difference
between a product stream having come from an MgO regeneration
facility or from a sulfur furnace. This will include general descrip-
tions of the processes involved, but since the technology of the
1010
-------
individual sub-processes is already well developed and in use in
the chemicals industry, extensive detail will be omitted.
.Later, some of the ecoaomic factors governing the use of MgO
process F. G, D. will be discussed.
Sulfuric Acid Operations With Feed from an MgO Regeneration Process
1. Calciner Off-Gas
After dust removal, the calciner off-gas is a gas stream at
approximately 100° F, saturated with water and containing 8 to
10% SO2, 4 to 5% D£ and the rest predominantly CO;? and N£?
gases that are inert as far as the acid plant is concerned. Table 1
compares typical calcine? oii-gas and t-ypical burner gas from a
conventional sulfur burner.
Table 1
Comparative Cornposition of Feed Gas to Sulfuric Acid Unit
Overall Composition, Mole
Type of Gas
Regeneration Facili
N2 COz Oz
ty 73 6 5
HZO SO2 Total
7 9 100
offgas
Sulfur Burner 79 - 12 - 9 100
offgas
The proportions of offgas components can be controlled well
within the parameters previously indicated. at the regeneration
facility,
Cooling, scrubbing to remove participates and SOy and con-
densing water out of the hot calciner exit gas is accomplished by
1011
-------
a "weak acid" scrubber and packed column after the calciner
and before the acid plant. At Rumford the liquid effluent from
this step is disposed of outside the plant. However, in a full
scale plant with MgO sulfur oxides-in-tail-gas control, this
stream could be recycled.
2. Gas Pre-Treatment
In the calciner, a low oxygen level is desirable in that it
prevents the oxidation of SC>2 to $03 (which would be lost in the
gas cleaning step just mentioned) and also prevents the oxidation
of MgSOo to MgSO^. Likewise it is important not to allow the
existence of conditions on the reducing side which would aid in
the formation of elemental sulfur or of soot, unburned fuel.
However, in the acid plant's "conversion" or catalytic oxidation
step, it is desirable to have a ratio of C>2 to SC>2 in excess of
1. 0 - at least 1. 2 to 1. It is thus necessary to provide additional
oxygen as air, to the gas stream before the conversion step.
Air is used to strip SO^ from liquid streams at two points
in the process; first at the weak-acid scrubber, and then at
the dry tower stripper. Additional air is fed into the gas stream
at the dry tower inlet, with quantity regulated manually by
monitoring SOo and O^ concentrations. At Rumford, this was
readily accomplished using conventional wet chemical methods:
The Reich and Or sat analysis. Obviously, electronic instrumentation
is applicable here.
1012
-------
Drying the cool gas is the next step, and conventional sulfuric
acid driers are used. At Rumford, the gas passes through a 93%
H-,SO4 drying tower that was previously used only for drying air
to the sulfur furnace. The dried gas can now serve as gas feed tc
the sulfur burner if it is desired to burn elemental sulfur as a
supplement to the SC>2 from the calciner offgas. This was a
standard practice for most of our operations at the Rumford plant.
Since the catalyst's ignition temperature is approximately
800 F, the gas must now be heated to this temperature. Several
sources of heat are available and, depending on the acid plant heat
balance and whether or not sulfur burning is also a source of SC>2,
one or more of these sources may be used. If the regeneration
facility is sufficiently close to the acid plant, it might be a source
of heat. Since calciner offgas must necessarily be cooled to
reduce its water content, it would be an advantage to reduce the
heat load on the weak-acid scrubber by transferring the heat to
the catalyst feed gas. Other possible heat sources include heat-
exchangers after the converters and, the heat generated if supple-
mental sulfur is being burned. By using one or more of these
sources and applying proper heat conservation techniques, adequate
heat can be obtained.
At Rumford, dried gas from the blower enters the sulfur
furnace, where additional sulfur may be burned. The gas then
picks up additional heat in the secondary converter's heat ex-
changer and in the primary converter's heat exchanger, before
passing to the top of the first catalyst bed. Temperature control
1013
-------
is achieved by bypassing gas around the burner, existing boiler,
and the heat exchanger, in a manner readily understood by
operators of existing Sulfuric acid plants.
After SO^ - (X ratio adjustment, drying and heating,
the gas from the regeneration-facility is essentially the same as
that from a sulfur burner. The only significant difference is a
higher CO? content in gas from the regeneration facility, and
possibly a slightly lower oxygen level.
A pictorial comparison of example methods for treating sulfur
burner offgas vs. regeneration facility offgas in a sulfuric acid
plant is shown in figures 1 (Dry Gas Plant) and 2 (Wet Gas Plant)
respectively (SO£ - O^ ratio control is not shown).
3. Acid Plant Changes - Mechanical
The major acid plant changes required are those needed to
achieve SG^ - C^ ratio adjustment, drying and heating of the gas
from the regeneration facility. SOo - O ratio adjustment may be
accomplished as simply as installing a "T" and valve on the
suction side on the acid plant feed fan. Drying may involve the
installation of a new vessel for 93% H^SO. drying of either
regeneration offgas or sulfur furnace feed gas, or both, if ex-
isting equipment cannot be refitted for this use. For heating,
the proximity of the regeneration facility to the acid plant becomes
important if the heat available in the regeneration facility is to be
used by the acid plant. For heat conservation, insulation integrity
1014
-------
98%
H2S04
BLOWER
ABSORBING
TOWER
SULFUR BURNER BOILER
NO. I
ECONOMIZER BOILER CONVERTER
NO. 2
FIGURE I. DRY GAS PLANT
1
PRE HEATER
93°/«
0/(CALCINER)
S02
IN
DRYING
TOWER
COLD
EXCHANGER
98%
ABSORBING
TOWER
INTER I HOT
EXCHANGER EXCHANGER
BLOWER CONVERTER
FIGURE 2. WET GAS PLANT
1015
-------
is a must as is the operation of process flow controls such as
dampers and condensate traps. For both drying and utilization of
available heat, a considerable amount of repiping was required at
Rumford, and will be required in any retrofit installation.
Process control changes are also required. The advent of the
additional heat balancing requirements makes process control more
complex. Where sulfur burning is used simultaneously with
regeneration offgas, blending the streams is also a process control
complexing factor. The needed controls are all standard and,
as such, there is nothing unproven required. However, there is no
question that the control of the efficiency of each subprocess is more
critical than if the sulfur burner is the only feed source to the acid
plant.
To meet environmental requirements, most acid plants have
recently had to exercise greater controls or install abatement
equipment to reduce emissions of sulfur oxides to the atmosphere.
The control techniques employed are generally designed to either
add conversion steps to the process or to absorb or scrub the
sulfur oxides out of stack gases. With a regeneration facility
in close proximity to the acid plant, it becomes quite economical
to install an MgO scrubber as the sulfur oxide emissions control
for a new or existing acid plant.
In addition to controlling emissions this increases the sulfur
efficiency of the plant. This can make an MgO scrubbing installation
1016
-------
for an acid plant being fed by a regeneration facility much more
of an economic advantage since it can potentially eliminate the
need for capital investment in sulfur dioxide conversion steps.
One other change must be considered and this relates to pro-
visions in the event of an unexpected shutdown of the acid plant.
In that event, the regeneration facility normally has considerable
surge in the gas stream which cannot be stored. Under these
conditions there must be provisions made in the acid plant to
take that gas. If an MgO scrubber is installed on the acid plant,
the scrubber and appropriate emergency by-pass controls could be
designed and installed to control and recover the gas surge from the
regeneration facility.
Obviously other currently available technology could be
employed to control stack exit SO£ concentrations: an alkali
scrubber, molecular sieve absorbers, or a double catalysis -
double absorption flow sheet could be used. At Rumford, SO_
concentrations from this small obsolete sulfur burning plant were
approximately 3, 500 ppm before the MgO regeneration facility
was installed. Because of the relatively short life of the project,
Chemico installed'a simple scrubber on the acid plant stack, using
caustic soda as scrubbing liquor to control SO2 discharge which
might be as high as 4, 500 ppm when the calciner was operating.
No attempt has been made to maximize the performance of this
small scrubber, and a local contractor is used to haul away
the spent alkali.
1017
-------
4. Acid Plant Changes - Operational and End Product
All operational changes occur in or before the converters. As
indicated before, where a regeneration facility and a sulfur furnace
are used simultaneously, the blending of these streams places
additional requirements on control of the heat, SO-> and O-> content
of the gases fed to the converters. This is a new type of operation
which required the development of new techniques at Rumford.
Also, where the regeneration facility, the sulfur furnace, and/or
the converters are used to heat the dried gas from the regeneration
facility, new start-up techniques may be required. At Rumford,
for both start-up and operation, these requirements amount to ad-
justing the ratio of air to regeneration facility offgas being fed
to the drying tower since this is the atmosphere in which any
supplemental sulfur burning is done. From the sulfur furnace
through the rest of the acid plant, the start-up techniques at
Rumford have not changed and the techniques used to maintain heat
balance in the plant after start-up are the same type as those
used in a conventional sulfur burning contact sulfuric acid plant.
The water balance in the plant also becomes more critical
with the additional water input from regeneration offgas, but again,
the techniques used for control are the same type as those used in
a conventional sulfur burning plant. Past the converters, the
operation is identical to that of a conventional plant.
Since the gas fed to the converters is virtually the same as
with a conventional plant it would be expected that the end product
1018
-------
would also be the same. This has been the case in Essex1 experience
when the regeneration facility is operating on feed from either an oil
fired or a coal fired power plant.
Sulfuric acid meeting Federal specifications for electrolyte acid
(50 ppm Fe maximum) can be routinely produced from MgO regenera-
tion offgas, in any desired concentration up to the limit of the plant's
capability, and consistent-with water balance. At Rumford, we
historically produced acid with Fe content in the 12 to 14 ppm range
during the winter months. In summer, with higher metal temperatures
exposed to our product, it was rare for iron content to be much above
25 ppm.
In the opinion of our plant management, acid quality has not
varied significantly with the advent of the regeneration facility;
iron content runs approximately 20 ppm. There has never been
any finding of magnesium compounds in the product acid.
Rumford sulfuric is all sold in merchant market, and none is
used captively. We sell product to manufacturers of detergents,
dye-stuffs, pharmaceuticals, and .tanning chemicals; to steel
companies for pickling; to utilities for boiler water treatment; to
water and paper companies for pH control; to manufacturers of
aluminum sulfate; to galvanizers; to a wider range of textile applica-
tions; to manufacturers of lead-acid storage batteries, etc., etc.
By and large, Essex sales in the New England market are at the
full published price, which is at this writing $45. 95 per ton, basis
1019
-------
100% HUSO.. Producing sulfuric acid from MgO regeneration
offgas has imposed no restrictions on Essex marketing of that
sulfuric acid.
5, Costs
Capital costs for modification of a new or existing sulfuric
acid plant for use of MgO regeneration offgas as feed to the plant
typically run less than 15% of the battery limits capital cost of the
plant. Operating cost additions are minimal and consist mostly of
power to transport the additional gas and liquid streams. No additional
manpower is necessary and no significantly different maintenance is
required.
Economics of a Sulfuric Acid Plant Related to an MgO F. G. D. System
It has been demonstrated that the construction and operation of
MgO F. G. D. scrubbers and regeneration facilities represents an
economically attractive alternative to burning low sulfur fuels in
many electric power utility applications. What has not been
previously discussed is the economics of the sulfuric acid plant
portion of the system and the relationship this has to the overall
economics of the system. This section will undertake that discussion
and present an example of the acid plant's operating economics.
The approach used here will be from that of a sulfuric acid
manufacturing company constructing and operating the plant, selling
its output in the conventional manner and deriving the profits that
1020
-------
would provide the motivation for such a company to enter into
this business. It is recognized that through capital or other means,
the utility or other parties might participate in the acid plant
economics. However, it is not expected that utilities will desire
to operate a chemical process facility. In the example to follow it
will be assumed that the applicable acid plant economics are those
of a chemicals manufacturing firm.
Example Economics
To determine the acid plant size required in this example, it
will be assumed that the plant is to match an F. G. D. system on a
1, 000 MW electric power plant operating at 75% load factor and
burning 3% sulfur oil. From Appendix I, 90% sulfur oxides recovery
will result in the production of 126, 710 short tons of sulfuric acid
per year. This would indicate that the design basis should be for
a 400 ton per day acid plant. The current estimated capital cost
for such a facility; including battery limits and off-site facilities,
design, engineering, purchase and erection (escalated through 1977}
and not including land, working capital, and financing costs, is
$6, 750, 000. For this example a 15% return on investment after taxes
will be considered sufficient justification for the construction of the plant.
The operating costs will be based on current costs escalated
through 1977 at 7% per year. Conventional financing will be assumed
for all of the capital cost at 12% interest per year.
1021
-------
While th<», price per ton of sulfuric acid at this writing has been
quoted as $45.95, a more conservative position will be used here.
For this example, it will be assumed that the market exhibits some
saturation above 100, 000 short tons per year in this way: The first
100, 000 tons will return $30. 00 per ton net back at the plant and all
above 100, 000 tons returns $5. 00 per ton. Under these conditions, for
the 126, 710 tons under discussion, the gross revenue would be $3, 133, 550.
The object of this example will be to determine the dollar value of
the sulfur oxides required to provide the economics indicated above.
Actual Capital Required
Capital (as stated above) $ 6,750,000
Land (5 acres at $50, 000 per acre) 250, 000
Working Capital 1, 000, 000
Interest During Construction (2 years at 12%) 1, 680, OOP
Total $ 9, 680t OOP
ROI Required
9,680,000 X 15% = $1,452, 000 Per Year After Taxes
Interest Charges
$6, 750, 000 financed at 12% interest for 30 years would require yearly
capital charge payments of $710, 190 of which an average of $485, 190
would be interest.
1022
-------
Operating Costs
It is assumed that the acid plant would use MgO stack gas scrubbing for
SC>2 emissions control.
Unit
Unit
Cost
Amount
Variable Costs
Cooling and
Cooling Water
Treatment
Produced
Per Ton $.08
of H?SO4
Produced
(126,710)
Sub Total
Non-Variable Costs
Direct Labor
Office & Supervision
Maintenance Labor
& Material
Laboratory and Supplies
Equipment Rental
Insurance and Taxes (2% of Capital)
Depreciation (10 Year)
GS&A (7% of Sales)
Sub Total
Total
13 People
5 People
5% of Initial Capital
Annual
Cost
Process Water
Electric Power
MgO
Catalyst
Gal.
KWH
Ton
Per Ton
of H2SO4
$.655/1,000
$.034
$123.00
$.1225
69,394 M
9,784 M
1,972
(126,710)
$ 45,453
332,656
242,556
15,522
10, 137
$ 646, 324
230,778
100,375
337,500
21, 875
4, 000
135,000
968, 000
219,344
$ 2,016,877
$ 2,663,201
1023
-------
Total Cost vs. Return
Under these conditions, the total annual cost to operate the acid plant
would be:
Operating Costs $ 2, 663, 201
Interest Charges 485, 190
Income Taxes 1,340,308
Return on Investment 1, 452, OOP
Total $ 5,940.699
To balance this, returns to the acid plant operation would be required
as follows:
Sulfuric Acid Sales $ 3,133,550
Support from F. G. D. User (The Utility) 2,807, 149
Total $ 5, 940, 699
From data in Appendix I, it can be calculated that support from
the F.G.D. user is equivalent to $.304 per barrel of fuel oil used
or 0. 44 mils per KWH of power generated by the F. G. D. user.
It must be emphasized that this is one very specific example
and that wide variations from this example would be common.
Varying acid market price, transportation costs, fuel costs, capital
costs, and method of financing could all significantly affect the
sulfuric acid plant economics. Also, capital participation on the
part of the F.G.D. user could considerably reduce the expense of
support requirements on the F.G.D. user by reducing the return on
investment requirements of the sulfuric acid manufacturer. Depending
1024
-------
on the degree of participation (for example, if funds were provided
for the sulfuric acid plant from the same source that they were provided
for the rest of the F. G. D. system, making the investment one complete
package), in the specific example herein stated the support requirements
could be reduced by more than 80% or down to less than $. 06 per
barrel or -0. 09 mils per KWH.
Likewise if the sulfuric acid plant being used in the F. G. D. system
is an existing plant, the only requirement for capital is for retrofit
which, as was mentioned before herein, is a considerably smaller
investment. Under these circumstances also, the return on investment
requirements (and thus F. G. D. user support) would be considerably
less and, with F. G. D. user capital participation, still less than what
was shown in the example given.
Summary and Conclusion
The technical feasibility of the use of SC>2 gas recovered from
MgO regeneration as the feed to a sulfuric acid plant has been
demonstrated at Essex1 50 ton per day sulfuric acid plant at Rumford.
The techniques and changes required to accomplish this have generally
been described here and, for two years electrolytic grade sulfuric acid
has been produced and sold from gas coming from an MgO regeneration
facility. Technical feasibility is established.
As also described herein, the economics of the MgO process will
vary with location. However, the most favorable applications are
those in urban locations, where offgas can replace sulfur in captive
1025
-------
sulfuric acid plants, where significant cost advantages accrue from
power plant use of high sulfur fuel, where freight rates from an MgO
source are not prohibitive, where the installation is new or retrofit
is not complex, and where local regulatory agency rulings permit
equitable distribution of F.G.D. costs to the rates and promote
reasonable capital financing. Where enough of these conditions exist,
the MgO F. G. D. system can be a low cost environmental advantage
to the public, the most appropriate and economic F.G.D. system for
the utility and a viable service business for the sulfuric acid
manufacturer. Essex Chemical Corporation plans to participate
fully with Chemical Construction Corporation and Basic Chemicals
Inc. , in the installation and operation of future electric power
plant MgO Process Flue Gas Desulfurization Systems.
1026
-------
Appendix I
Sulfuric Acid Plant Requirements
Based on Utility Capacity
Data
Station Rated Capacity 1, 000 MW
Station Heat Rate 9, 000 BTU/KWH
Fuel 3% Sulfur Oil
Load Factor 75%
Fuel Heat Content 18, 500 BTU/lb.
338.1 Ib. Oil = 1 Bbl. Oil
90% of sulfur oxides removed by F. G. D. system.
96% of recovered sulfur oxides converted to H^SO^..
Total BTUs Required Per Year
1,000, 000 KW X 8,760 Hr. X 9,000 BTU/KWH X 75% =
59.13 X 1012BTU/Yr
Total Oil Required Per Year
59. 13 x 10 BTU/Yr = 9, 453, 464 Bbl/Yr
18, 500 BTU/lb X 338. 1 Ib
Bbl
Total Sulfur Recovered Per Year
9,453,464 Bbl/Yr X 338. 1 Ib/Bbl X * l°*g ^ X .03 X.9 =
38, 526 Long Tons/Yr
Total Sulfuric Acid Generated Per Yeaj1
_0 ,_, Long Tons „ 3.426 Short Tons H2SO4 _
3o. O£.b - — - A - , . A .70 -
Yr Long Ton Sulfur
126,710 Short Tons/Yr
1027
-------
Appendix II
References
(1) McGlamery, G. G., "Magnesia Scrubbing" Design Branch, Tennessee
Valley Authority, Muscle Shaols, Alabama, Presented at the EPA
Flue Gas Desulfurization Symposium, May, 1973.
(2) Farkas, M. D., and Dukes, R. R., "Multiple Routes to Sulfuric Acid. "
Sulfur and SC>2 Developments, American Institute of Chemical Engineers,
New York, New York, (1971).
(3) McGlamery, G. G., Torstrick, R. L., Simpson, J. P., and Phillips,
J. F., Jr., "Conceptual Design and Cost Study, Sulfur Oxide Removal
from Power Plant Stack Gas - Magnesia Scrubbing - Regeneration:
Production of Sulfuric Acid. " Springfield, Virginia 22151;
National Technical Information Service. (May, 1973) Pages 100-130.
(4) Farmer, M. H., "Long Range Market Projections For By-Products of
Regenerable Flue Gas Desulfurization Process." Esso Research and
Engineering, Linden, New Jersey, Presented at the EPA Flue Gas
Desulfurization Symposium, May, 1973.
1028
-------
SECOND GENERATION PROCESSES FOR FLUE GAS DESULFURIZATION
INTRODUCTION AND OVERVIEW
A. V. Slack
SAS Corporation
Wilson Lake Shores
Sheffield, Alabama
Prepared for Presentation at the
EPA Flue Gas Desulfurization Symposium
Atlanta, Georgia
November if-7,
1029
-------
SECOND GENERATION PROCESSES FOR FLUE GAS DESULFURIZATION
INTRODUCTION AND OVERVIEW
by
A. V. Slack
SAS Corporation
Wilson Lake Shores
Sheffield, Alabama
The earlier papers in this symposium dealt with processes that
have been operated on a full scale. There is also a large group that are
not as far along and therefore can be said to be in the "second generation"
category. Examples of these can be found in each of the three major
process groups: (l) lime-limestone scrubbing, (2) two-stage (or indirect)
lime-limestone (clear liquor scrubbing with lime-limestone regeneration,
often called "double alkali"), and (j) recovery of the sulfur as a useful
product. The present paper introduces the "second generation" category and
attempts to analyze the merits and drawbacks of each of the processes
involved.
What is a second generation process? This is a good question,
because process status ranges all the way from systems that are fairly well
developed to nebulous departures that have not yet reached the laboratory
bench. There is no sharp dividing line at any stage that allows a good
classification between the first and second generation, or that qualifies
a process for the second generation category rather than perhaps "third
generation." In this paper a process must have been tested at least on a
pilot-plant scale to qualify. At the other end of the spectrum, methods
that have been tested on a full scale are generally excluded but a few of
those developed recently in Japan are included because they are new to the
U. S.
Lime-Limestone Scrubbing
Since absorption of SOo by a lime or limestone slurry is the oldest
and most used technology, most of the process variations can be regarded
as first generation. The few that seem'to fit into the second generation
category are listed in Table 1.
1031
-------
Fuji Kasei (japan)
The main feature of the Fuji Kasei process is a special scrubber
(called "Moretana"), which contains four perforated plates each with the
hole diameters and plate thickness tailored to the specific scrubbing
situation. The system has been installed on three industrial boilers. A
special oxidizer is used to convert the calcium sulfite to gypsum.
Fuji Kasei claims an exceptionally high mass transfer rate for the
Moretana scrubber. The system has made little headway, however, against the
lime-limestone system of Mitsubishi Heavy Industries. Although 1J Fuji
Kasei installations are said to be planned, they are all on small industrial
boilers or other small units whereas MHI has many systems installed or
underway in plants of the major Japanese utilities.
Joy Manufacturing
A wet precipitator placed on top of a spray scrubber removes mist
and fine particulate. A pilot plant is operating at Pennsylvania Power and
Light's Sunbury station.
Another feature of the process is a very high L/G that gives high
scrubbing efficiency and utilization even though limestone is the absorbent.
Mist elimination and particulate removal are also good; the pilot plant has
a clear stack.
The high efficiency of wet precipitators in removing fine
particulate ( < 2 U ) may give the Joy system added significance in the
future, especially if the absorber can remove the bulk of the dust without
fouling. No precipitator would then be necessary ahead of the scrubber and
the wet precipitator would give better overall efficiency than the usual
combination of dry precipitation and wet scrubbing.
Kellogg
Magnesium compounds are added to the lime-limestone slurry in
fairly large amounts to improve absorption efficiency and promote oxidation
to sulfate. Excellent results were obtained in pilot-plant tests. Since
much of the absorbed SOo is present as dissolved magnesium sulfite and
bisulfite, oxidation can be carried out by air sparging without the
expensive pH control necessary in Japanese processes for oxidizing calcium
sulfite.
If the process is shown to be effective on a larger scale, it
may be widely adopted. The gypsum produced should be a superior throwaway
material.
1032
-------
Mitsui Miike
This is another process for getting sulfite oxidation without
using sulfuric acid to lower the pH as is done in most Japanese lime-limestone
processes. The bleed stream from the scrubber, on its way to the oxidizer,
is contacted with about 25$ of the unscrubbed stack gas. The resulting
partial SO^ absorption reduces the pH enough to get the calcium sulfite in
solution and available for oxidation. The gas stream then passes to the
main scrubber to complete the SOo absorption.
An unidentified catalyst is used to increase both oxidation rate
and absorption efficiency. Pilot-plant data have indicated good efficiency
with limestone even in venturi absorbers.
A prototype is just starting up in Japan and construction of
units for three boilers is underway. Further adoption depends on the
balance between the relatively high investment and the reduced operating
cost. (See further discussions in a paper by J. Ando in this symposium.)
National Lime Association
A rotary drum fitted with hanging chains inside serves as the
scrubber. Several advantages are claimed, including (l) no possibility of
scaling, (2) low pumping head, (3) low pressure drop, (k) no lime slaking or
gas prehumidification needed, (5) no nozzle wear, and (6) no demister
plugging.
Four pilot plants have been operated. The data do not appear
sufficient for fully evaluating the process. It is reported that an
industrial boiler will be fitted with the system, in which case a good
evaluation probably can be obtained. The potential advantages make the
process very attractive if absorption efficiency is adequate and if dust
evolution is not a problem.
TVA
The use of a small amount of benzoic acid improved limestone
scrubbing efficiency in pilot-plant operation. A cost estimate made by
TVA, however, indicates that the overall economics may not be favorable
(see paper by G. G. McGlamery in this symposium.)
1033
-------
Eva1gation
The second generation approaches in lime-limestone scrubbing are
not likely to have as much impact as in the double alkali and recovery areas.
The first generation systems are becoming reasonably well developed—especially
those at Paddy's Run and Mohave which seem to be operating fairly reliably.
Nevertheless, the promise of lower cost and superior product may give the
new processes an important role.
Lirae-Limestone: Two-Stage Operation (Double Alkali Type)
The scaling and plugging problems encountered in lime-limestone
scrubbing have led many investigators to 3ook for a better throwaway
method. Most of the effort has been on the approach of using a clear
solution in the scrubber and reacting the scrubber effluent solution with
lime or limestone to precipitate calcium sulfite and sulfate, the same
products as in straight lime-limestone scrubbing. The absorbent is usually
a solution of ammonia or a sodium salt (whence the term "double alkali11)
but several other materials have also been used.
The process list in Table 2 includes the first generation
processes for perspective. About half of the methods are in the second
generation category, indicating the high degree of development activity
in the field0
Chiyoda (Combined NOK-SOo Process)
The standard Chiyoda process (catalyzed weak acid scrubbing
followed by reaction with limestone) has been combined with NOX removal by
adding ozone to the entering gas stream. One of the products is nitric
acid, which is collected in the scrubber. The method is considered to be
especially suitable for use in conjunction with the main Chiyoda process
because otherwise the nitric acid would cause undesirable sulfite oxidation,
whereas Chiyoda is a full-oxidation process (see paper in this symposium).
The growing emphasis on NOX emission reduction in Japan may bring
processes such as the Chiyoda NOX-SC>2 combination into use. The method
seems to have some advantage over those that remove NO and S0£ in separate,
unrelated operations. Further tests on a larger scale are necessary for
adequate evaluation.
1034
-------
Fuji Kasei (Japan)
This is also an SO -NO combination method, with sodium hydroxide
used as the SO absorbent ana lime as the regenerant. Chlorine dioxide is
added just before the scrubber to oxidize NO. It is claimed that the
oxidation is selective and that little ClOp is lost in oxidizing SOg.
Since some phases of the process have not been worked out,
evaluation is difficult. Five installations on industrial boilers are
being considered.
Kurabo (japan)
In the absorber loop, a solution of ammonium sulfate is circulated
through the scrubber and then through an air oxidizer. A bleed stream from
the oxidizer is treated with lime to precipitate calcium sulfate.
Kurabo has tested the process at the company's pilot-plant
installation near Osaka (japan) with good results. The ammonium sulfate
provides enough basicity to absorb S0? efficiently at a medium L/G (in the
50-80 gal/mcf range, depending on process requirements). Thus the process
is in between Chiyoda and double alkali processes such as Kureha-Kawasaki
and Showa Denko, in which NapSO, is the absorbent, lime or limestone is the
regenerant, and the calcium sulfite is oxidized in a system requiring
addition of sulfuric acid. Kurabo has a lower L/G than Chiyoda but retains
the Chiyoda advantage of oxidizing sulfite while it is still soluble rather
than after conversion to calcium sulfite.
The method is also superior to ammonia-based double alkali
processes (NKK, Kuhlmann) in that the regenerated ammonia is recycled as
aqueous ammonia rather than in the gaseous form, thereby avoiding the need
for an ammonia recovery scrubbing system. In addition, there is no fume
formation in the Kurabo process; although it may be possible to avoid
fuming in ammonia scrubbing methods, special care will be required and some
additional expense involved.
Kurabo is building full-scale systems, one of which is to start
up this year. The system is described more fully in a paper by J. Ando in
this sypmosium.
1035
-------
Dowa Mining (japan)
The SOp is absorbed in a solution of Alp (SO, )-z'AlpO giving
(SO, )2 which is then oxidized to A^so^Jh and rgenerated
with limestone.
The process has advantages similar to the Kurabo method—lower
L/G than for water or weak acid scrubbing and easier oxidation than for
sodium or ammonia-based double alkali processes. As compared to Kurabo,
the ability to use limestone rather than lime is a definite advantage.
Dowa Mining has installed the aluminum sulfate process on two
of its own sulfuric acid plants. These were started up quite recently;
operation so far seems to be satisfactory. (See also discussion in the
symposium paper by J. Ando. )
The Lurgi double alkali process is a NaOH-Ca(OH)p system much
like the General Motors double alkali process in the U. S. (see paper by
N. Kaplan in this symposium), except that a special scrubber type is used.
Pilot-plant work has been carried out but apparently no commercial units
have been installed.
Monsanto
A water solution of ethanolamine absorbs the SOp followed by
regeneration with lime. A 1-mw pilot plant has been operated and a 125-mw
design prepared. The process is described in detail and the advantages
presented in a later paper in this symposium.
Toyo Engineering
The system is of the ammonia-lime type. The main distinctive
feature appears to be use of a special crystallizer design in reacting the
Ca(OH) with the ammonium sulfites and sulfate.
The status of the Toyo process is not clear. No units have been
sold and the company does not seem to be developing the process further.
No evaluation can be made until further information is available.
1036
-------
Evaluation
The impact of the second generation processes in the two-stage
throwaway field is not very clear. While some of them have significant
potential advantage over the first generation group, the situation is
clouded by the fact that two-stage processes in general may not be able to
compete with direct lime-limestone scrubbing. The two-stage type was
developed primarily to avoid scaling and plugging in the scrubber, but
recent improvements seem to have essentially eliminated this problem in
direct scrubbing.
Processes based on sodium or ammonium compounds as absorbents
have the advantage, however, that a very high degree of SO^ absorption can
be attained, thus giving the two-stage type an advantage in situations where
high removal is required. If the cost can be reduced by second-generation
departures, and added advantages such as NO removal developed, the position
of two-stage throwaway may improve in the future.
Recovery
The sludge disposal problem in the foregoing processes is enough
in-itself to make recovery of the sulfur in a salable form an attractive
course. In addition, sale of the product offsets to some extent (usually
very small) the cost of the operation, and a natural resource is conserved
that otherwise would be wasted as sludge. The last of these is especially
important because it is likely in the future that sulfur recovered from
waste gases will be needed to bolster the dwindling supply from other
sources.
Processes in the recovery category outnumber the throwaway type
by a considerable margin; a hundred or more can be counted if those
patented but relatively untested are included. The list in Table J includes
only the more significant ones. Again, the first generation methods are
listed to give perspective.
Development work on recovery processes has followed four major
lines of attack.
1. "Gathering" the S09 in a concentrating operation to get a stream
concentrated enough for economical conversion to sulfur or sulfuric
acid. Sulfur is the preferred product because storage is simpler
and shipping cost lower.
1037
-------
2. In situ oxidation or reduction of the SOp while still in the main
gas stream. Because of the dilute concentration, this has been a
difficult technology to develop.
3. Collection of the SOn as an alkali, sulfite, followed by reduction
in a high temperature process, separation of HpS in an intermediate
step, and conversion to a sulfur by the Glaus or other standard
process.
4. Reduction of alkali sulfite at low temperature while still in
solution. Elemental sulfur may be produced directly in the
solution, as in the citrate and phosphate methods, or BUS can be
evolved as an intermediate. Such processes generally depend on
formation of thiosulfate in the solution since it appears that
thiosulfate is easier to reduce to sulfur or hydrogen sulfide than
are most other sulfur compounds.
The first-generation recovery processes have several problems
that are either not encountered or are less troublesome in throwaway
methods.
1. Dust in the gas, even though mainly removed in a preceding
precipitator, can accumulate in closed-loop scrubbing and regenera-
tion systems, and interfere with operation. And if a catalyst is
used, blocking of the pores can be a problem.
2. Processes are relatively complicated and expensive because of the
regeneration requirement.
3. Energy requirement in the regeneration step can be quite high.
k. If the reducing agent is natural gas, it may well be not only
expensive but also unavailable. (EPA has a program underway on
developing ways to use less expensive reducing agents.)
5. If the reducing agent is hydrogen sulfide, it will have to be made
in most cases by reducing two-thirds of the product sulfur--which
is expensive and involves handling large quantities of a relatively
hazardous chemical.
6. Unless the process can reduce sulfate as well as sulfite, the
high solubility of the sulfate will cause a water pollution problem.
7- The product must be moved out to consumers on a fairly steady basis
since it cannot be discarded unless special provisions are made
(neutralization of acid or landfill of sulfur). This is much more
of a problem with acid than with sulfur because the latter can be
stored at relatively low cost.
1038
-------
On the other hand, recovery has an advantage over the lime-
limestone processes in that there is the opportunity for operating in the
dry state and thus avoiding the need for reheating the gas.
The ability to avoid one or more of the problems unique to recovery
is a major advantage for any recovery process. The main effort in second-
generation process development has been to avoid as many of them as possible
and thus to improve over the first generation type.
Gathering Processes
Methods that concentrate the SOo as a first step include both
alkali solution scrubbing and absorption in solids at elevated temperature.
Alkali Scrubbing: The Stone and Webster-Ionics process is based
on scrubbing with Na^SO-2 solution. The Na2SOv-NaHSO^ from the scrubber is
treated with a Na2SO^-HpSCK solution equivalent to sodium bisulfate to evolve
a rich stream of SOg and the coproduct Na£SO, electrolyzed to regenerate
the bisulfate. Power consumption is high but the cost can be minimized
by using off-peak power. Excess sodium sulfate caused by oxidation must
either be discarded or electrolyzed to give weak sulfuric acid, which is
probably usable if tUSO, is the main product. Further evaluation of the
process is given in another paper in this symposium.
In the ammonia scrubbing process being developed by EPA-TVA,
the ammonium sulfites produced in the scrubber are decomposed in -a way
similar to S and W-Ionics--by treating with ammonium bisulfate. The
bisulfate is regenerated thermally rather than by electrolysis. Excess
sulfate must be separated but since it is ammonium sulfate it can be used
as a fertilizer. This process is also evaluated further in another
symposium paper.
The IFP ammonia scrubbing method involves evaporation of the
scrubber solution to give a gaseous mixture of SC>2, NH and 1^0, which is
then mixed with E^S and passed through the IFP "wet Claus" reactor.
Elemental sulfur is formed directly and the ammonia, which is not affected,
is returned to the scrubber. A separate step is provided for reducing
excess sulfate.
The IFP method is being tested, both in prototype and full-scale
units, in France and Japan. None of the installations has been operating
long enough to provide any significant data. Like the EPA-TVA process, there
is a fume problem in the scrubbing step. Progress has been made in
preventing the fume but further demonstration is needed.
1039
-------
Absorption in Solids: The Esso-B and W method is reported to
be of the solid absorbent type. Pilot-plant work has been underway for
some time but no data have been reported. The dry operation avoids the
corrosion and other problems encountered in wet scrubbing and no reheat
is necessary. The equipment is large because of the high operating
temperatures, however, and the absorbent is subject to plugging by dust.
The Bergbauforschung carbon process is also of the gathering
type. The sorbed SOp, converted unavoidably to H^SOj, in the carbon pores,
is evolved as a concentrated stream by heating the loaded carbon to a
temperature high enough for reduction of the acid by part of the carbon.
The method will be tested in a prototype at Gulf Power's Scholz station.
Although the Bergbau process involves handling large tonnages
of moving solids, the system is dry, can be added at the end of the power
plant train, and removes dust from the gas without any adverse effects on
regeneration.
All the gathering processes require a second operation to convert
the SOp to a useful product. Making sulfuric acid is the simplest approach
but several methods for making sulfur are being explored. The first
generation method (but still not demonstrated for power plant operation)
is reduction by natural gas. Foster-Wheeler is working on a second-
generation process in which the reduction is accomplished by direct
reaction with anthracite coal. This will be tested in conjunction with
the Bergbauforschung test project.
Oxidation in Situ
The Westvaco carbon method appears to fit best into this category.
The sulfuric acid formed in the carbon pores by SOp oxidation is converted
directly to elemental sulfur by reduction with H S, after which the carbon
is heated to volatilize the sulfur.
The Westvaco method is dry, is not affected by dust, uses coal
(converted to producer gas) as the reductant, and has no sulfate problem.
The main question may be the carbon attrition in the fluidized bed adsorber.
Further evaluation is given in another paper in this symposium.
1040
-------
High Temperature Reduction
The main example is the Atomics International method in which
the SO is absorbed by a molten mixture of alkali carbonates and the
resulting sulfites reduced to sulfide by reaction with petroleum coke at
high temperature. The sulfides are then dissolved in water and COg added
to evolve H S which is converted to sulfur by standard procedures. Tests
of the method in a 10-mw system at Consolidated Edison's Arthur Kill station
have encountered severe corrosion problems.
The basic reduction step in the molten salt process has been
used by AI in another system (aqueous carbonate) that operates at lower
temperature and therefore can be added at the end of the power plant
train. Sodium carbonate solution is sprayed into a spray-dryer type of
absorber in which the solution absorbs SOp and is evaporated to dryness.
The solid sodium sulfite is reduced with coke and H^S produced as in the
molten salt process.
The water evaporation cools the gas somewhat but not as low as
a wet scrubber; reheat probably would not be required. Petroleum coke is
an economical reducing agent, all the sulfate is reduced by the coke, and
sulfur is the product. Thus the process avoids the problems mentioned
earlier; on the other hand, a particulate collection system must be
installed after the scrubber to recover the sodium sulfite.
The spray drying and reduction steps have been tested separately
in pilot plants and the HgS generation is a common process in the paper
industry. An integrated test program is needed.
In Japan, Tsukishima Kikai (TSK) has adapted the Billerud process
to SOp recovery from stack gases. The Billerud is a method used in the
paper industry to recover sulfur compounds from waste lignin-bearing liquors.
The liquor is sprayed into a furnace fired with heavy oil under reducing
conditions. The organic compounds in the liquor reduce sodium sulfite and
sulfate to sodium carbonate (solid) and HpS. The sodium carbonate is
separated, along with carbon formed because the furnace is operated at a
temperature low enough to prevent melting of the sodium carbonate, and the
H S converted to compounds for recycling.
In adapting the process to stack gas cleaning, TSK operates much
like the Billerud method. Stack gas is scrubbed with recycled sodium
carbonate solution and the resulting sulfite-sulfate solution sprayed into
the top of a standard Texaco or Shell oil gasification unit. -The solids-
laden gas passes through a waste heat boiler, a cyclone-scrubber combination
to separate the solids, and a Stretford unit to convert HgS to elemental
sulfur.
1041
-------
The method uses a low-cost reducing agent, has a relatively low
regeneration energy requirement, and produces sulfur. The main problems
are the .dilute concentration of H S in the gas and difficutly in separating
carbon. Pilot-plant tests are underway.
Reduction in Solution
The citrate process has been developed mainly by the U. S. Bureau
of Mines. Pfizer also is doing development work on the method. The S02
is absorbed in sodium citrate solution and the scrubber effluent treated
with HpS to give elemental sulfur directly. Although the chemistry is not
entirely clear, thiosulfate is formed during regeneration and probably is
the material actually reduced to sulfur. The citrate buffers the system and
keeps it in a desirable pH range.
In addition to work in the Salt Lake City laboratories, the Bureau
is operating a pilot plant (about 0.5-mw) at the Bunker Hill smelter in
Kellogg, Idaho. A unit for producing HpS from product sulfur is included.
Pfizer has a pilot plant at Terre Haute, Indiana, quite similar to the
Bunker Hill pilot except that a different flotation system is used to
separate the finely divided sulfur from the citrate solution and no HpS
production unit is included. Both pilot plants seem to be giving gooa
results; the Bunker Hill operation is described in one of the following
papers in this symposium.
Stauffer has a variation of this type of process, with phosphate
used as the buffering agent. A small pilot plant (about 300 cfm) is being
operated at Northeast Utilities' Norwalk Harbor station. This operation is
also much like that at Bunker Hill excpet for a different flotation system.
Results are reported to be good.
All three systems produce sulfur and have a relatively low energy
requirement. The drawbacks are very dilute scrubber solution (with
consequent large regeneration system size), use of H?S as the regenerant,
natural gas requirement to make the HpS, and a purge stream of sodium
sulfate (although not as much as in the Wellman-Lord process).
The Consolidation Coal method has been disclosed only in patents,
which indicate that the scrubbing solution contains potassium compounds such
as KgCO,, KOOCH, KSH, and K^S; that potassium thiosulfate (Kp_S 0,) is a
major product; and that the KpSoO, is reduced to HpS by use of CO. The
chemistry is complex and many side reactions occur. The scrubbing step
has been tested on a 10-mw scale and regeneration in a small pilot plant
(less than 0.5-mw).
1042
-------
UOP's Sulfoxel process has also been described only in patents,
These indicate scrubbing with an alkali such as Na CO,, treating the resulting
sulfites with H S to form thiosulfate, reducing the thiosulfate to sulfide
by use of CO, and carbonation of the sulfide to evolve H S and regenerate
NagCO,. Part of the H S is recycled and part converted co sulfur as the
product. A 22-mw test system installed at Commonwealth Edison's State Line
station was shut down recently and tests begun on a smaller scale.
Evaluation
From this it is clear that the second-generation recovery methods
avoid some of the problems that are drawbacks to the first generation.
Some of the processes are more successful than others in this respect.
In regard to plant cost, it is difficult to make an evaluation
because none have been developed far enough for definitve cost estimates. The
carbon processes have the drawback of very large absorbers, and the regenera-
tion solutions in the citrate and phosphate methods are so dilute that
large equipment will be required. The ammonia systems are superior in this
respect but extra capital may be required for fume control. The TSK
is quite straightforward but, again, the H?S stream is dilute. The
electrolytic cells in S and W-Ionics and tne spray drier in AI's aqueous
carbonate are expensive pieces of equipment. Both Consolidation Coal and
UOP appear to be relatively complex processes.
Energy requirement for regeneration appears to be relatively low
for AI's aqueous carbonate, citrate-phosphate, TSK, and Westvaco. Moreover,
AI, Westvaco, Bergbau, and Esso-B and W save energy by not requiring reheat.
In regard to type of reducing agent, the ability to use petroleum
coke by AI and heavy oil by TSK is a major advantage. Foster-Wheeler, which
has licensed Bergbau, also is developing an SOg reduction method based on
coal. Methods such as BM, Pfizer, Stauffer, and Westvaco, that use HpS as
a reducing agent, are under a burden because of the cost and care in Handling
required.
The water pollution problem from sulfate varies. Several
processes (AI, Bergbau, Esso-B and W, IFP, TSK, and Westvaco) have sulfate
reduction built into the system. Methods that have thiosulfate in the
scrubber circuit (Bureau of Mines, Pfizer, Stauffer, Consol Coal, UOP) have
a relatively minor sulfate problem because the thiosulfate inhibits
oxidation. The EPA-TVA ammonia process has full oxidation but ammonium
sulfate is an acceptable co-product. S and W-Ionics is probably the most
vulnerable on this point.
1043
-------
The processes with the most handicap in making sulfur appear to
be Bergbau and TSK--because the SOg or HpS is diluted by other gases.
All the second generation methods have the advantage that they
are capable of producing elemental sulfur, whereas seven of the 11 first
generation processes listed in Table J either cannot make sulfur or produce
such a dilute SCU stream that it would be expensive to do so.
At the present time there is not enough information for a
definitive comparison among the second-generation recovery methods. Each
has some good points and some drawbacks; the balance can be determined only
fay further testing. All of them deserve the further support necessary to
bring them to the point that an adequate evaluation can be made.
1044
-------
TABLE 1
Second Generation Lime-Limestone Processes
Developer
Absorbent
Fuji Kasei (japan) Lime
Joy Manufacturing Lime-limestone
Kellogg Limestone
Mitsui Miike (japan) Limestone
National Lime
Association
TVA
Lime
Limestone
Principal
distinctive feature
Scrubber type
Use of electrostatic
precipitator
Special additive
Oxidation system
Rotary drum scrubber
Development status
Full-scale operating
Pilot
Pilot
Full-scale under
construction
Pilots
Additive (benzole acid) Pilot
1045
-------
TABLE 2
Status of Two-Stage Lime-Limestone Processes (Double Alkali)
Developer
First generation
A. D. Little
Chiyoda
Envirotech
General Motors
Kureha -Kawasaki
NICK (Japan)
Showa Denko
FMC
Second generation
Chiyoda (NC^-SOg
combination)
Fuji Kasei
Kurabo (Japan)
Dowa Mining
Lurgi
Monsanto
Toyo Engineering
Absorbent
NaOH
Weak HgSO,
NaOH
NaOH
Na2SO,
NH^OH
Na2S05
NaOH
Weak H2SO, -HNO,3
NaOH3
(mk)sok
A12(S04)3
NaOH
Ethanolamine
NH^OH
Precipitant
Ca(OH)2
CaCO,
Ca(OH)2
Ca
-------
TABLE
Status or Recovery Processes
Developer
First generation
Chernico-Basic
Grille
Hitachi
Lurgi
Mitsubishi Heavy
Industries
Monsanto
NIIOGAZ
Shell
Sumitomo
United Engineers
Wellman-Lord
Absorbent
MgO
MgO
C
C
MnOg
Water
(NH^)2S05,
CuO
C
MgO
NagSO,
Regeneration
method Intermediate
Thermal SOg
Thermal S02
None None
None None
Reaction Mn2(SOj^),
with NH ^
None None
Thermal SOg
Reduction (Hg) SOp
Thermal SO
Thermal SOg
Thermal SOg
Product
HpSO^
or S
or S
Dilute
Dilute
(NHjt)2SO
Coned.
S or
S or
H2SOU
S or
2
S or
H SO,
Development
status
Full-scale
Prototype
Full-scale
Full-scale
. Full-scale
abandoned
Full-scale
Full-scale
Full-scale
Full-scale
Full-scale
under
construction
Full-scale
1047
-------
TABLE 3 (Cont'd. )
Regeneration
Development
Absorbent
Intermediate Product
Second generation
Atomics International
1 1
Bergbauforschung
Bureau of Mines
Consolidation Coal
EPA-TVA
Esso-B and W
IFF
Pfizer
Stauffer
S and W-Ionics
TSK (Japan)
UOP
Westvaco
NagCO
Mixed
carbonates
C
Sodium
citrate
KgCO,
(plus
formate)
NHjjpH
CuO
NH.OH
Sodium
citrate
Sodium
phosphate
NaOH
NagCO,
Na2C05
C
Reduction
(coke)
Reduction
(coke)
Thermal
Reduction
(HgS)
Reduction
(CO)
Acidulation
Reduction
Thermal
Reduction
(HgS)
Reduction
(HgS)
Electrolysis
Reduction
(heavy oil)
Reduction
(CO)
Reduction
(H S)
2
HgS
HgS
so2
None
HgS
so2
so2
SOg
None
None
SOg
HgS
H2S
None
S
S
S or
HgSO^
S
S
S or
HgSO^
S or
HgS04
S
S
S
S or
V°4
S
S
S
Pilot
Prototype
Prototype
Pilot
Pilot
Pilot
Pilot
Pilot
Pilot
Pilot
Pilot
Pilot
Prototype
Pilot
1048
-------
PILOT PLANT TESTING OF THE CITRATE PROCESS
FOR S02 EMISSION CONTROL
W. A. McKinney, W. I. Nissen
D. A. Elkins, and J. B. Rosenbaum
Bureau of Mines
Salt Lake City Metallurgy Research Center
Salt Lake City, Utah
ABSTRACT
The Bureau of Mines citrate process for removing S02 from industrial
wast; gases comprises absorption of S02 in a solution of sodium
citrate, citric acid, and sodium thiosulfate followed by reacting the
absorbed 302 with H2S to precipitate elemental sulfur and regenerate
the citrate solution for recycling. Two pilot plants operated'to
assess the feasibility of the citrate process for S02 emission control
have produced tail gas containing less than 50 PPm S02 from stack gas
containing 1,000 to 5,000 ppm S02. A pilot plant constucted by the
Bureau of Mines and operated jointly by the Bureau and The Bunker Hill
Co. at a lead smelter in Kellogg, Idaho, is described in detail.
Nominal capacity of the plant is 1,000 scfm of 0.5-percent S02 gas
yielding 1/3 ton sulfur per day. General features and operating re-
sults are summarized for a pilot plant designed, assembled, and
operated by Pfizer, Inc., Arthur G. McKee and Co., and Peabody
Engineering to treat 2,000 scfm of 0.1- to 0.2-percent S02 gas from
a coal-fired steam generating station at Pfizer's Vigo Chemical plant
in Terre Haute, Ind. Process economics for coal-fired boiler flue gas
desulfurization are projected on the basis of the pilot plant
operations.
1049
-------
PILOT PLANT TESTING OF THE CITRATE PROCESS
FOR S02 EMISSION CONTROL
INTRODUCTION
Research on techniques for removing S02 from waste gases was initiated
at the Bureau of Mines Salt Lake City Metallurgy Research Center in
1968, Pioneering research had indicated that effective removal of S02
and recovery of sulfur from waste gases might be achieved by absorbing
the S02 in a suitable solution and then reacting the absorbed S02 with
gaseous H2S to precipitate sulfur and regenerate the solution for re-
cycling. After a year of screening many possible reagent combinations
of inorganic and organic solutions, we established that a solution of
citric acid and sodium citrate was a very effective absorbent for S02
and had most of tfca desirable characteristics that we had been search-
ing for. Among the factors affecting the choice of citrate were
chemical stability, low vapor pressure, adequate pH buffering capacity,
and the purity ar:d physical character of the precipitated sulfur.
As first studied in the laboratory, the process comprised absorbing
S02 in citrate solution, reacting the absorbed S02 with bottled H2S,
filtering and melting the precipitated sulfur, and recycling the re-
generated citrate solution to the S02 absorption step. Water produced
by the sulfur precipitation reaction was evaporated by the hot gas in
the absorption column. Preparation of H2S from recycle sulfur by
reacting -lulfiir vapor with natural gas and steam over an alumina
catalyst was examined separately.
Subsequently a pilot plant to process up to 300 cfm of reverberatory
furnace gas was placed into operation in November 1970, jointly by the
Eurit>.ii of Mines and Magma Copper Co, at the San Manuel smelter in
Arizc;-.ia, Sulfur conversion to H2S was omitted as being premature.
Owing in part to hasty procurement and assembly, intermittent operation
of the pilot plant over a 6-month period was troubled by failures of
the gas cleaning system, pump breakdowns, and plugging of flow lines
by precipitated and melted sulfur. Useful data on consumption of
citric acid and other reagents were not obtained, but the S02 absorp-
tion and regeneration system proved readily manageable for removal of
93 to 99 percent of the S02 from the smelter gas. Findings of the
initial laboratory and pilot plant research were reported in 1970 and
1971 (2, 5).i/
The preliminary Bureau of Mines laboratory and pilot plant research
demonstrated that the citrate process is capable of substantially
complete removal of S02 from industrial waste gases. No scaling
problems are encountered in the absorber because there are no solids
Underlined numbers in parentheses refer to items in the list of
references at the end of this report.
1050
-------
in the clear citrate absorbent liquor. Most of the S02 is converted
to sulfur.with only about 1 percent converted to sulfate regardless
of the SO2 and oxygen content of the feed gas. The process produces
an end product of elemental sulfur that can be marketed or readily
stored with a minimum of environmental disturbance.
After the encouraging preliminary results, two pilot plant investiga-
tions were undertaken to further test the process and obtain useful
data for engineering evaluation and cost estimates. This report
describes the Bureau of Mines pilot plant operation at The Bunker Hill
Co. lead smelter in Kellogg, Idaho, and briefly summarizes the Pfizer-
McKee-Peabody pilot plant operation treating stack gas from a
coal-fired steam-generating station in Terre Haute, Ind, Estimates of
capital and operating costs for removing S02 from a 1,000-MW power-
plant burning high-sulfur coal also are presented.
PROCESS DESCRIPTION
As a result of further process development in the laboratory leading
to an improved method for separating the sulfur from the dilute
citrate-sulfur slurry, the citrate process as now envisioned and shown
in figure 1 comprises the following steps:
1. The S02-bearing gas is cooled to between ^5° and 65° C
(11J° and 1^9° F) and cleaned of H2S04 mist and solid
particles.
2. The S02 is absorbed from the cooled and cleaned gas by
a solution of sodium citrate, citric acid, and sodium
thiosulfate.
3. Absorbed S02 is reacted with added H2S at about 65° C
(1^9° F) and atmospheric pressure, precipitating
elemental sulfur, and regenerating the solution for
recycle.
IK Sulfur is separated from the solution by oil flotation
and melting.
5. The H2S for step 3, if not otherwise available, is made
by reacting two-thirds of the recovered sulfur with
natural gas and steam.
CHEJ^STRY OF THE PROCESS
Absorption of S02 in aqueous solution is pH-dependent, increasing at
higher pH. Because dissolution of S02 forms bisulfite ion with
resultant decrease in pH by the following reaction,
1051
-------
GAS CLEANING
AND
COOLING
._ .._
/^ j^j y^
M
o
ui
Flue
gos-
' cooled gas
HoO—•>
SO2 ABSORPTION I SULFUR PRECIPITATION |
I AND !
(SOLUTION REGENERATION
To atmosphere
SULFUR SEPARATION
H2S GENERATION
Steam
CH4
FIGURE I.-Generalized citrate process flowsheet.
D-26I4-SL
-------
S02 t H20 * HSOg <- H, (1)
the absorption of S02 in aqueous solution is self -limiting. However,
by incorporating a buffering agent in the solution to inhibit pH drop,
high-S02 loadings and substantially complete S02 removal from waste
gases can be attained. The principal function of the citrate or other_
carboxylates we have tested is to serve as a buffering agent during S02
absorption.
The chemistry is complex for the production of sulfur and regeneration
of absorbent by reacting H2S with the S02 in the aqueous solution, but
the overall reaction is as follows:
S02 -i- 2H2S - 5S° + 2H20. (2)
Actually, thiosulfate and polythionates are found in solution at
equilibrium concentrations after several S02-absorption and H2S-
regeneration cycles. Oxidation of S02 in the aqueous solution is
sharply depressed by complexing of HSOs and H4" from reaction 1 by the
thiosulfate ion, according to the following reaction:
H4" •*- HSOg -t- SaOg5* ~ (S02-S203)= * H20. (3)
Reaction J shows how this complex might react with H2S to form elemental
sulfur and thiosulfate ion.
(S02-S203)= + 2H2S s 3S° * 2H20 + S203~. (U)
To assure satisfactory operation of the system on startup, sodium
thiosulfate is added to the initial absorbing solution.
Hydrogen sulfide for regeneration of the absorbent and precipitation of
elemental sulfur can be produced by reacting sulfur with methane and
steam as shown in reaction h.
CH4 + kS * 2H20 -. C02 + UH2S. (5)
Other reducing gases such as hydrogen and carbon monoxide can be used
in place of methane. More detailed information on the chemistry of the
citrate process is provided in a Bureau of Mines publication (6) and a
paper presented at the American Chemical Society National Meeting in
April of this year (J ) .
THE BUNKER HILL CITRATE PILOT PLANT
Nominal capacity of the Bunker Hill pilot plant is 1,000 scfm of 0.5
percent SO 2 gas yielding about 1/3 ton of sulfur per day. The plant
is to be operated in three phases. Because persistent mechanical
1053
-------
failures of the gas cleaning system at the San Manuel copper smelter
pilot plant were a principal cause of intermittent operation. Phase I
of the Bunker Hill plant operation is designed to treat cleaned, k- to
5-percent S02 gas diverted from the Lurgi sintering furnace feed to the
lead smelter acid plant and diluted with air to 0.5 percent S02.
Commercially produced hydrogen sulfide from a tank trailer is used for
the sulfur precipitation reaction. In the Phase II operation, an H2S
generation plant producing ?6 to ?8 percent H2S gas by reacting product
sulfur with natural gas and steam will replace the nearly pure H2S
from the tank trailer. In Phase III, the lead smelter sinter plant
tail gas, which contains dust, acid mist, and from 0.3 to 0.9 percent
S02, will be used as pilot plant feed. This sinter tail gas presently
passes through a baghouse and then is discharged to the atmosphere
through Bunker Hill's main stack. In the Phase III operation, in line
with conventional lead smelter practice, most of the valuable dust
from the tail gas will be recovered in a baghouse, the gas will be
cooled in a packed scrubber, and H2S04 mist and traces of particulate
matter will be removed by a wet electrostatic precipitator. One of
the goals in the test plant operation is to determine the minimum gas
cleaning requirement compatible with the citrate process. A block
diagram of the complete Bunker Hill pilot plant is shown in figure 2.
Contracts for the design fabrication, and installation of the Phase I,
II, and III plants were awarded to Morrison-Knudsen Co., of Boise,
Idaho. Specifications and preliminary design information were
provided by the Bureau of Mines. Thio-Pet Chemicals, Ltd., of Calgary,
Alberta, Canada, a company with industrial experience in production of
hydrogen sulfide and carbon disulfide by the sulfur-methane-steam
reaction, acted as subcontractor to Morrison-Knudsen and provided the
design for the H2S generation plant. Construction on the Phase I S02
absorption and sulfur recovery plant was completed in early January of
this year. After shakedown runs, modifications, and acceptance test-
ing, plant operation began in mid-February. In late September,
construction was completed on the H2S generation plant and pretesting
began. Construction is underway on the gas cooling and cleaning •
plant. This plant is scheduled for completion on January 7, 1975.
This report covers the Phase I operation only.
Phase I Plant Description
Figure 3 shows the flowsheet for this phase of the operation. Strong
gas containing about k.5 percent S02 from the Bunker Hill Lurgi up-
draft lead sintering furnace passes through a baghouse, scrubber, and
wet electrostatic mist precipitator for removal of particulate matter
and H2S04 mist before entering the lead smelter acid plant. 'Clean
acid plant feed gas for the citrate plant is drawn from either of two
connections, one between the mist precipitator and the acid plant
drying tower and one downstream from the drying tower. This gas is
diluted tenfold with air to provide 1,000 scfm of gas containing
1054
-------
300~gal absorbent makeup tank
Sinter plant tail gas
"1
1,200-sq-ft boghouse
•Flue dust
J_
2.5-ft-diom by 18-ft-high packed scrubber tower
•H-0
• Porticulate sludge
Electrostatic mist precipitator
300-gal absorbent feed tank
• H2S04 mist
2.5-ft-diam by 30-ft-high packed absorption tower
Absorbent solution
Oil
Treated flue gas
- to atmosphere
Three 100-gal HjS precipitation reactors
•CO,
-Oil
100-gal oil-sulfur contactor
H2S-C02
50-gal sulfur flotation vessel
200-cu-fl sulfur float product storage bin] Steam
I
l-ft-diam by 4-ft-long
sulfur autoclave settler
— Natural gas
1/3 molten sulfur .
cast into 100-Ib molds
FIGURE 2.-Bunker Hill pilot plant.
-------
Dilute
To atmosphere
/ Humidifier
2 SOg absorber tower
J Sulfur precipitation reactor
4 H2S storage tank and vaporizer
5 Kerosine contactor
6 Sulfur flotation tank
7 Heat exchanger
8 Autoclave separator
9 Citrate makeup tank
Flue gas
Dilution air
H,0
Recycle liquor
FIGURE 3-Bunker Hill citrate pilot plant-Phase
D-2465-SL
-------
approximately 0.5 percent S02. The dilute S02 feed gas passes through
a 2.5-foot diameter by l8-foot-high Fiberglas-reinforced polyester
(FRP) humidifying tower containing a 6-foot section packed with 1-inch
polypropylene Intalox±/ saddles and a stainless steel mist eliminator.
Humidification of dilution air and the dry gas from the acid plant is
required to prevent excessive evaporation of citrate solution in the
absorption tower. During huaidification, the cool air-gas mixture is
heated by steam to raise the temperature of the air. For final
temperature control, the diluted gas is heated after humidification to
between U5° and 65° C (113° and 1^9° F) in a steam-to-air heat exchanger
with stainless steel tubes. This temperature range corresponds with
temperatures to which the lead sintering furnace tail gas would be
cooled before treatment in the citrate plant.
The gas stream then is passed upward through a 2.5-foot diameter by
30-foot-high FRP-packed absorption tower countercurrent to the citrate
solution, which absorbs over 95 percent of the S0£. Citrate solution
flows through the system at a rate of 10 gallons per minute. The
absorption tower contains three 6-foct sections packed with 1-inch
polypropylene Intalox saddles and a stainless steel mist eliminator.
Other gas-liquid contacting techniques may be applicable, but nearly
all our experience in the laboratory has been with packed towers.
From the absorption tower the citrate solution, at a pH of k.O to h-,5
and containing about 10 grams SOS per liter, flows by gravity to closed
stirred vessels for reaction with 2/3 ton of RSS per day to form 1 ton
of elemental sulfur. Three 100-gallon stainless steel reactor vessels
arranged for countercurrent flow of citrate solution and H2S gas are
available for the sulfur precipitation step. During the Phase I
operation, when nearly pure commercial-grade H2S is used, the 10-minute
retention time available in one reactor usually is sufficient for the
sulfur precipitation step. When the K2S-C02 mixed gas prepared from
recycled sulfur is used during Phase II operation, three reactors
probably will be necessary to insure adequate retention time for contact
of the gas and loaded citrate solution. A 1- to 3-percent-solids
slurry containing elemental sulfur and regenerated citrate solution
overflows the reactors and passes through a common header to a stain-
less steel reactor effluent tank. From this tank, the dilute slurry
is pumped to a 100-gallon FRP conditioner tank where kerosine or other
hydrocarbon oil is added for the sulfur flotation-separation step.
The oil-conditioned slurry flows to the 50-gallon-capacity feed well of
a specially designed sulfur flotation skimming device. This stainless
steel apparatus resembles an Esperanza drag classifier. The sulfur
separates from the citrate solution by floating to the surface as a
35- to U5-percent-solids product, leaving clear regenerated citrate
for recycle to the absorption tower. A fabric-reinforced, neoprene
~2jReference to specific trade names is made for identification only
and does not imply endorsement by the Bureau of Mines.
1057
-------
conveyor belt with stainless steel flights skims sulfur off the sur-
face of the citrate solution in the feed well. The sulfur is pulled
up an inclined chute and is discharged to a 200-cubic-foot, stainless
steel conical storage bin.
Regenerated citrate solution from the feed well of the sulfur skimmer
passes through a 50-gallon FRP settling tank and then to a JOO-gallon
FRP absorber feed tank. Any sulfur solids carried over to the settler
from the sulfur flotation skimmer are periodically pumped back to the
oil-conditioning tank with a diaphragm pump. Citrate solution from
the absorber feed tank is pumped through parallel backwash clarifica-
tion filters and a water-cooled heat exchanger to the absorption
tower.
On day shift only., the sulfur float product is withdrawn from the
storage bin with the aid of an attached vibrator and pumped by a
Moyno positive displacement auger-type pump through a single-tube,
steam-jacketed heat exchanger where the sulfur is melted at about 135°
C (275° F). Molten sulfur and citrate solution pass into a closed
settler tank at 155° C and under a pressure of 35 psi. Molten sulfur
is tapped from the bottom of the autoclave settler and cast in 100-
pound blocks. During the Phase II operation, a bleed stream of molten
sulfur will flow to the H2S generating plant. Citrate solution and
the oil used for flotation are withdrawn from the top of the settler
through a sulfur knockout pot, filter, and water-cooled heat exchanger
into a 50-gallon FRP decanting vessel for separation and reuse. Citrate
solution from this tank drains to the absorber feed tank.
Materials of construction used in the Bunker Hill pilot plant were
chosen to resist corrosion by the S02 gas and citrate liquor process
streams. Process piping is stainless steel, FRP, polyvinyl chloride
(PVC), or chlorinated polyvinyl chloride (CPVC), as applicable. All
stainless steel in the plant is type 316. An H2S incinerator is
provided outside the main building to burn H2S vented from the sulfur
precipitation reactors, oil flotation conditioner, and H2S tank trailer
areas, and to incinerate gas released under emergency or upset condi-
tions. The small amount of strong S02 gas from the incinerator passes
through a duct to the main baghouse of The Bunker Hill Co. lead
smelter where it is diluted by the nearly 200,000 cfm of gas from the
lead blast furnaces going up the main stack. A 40-foot-long van
modified to serve as a laboratory and instrument trailer is connected
to the main building. The pilot plant is completely instrumented and
controlled from the panel in the instrument trailer.1 A paper pro-
viding more detail on the design of the Bunker Hill pilot plant was
presented at the 19?4 AIME Annual Meeting in Dallas, Tex. (k).
1058
-------
Phase I Plant Operating Results
Phase I operation of the Bunker Hill citrate pilot plant was started
on February 15, 19?4. The plant has operated for a total of 1,300
hours through August 29, producing about If5 tons of bright yellow,
high-quality sulfur. Because of interruptions resulting from chang-
ing work crews, mechanical failures, and unavailability of feed gas,
the longest continuous operation to date has been about 160 hours.
Citrate loss over this period was 7.5 pounds per net long ton of
sulfur recovered from feed gas. Sulfur dioxide removal from feed gas
containing 0.3 to 0.1*5 percent S02 ranged from 93 to 99 percent when
operating at the design gas flow rate while varying gas temperature
and citrate concentration. Regeneration of the citrate solution and
precipitation of sulfur with H2S has been easily controlled in a single
reactor. The precipitated sulfur has been successfully recovered as
a high-purity product by oil flotation and melting.
Table 1 summarizes results obtained under reasonably steady-state
continuous operation at gas flow rates of 1,000 and 1,250 scfm and
S02 concentrations of 0,3 to 0.^5 percent. Gas temperatures ranged
from 35° to 50° C (953 to 122° F). Citrate solution concentration
was 0.5 M for most of the tests, the sodium-to-citric acid molar ratio
in the citrate solution was 2:1, and the pH of the citrate feed solu-
tion to the absorption tower was about 4.5.
TABLE 1. - Results of Bunker Hill citrate pilot plant
Gas
flow
rate,
scfm
1,000
1,000
1,000
1,000
1,000
1,250
Feed gas
concen-
tration,
pet SOp
0.32
Al
.34
.kk
.45
.42
operation -
Gas
temperature ,
0 F
95
100
116
122
109
119
February
Citrate
solution
flow,
gal/min
1/10
1/10
10
10
8
10
to September 1974
Citrate
solution
loa,ding ,
g/1 SOP
6.3
3.8
6.7
8.9
10.2
9-9
Off-
gas,
ppm S02
32
57
190
300
70
340
S02
removal,
pet
99-0
98.6
9^.4
93.2
98.4
91.9
_ 0.25 M citrate solution.
The test results show that at the design flow rate of 1,000 scfm, S02
absorption decreased from 99 percent at a gas temperature of 35° C
(95° F) to 93 percent when the gas temperature was 50° C (122° F).
Exit gas from the pilot plant ranged from 300 ppm S02 at the higher
temperature to only 32 ppm at a gas temperature of 95° F. Excellent
S02 absorption was obtained with the more dilute citrate solution at
the lower temperatures, even, though the solution loading represents
1059
-------
65 percent of the maximum equilibrium loading of the 0.25 M citrate
solution as compared with about 55 percent of the maximum loading for
the 0.5 M absorbent liquor.
The S02 removal efficiency was still over 98 percent, and the offgas
contained less than 100 ppm S02 when the citrate solution flow rate was
decreased to 8 gallons per minute, thus increasing the loading to 10
grams S02 per liter. This loading represents 60 to 65 percent of the
maximum equilibrium loading of 0.5 M citrate solution at the test
temperature of 109° F. In the test made at a gas flow rate of 1,250
scfm at the higher temperature of 119° F and operating with the higher
S02 loading of the citrate liquor, the S02 content of the offgas in-
creased to 3^0 ppm, but the S02 removal efficiency still exceeded 90
percent.
At the design capacity of the plant, a gas flow rate of 1,000 scfm and
a solution flow rate of 10 gallons per minute, the total pressure drop
through the absorption tower was 6 inches of water. As the solution
flow rate was reduced to 8 gallons per minute to increase the S02
loading of the citrate solution, the pressure drop decreased to 5.6
inches of water. Operation of the plant at a gas flow of 1,250 scfm
with 10 gallons per minute of citrate solution flowing through the
absorption tower increased the pressure drop to 9 inches of water.
Precipitation of sulfur with commercially produced, nearly pure H2S
and regeneration of the citrate solution for continued absorption took
place in a single reactor. No sulfur buildup occurred along the walls
of the stainless steel reactor or on the impellor provided the tip
speed of the impellor was at least $00 feet per minute. In the early
stages of plant operation, excess H2S absorbed in the citrate solution
resulted in cloudy recycle solution recovered from the kerosine flota-
tion step, apparently due to delayed precipitation of colloidal sulfur.
In addition, some of this absorbed H2S escaped at times from the sulfur
skimmer into the plant building. This problem was corrected by using
a second stirred reactor as a delay tank to allow more contact time
with the H2S and bypassing about 5-volume percent of the S02-loaded
liquor from the absorption tower to the reactor effluent tank to react
with the excess absorbed H2S. These measures have resulted in con-
sistently clear citrate solution from sulfur flotation for recycling
to the absorption tower and have stopped the escape of H2S from the
sulfur flotation equipment.
Ho problems have been encountered in plugging of lines between reactors
and the reactor effluent tank. A problem existed with sulfur buildup
in the automatic level-controlled reactor effluent tank until a small
agitator was installed to keep the sulfur in suspension. The kerosine
conditioner tank operates well with no sulfur buildup at the design
liquid flow rate, provided the impellor tip speed is at least 700 feet
per minute. Some trouble was experienced initially with holdup of
1060
-------
floated sulfur in the freeboard of the tank necessitating occasional
cleanout of the 3-inch-diameter overflow line to the skimmer. How-
ever, the addition of a second impeller operating just beneath the
liquid surface prevented buildup of the large lumps of powdery floated
sulfur that were blocking the overflow line.
A powder-like sulfur product of about 50 percent solids has been ob-
tained by adding between 55 and kO pounds of kerosine per ton of
sulfur produced. About 20 percent of the kerosine added for flotation
has been recovered from the melting operating for reuse. Most of the
kerosine loss can be attributed to volatilization from the hot sulfur
slurry in the kerosine conditioner and sulfur skimmer. Because of this
high volatilization loss, kerosine consumption during the pilot plant
operation to date has averaged 90 pounds per net ton of sulfur
recovered from the feed gas. Laboratory tests have indicated that this
hydrocarbon consumption can be reduced considerably by using low-
volatile motor oil in place of kerosine. In a 20-scfm continuous test
plant at the Salt Lake City Metallurgy Research Center, the use of SAE
10 motor oil resulted in a sulfar float product equivalent to that
produced with kerosine, and the oil consumption was one-fourth that of
kerosine. Various oils will be investigated in future campaigns at
the Bunker Hill pilot plant.
The sulfur melting step has functioned satisfactorily at the design
capacity. The melting rate seems to be limited to about 500 pounds of
sulfur- per hour by the maximum speed of the drive mechanism on the
Moyno sulfur pump. Some plugging problems have been encountered in
the citrate liquid lines from the autoclave settler, apparently because
of-sulfur being dissolved in the kerosine flotation reagent and then
crystallizing out upon cooling. Possibly, a dual filter downstream of
the liquid cooler or substitution of motor oil for kerosine will solve
this problem.
The sulfur produced by the Bunker Hill citrate pilot plant has been
bright yellow and of better than 99-5-percent purity. Carbon content
has ranged from 0.2 to 0.5 percent. In the laboratory, continuous
test plant sulfur recovered by motor oil flotation contained 0.01 per-
cent carbon.
During operation of the Bunker Hill citrate pilot plant, the rate of
oxidation of S02 to S04 was determined to be about 1.5 percent. This
is quite low considering that the feed gas, which is predominately air,
contained about 20 percent 02. Since the plant was started up in
February, the thiosulfate concentration of the citrate solution has
ranged from 20 to h-Q grams per liter. The sulfate concentration has
built up to about kO grams per liter. However, this figure is not
representative of the greater sulfate buildup expected because of a
few plant upsets resulting in large solution losses, which required a
1061
-------
makeup of fresh citrate solution. These losses occurred when the
reactor effluent tank or kerosine conditioner tank plugged causing
citrate solution to back up and flow through vent lines to the H2S
incinerator. These large solution losses appear to have been eliminated
by the modifications to the reactor effluent tank and kerosine condi-
tioner. In addition, large collection tanks have been installed in the
vent lines to the incinerator should plugging problems occur again.
In Phase II of the Bunker Hill pilot plant operation, the H2S genera-
tion plant will be operated utilizing both a one-stage and two-stage
procedure for production of H2S by the sulfur-methane-steam reaction
to provide data for engineering evaluation of this important step in
the citrate process. The H2S generation plant also will be operated in
conjunction with the sulfur dioxide absorption and sulfur recovery
section to determine the influence of the impure H2S gas on sulfur
precipitation and S02 removal efficiency.
In addition to operating the gas cooling and cleaning plant to determine
gas cleaning requirements during Phase III operation of the pilot plant,
test campaigns will be run on the Lurgi sinter furnace tail gas to
further relate feed gas temperature, citrate concentration, and citrate
solution loading to S02 absorption.
THE TERKE HAUTE CITRATE PILOT PLANT
The Pfizer-McKee-Peabody citrate pilot plant at Terre Haute, Ind., is
described in two recent publications (l,_3). -Briefly, the skid-mounted
unit treated 2,000-scfm of gas from a coal-fired spreader stoker-type
boiler with a 25,000-pound per hour steam-rated capacity. The flue gas
was adiabatically cooled with quench water and passed into a Venturi-type
water scrubber to remove fly ash. Absorption of S02 took place in an
impingement plate scrubbing tower using an aqueous solution of sodium
citrate and citric acid. The S02-rich citrate solution flowed from the
absorber through a steam-heated heat exchanger to a three-stage con-
tinuous stirred tank reactor system countercurrent to a flow of pure.
H2S from a tank trailer. The S02 was reduced to sulfur and the citrate
solution regenerated. Sulfur slurry was pumped from the reactors to a
surge tank and then to a sulfur flotation separation system. Citrate
liquor was recycled from the flotation unit to the absorption system.
The sulfur flotation product was pumped as a slurry through a heater at
a temperature above 125° C (257° F) and a pressure of 70 psig to melt
the sulfur. Liquid phases were separated in a pressure decanter where
the bottom layer was drawn off as high-quality molten*yellow sulfur and
the citrate solution top layer was discharged to a flash drum at reduced
pressure.
The following statement on plant operation was furnished by Louis Korosy
of Pfizer Inc., Frank Chalmers of Arthur G. McKee and Co., and Srini
Vasan of Peabody Engineering Systems:
L062
-------
"While generally similar to the Bureau of Mines unit, there
are two major design differences%in the Terre Haute plant:
(1) An impingement plate tower is\used rather than"slacked
tower. This permits the use of higher gas velocities and
thus smaller tower diameters; (2) the sulfur separation is
based on a flotation principle, but no -hydrocarbon addition
is made.
Operation under the final equipment configuration started
on March 15. Between then and September 1, 2,330 hours
of operation were logged. Runs were generally five to
six days in length.
During this period the S02 removal efficiency of the
system was well above 95 percent. SOS in the exit
gas consistently was less than 50 ppm and actually
averaged 3^ ppm. At this time the inlet gas to the
system averaged about 1,000 ppm S02,
The trouble-free operation of the absorber in the citrate
process was highlighted during one phase of the pilot
plant program. While clear liquor is normally returned
to the absorber from the surfur separation step, for a
period of weeks sulfur slurry was deliberately returned
to the absorber. No problems were encountered.
At this point the test unit has provided sufficient data
to prepare scale-up designs for further projects on a
guaranteed S02 removal basis."
COST ESTIMATE FOR CITRATE PLANT AT
COAL-BURNING POWERPLAIfT
Based on results of the Bunker Hill and Terre Haute pilot plant
operations, a cost estimate was prepared by the Process Evaluation
Group at the Salt Lake City Metallurgy Research Center to determine
the cost of removing 95 percent of the sulfur from the tail gas of a
1,000-MW coal-burning powerplant. Assuming that the powerplant burns
8,UOO tons per day of coal containing 3 percent S, the gas flow would
be 1,730,000 scfm (60° P) with an S02 content of 0.2k percent. The
yield of sulfur would be about 214 long tons per day. The plant is
assumed to operate 7,000 hours (292 days) per year.
Annual operating costs for the plant are itemized in table 2. A
summary of operating labor requirements, fixed capital cost, annual
operating costs, and unit production costs for each of the unit opera-
tions and for the entire plant are presented in table 3. The fixed
capital costs were estimated by standard chemical engineering cost
estimating procedures for a "study estimate" and are for the second
quarter of 197^ or a Marshall and Swift (M&S) index of 386.
1063
-------
The operating cost of the citrate plant includes reheating the final
tail gas from 50° to 170° C (122° to 3^0° F) with a natural gas burner
at the base of the powerplant stack. Operating losses are assumed at
11.2 pounds of citric acid and 16 gallons of kerosine per long ton of
product sulfur. Costs for Na2S04 removal were estimated on the basis
of 1 percent of the removed sulfur being converted to sulfate.
TABLE 2. - Annual operating cost of S0? removal from
1,000-MW powerplant offgas
Unit cost
Total cost
Direct cost
Raw materials
Citric acid
Soda ash
Kerosine
Utilities
Electricity
Natural gas
Process water
Cooling water
Steam
Direct labor
Labor
Supervision
Plant maintenance
Labor
Supervision
Materials
Payroll overhead
Operating supplies
Total direct cost
Indirect cost (administration and
overhead)
Total capital charges
Total annual operating cost
$3^9,100
48,600
648,400
317,600
2,546,000
13,900
223,700
941,600
151,500
22,700
563,900
112,800
655,800
11,046,100
4,042,800
174,200
1,332.500
212,800
266,500
7,074,900
709,500
6,084,800
13,869,000
1064
-------
TABLE 3- - Summary of costs for S0g removal from a 1,000-MW
powerplant offgas
Number of
operators
Capital
cost
Unit
Annual production
operating cost per long
cost ton of sulfur
Gas cooling and S02
absorption k.2
Sulfur precipitation 2.1
Sulfur recovery 2.1
H2S generation 2.3
Na2S04 removal 2.3
Facilities
Utilities
Fixed capital
Working capital
Total 13.0
$19,951,000 $8,096,000 $129.86
611,000 1,183,000 18.98
3,217,000 1,680,000 26.95
3,056,000 2,685,000 U3.06
1^83,000 225,000 3.61
2,71^,000
,3,257,000
33,289,000
3,100,000
36,389,000 13,869.000 222 A6
Direct costs include materials and utilities at the following unit
costs: Citric acid at $1,000 per ton; soda ash at $^0 per ton;
kerosine at $0.65 per gallon; electric power at $0.01 per kW-hr;
natural gas at $1.00 per thousand cubic feet; and water at $0.17 per
thousand gallons for process use, and $0.03 per thousand gallons for
cooling. These costs also include direct labor at $5-60 per hour,
plus supervision at 15 percent of direct labor; plant maintenance
consisting of maintenance labor at 2.2 percent of fixed capital costs,
plus supervision at 20 percent of maintenance labor and maintenance
material at 116 percent of maintenance labor; payroll overhead at 25
percent of the cost of labor and supervision for operations and
maintenance; and operating supplies at 20 percent of plant maintenance.
Indirect costs are hO percent of direct labor, plant maintenance, and
operating supplies.
Capital charges, which include amortization, taxes, and insurance are
17 percent of the total capital which is the average annual fixed
capital charge for privately financed steam-electric plants.
As shown in table 3, the capital investment for a citrate plant at a
1,000-MW coal-burning powerplant is estimated at $36.^ million.
Assuming no credit for sulfur, operating cost for removing 95 percent
of the S02 from the stack gas is about $13.9 million. This is equiva-
lent to $5.66 per short ton of coal or 1.98 mills per kW-hr.
1065
-------
FUTURE PLANS
Plans are underway for one or more large-scale plants to demonstrate
the citrate process for S02 emission control at powerplants or steam-
generating facilities burning high-sulfur coal or oil. The
demonstration plants will operate on 30- to 60-MW powerplants or
steam-generating plants of equivalent capacity. They will be provided
and operated under a cooperative arrangement and cost sharing basis
between the Bureau of Mines and the Environmental Protection Agency
and interested industrial firms. Proposals for demonstration plants
including preliminary engineering estimates are to be submitted by
early December 197^. Contracts will be awarded, after negotiations,
probably early in 1975-
Work on the citrate demonstration plant contracts will be divided into
four phases. Phase I, consisting of process design and definitive
cost estimates, should be completed by July 1975- Phase II, which
includes detailed engineering design, construction, and mechanical
acceptance of the plants, should be completed by July 1978. Phase
III consisting of startup and performance acceptance testing would
take place at the conclusion of Phase II. This would be followed by
Phase IV, comprehensive emission testing programs to be conducted at
the demonstration plants by independent contractors for 1 year.
1066
-------
REFERENCES
1. Chalmers, Frank S., Louis Korosy, and A. Saleem. The Citrate
Process to Convert S02 to Elemental Sulfur. Pres. at Industrial
Fuel Conf. Purdue University, West Lafayette, Inc., Oct. 3, 1973,
6 pp. (Available upon request from Arthur G. McKee & Co.,
Cleveland, Ohio).
2. George, D. R., Laird Crocker, and J. B. Rosenbaum. The Recovery
of Elemental Sulfur from Base Metal Smelters. Min. Eng., v. 22,
No. 1, January 1970, pp. 75-77.
3. Korosy, L., H. L. Gewanter, F. S. Chalmers, and S. Vasan.
Chemistry of S02 Absorption and Conversion to Sulfur by the
Citrate Process. Pres. at l67th ACS Meeting, Los Angeles,
Calif., Apr. 5, 197^, 32 pp. (Available upon request from
Pfizer, Inc., New York, N. Y.).
4. McKinney, W. A., W. I. Nissen, and J. B. Rosenbaum. Design and
Testing of a Pilot Plant for S02 Removal From Smelter Gas. Pres.
at AIME Ann. Meeting, Dallas, Tex., Feb. 23-28, 19?U, AIME
Preprint A-7^-85, 12 pp.
5. Rosenbaum, J. B., D. R. George, and L. Crocker. The Citrate
Process for Removing S02 and Recovering Sulfur From Waste Gases.
Pres. at AIME Environmental Quality Conf., Washington, D. C.,
June 7-9, 1971, 2.6 pp. (Available upon request from the Salt
Lake City Metallurgy Research Center, Salt Lake City, Utah).
6. Rosenbaum, J. B., W. A. McKinney, H. R. Beard, Laird Crocker, and
W. I. Nissen. Sulfur Dioxide Emission Control by Hydrogen Sulfide
Reaction in Aqueous Solution - The Citrate System. BuMines
RI 777^, 1973, 31 PP-
1067
-------
TVA-EPA PILOT-PLANT STUDY OF THE AMMONIA ABSORPTION
AMMONIUM BISULFATE REGENERATION PROCESS
Claude E. Breed
Tennessee Valley Authority-
Muscle Shoals, Alabama
Gerald A. Hollinden
Tennessee Valley Authority
Chattanooga, Tennessee
Prepared for Presentation at
Flue Gas Desulfurization Symposium
Sponsored by the Environmental Protection Agency
Atlanta, Georgia
November k-1, I9lh
1069
-------
TVA-EPA PILOT-PLANT STUDY OF THE AMMONIA ABSORPTION
AMMONIUM BISULFATE REGENERATION PROCESS
Claude E. Breed
Tennessee Valley Authority
Muscle Shoals, Alabama
Gerald A. Hollinden
Tennessee Valley Authority
Chattanooga, Tennessee
ABSTRACT
The Tennessee Valley Authority and the Environmental Protec-
tion Agency began studies of an ammonia scrubbing program at TVA's
Colbert pilot plant in 1969. The principal advantage in the TVA-EPA
process is economic regeneration by acidulation with ammonium bisul—
fate produced from thermal decomposition of ammonium sulfate. This
paper highlights the pilot-plant activities since the 1973 Flue Gas
Desulfurization Symposium.
Absorber product liquors, that are amenable to regeneration,
have been produced while obtaining good S02 removal and low ammonia
losses. However, the following problem areas have been established:
• Open—loop operation of the prewash section.
• Fume formation in the absorber and after discharge from
the stack.
• Solids precipitation in the absorber liquors.
• Incomplete S02 release in the acidulation and stripping
operations.
The prewash section was operated open loop to minimize cor-
rosion but will have to be closed in a final process design. Under
proper operation of the system, it is possible to control fume forma-
tion inside the absorber while operating at relatively high salt
concentrations. Avoiding fume formation on days of low temperature
and high relative humidity may be impractical to achieve. Precipi-
tated solids in the absorber liquor may present problems in the
regeneration section. Most of the absorbed S02 can be released in
the acidulator and stripper, but the remaining S02 creates problems
in the evaporator. The acidulator and stripper are being redesigned
to achieve a complete release of S02. An ammonium sulfate decomposer
has been designed.
1071
-------
TVA-EPA PILOT-FLAM1 STUDY OF THE AMMONIA ABSORPTION -
AMMONIUM BISULFATE REGENERATION PROCESS
Absorption of sulfur dioxide from plant flue gases using
aramoniacal solution and subsequent regeneration of the absorbing solu-
tion has teen proposed or demonstrated in several processes. Lepsoe
and Kirkpatrick1 report that one of the earliest of these is the Cominco
process in which S02 from a smelter operation is absorbed in an aramonia-
cal solution, the solution is acidulated to produce ammonium sulfate
for disposal as a fertilizer and the evolved S02 is sent to an acid
plant. TVA piloted a similar process on power plant stack gases in
the 1950's according to Hein, et al.2 Other processes which produce
fertilizers as the end product have been developed and are in full-
scale operation on sulfuric acid plant tail gases in Czechoslovakia
and Romania.3;4 These processes use nitric or phosphoric acid to
release the S02 and produce marketable fertilizers. Limited long-term
markets for ammonium sulfate and the constraints on the location of
power plants to fertilizer manufacturing centers limit their usage in
the U.S.
Regeneration processes that do not rely on fertilizer mar-
keting have been studied extensively. H. F. Johnstone^ developed a
process in the 1930's that produced only S02. Steam stripping was
employed to recover S02 and to regenerate the ammoniacal solution for
reuse-in the scrubber. Since concentrated S02 is the major product,
either sulfuric acid, liquid S02, or elemental sulfur can be the
final product, depending on the need of the user. Although the
Johnstone process has been operated successfully on a small pilot
scale and presupposes no link to a fertilizer plant, it possesses
some undesirable characteristics. Energy requirements for the strip-
ping step are relatively high, 10—15 pounds of steam per pound S02. Oxi-
dation products are difficult to purge from the system without loss
of active species as well. The occurrence of undesirable dispropor-
tionation reactions in the steam stripper further aggravates the
oxidation problem.
Development of the Johnstone process has been vigorously
pursued in the USSR. Several pilot-plant studies led to installation
of a large-scale (about TO MW) system on a coal-fired power plant in
Moscow in 1952 and the system was operated until the power plant was
converted to burn gas in 196l. Regeneration was based on steam strip-
ping. A wet electrostatic precipitator was reported effective in con-
trolling fume emission. During the period 1968-1972, an ammonia
scrubbing process was used on a 15-megawatt oil-fired power plant
located at Ufa in the Ural Mountains. Both direct oxidation to produce
ammonium sulfate and regeneration by the autoclave process were tested.
Corrosion was a serious problem.
Design work has been started on an ammonia scrubbing system
for a full-scale (about 250 MW) oil-fired boiler near Moscow. Completion
1072
-------
in 1977 is planned. Thermal regeneration will be used for recovery of
S02, "but the TVA-EPA ammonium Msulfate process is also being considered.
Ammonia scrubbing is included in the agreement for formal exchange of
technology between the U.S. and the USSR.
Other regenerative schemes not dependent on the fertilizer
market structure are mentioned below.
The Institut Francais de Petrole (IFF) in France is develop-
ing a process based on ammonia scrubbing. The various steps of the
process have been piloted separately at various locations. Two inter-
grated systems are now being installed in Japan; one on a Glaus plant
and one at a refinery.
In March 197^, Nippon Kokan KK began operating a full-scale
ammonia/lime double-alkali unit for treating 150,000 Nm3/h waste gas
from an iron ore sintering plant (300,000 Nm3/h capacity) at the
Keihin Works, Kawasaki, Japan. An ammonia-based fume from the absorber
is considered a major problem. The calcium salts in the system are
converted to gypsum and discarded. No additional units are currently
being installed by Nippon Kokan KK using the ammonia double-alkali
scrubbing process.
In September 1973, TVA modified its limestone - wet-scrubbing
pilot plant to obtain preliminary data on an ammonia-limestone/lime
(double alkali) scrubbing process. Good S02 removal was obtained
(90$ +). Two major problems identified during the initial period of
this pilot-plant operation were (l) excessive losses of soluble ammonia-
sulfur salts with the discarded solids (calcium sulfite, fly ash, and
unreacted limestone) from the filtration step and (2) formation of a
dense, persistent plume by the scrubbed gases exhausted to the atmosphere.
THE TVA-EPA AMMONIA ABSORPTION -
AMMONIUM BISULFATE REGENERATION PROCESS
In 1968, EPA contracted with TVA to begin a pilot-plant
program to study in depth the ammonia absorption process as applied
to power plants. The pilot plant was located at TVA's Colbert Power
Plant, a pulverized coal-fired installation in northwest Alabama. The
initial phase of study covered a number of variables in the absorption
step. The results led EPA and TVA to amend the program to include a
process for regenerating the liquor from the absorber operation. Both
TVA and EPA have shared in funding the amended study. The regeneration
process selected for the second phase of pilot-plant study was the
ammonium bisulfate process. A topical report recently issued by TVA6
and a paper given at the Flue Gas Desulfurization Symposium in New
Orleans by Hollinden, et al.,r last year covers the early absorption
study and the first year's work on the regeneration process. The pur-
pose of this paper is to update those reports to cover the past year's
pilot-plant work.
1073
-------
The TVA-EPA ammonium bisulfate regeneration process (ABS
process) uses a version of the HixonnMiller8 scheme for release of
S02 from ammoniacal solutions by acidification and subsequent decom-
position of the ammonium sulfate product to ammonia for recycling to
the absorber and ammonium bisulfate for recycling to the acidulator.
The net absorption reactions are:
(1) WH3 + S02 + H20
(2) NEiHSOs + WH3
(5) (NH4)2S03 + S02 + H20 - »'2NH4HS03
H20 + S02f
2(NH4)2S04 + H20 +
The net regeneration reactions are:
As in the Johnstone process, the only product from the ABS process is
a gaseous stream of S02. The ABS process has advantages over the sul-
fite steam stripping process in that disproportionation reactions are
avoided and sulfites are not purged from the system. Also, energy
requirements for the ABS process are estimated to be approximately
one— half to two-thirds that of the Johnstone process. For simplicity, the
pilot plant is divided into the absorption and the regeneration sections.
Absorption Section
A flowsheet of the absorption section of the pilot plant is
shown in Figure 1. The pilot plant is designed to treat approximately
5000 cfm of flue gas at absorber conditions (saturated at about 125°F).
The gas flows through a prewash section containing a venturi-type
element. The conditioned gas is then routed to the absorption tower.
The absorber consists of as many as four independent absorbing stages
(valve tray elements).
The makeup water to the absorber is added to the top recircu-
lating stage (G-U) to maintain a low salt concentration at the top of
the tower to control the ammonia loss from the system. Ammonia is added
to the second absorption stage (G-2). A portion of the liquor from the
bottom tray is pumped to G-2 to decrease the pH of the liquor on the second
stage. Decreasing the pH reduces the vapor pressure of ammonia which
decreases the possibility of gas phase reactions of ammonia, S02, and
water to form a fume in the absorber. Product liquor is withdrawn from
the first stage (G-l) and is stored in surge tanks for use in the regenera-
tion section. The scrubbed flue gas is exhausted to the atmosphere either
with or without reheating.
A discussion of the individual steps follows:
Gas Pretreatment. The inlet flue gas enters the pretreatment
section (Figure 2) through a 1- by 1-foot duct which contains a venturi
1074
-------
-TO STACK
CHEVRON-TYPE
MIST ELIMINATOR
FLUE
GAS
PRODUCT
J-2
J-3
Figure 1. Absorption section.
-------
I FLUE
I INI
FLUE GAS
INLET
pooo
INLET
WATER
VENTURI
ELEMENT
1
SUMP
RECIRCULATION
PUMP
EXIT GAS
TO ABSORBER
LIQUOR OVERFLOW
TO DRAIN
Figure 2. Prewash section.
1076
-------
element similar to Environeering Inc. 's Ventri—Rod element. The element
consists of a number of 3/^-inch rods mounted perpendicular to the gas
flow. The number of rods can be varied to increase or decrease the pres-
sure drop (AP) across the element. Recirculating wash water from a sump
beneath the venturi element is sprayed cocurrently with the gas stream
immediately upstream from the element.
The gas and spray water passes through the venturi element
and into the sump. The velocity of the flue gas decreases in the sump
to 8 feet per second allowing some of the spray water entrained in the
venturi to separate. The gas travels the length of the sump and is
routed through a 1^-inch-diameter duct to the absorber.
The purpose of the gas pretreatment is to humidify and cool
the flue gas and to remove fly ash and chlorides from the gas stream.
Humidification, cooling, and chloride removal are considered necessary
for reducing the plume emission from the absorber. Fly ash is removed
before the absorption section to decrease the amount of solids in the
scrubber liquor. These solids could cause major problems in the
absorption and regeneration sections.
In the test program, the pressure drop across the venturi and
the liquid-to-gas ratio (L/G, gal/1000 ft3) of wash water to the element
were varied to determine their effects on the particulate, sulfur dioxide,
and chloride removal, and on mist carryover. The pressure drop ranged
from 0.5 to 15.0 inches of water and the L/G from 10 to 30. The flue
gas used was taken both upstream and downstream from an electrostatic
precipitator. The results of these tests are summarized below.
e The inlet fly ash loading was reduced an average of 93$
when the inlet dust loading was 4 to 6 grains of dust per
dscf. No reduction in dust loading was obtained when the
inlet loading was low (0.01-0.1 gr/dscf).
9 No direct relationship was established between the S02
removal and the L/G, pressure drop, or pH of the wash
water in the prewash section. Between 7 and. 23$ of the
inlet S02 (2600 ppm) was removed in the section.
® An increase in the pressure drop across the venturi ele-
ment and the L/G to the element increased the mist carry-
over into the absorber.
9 Essentially all of the chloride in the flue gas (35 ppm
average) was removed in the venturi section at the antici-
pated normal operating conditions.
In order to minimize mist carryover during absorber tests,
the venturi prewash section was operated at as low a pressure drop
and L/G as possible consistant with good humidification. Based on
actual operation, these conditions are a pressure drop of 5 to 6 inches
of water and an L/G of about 10.
1077
-------
There were three major problem areas identified during the
operation of the prewash section (l) low pH of sump liquor (corrosion),
(2) disposal of low pH purge water from prewash sump, and (3) mist carry-
over from the prewash section to the absorber section.
During initial operation, the quantity of makeup water added
to the prewash section was that consumed in the process for humidifi—
cation and absorber product liquor bleedoff, approximately 1. 2 gpm.
At this flow rate, the pH of the wash liquor was 1. 0 or less. The
makeup water throughput was increased to 15-30 gallons per minute to
increase the pH of the wash water and to reduce its corrosion potential.
This raised the pH of the wash water to approximately 2. 5, but corro-
sion continued.
An inspection of the gas pretreatment section was made after
2000 hours of operation. The entire pretreatment section, constructed
of Type 3l6L stainless steel, was badly damaged by corrosion. The rods
across the throat of the venturi were pitted badly, especially on the
ends, and several needed replacing. The liquor sump walls were corroded
with the most severe attack above the normal gas—liquor interface. In
this area, pits ho to 60 mils deep were common. The gas duct from the
sump to the absorber and the sample probes inside the duct (316 S. S. )
were also severely corroded.
The gas prewash section was operated "open loop" to minimize
corrosion of the equipment. The problem of disposing and/or utilizing
low pH water will be common to most S02 removal processes that require
a prewash. It is recognized that in a commercial operation, this loop
would have to be closed to meet environmental regulations. Attempts
to close the loop around the prewash section will be studied in future
operations.
Mist carryover from the venturi section to the absorber dilutes
the absorber product liquor and transports dissolved and undissolved
solids into the absorber section. Both of these problems are discussed
in more detail in the absorber section of this paper.
Currently, the prewash section is being redesigned to resolve
the corrosion and mist carryover problems. To protect against corrosion,
the sump and exit gas ducts will be constructed of FRP; the venturi
throat will probably be lined with neoprene. Several types of corrosion-
resistant materials will be tested for use as rods in the venturi. The
new prewash section will include a mist eliminator (prpbably a chevron
in the horizontal position) to reduce the amount of carryover to the
absorber.
SOg Absorber. The absorber shell consists of 32-inch by 32-inch
by h-foot sections. At normal operating conditions, the gas velocity
in the tower is approximately 7 feet per second. Two types of absorp-
tion elements, marble beds and. valve trays, have been tested.
1078
-------
A two-pass chevron mist eliminator is located in the vertical
position above the top absorption stage.
Two series of tests were run during the past year. During
the first test period, the absorber consisted of one valve tray followed
by two marble beds, Figure 3- Tests were made to determine the highest
liquor concentration that could be obtained using this configuration
while limiting the outlet S02 concentration to < 250 ppm and the ammonia
concentration to < 50 ppm. Product liquors with high concentrations
require less energy for regeneration and smaller regeneration equipment.
The maximum product liquor concentration obtained using this
configuration was a C& (mols of active ammonia present as sulfite and
bisulfite per 100 mols of total water) value of l4. Factors limiting
the CA in the initial series were (l) the diluting effect of the mist
carryover from the gas pretreatment section and (2) the S0a and NH3
vapor pressure over the liquor on the top stage. An increase in the
concentration of the product increases the liquor concentrations
throughout the tower. An additional absorption stage was required to
reduce the ammonia and S02 vapor pressure over the liquor on the top
stage to meet the arbitrarily set limits.
The second series of tests was made with the fourth stage
added to the absorber. Figure k shows the absorber arrangement after
addition of the fourth absorber stage. The two marble—bed absorption
elements were replaced with valve trays during the modification
because of occasional excessive stage-to-stage weepage which upset
the liquor concentrations throughout the absorption tower. Experience
showed that this problem does not occur with the valve tray when the
gas flow rate remains relatively constant although weepage from the
valve trays is a problem when the gas flow varies.
Tests using this absorber arrangement indicated that product
liquor could "be produced that had C.'s of 20 and above. Table 1 shows
data from the absorber operation while producing liquor with a C. of 21.
No absorber operational problems were encountered during operation at
the high concentration.
After determining that high liquor concentrations could be
produced in the modified absorber, it was operated to produce liquor
with CA'S of 12 to 15 for use in the regeneration section. Table 2
shows typical data from this operation. Solutions with concentrations
in this range are acceptable for regeneration in the ammonium bisulfate
process and are much easier to achieve within the limits of S02 and
ammonia loss and plume emission than are the more concentrated solutions.
The most apparent problem in the absorption step is that of
fume formation, Figure 5, From results previously reported by TVA on
1079
-------
INLET
GAS
G-3
(MARBLE BED)
G-2
(MARBLE BED)
V-l
(VENTURI-L_
TYPE
ELEMENT)4-
G-l
[VALVE TRAY)
EXIT GAS TO ATMOSPHERE
.
' -
REHEAT
V
=u
r- MAKE-UP
. _ .j WATER
VS-I
PURGE TO
ASH POND
F-l
AMMONIA
I
n
F-2
PRODUCT
Figure 3. Three-stage absorption.
PROCESS
MAKE-UP
WATER
F-3
1080
-------
FLUE GAS
OUTLET
FLUE GAS
OUTLET
LIQUOR
INLET
FLUE GAS
FROM VENTURl
SECTION(V-I)
LIQUOR
INLET
LIQUOR
INLET
LIQUOR
INLET
G-l
VALVE TRAY)
LIQUOR
INLET
FLUE GAS
FROM VENTURl
SECTION(V-I)
SECTION
6
SECTION
5
LIQUOR
OUTLET
CHEVRON MIST
ELIMINATOR
C-4
•(VALVE TRAY)
LIQUOR
OUTLET
G-3
(VALVE TRAY)
LIOUOR
OUTLET
G-2
(VALVE TRAY)
LIOUOR
OUTLET
G-l
(VALVE TRAY)
LIOUOR
OUTLET
Figure h. Three- and four-stage absorber.
1081
-------
TABLE 1. TYPICAL ABSORBER LOOP TEST—AMMONIUM BISULFATE PILOT PLANT
Test Conditions
Gas to absorber
Flow rate, scfm at 32°F 2350
Temperature, °F 295
S02, ppm 2800
Gas leaving venturi
Temperature, °F 120
S02, ppm 2720
Gas leaving first stage
Temperature, °F 123
S02, ppm 1160
Gas leaving second stage
Temperature, °F 12k
S02, ppm iijO
Gas leaving third stage
Temperature, °F 115
S02, ppm 510
Gas leaving fourth stage
Temperature, °F ilk-
S02, ppm 390
S02 removal, % Qk
Fresh water feed to absorber, gpm 0.2
Ammonia feed to absorber, Ib/hr 22.5
Liquor flow from first stage to second stage, gpm 1.6
Liquor to first stage (product)
CA 21.1
S/CA o. 73
PH 6.2
Specific gravity 1. 286
Liquor to second stage
CA 21.5
S/CA o. 67
PH 6.3
Specific gravity 1. 282
Liquor to third stage
c, 6.6
S/CA 0.78
PH 5.9
Specific gravity 1.138
Liquor to fourth stage
C PS
s^c o. 86
PH 5.7
Specific gravity 1. 086
1082
-------
TABLE 2. TYPICAL ABSORBER LOOP TEST—AMMONIUM BISULFATE PILOT PLANT
Test Conditions
Gas to absorber
Flow rate, scfm at 52 °F 2550
Temperature , °F 305
S02, ppm 2680
Gas leaving venturi
Temperature, °F 125
S02, ppm 2520
Gas leaving first stage
Temperature, °F 121
S02, ppm 12k)
Gas leaving second stage
Temperature, °F 122
S02, ppm 530
Gas leaving third stage
Temperature, °F 116
S02, ppm 520
Gas leaving fourth stage
Temperature, °F 115
S02, ppm 250
MS, ppm 10
S02 removal, 91
Fresh water feed to absorber, gpm 0.5
Ammonia feed to absorber, Ib/hr 21.5
Liquid flow from first stage to second
stage, gpm 0.74
Liquor to first stage (product)
CA 14.7
S/CA 0.80
pH 5-9
Specific gravity 1.256
Liquor to second stage
CA 15.5
S/CA 0.66
pH 6.4
Specific gravity 1.220
Liquor to third stage
CA 4.0
s/c. 0.84
PH 5.9
Specific gravity 1.095
Liquor to fourth stage
CA 1.6
S/CA 0.94
pH 5.6
Specific gravity 1.042
1083
-------
o
CO
Figure 5. Unacceptable plume during routine operation; product from
absorber: C^ = 10.0, S/CA = 0.8; no reheat.
-------
the use of a single—stage marble—bed absorber, it was concluded that an
acceptable plume (defined as < 5$ opacity for this pilot plant) could
be obtained when:
• A water wash is used ahead of the absorption stage.
« The absorber and all ducts are insulated.
• Reheat is applied as required to dissipate the steam plume.
Additional tests were run using a single—stage valve tray-
absorber, Figure 6, to obtain data on the plume emitted from an uninsu-
lated absorber as compared with an insulated absorber. The absorption
stage was sandwiched between two water—wash stages. Either or both of
the water stages were turned off to determine the effect on plume
opacity. The CA values of the absorption liquor tested were 2, 7, and
12. The S/CA ratio (mols of S02 present as sulfite and bisulfite per
mol of active ammonia present as sulfite and bisulfite) was controlled
near 0.8. The exit gas was reheated to 1^0° and 175°F in each test.
The water wash from the bottom decreased the plume opacity
in all cases. The water wash was not used in Figure 7 and the plume
opacity was 5 to 10$. The water wash was activated and the opacity
decreased to 0$; Figure 8. The same amount of reheat (200°F) was applied
in both cases to eliminate the steam plume. The water wash from above
lowered the opacity only at a CA of 12. The combination of a water wash
on the bottom tray and reheat resulted in an opacity of 5$ or less for
all three levels of CA'S, Figures 9, 10, and 11.
The plume was also observed during multistage absorber operation.
The insulated absorber shown in Figure 3 contained three absorption stages
(one valve tray and 2 marble beds) preceded by a prewash section.
A steam and particulate plume was emitted from the absorber in
these multistage absorber tests. In most cases, a reheated exit gas tem-
perature of 10° to 20°F above the calculated value at which the steam plume
should dissipate was required to reduce the plume to less than or equal to
5$ opacity, Figure 12. A previous test series, reported at the Flue Gas
Desulfurization Symposium in New Orleans last year, indicated that a
single—stage absorber could be operated with an acceptable plume while
reheating to or just above the calculated temperatures. The only opera-
tional changes made in comparison to the previous series were three
stages of absorption versus one and higher liquor concentration on all
stages. Mist carryover from the more concentrated liquor on the top
stage probably accounts for the increase in reheat temperature required
to maintain the same opacity. During reheat, the mist is evaporated,
thus increasing the vapor pressure of S02 and NH3 in the exit gases.
A plume then results from the gas phase reaction of S02 and ammonia in
the presence of water in the gas duct and/or in the atmosphere.
1085
-------
CHEVRON -TYPE
MIST ELIMINATOR
G-4
(VALVE TRAY)
6-3
(VALVE TRAY)
G-2
(VALVE TRAY)
G-l
(VALVE TRAY)
INLET GAS
MAKE-UP
WATER
t
EXIT GAS TO ATMOSPHERE
v_
TO DRAIN
REHEAT
MAKE-UP
WATER
AMMONIA
F-2
PRODUCT
J
MAKE-UP
WATER
TO
DRAIN
Figure 6. Single-stage valve-tray absorber.
1086
-------
o
00
Figure 7. Opera i
v/j_uu no water wash ahead of first absorber stage;
200° F reheat; 5 to lO/o opacity.
-------
a
Co
v
I
Figure 8. Operation with water wash ahead of first absorber stage;
200°F reheat; 0$ opacity.
-------
c
00
•JD
Figure 9. Plume from operation with product solution having c^ » 2; one
absorber stage sandwiched between two water wash stages; no insulation;
lU2°F reheat; 0% opacity.
-------
o
iO
o
Figure 10. Plume from operation with product solution having C^ = 1; one
absorber stage sandwiched between two water wash stages; no insulation;
°F reheat; 0$ opacity.
-------
Figure 12 . Plume from operation with product having C^
three absorber stages preceded by a water wash stage; 1&0°F
reheat; 5% opacity.
1092
-------
When the fourth absorption stage was added to the tower, the
insulation was removed and not replaced. As in the 5—stage insulated
absorber operation, the plume from the uninsulated l*-stage absorber was
controlled at % opacity in most cases by reheating the exit gas 10° to
20° F above the temperature required to dissipate the steam plume, Figure 13.
The use of the prewash, the reheat, and dilute solutions at the
top of the tower are considered necessary in controlling the plume from
the absorber. The value of insulating the absorber has not been fully
evaluated by TVA and no recommendations can be made at this time.
In all cases, the ambient weather conditions have an overriding
effect on plume emission. On cold wet days, the plume cannot be reheated
to a high enough temperature to eliminate the plume. On warm days with
high humidities, the plume reforms outside the absorber 10 to 20 feet
downwind of the stack.
In two sampling and analysis series conducted about two months
apart, EPA measured the levels of ammonia and S02 in the exit gas as 1.k&
+ k.n ppm and 2j4 + 88 ppm, respectively. Particulate mass emissions
and particulate size distribution of the effluent particulate were also
determined by EPA. The measured effluent particulate emission concen-
tration was 0. 057 + 0.015 gr/dscf and 0.127 + 0. 015 gr/dscf. Calcula-
tions indicate that the measured particulate concentration in the exit
gas could not totally be accounted for via the gas phase reaction between
NE3 and S02 but was primarily due to mist carryover from the absorber.
Size distribution analysis indicated that 88.1$ + 2. 6 of the particulate
was finer than about 2 microns and that 6l. 7 + 9$ o~f the emitted particu—
late was finer than about 0. 8 micron. Analysis of the emitted particu—
late by X-ray defraction technique and wet chemical methods indicated
that the particulate was pure ammonium sulfate. As stated earlier, this
particulate loading is believed to be the result of carryover.
The precipitation of solids in the absorption section may
present another major problem. A yellow solid, tentatively identified
petrographically as a homogeneous iron-ammonia—sulfur compound was pre-
sent in all of the absorber liquors. Fly ash is the probable source of
iron in the solid. Although most of the fly ash is removed in the pre—
wash section, some passes through the venturi and is removed in the
absorber. The mist carryover from the prewash section also contains
dissolved iron along with other metals.
A small quantity of the solids in the product liquor may be
sufficient to adversely affect the production of crystalline ammonium
sulfate in the regeneration section. Also the solids contain iron which
may catalize the decomposition of ammonia during thermal decomposition
of ammonium sulfate. Attempts to remove the solids by filtration were
unsuccessful because the precipitated solids and fly ash form a gelatinous,
thixotropic material (Figure 14) that blinded the filter media. The
material is removed from the product liquor in the pilot plant by set-
tling in the product storage tanks and then decanting the clear super-
natant liquor for use in the regeneration section. This method of solids
separation is acceptable for the pilo1>-plant operation but may prove
impractical for a commercial operation.
1093
-------
o
<£>
JS
Figure 1J. Plume from operation with product liquor having CA = 14^ four
absorber stages preceded by a water wash stage; 155°F reheat; % opacity.
-------
Figure 14. Photomicrograph of solids precipitated is the absorber loop
mixed with fly ash.
-------
Oxidation of sulfite to sulfate occurred in the absorption
section. In the range of liquor concentrations tested, the oxidation
ranged from 10 to ~L% of the absorbed S02. An average value of lyjo
was obtained over an extended operating period using the lj—stage
absorber. This oxidation byproduct would have to be purged from a
closed—loop system and probably sold as a fertilizer.
An equipment inspection of the absorption tower was made to
locate the areas in which corrosion is a problem.
The absorber shell, shown in Figure 15, is a combination of
six it—foot sections salvaged from various absorbers. Data pertinent
to each section are:
Section
Description
Transition piece for gas
inlet and liquor outlet
Section between first and
second absorption stages
Section between second and
third absorption stages
Section between third and
fourth absorption stages
Section between fourth
stage and demister
Demister housing
Material of
Exposure, construction,
hr Type stainless steel
14,000 316
10,000 304
k,000 304
k,000 304
,ooo 316
10,000 304
The first section, which has been in service 1^,000 hours, had
only a few small pits. The second and sixth sections, exposed for 10,000
hours, were severely pitted. The third and fourth sections, exposed for
4000 hours, contained very few pits. The fifth section was used for
about 13,000 hours as an exhaust duct after the demister and for about
1000 hours as a housing for an absorbing element. The walls of this
section were severely pitted; most of the pitting occurred in the location
above the demister.
The tower contains four Type 316 stainless steel absorption
trays (valve trays) spaced 4 feet apart. The bottom tray (G-l) was used
for about 3100 hours, and the three upper trays (G-2, 3, and 4) were used
about TOO hours. The bottom side of G-l was pitted, but the top side
showed little corrosion. The support angles located below the tray
(Type 316 or 304) were severely pitted. Some of the internal piping
1096
-------
LIQUOR
INLET
LIQUOR
INLET
LIQUOR
INLET
FLUE GAS
FROM VENTURI
SECTION(V-I)
SECTION
6
SECTION
5
G-4
-(VALVE TRAY)
SECTION
2
CHEVRON MIST
ELIMINATOR
G-3
(VALVE TRAY)
G-2
(VALVE TRAY)
LIQUOR
OUTLET
G-l
(VALVE TRAY)
Figure 15. Absorber shell.
1097
-------
(probably Type 304) to the tray was severely pitted while other sections
(Type 316) were only slightly affected. The temperature of the liquor
normally in contact with the bottom tray was 125°F (52°C) with an average
pH of 5.8 and a CA value of 8-20.
The bottom tray was used for about 1200 hours of its total
exposure time, as a water wash stage and for humidification of the inlet
flue gas. During this time, the operating conditions were similar to
the operating conditions of the venturi. Most of the corrosion probably
occurred during this time.
The second, third, and fourth trays (G—2, 3, and 4), supports,
and internal piping showed no signs of corrosion after the 700 hours of
operation. The temperature of the absorbing liquor on these stages was
approximately 120°F; the pH value ranged from 5. h to 7. 0.
The chevron mist eliminator is located in the uppermost 4-foot
section of the absorber. This section of the absorber has been in
service about 10,000 hours and was corroded severely. Severe pitting,
JO mils deep had occurred where drainage from the mist eliminator flowed
down the walls. The top and bottom of the mist eliminator blades were
badly pitted.
The transition piece (Type 3161 stainless steel) that joins
the top of the absorber to the I.k—inch— diameter flue gas exit duct has
been in service for about 700 hours with no apparent corrosion.
Conclusions drawn from the equipment inspection are:
* Type 316 stainless steel is not suitable for use in the
prewash (venturi) section.
• Type 316 stainless steel is suitable for service in the
absorber shell and in the internal piping.
• Use of Type 30^ stainless steel, which pitted badly in
the absorber shell and piping when exposed for 10,000
hours, is marginal.
• The demist section is subject to severe corrosion, and use
of both Types 30^ and 316 stainless steel in this area is
marginal.
Regeneration Section
A flowsheet of the regeneration section of the pilot plant
is shown in Figure 16. The absorber product liquor is fed to an acidu-
lator. Here the liquor is mixed with an acid ion to chemically release
the S02. A portion of the released S02 is evolved in the acidulator.
In past operation, the acid ion was supplied by plant grade sulfuric
acid. In an intergrated ABS system, the acid ion will be supplied by
ammonium bisulfate generated by thermal decomposition of ammonium sulfate.
1098
-------
SULFURICACID
STORAGE TANK
F-IO
PRODUCT LIQUOR
FROM
ABSORBER SECTION
F-5
COOLING
WATER
TEAM
D~2 r*
w
IACIDULATOR
I
DRAIN
EVAPORATOR
CRYSTALLIZER
FILTER
E-l
AMMONIUM
SULFATE
STEAM
STRIPPER
STRIPPING
*- GAS
F-7
Figure 16. Regeneration section.
1099
-------
The liquor from the acidulator overflows into a stripper where
the remaining S02 is stripped from the liquor with a countercurrent flow
of air.
The stripped liquor (ammonium sulfate solution) is metered
into the evaporator-crystallizer where water is evaporated to produce
a slurry of ammonium sulfate crystals. The crystals are separated from
the mother liquor in a continuous belt filter and dried. The crystals
will be fed to the ammonium sulfate decomposer when it is available.
A discussion of the individual steps follows.
Acidulation_ and Stripping. The acidulation and stripping
vessels are shown in Figure 17. The acidulator is constructed from a
6-foot section of a 12—inch stainless steel pipe and is coated internally
with DuPont's TFE Teflon for corrosion protection. A cone mixer, located
at the top of the acidulator, is used for mixing the absorber product
liquor with sulfuric acid. The acidulator is mounted so that the point
of gravity overflow to the stripper can be raised or lowered to vary the
retention time of the material in the acidulator.
The stripping vessel was constructed of a 6—foot length of
12—inch stainless steel pipe and the inside was coated with Teflon. The
stripper contains a ij—foot section of packing (polypropylene Pall rings).
Acidulated liquor enters the top of the vessel and flows countercurrent
to a stream of stripping gas (air) entering the vessel near the bottom.
The S02 stripped from the liquor joins the S02 released in the acidulator
and is routed back to the power plant ductwork. The stripped liquor
(ammonium sulfate solution) is then fed to the evaporator crystallizer.
Typical data from a series of acidulation and stripping opera-
tions are shown in Table 3. In these tests, liquor from the absorption
section having a C. value of l4. 7 was metered to the acidulator. Sulfuric
acid (93$) was also metered to the acidulator at a rate to give an acid
ion to ammonia ion ratio of 1. 5. The liquor and acid were at ambient
temperature (about 80°F). Stripping gas (air) also at ambient tempera-
ture was used to remove the chemically released S02 from the solution.
The airflow rate was 10 cfm (30 cfm/gal of solution). Under these
conditions, 96$ of the S02 in the ammonium sulfite — bisulfite liquor
to the acidulator was removed during the acidulation and stripping
operation.
The S02 remaining in the acidulated and stripped liquor is
released in the evaporation-crystallization step and leaves the process
with the steam from this equipment. Because this is a potential pollu-
tion problem additional efforts are being made to further decrease the
S02 remaining in the acidulated and stripped liquor. New designs for
both the acidula_tor and stripper are being considered to improve the SO?
release from this operation.
The acidulator and stripper were inspected after l800 hours
of operation. The Teflon coating in the acidulator had failed and was
1100
-------
SULFURIC
ACID
PRODUCT -
LIQUOR FEED
•*c
ACIDULATOR
(D-2)
1
GAS OUTLET
»
V
i
ACIDULATED
LIQUOR
OUTLET
•*• S02 TO POWER PLANT
DUCT SYSTEM
GAS OUTLET
T
STRIPPER
(D-3)
STRIPPING GAS
ACIDULATED-
STRIPPED LIQUOR
Figure 17. Acidulator-stripper.
-------
TABLE 3. TYPICAL REGENERATION LOOP OPERATION
AMMONIUM BISULFATE PILOT PLANT
Test Conditions
Acidulator
Liquor feed in
CA Iho7
s/cA o. 81
pH 6.2
Specific gravity 1. 2kk
Flow rate, gpm 0. 3
Sulfuric acid
Flow rate, gpm 0. 07
Percent sulfuric acid 93
Stoichiometrya 1.5
Liquor flow out
pH 1.8
Specific gravity 1.2kh
Percent S02 release 90
Stripper
Stripping gas
Type gas Air
Flow rate, cfm at 70°F 10
Liquor flow out
pH 1.9
Specific gravity 1. 2^0
Percent S02 release 59
Overall % S02 release 96
Q
Stoichiometry is the ratio of acid ion to mols WH3 as ammonium
bisulfite and ammonrum sulfite.
1102
-------
peeling from the cone mixer and acidulator walls. The exposed stainless
steel bottom of the cone mixer was severely damaged. The bare stainless
steel acidulator walls were coated with a rust— colored deposit.
The stripper was found to be in good condition. Ho peeling
of the Teflon coating was noted and the Pall rings showed no signs of
wear.
Evaporator-Crystallizer. The evaporator— crystallizer used
to produce ammonium sulfate from the acidulated and stripped liquor
is ^ JO" feet high, 2 feet in diameter at the bottom., and has a rated
capacity of J50 pounds of ammonium sulfate crystals per hour. The entire
unit is constructed of Type Jl&L stainless steel.
The crystallizer was received and installed near the end of
the fiscal year and most of the operating time was spent in getting the
equipment to perform according to specifications. Limited tests after
operation was established indicated that the evaporator— crystallizer
was adequately sized to remove water from the liquor at the highest
flow rate anticipated during the test program. Crystals of ammonium
sulfate produced in the crystallizer were sent to a belt filter.
Ammonium sulfate removal was accomplished by use of a belt filter
and a product dryer which removed the approximately 150 pounds per hour of
ammonium sulfate crystals produced in the crystallizer. The belt filter
was borrowed from another pilot plant and proved to be greatly oversized
for the ABS study. The belt has an effective filtration area of 10
square feet and was operated at its lowest operating speed of about 2
feet per minute. Intermittent operation was required even at the
lowest belt speed in order to build a cake thick enough to maintain a
vacuum in the system. Hinderance from ferrous ammonium sulfite, the
subject of earlier concern, did not materialize and satisfactory
filtration rates were obtained.
The crystals were dewatered from about 80$ to about % water on
the belt and were then fed to a propane gas— fired dryer (l ft dia by 12 ft
long). Product from the dryer contained less than 1$> water and was free
flowing.
Figure l8 is a photograph of a typical group of crystals.
Approximately 70$ of the crystals were retained on a 35-mesh screen.
Ammonium Sulfate Decomposer. The ammonium sulfate decom-
poser will be used to decompose the sulfate crystals to ammonia and
ammonium bisulfate. The ammonia will be returned to the absorption
section and the ammonium bisulfate to the acidulator.
The preliminary project plans for a completed ABS process
included a thermal decomposer to be supplied by an outside company.
However, the piloting and scaleup of that decomposer which uses direct
combustion of petroleum products as the heat source has not been
completed. The plans have now been modified to include an electrically
heated decomposer. A preliminary design for the pilot unit has been
completed using the design criteria listed on page J>k .
1103
-------
Figure 18. Crystals of ammonium sulfate produced in the ABS process;
one scale division = 1 millimeter.
-------
Capacity e* 2 MW
Ammonium sulfate feed rate 200-250 Ib/hr
Melt temperature 650°-750°F
Melt volume 12-lU cu ft (1500 Ib)
Flow rate of melt 2 cu ft/hr (0.25 gpm)
Steam feed rate ^0-75 Ib/hr at process temp.
Off-gas rate &. 3000-5000 cu ft/hr
Net electrical power input ^5 kW ( ^.550 kWh/ton A/S
to process decomposed)
The design of the decomposer is based primarily on information supplied
by Plancor 1865, a United States Government Document9. The antici-
pated delivery date for the unit is fall 1975. Once the decomposer is
installed and is on stream, the ABS system will be operated on a cyclic,
closed-loop basis to determine long—term operating characteristics of
the intergrated process.
Future Studies
Results of past operations have shown that the ammonium
bisulfate process is a promising candidate for second generation S02
removal systems. Continuing pilo1>-plant activities will be directed
toward long—term operation of an intergrated ABS system and demonstrated
performance of 9C$+ S02 removal. The approach to this objective will
include:
• Closing the water balance around the gas prewash section
and reducing carryover into the absorber.
• Further reducing the plume emissions from the absorber.
• Improving the acidification and stripping steps to obtain
a complete release of S02 using ABS melt for acidulation.
• Improving the solids separation from the absorber product
liquor.
« Studying ammonium sulfate crystallization (energy required,
crystal growth, crystal size, etc.).
1105
-------
CONVERSION OF ENGLISH UNITS TO METRIC EQUIVALENTS
Multiply To Obtain
English Unit By Metric Equivalent
ft3/min 0.0283 m3/min
Ib O.k$k kg
°F °C = | (°F - 32) °C
gr/ft3 2.288 g/m3
in 2.54 cm
ft 0.30*t7 m
gal/min 3-785 1/min
1106
-------
REFERENCES
Mo.
1 Lepsoe, R. and Kirkpatrick, W. S. "S02 Recovery at Trail, "
Trans. Can. Inst. Mining Met. XL, 399-l;o4 (1937).
2 Hein, L. B., Phillips, A.B., and Young, R.D. "Recovery of S02
from Coal Combustion Stack Gases. " In Problems and Control of
Air Pollution (Frederick S. Mallatte, ed), Reinhold, New York
(1955) PP 155-69.
3 Sulfur 80 (1), 36-37 (Jan.-Feb. 1969).
h Rumanian Ministry of Petroleum Industry and Chemistry. "Ammonium
Sulfate" Brit. Pat. 1,097,257 (Jan. 3, 1968).
5 Johnstone, H. F. "Recovery of S02 from Waste Gases." Ind. Eng.
Chem, 29 (12), 1396-98. (Dec. 1937).
6 Tennessee Valley Authority, "Pilot-Plant Study of an Ammonia
Absorption - Ammonium Bisulfate Regeneration Process, Topical
Report Phases I and II. " Report Y-83, Prepared for U. S. Environ-
mental Protection Agency (EPA), Environmental Protection Technology
Series, EPA-650/2-71)-049-a (June 197*0.
7 Hollinden, G. A. , Moore, N. D. , Williamson, P. C. , and Denny, D. A.
"Removal of Sulfur Dioxide from Stack Gases by Scrubbing with
Ammoniacal Solutions: Pilot—Scale Studies at TVA." Proceedings:
Flue Gas Desulfurization Symposium (1973) PP 961—96, Environ-
mental Protection Technology Series, EPA-650/2-73-038 (December
1973).
8 Hixon, A. W. and Miller, R. "Recovery of Acidic Gases." U.S.
Pat. 2,405,7^7 (Aug. 13, 19^6).
9 The Chemical Construction Company, Plancor 1865, "Alumina—From-Clay
Experimental Plant (Plancor 1865) at Salem, Oregon" (19^3).
1107
-------
DESCRIPTION AND OPERATION OF THE STONE & WEBSTER/IONICS
S02 REMOVAL AND RECOVERY PILOT PLANT AT THE
WISCONSIN ELECTRIC POWER COMPANY, VALLEY STATION, IN MILWAUKEE
K. A. Meliere and R. J. Gartside
Stone & Webster Engineering Corporation
Boston, Massachusetts 0210?
and
W. A. McRae and T. F. Seamans
Ionics, Incorporated
Waltham, Massachusetts 0215*4-
ABSTRACT
The first phase of the EPA-WEPCO-sponsored
program to evaluate the S&W/Ionics closed cycle
S02 removal system has been completed. The
technical feasibility of the process was demon-
strated in a 2000 ACFM pilot plant. Average
S02 removal was 85-95$ at exit concentrations
of 200-300 ppm S02. Oxidation of S02 in the
absorber varied from 7-25$ with a maximum error
of measurement of 1%. Other pilot plant results
and technical considerations in the design of a
prototype plant are discussed.
A brief summary of some of the operating
problems encountered is presented. Estimated
annual operating costs Exclusive of maintenance
and fixed charges) for a typical 500 MW power
plant are presented.
prepared for
i/^vironmental Protection Agency
Flue Uas Desulfurization Symposium
Atlanta, Georgia
7 November
1109
-------
DESCRIPTION AND OPERATION OF THE STONE & WEBSTER/IONICS
S02 REMOVAL AND RECOVERY PILOT PLANT AT THE
•WISCONSIN ELECTRIC POWER COMPANY, VALLEY STATION, IN MILWAUKEE
1. INTRODUCTION
The Stone & Webster/ Ionics S02 Removal and Recovery Process Is based
on absorption of sulfur dioxide in aqueous caustic which is subsequently
regenerated in Ionics' SULFOMAT™ electrolytic cells (electrolyzers) .
This patented process is applicable to gaseous effluents from station-
ary power plants burning fuels containing sulfur and to tail gases from
sulfur recovery plants, smelters and sulfuric acid plants.
The process consists of three essential steps arranged in a closed
loop:
a. Sulfur dioxide (802) is absorbed from flue gas in an aqueous
caustic soda (NaOH) -sodium sulfate (Na2SOi|) solution to pro-
duce aqueous sodium bisulfite (NaHSOj) containing some sodium
sulfite
b. This aqueous bisulfite- sulfite is mixed with dilute sulfuric
acid resulting in formation of aqueous sodium sulfate and
gaseous, wet sulfur dioxide. The latter can be recovered for
sale as commercially pure S02> converted to commercial 66°
Baurae sulfuric acid or to elemental sulfur.
c. The aqueous sodium sulfate is converted by electroylsis into
caustic soda and sulfuric acid. The caustic is recycled to
Step 1; the sulfuric acid is recycled to Step 2.
The process can be summarized by the following reactions:
Step 1, Absorption
2 NaOH + S02 -* Na2S03 (l)
Na2S03 + S02 + H20 — * NaHS03 (2)
Step 2, Recovery
Na2S03 + H2SOl^ — > ^SOl^. + H20 + S02 (3)
2NaHS03 + H2SOU — > Na2SOl; + 2H20 + 2S02 (U)
Step 3, Electrolytic Regeneration
2 Na2S01f + 6 K20 -> U NaOH + 211230^ + 2H2 + 02 (5)
1111
-------
The primary side reactions are absorption of sulfur trioxide (803) and
the oxidation of sodium sulfite and/or bisulfite. Either reaction
forms sodium sulfate directly from which sulfur oxides vrill not be re-
leased upon acidification. These side reactions, therefore, result in
a sulfate ion increase in the system. Special electrolyzers are used
.to purge excess sulfate as pure, aqueous sulfuric acid (10% wt). The
quantity of such acid is quite small compared to the amount of sulfur
dioxide recovered. It may be used in the power plant,for example, for
regeneration of ion exchange deionizers.
Under a joint program cosponsored by the Environmental Protection
Agency and Wisconson Electric Power Company, the process has been
tested at Wisconsin Electric's Valley Station in Milwaukee. Phase I
of a planned three-phase demonstration program has been completed. It
consisted of design, installation and operation of an integrated pilot
plant, development of a full-scale electrolyzer system and preliminary
design of a prototype system. It is planned that Phase II of the pro-
gram will include design, procurement and installation of a prototype
facility treating all the flue gas from one of the four 75 MW coal
fired boilers at the station. Phase III would involve startup and
long-terra operation of the prototype facility. A flow scheme is shown
in Figures 1 and 2 for the 75 MW plant.
2. PROCESS DESCRIPTION - PILOT PLANT
2.1 Absorption
Entering flue gas was cooled from 290° + 20°F to 120°F by direct water
quench in the bottom of the absorber, after which the gas was contacted
with aqueous caustic containing sodium sulfate and about 8% sodium
hydroxide. Flue gas was returned to the stack from the top of the
absorber. The caustic was converted to an aqueous mixture of sodium
bisulfite, sulfite and sulfate containing the S02 and 803 removed from
the flue gas. Above the quench section there were three packed sections
each 10 feet high, that contained 2 inch Tellerettes. For most of the
test program the uppermost stage was used as a demister. Absorbent
was recirculated around each absorbing stage.
The flue gas rate was varied from 1*500 to about 9000 pounds per hour.
SOa concentration varied from about 1000 to about 3600 ppm. Percentage
removal varied from 85 to 95$ at effluent gas concentrations containing
200-300-ppm S02.
2.2 Sulfur Dioxide Recovery
Prior to entering the stripper, net effluent from the absorber was
reacted with the sulfuric acid-sodium suli'ate mixture recycled from the
1112
-------
STACK
STRIPPER
QUENCH
CIRCULATION
PUMP
ABSORBER CIRCULATION
PUMPS
STRIPPER
REFLUX PUMP
Figure 1 S02 removal section.
Stone & Webster/Ionics S02 Removal & Recovery Process
-------
ELECTROLYTIC CELL SYSTEM
"A" SANK "B'BANK
CELL FEED TANK
RtCYCLt CAUSTIC-
K
AM
"A" BANK
ANOLYTE
DRUM
(
"7—
—
i
30C CAT
I I
0 M
J
<0« CAT
«-
ANOOt CATHOOC
-
AN
I I
0 M
I I
J
—
)OE CATHODC
I I
0 M
I I
J
^
"A"/ "8" BANK
CATHOLYTE
DRUM
V BANK
ANODE
COOLER
\
r
I
-^
1
CAT>
CAT*
400E [ AHC
I I I
MOM
I I I
J L»
•OOt 1 AHC
I I
MOM
I I 1
J U,
DC
JOE
—
one ( ANODE
1 1 1
MOV
1 1 1
J L
"A"/"B"BANK "a" BANK
CATHODE ANODE
COOLER COOLER
£
J)
-B'B
ANOL
DR
1
1 r
ANK
YTE
JM
]
c
A
R
0
^1
1
W
CIDSULFA1
ECYCLE
RUM
E
1
^
CAUSTIC
ri RECYCLE
f ORUM
LJ
101, ",10, >«!!««
^
CELL FEED VCELL ANODE
PUMP CIRCULATION
PUMP
"A'/'B" CELL CATHODE
CIRCULATION
PUMP
"B" CELL ANODE ACID SULFATE CAUSTIC
CIRCULATION RECYCLE RECYCLE
PUMP PUMP PUMP
Figure 2 Regeneration system.
Stone & Webster Ionics SOg Removal & Recovery Process
-------
electrolyzers. The following reactions occurred in situ;
NaS03 + H2S01,. - >.' HgO + SOa + Na2SOl}. (£)
2NaHS03 + H2SO^ - > 2H2.0 + 2S02 + Jfe^SOl^ (7)
The operating pressure was 10-15 psig. The reboiler' temperature was
2^0-250°F. Recovered S02 in the stripper overhead was at a purity of
98$(vol) or greater. The stripper bottoms stream was controlled to a
pH of about 3.5 and contained 50-100 ppm of dissolved S02.
2.3 Absorbent Regeneration and Oxidation Product Rejection
Aqueous sodium sulfate recovered from the stripper bottoms was adjusted
to about pH 8.5 and hydrogen peroxide was added to precipitate iron as
ferric hydroxide. After filtration, sodium sulfate was sent to two
types of electrolyzers (as described in 2.U.3). In each type of elec-
trolyzer, sodium hydroxide containing sodium sulfate was generated and
was recycled to the absorber. A mixture of sodium sulfate and sulfuric
acid was also generated and was recycled to the sulfur dioxide recovery
step described above. Oxygen was generated as pure humid gas at the
anodes of each electrolyzer. Pure humid hydrogen was generated at each
cathode. Both gases were diluted with air and vented to atmosphere.
In the four-compartment ("B" type) electrolyzers, anode product was
substantially pure, approximately 10% sulfuric acid equivalent in
quantity to the SO^ absorbed and S02 oxidized. In a commercial- scale
plant this sulfuric acid could be withdrawn from the system for regen-
eration of mixed bed ion exchange deionizers, for marketing or for
other disposal.
In the pilot plant the volume was small and the sulfXrric acid was not
recovered. During the initial operations of the pilot plant, the
sulfuric acid sometimes contained substantial amounts of sodium sulfate
due to upsets in cell operating conditions.
Sulfate ion so removed as dilute acid is a preferred method by which
803 contained in the entering flue gas plus S02 oxidized to sulfate in
the absorber are removed from the system. The total amount of sulfate
formed in the absorber determines the required number of four- compart-
ment electrolyzers and the amount of acid produced. In the pilot plant,
36$ of the electrolyzers had four compartments.
2.U Technical Considerations
2.^.1 Absorption., The overall process efficiency of sulfur
dioxide removal from the flue gas is determined by three factors;
they are:
a. The ratio of sodium sulfite to bisulfite in the net effluent
liquor.
1115
-------
b. The amount of oxidation of sodium sulfite and/or "bisulfite to
sodium sulfate that occurs in the absorber, and
c. The amount of recycle caustic required for electrolyzer feed
treatment.
Each of these factors can significantly affect the efficiency of the
process.
The ratio of sodium sulfite to bisulfite is conveniently expressed in
terms of the S/C ratio. The S/C ratio is defined as the ratio of
moles of sulfur as sulfite and bisulfite to the moles of sodium asso-
ciated with this sulfur. This ratio varies from 0.50 (100$ sodium
sulfite) to 1.00 (100$ sodium bisulfite). The higher the S/C ratio,
the lower the required amount of caustic to effect a given sulfur
dioxide removal.
Oxidation is defined as the amount of sulfur dioxide absorbed as
sodium sulfate compared to the total sulfur dioxide absorption.
Absorption as sodium sulfate will consume two moles of caustic per
mole of sulfur dioxide.
Sulfur dioxide absorbed as the sulfate will not be released by acid-
ification with sulfuric acid and will result in a net sulfate ion
increase in the system. The amount of oxidation will set the number
of "B" cells required for removal of excess sulfate ion. "B" cells
require more capital investment than "A" cells and consume slightly
more power. This provides further incentive for minimizing- oxidation.
One of the primary goals of the test program was to study the influ-
ence of both operating and chemical parameters on the amount of
oxidation.
The amount of recycle caustic required for feed liquor pH adjustment
is determined by the extent to which acid addition at the stripper
exceeds the stoichiometric requirement.
A. Caustic Utilization
Overall caustic utilization is defined as the ratio of the moles
of sulfur dioxide absorbed per mole of caustic generated. A
practical maximum caustic utilization is 0.90 moles of sulfur
dioxide removed per mole of caustic generated. This is based
upon a maximum S/C.ratio of 0.925 in the absorber draw, a min-
imum caustic recycle of 3«2$> and assuming no oxidation.
The maximum S/C ratio is set by the equilibrium vapor pressure
of S02 over the sodium sulfite-bisulfite-sulfate solution and
is a function of the temperature and pressure of the system.
The minimum caustic recycle required occurred at a stripper
bottoms pH of about 3«50« Operation at this pH represents a
1116
-------
"balance between residual S02 in the. stripper bottoms and the amount
of excess acid added above the stoichiometric requirement.
The actual overall caustic utilization for the test program
averaged about 0.?8 or 86.9$ of the practical maximum. This
corresponds to an S/C ratio of 0.925, a caustic recycle of
5.0$ and an oxidation level of 13.0$.
B. Oxidation
Oxidation levels in the absorber varied from 7 to 25$ over the
test program with the majority of measured values under 15$. The
maximum uncertainty in this result was such that the best we
could determine oxidation to was in the amount of 7%. That is,
results indicating oxidation in the amount of 7% could also be
interpreted as oxidation in the amount of 0$.
In general, oxidation was found to be affected by:
a. The amount of liquid recirculated to each stage. Higher
rates increased oxidation.
b. The (S/C) of the effluent liquor. High S/C's favored
lower oxidation.
c. The concentration of S02 in the inlet gas. Higher SOg
partial pressure favored lower oxidation.
Additionally, varying the concentration of sodium sulfate did
not appear to repress the amount of oxidation.
2.h.2 StLLFur Dioxide Recovery. The variable of primary concern
in the stripping section is the minimum steam rate for a given S02 con-
centration in the bottoms. The concentration of S02 in the bottoms is.
determined by:
a. Stripping Steam Rate
b. Acidity
c. Temperature
d. Pressure
e. Number of Stripping Trays
There is an optimum bottoms concentration of S02 that is determined by
economics. For example, recycling more acid will save steam by repres-
sing the ionization of sulfurous acid. However, this also increases
the downstream recycle caustic required for adjustment of the cell feed
liquor acidity. Lowering the system pressure also enhances the strip-
ping operation. However, if the bottoms temperature becomes too low,
then heat exchange between feed and bottoms becomes impractical.
1117
-------
Heat exchange between feed and bottoms was not practiced in the pilot
plant. Also, the reboiler and overhead condenser were generously sized
making it difficult to control the stripping operation at low steam
rates. Therefore, pilot plant results are not valid for extrapolation
to a commercial scale.
For the S02-H20 system at 10 psig and 50 ppm of SOg in the stripped
solution and with heat economy, a stripping steam rate of k Ib/lb S02
has been calculated. A minimum rate of about 2 Ib/lb S02 is required
for heating the feed stream.
However, the real situation is the system S02-H2SOl|-Na2SOl|-H20. The
amount of ionized S02 in an acid salt solution should be less than in
the S02-H20 system and would result in easier stripping.
Some data available for the solubility of S02 in 0.10 N IfeSOLt at 90°F
indicates about a 50$ decrease in S02 solubility. Thus, the minimum
steam rate would probably be in the order of 3 Ib/lb S02.
2.U.3 Absorbent Regeneration and Oxidation Production Rejection.
The unique feature of the Stone & Webster/Ionics process is the elec-
trolyzer system in which the caustic is regenerated. Reference has
been made above to use of two different designs, a three -compartment
electrolyzer and a four -compartment electrolyzer. The former is the
basic design that converts sodium sulfate into sulfuric acid and caustic
soda. The four-compartment design is the means by which excess sulfate
is removed from the recirculating liquid system as pure, dilute sulfuric
acid.
A schematic diagram of the three -compartment (Type "A") electrolyzer is
shown in Figure 3- The main components are an anode, a microporous
diaphragm, a cation-selective membrane and a cathode. These components
are separated from each other by flow directing spacers which also pro-
vide required gasketing.
Stripped sodium sulfate solution from the stripper was cooled, pH
adjusted to 8.5, and several ppm of hydrogen peroxide added to precip-
itate ferric hydroxide. The latter and other heavy metal oxide and
hydroxides were removed by filtration. The sodium sulfate was -fed to
the central compartment of the electrolyzer s at a concentration of
about 20^. Sodium ions migrated through the cation-selective membrane
toward the cathode under the influence of a direct current voltage
impressed across the electrolyzer. At the cathode, water was elec-
trolyzed to hydrogen gas and hydroxide anion:
H20 + e~ — > 1/2 H2 + OH" (8)
The hydroxide anions were electrically balanced by sodium cations
entering the cathode compartment through the cation selective membrane.
The effluent catholyte consisted of NaOH plus HaaSOli, which after dis-
engaging hydrogen was sent to a surge tank and then to the scrubber.
1118
-------
1/2 C
H2S04
Na2S04
©
2e —
D CM
1/2 02 |Na2S04
/__ 1
t
i
, •^•^•^ 5 04 = *^*
•*• 2H"*" '
1
0
H2
X
J.
^^ icNo
20H"""*1
1 III
l*-2e
NaOH
D" POROUS DIAPHRAGM
CM» CATION SELECTIVE MEMBRANE
N02S04
"A" Cell
H2S04
\/c U2 I\a2su4 n2
i ,
H2S04 •*-
* t ^ ^
© AM D CM ©
2e —
7
/
/
/
/
1/2 02
f
\
+ |Na2S04
*-H •*-;
c_. 1 c>/«i _
4 | 4
I
i
H2
^t
«». 4-
20 H~*-
— 2e
t ' t It
.
H20 — *• • A
D - POROUS DIAPHRAGM Na2S04
CM - CATION SELECTIVE MEMBRANE
AM = ANION SELECTIVE MEMBRANE
NaOH
Na2S04
"B" Cell-
Figure 3 Schematic diagram of "A" and "B" electrolytic cells.
1119
-------
The function of the cation-selective membrane was to prevent physical
mixing of the catholyte and center compartment feed streams.
Essentially, only sodium cations from the center compartment pass
through the membrane to combine with hydroxide anions produced at the
cathode. Water may be fed to the catholyte compartment in sufficient
quantity to produce 6 to 20$ caustic. However, the Stone & Webster/
Ionics process is an essentially closed system. Therefore, to maintain
water balance in the system it was necessary to feed recycled sodium
sulfate solution to the catholyte compartment instead. This procedure
does not introduce any problems and was the source of the sodium sul-
fate in the scrubber feed.
The center compartment feed passed through the microporous diaphragm,
into the anode compartment. At the anode, water was electrolyzed to
hydrogen cations and oxygen:
H20 ?• ^02 (s) + 2H+ + 2e~ (9)
The hydrogen cations combined with sulfate anions to form sulfuric acid.
The diaphragm flow was designed to prevent hydrogen ions from migrating
across the cation-selective membrane. Such flow must give enough
linear velocity through the diaphragm to sweep hydrogen ions back into
the anolyte compartment. The diaphragm should have an hydraulic re-
sistance adequate to insure that such flow is substantially uniform
over the entire surface. A flow which is sufficient to sweep back most
of the hydrogen ions carries with it about half of the sodium in the
center compartment. Thus, at the anode, only part of the sodium sul-
fate in the feed stream was electrolyzed to sulfuric acid. The anode
product was therefore a mixed solution containing both sulfuric acid
and sodium sulfate. This product was the dilute acid solution that was.
recycled to release S02 in the stripper.
In the four-compartment (Type "B") electrolyzers, a schematic diagram
of which is presented in Figure 3, the reactions and operation of the
catholyte compartment were exactly those described for three compart-
ment electrolyzers. Catholyte effluents from both electrolyzer types
were combined before being sent to the absorber. The center compart-
ment feed flowed through the porous diaphragm, into the "mid-anolyte
compartment", from which it was discharged from the electrolyzer.
This stream contained some sulfuric acid due to the inability of the
anion selective membrane completely to exclude hydrogen ion. The
mid-anolyte effluent was combined with the anolyte stream from the
three-compartment electrolyzers. The unique reactions of the four-
compartment electrolyzer occurred at the anode, where water was elec-
trolyzed to hydrogen cations and oxygen gas. These hydrogen cations
combined with sulfate anions entering the anode compartment to form
sulfuric acid. The function of the anicn-selective membrane was to
exclude sodium from the anode compartment while allowing sulfate to
enter. It also, to a considerable extent, prevented hydrogen cations
from leaving the anode compartment. Thus, the anode reactions per-
mitted removing sulfate without losing sodium.
1120
-------
Each electrolyzer consisted of an anode, a cathode, at least one
membrane, and one diaphragm with appropriate separators and internal
flow distributors. A second membrane was placed between the diaphragm
and the anode in the Type "B" electrolyzers. Fluid was internally man-
ifolded and was in parallel. The electrolyzers were arranged into a
number of modules within a single structural frame. The electrolyzers
within a module were in parallel electrically and the modules were in
series. The internal fluid manifolding was brought to headers in the
frames. Headers were provided with quick-disconnect fittings to the
electrolyzer fluid distribution and collection mains.
A large number of materials of construction have been tested in order
to optimize life and performance of electrolyzer components. Both lead
alloy and noble metal coated titanium anodes were tested. The latter
were found to be more cost-effective. Preferred cathodes were nickel
plated carbon steel and fluorcarbon fabric backings were preferable to
acrylic backings for ion selective membranes. The life of the ion
exchange resin component of such membranes appears to be entirely ade-
quate. During the initial operations of. the pilot program problems
with other equipment in the system heavily loaded the electrolyzers
with corrosion products and other atypical insoluble materials such as
magnesium hydroxide and calcium carbonate. Such problems were elimin-
ated in due course. We were pleased to observe that the electrolyzers
readily returned to normal operation after such upsets.
Commercial-scale electrolyzer components, tested at Ionics in
Watertown, Mass., were designed to be even more rugged. This scaling
up has been accomplished with a reduction in electrical energy con-
sumption. The specifications of the electrolyzers are essentially as
follows:
"A" Cells "B" Cells
Effective area per cell 5 sq. ft. 5 sq. ft.
Operating Temperature l6o°F ^O^
Effective Cell Height, inches 20 20
Interelectrode Distance, inches 0.317 0.3^5
Current Efficiency, overall 85$ 85$
Current Efficiency, pure acid M?$
2.5 Operating Experience
The vast majority of the operational problems experienced at the test
site were mechanical in nature and typical of any pilot plant opera-
tion. Once these problems were resolved the process had an overall
operation availability of greater than 90$ during the latter stages of
the test program.
Forced draft fan vibrations were a frequent cause for shutdown. This
was finally traced to a buildup of acid sludge on the blades occurring
1121
-------
only at gas inlet temperatures "below 225°F. Water washing the fan
"blades and flue gas temperature control reduced the vibrations! prob-
lems to a manageable level.
During the initial phases of the test program, the extent of feed
liquor cleanup required for smooth electrolyzer operation was not well
understood. Improper liquor cleanup caused a number of shutdowns until
an adequate system was installed.
Pressure control on the stripper overhead was hampered by sulfur
dioxide hydrate formation in the instrument lines, the flow measuring
orifice, condenser tubes, or a combination of all three. This problem
is correctable by maintaining temperatures above 6o°F at all points in
the system.
Absorption liquor entrainment losses were a problem when operating with
three packed stages. The high pressure drop nozzles initially
installed in the liquor recirculation lines produced a fine atomizing
spray which in combination with the relatively high gas velocities was
responsible for the entrainment. Replacement of the spray nozzles and
operating with the third packed stage as a demister corrected the
problem.
The system suffered from a continual water loss throughout the test
'program. This loss was traced to the evaporation of water from the
process liquids due to incomplete saturation of the flue gases in the
water quench section. This loss was in addition to the known water of
decomposition required for the electrolysis operation. A reasonably
constant feed liquor density and purity is required for good cell volt-
age control, hence continuous water addition to the cell feed liquor
is necessary.
3. 75 MW PROTOTYPE FACILITY
A 75 MW prototype facility has been designed. It is proposed that it
be installed at the Valley plant of Wisconsin Electric Power Company.
The basis of design is:
a. Process all the flue gas from one 75 KW boiler at 100$ load
factor, approximately 260,000 ACFM at 320°F.
b. Based on statistical analyses of recent coal'burn, the plant
will be designed to remove an average of 3100 pounds of S02
per hour on a continuous basis.
c. Sufficient "B" Type four-compartment electrolyzers will be
used to provide for a maximum oxidation quantity in the
amount of 15%.
1122
-------
d. Flue gas to be desulfurized "by a minimum of 90$.
e. 1152 "A" Type electrolyzers will be vised having a total
active area of 5760 square feet. 3Bh "B" Type electrolyzers
will "be used having a total active area of 1920 square feet.
U.O PROCESS ECONOMICS
Since many large boilers supplying steam for power generation, have
capacities of the order of 500 MW, we have developed operating costs
for a 500 MW plant instead of the smaller 75 MW plant proposed for
Milwaukee. We believe these costs are more representative of the
average power plant installation. Costs of utilities do not neces-
sarily reflect WEPCO costs, but are those generally used in assessing
costs of competitive processes.
Estimated annual operating costs, excluding fixed charges and
maintenance, are shown in Tables 1 and 2 for a coal fired plant assum-
ing 3.5% S coal, 90$ S02 removal and 80% boiler load factor. Table 1
assumes that the stack gas is reheated using the hydrogen generated by
the electrolyzers and low sulfur No. 6 fuel oil. Table 2 assumes no
stack gas reheat, but assumes the hydrogen generated is burned for its
heating value.
Fixed charges for the plant have not been included for two reasons.
First, because of the present high rate of inflation affecting equip-
ment, materials and labor costs, capital cost can be made to vary
depending on the amount of optimism or conservatism applied to the
plant cost estimate.
Second, there is still not agreement in the industry on the annual
fixed charge rate required to amortize the capital investment of pollu-
tion control facilities. Typically, utilities write off capital for
generating facilities at lh% per year. Such a policy has the effect
of increasing annual pollution control charges compared to other
methods of financing. For example, if the plant were instead financed
with 20-year 9$ bonds, annual debt service would be 11$, not l*t$. On
the other hand, if one took full advantage of special financing per-
mitted for pollution control facilities, then the effective annual
capital charge would be only about 7%. Such a rate assumes:
a. 100$ financing by industrial revenue bonds at 6$ interest
b. Straight line pay-backs of principal
c. Accelerated depreciation under Internal Revenue Service
d. 7% discount rate for all cash flows
1123
-------
TABLE 1 APPROXIMATE ANNUAL OPERATING COST WITH STACK GAS REHEAT
STONE & WEBSTER/IONICS S02 REMOVAL PROCESS
Basis: 500 MW; 1,355,000 tons per year of 3.5$ S Coal;
ft, Load Factor; 90$ S02 Removal
Quantity
260 T/Y
260 T/Y
27,500 KW
10,600 KW
68,000 Ibs/hr
112 GPM'
8,000 GPM
Unit Price
Total
$ 60
$ 300
$
15,600
78,000
62,^00
11 mils 2,115,000
11 mils 816,000
$ 1.50/1000 Ibs 715,000
$ 0.30/1000 gals 1H,000
$ 0.03/1000 gals 101,000
6 bbl/hr $ 12.10 bbl/hr 503,000
Utilities & Chemicals
Soda Ash
hydrogen Peroxide
Filter Aid
Electrieity-Electrolyzers
Electricity-Auxiliaries
Steam (100 psig)
Condensate
Cooling Water
No. 6 Fuel Oil, Low S
Subtotal, Utilities &
Chemicals
Operating Labor
Administration, Supervision
and Laboratory
Credit for S in S02
(Includes credit for 10$ sulfuric acid at sulfur value)
Total Operating Cost (excluding maintenance) $ U,290,000
Cost per KWH, (3-5 x 109 KWH) 1.2 mils
Cost per Ton of Coal (1,355,000 tons) $ 3.17
3/shift
11/hr
$ U,teO,000
$ 290,000
$ 150,000
38,000 LT/yr $15/LT Sulfur -$ 57.0,000
1124
-------
TABLE 2 APPROXIMATE ANNUAL OPERATING COST WITHOUT STACK GAS REHEAT
STOlffi & WEBSTER/IONICS S02 REMOVAL PROCESS
Basis: 500 MW, 1,355,000 tons per year of 3.5$ S Coal;
Load Factor; 90$ SOg Removal
Utilities &_Chemicals
Soda Ash
Hydrogen Peroxide
Filter Aid
Electricity-Electrolyzers
Electricity-Auxiliaries
Steam (100 psig)
Condensate
Cooling Water
Quantity
260 T/Y
260 T/Y
27,500 KW
10,600 KW
68,000 Ib/hr
112 GPM
8,000 GPM
Unit Price
$ 60
$ 300
11 mils
11 mils
$ 1.50/1000
$ 0.30/1000
$ 0.03/1000
Total
$ 15,600
78,000
62,1*00
2,115,000
8l6,000
Ibs 715,000
gals lU,100
gals 101,000
Subtotal, Utilities &
Chemicals
Operating Labor
Administration, Supervision
and Laboratory
3/shift
$H/hr
$ 3,917,000
$ 290,000
$ 150,000
Credit for S in S02 38,000 LT/yr $15/LT Sulfur -$ 570,000
(Includes credit for 10$ sulfuric acid at sulfur value)
Credit for H2
27 MM BTU/hr $1.59/MMBTU
(LHV)
Total Operating Cost (excluding maintenance)
Cost per KWH, (3-5 109 KWH)
Cost per Ton of Coal (1,355,000 tons)
-$ 300,000
$ 3,^7,000
1.0 mils
$ 2.57
1125
-------
Such' industrial revenue bonds and accelerated depreciation have become
very popular for financing pollution control plants.
5. CONCLUSION
Operation of the WEPCO pilot plant of the electroJytically regenerated
S02 removal and recovery process has been completed. The process has
been shown to be technically feasible and it has been demonstrated that
process reliability can be designed into the process to maintain the
continuity of operation required by the power generation industries.
The electrolyzers (electrolytic regeneration cells) performed somewhat
better than expectations and were found to be surprisingly forgiving
of upsets in ancillary systems and components. Problems with some com-
ponents and materials of construction in the ancillary systems were
found and solved during the pilot program.
Preliminary design of a 75 MW prototype system has been completed.
Experience from the pilot plant has been incorporated and it is
believed that the prototype will demonstrate the reliability re-
quired for power plant operation.
1126
-------
CALSOX SYSTEM DEVELOPMENT PROGRAM
PRESENTED AT THE
EPA FLUE GAS DESULFURIZATION SYMPOSIUM
November 4-7,1974 — Atlanta, Georgia
Authors
R. E. Barnard, Monsanto Enviro-Chem Systems, Inc.
St. Louis, Missouri
R. K. league, Monsanto Enviro-Chem Systems, Inc.
St. Louis, Missouri
G. C. Vansickle, Indianapolis Power & Light Company
Indianapolis, Indiana
ABSTRACT
The CALSOX system was developed by Monsanto to overcome the
scaling problems encountered by systems scrubbing SCb bearing
gases with lime/limestone slurries. The CALSOX system removes
the SO2 from the gas by absorption in an aqueous ethanolamine
solution, using a relatively simple gas/liquid contactor. Removal of
the sulfur values from solution occurs externally of the contactor by
precipitation with lime in a system designed specifically for this
purpose. After separation of the insoluble calcium salts by filtration,
the regenerated absorbent solution is returned to the system. Thus,
by minimizing precipitable calcium ions in the gas/liquid contactor,
scaling is prevented.
A 3,000 acfm pilot unit began operation in February 1973 in a joint
program with Indianapolis Power & Light Company and continued
until the end of October 1973. Design criteria originally established
for the process included:
SO
-------
CALSOX SYSTEM DEVELOPMENT PROGRAM
PRESENTED AT THE
EPA FLUE GAS DESULFURIZATION SYMPOSIUM
November 4 - 7,1974 — Atlanta, Georgia
PROCESS CHEMISTRY
The CALSOX1 system presents a new concept in SOa scrubbing of
boiler flue gases. The flue gas is scrubbed with an aqueous etha-
nolamine solution which absorbs the SOz, which when precipitated
with lime produces calcium sulfite and calcium sulfate. Two of the
unique concepts of the CALSOX system are the use of the etha-
nolamine absorption solution and the two step regeneration which
prevents calcium scaling in the absorber.
Ethanolamine was chosen as the alkaline agent for absorption be-
cause it allows complete regeneration with lime. Figure 1 shows the
ethanolamine neutralization curve for absorption of SOa. As indicated
on this curve the original pH of an aqueous ethanolamine solution is
11.3. Lime has a very high solubility over the entire neutralization
range. All the lime added will be utilized in precipitating sulfur
species up to the neutralization point.
After precipitation of the sulfur species as calcium sulfite and cal-
cium sulfate, the resulting solution will have an equilibrium concen-
tration of calcium and sulfate ions, approximately .02 molar. This
concentration of calcium ions is greater than can be tolerated in the
absorption system without the potential for scaling as a result of
calcium sulfite precipitation. For this reason the two step regenera-
tion is used to lower the calcium ion to the calcium sulfite solubility
and thereby eliminate the possibility of scaling in the absorption step.
After complete regeneration with lime the solution is mixed with the
scrubber return liquor in a reaction tank and the calcium ion is pre-
cipitated as calcium sulfite. The clear liquor from this step is then
used in the scrubbing operation.
The CALSOX process that applies this chemistry is presented in Fig-
ure 2. Flue gas from the boiler's existing induced draft fan flows to a
booster blower which directs the gas to a single stage absorption
system for removal of the S02. The cleaned gas is then reheated, if
required, to minimize steam plume formation, and then goes to the
stack. The rate of SOz removal can be controlled by changing the
operating pH of the absorber.
1128
-------
The SOa rich absorbent from the absorber then flows to a two step
regeneration system where the sulfite and sulfate ions are precipi-
tated as the calcium salts and the regenerated absorbent is returned
to the system. In the first step, part of the sulfite is precipitated as
calcium sulfite thereby lowering the soluble calcium level to below
50 parts per million and eliminating the possibility of scaling in the
absorber. In the second step the remaining sulfate and sulfite ions
are precipitated with lime; the calcium addition is controlled by pH,
Lime slurry is prepared by feeding quicklime from a storage silo and
absorbent solution from the regeneration system to a conventional
lime slaker. The slaked lime slurry is introduced on demand to the
second regeneration reactor.
The underflow slurry of calcium sulfite and sulfate crystals is fed
from the thickener to a vacuum filter for dewatering and washing.
The amount of wash water is controlled to equal the amount of water
lost by evaporation in the adiabatic cooling of the gas stream in the
absorber. The filtrate and wash water are returned to the process
providing a closed loop system. The separation of the absorption
and regeneration steps and the buffering capacity of the ethanol-
amine solution provide for good control of the entire system.
PILOT PLANT PROGRAM
In April 1972, Indianapolis Power & Light Company and Monsanto
entered into a joint program on the design, construction and opera-
tion of a 3,000 acfm pilot unit. This pilot unit was installed at the
Elmer W. Stout Generating Station in Indianapolis, Indiana and began
operating in February 1973. The operating and testing program con-
tinued until the end of October 1973, at which time it was concluded
that sufficient information had been obtained for the design of a
demonstration unit in the 100-125 Mw range. The design criteria
originally established for the process included: SOa removal, 90%
minimum; particulate emission, 0.02 grains per scf maximum; cal-
cium oxide usage, 120% stoichiometric maximum; absorbent usage,
2 Ibs. per megawatt hour maximum. Performance experienced dur-
ing sustained operation of the pilot plant system exceeded all the
preceding criteria.
Figure 3 is a flow diagram of the pilot plant. The pilot plant as con-
structed at Indianapolis Power & Light Company differed from the.
process description presented in Figure 2 in three areas. In the pilot
plant, gas was to be taken both before and after the electrostatic
precipitator. Therefore, a scrubber was installed ahead of the absorber
for particulate removal since it was not known how much particulate
the absorber could tolerate. The second change was that the pilot
plant used an ID fan instead of an FD fan. The fan was a 3600 rpm
overhung fan installed between the absorber and the reheater. The
1129
-------
other change in the pilot plant from the process design shown in
Figure 2 is the use of reactor clarifiers instead of reactors sepa-
rated from thickeners. These are the areas of importance in the
discussion of the problems encountered in the pilot plant and the
resolutions that will be used in the full scale design.
A Ventri Rod2 scrubber was used in the pilot plant for particulate
removal. It was operated at 10 inches water pressure drop with a
liquid to gas (L/G) ratio of 13.5 gallons per thousand standard cubic
feet. When the gas stream was taken before the electrostatic pre-
cipitator (but after a mechanical collector) the overail average ash
loading was 0.86 grains per standard cubic foot. The average outlet
loading of the scrubber was 0.036 grains per standard cubic foot.
This represents approximately 96% removal of the fly ash. There
was additional fly ash removal in the cross flow absorber. The aver-
age outlet loading from the absorber was 0.004 grains per standard
cubic foot. The fly ash removal across the absorber alone was ap-
proximately 87%.
The absorber was able to achieve 90% SOa removal under all con-
ditions even when the inlet SCb dropped to as low as 1500 parts per
million. This was with the absorber operating at an L/G ratio of 25
gallons per thousand standard cubic feet. When the liquid across the
absorber was reduced to 12 gallons per thousand standard cubic
feet the absorption efficiency was approximately 80%. The lime uti-
lization fluctuated in the regeneration reactor as a result of fluctua-
tion in the pH control. The average lime utilization was approximately
102% of stoichiometric, significantly better than the target of 120%.
Washing efficiency was good on the horizontal vacuum filter. The
absorbent utilization, represented by loss in the cake, averaged 0.5%
in the dry cake, which is equivalent to 0.6 pounds per megawatt hour,
a significant improvement over the target of 2 pounds per mega-
watt hour.
The operating program for the pilot plant was divided into three
phases. Phase I was the start-up and shake-down period from Feb-
ruary 6 to March 31, during this period the unit operated approxi-
mately 75% of the time. A summary of the operation is presented in
Table 1. In general, the problems were normal to start-up and shake-
down. There was a carbon steel valve that was incorrectly installed
and had to be replaced with a stainless valve. There were other
operating problems associated with learning the peculiarities of the
system. There was a major outage associated with repair of the ID
fan. It was cleaned and balanced and one of the bearings was
replaced.
1130
-------
A major change that came out of the-operation in Phase I was the
relocation of the lime slurry system. In the original pilot plant,
lime as calcium hydroxide was slurried on the first level and then
pumped by a Moyno3 pump up to the reaction chamber in the reactor
clarifier. There were several problems associated with this system.
The package unit purchased had a 5 minute residence time in the
slurry tank. This was too short a time and provided excessive scaling.
Furthermore, scaling occurred in the Moyno pump and in the line up
to the reaction zone. The scale built up as small pieces of material
that would eventually plug the line since there was insufficient fluid
velocity to keep the line clean. Recirculation was attempted but was
unsuccessful. During the period of April 1 -16, the lime slurry sys-
tem was relocated to the second level where it could gravity over-
flow into the reactor ciarifier. Clear liquor from the clarifier overflow
was pumped continuously through the slurry chamber and on demand
from the pH controller lime was added. Excellent automatic pH con-
trol was maintained in the absorber and recycle clarifier; however,
significant problems were encountered with scaling on the pH elec-
trodes in the reaction chamber of the reactor clarifier. This was not
successfully corrected during Phase I and the majority of operation
during this period was on manual control. Later improvements al-
lowed better monitoring. Fortunately, the system had a very large
capacity and the pH did not change rapidly.
Deposits on the ID fan were handled by the introduction of an inter-
mittent water wash into the inlet of the !D fan, 2 gpm for a period of
two minutes every two hours. This was sufficient to keep the majority
of the deposits off the fan and did allow it to operate without exces-
sive vibration.
Table 2 is a summary of Phase II operation. The objective of Phase II
was to demonstrate on-stream operability. The system was to be run
at a steady state condition for a period of at least 30 days continuous
operation. The period covered is from April 17 through May 29.
During this period the unit operated with a 99.7% on-stream time.
The outages were associated with electrical malfunctions, primarily
the control panel which had a faulty circuit breaker. This could not
be replaced until the unit was taken off line, so the unit was continued
in operation with this problem. In this test period, the unit was con-
sidered on-stream if the absorber was handling flue gas and remov-
ing SO2 to the code limit. Maintenance was required on the lime
handling and filter systems, but these did not affect the operation of
the absorber. One of the important features of the CALSOX system
is the ability to have the lime handling and filter operations off
stream for periods of up to 8 hours without affecting the rest of
the plant. This capability is associated with the large absorption
capacity of the ethanolamine solution, and of course the storage
capacity of the reactor clarifier (thickener).
1131
-------
The major process problem encountered in Phase II was pluggage
of the reactor underflow line from the reactor clarifier to the filter.
Steady state flow in this line was approximately 0.5 gattons per min-
ute. At this flow the linear velocity through the line was very low.
A Moyno pump was used and in general this supplied satisfac-
tory control of the flow of the material, but any accumulation of
large chunks would block the underflow line. It was also found that
the scrapers on the rake in the clarifier zone had too much clear-
ance and would allow accumulation of compacted sludge in the bot-
tom of the clarifier. This compacted material would break off in large
pieces from the wall and plug the underflow to the filter. This problem
was alleviated by the installation of neoprene wipers on the ends of
the clarifier scrapers and the relocation of the discharge opening.
In addition, the sludge level was maintained at a fairly low value dur-
ing most of the operation to eliminate the possibility of compacted
material. The slurry was also recycled continuously when it was not
being fed to the filter: The filter was oversized for the total slurry
generated and it could not be operated continuously. This caused
some problems at first in maintaining stable filter operation, but with
more experience it was possible to start up and shut down the filter
without significant problems.
In Phase III the operation of each unit of equipment was studied to
determine its optimum performance capabilities and to provide ac-
curate information for sizing of a full scale demonstration plant. This
period of operation covered from June 9 through October 14, In-
cluded was a 3 week period when the IPALCO personnel operated
the plant. There were several outages during Phase III associated
with rearrangement of equipment and special testing, but the major-
ity of the outages were caused by operating problems with the. ID
fan. In Table 3 is presented the history of the ID fan operation during
Phase III. During the early part of Phase III the intermittent washing
operation of the ID fan was discontinued because of unsatisfactory
results.
At the end of Phase II a build up of vibration was noted in the ID
fan; it was removed from its housing, sandblasted, checked for
cracks, balanced and reinstalled. When reinstalled, a continuous
washing operation was attempted. The result was disastrous as the
fan blades corroded completely through and the fan wheel separated.
Because of delivery time of a new fan wheel, a replacement ID fan
was obtained. During the down time to install the replacement fan,
hot gas recirculation was installed to allow the reheated gas from
the exit of the reheater to be recirculated to the inlet of the blower.
This was not completely successful but was continued throughout
the remainder of the operation. The inlet temperature to the fan was
1132
-------
maintained at 160°F or higher in an attempt to minimize the deposi-
tion of particles on the fan blades.
After a replacement wheel for the original fan was obtained it was
epoxy coated and reinstalled. The washing program established was
to wash for 15 seconds with hot water every 2 hours. It was found
after approximately one month operation that the epoxy coating was
beginning to flake off and that some fly ash was depositing on the
fan. Figure 4 shows the condition of the fan wheel. In Figure 5 it can
be seen that not only is fly ash present, but also some traces of
crystalline calcium sulfate material. The next fan coating attempted
was a 15 mil neoprene coating. This was not successful as the neo-
prene shredded after a very short period of operation. The last coat-
ing attempted was epoxy and the time between washing was length-
ened to 15 seconds every 4 hours. The final inspection after three
weeks of operation indicated that the fan wheel was in good con-
dition and very clean. Nevertheless, in future plant installations, it
is proposed that an FD fan will be used to eliminate the problems
associated with an ID fan. The FD fan concept was not demonstrated
in the pilot plant; however, it has been well proven in other applica-
tions.
During the third phase of the pilot plant operation, the unit was on-
stream approximately 72% of the time. At the end of the third phase
of operation it was concluded that sufficient information had been
obtained for the design of a large scale demonstration plant. In
November 1973, a program was entered into between Monsanto and
Indianapolis Power & Light for the design and cost estimation of a
125 megawatt demonstration plant.
LANDFILL STUDIES
During the pilot plant operation, filter cake from the vacuum filter
was used in landfill studies at the IPALCO site. A schematic of the
landfill test arrangement is shown in Figure 6. One inch flexible
plastic tubing was set into the ground at different levels and the
leachate material was collected from these tubes periodically. Four
different test plots were set up: 1) A control site with no cake, 2) A
cakefill, 3) A cake fly ash fill where the cake was mixed approximately
50% with fly ash, and 4) An activated sewage sludge treated cakefill.
In order to increase the amount of data obtained, the sites were
irrigated with fly ash pond water over a period of one year. In this
way the landfill sites were exposed to the equivalent of two years of
normal rainfall for the Indianapolis area over a single year of seasonal
cycles. During the exposure leachate samples were taken at different
depths below the fill material. After exposure, landfill plot corings
were taken. Additional studies on biodegradation of leachate, solu-
bilization of metal ions and cake and leachate toxicities were carried
out. Figure 7 shows a photo micrograph of fresh cake taken from
1133
-------
the filter belt, and Figure 8 shows typical landfill material after
exposure to one year of seasonal cycles. One of the major findings
of the landfill studies is that the cake from the CALSOX system
reduces the permeation rates through normal soil. Some of the
general conclusions obtained from the landfill studies is that the pH
of the leachate is neutral to slightly basic. The soluble salts level
in the cake leachate is of an order two to four times greater than
was found in the control, but the soluble salts level is still below the
potable water standards. Fly ash addition most significantly affects
the dissolved chloride ion concentration in the leachate. The trace
element levels below the fill were not significantly affected by the
cake.
The use of sewage sludge on the surface of site four resulted in a
high organism at the time of the cake landfilling. This represents the
most severe condition for biological activity encountered in any
normal landfill. The presence of many varieties of organisms allowed
for selective growth in the landfill media. This resulted in observa-
tion of a process which would normally take very long periods to
establish in the absence of the organism treatment. The leachate
supports organism growth below the ground resulting in higher
organic carbon loading in the leachate. The organisms in the ground
appear to metabolize leached sulfur compounds producing thio com-
pounds and possibly some sulfide ion. Separate studies were under-
taken to determine some of the additional parameters associated
with landfilling of the cake. It was found that the leachate is readily
biodegradable under aerobic conditions and probably also under
anaerobic conditions. It was found that the leachate material did not
significantly affect the solubilization of metal ions in the specific
cases studied: nickel, copper, cadmium and lead. In all cases at pH
around 11 the solubility in the leachate was less than in an aqueous
solution whereas in neutral solution the solubility in the leachate
was slightly higher than in the aqueous solution. It was also found
that the cake and cake leachate are relatively inert and have no
significant toxicity. This was measured in the normal terms of the
irritation properties to laboratory animals and acute toxicity by
aquatic tests with both bluegills and trout.
DESIGN AND ESTIMATION
OF THE 125 Mw DEMONSTRATION PLANT
As a result of a successful completion of a pilot plant operation, it
was jointly agreed between Indianapolis Power & Light Company
and Monsanto Enviro-Chem Systems that a demonstration plant of
approximately 125 megawatts would be designed using CALSOX
engineering and a cost estimate prepared for a CALSOX unit that
would handle approximately one-fourth the flue gas from the No. 7
unit of the Elmer W. Stout Generating Station in Indianapolis. Figure
1134
-------
9 shows the CALSOX system Plot Plan at IPALCO. Basic design
criteria for the plant included a new independent stack to handle the
flue gas from the plant, use of pond water to slurry the cake to land-
fill area and use of a railway system for delivery of lime and the etha-
nolamine absorbent.
The demonstration plant design incorporated the changes that were
developed during the pilot plant operation; namely, the relocation of
the booster blower from an induced draft to a forced draft and the
installation of the lime slurry facilities at such a level that it would
allow for gravity overflow into the reactor. In addition, since this unit
would be receiving the gas after the electrostatic precipitator, only
the absorber section was included as a scrubber module was not
required. Flue gas reheat (if required) will be by the mixing of heated
ambient air with the flue gas stream providing the capability of ap-
pr.oximately 50°F of reheat. A testing program would be carried out
to determine any need for reheat.
An additional change in the demonstration plant is that the reaction
zones will no longer be integral with the reactor cSarifiers as was
used in the pilot plant. Separate reactor vessels will be installed so
that a good level of agitation can be provided in the continuous
stirred tank reactors. This should improve agitation and give a better
degree of reaction completion than was found in the piJot plant. The
reactor clarifier will be replaced with a heavy duty thickener to better
handle the sludge. Duplicate lime slaking facilities are also included
in the plant design to provide for maximum reliability.
The cost of the 125 megawatt demonstration plant, including escala-
tion during the construction period, is estimated at about $9-million
which is approximately $72 per kilowatt installed capacity. The
operating costs, taking advantage of the raw material performance
demonstrated by the pilot plant, is estimated to be about 3.0 mills
per kilowatt hour. Figure 10 shows the Estimated Operating Costs.
SUMMARY
The feasibility of the CALSOX process has been demonstrated in
pilot operation. The concept of absorbing SOj with ethanolamine and
avoiding scaling in the gas handling equipment is a very forward
step to a reliable system. The low liquid to gas ratio should reduce
capital and operating costs. The buffered absorbent will allow flex-
ibility in the system and offer easier pH control. We look forward to
the future when the development program on CALSOX is more com-
plete and a full scale demonstration unit is in operation.
1135
-------
TABLE 1
Summary of Operation
Phase 1 Start-Up
OPERATING
PERIOD HOURS
FEB. 6-
FEB. 9-
FEB. 19
MARCH
MARCH
MARCH
MARCH
APRIL 1
- FEB. 8 53
-FEB. 18 — 0—
— FEB. 28 221
1 — MARCH 2 — 0—
3 — MARCH 10 121
10 — MARCH 14 — 0—
14 — MARCH 31 395
— APRIL 16 — 0—
OUTAGE
HOURS OUTAGE
4 FORCED OUTAGE — PLUGGAGE OF ABSORBER
SPRAY NOZZLES
SCHEDULED OUTAGE #5 BOILER
SEMI-ANNUAL MAINTENANCE
4 FORCED OUTAGE — REPAIR SCRUBBER FLOW
CONTROL VALVE; REPAIR ABSORBER PIPING
67 FORCED OUTAGE — REPLACE SCRUBBER
FLOW CONTROL VALVE WITH STAINLESS STEEL
32 FLUE GAS FLOW CONTROL REPAIR;
INSTRUMENT AIR REPAIR; ID FAN REPAIR
96 FORCED OUTAGE — CLEAN AND BALANCE ID
FAN; REPLACE BEARING
7 PLUGGAGE IN REACTOR OVERFLOW LINE;
FOULING IN ABSORBER PUMP
DOWN FOR EQUIPMENT MODIFICATION
-------
TABLE 2
Summary of Operation Phase II
On-Stream Operability
APRIL 17 TO MAY 29
TOTAL DAYS OPERATED — 42.33
TOTAL HOURS OPERATED — 1012.34
TOTAL HOURS OUTAGE — 2.66
PERCENT ON-STREAM — 99.7
4-29-73
OUTAGES
DATE
4-21-73
4-24-73
TIME
1500-1630
0140-0210
0425-0445
DURATION
HOURS
1.5
.83
REASON
ID FAN OFF — SCRUBBER PUMP DRY — SCRUBBER
RECIRCULATION STOPPED. CAUSE — UNKNOWN
POWER TO CONTROL PANEL OFF — FLUE GAS FLOW
FLUCTUATION. CAUSE — UNKNOWN (ELECTRICAL
STORM).
2300-2320
.33 POWER TO CONTROL PANEL OFF — RESET
BREAKERS — RESTART FAN. CAUSE — UNKNOWN.
-------
TABLE 3
Phase III Parametric Studies
ID Fan Operation
LO
OO
DATE
6-9 TO 6-1 6
6-28 TO 7-1 3
7-1 9 TO 8-2
8-5 TO 9-7
9-1 4 TO 9-25
UNIT
ORIGINAL FAN
REPLACEMENT
FAN
REPLACEMENT
FAN
ORIGINAL FAN
ORIGINAL FAN
COATING
NONE
NONE
NONE
1 MIL EPOXY
15 MILS
NEOPRENE
WASHING
CONTINUOUS 0.5
GPM
NONE
NONE
15SEC./2 HRS.
15 SEC. 12 HRS.
COMMENTS
CATASTROPHIC CORROSION
FLY ASH DEPOSITS
FLY ASH AND ABSORBENT
DEPOSITS — INSTALLED
HOT GAS RECIRCULATION
TO BLOWER INLET
EPOXY FLAKED OFF —
FLY ASH DEPOSITS
NEOPRENE SHREDDED
9-29 TO 10-14 ORIGINAL FAN 1 MIL 15 SEC./4 HRS.
SPECIAL EPOXY
FINAL INSPECTION SHOWED
VERY CLEAN WHEEL.
-------
CO
VD
12.0
11.0
10.0-
1 8.0-
I
0.1
FIGURE 1
Ethanolamine Neutralization Curve
I
0.2
I
0.3
RATIO
MOLES SO,.
MOLES DEA VERSUS pH
I
0.4
RATIO
I I
0.5 0.6
MOLES SO.
MOLES DEA
I
0.7
0.8
0.9
1.0
-------
FIGURE 2
CALSOX Process
FROM
ID FAN
RECYCLE
REACTOR
TO ATMOSPHERE
AIR
ABSORBER IDEMISTER
CLARIFIER
REGENERA
TION
REACTOR
THICKENER
POND WATER
SLURRY
TO POND
-------
FIGURE 3
CALSOX System Pilot Plant Flow Sheet
REHEATER STACK
\
RECYCLE
CLARIFIER
\ /
VT
REACTOR
CLARIFIER
\
SLAKER LIME
CONVEYOR FEEDER
VACUUM FILTER
_o
AFRESH WATER)
_
o
"nrrUCK
I A u—r" x.
CAKE
TO
LANDFILL
-------
FIGURE 4
" JBiilry* *• ~*A '• * :'i-
Inlet to Fan Wheel
Showing Hub and Front Flange
Between Radial Blades
Showing Edge of Blades and Front Flange
1142
-------
r
X1,000
X 2,000
X 5,000
X 5,000
FIGURE 5
Deposits on Fan
1143
-------
FIGURE 6
Landfill Schematic at IPALCO
Layered fill
1144
-------
FIGURE 7
Photo Micrograph — Fresh Cake
X 1,800
1145
-------
FIGURE 8
Photo Micrograph — Aged Cake
V *
.
'
,
*r
'- , -i *• '
> > ,-• •
. — " -V '" • r
' - ' . * » l» ' — • ^ v
-"Vy. .>• / .
X
"k V,^^Y •*, »
. ^ *>>• '
<:.
•*^-' ••'•-'
X 2,500
1146
-------
FIGURE 9
CALSOX System Plot Plan IPALCO #7 Unit
BOILER AREA
LIME
UNLOADING BUILDING
RAILROAD SPUR
m m-
REGENERATION
REACTOR
/ CALSOX
HUMIDIFIER/ STACK
ABSORBER
STREET
-------
FIGURE 10
Calsox Process
Estimated Operating Costs
(125 Mw — 3.2% Sulfur Coal; 90% SO2 Removal)
00
ANNUAL QUANTITY
UNIT COST—$
ANNUAL
COST—$ Mills/kwhr.
DIRECT COSTS
Delivered Raw Materials
DEA
Quicklime
Subtotal Raw Materials
CONVERSION COSTS
Operating Labor
Supervision
Steam
Process Water
Electricity
Instrument Air
Maintenance, Labor & Materials
Laboratory
Subtotal Conversion Costs
INDIRECT COSTS
Average Capital Charges at 15% of
Total Capital Invested
Overhead
Plant, 20% of Conversion Costs
Administrative, 10% of Operating Labor
Subtotal Indirect Costs
TOTAL OPERATING COST
518,000 Lbs.
18,200 Tons
$.18/Lb.
$26/Ton
11,650 Man Mrs. $8/Man Hr.
1,450 Man Hrs.
218,750MLbs.
70.000M Gal.
18,900,000 kwhr.
26.300M Ft3
$10/Man Hr.
$.50/M Los.
$.1/MGal.
$.0038/kwhr.
S.15/M Ft3
.04 x $6,000,000 — (Fixed Investment)
93,240 .108
473,200 .541
566,440 .648
93,200 .107
14,600 .017
109,400 .125
7,000 .008
72,000 .082
4,000 .004
240,000 .274
20,000 .022
560,200 .639
1,350,000 1.543
112,000 .128
9.300 .011
1,471,300 1.682
2,597,940 2.969
BASIS:
Remaining Power Plant Life-— 30 Years
Stack Gas Reheat to 170°F
Power Plant On-Stream 7,000 Hr./Year
Capital Investment $9,000,000
-------
ACKNOWLEDGEMENTS:
1 Proprietary term and trademark of Monsanto Enviro-Chem Systems,
Inc.
2 Trademark of Riley Company.
3 Trademark of Bobbins and Myers, Inc.
1149
-------
WESTVACO ACTIVATED CARBON PROCESS
FOR SOX RECOVERY AS ELEMENTAL SULFUR
F. 0. Ball
6. N. Brown
A. J. Repik
S. L. Torrence
Westvaco Corporation
Research Center
North Charleston, South Carolina
Abstract
An all dry, fluidized bed process using activated carbon for the effec-
tive recovery of S02 as elemental sulfur from stack gas has been demon-
strated in a 20,000 cfh integral pilot plant. The granular carbon was
continuously recycled over 20 times between a flue gas slipstream of an
oil fired boiler for S02 removal and the sulfur recovery steps. The
performance of the carbon remained at a high level over the 300 hour
test with no undue chemical or mechanical loss of carbon. Over 90% of
the 2,000 ppm sulfur oxides was removed from the flue gas as sulfuric
acid by catalytic oxidation and subsequent hydrolysis within the carbon
granule. In the first of two recovery steps, the acid loaded carbon was
initially contacted at 300°F with internally produced hydrogen sulfide
for conversion of the acid to elemental sulfur. The by-product sulfur
was then thermally stripped from the carbon and the required H2S pro-
duced by reacting the remaining sulfur on carbon with an external source
of hydrogen at 1000°F.
Sufficient process and design information was developed from the data
obtained in the integral run and prior stepwise pilot equipment opera-
tion to permit scale-up to a 15 MW prototype, the next anticipated
development stage. The preliminary design includes installation on a
coal fired boiler and the use of a coal fed gas producer to supply the
necessary reducing gas. An economic assessment of a conceptual design
for the S02 removal process as applied to a 1,000 MW coal fired boiler
indicated capital and operating costs competitive to costs of other
regenerable systems.
Much of this information was developed under a contract partially funded
by the Environmental Protection Agency with Mr. Leon Stankus acting as
project officer. However, EPA does not necessarily endorse the product
or process.
1151
-------
WESTVACO ACTIVATED CARBON PROCESS
FOR SOX RECOVERY AS ELEMENTAL SULFUR
INTRODUCTION
Activated carbon as used in dry regenerable S02 processes avoids
the critical control of chemical reactions necessary in wet processes
and the costs and problems involved in separating water from by-product,
either for recovery or disposal. Additionally, in the wet processes,
flue gas reheating may be necessary for fan protection and plume control.
Carbon processes which have or are being used all depend upon the
catalytic and sorptive character of the carbon which is utilized for
conversion of the S02 to sulfuric acid within the carbon granules. The
processes generally differ in the mode of removal and recovery of the
sulfuric acid from the carbon, i.e. either by thermal regeneration in
which the acid reacts chemically with the carbon to produce a S02 rich
by-product off-gas or by washing the acid loaded carbon with water to
produce a weak sulfuric acid. Further differences exist in the addi-
tional methods of upgrading the by-product streams through add-on steps
for conversion of the S02 gas stream to elemental sulfur or concentrated
sulfuric acid. The method of flue gas-granular carbon contacting also
varies in that fixed beds or moving beds with an upflow or crossflow gas
pattern are used. Particle size and characteristics of the carbon
granules with respect to the rate of S02 removal may differ, affecting
pressure requirements and equipment size.
Westvaco, as a major producer of activated carbon, embarked on a
program in which carbon, with a high S02 pickup rate capability, is
recycled with regeneration of the carbon achieved by reducing the
sulfuric acid chemically within the process to elemental sulfur without
the carbon being consumed. These techniques would a]so have potential
application to regeneration in any other processes which adsorb S02 as
sulfuric acid. Furthermore, the fluidized bed was selected for gas-solid
contacting in this development as one approach to permit handling rela-
tively large volume rates of gases in contact with recirculating carbon
solids. Other contacting methods may also have merit also but the
effectiveness of fluidized carbon bed systems has already been
demonstrated in large commercial units in existence ,
1152
-------
handling gas rates up to 540,000 cfm. The feasibility of using such a
carbon system was confirmed in bench scale and small pilot equipment
whereby H2S in contact with the sulfuric acid on the carbon resulted in
conversion to elemental sulfur which was then stripped off the carbon by
heating. Part of the sulfur was reacted with an outside source of '
hydrogen to produce the needed H2S.
The selection of a granular activated carbon and identifying the
major variables in the process chemistry served as a basis for the joint
work under an EPA contract which essentially involved scaling up the S02
removal sorption and the regeneration-sulfur recovery steps to a 20,000
cfh pilot plant. The objectives of the contract were to develop further
information on each process step initially and to finally demonstrate
the technical feasibility of the entire process and to evaluate the
performance of the carbon under extended recycling conditions in an
integrated pilot plant using flue gas from an oil fired boiler.
PROCESS CONCEPT
In the Westvaco Process dry granular activated carbon is contacted
with flue gas at stack gas temperatures. The S02 is removed through
catalyzed oxidation to SOs and a subsequent hydrolysis to sulfuric acid
which remains sorbed in the carbon granules, i.e.
$02 + 1/2 02 + HzO " H2S°4 (Sorbed) 0)
Sufficient water vapor and oxygen are present normally in the flue gas
for the reaction. This reaction takes place in a staged fluidized bed
vessel with provisions for adjusting the temperature for optimum S02
removal rates.
The sulfuric acid loaded carbon is transported mechanically to a
second fluidized bed reactor wherein the acid comes in contact with
hydrogen sulfide to produce elemental sulfur, which remains in the
1153
-------
carbon granules, and water vapor which is exhausted. Temperatures near
300°F are required for the reaction, i.e.
H2S04 + 3 H2S 4 S + 4 H20 (2)
Generation of the required hydrogen sulfide and the removal of the
elemental sulfur for recovery is accomplished in a third fluidized bed
reactor according to:
4 S
The thermal stripping of the sulfur and the reaction to produce H2S
requires temperatures near 1000°F.
In essence the entire process chemistry could be depicted by the
following in which the reductant hydrogen is attached to a recycled
sulfur loop to become more reactive as intermediate product, H2$, for
reduction of the acid to sulfur and water vapor.
REMOVAL REGENERATION
S02 + 1/2 02 + H20 — *- H2S04 + 3 H2S ^ — *> 4 HzO + S (Product) (4)
The hydrogen may be supplied through a number of commercially avail-
able gasifiers utilizing coal or other fossil fuels. Heating of the
regenerating reactors may be provided by conventional fuel burning
units.
The carbon serves as a vehicle for promoting the reactions effi-
ciently but does not directly take part. It is recycled between the S02
removal vessel to the regeneration vessels where its activity is
restored to the initial level.
1154
-------
The steps in this process concept were studied separately and
finally in combination for a completely integrated demonstration in
pilot plant equipment.
PILOT EQUIPMENT AND OPERATION
Integral Operation
The pilot plant evolved finally into three pieces of equipment, a
sorber, an acid converter and a sulfur stripper/H2S generator through
which granular carbon flows by gravity as essentially shown in Figure 1.
Reacting gases flow counter-current to the solids in each of the vessels
at the appropriate temperatures and rates to carry out the necessary
functions. Flue gas from an oil fired boiler is pumped through the S02
sorber at rates near 20,000 cfh at stack gas temperatures for removal of
S02 and SOs. The only other constituent introduced to the pilot plant
is the gas stream containing cylinder hydrogen which is fed to the
sulfur stripper/H2S generator reactor for production of H2$ and strip-
ping off of the by-product sulfur. The sulfur is ultimately removed
from the system in molten form in the condenser. The sulfur-free gas
containing H2S then passes through the sulfur generator for reaction
with the sulfuric acid on the carbon for conversion to elemental sulfur.
The spent gas from the sulfur generator is vented.
Heating and cooling requirements for the gases and carbon during
integral pilot operation are met with electrical resistance heating in
the case of the acid converter and sulfur stripping/H2S generator and
direct water spray for cooling the S02 sorber. All are automatically
adjusted except for the acid converter.
Recirculation of the carbon through the system is accomplished by
gravity flow through the reaction vessels and raised for recycle
mechanically by bucket elevator. The recirculating rate of the carbon,
normally about 30 lbs./hr., is controlled automatically by a gravimetric
solids rate feeder. For material balance purposes, any dust in the
major streams is collected through cyclones and bag filters.
1155
-------
FIGURE I. WESTVACO PROCESS INTEGRAL PILOT PLANT
CAMOM
•AS
FLUE GAS
S0 '2000 PPM
CFH
FIR1T STAK
FLUID BED
SOg SORBER
I8"DIA.'X 17.5 FT
I7S «F
SECOND STAGE
FLUID BCD' '
3A3
DISTRIBUTE
ACID CONVERTER
8" OIA. X 6 FT
300 °F
S. STRIPPER
4 DIA. X 19 FT.
B STAGES
IOOO°F
Ho -4- N,
IFLUIDIZING GAS
FLUID BED DETAIL
REGENERATED CARBON .
RECYCLE. APROX. 30Lfc/HR
SOLIDS
RATE
CONTROLLER
1156
-------
Main Reactors - Description
S02 Sorber
The sorber consists of an 18 inch diameter x 17.5 ft. high vessel
with 5 fluidized beds of carbon, each bed having an expanded bed depth
of 12 inches of carbon with the exception of the bottom bed which has 8
inches. The carbon is fed at a controlled rate to the top stage and
flows by gravity through the overflow weirs/downcomers from stage to
stage through the column. A uniform gas velocity across each stage of
the reactor is maintained through a perforated gas distributor plate,
as shown in the cutaway detail in Figure 1. Flue gas containing ^2,000
ppm S02 flows upward at a rate of 20,000 cfh. The inlet flue gas
temperature is at stack gas conditions of near 300°F which is the
temperature maintained in the bottom fluid bed stage where the $03 is
removed. The bed temperatures of the next stage are decreased to the
desired level, 175°F, through a water spray injected directly into a
carbon bed. Temperatures of the upper carbon bed stage are allowed to
seek their own level. Figure 2 is a photograph of the actual reactor
installation.
Acid Converter
In the integral pilot runs the acid converter consisted of a
moving bed unit 8 inches in diameter containing a carbon bed depth of
about 6 feet. Plug flow of solids through the reactor was assured
through a specially designed cone bottom section and particle residence
time testing prior to use. Temperature control in the range of 300°F
was effected by external electrical heating of the reactor walls and
by adjusting the moisture content of the acid loaded carbon feed.
The H£S containing gas enters the bottom of the reactor through a dis-
tributing zone for countercurrent contact with the carbon flowing
downward. The H2S depleted off-gas was vented from the top of the
reactor.
1157
-------
Figure 2. Continuous 18 inch diameter, 5 stage S02 and
adsorber operating on flue gas from a 50 MW
oil fired boiler.
503
1158
-------
Early in the program, studies showed that sulfur generation could
be acceptably performed in staged fluidized beds. However, the reactor
size needed for accommodating the gas and carbon flows of the
integral pilot design was prohibitively small, about 2 inches in
diameter, to be practically operated. A moving bed reactor was designed
for the integral pilot plant using preliminary information developed on
a bench scale.
Sulfur Stripper/H2S Generator
Removal of the by-product sulfur and reaction of the hydrogen with
the remaining sulfur on the carbon is performed in a 4 inch diameter, 8
stage fluidized bed reactor. The carbon loaded with sulfur is fed to
the third stage from the top and flows downward through overflow weirs
maintaining bed heights of five inches. Batch carbon beds are maintained
in the upper two beds to promote the conversion of sulfur to H2S.
Hydrogen containing gas is fed to the reactor bottom at rates near 220
scfh while reactor temperatures near 1000°F were maintained with external
electrical heaters. The off-gas containing vaporized sulfur and H2S is
passed through a dust removing cyclone and then the sulfur condenser
before use in the acid converter. Regenerated carbon discharging from
the reactor was cooled in an indirect heat exchanger for recycling back
to the S02 sorber by bucket elevator.
Process Unit Operations
In developing the design and operating information needed for the
integral pilot plant each of the process steps was studied separately in
the described main reactors and in other equipment. Two smaller fluid bed
S02 sorbers, one of a 6 inch diameter and one of a 4 inch diameter, in addition
to a thermogravimetric analyzer were used prior to testing of the 18 inch
diameter reactor. Batch pilot and bench scale equipment was used for
initial study of the sulfur generation and sulfur stripping/H2S genera-
tion steps. This information was used for developing rate equations for
the reactors and conditions involved.
1159
-------
Fluidization Mechanics
Some interrelation exists between the conditions for maintaining
the process chemistry and fluidization requirement in operating the
fluid bed reactor. Information was developed in room temperature
mock-up equipment of pilot plant size prior to the integral run on
fluidization, including required gas velocities, pressure drop, gas
distributor design, entrainment, and dust generation.
Instrumentation and Control
Sufficient instrumentation is available to maintain the desired
operating conditions during steady state conditions and to collect the
data necessary for performance evaluation. All input gas flow rates are
monitored through meters and checked by gas analysis instruments.
Temperatures and pressures at appropriate points within the system and
reactors are either indicated or recorded. The control and instrument
panel for the pilot plant is shown in Figure 3.
Sample Points and Analysis
Ports were positioned on the inlet and outlet of each of the three
reactors for sampling the granular carbon to determine the amount and
form of sulfur and moisture content. Gas sample ports were also
positioned so that various inlet and outlet points in the system were
analyzed chromatographically for H2, 02, HzS, S02, N2, C02, CO and H20
at the desired time. Samples of the solid were analyzed using standard
tests for measuring the physical and adsorption properties.
Granular Carbon
The carbon used in the integral pilot operation is a commercially
producible coal based carbon with a nominal 12x40,mesh size. The bulk
density of the carbon was about 40 Ibs./cu. ft. and the S02 removal
properties and attrition resistance were 60 minimum and 97 maximum,
respectively, as determined by specially designed tests. Although this
is an improved carbon with respect to attrition resistance and is satis-
factory for pilot testing, other carbons are being developed having
superior properties-for commercial use.
1160
-------
H
H
ST
Figure 3. Control panel for integral pilot plant operating on oil fired boiler.
-------
Flue Gas Characteristics
Gas used in the integral testing was flue gas from the stack of a
50 MW oil fired boiler having a mechanical dust collector. The sulfur
content of the oil was about 1.8-2.0% S which produces about 1100 ppm
S02 in the flue gas. In order to avoid variability at this stage of
operation, provisions were made for injecting additional S02 into the
flue gas to maintain a uniform level of 502 to tne pilot plant. The
temperature of the flue gas was kept at a stack gas temperature level
of 300°F before introducing into the S02 sorber.
Reducing Gas Composition
The reducing gas was a mixture of hydrogen and nitrogen from gas
cylinders. Hydrogen input rates were varied from 40 to 48% of the total
flow to establish process requirements.
General Operating Procedure
In starting up the integral system a known quantity of carbon to be
tested, about 500 pounds, was placed in the system to be recirculated
while preheating with a start-up heater to approach the desired operating
temperatures. The switch to flue gas was then made and the temperatures,
carbon flow rate and gas flows, and compositions were adjusted to the
estimated conditions needed. Manual adjustments were made to the $02
added to the flue gas above the actual oil produced S02 to maintain a
constant level. The amount of carbon placed in the system was sufficient
to minimize adding fresh carbon during the integral run and represents
about 40% above that needed to fil the reactors and conveying system.
Selection of Conditions for the Integral Run
The general intent of the integral run was to maintain constant
conditions over an extended period in which the granular carbon would be
exposed to repeated sorption and regeneration conditions with H2S pro-
duced in the process. Up to this time carbon had been exposed to flue
gas during sorption but in the acid converter step only cylinder H2S had
been used. An arbitrary time of 30 cycles was initially selected for
1162
-------
the integral run during which time any trend would be detectable and
indicative of longer term effects. In addition, a 90% S02 removal
efficiency was to be maintained with a sulfuric acid loading on the
carbon of at least 18 Ibs. acid/100 Ibs. carbon. Other limits on acid
conversion, sulfur recovery and operating conditions were selected based
on pre-integral pilot and bench scale test results.
RESULTS AND DISCUSSION
Integral Pilot Plant Run
Overall Operation
A single batch of carbon was exposed to 21 sorption-regeneration
cycles for steady state periods over a 300 hour operating time. Uniform
conditions were maintained during this time for the S02 sorption opera-
tion; however, changes in hydrogen input to establish process require-
ments led to 3 steady state periods for regeneration. During this time
the carbon was regenerated to its original activity in all cases. One
major interruption was experienced in carbon handling which caused shut-
down of the pilot plant, but care was taken in removing and replenishing
the carbon from the equipment to assure minimum process disturbance.
Run Conditions
The run conditions in Table 1 were predetermined to meet the target
goal of 90% S02 removal with an acid loading of at least 18 Ibs. acid/
100 Ibs. carbon. The inlet flue gas was controlled at the rate of
22,000 acfh to the S02 sorber. The S02 content was adjusted as neces-
sary to maintain 1900 to 2000 ppm and the inlet temperature was main-
tained at 300°F on the first stage of SOs removal with the next stage
cooled by water spray to 175°F for S02 removal. The temperatures of the
remaining stages were not controlled and were allowed to rise due to
heat of reaction during S02 removal. Carbon bed depths of 3.5 inches on
1163
-------
TABLE 1. Range of operating conditions for
integral pilot plant run.
INLET FLUE GAS
Gas Rate:
Temperature:
Composition, S02:
S03:
NO:
02:
H20:
Inert Gas (C02, N2):
CONDITION
22,000 SCFH
300°F
1900-2000 PPM
50 PPM
150 PPM
4.5 Vol.
13 Vol.
Balance
$02 SORBER
Temperature, Stage 1 (Bottom):
Stage 2 (H20 Spray);
Carbon Bed Depth (Expanded):
Fluidizing Velocity:
Space Velocity:
300°F
175°F
Stage 1 - 8"
Stages 2 to 5 - 12"
3.5 Ft./Sec. (3 300°F
3400 SCF Gas/CF Carbon-Hr.
REGENERATORS
Acid Converter
Temperature:
Inlet Rate:
Space Velocity:
$ Stripper/H2S Generator
Temperature:
H2 Inlet Rate:
Space Velocity, Stripper:
H2S Generator:
Fluidizing Gas Velocity:
290°F (Avg.)
Output from H2$ Gen. Range =
2.5-2.9 mol/moles
100 SCF Gas/CF Carbon-Hr.
1000-1100°F
3.4, 3.9, 4.3 moles H2/mole
S02 Sorbed
2000-3100 SCF Gas/CF Carbon-Hr.
6200-9300 SCF Gas/CF Carbon-Hr.
1.8-2.7 Ft./Sec. @ 1000°F
CARBON RECYCLE RATE:
29-30 Lbs. C/Hr.
1164
-------
the bottom stage and 6 inches on each of the remaining stages were set
for a total of 56 inches in an expanded state.
The recycle rate of carbon was set to achieve the desired acid
loading based on the previous relationships developed between the
operating parameters in the sorber.
Temperature and space velocity conditions for the acid converter
and sulfur stripper/H2S generator were selected based on earlier process
unit test results. The amount of hydrogen flow to the generator was
varied above the stoichiometric requirement of 3 moles/mole of S02 sorbed
on the carbon. The amount of H2S entering the sulfur generator was pre-
determined by the hydrogen input with no attempt to control this rate.
The pilot plant was started and operated over the extended period
under the above conditions using the previously described procedures.
Overall Process Performance
The main factors observed in the pilot operation were the level of
S02 removal during cycling of the carbon, the effects of hydrogen
supplied to the system for reduction in terms of stoichiometric require-
ment, and the conversion of S02 pickup to elemental sulfur. Other
factors were the detection of any carbon loss by chemical and/or
mechanical means and the disposition of any excess hydrogen.
S02 Removal
The removal of S02 during the integral run over the 300 hour period
is given in Figure 4. The flue gas, containing 1900 to 2000 ppm S02,
was desulfurized to well above 90% with a maximum of 97% or correspond-
ing to 60 ppm remaining in the effluent gas. By inspection of the plot,
there does not appear to be any reduction in removal efficiency of the
carbon. This has also been substantiated by analysis of the recycled
carbon in laboratory tests. During the integral run the carbon was
cycled through the system some 21 times based on a calculated carbon
1165
-------
FIGURE 4. S02 removal efficiency during integral pilot tests.
100 4-
954-
90
85 +
30 60 90
_| 1 1 1 L_
120
RUN TIME, HOURS
150 180 210
GOAL = 90X
INLET S02 = 1900-2000 PPM
4-
240
270
300
330
360
-4
10 12 14
NUMBER OF CARBON CYCLES
16
18
20
22
residence time of 15 hours in the integral system. The amount of S02
picked up by the carbon in terms of sulfuric acid averaged 24 lbs./100
Ibs. carbon, substantially above target.
No corrosion or dew point problems were noted in operation at
175°F since the 30-50 ppm gaseous SOs in the flue is adsorbed on the
carbon. This removal of SOs with carbon was demonstrated in previous
studies.
The 150 ppm NO in the flue gas is not directly affected by the
carbon and as such remains in the flue gas. The initial effect of the NO
is to suppress the S02 pickup. This effect appears up to a NO concentra-
tion of about 150 ppm but not beyond. This aspect is covered more fully
in a later section.
Regeneration Results
In the integral runs the intent was to demonstrate that the carbon
could be repeatedly regenerated for reuse and to maximize the amount of
elemental sulfur produced within the limitations of the present pilot
equipment. The only deliberate change in regeneration conditions was
1166
-------
in the hydrogen input. Other conditions were ore-set based on on'or
work.
As discussed in the preceding section, the activated carbon retained
its adsorptive capabilities throughout the run, attesting to the suit-
ability of regeneration under all hvdroaen inout conditions.
Sulfur By-product
It is important that the sulfur by-product from the regeneration
system be a salable commodity. The element sulfur recovered from the
pilot tests had characteristics as shown below:
Properties of Sulfur Product
Sulfur 99.7%
Ash 380 ppm
Carbon 2500 ppm
Acidity 2 ppm
Chloride <2 ppm
These properties, measured for Westvaco by a sulfur producer, classify
the sulfur collected as a commercial grade.
The small amounts of carbon in the sulfur, a result of fines carry-
over, from the regenerator gave the sulfur a greenish cast. It was
demonstrated that these fines could be readily filtered to give a bright
sulfur product of 99.9% purity.
During these integral tests as with prior work, there was limited
temperature control in the moving bed acid converter. As a result of
higher than desired temperatures, a part of the sorbed acid decomposed
to S02 in the upper part of the unit and was not readily available for
conversion to sulfur. Thus a maximum of 85% conversion of the sorbed
acid to sulfur was obtained with this equipment. Prior testing had shown
that with proper temperature control essentially 100% conversion to
sulfur is possible and this should be readily attainable in larger equip-
ment where fluid beds will be used.
1167
-------
Effect of Hydrogen Input
Three levels of hydrogen input were evaluated during the integral
runs and analysis on all the process streams were used in preparing the
material balance presented in Table 2.
Table 2. Effect of hydrogen input on by-product recovery.
Condition
A B
TOTAL HYDROGEN INPUT 4.6 4.3 3.9
(moles/mole available acid)
HYDROGEN USAGE
(moles/mole available acid)
1. Formation of by-product 2.9Q 2 88 30
sulfur
2. Reaction with by-product n 9r n -,-, n
_.. . r i i r- U • t- J \J » I I \J
sulfur to form
Reaction with chemisorbed
oxygen to form H20
TOTAL MEASURED HZ OUTPUT 4.05 3.99 3.90
Condition C essentially represents the process hydrogen input neces-
sary for conversion of the available acid to elemental sulfur product.
The hydrogen input above the stoichiometric ratio of 3 reacted with chemi-
sorbed oxygen to form water and H2S did not appear in the sulfur generator
vent gas. The reaction of a part of the inlet hydrogen with chemisorbed
oxygen had been observed in previous work and is apparently instrumental
in retaining the activated carbon's activity upon cycling. As the
hydrogen was increased from 4.3 to 4.6 a part of the product sulfur
1168
-------
reacted to form H2S which appeared in the vent gas, while formation of
water essentially remained constant. The difference between the measured
hydrogen input and output amounts to about 12% and could be the result of
analysis error or, possibly, chemisorption of these small amounts of
hydrogen on the carbon itself.
If the hydrogen ratio were lowered below those of Condition C, S02
formation would be expected at the expense of part of the product sulfur
product. This would be the desired direction if the process is slightly
out of balance since 502 can be readily recycled to the sorber.
The gas residence time in the regenerators is only about 15 seconds;
therefore, response of the system to hydrogen input is very rapid.
Thus, control should be readily achieved by monitoring regeneration
off-gases and adjusting the hydrogen input.
Carbon Attrition
The activated carbon used in integral tests was improved compared
to normal plant materials. The attrition rate measured with this
material, Figure 5, showed an initial decrease, probably due to a
FIGURE 5. Activated carbon attrition rate
during integral pilot tests.
1.0--
12 16 20
NUMBER OF CARBON CYCLES
28
32
1169
-------
rounding off of rough edges and then a stabilization at a rate of 0.26
Ibs./hr. The data indicated nearly all of this attrition occurred in
the fluidized beds of the S02 sorber. Additional work has shown that the
combination of larger particle sizes and carbons with improved hardness
will reduce the attrition rates to about 10% of the values measured here.
These improvements will be incorporated in future scale-up work.
It is significant that there is no apparent increase in the attri-
tion rate as the carbon was recycled thermally and chemically as has been
observed with other solid adsorbents. The nature of S02 recovery with
carbon, that it only provides a surface for catalysis and adsorption
rather than actually chemically participating in the reactions as is
done with metal oxides, probably results in the maintenance of structural
integrity and strength of the carbon.
Carbon Burn-off
In passing through the regeneration sequence the activated carbon
is exposed to temperatures progressively increasing from 300°F to 1000°F.
To -prevent chemical consumption at 1000°F in the regeneration sequence
the sulfuric acid is reduced to elemental sulfur at 300°F. In addition
to production of elemental sulfur, a goal of the Westvaco Process is to
minimize the amount of carbon reaction to produce C02- Measurements
were made on the C02 content of the regeneration off-gases to estimate
the amount of "burn-off" that could be occurring in the process.
As shown in Figure 6, the carbon burn-off as calculated from C02
evolution reached a stable value of about 10 - 12 Ibs. per ton of S02
sorbed from the flue gas. As shown by the dotted line this compares to
a "burn-off" of 187 Ibs. /ton if the carbon were consumed by reacting
with all the sorbed acid under thermal regeneration'conditions. This
1170
-------
FIGURE 6. Carbon burn-off during integral pilot tests.
190 • •
185 • •
180-.
MAXIMUM THEORETICAL BURN-OFF WITH THERMAL REGENERATION
u_
O
cz.
CD
1
o;
2
f
zs-
20 •
15.
10-
5-
0.
^
\°
°\ o R
^<^_ O Oa o O O • ^^* n -H « — °Y»
"• OO O O ^
O
1 1 1 j 1 { 1 1 1 ( 1 1 ! 1 1 l_J_J 1 1 1 1 tJ
10 1? 14
NUMBER OF CARBON CYCLES
16
18
20
22
24
reduction in burn-off of about 95% shows that the original objectives
were achieved. By inspection of the data there was no apparent effect
on burn-off when the hydrogen input was varied in the range of 3.9-4.6
moles/mole acid discussed earlier.
Complete conversion of the acid to sulfur was not required to pre-
vent burn-off. Earlier experiments on the bench scale verified this fact,
in that the addition of sulfur by various means considerably reduced the
chemical consumption of the activated carbon during regeneration. This
may have application in eliminating carbon burn-off in other carbon
processes that produce SO? as a product or product intermediate.
As discussed earlier there was some thermal decomposition in the
acid converter which would probably explain the small amount of burn-off
measured.
If all of the C02 measured is a result of burn-off the low values
measured here would correspond to a complete replacement of the inventory
only about once every two years.
Effect of Variables and Design Relationships
Prior to the integral tests extensive variable studies were con-
ducted on each process step to establish design relationships.
1171
-------
Considerable attention was given to the S02 sorber due to the potenti-
ally large variations in flue gas compositions and to the effect on the
size of this reactor.
$02 Sorption
Bench scale tests of the effects of temperature, acid loading and
S02, oxygen and water concentration on S02 removal are summarized in
Figure 7. Nitric oxide, which is also normally present in flue gas, is
FIGURE 7. S02 sorption rate model.
50
O
X
s:
LU
t—
Z
»—t
O
O
UJ
s
10
5
150°F
1
0.01
V
SULFUR DIOXIDE
OXYGEN
WATER VAPOR
0.05 0.1 0.5 1 5
GAS CONCENTRATION, VOLUME PER CENT
10
not picked up on the carbon; however, it does affect the S02 sorption
rate as shown in Figure 8. The effect of NO is constant above 100 ppm,
well above that normally present in flue gas. A multiple regression
analysis of all of the bench scale data resulted in a rate expression
1172
-------
FIGURE 8.
O.Z5
0.0
Effect of nitric oxide on S02
sorption at 200°F.
100 200
NITRIC OXIDE. PPH
relating all of the variables affecting S02 sorption as given by
Equation (1):
5520
v - 1.59(l
-------
where y, and y^ are the S02 concentrations entering and leaving any
particular stage and S is the space velocity for that stage.
The expression derived from bench scale results were compared with
pilot tests in the fluid bed S02 sorber operating on actual flue gas and
gave good agreement as shown in Figure 9.
100
80
60
40
P
to
' 20
5. 10
FIGURE 9. Comparison of rate model
to pilot data.
o
-------
FIGURE 10. Effect of sulfur dioxide concentration in
flue gas on pressure drop and space
velocity for various levels of S02 removal
89
PERCENT S02 REMOVAL
50 -
45 .
40 -
35 •
30 .
25 •
20
15 .
10 -
5 -
0
NUMBER OF STAGES: 5
TEMPERATURE: 175°F
0? CONCENTRATION: 4*
H20 CONCENTRATION: 10*
1" CARBON =0.5" W.G.4P
3000 PPM S02 . .
2000 PPM SOg
1000 PPM SO?
1 11 -t— 1 1 1 1 1 1
An m QO O1 QA OC O£ Q7 DO ftn
- 1000
*/>
5
u_
.1500 £
. 2000 I
I
- 3000 :
L
t.
• 4000 I
-------
An important characteristic of the system, as demonstrated by the
curves, is the little penalty required in terms of pressure drop in order
FIGURE 12.
Effect of oxygen concentration in flue
gas on pressure drop and space velocity
for various levels of S02 removal.
50- •
^45--
X
£ 40--
rc
<_>
* 35--
§ 30
ce.
o
uj 25+
o:
y, 20..
5"-
NUMBER OF STAGES: 5
TEMPERATURE: 175°F
H20 CONCENTRATION: 10%
SO, CONCENTRATION: 2000 PPM
1" CARBON = 0.5" W.G.AP
tt 0
89
90
91
92
93
94
95
96
97
98
99
- -1000
••1500
. - 2000
3000
4000
100
PERCENT S02 REMOVAL
to increase S02 removal from 90-99%. More details of the rate data and
f\ i
rate model for S02 sorption are included in previous publications ' .
Acid Conversion
The activated carbon from the S02 sorber contains sulfuric acid
sorbed in the pores of carbon and the acid is converted to elemental
sulfur with hydrogen sulfide in the acid converter. Bench scale rate
measurements, made over a wide range of conditions, led to a rate model.
Equation (4):
Rate of Sulfur _
r- j. • ~
Formation
Q
.O
-2644/T /Y xO.67
6 \«V/
v
(*H2S}
0.58
(4)
1176
-------
that represented the rate data shown in Figure 13. Using the same assump-
tions made for the S02 sorber as a multistage fluid bed, Equation (4)
FIGURE 13. Comparison of the sulfur generation rate
model to the experimental data for 250
to 325°F.
RATE = K0 e'E/RT (H2S)a (H2SO,,)b
6 8 10 20
H2S CONCENTRATION, VOLUME
60 80
was used to develop the design relationship, Equation (5):
Space Velocity . 9.05 a'2644/1 Xv°'67 [y/'42 - y,0'"2] (5)
where Xv is the acid loading on a stage and y1 and y2 are the
concentrations entering and leaving the stage. The design relationship
1177
-------
was compared to actual pilot fluid bed operation and predicted results
within about 25%.
The relative effects of hydrogen sulfide concentration and tempera-
ture on the rate of reaction below 350°F and 60% H2S are shown in
Figure 14. As can be seen, as the temperature or H2S concentration
FIGURE 14. Effect of temperature and H2S concentration
on the relative rate of sulfuric acid con-
version to elemental sulfur.
o
ID
UJ
i
1.0. .
0.9- •
0.8- •
0.7. .
0.6- •
0.5- •
0.4-•
0.3- •
0.2- •
0.1' •
0 - -
H2S CONCENTRATION
I I I I I I I I I
175
200 225 250 275 300
TEMPERATURE, °F
325
350
decreases, the relative rate of reaction also decreases. Although the
reaction is not as temperature sensitive as the S02 removal,an operating
temperature as high as possible is desirable to maximize the space
velocity. The upper limit to temperature is about 350°F because acid
conversion begins at this temperature; and, for this reason,an operating
temperature of about 300°F is normally used. The H2$ concentration on
1178
-------
the other hand is a function of the hydrogen concentration to the sulfur
stripper/H2S generator. Typical concentrations from a reformer after
being shifted are about 60% and from a gasifier after being shifted
about 20%. More information on development^of Equation (5) is given in
a previous publication .
Sulfur Stripping/H2S Generation
During sulfur stripping and H2S production the sulfur product is
vaporized from the carbon and the remainder is reacted with hydrogen to
produce the H2$ required for acid conversion. Bench scale testing with
sulfur loaded carbon showed that about 7% of the sulfur was chemisorbed
and could be removed by reaction with hydrogen. The remainder of the
sulfur could be vaporized and recovered as elemental sulfur.
Equilibrium experiments showed that the physically adsorbed sulfur
isotherm followed the familiar Polanyi-Dubinin relationship as shown
below:
ln(L - 7.3) = 4.1 - 0.179(T log ^)2 x 106
where L = Ib. S/100 Ibs. C
T = °R
PS = vapor pressure of sulfur, torr
P = vapor pressure of sulfur over carbon,
torr.
The effect of loading and temperature upon the concentration of sulfur
in the vapor phase over carbon is shown in Figure 15.
1179
-------
Figure 15. Effect of temperature on the equilibrium
data of sulfur vapor over activated carbon.
i.o-
0.11
.01-
o
(_>
oi
.001-
.0001
1200°F
10 15 20 25
SULFUR LOADING ON CARBON, IBS. S/JLB. C
30
35
In a continuous system the removal of physically adsorbed sulfur
is a function of temperature and residence time and the removal of the
chemisorbed sulfur is a function of temperature, hydrogen concentration
and contact time.
1180
-------
The results of fluid bed tests summarized in Figures 16 and 17
FIGURE 16.
Effect of temperature on sulfur removal
in a continuous fluid bed reactor.
0.3 h
o
CO
CO
—I
o
«=c
o
I
o:
Z3
U.
0.2
0.1
INLET SULFUR: .26 LB./LB.C
H2 CONCENTRATION: 27-32 VOL %
SPACE VELOCITY: 1300=1700 HR71
CARBON RESIDENCE TIME: 10-13 WIN.
I
800
1000
TEMPERATURE, °F
1200
showed that temperatures of 1000-1200°F were desirable to reduce the
residual sulfur loading to acceptable values. Removal of the sulfur
is not a strong function of hydrogen concentration as long as suffi-
cient hydrogen is available for reaction with the chemi s.orbed sulfur.
1181
-------
FIGURE 17.
0.3 -
Effect of hydrogen concentration on
sulfur removal from active carbon in a
continuous fluid bed reactor.
INLET SULFUR LOAD = .26 LB.S/LB.C
AVG. TEMP. * 1200°F
SPACE VEL = 3000 MR"1
RES. TIME = 6 MIN.
SPACE VEL = 1300 HIT1
RES. TIME = 13 MIN.
20 30
HYDROGEN CONCENTRATION, %
40
Once the sulfur is in the gas phase with hydrogen, reaction occurs
to form H2S required in the acid converter. Studies showed that the
conversion rate is expressed by Equation (6):
H2S =
on
where v
T
S
H2
208(107)(v)°'5
[299 T e-30645/T + (v)0.5_
= linear gas velocity,
= temperature, °R
= sulfur concentration
= hydrogen concentratio
e-30645/T (s,l/2 (H2)
ft. /sec.
as S-| , volume fraction
n, volume fraction.
(6)
1182
-------
The interrelationships of stripping equilibrium and rate and H2S
formation rate are now being combined to develop a method for design of
this reactor. The space velocities/conversion relationship measured
during separate unit and integral tests, however, provide sufficient
information for scale-up under expected conditions which can be refined
as modelinq techniques improve.
CONTINUED PROCESS DEVELOPMENT - PROTOTYPE
Sufficient information has been developed with the initial integral
pilot operation to consider scale-up to a larger unit. The mechanical,
adsorptive and catalytic character of the granular activated carbon has
been maintained during steady state recycle conditions over an extended
period without any significant departure from the initial process
concept. Fluidized beds, as used in the integral and pre-integral pilot
operation, provide an effective means for gas-solid contacting, even
through some solids flow problems were experienced, not unexpected for
this size equipment. Undesired side reactions or buildup of trace con-
taminants were not evident over the 300 hour carbon recycle test period.
Although not discussed, the demands on the materials of construction are
within those normally encountered in the petroleum industry. Information
on basic reaction rates and heats of reaction is available for scale-up
which essentially involves expansion of the cross-sectional area of the
fluidized reactors since the carbon bed depth should remain essentially
the same as in the pilot plant.
In selecting the next size for scale-up, consideration must be
given to the type of information required and minimizing risks without
committing excessive money. The size of the intermediate unit should
permit getting the technical and economic information needed for deci-
sions to proceed to a full plant scale. The prototype should provide
specific data on validity of scale-up, performance of materials of
construction, process dependability and controllability and performance
of off-the-shelf process equipment.
1183
-------
With these factors in mind, a scale-up of about 100 from the
present pilot size or a 15 MW equivalent prototype installation is
proposed. The flue gas volume rate for the prototype would be 1.2
million scfh compared to the 15,000 scfh of the pilot plant or a 16.3
ft. diameter S02 sorber compared to the pilot plant 1.5 ft. diameter
sorber.
A parallel continuing process development program while the proto-
type unit is being constructed and installed would serve a useful
purpose. Although major design features of the reactors can be scaled
up with some confidence, the effect of internals on the fluidizing
mechanics should be prudently determined in mock-up or shop tests.
Since there is limited experience in operation of the integral pilot
plant, additional performance testing under a variety of conditions
would be of value in design and operation of the prototype.
A preliminary prototype design has been prepared based on calcula-
tions and the information available from the pilot plant operation. The
prototype plant consists of four major components or 'ireas as with the
pilot plant:
1. Flue gas scrubbing (dry S02 removal)
2. Carbon regeneration and sulfur by-product recovery
3. Reducing gas (hydrogen) production
4. Materials handling and storage.
The basis for the nominal design is 90% removal of S02 from a coal
fired boiler using 3.5% sulfur coal. The sizes for the major pieces
of equipment for the 15 MW prototype have been determined and are pre-
sented in Figure 18 in elevation form.
Carbon is transferred at a rate of 3.3 tons/hour between the S02
sorber for treatment of 30,000 scfm flue gas and the regeneration
reactors for removal and recovery of the elemental sulfur produced at a
rate of 1,328 Ibs./hr. The transfer system consists of 3 bucket eleva-
tors and 2 hoppers for 10 hours of storage for regenerated and acid
loaded carbon. The reducing gas is produced in a gas producer with a
coal usage of 716 Ibs./hr. The equipment and information are shown on
the following pages.
1184
-------
FIGURE 18. WESTVACO PROCESS 15 MW PROTOTYPE UNIT
REGENERATION AND
PRODUCT RECOVERY
Co
HYDROGEN PRODUCTION
SOLIDS HANDLING AND STORAGE
SULFUR STRIPPER/
H2S GENERATpft
-------
FLUE GAS S02 SORBER
MATERIALS HANDLING
AND STORAGE
REGENERATION
HYDROGEN PRODUCTION
US KATE: 30,000 SCFX
CAS COMPOSITION: f.610 m SO?
40 PPM SOI
41 0;
iWHrO
REACTION: SO; • l/t 0? — SO)
SOj • «/> — •• H?S04 (SortwO)
MS 1 01 SO; KtMML: fO T.
OUTLET CMOOM LOW IK: O.ft IB, ACIO/L8. C
HEAT RELEASE: -117.000 8TV/MOL SO?
WCUTIK TDTCUTWt: ISO-JOO'r (SO? UNBVAL.)
30&-3Wt (SO) ttlDMl)
,, GAS/saiO CONTACT : CAS/SOLID FLUID KO CWfTEKCUttElrt.
)^ SSTAOS
QO RUIOIZIK GW VILOC.: J FT./S«.
-------
Preliminary process flowsheets and heat and material balances have
also been developed for the prototype to permit a detailed design and
cost estimate to be prepared.
PROCESS ECONOMICS
The projection of full scale flue gas desulfurization system costs
from pilot data is risky at best due to the many uncertainities associ-
ated with full scale commercial units. Such an evaluation may, however,
provide an indication of the process standing relative to others under
development and can help identify critical areas for future development.
An economic evaluation was prepared for the Westvaco Process by
scaling up the 15 MW flowsheet presented earlier to a 250 MW module. A
250 MW module was chosen because this given fluid bed S02 sorber sizes
comparable to those that have been constructed and operated commercially
in solvent vapor recovery. The costs derived from this conceptual
design for a 1,000 MW installation comprised of four 250 MW modules is
summarized in Table 3.
TABLE 3. Estimated capital and operating
costs for 1,000 MW utility
boiler.
Investment
$35/KW
Annual Costs
2 Mil/KWH
It is recognized that there are many factors which will undoubtedly
affect costs as the process is scaled up. This estimate would seem,
however, to indicate that the Westvaco system is competitive with others
under development. The cost distribution shown in Figure 19 provides a
basis for defining the areas of emphasis in future development.
1187
-------
FIGURE 19. Westvaco process cost distribution.
100
90
80
60
50
30
20
10
CAPITAL INVESTMENT
ANNUAL COSTS
S02 REMOVAL
31*
REGENERATION 4
PRODUCT RECOVERY
17.41
HYDROGEN PRODUCTION
STORAGE. HANDLING.
AUXILIARIES
9.5*
CATALYST
51
INDIRECT
(ENGINEERING, CONTRACTOR,
ETC.)
32.4*
HATER
0.8* "
LABOR S OVERHEAD
7.6%
MAINTENANCE
11.0*
POWER
8.6%
FUEL
13.0t
ACTIVATED CARBON
5.3*
COAL
10.0?
CAPITAL
CHARGES
43.6X
The largest direct cost items in the investment are the S02
removal and carbon regeneration with 31% and 17.4% of the total cost,
respectively. These are also the system components for which there
is the least full scale knowledge. This suggests that significant
i
attention to the details in this equipment may result in cost
reductions.
As might be expected, the charges for the original plant investment
are by far the largest annual cost, reaching nearly half of the total.
Any possible reduction in the sorber and regenerator costs in particular
1188
-------
would have a significant effect on this. Activated carbon make-up
charges account for 5.3% of the annual cost under the assumed losses.
Power and fuel costs in the current design comprise nearly 25% of
the annual costs. The power costs are, of course, primarily a result
of the pressure drop in the S02 sorber. A reduction in the power con-
sumption would be most affected by further improvements in the efficiency
of the activated carbon, since the carbon inventory and consequently
the pressure drop of the S02 sorber are directly related to carbon
efficiency.
The fuel costs result primarily from the need to heat the activated
carbon and sulfur to the stripping temperature. No credit was taken in
this evaluation for steam produced during sulfur product condensation
which will partially offset fuel costs. There is also the potential for
heat recovery from the activated carbon but this must be investigated
for the trade-off between the additional equipment costs and the fuel
savings. No credit for sulfur recovery was assumed.
Consideration of these cost factors should be included in the
design and operating test objectives of the prototype program.
CONCLUSIONS
The general objectives of the program were to develop further infor-
mation on each process step, to demonstrate the technical feasibility
of the entire process and to evaluate the performance of the carbon under
extended recycling conditions in an integrated pilot plant using flue
gas from an oil fired boiler.
The basic conclusions reached on the results and information to
date are:
1. Granular activated carbon of the type used can
effectively remove S02 from flue gas and can be
regenerated satisfactorily over a repeated number of
1189
-------
cycles without reduction in activity or an unacceptable
physical loss through chemical reaction or mechanical
attrition.
2. Information has been developed on each of the three
unit process steps, S02 sorption, acid conversion and S
stripping/H2S generation,to define the principal
variables affecting the process chemistry and their
correlations in regard to rate of reaction.
3. An acceptable sulfur product can be produced by the
process with H2S as an internally generated intermediate
reductant.
4. Use of fluidized beds present a viable and attractive
method of gas-solids contacting although other contacting
means are also applicable.
5. Operation of the integra"1 pilot plant over the limited
time did not appear to present any problems in regard to
control of the process.
6. Sufficient information has been generated on the perform-
ance of the activated carbon, process chemistry and pilot
operation to proceed to the next stage of development.
RECOMMENDATIONS
Based on the above conclusions it is proposed that scale-up to a
larger prototype plant be considered as the next step toward a commer-
cial plant.
ACKNOWLEDGEMENTS
The process development of the Westvaco S02 Recovery Process has
been partially funded by the Environmental Protection Agency.
1190
-------
BIBLIOGRAPHY
1. Avery, D. A., and D. H. Tracey, "The Application of Fluidized Beds
of Activated Carbon to Solvent Recovery from Air or Gas Streams"
Tripartite Chemical Engineering Conference - Symposium on
Fluidization, 1968, p. 21.
2. Levelspiel, 0., CHEMICAL REACTION ENGINEERING, 2nd Edition, John
Wiley & Sons, Inc., New York, 1964.
3. Brown, G. N., et al., "S02 Recovery Via Activated Carbon", Chemical
Engineering Progress 68(8):55-56 (August 1972).
4. Brown, G. N., et al., "Conversion of Stack Gas S02 to Elemental
Sulfur by an Activated Carbon Process", Presented at 71st National
AIChE Meeting, February 20-23, 1972, Preprint 25B.
5. Ball, F. J., et al., "Recovery of S02 from Stack Gases as Elemental
Sulfur by a Dry Fluidized Activated Carbon Process", Presented at
164th National ACS Meeting, August 31, 1972.
1191
-------
TECHNICAL REPORT DATA
(Please read Inuruclions on the reverse before completing)
l. REPORT NO.
EPA-650/2-74-l26-b
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Proceedings: Symposium on Flue Gas Desulfurization-
Atlanta, November 1974
s. REPORT DATE
December 1974
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Miscellaneous
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
1AB013; ROAP 21ACX-AA
NA
11. CONTRACT/GRANT NO.
In-House
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings; 11/4-7/74
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The proceedings document the presentations made during the symposium, which
dealt with the status of flue gas desulfurization technology, both in the U.S. and
abroad. The presentations emphasize process costs, both regenerable and non-
regenerable processes, second generation processes, and byproduct disposal/
utilization. Aim of the symposium was to provide potential users of sulfur oxide
control technology with a current review of progress made in applying processes
for the reduction of sulfur oxide emissions at the full- or semi-commercial scale.
The symposium was the sixth of such EPA-sponsored meetings, dating back to
1966, when the topic of principal concern was the use of limestone to control sulfur
oxide emissions.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.tOENTlFIERS/OPEN ENDED TERMS
c. COSATl Field/Group
Air Pollution Byproducts
Flue Gases Disposal
Desulfurization Marketing
Sulfur Oxides Sludge
Cost Effectiveness
Regeneration (Engineering)
Air Pollution Control
Stationary Sources
13B
21B
07A, 07D, 05C
07B
14A
8. DISTRIBUTION STATEMENT
19. SECURITY CLASS (Thil ReponJ
Unclassified
21. NO. OF PAGES
531
Unlimited
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
1192
-------
ENVIRONMENTAL PROTECT!ON AGENCY
Technical Publications Branch
Office of Administration
Research Triangle Park, N.C. 27711
OFFICIAL BUSINESS
AN EQUAL OPPORTUNITY EMPLOYER
POSTAGE AND FEES PAID
ENVIRONMENTAL PROTECTION AGENCY
EPA - 335
SPECIAL FOURTH-CLASS RATE
BOOK
Return this sheet if you do NOT wish to receive this material [ 1,
or if change of address is needed [~1. (Indicate change, including
ZIP code.)
PUBLICATION NO. EPA-650/2-74-126-b
------- |