Topics: EPRI CS-3706
Flue gas desulfurization Volume 2
Sulfur oxides Project 982-31
Nitrogen oxides Proceedings
Wet scrubbers November 1984
Dry scrubbers
Pollution control equipment
Proceedings: Eighth Symposium
on Flue Gas Desulfurization
Volume 2
-------
RJE PORT SUM M _A_ R_ Y
SUBJECTS SOX control / NOX control / Solid by-product disposal/reuse / Integrated
environmental control
TOPICS Flue gas desulfurization
Sulfur oxides
Nitrogen oxides
Wet scrubbers
Dry scrubbers
Pollution control equipment
AUDIENCE Environmental engineers / Generation planners
Proceedings: Eighth Symposium on Flue
Gas Desulfurization
Volumes 1 and 2
Timely exchanges of technical and economic information on
flue gas desulfurization (FGD) systems are essential to coal util-
ities that must meet strict emissions standards. These proceed-
ings constitute a valuable resource for utility, architectural-
engineering, and system-supplier personnel who must make
decisions about the design, installation, and operation of FGD
systems.
BACKGROUND
OBJECTIVE
APPROACH
KEY POINTS
Sulfur dioxide (SO2) emissions from coal-fired generating plants must be
carefully controlled to comply with government regulations. Compliance,
however, frequently means that utilities must install expensive and compli-
cated FGD systems. Therefore, utilities faced with limiting S02 emissions
need up-to-date information on this rapidly evolving technology in order to
select the most reliable and cost-effective process.
To provide a forum for exchanging information on the scientific, technical,
and regulatory developments related to SO2 control.
The EPA and EPRI cosponsored a four-day symposium that featured the
presentation of 40 technical papers and a major panel discussion. Utility and
industrial users and representatives of FGD system suppliers, research insti-
tutions, and government agencies were invited to contribute papers empha-
sizing progress in SO2 control, recent experience with installed systems, and
pertinent test results. Some 730 persons attended.
In the keynote address, the executive director of the National Acid Precipita-
tion Assessment Program examined the program's purpose, scope, and
status and its focus on providing Congress with a better scientific basis for
legislative and regulatory decisions. Nine other sessions included such
diverse topics as economics, construction materials, absorbent injection,
dual alkali systems, flue gas treatment (combined SOX/NOX), FGD chemistry,
limestone and organic acid, waste disposal and utilization, and dry FGD
systems (both pilot- and full-scale). The role of the architect-engineer in
EPRI CS-3706S Vols. 1 and 2
-------
EPRI PERSPECTIVE
PROJECT
constructing FGD systems for utilities was the topic of a panel discus-
sion conducted by representatives from seven architectural-engineering
firms. The purpose of the architect-engineer, all agreed, was to serve as
an extension of the utility's own engineering staff.
Of eight flue gas desulfurization symposia that have been held, this is
the second EPRI has cosponsored. The meetings, held approximately
every 18 months, bring together FGD vendors, government regulators,
researchers, and architect-engineers. In this relatively new and contin-
ually changing technology, the symposia present an excellent opportu-
nity for a wide-ranging interchange of FGD information and experience.
The meetings are well attended, and the published proceedings provide
a comprehensive and useful source of up-to-date happenings in SO2
control. The next symposium is planned for June 1985 in Cincinnati.
RP982-31
EPRI Project Manager: Thomas M. Morasky
Coal Combustion Systems Division
Contractor: Research Triangle Institute
For further information on EPRI research programs, call
EPRI Technical Information Specialists (415) 855-2411.
ORDERING INFORMATION
EPRI Members
EPRI CS-3706 Vols. 1 and 2, Proceedings, November 1984.
V1, 594 pages. V2, 578 pages.
If this report is not available from your company libraries or your
Technical Information Coordinator, you can order it from
Research Reports Center
P.O. Box 50490
Palo Alto, CA 94303
(415) 965-4081
Nonmembers You can order this report in print or microfiche from
Research Reports Center.
Price: V1 $41.50; V2 $41.50 Overseas price: V1 $83.00; V2 $83.00
(California residents add sales tax.)
Payment must accompany order.
;>•: Po.'.er Research institute PO BOA 10412 Palo Alto, CA 94303 All rights reserved
-------
Proceedings: Eighth Symposium on Flue Gas
Desulfurization
Volume 2
CS-3706, Volume 2
Research Project 982-31
Proceedings, November 1984
New Orleans, Louisiana
November 1-4, 1983
Prepared by
RESEARCH TRIANGLE INSTITUTE
Cornwallis Road
Research Triangle Park, North Carolina 27709
Compiler
F. A Ayer
Prepared for
Environmental Protection Agency
Office of Research and Development
401 M Street SW
Washington, DC 20460
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
EPA Project Officer
J. W. Jones
and
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, California 94304
EPRI Project Manager
T. M. Morasky
Desulfurization Processes Program
Coal Combustion Systems Division
-------
ORDERING INFORMATION
Requests for copies of fhis reporf should be directed to Research Reports Center
(RRC), Box 50490, Palo Alto, CA 94303, (415) 965-4081. There is no charge for reports
requested by EPRI member utilities and affiliates, U.S. utility associations, U.S. government
agencies (federal, state, and local), media, and foreign organizations with which EPRI has an
information exchange agreement. On request, RRC will send a catalog of EPRI reports.
Reseaich Calegones SOV control
NO. control
Solid by-broduct disposal/reuse
Integrated environmental contra
Copv'ignt - '9S- Electric Pov.er Research Institute, Inc All rights reserved
NOTICE
T"is -ecor! >,vas prepared as an account ot work sponsored in part by the Electric Power Research Institute, Inc
EPRi' N-E- "•?• EPRI rremoers of EPRI, nor any person acting on their behall (a) makes any warranty, express or
mpi'ec1 .vth respect 'o :ne use of any information apparatus, method or process disclosed in this report or that
sucn use mav not infringe privately owned rights, or (b) assumes any liabilities with respect to the use of, or for
dar-.ages 'esuiting from ire use of, any information, apparatus, method, or process disclosed in this report
-------
ABSTRACT
These proceedings are of the Eighth Symposium on Flue Gas Desulfurization,
held November 1 to 4, 1983, in New Orleans, Louisiana. The symposium was
sponsored by EPA's Industrial Environmental Research Laboratory, located in
Research Triangle Park, North Carolina, and the EPRI Coal Combustion Sys-
tems Division, located in Palo Alto, California.
The objective of the symposium was to provide a forum for supplier, user,
service, and regulatory groups to discuss the technical and regulatory
aspects of SOp control. The emphasis was on progress in SO^ control
technology, recent experience, and test results, not on future plans.
Volume 1 contains 24 papers from days one, two, and three, plus abstracts
from panel members. Volume 2 contains 16 papers from days three and four,
plus 6 unpresented papers.
111
-------
PREFACE
These proceedings for the Eighth Symposium on Flue Gas Desulfurization
(FGD) constitute the final report submitted to the Industrial Environmental
Research Laboratory; EPA; Research Triangle Park (IERL-RTP), North
Carolina; and EPRI's Coal Combustion Systems Division, Palo Alto,
California. The symposium was conducted at the Sheraton Hotel in New
Orleans, Louisiana, November 1 to 4, 1983.
The meeting served as a forum for the exchange of technical and regulatory
information and developments regarding systems and processes applicable to
utility and industrial boilers. At the opening session, the keynote
address examined the status and outlook for the National Acid Precipitation
Assessment Program: its present status and outlook for the future. Pre-
sentations were also made on the state of air quality legislation and regu-
lations, current and projected regulations of the Resource Conservation and
Recovery Act, and trends in commercial application of FGD technology. Sub-
sequent technical sessions dealt with economics, construction materials,
dry furnace absorbent injection, dual-alkali, flue gas treatment (combined
SOX/NOX). Other sessions included FGD chemistry, limestone/organic acid,
waste disposal/utilization, and dry FGD systems, pilot plant test results,
and full-scale installations. Participants also discussed the role of
architect-engineer as middleman between the utility and FGD suppliers.
Representatives from electric utilities, state environmental agencies,
equipment and process suppliers, coal and petroleum suppliers, EPA and
other federal agencies, and research organizations attended the sessions.
The following people contributed their efforts to this symposium.
• Julian W. Jones, Chemical Engineer, Emissions/Effluent
Technology Branch, Utilities and Industrial Power Division,
IERL-RTP, Research Triangle Park, North Carolina; EPA
symposium general chairman and project officer
• Thomas M. Morasky, Manager, Reliability and Nonrecovery
Systems, Coal Combustion Systems Division, Palo Alto,
California; EPRI symposium general chairman and project
manager
• Franklin A. Ayer, Manager, Conference Planning Office,
Center for Technology Applications, Research Triangle
Institute, Research Triangle Park, North Carolina;
symposium coordinator and compiler of the proceedings
Thomas M. Morasky, Project Manager
Coal Combustion Systems Division
-------
TABLE OF CONTENTS
VOLUME 1
Section „
Page
SESSION 1: OPENING SESSION
Julian W. Jones, Chairman
Keynote Address: National Acid Precipitation
Assessment Program: Status and Outlook 1-1
J. Christopher Bernabo
Remarks 1-21
Sheldon Meyers
The Resource Conservation and Recovery Act: Current.
and Projected Regulations 1-27
Stephen A. Lingle
Trends in Commercial Applications of FGD 1-29
Bernard A. Laseke,"' Michael T. Melia, and Norman Kaplan
SESSION 2: ECONOMICS
Thomas M. Morasky, Chairman
Computer Economics of Physical Coal Cleaning and
Flue Gas Desulfurization 2-1
Charles R. Wright,"' Terry W. Tarkington, and
James D. Kilgroe
Economic Evaluation of FGD Systems 2-27
Jack B. Reisdorf,* R. J. Keeth, C. P. Robie,
R. W. Scheck, and Thomas M. Morasky
Estimating Procedure for Retrofit FGD Costs 2-47
R. R. Mora, P. A. Ireland," R. J. Keeth, and
T. M. Morasky
Comparative Costs of SO^ Removal Technologies 2-63
John 0. Milliken
SESSION 3: MATERIALS OF CONSTRUCTION
Charles E. Dene, Chairman
""Denotes speaker
VII
-------
Section Page
EPRI Research on Corrosion and Degradation of
Materials for FGD Systems 3-1
Barry C. Syrett
Simultaneous Design, Planning, and Materials of
Construction Selection for FGD Systems 3-15
Alex Kirschner, Norman Ostroff,* R. F. Miller,
and W. L. Silence
Acid Deposition in FGD Ductwork 3-47
Daniel A. Froelich,-'" Carl V. Weilert, and Paul N. Dyer
In Situ Evaluation of High Performance Alloys in
Power Plant Flue Gas Desulfurization Scrubbers 3-61
R. W. Schutz and Charles S. Young*
SESSION 4: DRY FUMACE ABSORBENT INJECTION
Randall E. Rush, Chairman
Results from EPA's Development of Limestone
Injection into a Low NO Furnace 4-1
Dennis C. Drehmel,* G. Blair Martin, and
James H. Abbott
Review of EPRI Research on Furnace Sorbent
Injection S02 Control 4-19
Michael W. McElroy
Direct Desulfurization Through Additive
Injection in the Vicinity of the Flame 4-31
M. Yaqub Chughtai-" and Sigfrid Michelfelder
SESSION 5: DUAL ALKALI
Norman Kaplan, Chairman
Utility Double Alkali Operating Experience .5-1
Dennis L. Clancy, Richard J. Grant, L. Karl Legatski,*
James H. Wilhelm, and Beth A. Wrobel
Pilot Evaluation of Limestone Regenerated Dual
Alkali Process 5-21
John C. S. Chang"" and Norman Kaplan
SESSION 6: FLUE GAS TREATMENT (COMBINED SO /NO )
J. David Mobley, Chairman x x
Status of the DOE Flue Gas Cleanup Program /. 1
John E. Williams
"'-"Denotes speaker
VI11
-------
Section Page
Status of SO^ and NO Removal in Japan 6-37
Jumpei Ando
PANEL: THE ARCHITECT-ENGINEER - MIDDLEMAN BETWEEN
UTILITY AND FGD SUPPLIER 6-43
A. V. Slack, Chairman
Edward W. Stenby, Gene H. Dyer, Paul R. Predick,
Michael L. Meadows, Douglas B. Hammontree,
Christopher P Wedig, and Richard Rao, Panel Members
SESSION 7: FGD CHEMISTRY
Dorothy A. Stewart, Chairwoman
'Influence of Chlorides on the Performance of
Flue Gas Desulfurization 7-1
William Downs,"" Dennis W. Johnson, Robert W. Aldred,
L. Victoria Tonty, Russell F. Robards,* and
Richard A. Runyan
Effect of High Dissolved Solids on Bench-Scale FGD
Performance 7-19
James B. Jarvis,* Timothy W. Trofe, and
Dorothy A. Stewart
Pilot Plant Tests on the Effects of Dissolved Salts
on Lime/Limestone FGD Chemistry 7-37
Dennis Laslo,* John C. S. Chang, and
J. David Mobley
Modeling of SO-^ Removal by Limestone Slurry
Scrubbing: Effects of Chlorides 7-57
Pui K. Chan and Gary T. Rochelle*
Influence of High Dissolved Solids on Precipitation
Kinetics and Solid Particle Size 7-79
Frank B. Meserole, Timothy W. Trofe, and
Dorothy A. Stewart"
Effect of Limestone Grinding Circuit on FGD
Performance and Economics 7-105
J. David Colley,"' 0. W. Hargrove, Jr., and
Dorothy A. Stewart
VOLUME 2
SESSION 8: LIMESTONE/ORGANIC ACID
J. David Mobley, Chairman
"'Denotes speaker
-------
Section Page
Process Troubleshooting at a Utility Limestone
FGD System 8-1
J. David Colley, Robert L. Glover,
Temple E. Donaldson,* and Dorothy A. Stewart
Technical/Economic Feasibility Studies for Full
Scale Application of Organic Acid Technology for
Limestone FGD Systems . 8-23
James C. Dickerman* and J. David Mobley
SESSION 9: WASTE DISPOSAL/UTILIZATION
James D. Kilgroe, Chairman
Full-Scale Field Evaluation of Waste Disposal
From Coal Fired Electric Generating Plants 9-1
Julian W. Jones,* Chakra J. Santhanam, Armand
Balasco, Itamar Bodek, Charles B. Cooper,
John T. Humphrey, and Barry K. Thacker
Operations History of Louisville Gas & Electric
FGD Sludge Stabilization 9-25
Robert P. Van Ness,"" John H. Juzwiak, and
William Mclntyre
Coal Waste Utilization in Artificial Reef Construction 9-37
Jeffrey H. Parker,* Peter M. J. Woodhead, and
Dean M. Golden
Solid Waste Environmental Studies at Electric
Power Research Institute 9-49
Ishwar P. Murarka
Presented by Karen Summers
SESSION 10, PART I: DRY FGD: PILOT PLANT TEST RESULTS
Theodore G. Brna, Chairman
Current Status of Dry SO^ Control Systems 10-1
Michael A. Palazzolo,* Mary E. Kelly,
and Theodore G. Brna
Acid Rain Prevention Thru New SO /NO Dry
Scrubbing Process * . * 10-23
Karsten S. Felsvang,* Per Morsing,
and Preston L. Veltman
"''Denotes speaker
-------
Section Page
Process Characterization of SO^ Removal in
Spray Absorber/Baghouse Systems 10-41
Eric A. Samuel,"' Thomas W. Lugar, Dennis E. Lapp,
Kenneth R. Murphy, Owen F. Fortune, Theodore G. Brna,
and Ronald L. Ostop
Dry Scrubber, Flue Gas Desulfurization on High Sulfur,
Coal-Fired Steam Generators: Pilot-Scale Evaluation 10-61
Bryan J. Jankura,* John B. Doyle, and Thomas J. Flynn
EPRI Spray Dryer/Baghouse Pilot Plant Status
and Results 10-81
Gary M. Blythe* and Richard G. Rhudy
SESSION 10, PART II: DRY FGD: FULL SCALE INSTALLATIONS
Richard G. Rhudy, Chairman
Field Evaluation of a Utility Dry Scrubbing System 10-109
Gary M. Blythe,"- Jack M. Burke, Theodore G. Brna,
and Richard G. Rhudy
Overview and Evaluation of Two Years of Operation
and Testing of the Riverside Spray Dryer System 10-131
John M. Gustke, Wayne E. Morgan,""
and Steven H. Wolf
Design and Initial Operation of the Spray Dryer
FGD System at the Marquette, Michigan, Board of
Light and Power - Shiras #3 Plant 10-161
0. Fortune,- T. F. Bechtel, E. Puska, and J. Arello
Start-Up and Initial Operating Experience of the
Antelope Valley Unit 1 Dry Scrubber 10-181
Robert L. Eriksen,* Frederick R. Stern,
Richard P. Gleiser, and Stanley J. Shilinski
Characterization of an Industrial Spray Dryer at
Argonne National Laboratory 10-199
Paul S. Farber* and C. David Livengood
UNPRESENTED PAPERS
An Economic Evaluation of Limestone Double Alkali
Flue Gas Desulfurization Systems 11-1
Gerald A. Hollinden, C. David Stephenson, and
John G. Stensland
""Denotes speaker
XI
-------
Section „
Paj|e_
Developments and Experience in FGD Mist Eliminator
Application 11-39
Richard T. Egan and William Ellison
FGD Gypsum: Utilization vs. Disposal 11-61
William Ellison
Operating Experience with the Chiyoda Thoroughbred 121
Flue Gas Desulfurization System 11-75
Seiichi Kaneda, Mitsuhiro Nishimura,
Hitoshi Wakui, Ikuro Kuwahara, and
Donald D. Clasen
Operation Experience with FGD Plant II at
Wilhelmshaven Power Plant, West Germany 11-91
B. Stellbrink, H. Weissert, and P. Kutemeyer
The SULF-X Process 11-111
Edward Shapiro and William Ellison
APPENDIX: ATTENDEES A-l
XII
-------
SESSION 8: LIMESTONE/ORGANIC ACID
Chairman: J. David Mobley
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
PROCESS TROUBLESHOOTING AT A UTILITY LIMESTONE
FGD SYSTEM
J. D. Colley, R. L. Glover,
T. E. Donaldson, D. A. Stewart
-------
PROCESS TROUBLESHOOTING AT A UTILITY LIMESTONE FGD SYSTEM
by: J. David Colley
Radian Corporation
Austin, Texas
Robert L. Glover
Radian Corporation
Austin, Texas
Temple Donaldson
Central Illinois Light Company
Peoria, Illinois
Dorothy Stewart
Electric Power Research Institute
Palo Alto, California
ABSTRACT
Central Illinois Light Company's Duck Creek Unit No. 1 experienced sig-
nificant reliability and operating problems with its limestone FGD system
following start-up in 1979. CILCo entered into a testing and evaluation pro-
gram co-funded by the Electric Power Research Institute in late 1981 to
verify the feasibility of using additives to improve the system's low S02
removal efficiency. A second objective of the work was to improve the sys-
tem's reliability, which averaged slightly better than 60 percent prior to
testing. Severe mist eliminator scaling that was causing routine scrubber
outages was primarily responsible for the reliability problems. The eight
month program that followed involved extensive chemical process troubleshoot-
ing.
Two types of organic acid buffers were tested along with the addition of
magnesium oxide during the period. Both dibasic acid (DBA) and the magnesium
proved capable of enhancing S02 removal to levels sufficient to maintain the
unit in compliance with the 1971 S02 NSPS. An economic analysis was performed
based on the data collected during this testing. The cost study compared
capital and annual operating and maintenance costs for each option over the
remaining life of the plant.
Work was conducted concurrent with the additive testing to solve the mist
eliminator scaling. The cause of the scaling was identified and effectively
stopped by switching to a fresh water wash, adjusting the wash sequence, and
improving limestone utilization. Improvements in limestone utilization were
accomplished by optimizing the operation of the mill circuit to provide a
finer, more reactive product. Since completion of the program, the FGD sys-
tem has consistently achieved reliability numbers greater than 95 percent
while at the same time lowering operating and maintenance costs.
-------
INTRODUCTION
Duck Creek #1 is a bituminous coal fired unit which is owned and oper-
ated by Central Illinois Light Company (CILCo). The boiler is rated at 400
megawatts and began commercial operation in 1976. It fires a high sulfur
Illinois coal and is subject to the 1971 utility NSPS which limits the units
SOz emissions to 1.2 pounds per million Btu boiler heat input. The FGD sys-
tem consists of four Riley Environeering rod deck scrubber modules each de-
signed to treat 25 percent of the maximum gas flow.
Since startup of its scrubbers, CILCo has experienced many of the same
operating problems seen at other utility limestone FGD systems. The primary
difficulties have been low S02 removal, chemical scaling in the mist elimi-
nators, materials and liner failure, and low reagent utilization. By 1981,
the utility was facing regulatory pressure because of S02 compliance vio-
lations and was reporting FGD reliabilities of only slightly more than 60
percent. In addition, the low reagent utilization was leading to accelerated
filling of the FGD waste disposal pond.
CILCo contracted Radian Corporation in late 1981 to conduct a test pro-
gram whose primary objective was to demonstrate the feasibility of using an
organic buffer to improve the FGD system S02 removal efficiency. Severe
weather hampered the testing but the initial results were encouraging.
Further test work was delayed until April following; a month long plant outage.
As the testing progressed after startup in April, the scope of work expanded
to include investigating the mist eliminator scaling problem and ways to im-
prove limestone reagent utilization. The Electric Power Research Institute
became involved at this stage and funded the cost of the remainder of the
program. This included conducting a test to determine the feasibility of
using magnesium as an alternative to the organic buffers and completing a com-
parison of the economics of the available options for improving S02 removal
efficiency. The results of the test work and cost study are discussed in this
paper along with a description of the FGD system. Finally, observations are
presented from a recent Radian sampling trip to document the systems perfor-
mance a year after the end of the Radian/EPRI program.
PROGRAM OBJECTIVES AND SCHEDULE
The overall objectives of the test program were to improve both the S02
removal efficiency and reliability of the Duck Creek FGD system. The organic
buffer and magnesium testing were intended to demonstrate effective means of
increasing the removal efficiency of the scrubbers. This work gathered pro-
cess and chemical data over a variety of system operating conditions and pro-
ceeded in three phases. The first phase involved gathering baseline data over
a one week period to characterize the performance of the scrubbers before ad-
ditive testing began. The information was collected to quantify the magnitude
of the S02 removal problem and to serve as a point of reference for data col-
lected later.
The organic acid demonstration test work which made up phase two began in
early May, 1982 and continued for four months until the end of the testing in
-------
August. Phase three, the magnesium testing, began the first of August and
concluded four weeks later.
To accomplish the second objective to increase the reliability of the FGD
system, work to solve the severe mist eliminator scaling was undertaken in
early May. This proceeded concurrent with the additive testing and lasted
throughout the 1982 test phase. Improving limestone reagent utilization was
important in solving the mist eliminator scaling problem. This work also ran
concurrent with the additive testing and lasted through the program.
SYSTEM DESCRIPTION
The limestone FGD system at Duck Creek was designed to treat the entire
volume of flue gas from the boiler and limit S02 emissions to below 1.2 Ibs
per MMBtu. At full load operation, approximately 400 MW of power are being
generated resulting in a total flue gas flow of 1.4 million acfm at scrubber
exit conditions (130°F, saturated). The boilers induced draft fans discharge
into a common plenum to which the inlet ducts of the four modules are con-
nected. Distribution of the flue gas is determined by the pressure drop
across each module and is not regulated by dampers. The flow to each may be
adjusted by manually removing rods from or adding rods to any of the seven
decks inside each tower. The SC>2 concentration of the 300°F gas as it enters
the towers averages about 2500 ppm. A simplified flow diagram showing one of
the four modules and the reagent preparation plant and waste disposal is shown
in Figure 1. Reagent preparation and waste disposal are common to the FGD
system.
Limestone preparation is accomplished with a 10' x 18' wet ball mill which
was originally designed to grind 40 tons per hour of stone. Four 6 inch dia-
meter hydroclones are used to control the product size distribution. Columbia
Quarry supplies the stone which is mined underground from the Kimmswick for-
mation. Product slurry at about 25 weight percent solids is stored in a
125,000 gallon tank prior to being fed to the scrubber reaction tanks.
Limestone feed to each reaction tank is controlled based on a pH feed-
back loop. The pH setpoint in the reaction tanks can be maintained within
±0.05 units with control of the limestone slurry in an on/off mode. The
150,000 gallon working volume of the reaction tanks provides the necessary
time for precipitation of the sulfur salts and dissolution of the limestone.
Reaction tank slurry is pumped to the top of the contactors at a rate of
about 15,000 gpm where it passes countercurrent to the flue gas. At full
load conditions, this results in a liquid-to-gas ratio of slightly more than
40 gals/1000 acf. Gas-liquid contacting to promote S02 mass transfer is
achieved with the seven rod decks. Gas side pressure drop across the scrub-
bers is approximately 12 inches H20 at design flow rates.
A two-stage horizontal mist eliminator (constructed of Hastelloy G) is
located downstream of the last rod deck in each absorber to remove entrained
slurry. Periodic washing of the mist eliminator removes collected solids.
The majority of the wash water is routed to the disposal pond with a small
amount entering the recycle loop. Flue gas exits the system without reheat
through a lined wet stack.
5-3
-------
00
i
-p-
FLUE GAS
(1.4 x 106acfm
AT 2500 ppm SO2)
DBA/
MAGNESIUM"
HORIZONTAL
MIST ELIMINATORS
PRODUCT
A A A A A
__
f FLUE GAS TO
(STACK @ 400 ppm so2
\
ABSORBER
SCRUBBER
RECYCLE
SLURRY
CYCLONES
I
V/ i ^WATER
I 1 r-^i
LIMESTONE
SLURRY
STORAGE
BALL MILL
RAW
LIMESTONE
- DISPOSAL -
~ POND -
1
WATER
RECLAIMED WATER
TO SCRUBBERS
AND ASH SLUICE
Figure 1. Duck Creek FGD System
-------
Solids content in the reaction tank is controlled at 12 weight percent
by adding dilution water reclaimed from the disposal pond. Waste slurry is
pumped from each reaction tank to the pond. This pond also serves as dis-
posal for the boiler bottom and fly ash as well as a final repository for all
other plant wastes.
DISCUSSION OF RESULTS
This section presents the key results of the four month period of
testing in 1982 and the 1983 evaluation trip. The results of the additive
testing are presented first in the S02 Removal subsection. This is followed
by results of the system reliability improvement work which are presented in
the Mist Eliminator Scaling and the Reagent Utilization subsections. The
Cost Study which investigated the economics of the alternative additives is
discussed next. Finally, observations made one year following the 1982
testing are presented in the 1983 Evaluation Trip subsection.
SO2 REMOVAL
Baseline Testing
A characterization of the FGD system was completed at the beginning of
the program to document the operation and performance of the FGD system, x^ith
particular emphasis on SOa removal. The testing was done prior to the ad-
ditive testing to define the baseline characteristics of the scrubbers. Data
were collected at various reaction tank pH setpoints and over a range of unit
loads. The results are plotted so that the effect of pH and gas flow on the
scrubbers S02 removal can be evaluated. Figure 2 presents the relationship
observed between reaction tank pH and the S02 removal efficiency of the scrub-
bers. There is a strong correlation between pH and removal up to about
93 percent removal indicating that liquid phase alkalinity is limiting
the scrubbers removal efficiency- The data shown in the figure were
taken with the unit operating at or above 90 percent of its maximum to mini-
mize the confounding effect of gas flow on the results. The figure shows
that controlling at a reaction tank pH of 5.9 results in a removal efficiency
of roughly 85 percent. This removal efficiency is the approximate level the
scrubbers are required to operate at to achieve the compliance emission limit
of 1.2 Ibs S02 per million Btu. As will be discussed in the next two sub-
sections, this high of a pH leads to scaling in the mist eliminators (ME) and
poor reagent utilization.
Figure 3 shows the effect of unit load (actually gas flow rate) on the
stack S02 emission rate. Above a load of about 320 MW, the emission rate
begins to exceed the compliance limit. At a load of 200 MW, the average S02
removal efficiency approaches 90 percent. At full load (400 MW), the removal
drops off to about 75 percent. The data were taken at an average pH setpoint
of 5.7.
Organic Acid Testing
Pilot scale testing at the EPA-IERL facility, prototype testing at the
Shawnee Alkali FGD Test Facility, and the full scale demonstration program at
3-5
-------
90
o
c
Q)
"o
LU
O
E
-------
2.0-
1.5-
ID
-I— •
GO
CO
o
1.0-
o
Duck Creek SO2
Compliance
Limit
o
05
H— «
C/D
0.5-
0
200
250 300 350 400
Unit Load, MW
Figure 3. Unit load vs. stack S02 emissions - baseline test period
3-7
-------
City Utilities Southwest Power Plant documented the feasibility of using
organic acids to enhance the S02 removal efficiency of lime/limestone scrub-
bers. The organic acids act as buffers when dissolved in the liquid phase of
the reaction tank slurry. As buffers, they prevent a large drop in pH at the
gas-liquid interface which naturally occurs due to the absorption of SC>2 •
This improves the mass transfer rate of S02 from the gas to the liquid by re-
ducing the liquid phase resistance. The higher the dissolved concentration
of the organic acids, the greater the improvement in mass transfer and there-
fore the better the S02 removal efficiency. The upper limit in improvement
is set by the gas-liquid contacting efficiency which determines the magnitude
of the gas phase resistance. Of course, the higher the dissolved buffer con-
centration, the higher the addition rates and therefore annual costs of the
raw material. Testing was conducted to determine the concentration of ad-
ditive required to achieve the desired S02 removal efficiency.
To accomplish this, the organic buffers tested were pumped directly into
two of the four reaction tanks at measured rates. The concentration in the
liquid phase in each was measured by an acid-base titration procedure. At
the same time, S02 removal, unit load, and pH data were collected for cor-
relation purposes.
Two types of buffers were tested. The majority of the testing was with
a material commonly referred to as dibasic acid or DBA, which is a coproduct
resulting from the manufacture of adipic acid. It is an aqueous mixture of
adipic, glutaric, and succinic acids, all equally as effective a buffer in a
limestone scrubber. The other type of organic additive tested was composed
primarily of adipic acid and hydroxycaproic acid, a monobasic molecule.
Testing with this material, commonly called acid water, was limited due to
severe reaction tank foaming caused by a component in the mixture.
Figures 4 and 5 present the observed effect of DBA concentration on S02
removal for Modules A and B, respectively. All data shown were taken at unit
loads above 360 MW (90% of maximum) . The data taken while the reaction tanks
were being controlled at a pH of 5.4 to 5.7 is separated from that taken
during operation in the 5.7 to 6.0 range. For operation in the higher pH
range, Module A required approximately 200 ppm DBA to maintain removal
efficiencies above 85 percent. Less DBA was needed in Module B (100 ppm)
to get 85 percent removal in the higher pH range. Data were collected at
the lower pH's for estimating the trade-off in limestone versus DBA consump-
tion. For operation in the lower pH range, approximately 400 ppm DBA is
required for Module A and about 300 ppm for Module B for 85 percent removal.
The reason for the better performance for B is unknown since liquid recycle
rate and gas flow measurements for each were virtually identical.
Because Duck Creek uses a disposal pond instead of a thickener/filtra-
tion step for dewatering waste scrubber sludge, little of the buffer added
is returned to the scrubbers. The pond has roughly 100 times the volume
of a typical thickener. Estimates of the loss rate of DBA from the reaction
tanks in the waste slurry were compared with the actual feed rates to deter-
mine the extent of DBA losses through chemical degradation, coprecipitation
with the sulfur salts, or in fugitive streams. The results showed that
-------
toon
o
E
CD
CC
c
CD
CD
Q.
95.
90
y 85.
80
75
• pH 5.4 to 5.7
• pH 5.7 to 6.0
—i 1 i i 1 1 > i ' 1 1 1 1 1 i
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500
DBA Concentration (ppm equivalent adipic acid)
Figure 4. Effect of DBA concentration on percent S02 removal—Module A
3-9
-------
100'
95.
o 90
o
CO
"c
0)
a ss
0)
Q_
80
75
pH 5.4 to 5.7
pH 5.7 to 6.0
0 100 200 300 400 500 600 700 800 900 1000
DBA Concentration (ppm equivalent adipic acid)
Figure 5. Effect of DBA concentration on percent SC>2 removal—Module B
-------
within the accuracy of the measurements all of the material fed could be ac-
counted for in the waste slurry which left the reaction tank. This informa-
tion is important because for closed loop FGD systems (those designed with
thickener/filtration solids dewatering) loss rates over 5 times higher than
filter cake liquor loss rates have been observed for nearly equivalent DBA
concentrations. This difference is probably due to the relatively short
liquid phase residence time in the CILCo system (10 hours) compared to a
closed loop system (weeks).
Based on these results, annual DBA consumption rates can be estimated
for the Duck Creek FGD system. Required concentrations for the two modules
not tested (C and D) are assumed to be the same as for Module A. Assuming
300 days per year online and 12 hours per day operation of the unit at or
above 90 percent of full load during which time DBA must be fed, 1.26 x 106
pounds of DBA (dry basis) are required for a low pH setpoint (5.5) and
630,000 pounds for a high pH setpoint (5,8). Using a DBA delivered cost of
29.5 cents per pound, dry basis, the annual cost difference between the two
is roughly $200,000. The section on Limestone Utilization discusses the
tradeoffs with annual limestone comsumption.
Magnesium Testing
Testing to examine the effectiveness of using magnesium to enhance
scrubber S02 removal efficiency was conducted in August, 1982. Dolomitic
lime was used as the source of magnesium to determine if it was a cost ef-
fective alternative to magnesium hydroxide. A detention slaker was installed
to hydrate the lime (calcium fraction only) prior to its being fed to the
scrubber reaction tank tested. In the reaction tank the magnesium oxide
dissolves, increasing the slurry liquid phase magnesium concentration.
Because magnesium sulfite is much more soluble than calcium sulfite, the
liquid phase sulfite concentration increases depending upon how much magne-
sium is present. At the same time, the magnesium also pairs with the dis-
solved chloride and sulfate ions which reduces its effectiveness. The in-
crease in sulfite_improves the liquid phase alkalinity of the slurry by
increasing the SQj species concentration and the MgSOs complex concentration.
Both act as buffers in the same way as the organic acids, promoting the mass
transfer of 862 from the gas to liquid phase.
Figure 6 presents the relationship observed between 862 removal and
dissolved magnesium concentration. No noticeable improvement in S02 removal
was seen below about 1000 ppm of dissolved magnesium. This is probably due
to the chloride and sulfate association. Approximately 1500 ppm
of dissolved magnesium was sufficient to maintain Module A at 85
percent S02 removal. Mass balance calculations verified that all of the
solid magnesium added to the reaction tank dissolved. The dolomitic lime
appears to be a viable means of adding magnesium to the scrubbers, therefore.
During the testing, the reaction tank pH was controlled at a relatively high
pH (5.8 to 6.0) to maximize the SOf and MgSOs complex species concentrations.
Dropping the reaction tank pH to 5.4 would theoretically shift the sulfite-
bisulfite equilibrium such that about a 30 percent reduction in concentration
of the S03= and MgSOs species would result. Actual test data were not avail-
able to verify this because of difficulties in operating the slaker for
8-11
-------
100 —i
90 —
(O
>
o
-------
extended periods. However, if accurate, a 30 percent increase in the magnesium
feed rate would be required to offset the reduction. This would be in ad-
dition to the increase in magnesium required to make up for the loss in scrub-
bing efficiency due to the lower pH operation. Recall that roughly a two fold
increase in DBA was required in dropping from a pH setpoint of 5.8 down to
5.5. The reason for this drop in efficiency is a combination of the reduction
in liquid phase alkalinity and of the reduction in solid phase alkalinity
(limestone) due to less excess limestone present in the slurry at lower pH's.
The tradeoff of low pH operation/improved limestone utilization is not favorable
for magnesium as it is for DBA.
Mass balance calculations also showed that due to the method of solids
disposal at Duck Creek, high feed rates of magnesium would be required until
steady state could be reached in the entire system. Depending upon the exact
volume of the disposal pond, this could require a number of years. To im-
prove the economics of magnesium-enhanced scrubbing, recycle of scrubber
blowdown liquor was investigated. The ability of hydroclone classifiers to
partially dewater the scrubber wastes was tested. Because of the extremely
fine size of the solids, small diameter classifiers were necessary to make
the desired separation. Results showed that the slurry could be concentrated
from 10% solids up to about 40% solids, reducing the magnesium consumption by
a factor of 4. The effect of a recycle step on the operation of the scrub-
bers is unknown and could not be investigated short of major mechanical modi-
fications. The reduction in DBA consumption would be somewhat less than that
for magnesium since the fraction of organic buffer chemically degrading or
coprecipitating with the sulfur would expect to increase as its residence
time in the absorber loop increased. The effect of this recycle step on the
economics of these two alternatives is discussed in the Cost Section later.
MIST ELIMINATOR SCALING
Prior to the test program, mist eliminator (ME) scaling was causing a
severe reliability problem for the Duck Creek FGD system. In an attempt to
control the scale, each module's mist eliminator was cleaned weekly by a crew
of maintenance men. Usually the scale growth rate was so fast that within a
few days, solids pluggage would cause a significant increase in the ME gas
pressure drop. The restriction resulted in a maldistribution of gas flows be-
tween the four absorbers. Higher gas flows through a recently cleaned tower
would result because of the lower resistance to flow through its ME. The
higher gas flows would reduce that absorbers SC>2 removal efficiency as well as
carry up greater quantities of mist to the ME. The increase in removal ef-
ficiency due to the lower gas rates through the restricted towers did not com-
pletely offset this reduction and overall the removal would drop. Trouble with
completely isolating the flue gas from the inside of the towers during cleaning
periods required that some gas bypass the system. Therefore, the mist elimina-
tor scaling was not only causing reliability problems, but it was also in-
directly causing compliance problems.
A brief period of time was spent early in the program testing the ef-
fectiveness of adding a gypsum scale inhibitor to the ME wash water. Even at
higher than recommended dosage rates, no effect on either the formation rate
or strength of the scale were seen.
8-13
-------
At the same time, work began to identify the causes of the scaling and
potential solutions for eliminating it. Regular inspections of the mist
eliminators verified the rapid buildup of solids on their surface. Analysis
of the solids showed that they were primarily gypsum with small percentages
of calcium carbonate and calcium sulfite present. The wash rates were
estimated to be about 1100 pgm with the front and back of the first stage ME
and front side only of the second state being washed. With the ME's having a
front face area of about 480 ft2, this resulted in a specific wash rate of
about 0.75 gpm/ft2. Each mist eliminator was washed for 5 minutes every
twenty minutes.
During the plant outage prior to the test program each ME was thoroughly
cleaned and all wash nozzles and headers were cleaned or replaced. Several
weeks into testing, a significant number of spray nozzles were requiring re-
placement each week. The nozzles were plugging with scale themselves. Inspec-
tion of the inside of the spray headers showed that a thick gypsum scale was
growing on the walls. The pluggage was reducing wash rates and this combined
with the plugged nozzles was accelerating scale formation on the ME's. The
wash water which was reclaimed disposal pond liquor was analyzed for major ca-
tions and anions and the results input to the Radian Inorganic Equilibrium
Program. The model predicted a gypsum relative saturation of 1.9 which is
significantly higher than the level at which gypsum scaling usually initiates
(1.4 relative saturation). From this analysis, it was obvious that the re-
claimed pond liquor could no longer be used for washing the ME's.
The only other source of water available was fresh water from the
cooling lake. Using it at a rate of 1100 gpm would cause significant over-
flow of the disposal pond which was not allowed. Results of previous work
with mist eliminators at the Shawnee Test Facility indicated that minimum
ME wash rates could be achieved using a combination of fresh water (or a
water low in TSS, dissolved calcium, and dissolved sulfate) , good limestone
utilization, and sufficient wash intensity. Based on the recommended wash
rates and durations and on the dimensions of the Duck Creek ME's, a total of
only 100 gpm of fresh water would be needed to keep them scale free if the
limestone utilization was kept above 85 percent.
After the necessary piping changes were made to switch to fresh water,
durations and frequencies were slowly reduced to find the minimum required
to keep the mist eliminators clean. To help offset the impact on the plant
water balance, scrubber recycle pump seal water was closely monitored to
eliminate excess flush rates and hoses used for general cleaning around the
plant were switched from fresh water to pond reclaim water. The wash rate had
been reduced to about 300 gpm by the end of August. Scaling and pluggage of
the mist eliminators were virtually eliminated.
Further optimization of the washing was possible, though. The front of
the first mist eliminator was not being washed with the recommended intensity.
More extensive piping and valve changes were required for which there was not
sufficient time during the 1982 testing. The Shawnee results called for six
minutes of washing every four hours at a specific wash rate of 1.5 gpm per
ft" of ME face area for the front of the first ME (nozzle pressure 41 psig).
The recommendation for the back of the front ME was 4 minutes of wash every
8 hours at a specific rate of 0.55 gpm/ft2 (nozzle pressure 13 psig). There
-------
was no recommendation for the second bank ME wash, but it was assumed to be
equal to or less than that for the back of the first bank. The piping con-
figuration for the Duck Creek fresh water wash resulted in the front and back
of the first ME and the front of the second being washed simultaneously at a
specific rate of 0.7 gpm/ft2 and a nozzle pressure of less than 10 psig.
Later the valving was changed so that the first ME was washed separate from
the second ME. This increased the pressure available for the front wash but
not up to the recommended level.
LIMESTONE UTILIZATION
Initial measurements of the limestone utilization in the Duck Creek
scrubbers showed that above a reaction tank pH of 5.4, the utilization dropped
rapidly. The measurements were made by analyzing a sample of the reaction
tank solids for calcium, sulfite, sulfate, and inert material. Utilization
is defined as the mole ratio of total sulfur to calcium. It therefore serves
as a good indicator of how efficiently the scrubbers are using the limestone
fed to the reaction tanks. Including liquid phase calcium and sulfur
species results in less than one percent change in the reported utilization.
Therefore, all numbers are based on solid phase analysis only.
The low utilizations resulting from operation above pH 5.4 were hurting
the FGD systems operation and performance in the following ways:
• reliability — excess limestone in the slurry contributes
to mist eliminator scaling,
• operating costs — excess limestone could substantially in-
crease annual limestone costs, and
• waste disposal — presence of the excess limestone in the
scrubber blowdown was consuming a significant volume of
the disposal pond.
There were two options available for increasing the limestone utiliza-
tion. First, the reaction tank pH could be controlled at or below 5.4. To
do so though, would lower the SOz removal efficiency of the scrubbers sig-
nificantly (see Figure 2) and require higher DBA or magnesium concentrations
to make up the difference. The second option would be to modify the ball
mill circuit operation to produce a finer and therefore faster dissolving
limestone.
During the baseline testing, the product from the grinding circuit was
sieved with a 200 mesh and a 325 mesh screen. The mill circuit was original-
ly designed to grind 40 tons per hour of pebble limestone and produce a
material of which 90 percent would pass a 200 mesh screen. Sieve results
during baseline testing showed that on the average 86 percent passed the 200
mesh screen and 70 percent passed the 325 mesh screen. Review of past sieving
data showed that the product varied anywhere from 75 percent passing 200 mesh
to 90 percent passing 200 mesh (no data on the percentage passing 325 mesh) .
The first and most obvious change examined to improve the grind was to
decrease the limestone throughput. Based on calculated 862 absorption rates,
3-15
-------
the mill throughput could be decreased from 40 tons per hour to 20 tons per
hour and enough reagent be prepared to keep up with the scrubber demand. After
several days of operation at the lower throughput, no measurable change was
seen in the product particle size. This suggested that the classifiers were
limiting the product fineness.
A series of operational and minor equipment changes were made during June
and July with the classifiers and the mill which resulting in a significant im-
provement in the product particle size distribution. The testing was not de-
signed as a parametric investigation but instead it was conducted as an opti-
mization program. The objective was to produce the finest particle size with
only relatively small changes to the existing circuit. The result was a mill
circuit that consistently produced a material that greater than 99 percent
passed 200 mesh and 90 percent passed 325 mesh.
Although quantification of the effect of each change on the produce
particle size is not possible, the following variables were changed in the
mill/classifier circuit optimization:
« mill throughput,
• mill slurry solids content,
• level of ball charge in mill,
• top-size of balls in mill charge,
• raw limestone feed size,
« diameter of classifier barrel,
• classifier pressure drop (or throughput),
• solids content in classifier feed, and
• diameter of classifier vortex finder and apex.
The most extensive equipment change made during the program was to re-
place the 10 inch diameter classifiers with six inch diameter models. De-
creasing the diameter of the cyclone barrel theoretically would reduce the
size of the largest particle in the product stream by as much as 10 to 30
microns. For comparison, the 200 mesh screen has square openings that are
74 microns in dimension and the 325 mesh screen has openings that are 45
microns square.
The effect of limestone particle size on limestone utilization in the
scrubbers was significant. Utilization data taken during the grinding cir-
cuit testing was separated into three categories according to limestone
particle size. Figure 7 presents the results. The data show that as scrub-
ber pH increases, limestone utilization drops off dramatically. It also shows
that at a constant pH, decreasing limestone particle size improved limestone
utilization significantly. For example, at pH 5.8, the utilization averaged
about 58% with the coarse limestone being fed to the scrubbers, while it
-------
90
80-
c
o
(O
N
* 70
o
0)
CD
E
60.
50-
40.
• Coarse Grind
A Medium Grind
• Fine Grind
d = DBA present
m - magnesium present
5.0
5.2 5.4
5.6 5.8
Scrubber pH
6.0
—i
6.2
Figure 7. Effect of limestone grind and scrubber pH on limestone utilization
8-17
-------
averaged about 92% with the fine limestone. Representative particle size
distributions for the three grinds are shown in Table 1. The scatter in the
data presented in Figure 7 is primarily due to the variation in the particle
size distribution of the limestone within these three categories.
Another important observation made is that the presence of either DBA or
magnesium in the reaction tank had no effect on measured limestone
utilizations. Samples taken during either DBA or Mg testing are identified
in the figure.
TABLE 1. REPRESENTATIVE PARTICLE SIZE DISTRIBUTIONS
FOR COARSE, MEDIUM, AND FINE GRINDS
Particle Size Distribution
Limestone Size % Passing 200 Mesh % Passing 325 Mesh
Coarse 86 73
Medium 94 81
Fine 97 87
As discussed earlier, the improvement in limestone utilization was
achieved by relatively minor operational and equipment changes. The primary
costs associated with these changes include the capital for the six inch dia-
meter classifiers which totaled less than $10,000 installed, an additional
$40,000 per year for electrical power to operate the mill at lower through-
puts, and an increase in maintenance costs which can be determined only by
time. The benefits by comparison are much more significant. The Duck Creek
scrubbers are now being operated at a pH of 5.8. By controlling at this set
point the scrubbers not only operate at a higher S02 removal efficiency
(nearly 82%) compared to previous operation at pH 5.6 but also achieve
greater than 90% limestone utilization. In terms of limestone consumption,
the Duck Creek scrubbers will use almost 30 million pounds per year less than
before. And since this limestone was passing through the scrubbers unused,
the waste disposal pond will see an equivalent reduction in material sent to
it, thereby extending its useful life. Over a 25-year operating period this
would mean a reduction in volume of solid wastes of almost 350,000 cubic
yards. This is roughly equivalent to the settled volume of FGD waste pro-
duced by the scrubbers in a 5 year period.
COST STUDY
Based on the results of the 1982 test program, a cost study was com-
pleted on the economics associated with using DBA and magnesium to improve
the Duck Creek scrubbers S02 removal efficiency. A total of six options were
studied-three each for DBA and magnesium. They investigated the difference
between automatic and manual control of additive feed as well as the effect
of recycling scrubber blowdown to reduce makeup rates. To simplify the work,
a constant reaction tank pH of 5.8 was assumed.
Capital and annual costs were estimated for each case. Based on these
costs, a present worth analysis of the various options was performed using
the following equation:
3-18
-------
where S is the cost for any year n and I equals the discount rate. For this
study, CILCo's -long term escalation factor of 8 percent was used to escalate
both capital and operating expenses. CILCo's long term discount rate of
10.75 percent was chosen for I.
The results of the present worth analysis which is based on a fifteen
year period are shown in Table 2. Also shown in the table are the capital
investments for each option. The numbers indicate that without the hydro-
clones to recycle some of the reaction tank blowdown liquor, the magnesium
options are not competitive. The reason for this is that much higher liquid
phase magnesium concentrations are required than for DBA. Therefore the
losses in the waste stream are proportionally higher.
The lowest cost option involves the use of DBA in conjunction with an
automatic feed controller and hydroclones for recycle. Both Case 2 and 6
follow next and have about the same present worth cost over the 15 year
period at $2.8 and $2.4 million each. Case 2 is the DBA option with auto-
matic feed control. The automatic system consists of a microprocessor that
receives a signal from the stack gas SC"2 monitor. The microprocessor con-
trols the feed of DBA based on the S02 emission level. As the emissions
approach the 1.2 lb/105 Btu standard the microprocessor would operate a valve
to feed DBA to the reaction tanks. As the 862 emission level drops below a
predetermined value, the microprocessor shuts off the feed. This system
would operate the same for the magnesium option. The constant DBA addition
option, Case 1, follows at a cost of $3.0 million.
From an operation and control standpoint, the constant DBA addition
option, Case 1, is the easiest to operate. Addition of an automatic con-
troller complicates the system to a small degree but would pay itself off in
a short time (approximately two years). Addition of hydroclones would in-
crease the mechanical complexity of the system. Because of the large number
of hydroclones which would be required, operating and maintenance related
problems would increase as well. Operation of a slaker to prepare the dolo-
mitic lime prior to addition to the reaction tanks makes the magnesium auto-
matic feed/hydroclone recycle option the most mechanically complex system.
The operation of a slaker requires close operator supervision. For this
reason, purchase of a hydrated lime would be advantageous. With this
approach, the material could be fed dry to the reaction tanks or could be
mixed with water and pumped as a slurry. Nonetheless, bulk solids handling
is inherently more difficult than handling small volumes of an aqueous buffer
solution making DBA more attractive from an ease of operation standpoint.
EVALUATION TRIP
A return trip was made to Duck Creek approximately one year following the
'•ompletion of the 1982 test program. The objective of the trip was to document
the performance of the FGD system for comparison to the observations made
during the 1982 testing. The results were used to evaluate the ability of the
plant personnel to maintain the recommended operating setpoints. They also
served as a basis for recommending any changes in controlling the scrubbers
TO further improve their operation.
8-19
-------
TABLE 2. COST ANALYSIS OF DUCK CREEK S02 REMOVAL ENHANCEMENT OPTIONS^
No.
1
2
3
4
5
6
00
Case Description
Constant DBA Addition
DBA Feed Controller
DBA Feed Controller Plus Hydroclones
Constant Magnesium Addition
Magnesium Feed Controller
Magnesium Feed Controller Plus Hydroclones
Thousands of
Capital Investment
110
150
240
140
180
270
Dollars (January 1983)
Cumulative Present Worth3
3,000
2,800
1,400
10,800
10,000
2,400
ho
o
*3
Escalation factor of 8% was used for both fuel and non-fuel related items.
Discount rate assumed was 10.75%.
-------
There were several significant observations:
1) The scrubbers were needing less DBA than anticipated
to maintain compliance level S02 removal ef-
ficiencies .
2) No mist eliminator scaling or pluggage had been
seen in the past year.
3) Limestone utilization was not as high at a given
pH as had been seen in 1982.
4) FGD system reliability was 96 percent over the period
July, 1982, through June, 1983.
5) Limestone product screen analyses consistently
showed better than 90 percent passing 325 mesh.
6) A five ton per hour increase in limestone feed to the
ball mill had little effect on product size distribution
and no noticeable effect on limestone utilization.
The DBA usage and limestone utilization observations were the most
interesting. The tradeoffs between low pH/high DBA and high pH/low DBA
operation were reviewed. Based on the data collected during the trip, it
would be to CILCo's advantage to lower the reaction tank pH setpoints to 5.6
and increase the DBA feedrate. This is because the savings in limestone due
to improved utilization at the lower pH more than offsets the increased DBA
consumption. Operation below pH 5.6 was not recommended because the pH/
utilization relationship begins to flatten below that point (see Figure 7).
The screen analyses showed that the mill circuit was consistently pro-
ducing a product with over 90 percent passing 325 mesh. Mill throughput was
increased from 20 tons per hour to 25 tons per hour to see if this affected
the product fineness or limestone utilization in the scrubbers. The amount
passing a 325 mesh screen dropped from 92 percent down to 89 percent. The
affect on utilization could not be measured if there was any. The resulting
25 percent decrease in mill operating time not only will save on power costs,
but should also reduce equipment maintenance.
In general, the control of the FGD systems key setpoints was very good.
As mentioned, the limestone mill circuit product fineness could be consis-
tently maintained depending on throughput. Reaction tank pH's were also
being controlled within a very small range with good accuracy. Overall, this
kept limestone utilization as constant as possible with a single loop (reac-
tion tank) design. Measurements of the solids content in the reaction tanks
indicated that the density control system was generally working well. Almost
all the measurements fell between 10 weight percent and 13 weight percent
with the average being 12 percent solids. Partly as a result of this, the
FGD system reliability has averaged slightly above 96 percent since July of
1982 and over 99 percent since January of 1983. At the same time, FGD sys-
tem operating and maintenance costs have dropped about 30 percent.
5-21
-------
TECHNICAL/ECONOMIC FEASIBILITY STUDIES FOR FULL
SCALE APPLICATION OF ORGANIC ACID TECHNOLOGY FOR
LIMESTONE FGD SYSTEMS
J. C. Dickerman, J. D. Mobley
-------
TECHNICAL/ECONOMIC FEASIBILITY STUDIES FOR FULL SCALE APPLICATION
OF ORGANIC ACID TECHNOLOGY FOR LIMESTONE FGD SYSTEMS
by: James C. Dickerman
Radian Corporation
3024 Pickett Road
Durham, NC 27705
J. David Mobley
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
The application of organic-acid buffer enhancement to flue gas
desulfurization (FGD) systems is a recent development that has resulted in
lowered costs and improved performance for those systems that have adopted
its use. A process which uses organic acids as an additive has several
advantages over conventional limestone scrubbing systems. These advantages
include improved SO. removal, decreased limestone consumption, increased
system flexibility fe.g., ability to respond to unplanned fluctuations in
coal sulfur content), and improved process reliability. This paper
summarizes the results of several cost analyses which were performed to
evaluate the potential economic benefits of converting operating FGD systems
to organic-acid-enhanced limestone scrubbing systems. Also, since the last
FGD symposium, two full-scale utility limestone scrubbing systems have
converted to organic acid enhanced operations. A summary of the first year
of operation for one of those systems—City Utilities Southwest Power Plant
(SWPP)—is also included.
BACKGROUND
The ability of a limestone FGD system to remove SO- is limited by the
absorptive capacity of the slurry liquid that makes contact with the flue
gas. Adding an organic acid to the system expands the capacity of the
liquid to absorb S0_ by buffering the pH in the absorber. These are the
reactions that take place:
Sulfur dioxide + Water «- Acidity + Sulfites (1)
Organic buffer anion (carboxylate ion) + Acidity + Organic acid (2)
8-23
-------
As the buffer reacts with the acidity formed in the SO absorption reaction,
a weak organic acid is formed (Reaction 2) and SO is removed (Reaction 1).
The buffer is regenerated in the reaction tank by adding limestone:
Organic acid + Sulfites + Limestone •*• Organic buffer anion
+ Calcium sulfite solids + Carbon dioxide (3)
Adding organic buffers can result in increased S02 removal, a decreased risk
of chemical scaling, potentially decreased limestone consumption, and
potentially reduced mist eliminator fouling because of increased utilization
of limestone.
When EPA realized that organic buffers could possibly to used to solve
some of the maintenance and cost problems that utilities were experiencing
with their lime/limestone FGD systems, the Agency began an extensive
research program, designed to provide a thorough evaluation of various
organic acid additives and how they can be used to improve lime/limestone
scrubbing. It consisted of four phases—laboratory-, pilot-, prototype-,
and commercial-scale.
The EPA began its research by sponsoring theoretical and laboratory
investigations of a number of different acids. As a result of these
studies, adipic acid appeared to be the most promising buffer candidate
because of its solubility in water, low volatility, chemical stability,
non-toxicity, high availability, and low cost. Also, the region of maximum
buffering for adipic acid, between pH 4 and 6, is ideal for limestone FGD
systems.
Two full-scale test programs were also sponsored by EPA which
demonstrated the technical and economic viability of this approach on
operating utility and industrial FGD systems. Results of both of these
demonstration programs have been reported in previous FGD symposia.
TECHNICAL AND ECONOMIC ANALYSES OF OPERATING SYSTEMS
Technical and economic analyses were made for four FGD installations to
determine the nature and cost of converting each to an organic acid enhanced
limestone scrubbing system. The four installations are: City Utilities'
Southwest Power Plant, Central Illinois Light Company's Duck Creek Station,
San Miguel Electric Cooperative's San Miguel Station, and Big Rivers
Electric Corporation's R. D. Green Station. The cost analysis of the
R. D. Green Station was not completed in time for this paper. The results
will, however, be presented at the FGD Symposium.
The approach used in conducting these process evaluations was basically
the same for all systems. First, typical process operating conditions for
each system were identified through a series of questionnaires and meetings
with plant personnel. Operating parameters such as pH, limestone
utilization, liquid to gas (L/G) ratios, and SO removal were obtained from
3-24
-------
each system to establish baseline process operations and baseline operating
costs. Next, technical evaluations were made to identify any process
modifications that would be required to convert each system to an organic
acid enhanced system, and to establish organic acid makeup requirements.
Finally, the capital investment costs for all required process
modifications, and the annual operating costs of organic acid enhanced
operations were determined. The analysis results for each system are
summarized below.
CITY UTILITIES—SOUTHWEST POWER PLANT (1)
City Utilities' (CU) major interest in converting their system to an
organic-acid-enhanced limestone scrubbing system was to improve the S0~
removal performance so that they could comply with the applicable S09
removal requirements. Several process alternatives were considered, each of
which had the potential for achieving the utility's goal of increased S0?
removal. The process alternatives considered included: increasing the L/G
ratio to the tray tower, adding adipic acid, and adding a mixture of
by-product organic dibasic acids (DBA). The DBA used at CU is a by-product
of adipic acid production and consists of a mixture of adipic, glutanic, and
succinic acids. The higher L/G option represented a capital cost intensive
solution; whereas, the adipic acid and DBA options have lower capital costs
but higher annual operating costs. A present worth analysis was performed
using escalation factors consistent with CU's long range planning
activities. The escalation of adipic acid and DBA closely approximated CU's
assumed escalation rate of fuel oil, and the escalation of cost of
electricity (production costs only) closely approximated CU's assumed coal
escalation rate. The discount factor for the analysis was assumed to be
equal to the general inflation rate.
The results of this study are shown in Figure 1. Note that the adipic
acid option is a lower cost option than increased L/G for about 6 years.
The technical uncertainties in making the L/G modification, coupled with the
demonstrated adipic acid system flexibility, make the adipic acid option a
good choice. However, the DBA option remains the economic choice through
more than 15 years. Consequently, City Utilities elected to convert their
FGD system to DBA enhanced operations in December 1981. Results of the
first year of operation of that system are reported later in this paper.
CENTRAL ILLINOIS LIGHT COMPANY—DUCK CREEK STATION (2)
Central Illinois Light Company's (CILCo's) main interest in considering
a conversion to an organic-acid-enhanced limestone scrubbing system was also
related to improving the system's performance so that it would achieve
compliance with the applicable SO,., regulations. As with City Utilities,
several process alternatives were considered. The options identified for
CILCo to achieve compliance included:
3-25
-------
constant DBA addition,
intermittent DBA feed,
intermittent DBA plus recycle,
constant magnesium oxide addition,
intermittent magnesium oxide feed, and
intermittent magnesium oxide feed plus recycle.
Adlplc Add
Addition
Increased
LK3 Ratio
DBA
Addition
Time (Years)
Figure 1. Cumulative Present Cost for Southwest Power Station
The CILCo system is significantly different than the other systems
evaluated in that a waste disposal pond is used for solids disposal. In
this system, a slurry waste stream of approximately 10 percent solids is
sent to the disposal pond, thereby significantly increasing the DBA or
magnesium makeup requirements. For this reason, a process alternative,
which included the addition of a hydroclone to concentrate the blowdown
stream to 40 weight percent solids and thus reduce the additive losses, was
also examined.
5-26
-------
Results of the present worth analysis over a 15-year period are shown
in Figure 2. These results show that the DBA options cost significantly
less than the magnesium addition options, except for the case in which a
hydroclone is used to reduce the amount of scrubber blowdown sent to the
waste disposal pond. Although the process options using the hydroclone
showed cost advantages, the use of a hydroclone has not been demonstrated in
the field, and thus, technical uncertainties associated with its use exist.
For this reason, CILCo has elected to convert their process to a
DBA-enhanced limestone scrubbing system with intermittent feed controls so
that DBA is added only during high load operations.
12 -i
11 -
10 -
9 -
7 -
s -
1 -
]= Constant DBA Addition
]= Intermittent DBA Addition
]= Intermittent DBA Addition Plus Recycle
]= Constant Magnesium Addition
]= Intermittent Magnesium Addition
]= Intermittent Magnesium Addition Plus Recycle
15
Figure 2. Cumulative Present Worth of Converting ClLCo's Duck Creek Station
SAN MIGUEL ELECTRIC COOPERATIVE (3)
San Miguel's interest in converting their system to an organic-acid-
enhanced system stemmed primarily from a desire to reduce their FGD
operating costs rather than from regulatory pressures. Normal operating
conditions at San Miguel required the use of fairly high limestone reagent
3-27
-------
ratios in order to meet compliance. Reagent ratios on the order to 1.2 to
1.3 are routine, and as high as 1.4 have been measured. Besides the obvious
costs associated with excess limestone use, San Miguel also experienced
chemical scaling in their absorber tower which was probably related, at
least in part, to the excess limestone present in the system.
This paper presents the results of a cost analysis to determine the
potential cost savings associated with adding DBA to achieve an SCL removal
level of 85 percent. Due to the uncertainties in the DBA feed requirement
and the current limestone reagent ratio at San Miguel, a range of annual
operating and maintenance (O&M) expenses was used in the cost analysis.
These expenses reflect the incremental change in O&M expenses from the
current FGD operation. Maximum O&M expenses correspond to the maximum
estimated DBA requirement (400 ppm) and the minimum limestone savings (which
result if San Miguel's current limestone reagent ratio is 1.2). Minimum O&M
expenses correspond to the minimum estimated DBA requirement (300 ppm) and
the maximum limestone savings (which occur if San Miguel's current limestone
reagent ratio is 1.3). Average O&M expenses assume a DBA feed requirement
of 350 ppm and a current limestone reagent ratio of 1.25.
A present worth analysis was performed based upon San Miguel's long
term escalation and market interest rate assumptions and is presented in
Figure 3. Taking into account the initial capital investment for the DBA
system, the cost savings for the minimum and average O&M expenses were
estimated over a 15-year period at $1.5 and $0.4 million in terms of present
worth. Payback periods were calculated to be 1.2 and 2.5 years from the the
beginning of 1983, respectively. Similarly, if the maximum operating
expenses are realized, the DBA system will cost San Miguel $657,000 over the
15-year present worth period.
Note that no credit for increased reliability was assumed in any of the
cost calculations. For San Miguel, which has experienced chemical scaling
problems, the savings could be very significant. Worst case assumptions
predict a cumulative cost for a DBA system of $657,000 over a 15-year
period. If DBA allows San Miguel to remain on line for only 2 extra days
during the 15 years, it will have saved the total present worth cost of the
DBA system for the entire 15-year period.
Because of the potential cost savings and system performance
improvements associated with the use of a DBA system, San Miguel has decided
to conduct a test program to verify the results of this engineering
evaluation. Field testing of DBA-enhanced limestone scrubbing operations
is to be conducted in late 1983.
-------
1000 -
500 -
I
5
1
I
-500 -
-1000 -
-1500
-2000 -
Figure 3. Cumulative Present Worth of Converting San MlgueTs FGD
System to DBA Assisted Operations
RESULTS OF THE FIRST YEAR OF COMMERCIAL OPERATION OF AN
ORGANIC-ACID-ENHANCED FGD SYSTEM
In 1980, Radian Corporation and City Utilities entered into an
agreement to participate in a demonstration project which was sponsored by
U.S. EPA's Industrial Environmental Research Laboratory at Research Triangle
Park, NC. Adipic acid and dibasic acid (DBA) were tested as buffering
agents in the Southwest Unit 1 limestone scrubber to determine their
effectiveness in increasing utilization and S0_ removal. This testing was
conducted during 1981 and verified the effectiveness of organic acid buffers
in a commercial limestone scrubber.
It was quite important to CU officials that improved scrubber operating
efficiencies be obtained immediately. In late 1980, Region 7 of the
U.S. EPA cited the operation of Southwest Power Station Unit 1 for failure
to comply with flue gas emission regulations. By October 1981, upon
-------
completion of Radian's testing, negotiations were ongoing with EPA in an
attempt to resolve the dispute. In efforts to reach an agreement on a
"Consent Decree" City Utilities agreed to use a material, such as adipic or
dibasic acid, as an additive to enhance the reactivity of the CaCC- and thus
increase the SO removal from the flue gas. City Utilities embarked
immediately on a plan to implement the use of DBA (4). This section of the
paper describes the results of the first year of DBA operations.
DBA FEED SYSTEM
A temporary DBA feed system was installed in late December 1981 to add
the DBA solution on a continuous basis. This temporary system used a
26,000 I (6,000 gal) stainless steel tanker trailer equipped with thermal
jacketing as a storage vessel. Two residential-sized electrical water
heaters provided a source of hot water at -\> 60°C (140°F) for circulation
through the thermal jacket of the truck. The heater return lines were run
in parallel with the DBA feed line through a 10 cm (4-in.) PVC conduit in
order to prevent feed line cooling. Maintaining high DBA temperatures was a
major concern because DBA tends to easily precipitate out of aqueous
solution. As an example, a 20 percent solution of DBA will begin to show
crystal formation at 22°C (72°F); higher concentrations require even higher
temperatures to maintain solubility. The entire temporary system was
assembled on-site and tied into the scrubber system in a matter of only a
few days.
Initially, DBA was pumped from the tank truck to the ball mill sump.
The feed rate was set manually. After the DBA was mixed with the freshly
ground limestone slurry from the ball mills, the mixture was pumped to the
limestone storage tank. Since the limestone handling equipment was common
to both scrubber modules, both scrubbers operated with the same DBA
concentration. Late in September, the DBA feed system was modified to allow
DBA to be fed separately into the reaction tank of either scrubber. (5)
S02 REMOVAL
Results of a year's worth of testing using DBA as an FGD additive were
consistent with those anticipated based on the demonstration project
conducted in 1981. On the whole, FGD performance was enhanced to such an
extent that Southwest Power Station consistently operated in compliance with
the 1.2 Ib S02/10 Btu (516 ng/J) federal emission limitation on a
day-to-day basis. The few emission exceedances experienced were generally
associated with start-ups, shutdowns, and other periods of allowable
operational curtailment.
8-30
-------
Figure 4 displays a monthly history of FGD performance, expressed as
overall SCL removal efficiency, for the 1982 calendar year. It can be seen
that efficiencies were consistently greater than the approximately
82 percent removal generally required at the Southwest Station to maintain
compliance. For comparative purposes, Figure 4 also displays the
performance history for calendar year 1980, wherein removal efficiencies
were erratic and averaged only 26.0 percent. Figure 5 shows these data in
terms of average monthly emission rates. This figure illustrates the
significant improvement in the ability of the system to comply with the
regulatory emission standard after the conversion to a DBA enhanced
system. (6)
100
90
BO
70
60-
50-
40 H
30
20
10
Note: Expressed as % Removal Over Alt Hours In the Period.
1982
/\
1980
Jan Feb Mar Apr May Jun Jul Aug Sep Ocl Nov Dec
Figure 4. Comparison of Southwest Power Station SO2 Removal Efficiencies,
1980 vs. 1982
The addition of DBA also improved the system's ability to respond to
various operational changes. During a 2-week period in March, the coal
supplied to the system contained 4.5 percent sulfur instead of the
3.5 percent sulfur level for which the system was designed. During this
period, the unit was able to maintain an in-compliance status by increasing
the amount of DBA feed to the system. (7)
3-31
-------
7.0
S 6.0 •
•o
^
fl 5.0 •
•
8. 4.0-
1 3.0-
1
r« „ ..
O 2.0
en
1.0
/ 1980
x"
^"\
-' \
\ X
\ /»x x x ^x
\ / ^*v ^* x** — •*'^
\ / "
N-y
^^^ ~--^__— 19B2
- 3010
- 2580
• Z150
- 1720
• 1290
• B60
- 430
Jan F«b Mar Apr May Jun Jul Aug S«p Ort Nov Dec
Figure 5. Comparison of Southwest Power Station Monthly Average
S02 Emission Rates, 1980 v». 1982
RELIABILITY
The overall reliability of the system was also increased significantly
in 1982 in comparison with other years. Figure 6 shows the monthly average
FGD reliability for 1982 compared with the average reliability for 1979 and
1980. (8) This reliability increase may have resulted in part from DBA
addition, but was also related to a number of process improvements made
during 1981 and 1982, such as the conversion from turbulent contact
absorption towers to tray towers.
CONTINUOUS MONITORS
The performance of the continuous emission monitoring system (GEMS)
used by CU to determine their S0_ emissions was also evaluated to determine
the ability of the GEMS to gather the type of data required by the 1979
revision to the utility New Source Performance Standard. To meet EPA
requirements for data capture, at least two data points must be taken each
hour for a minimum of 18 hours each day for 22 days per month. The goal of
the CEMS at SWPP was to collect at least four data points per hour for
24 hours.
-------
£
3
i
I
8,
E
I
100
40 -
20
o
1~
o
c
1982
Average 1979-1980
Jan
Feb Mar Apr May Jun Jul Aug Sep O=t Nov Dec
Figure 6. Comparison of Southwest Power Station Process Reliability,
1982 vs. Average Reliability for 1979-1980
Monitor reliability data were collected during the last 6 months of
1982. During this time, the GEMS proved to be very reliable, capturing in
excess of 97 percent of the available data for each scrubber module.
Table 1 presents the data capture percentage for each module by month and
for the program as a whole. The lowest data capture efficiency was
94 percent during July on the S-l Module. The majority of the downtime
encountered by the GEMS was caused by power interruptions. (9)
TABLE 1. SUMMARY OF CONTINUOUS MONITOR RELIABILITY
Average
Module S-l
97.3
Module S-2
June
July
August
September
October
November
December
98.9
94.0
95.9
99.0
100.0
97.4
96.2
100.0
96.8
96.1
98.4
99.9
97.6
100.0
98.4
3-33
-------
DBA CONSUMPTION
The DBA consumption rate increased significantly during the last half
of 1982 to where it was nearly twice that observed during the 1981
demonstration tests. The increased DBA consumption was due to two main
factors: changes in the recycle slurry pH and deterioration of the slurry
recycle pumps. Slurry pH was decreased in 1982 from 5.4 to 5.1 to improve
limestone utilization in an attempt to improve mist eliminator reliability.
(Mist eliminator performance has been greatly improved by this technique and
by revising the system water balance to wash with greater amounts of fresh
water.) While the lower pH does improve limestone utilization, a higher
concentration of DBA is required to achieve the desired SO- removal. (10)
The deterioration in the recycle slurry pumps over time resulted in an
actual L/G ratio less than indicated. In order to maintain the same SO^
removal at a decreased L/G, the concentration of DBA was increased. The
pumps underwent a complete overhaul at the plant's scheduled outage in early
1983 to correct this problem.
FUTURE ACTIVITIES
Several operational improvements were implemented during the 1983
extensive maintenance outage. An example is a limestone classification
system and a permanent DBA addition system. Based on the experiences of the
Duck Creek Station of the Central Illinois Light Company in their DBA
program, it was found that by improving the efficiency of the limestone
classification system they could reduce DBA consumption on the order of
magnitude of 10 percent. During a short period of classifier testing during
November 1982, similar, yet inclusive, results on DBA usage were obtained by
CU.
With the improved operator control over the DBA feed system and the
direct feed to each hold tank, a much quicker chemical response will be
obtained which should allow consistent S0? emissions rates at or slightly
below the required 1.2 lb/10 Btu (516 ng/J) level. An increased reduction
of DBA usage is projected as the operator exercises more precise system
control. When considering the beneficial effects of the classifier system
as well, overall DBA consumption should be reduced significantly as compared
to 1982 values.
-------
REFERENCES
1. Hargrove, 0. W. , J. D. Colley, and J. D. Mobley. Adipic Acid-Enhanced
Limestone Flue Gas Desulfurization System Commercial Demonstration.
Paper presented at APCA/ASME Information Exchange, Research Triangle
Park, North Carolina. December 8-9, 1981.
2. Colley, J. D., P. A. Nassos, and S. T. Litherland. Field Investigation
of FGD Chemistry. Radian Corporation, Austin, Texas. Draft EPRI
Report. March 1983.
3. Glover, R. L. A Cost Analysis of the Conversion of San Miguel Electric
Cooperative's Limestone FGD System to Utilize Adipic Acid. EPA
Contract No. 68-02-3171, Task No. 56. Radian Corporation, Austin,
Texas. Draft Report. October 1982.
4. Hicks, N. D. and D. M. Fraley. Addition of Adipic and Dibasic Acids to
a Conventional Flue Gas Scrubber: Costs Operating and Design
Experiences. Paper presented at Twenty-Seventh Annual APPA Engineering
and Operations Workshops. San Antonio, Texas. February 14 - 17, 1983.
5. Brown, G. E., J. C. Dickerman, and 0. W. Hargrove. Results of the
First Year of Commercial Operation of an Organic Acid Enhanced FGD
System. EPA Contract No. 68-02-3171, Task No. 53. Radian Corporation,
Austin, Texas. Draft Report. May 1983.
6. Reference 4.
7. Reference 5.
8. McMahan, J. L. City Utilities Quarterly Compliance Status Reports.
Submitted to Missouri Department of Natural Resources. Calendar
Quarters 1 - 4, 1982.
9. Reference 5.
10. Reference 5.
8-35
-------
SESSION 9: WASTE DISPOSAL/UTILIZATION
Chairman: James D. Kilgroe
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
FULL-SCALE FIELD EVALUATION OF WASTE DISPOSAL
FROM COAL FIRED ELECTRIC GENERATING PLANTS
J. W. Jones, C. J. Santhanam, A. Balasco,
I. Bodek, C. B. Cooper, J. T. Humphrey, B. K. Thacker
-------
ABSTRACT
This paper summarizes results of a 3-year study of current coal ash
and flue gas desulfurization (FGD) waste disposal practices at coal-
fired electric generating plants. The study was conducted by Arthur D.
Little, Inc., under EPA contract 68-02-3167, and involved characterization
of wastes, environmental data gathering, evaluation of environmental
effects, and engineering/cost evaluations of disposal practices at six
selected sites in various locations around the country. Results of the
study are expected to provide technical background data and information
to EPA, state and local permitting officials, and the utility industry
for implementing environmentally sound disposal practices.
Data from the study suggest that no major environmental effects have
occurred at any of the six sites; i.e., data from wells downgradient of
the disposal sites indicate that waste leachate has resulted in concen-
trations of chemicals less than the EPA primary drinking water standards.
A generic environmental evaluation based on a matrix of four waste
types, three disposal methods, and five environmental settings (based on
climate and hydrogeology) shows that, on balance, technology exists for
environmentally sound disposal of coal ash and FGD wastes for ponding,
interim ponding/landfilling, and landfilling. For some combinations of
waste types, disposal methods, and environmental settings, mitigation
measures must be taken to avoid adverse environmental effects. However,
site specific application of good engineering design and practices can
mitigate most potentially adverse effects of coal ash and FGD waste
disposal. Costs of waste disposal operations are highly system and site
specific.
9-1
-------
Introduction
This study—of current coal ash and flue gas desulfurization (FGD) waste
disposal practices at coal-fired power plants—was conducted for the
U.S. Environmental Protection Agency (EPA) by Arthur D. Little, Inc.
(ADL) under EPA contract 68-02-3167. The study involved characterization
of wastes, environmental data gathering, evaluation of environmental
effects, and engineering/cost evaluations of disposal practices at six
selected sites at various locations around the country. Results of the
study are expected to provide the technical background data and information
needed to help EPA determine the degree to which disposal of these
wastes needs to be managed in order to protect human health and the
environment. The study results will also assist EPA in preparing a
report to Congress required under the 1980 Amendments to the Resource
Conservation and Recovery Act (RCRA), and should provide useful technical
information to federal, state, and local permitting officials and utility
planners on methods for environmentally sound disposal of coal ash and
FGD wastes. Results of this 3-year effort are summarized in this
paper.
Background on Waste Generation/Disposal Methods
Coal-fired power plants using conventional combustion technology generate
two major categories of waste materials. Coal ash (fly ash, bottom ash,
or boiler slag) and FGD wastes are generated in large amounts relative
to other wastes generated at these plants and, therefore, are usually
referred to as "high volume wastes." Numerous other wastes, generated
in smaller quantities, are associated with other processes or maintenance
operations in a power plant such as coal pile runoff, boiler blowdown,
cooling tower blowdown, water treatment wastes, maintenance cleaning
wastes, general power plant trash, and plant sanitary wastes. This
project primarily focused on the high volume wastes.
Fly ash from coal-fired utility boilers is collected by mechanical
collectors and/or electrostatic precipitators, fabric filters, or wet
scrubbers. By late 1982, approximately 103,000 MW of coal-fired generating
capacity—operational units, units under construction, and units at
various stages of planning—had been committed to FGD systems. Flue gas
desulfurization can be accomplished by nonregenerable throwaway systems,
which result in FGD wastes, or by regenerable systems, which produce a
saleable product (sulfur or sulfuric acid). Operational nonregenerable
FGD systems are currently predominated by wet scrubbing technology
?o«r?n«,S°me ^ FG° scrubbinS systems were becoming operational in
1982-1983 The principal types of systems used in utility power plants
are those based on direct limestone, direct lime, alkaline fly ash dual
alkali, and lime- or sodium-based dry FGD systems/''
9-2
-------
Some projections on the generation of coal ash and FGD wastes [together,
these are designated as flue gas cleaning (FGC) wastes] in the U.S. are
presented in Table 1. Most of the coal ash and all of the FGD wastes
generated are sent to disposal.(2,3) Considering the expected increase
in coal consumption in the U.S., this is likely to be the case for many
years. Utilization of FGC waste is expected to grow but at a slower
rate than FGC waste generation. A significant fraction of the total
coal ash generated is used for such purposes as soil stabilization, ice
control, and as ingredients in cement, concrete, and blasting compounds;
however, there is currently no utilization of FGD wastes in the U.S. On
balance, disposal will continue to be the major option for FGC waste
management in the U.S. for the foreseeable future.
Currently, all FGC waste disposal is on land. At-sea disposal may be a
future alternative if it can be practiced under environmentally and
economically acceptable conditions. The principal methods of disposal
on land are: ponding, landfilling (including disposal in surface mines),
and interim ponding followed by landfilling. Table 2 presents data on
current practices based on data obtained on 176 plants.
Ponding of FGC waste is more widely practiced than any other disposal
method. Ponding can be employed for a wide variety of coal ash and FGD
wastes including chemically treated FGD wastes. Ponds can be designed
based on diking or incision, but the construction of dams or dikes for
ponds is usually expensive. In the future, particularly if chemical
treatment of FGD wastes is widely practiced, ponding will probably be
limited to those sites that would involve minimal construction of dams
or dikes. One exception could be a special case of wet ponding—FGD
gypsum "stacking." In this case, gypsum slurry from a forced oxidation
system would be piped to a pond and allowed to settle and the supernate
recycled. Periodically the gypsum would be dredged and stacked around
the perimeter of the pond, thus building up the embankments.
Landfilling of FGC waste is also widely practiced, and can involve one
or more of a variety of handling operations prior to the disposal
operation. For example, bottom ash is almost always sluiced from the
plant, so it must be dewatered (e.g., via hydrobins) before it is
transported. Dewatering must also be applied to fly ash that is sluiced
from the plant or is wet-scrubbed from the flue gas—with or without
significant quantities of SO,.,. Wet FGD waste must also be dewatered via
thickening, vacuum filtration, and, if necessary, blending with dry fly
ash for stabilization or other chemical treatment ("fixation") additives
such as lime. On the other hand, fly ash slated for landfill is typically
transported directly from the plant in a dry state, with only enough
moisture added as required for dust control and compaction in the
landfill. Wastes from a spray dryer FGD system can also be transported
directly; during this project, commercial operation of these systems on
utility boilers was just beginning.
9-3
-------
TABLE 1. PROJECTIONS OF FGC WASTE GENERATION
BY UTILITY PLANTS IN THE UNITED STATES
(1980-1995)
Waste Type
Waste Generation (10 Metric tons/yr)
1980 1985 1995
Coal Ash*
FGD Wastes
TOTAL
62.4
8.6
71.0
(78.3)c
83.2
26.9
110.1
(121. 4)c
110.0
48.6
158.6
(174. 8)c
Coal ash quantities are shown on a dry basis.
FGD waste quantities are shown on a wet basis (50% solids).
10 tons/year.
Source: Reference 2
TABLE 2. CURRENT FGC WASTE DISPOSAL METHODS UTILIZED AT
UTILITY COAL-FIRED POWER PLANTS IN THE U.S.
(Data Base: 176 Plants > 200 MW)a
Type of Waste
Fly ash only
Bottom ash only
Combined fly and bottom ash
FGD waste only
Mixed fly ash and FGD waste
Mixed bottom ash and FGD waste
Mixed fly ash and FGD waste (stabilized)
Mixed fly ash, bottom ash, and FGD waste
Pondb
18
29
69
5
7
1
2
2
Number
Landfill0
46
13
9
-
7
-
6
1
of Plants
Interim Pond/Landf 111°
6
29
16
_
_
1
_
1
Coal-fired plants on which data were available (>_80% of their power generated from
coal in 1977) which have generating capacities >_200 MW with the exception of four
plants employing FGD systems. Figures represent the number of plants at which each
waste type/disposal method is practiced. (Note: Many plants utilize more than one
method.)
Includes direct ponding and interim/final ponding methods.
Includes managed and unroanaged fills and mine disposal.
Source: Reference 4
9-4
-------
In a landfill disposal site, the wastes are spread on the ground in 0.3
to 1 m (1 to 3 ft) lifts and compacted by wide track dozers, heavy
rollers, or other equipment. Layering proceeds in 0.3 to 1m lifts in
segments of the site. The ultimate height of a disposal fill is site
specific, but can range from 10 m (30 ft) to as high as 76 m (250ft).
A properly designed and operating dry impoundment system can potentially
enhance the value of the disposal site after termination or at least
permit post-operational use.
Mine disposal is a variation of landfilling that is receiving increased
attention. Surface coal mines, particularly those serving "mine-mouth"
plants, offer the greatest capacity and economic attractiveness for
disposal of wastes from power plants.(5,6) Since the quantity (volume)
of FGC wastes produced is considerably less than the amount of coal
burned, many mines would have the capacity for disposal throughout the
life of the power plant. Several plants, particularly in the Plain
States (e.g., North Dakota), have practiced this disposal method in
recent years.
Interim pond/landfill has been an important waste disposal method in the
past, but is likely to decline in importance in the future, particularly
since dry ash handling and disposal is being more widely practiced.
SITE SELECTION AND TEST PLAN PREPARATION
Candidate Site Selection Process
The overall objective of the candidate site selection process was to
evaluate available data on coal-fired power plants and recommend a
number of candidate and backup sites. The selection process consisted
of two steps:
First, the contiguous 48 states were divided into 14 physiographic
regions, and the plants in each region were screened to develop a list
of plants that would be suitable for consideration as candidate and
backup sites. A total target of 25 to 30 sites, including 18 candidate
and 7 to 12 backup sites, was desired. Based on an assessment of
present and future FGC waste disposal practices, a preliminary distribution
of the targeted number of candidate sites in each region was agreed
upon. In screening selections, the investigators remained cognizant of
the targeted number in each region, but were not absolutely limited by
that number. The attempt was to choose desirable plants in as many
regions as possible. The list of plants in all the regions that came
through this filtration process amounted to 26.
Second, these 26 were then ranked in iterative group discussions leading
to the nomination of 18 as candidate sites and the remainder as backup
sites.
9-5
-------
Final Selection Process
The candidate and backup sites were then subjected to a more detailed
evaluation. These evaluations included one or more detailed site visits
by engineering, environmental, and hydrogeologic specialists assigned to
the project. Their findings, together with mid-course evaluations that
were continuously taking place, supported an iterative process that
resulted in the selection of the final six sites. Table 3 provides
overall information on the final six sites that were selected for
evaluation under this project; Figure 1 indicates the site locations.
Test Plan Preparation
Detailed test plans providing background information on each of the
sites, together with a description of the proposed program of site
development, physical and chemical sampling, and analysis and engineering/
cost assessments, were developed. The test plans were reviewed by EPA
and the utility involved, and their comments were incorporated. The
finalized test plans guided the work at each site.
SITE DEVELOPMENT AND PHYSICAL TESTING
After approvals from the utility and, in some cases, from state regulatory
agencies, site development was begun. Site development and physical
testing were governed by procedure manuals^ ' developed for this project.
The activities involved in site development included the drilling of
borings; excavating test pits; collecting waste, soil, and water samples;
conducting field permeability tests; installing ground water monitoring
wells and piezometers; and documenting each activity. These activities
took place at each of the six sites in time periods of 2 to 4 weeks.
Table 4 indicates the timing under which the six sites were developed
and the extent of the activities at each site. The table also gives the
number of physical tests performed; i.e., laboratory soil classification
and permeability tests on waste samples from the sites. Preliminary
water balances were also developed for each site.
CHEMICAL SAMPLING AND ANALYSIS
At each site a program of chemical sampling and analysis was undertaken
'
sampling and analysis program.
obtained from a series of
relativeiy dr-
* BWm&T^ °f the
-------
TABLE 3. WASTE DISPOSAL SITES SELECTED FOR EVALUATION
Plant (Utility)
Allen
(Duke Power)
Elrana
[Duquesne Light
(waste disposal
by Conversion
Systems, IDC.))
Dave Johnston
(Pacific P&L)
Sherburne
County
(Northern
States Power)
Powerton
(Commonwealth
Edison)
Smith
(Gulf Power)
UL - Unllned.
CL - Clay-Lined.
AL - Artificially
Nameplate
Generating
Location Capacity
State (County) (MW)
N. Carolina 1155
(Gaston)
Pennsylvania 510
(Washington)
Wyoming 750
(Converse)
Minnesota 1458
(Sherburne)
Illinois 1786
(Tazewell)
Florida 340
(Bay)
Lined (Por-0-Tec).
Startup Date
(mo/yr)
Plant (FGD)
-/57
6/52
(10/75)
V59
5/76
(5/76)
V72
6/65
High Priority Issues
Under Study
Employment
of a
Waste Site Under Study Ground Surface Potentially
Waste Type
Combined fly
and bottom ash
Stabilized
FGD waste
Combined
fly and
bottom ash
Fly ash
Fly ash/FGD
Combined fly
and bottom ash
Combined fly
and bottom ash
Disposal Water Water Mitlgatlve
Method Quality Quality Practice
Pond (UL) XX X
Landfill XX X
(UL; offsltc)
Landfill
(UL; offaite)
Landfill (UL) X X
Pond (CL) X - X
Landfill (AL) XX X
Pond (UL) XX X
-------
DAVE JOHNSTON
SHERBURNE COUNTY
-V"^
POWERTON /"' \ f .
\^
/-->;^
ELRAMA
AU.CN
FIGUBJE 1. LOCATION Of WASTE DISPOSAL SITES SELECTED FOR EVALUATION
9-8
-------
TABLE 4. SUMMARY OF SITE DEVELOPMENT/PHYSICAL TESTING
Plant
Allen
Elrama
Johnston
Sherco
Powerton
Smith
Date
Development
Completed
(mo/yr)
01/81
03/81
05/81
08/81
11/81
12/81
Number of
Laboratory Physical Tests
Borings
20
20
14
13
11
25
Number
Wells
20
16
12
11
9
24
of
Test
Pits
2
4
10
-
1
-
Soil
Samples
152
199
154
178
112
146
Unified Soil
Classification
Series (USGS)
18
17
12
20
30
15
Permeability
4
13
7
6
8
8
-------
TABLE 5. SUMMARY OF CHEMICAL SAMPLING AND ANALYSIS PROGRAM
Samples" Trip 1
Site
Allenc
Elrama
Sherco
Smith
Powerton
Dave Johnston
Trip 1
wells
ash solids
interstitial liquors
soils
wells
waste solids
soil
waste extracts
wells
waste interstitial
liquors
waste solids
liner solids
liner liquor
soil solids
soil extracts
liquids
waste solids
interstitial waste
liquors
soil
soil liquors
wells
waste solids
wells
waste solids
waste extracts
soils
Trips 2, 3, and 4
wells and
surface waters
wells, lysimeters,
surface waters
wells and
surface liquors
wells and
surface waters
wells and
surface waters
wells and
surface waters
I CAP
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Analyses
Trips 2, 3, and 4
1C As/Se Field Data Other
X
X
X
X
X
X
X
X
X
X
X
X
X
X X
X
X
X
X X
J
X Xd
X
X X
X
X X
X X
X X
X
aSamples obtained during site development and subsequent sampling and analysis trips.
Analyses performed are abbreviated as follows!
ICAP - Ag, Al, B, Ba, Be, Ca, Cd, Co, Cr, Cu, Fe, K, Mg, Mn, Mo, Na, Ni, P, Pb,
Si, Sr, Th, Ti, V, Zn, Zr. (Does not Include B, Ba, and SI
for solids-)
1C - F~, Cl~, N05, S0£, Br~, PO^3".
As/Se - either or both on selected samples.
Field Data ground water level, pH, dissolved oxygen, conductivity, temperature.
^ther samples were obtained (boiler cleaning wastes). Analysis was limited to ICAP,
pH, and bromate.
dlncludes solids characterization for SO^2 , total oxidlzable sulfur, slurry pH,
acid insolubles.
9-10
-------
Chemical samples are subjected to several types of analyses: ion
chromatography (1C) for six anions, inductively coupled argon plasma
emissions spectroscopy (ICAP) for 26 metals; and atomic absorption
spectroscopy (AA) for selected metals. As shown in Table 5, these types
of analyses were performed on a mix of solid and liquid samples for each
site. In addition, a limited number of experiments were performed to
assess the attenuative capacity of various soils obtained at the sites.
Furthermore, during the initial phase of this project, 23 grab samples
of wastes from 18 plants were obtained and analyzed using the EPA
Extraction Procedure (EP); a summary of results from these tests is
given in Table 6. Further details on these tests, as well as results of
radioactivity measurements, are included in Reference A.
SITE-SPECIFIC ENVIRONMENTAL EVALUATIONS
The data and information from site development and sampling/analysis
were subjected to environmental effects evaluation throughout the project.
The individual site evaluations were developed by a series of five
sequential steps.
First, a review and evaluation was made of available background information
on the disposal operation and its environmental setting. Second, present
disposal-related water quality effects were identified and described
based on evaluation and measured information developed in this project.
Third, apparent cause/effect relationships were hypothesized to explain
the findings at the sites. Fourth, potential future ranges of water
quality effects were considered to the extent that suitable data were
available. Finally, industry-wide implications of the findings at the
individual sites were considered in the generic assessment, discussed
later in this paper.
Environmental evaluation of all six sites has generated a significant
amount of data and information. The following general items can be
reported:
1. Data suggest that no major adverse environmental effects have
occurred at any of the sites. For example, data from wells
downgradient of the disposal sites suggest that the contribution
of waste leachate to the ground water has resulted in concen-
trations of chemicals less than the primary drinking water
standards established by EPA.
2. The results from the sites are internally consistent. In
other words, the analyses of samples taken on different dates
at the same locations are very similar.
3. The total integrated evaluation of data from site development,
site water balances, physical testing of wastes samples, and
chemical sampling and analysis is providing a large significant
data base to explain many of the environmental effects that
can result from coal ash and FGD waste disposal.
A brief account of the results at each of the sites is presented below.
9-11
-------
TABLE 6. SUMMARY OF RESULTS OF EXTRACTION PROCEDURE (EP) TESTS
OF 20 FLY ASH AND 3 FGD WASTE GRAB SAMPLES
I
I—'
ro
Metal Overall Range Observed, \i g/1
Fly Ash FGD Waste
Arsenic <2 - 410 <2 - 65
Barium <100 - 700 <150 - 230
Cadmium <2 - 193 <2 - 20
Chromium <8 - 930b <11 - 26b
Lead <3 - <36 <5
Mercury <2 <2
Selenium <2 - 340 8-49
Silver <1 <1
Interim Primary Drinking
Water Standards3, jj g/1 Ratio of Range Observed to Standards
50
1000
10
50b
50
<1
10
50
Fly Ash FGD Waste
<0.04 - 8.2 <0.04 - 1.30
<0.1 - 0.7 <0.15 - 0.23
<0.2 - 19.3 <0.2 - 2
<0.16 - 18.6 <0.22 - 0.52
<0.06 - 0.72 0.1
<1 <1
<0.2 - 34 0.8 - 4.9
<0.02 <0.02
aReference 7 gives these standards ...for use in determining whether solid waste disposal activities comply
with ground water criteria." Standards included in Reference 7, but not measured in these tests, are for
fluoride: 1400-2400 pg/1 (depending on temperature), and for nitrate (as N): 10,000 yg/1.
"Reference 8 contains an amendment to the chromium criteria- for the EP, revising it from total chromium to
Cr(VI); since the total chromium values were measured by atomic absorption (AA), the measured ranges
represent upper limits for Cr(VI) in the samples.
-------
Allen - The results indicate the following:
1. Leachate generated within the ash ponds contain elevated (over
background) concentrations of several waste-related chemical
constituents (e.g., boron, sulfate, calcium, arsenic). However,
the surrounding soils have attenuated significant fractions of
leachate constituent contaminants within the immediate vicinity
of the ponds.
2. Leachate water from the upgradient portions of the ash ponds
has not moved sufficiently to create steady-state concentrations
of unattenuated constituents (e.g., sulfate) in the downgradient
wells. However, concentrations of these constituents are
expected to only reach or barely exceed secondary drinking
water standards (e.g., for sulfate, 250 ppm).
Elrama - The results indicate the following:
1. Prior to disposal of FGC wastes, much of the site was contaminated
by acid mine drainage, resulting in low pH (4.5 to 5) and high
concentrations of chemical constituents (e.g., approx. 2000 ppm
sulfate) in the ground water.
2. The landfill and runoff collection ponds may serve as potential
sources of some constituents via leachate and overflow, including
chloride and calcium (both at approx. 500 ppm). However,
neither chloride nor calcium is at a level to cause major
concern. In addition, an elevated level (approx. 0.2 ppm) of
arsenic was measured at one waste/soil interface lysimeter;
however, it does not appear to be a general problem. In any
event, substantial attenuation of arsenic by soils at the site
is expected.
3. The relative absence of elevated levels of these waste-related
constituents in downgradient ground water may be explained by
the relatively short time the fill has been in operation
(4 years), chemical/physical attenuation phenomena (including
the effects of the treatment/disposal process), or a combination
of these factors.
4. The landfill does not appear to alter significantly the local
concentrations of some constituents (such as sulfate) potentially
available from both mine drainage and FGC wastes.
Dave Johnston - The results indicate the following:
1. The water balance and estimates of plume arrival time indicate
that the widespread measurement at the site of what might
elsewhere be considered elevated chemical constituent levels
(e.g., sulfate, approx. 1000 ppm) is not due to the waste
landfills. The estimates of plume arrival time for the
9-13
-------
peripheral wells downgradient from (not directly under) the
active landfill are in excess of 100 to 300 years considering
only travel time in the saturated zone. Travel time from the
20-year old inactive landfill to a much closer (to the landfill)
downgradient well is estimated to be in excess of 20 years,
accounting for both unsaturated and saturated zone travel.
2. Most of the "elevated" measurements reflect pervasively high
background levels characteristic of highly mineralized ground
water in many western settings. However, lower measured
values (e.g., sulfate, approx. 100 ppm) at one background and
one peripheral well indicate that even in highly mineralized
arid areas there may be areas of good water quality that
require protection in waste disposal site planning and management.
Sherco - The results indicate the following:
1. Leachate movement from the ponds has thus far been sufficiently
retarded by the clay liner to preclude development of significant
elevations of chemical constituents in the leachate at down-
gradient wells.
2. There is a waste-related influence reflected in the slightly
elevated levels (of boron and sulfate) measured in the peri-
pheral/downgradient wells to the west and southwest, but it is
not clear whether this is due to past leakage from the sheet
piling/conduit area, to leachate that has moved through the
liner, or to a combination of these two sources.
3. Because of the pervious soils in the area of the site, significant
increases in concentrations of major soluble species are
expected to occur in downgradient wells in the next few years.
Secondary drinking water standards are expected to be exceeded
in these wells. However, any effects of movement of these
species off-site will be mitigated (diluted) by the Mississippi
River, which flows by the plant.
4. The higher concentrations of waste parameters in FGC pond
supernatant versus underlying waste interstitial waters may be
due to two factors: first, the conversion by the utility to a
system involving recycle of the FGC waste transport water
would have resulted in increased concentrations of chemicals
in the water; and second, the evaporation of water in the pond
would also increase remaining chemical concentrations.
9-14
-------
Powerton - The results indicate the following:
1. Although the completed landfill was supposed to have a 0.25 m
(8 in.) Poz-0-Tec liner, during the coring operation a general
absence of liner material was observed. This observation is
consistent with the practical difficulty of achieving uniform
placement of such a relatively thin layer of soil-like material
over a large area. Current engineering practice suggests that
a minimum thickness of 0.45 to 0.60 m (18 to 24 in.) of liner
placement would be desired to ensure full effectiveness.
2. The surface water analytical results for Lost Creek are consistent
with the water balance calculations. Both sets of results
indicate that the stream has adequate assimilative/dilution
capacity to lower current concentrations of chemical constituents
in leachate reaching Lost Creek to insignificant levels.
3. The results also suggest that the stream, if an effective
ground water flow divide, may limit the extent of further
downgradient ground water contamination by the waste plume.
4. The general lack of elevated trace metal concentrations in
ground water suggests that a combination of chemical attenuation
(especially for chromium and lead) and dilution is preventing
the release of significant quantities of these elements and/or
elevation to significant concentrations at downgradient locations.
5. Elevated concentrations of nitrate at various sampling locations
at the site can be attributed to local agricultural and urban
nonpoint source activities and not the coal ash landfill.
Smith - The results indicate the following:
1. There appears to have been a steady state achieved between the
concentrations of soluble species in the pond and in the
immediately adjacent downgradient areas.
2. There appears to be little nor no chemical attenuation of the
major tracer species such as calcium and strontium, but rather
a progressive reduction in concentrations in the downgradient
direction. This is consistent with what would be expected due
to admixing of leachate with the greater amounts of dilution
water.
3. The use of high total dissolved solids Bay water in the pond
for makeup and its presence in adjacent downgradient areas
create a situation where little incremental effect is detectable
from such typical ash pond "tracer" species as sulfates,
chlorides, and boron.
9-15
-------
GENERIC ENVIRONMENTAL EVALUATION OF COAL ASH AM) FGD WASTE DISPOSAL
The environmental effects of solid waste disposal practice is determined
by three factors: waste type, disposal method, and the environmental
setting. The data base from this project and other related projects
suggests that present and future practices of coal ash and FGD waste
disposal may be effectively evaluated through a matrix consisting of
four waste types, three disposal methods, and five environmental settings.
The four waste types are:
1. Fly ash or fly ash admixed with other materials. A significant
body of literature suggests that the majority of trace metals
available for leaching from utility solid wastes may be associated
with those containing fly ash. Thus this category of wastes
includes fly ash or fly ash mixed with bottom ash and fly
ash/bottom ash/FGD waste mixtures (excluding chemically treated
FGD wastes; see item 3, below).
2. Non fly ash materials. In this category are included bottom
ash (or boiler slag) and FGD wastes that are disposed of
separately from fly ash (including forced oxidation wastes).
This category usually contains lesser concentrations of trace
metals, but can result in higher concentrations of major
species (e.g., chlorides from FGD waste).
3. Stabilized FGD wastes. FGD wastes may be processed or stabilized
for full-scale disposal by a variety of processes; the processes
presently in commercial practice involve the addition of lime
and fly ash, or processed slag. Lime/fly ash stabilization
for landfill disposal is presently practiced at some power
plants and is expected to grow in importance. Processed FGD
wastes are a separate category because of the differences in
their physical and chemical properties created by the stabili-
zation process.
4. Dry FGD wastes. Several dry FGD systems are, expected to come
into commercial use over the next 3 years. It appears that
calcium-based dry FGD systems are anticipated to grow more
than the sodium-based systems. By either process, dry FGD
systems provide a combined waste containing fly ash and the
sulfur compound in a relatively dry form that is likely to be
sent for disposal to a managed landfill. The physical and
chemical properties of these wastes are expected to be different
from the other categories discussed above; additionally, there
is a relative lack of even limited field scale information on
their leaching characteristics to date.
9-16
-------
There are three disposal methods for coal ash and FGD wastes that are in
practice and expected to continue in the future: (1) pond disposal;
(2) interim ponding followed by landfill disposal; and (3) landfill
disposal (including disposal in mines, which is considered a special
case of landfilling).
Three of the five environmental settings for solid waste disposal are
based on major differences in climate and hydrogeology. These are:
(1) coastal areas, specifically those areas where surface water and
ground water are influenced by the ebb and flow of tides; (2) arid
areas, characteristic of much of the western U.S. where net evaporation
generally exceeds precipitation by a significant margin; and (3) interior
areas, characteristic of the non-coastal portions of the eastern U.S.
where there tends to be more of a balance between precipitation and
evaporation and where permanent surface water bodies are in such abundance
as to be near many disposal sites.
Further evaluations during this project suggested that a further breakdown
of two special categories would be useful because of their significant
characteristics. These are: (1) arid areas in the west where ground
waters and surface waters are very highly mineralized, and (2) interior
areas subject to acid mine drainage. Both of these last two types of
settings and the coastal setting tend to have water quality characteristics
that can potentially show less of an incremental effect from coal ash
and FGD waste leachates. This is because the waters in these areas
already contain a number of chemicals found in the leachate.
Table 7 is a matrix of waste types, methods of disposal, and environmental
settings and indicates combinations for which field-scale and other
information is available. Sources of data and information other than
this study included the Utilities Solid Waste Activities Group (USWAG),
the Electric Power Research Institute (EPRI), and the Department of
Energy (DOE)- DOE is currently sponsoring a study of disposal FGD
wastes in a surface mine; the study was originally sponsored by EPA. As
is clear, some information is available for most of the combinations
that are being practiced today or are likely to be practiced in the
future.
It appears that, on balance, technology exists for environmentally sound
disposal of coal ash and FGD wastes using any of the modes of disposal.
Potential environmental effects are highly site and system specific.
For some combinations of waste types, disposal methods, and environmental
settings, mitigative measures must be taken to avoid ground water and/or
surface water contamination. However, site specific application of good
engineering design and practice can mitigate most potentially adverse
environmental effects of waste disposal.
9-17
-------
I
M
oo
TABLE 7. SUMMARY OF INFORMATION AVAILABLE FOR COMBINATIONS
OF WASTE TYPES, DISPOSAL METHODS, AND ENVIRONMENTAL SETTINGS
Ponding Interim Ponding/Lsndf i 1 ling
Fly
Ash*
COASTAL SETTING X
Smithc
Non-Fly Processed Dry Fly
Ashb FCD FGD Asha
P NA NA X/Pd
Chistnan
Cr.
(USUAC)
Non-Fly Processed Dry Fly
Ashb FGD FGD Ashb
P P NA P
Landfilllng
Non-Fly Processed Dry
Ashb FCD FCD
P P Pe
ARID WESTERN SETTINC-
Not Highly Mineralized
ARID WESTERN SETTING -
Highly Mineralized
INTERIOR SETTING -
Not Highly Acidic
INTERIOR SETTING -
Highly Acidic
(nine drainage)
Notes: a. Includes
X P X NA X/Pd P
Allen, Bruce Ballly
Sherco, Mansfield (USWAC)
Michigan City
(USWAG) ,
Walllngf ord
(USWAG)
P P P NA P P
co-disposal of fly ash with other wastes.
Johns ton;
Hilton Young
(DOE/EPA)
P NA X
Powerton,
Zue 11 inger
(USWAC)
(USWAG)
Dunkirk (DOE)
P NA P
P X Pe
Coneevllle
(EPRI/USWAC)
P it Pd
Elrama
c. Plants for
appropria te poa1tions.
Either the interim pond or landfill ?•*-act of operation studied at field scale, but not both.
Laborabory da ta only.
e. Laborabory da ta only.
Key:
X - Data available from full-scale field studies.
P - Data aval lab le from la bora tory and/ or 11 mi ted- scale field a tudies for projec tloo purposes .
NA = Matrix combination not applicable due to lack of present and future practice.
-------
ENGINEERING/COST EVALUATIONS
The first major efforts in the engineering/cost evaluations involved
development of site-specific conceptual engineering designs and costs
(capital and first year operating and maintenance costs) for the current
solid waste handling and disposal operations at the six study sites. To
facilitate the ultimate use of the cost data, the estimates were developed
by breaking down the waste handling and disposal operations into five
modules: (1) raw material handling and storage; (2) waste processing
and handling; (3) waste storage; (4) waste transport; and (5) waste
placement and disposal (including site monitoring and reclamation).
Based on the site-specific cost estimates and other studies by TVA,
EPRI, and other organizations, generic capital and O&M cost estimates
were then prepared for individual modules comprising waste handling and
disposal for coal ash and FGD wastes. Tables 8 and 9 provide a summary
of the results of this effort.
The range of costs given represents variations in specific plant operations
as well as variations in the several cost estimates used in preparing
these estimates. For example, the higher end of the range for FGD waste
handling/processing might include thickening, vacuum filtration, and
mixing with lime and fly ash, while the lower end could represent a
simpler operation with little or no processing. Figures 2 and 3 show,
in graphical form, the estimates for the FGD waste handling/processing
"module." Similar figures for all the modules listed,in Tables 8 and 9
will be included in the final report for the project.
CONCLUDING REMARKS
Results from this 3-year study of disposal of coal ash and FGD wastes
from coal-fired electric generating plants should provide major technical
guidance for regulatory bodies and the utility industry. However,
results from field studies of this type are limited, and predictive
tools (e.g., computer models) for evaluating interactions between these
wastes and site-specific hydrogeologic systems are, in many cases,
inadequate. For this reason, additional efforts sponsored by the industry
are currently underway to develop more sophisticated tools for predicting
and analyzing the potential environmental effects of coal ash and FGD
waste disposal. These efforts will be described later in this session.
9-19
-------
TABLE 8. GENERIC CAPITAL COST ESTIMATES FOR FGC WASTE DISPOSAL
(Late 1982 Dollars)*
Capital Cost Range
($/kH)
Fly
Fly
Ply
Fly
Module
Submodule
ash handling/processing Wet handling
Wet handling
ash storage
ash transport
ash placement/disposal
Bottom ash handling/processing
Bottom ash transport
Bottom ash placement/disposal
Raw
PCD
"FCD
FGD
"Er
>>Rt
materials handling/storage
waste handling/processingc
waste placement/disposal
Dry handling
Dry
Wet sluicing
Dry trucking
Unlined pond
Landfill
Wet handling
Wet handling
Wet sluicing
Dry trucking
Unlined pond
Landfill
Dry (lime and
Wet handling
Dry trucking
Unlined pond
Landfill
igineering News Record (ENR) Index - 3931.11
^lationship between plant size a
• 365.97
250b
w/o recycle 2.3-4.3
w/recycle 3.7-6.8
2
4
3
0
15
4
w/o recycle 2
w/recycle 2
3
0
6
1
fly ash) 2
18
0
0
10
4
(1913 100)
(1967 100)
nd waste generation for typical
.2-4.1
.7-8.8
.5-6.4
.3-0.5
.1-27.8
.3-8.1
.2-4.1
.5-4.6
.0-5.6
.2-0.4
.4-11.8
.3-2.4
.4-4.5
.1-33.6
.7-1 .3
.4-0.7
.0-18.6
.1-7.6
case:
Plant
Size (MU)
500b
1
3
1
4
2
0
12
3
1
2
2
0
5
1
2
15
0
0
8
3
.9-3.5
.0-5.5
.8-3.3
.2-7.7
.7-5.1
.3-0.6
.9-23.9
.3-6.1
.7-3.2
.0-3.7
.4-4.5
.2-0.3
.1-9.6
.1-2.0
.1-3.9
.2-28.3
.5-1 .0
.3-0.6
.9-16.6
.3-6.2
1
2
1
3
2
0
11
2
1
1
1
0
4
0
1
12
0 .
1000b
.5-2.
.4-6.
.4-2.
.7-6.
.2-4.
.3-0.
9
4
7
8
0
5
.0-20.5
.5-4.7
.3-2.
.6-3.
.9-3.
.1-0.
.2-7.
.9-1.
.9-3.
.8-23
. 4-n.
0.3-0.
7
2
.9-14
.7-5.
5
0
6
2
7
6
4
.8
8
5
.7
0
2000b
1.3-2.3
1.9-3.6
1.2-2.2
3.2-5.9
1.7-3.2
0.2-0.5
9.4-17.5
1.9-3.6
1.0-1.9
1.3-2.4
1.5-2.8
0.1-0.2
3.4-6.2
0.7-1.3
1.6-3.0
10.8-20.0
0.4-0.7
0.3-0.5
7.0-13.1
2.2-4.0
Annual Waste Generation Rate
Fly Aah
Bottom Aah
FGD Waste
"Typical Case"
Coal Propertie
Load Factor:
(dry metric
Assumptions
s :
tons/HW of Plant Generati
280
70
240
21 S, 13Z Ash,
70Z
10,500
ng Capacity)
Btu/lb
(24.4 M
MJ/kg)
Heat Rate:
S02 Removal:
Lime Stoichiometry:
Fly Ash/Bottom Ash Ratio:
cAssumed FGD System: Wet Lime Scrubbing
10,250 Btu/kUhUO,8M MJ/kWh)
90Z
1.1
80/20
9-20
-------
TABLE 9. GENERIC ANNUAL COST ESTIMATES FOR FGC WASTE DISPOSAL
(Late 1982 Dollars)*
Annual Cost Range
($/dry metric ton)
Plant Size (Mw)
Module
Fly ash handling/processing
Fly ash storage
Fly ash transport
Fly ash placement/disposal
Bottom ash handling/processing
Bottom ash transport
Bottom ash placement/disposal
Raw materials hand ling /storage
FGD waste hand ling /pro cess ing
FCD waste placement /disposal
Subtaodule
Wet handling w/o recycle
Wet handling w/recycle
Dry handling
Dry
Wet sluicing
Dry trucking
Unlined pond
Landfill
Wet handling w/o recycle
Wet handling w/recycle
Wet sluicing
Dry trucking
unlined pond
Landfill
Dry (lime and fly ash)
Wet handling
Dry trucking
Unlined pond
Landfill
250b
2.5-4.6
3.7-6.8
2.5-4.7
3.3-6.1
4.2-7.6
1.7-3.1
11.5-21.3
7.0-13.0
11.3-21.0
12.3-22.8
9.2-17.1
3.4-6.3
9.2-17.1
5.4-10.0
4.1-7.6
17.2-31.9
1.1-2.1
2.9-5.4
8.5-15.8
4.0-7.5
500b
1.0-3.7
2.9-5.4
2.1-3.9
3.0-5.6
3.2-5.9
1.5-2.8
9.1-16.8
5.6-10.5
9.0-16.7
10.3-19.1
7.3-13.5
2.8-5.2
7.9-14.6-
4.7-8.8
3.7-6.7
13.8-25.5
0.9-1 .7
2.3-4.3
6.7-12.4
3.4-6.3
1000b
1.6-3.0
2.3-4.3
1.7-3.2
2.8-5.2
2.5-4.7
1.3-2.5
7.2-13.5
4.6-8.5
6.9-12.8
8.4-15.7
5.6-10.3
2.2-4;l
6.5-12.1
4.1-7.6
3.4-6.2
11.0-20.5
0.7-1.3
1.8-3.3
5.2-9.7
2.8-5.3
2000
1.3-2.3
1.8-3.6
1.5-2.7
2.5-4.7
2.0-3.7
1.2-2.2
5.7-10.5
3.7-6.9
5.3-9.9
6.9-12.8
4.3-7.9
1.8-3.3
5.4-10.0
3.5-6.5
3.0-5.6
8.8-16.4
0.6-1.1
1.4-2.6
4.1-7.6
2.4-4.4
afcngineering News Record (ENR) Index - 3931.11 (1913 100)
^Relationship between plane size
Fly Ash
Bottom Ash
FGD Waste
"Typical Case
• 365.97 (1967 100)
and waste generation for typici
Annual Waste Gener;
(dry metric tons/MU of Plant
280
70
240
Assumptions
jl case:
ation Rate
Generating
Capacity)
Coal Properties:
Load Factor:
Heat Rate:
S02 Removal:
Lime Stoichiometry:
Fly Ash/Bottom Ash Ratio:
cAssuraed FGD System: Wet Lime Scrubbing
21 S, 13Z Ash, 10,500 Btu/lb U4-4" MJ/kg)
70Z
10,250 BtuAWh (10.8M MJ/kWh)
90Z
1.1
80/20
9-21
-------
50.000
100 OOO
metric tons / year
200,000 3OO.OOO
400.00O
100,000 300,000 300,000 4OO.OOO
F 0 D WASTE GENERATION RATE ( ton* / y«ar )
Soviet- Arthur 0. Little,Inc. £iiimot«
500,OOO
FIGURE 2. FGD WASTE HANDLING AND PROCESSING!
ANNALIZED COSTS VERSUS FGD
WASTE GENERATION RATE
9-22
-------
eooo
60OO
CO
O
O
o
z
40OO
2000
z
z
100,000
metric tons / year
200,000 soopoo
400,000
Bads: 1982 Dollar*
I
I
O 100,000 200.OOO 3OO.OOO 4OO.OOO
F Q 0 WASTE GENERATION RATE ( tona / year )
' Source' Arthur 0. Little,Inc. Eilimotei
500,000
FIGORE 3. FGD WASTE HANDLING AND PROCESSING:
ANNALIZED COSTS VERSUS FGD
WASTE GENERATION RATE
9-23
-------
REFERENCES
1. Smith, M.P., et al. , "EPA Utility FGD Survey, July-September 1981,"
prepared by PEDCo Environmental, Inc., for EPA, Industrial Environ-
mental Research Laboratory, Research Triangle Park, NC, EPA-600/7-
81-012e (NTIS No. PB 82-23150), December 1981.
2. Santhanam, C.J., et al., "Waste and Water Management for Conventional
Coal Combustion: Assessment Report-1980." Prepared by Arthur D.
Little Inc., for EPA, Industrial Environmental Research Laboratory,
Research Triangle Park, NC, EPA-600/7-83-007 (NTIS No. PB 83-163154),
January 1983.
3. Santhanam, C.J., et al., "Waste and Water Management for Conventional
Coal Combustion: Assessment Report-1979," Vol. I-V, prepared by
Arthur D. Little, Inc., for EPA, Industrial Environmental Research
Laboratory, Research Triangle Park, NC, EPA-600/7-80-012a to e (NTIS
Nos. PB 80-158884, -185564, -222409, -184765, -185572), January
(Vol. I) and March (Vol. II-V).
4. Santhanam, C.J., et al., "Characterization and Environmental
Evaluation of Full-Scale Utility Waste Disposal Sites - Final
Report (draft), prepared by Arthur D. Little, Inc. under Contract
No. 68-02-3167 for EPA, Industrial Environmental Research Laboratory,
Research Triangle Park, NC, to be published.
5. Lunt, R.R., et al., "An Evaluation of the Disposal of Flue Gas
Desulfurization Wastes in Mines and the Ocean: Initial Assessment,"
prepared by Arthur D. Little, Inc., for EPA, Industrial Environmental
Research Laboratory, Research Triangle Park, NC, EPA-600/7-77-051
(NTIS No. PB 269270), May 1977.
6. Lunt, R.R., et al., "An Evaluation of the Disposal of FGD Wastes in
Coal Mines and at Sea: Refined Assessment," (draft) prepared by
Arthur D. Little, Inc., under Contract No. 68-03-2334, for EPA,
Industrial Environmental Research Laboratory, Research Triangle
Park, NC, to be published.
7. "Criteria for Classification of Solid Waste Disposal Facilities and
Practices: Final, Interim Final, and Proposed Regulations," Federal
Register, Vol. 44, No. 179, September 13, 1979.
y- "Hazardous Waste Management System; General and Identification and
Listing of Hazardous Waste," Federal Register, Vol. 45 No 212
October 30, 1980. '
9. Murarka, I.P., "Solid Waste Environmental Studies at the Electric
Power Research Institute," to be presented at the EPA/EPRI Symposium
on Flue Gas Desulfurization, New Orleans, November 1-4, 1983.
-------
OPERATIONS HISTORY OFrLOUISVILLE GAS & ELECTRIC
FGD SLUDGE STABILIZATION
R. P. Van Ness, J. H. Juzwiak,
W. Mclntyre
-------
OPERATIONS HISTORY OF LOUISVILLE GAS & ELECTRIC
FGD SLUDGE STABILIZATION
by: Robert Van Ness
Manager of Environmental Affairs
Louisville Gas and Electric Company
Louisville, KY
John H. Juzwiak, P.E.
Manager of Utility Operations
Conversion Systems, Inc.
Horsham, PA 19044
William Mclntyre
Senior Technical Advisor
Conversion Systems, Inc.
Louisville, KY
ABSTRACT
The Louisville Gas & Electric Company (LG&E) has been an industry leader
in the deployment of flue gas desulfurization technology from its very incep-
tion. LG&E was one of the first major utilities to install SO2 scrubbers;
and, at the present time, it operates seven scrubbers serving a combined gen-
erating capacity exceeding 2200 m.w. These scrubbers which have been purchased
from several different manufacturers, represent a broad spectrum of scrubber
technologies including dual alkali, lime and limestone systems.
The bleed stream from these scrubbers is dewatered, and the resulting
solids are chemically and physically stabilized in processing plants purchased
from Conversion Systems, Inc. (CSI), an early pioneer in the field. Unit #6
at Cane Run, a 270 m.w. unit with a dual alkali scrubber, is served by a
stabilization facility which began operations in April, 1980. Units #4/5 at
Cane Run have a combined capacity of 360 m.w. and are equipped with lime
scrubbers. The waste from these scrubbers is combined and treated in another
stabilization facility. The four units at Mill Creek have a combined capacity
exceeding 1600 m.w. and are also equipped with lime scrubbers. One large
stabilization facility was installed to handle the combined bleed from all
four Mill Creek scrubbers.
The stabilization plants have run for approximately three years and are
operated and maintained by LG&E personnel with advisory assistance supplied by
CSI. The knowledge gained from the experiences of LG&E and CSI in the last
three years of operation would be useful both to current operators of stabili-
zation facilities and to those who are anticipating the procurement of scrub-
bers and stabilization facilities.
This paper presents some of the operating and maintenance history which
has been obtained from these plants. Included will be discussions on reliabil-
ity of individual equipment and discussions of some modifications which were
9-25
-------
made to improve reliability. Operating problems will be addressed including
some of the inherent difficulties encountered in processing scrubber sludge
and fly ash.
In addition, a short discussion of the landfill operation which is an in-
tegral part of the stabilization process will be presented, as well as the
results of several years of landfill investigation. Discussions of the impact
of the landfill operation and plant control upon the environmental properties
of the final landfilled material will be offered.
BACKGROUND
Louisville Gas & Electric Company (LG&E) owns and operates two coal fired
generating stations in Jefferson County, Kentucky. These stations serve the
city of Louisville and its surrounding suburbs. During the early 1970's, when
sulfur dioxide emissions from coal burning power plants became a high environ-
mental priority, LG&E was required to install scrubbers on the newest of its
coal fired generating units. At that time there had been only a handful of
scrubbers installed on major utility power plants. LG&E became one of the
industry leaders with its early demonstration scrubber at their Paddy's Run
Plant. It has continued in the lead position by installing a total of seven
additional scrubbers at its Cane Run and Mill Creek Generating Stations.
During the early operating days of the Cane Run scrubbers, it became appar-
ent that some form of sludge stabilization was necessary to dispose of the
scrubber bleed and to prevent the creation of a surface or groundwater pollu-
tion problem. It was for this reason that LG&E in the late 70's contracted
with Conversion Systems, Inc. (then known as IU Conversion Systems) to design,
engineer and assist in the erection of three scrubber sludge dewatering and
stabilization facilities. Two are located at Cane Run; and, one at Mill Creek.
These three stabilization facilities are operated by manpower provided by LG&E
with the help of an advisory team provided by CSI. The disposal of the scrub-
ber waste is accomplished by dewatering and stabilization, utilizing the
Poz-0-Tec system developed by CSI together with controlled landfill disposal.
The LG&E/CSI operations team has acquired over three years experience in
operating these three stabilization facilities and has encountered many opera-
tion and design problems. Many of these problems will be encountered by other
firms as they begin operation of their own stabilization plants.
The first of these facilities was a relatively low budget blending unit
designed to handle waste generated by the Cane Run #6 unit. This unit is a
270 megawatt generator using coal with 12 to 16% ash and 4% sulfur content.
It is equipped with an electrostatic precipitator and one of the first dual
alkali scrubbers. The blending facility included fly ash and lime storage and
the feed systems necessary to blend fly ash and lime with dewatered sludge
from the scrubber. The interface on this plant was somewhat unusual when com-
pared to other facilities provided by Conversion Systems. LG&E had been de-
watering thickener underflow and disposing of this material in their bottom
ash ponds with no stabilization. When it was decided to install stabilization,
9-26
-------
a facility was designed which placed the blending equipment within the skirt
of the fly ash silo. This compact arrangement, although economical and expe-
dient, was later the cause of many operational and maintenance headaches.
(See Figure 1)
FIGURE 1 Exterior of Cane Run #6 showing Poz-O-Tec
being loaded for transport to landfill
UNDERFLOW HANDLING
In order to properly evaluate the experiences encountered in the operation
of the Cane Run facility, it is necessary to consider the system including not
only the blending unit, but also the vacuum filters and thickeners. Those
familiar with FGD thickener operation realize that it is difficult to maintain
a specific underflow density in normal operation. Usually the variance in
underflow density is alleviated by a surge tank between the thickener and the
vacuum filters in which the thickener underflow is blended to control the den-
sity and flow. In the Cane Run #6 facility the vacuum filters are fed directly
from the thickener underflow pumps. Since the thickener underflow will vary
from 5% to 30% solids, the resulting filtercake solids content is also highly
variable. Because the processing plant must handle upsets in the thickener
operation, the quantity of material is highly variable. Due to wide swings in
flow and solids content, the control system is sometimes unable to respond,
and the final blended product is not always consistent. At times the material
is very dry which causes stockpile and landfill dusting problems; and, at other
times, the material is wet and sloppy which causes difficulties in handling
and can turn the landfill into a quagmire. We have been able to mitigate the
9-27
-------
effects of this variability by careful stockpile and landfill management tech-
niques. These techniques include segregation and selective blending of stock-
piled material with different properties. These practices enable us to produce
a more consistent landfill material which minimizes handling and dusting prob-
lems, thus increasing landfill operating efficiency.
The absence of a surge tank has increased problems in the day-to-day man-
power allocation and management of the facility. The waste treatment plant is
a slave to the thickener operation. The plant must be ready to receive thick-
ener underflow whenever the torque on the thickener rake is high. The facility
also must sit idle or process at a minimal rate while a bed is developed in
the thickener.
The production rates of the processing facility have, at times, exceeded
the design rates due to high torque conditions in the thickener. In most cases
we have been able to handle these additional quantities of material, although
the resulting product has not always been entirely up to specification require-
ments, and the overload condition has put an excessive strain on the equipment.
We have been able to handle this off-spec material through the use of stockpile
and landfill management techniques. These techniques have enabled us to main-
tain environmentally safe landfill sites.
The Cane Run #4/5 and Mill Creek Processing Plants have been equipped with
underflow surge tanks which have produced a more consistent product and permits
greater flexibility for maintenance scheduling.
FLY ASH HANDLING
We have installed a total of four fly ash silos at the three LG&E facili-
ties. All of these silos are designed with mass flow bin bottoms and variable
speed tapered screw feeders. Although this design was initially more expensive
than the tradition air slide flat bottom fly ash silo installation, there are
almost no operating costs since it eliminates the need for compressed air for
air slide operation. We have also eliminated the high maintenance cost usually
encountered with star valve installations. This fly ash feeder and bin design,
however, is not without its operating problems. The bin bottoms are designed
with long, tapered sides which reduce the potential for bridging and rat-holing
by improving the flow characteristics of the fly ash. The result is a rela-
tively uniform delivery of fly ash to the tapered screw feeder.
The tapered screw removes the fly ash at a predictable and controllable
rate and feeds it into the process. During the start-up of the Cane Run #6
facility, we experienced several fly ash floods; that is, the contents of the
silo flowed through the feed system totally out of control. Subsequent experi-
ments and careful review of our operating logs indicated that this condition
was observed only at times when the fly ash level in the silo was very low or
when ash was being transferred into the system. Since the ash is transported
via pneumatic conveyor, we theorize that air was remaining entrained in the
ash and while in this state the ash was extremely flowable.
9-28
-------
Our solution to this problem was to modify our operating procedures and
our equipment. Our operating procedures were modified to require the slide
gate above the tapered screw feeder to be closed whenever the silo level is
below the intersection of the cone and the straight side of the silo, or when-
ever ash is transferred into the silo. In addition, we installed instrumenta-
tion which automatically closes the gate whenever a flood condition is detec-
ted. These measures proved successful in the Cane Run #6 plant; but because
of the size of the gates in the two later plants, it was decided to install a
smaller gate downstream of the feeder. This gate is designed to close whenever
a flood condition is detected.
FLY ASH DUST CONTROL
A problem encountered early in start-up was the operation of the mixer
dust collectors. These collectors are intended to control fly ash and lime
dust (and water vapor) which escapes (the equipment) to the atmosphere at the
mixer. Our original design proved inadequate due to plugging of the duct work.
The failure of these collectors led to high dust concentration in the air which
made it uncomfortable for operators and created problems for the equipment. A
revised design was installed on the Mill Creek plant which proved successful.
While earlier installations attempted to maintain a negative pressure in the
mixer in an effort to contain the dust within the mixer chamber, this new
design concentrated its efforts outside the mixer. Observation of the mixer
system revealed that the atmosphere of the mixer is extremely turbulent with
the turbulent action causing dust particles to be suspended. When this atmos-
phere was pulled into the dust collector, the collector soon became overloaded
and clogged. We soon learned that the dust particles tend to settle, if per-
mitted, in a still atmosphere. This was accomplished with a revised dust
collection hood which was designed with a long stilling chamber immediately
after the mixer. The dust collector suction was moved downstream of the
stilling chamber. This design provided sufficient time for dust particles to
settle and become captured by the slightly moist product. Our success with
this system was extremely gratifying, and efforts are now underway to install
the same system at the other Louisville plants.
MAINTENANCE
The fly ash laden atmosphere combined with the corrosive properties of
lime and the water in the filtercake all combine to make a maintenance head-
ache. Those who have never operated a sludge stabilization facility should
not be lulled into believing that these plants do not require operators, or
that maintenance will be minimal. Improvements in the fly ash dust collection
system and other equipment modifications have contributed greatly to improving
the service life of our equipment. However, additional maintenance routines
initiated by LG&E at these plants have been an even greater factor in the ser-
vice life of this equipment. A good deal of the maintenance problems encoun-
tered in a sludge stabilization facility are due to ineffective or inadequate
cleanup efforts. The theory of operation of the Poz-O-Tec process is that the
fly ash and lime chemically combine to form a matrix which binds the FGD
scrubber sludge into a cement-like material. This material will build up and
cure into a hard, abrasive mass which will destroy bearings, gears and other
9-29
-------
mechanical parts. This was evident in the early days at Louisville when mater-
ial had been allowed to cure in the mixer. The mixer paddles began to hammer
the cured material. The resultant stress was transmitted along the shaft and
eventually contributed to the premature failure of several gear teeth in the
mixer speed reducer. LG&E has instituted a program which keeps the material
build-up under check. Every day the plant is shut down for complete cleaning.
This includes cleaning of the mixers and hosing of conveyor parts. (See
Figure 2) This cleaning usually takes several hours but has proven worthwhile
in extending the life of the equipment. To speed cleanup at the Cane Run #6
processing plant, the mixer has been coated with a high density polyurethane
coating which has proved to be very useful in preventing material build-up.
It permits the mixers to be cleaned with high pressure water. This coating
has, however, become scored by the fly ash and its benefit will diminish. A
reapplication will hopefully restore the surface. If this coating is success-
ful, the other processing facilities will have this coating installed.
FIGURE 2 Interior of Mill Creek During Clean-Up
The lubrication program developed for the Louisville plants has recently
waf^/H 1° takVnt° account some of the lessons we have learned. As
was stated earlier the facilities are washed down at least once a day with
water at about 50-75 psig. Over a period of time daily washdown will flush
also conll , bea^^s' ™e "Y ash and lime present in the atmosphere may
also contaminate grease and oil. These contaminated lubricants if left in
Place act as abrasives, reducing bearing and gear life. To eliminate the e
?he applcationVof90ne "° * l»bricati™ Sched"le "hich reduces the time between
the application of grease and increases the frequency of gearbox oil changes.
9-30
-------
WINTER OPERATION
Those stabilization plants operating in cold climates are faced with the
additional hardship of winter operation. Cold weather hampers processing plant
operation by reducing the filterability of the thickener underflow. In addi-
tion, the problem of belt and idler freezing reduces conveyor belt and idler
life. The clean-up of outdoor conveyors is also a problem since the use of
high pressure water hoses increases the possibility of ice-up.
Cold weather further inhibits the fly ash/lime reaction. This particular
problem is resolved by stockpile management techniques which utilize the pro-
duct's insulating properties. With proper techniques, even if the outside
temperature is below freezing, temperatures approaching 100°F may be maintained
within the pile. Periodic monitoring of these temperatures, however, is neces-
sary in order to insure sufficient reaction before landfilling the product and
to guard against over-reaction which would result in the creation of boulders.
SOLUTIONS
In the design and construction of the processing plant that services the
Cane Run Unit #4/5, several modifications were made and improvements realized
over the Cane Run #6 design. This unit is about 3/4 mile from the thickeners
and approximately 1 mile from the fly ash silos. It was decided to include
100% redundancy for the filtration system, and a single surge tank was instal-
led. No redundant conveyor systems or redundant mixers were installed. Oper-
ation of the surge tank has enabled LG&E to schedule operations so that most
of the maintenance is performed during the daylight shift. It has also
smoothed the flow variations which have been seen in the Cane Run #6 unit.
Underflow feed from the surge tank varies approximately 2 to 3% over the dura-
tion of a pumping shift. We are able to process at a more consistent rate and
generate product with a more consistent solids content.
The processing plant servicing the Mill Creek Units #1, 2, 3 & 4 was
designed with almost 100% redundancy. The only common element at this plant
is the large surge tank. Redundant fly ash and lime systems are included in
this facility as well as separate conveyors and mixer trains. Separate mixers
allowed the plant to schedule routine cleaning and maintenance. This insures
efficient cleaning of each mixer since the cleaning crews are able to take
their time and do a complete job.
The Mill Creek facility is also equipped with an on-site laboratory which
is manned during the daylight shift. This provides immediate feedback to the
processing plant operator so that variances in design mix are discovered
quickly and are corrected. Because of this LG&E is assured of efficient util-
ization of lime and material with the proper in-place qualities.
OTHER RECOMMENDATIONS
The utility just beginning to operate its first FGD sludge processing
facility should realize that many elements will effect operating efficiency.
At LG&E several types of filtercloth were utilized based upon filter-leaf tests
performed by CSI's on-site advisors. The optimum cloth was chosen based upon
9-31
-------
filtercake production rates and properties. At Mill Creek we have utilized
several types of filtercloth on different filters. Each type is used for dif-
ferent properties of the thickener underflow. We have done this because we
understand that there are variabilities in the coal burned and the operation
of the scrubber which must be taken care of by the sludge stabilization
facility. Since all material must be processed by the sludge stabilization
facility, this facility must be adaptable to the variances which will be
encountered in the operation of the power plant, and it must be ready to
accept upsets with- out causing plant downtime. Therefore, we have tried to
adapt the plants to accept these variances. There are, of course, some
variances which go beyond the adaptability of the processing facility. We
have been able to handle these through the use of stockpiling and landfilling
techniques, most notably microencapsulation, which we have successfully
employed at the Louisville facilities.
LANDFILL
The above topics discussed the processing facilities required to process
the FGD scrubber sludge into a stabilized material. However, to complete pro-
perly the Poz-0-Tec system, the processed material must be conditioned and
placed within a permitted landfill. LG&E also contracted with Conversion Sys-
tems to provide a Waste Management Plan for the disposal of the material
produced by their processing facilities. This plan included all necessary
information and procedures for LG&E to obtain landfill permits from the State
of Kentucky for both of their generating stations. The plan included proposed
material characteristics (unconfined comprehensive strengths, permeability,
leachate), landfill and process plant quality control programs, and landfill
site development plans. With this information, LG&E was able to obain land-
fill permits at both power stations.
At this time, landfills are being developed at both sites in accordance
with the Waste Management Plan with monthly inspections by the State of Ken-
tucky. The permitted sites include borrow areas (areas where soil was removed
to construct the station), flood plain areas of the Ohio River and other areas
not suitable for the construction of power plant structure. One of the areas
not suitable for utility construction is currently being reclaimed by landfil-
ling with Poz-O-Tec at the Mill Creek generating station. Plans have been
made to construct a limestone crushing plant upon completion of this landfill
site.
GROUND AND SURFACE WATER MONITORING
Included in the Waste Management Plan were programs for ground and surface
water monitoring. The groundwater monitoring is accomplished by quarterly
sampling of wells located near the landfill sites. These locations were estab-
lished by the groundwater flow which flows toward the Ohio River. Efforts
were made to locate one monitoring well upstream of each landfill site and two
or three wells downstream of each landfill site in order to determine the
effect of the landfill upon groundwater quality. Initially, the wells were
sampled to obtain a composite standard against which subsequent water quality
data could be measured. Once the landfill operation started at each site, the
up-gradient wells have been used to monitor any changes in the quality of
9-32
-------
groundwater passing beneath the disposal area. Before each sampling, the
static water level in each monitoring well has been recorded. This data has
been used to determine the annual fluctuations in the groundwater elevation
for the disposal area, since the placement of processed material within the
disposal area should not affect the existing dominant direction of the ground-
water flow. Up to this time the water quality analysis have indicated no
change in the overall groundwater quality and dominant direction of the
groundwater flow.
The obtaining of the monitoring well samples has been simplified as each
well has had a submersible pump installed. This pump is energized by a port-
able generator which enables the laboratory technicians to sample each well in
less than one hour. Originally, the wells were sampled with a siphon pump
located on the back of a pick-up truck. Priming of each well posed problems
due to the wells isolated location which increased the time required to sample
each well. After obtaining the samples, they are preserved and sent to the
CSI's Technical Center for analysis in accordance with the permit guidelines.
All landfill run-off caused by precipitation is controlled and routed
through a pond constructed near each landfill site. The ponds were designed
to collect the run-off and discharge through a sandfilter. Sizing of the ponds
and associated draining channels were in accordance with the Louisville area
one hundred year rainfall criteria. Development of the landfill incorporated
drainage designs to control all run-off during construction. Efforts have
been made daily to control sediment due to errosion as an excessive amount of
sediment in the run-off increases the maintenance of the pond and sandfilter.
Up to this point, the sandfilter has had to be cleaned approximately twice a
year. Run-off samples have been obtained at weirs installed at each sandfilter
discharge, after each rainfall.
QUALITY CONTROL
In order to continue to meet the parameters of the landfill permits, the
Waste Management Plan also provides a quality control program for each proces-
sing facility and landfill site. These programs are maintained by LG&E's lab-
oratory personnel with CSI's Advisory Staff personnel providing any technical
advice needed to maintain these programs.
Daily process plant quality control activities include sampling of all the
raw materials sent to the process plant and determining the proper addition
rates for the process plant operators to utilize. The operators use these
recommendations to make minor set-point changes on automatic controllers. At
times, major set-point changes are required when one of the raw materials is
received out of specification. This usually occurs when the filtercake solids
are low due to low solids received from the thickeners. Other daily routines
include sampling of the stockpiled material to determine its curing rate.
This analysis also provides information on back-shift production since the
laboratory personnel are only assigned to the process plant during the daylight
shift. This minimum coverage of quality control is satisfactory as long as
the process plant personnel operators operate the plant in the automatic mode.
9-33
-------
Whenever non-specification material is produced due to operator error, equip-
ment breakdowns, or raw material availability, the quality control personnel
recommend set-point changes to produce an above specification material to blend
with the below specification material. Cooperation between both the operators
and quality control technicians has resolved most deviations smoothly. (See
Figure 3)
FIGURE 3 Exterior of Cane Run #4/5 Showing Stockpile Management Practices
During the start-up of each facility, ratio tests were performed to deter-
mine minimum addition rates of fly ash and lime. It was found that for these
particular installations, the fly ash addition rate could be reduced to 60% of
the design with a small increase in the addition of lime and still meet the
minimum material characteristics required in the landfill.
The landfill quality control program is also maintained by LG&E personnel
with assistance from the CSI Advisory Staff personnel. The same technicians
that perform the process plant analysis also maintain the landfill program
since the frequency of testing is dependent upon the amount of processed mat-
erial produced. In-place landfill density tests are performed for every 10,000
to 12,000 tons of material produced. These tests are used to control the
landfill placement technique and to insure that the material is placed and
compacted to a minimum density specification. When this minimum density
specification has been achieved, it has been found the landfill strength per-
meability and leachate criteria are met. These three criteria are continually
checked as landfill test cylinders are prepared routinely for testing at the
CSI's Technical Center. These test cylinders are molded to the landfill
in-place density test results by the LG&E technicians and, after curing, are
sent to the Technical Center for analysis. Results from these landfill test
9-34
-------
cylinders have shown that as long as the processed material of the proper mix
is properly conditioned, placed and compacted, all permit guidelines concerning
the materials physical characteristics have been met.
LANDFILL OPERATIONS
LG&E contracted a local heavy-equipment contractor to prepare the landfill
sites and place the processed material in accordance with the Waste Management
Plan. Since the landfills are located on the utility property next to the
generating station, the contractor has been able to utilize offthe-road equip-
ment. The size of his fleet of hauling equipment has grown as each of the
three processing facilities started up. At this time, he is using both 35
yard and 50 yard capacity trucks. (See Figure 4) The larger trucks have posed
a problem as the roads bearing this traffic must be well-constructed to support
their gross weight. In the landfill itself, additional effort must be made in
the site development to eliminate truck traffic on recently placed material as
it will not initially support this weight. Since curing rates are slowed
during winter operations, the active landfill must be increased accordingly.
FIGURE 4 Exterior of Mill Creek showing Poz-O-Tec
being loaded for transport to Landfill
-------
Low-ground pressure bulldozers place and compact the material with a self-
propelled vibratory roller providing additional compaction and surface sealing.
Maintenance repairs on the landfill equipment is actually less than the con-
tractor's normal equipment repairs on earth projects. This can be attributed
to the fact that although the processed material has abrasive additives, it is
less abrasive than the sandy soil in which his equipment usually operates.
The contractor operates five days a week, approximately eight hours a day.
Depending on which site he is actively working, his hauling rate varies between
200 to 500 tons per hour.
MATERIAL CHARACTERISTICS
The LG&E material characteristics are based upon the type of coal used,
type of scrubber and agent, fly ash collector size and collection efficiency
and available lime supply. The design specifications of process plant mixes
and ultimate landfill characteristics were achieved without any major changes
from the original Conversion Systems, Inc. equipment design. We did, however,
lower the filtercake solids content and production range, as it was found dur-
ing start-up at all three plants that the filtercake could be produced too
dry. Vacuum filter cloth changes corrected this problem.
Annual core borings of the landfills have been performed to insure that
the ongoing quality control results are accurate, and that what is actually
placed is meeting all the minimum criteria. All results have shown that when
the processed material is properly conditioned, placed and compacted, the per-
mit guidelines are satisfied by landfill disposal using the Poz-O-Tec system.
Typical data from these borings are shown in Table #1.
TABLE I
TYPICAL 28 DAY MATERIAL PROPERTIES
Mill Creek Cane Run
Unconfined Compressive Strength 75 psi 48 psi
Permeability 2.5 x 10~6 1.01 x 10~6
9-36
-------
COAL WASTE UTILIZATION IN ARTIFICIAL REEF CONSTRUCTION
J. H. Parker, P. M. J. Woodhead, D. M. Golden
-------
COAL HASTE UTILIZATION IN ARTIFICIAL REEF CONSTRUCTION
by: J. H. Parker and P. M. J. Woodhead
Marine Sciences Research Center
State University of New York
Stony Brook, New York 11794
and
D. M. Golden
Electric Power Research Institute
P.O. Box 10412
3412 Hi 11 view Avenue
Palo Alto, California 94303
ABSTRACT
The technology of coal combustion has improved greatly over the last few
decades. One result of the elaborate emission control equipment required on
coal plants to meet stringent air quality requirements is the large volumes
of wastes that must be utilized or disposed of safely. Flue-gas
desulfurization (FGD) sludge and fly ash may be produced at the rate of as
high as 1,000 tons per day at a typical coal fired power plant. Waste
disposal, especially in urban coastal areas, has become a major obstacle to
conversion to coal combustion for generating electricity.
To assess a possible solution to the waste disposal problem, 500 tons of
FGD sludge and fly ash were stabilized into blocks and placed in the ocean as
an artificial reef. Previous laboratory investigations had indicated that no
toxic chemical or physical effects should occur in the marine environment.
After three years in the sea, the coal waste blocks support a diverse
community of reef fish and invertebrates and have maintained their structural
integrity. No adverse environmental effects have been detected.
Assuming that the coal waste blocks continue to be environmentally
acceptable in the marine environment, the engineering and economic
feasibility of this method of disposal should now be confirmed.
INTRODUCTION
The increase in the combustion of coal for generating electricity has
resulted in the improvement in the technology required to burn coal cleanly.
Electrostatic precipitators are able to remove a? much as 99% of the
particulate ash and flue gas desulfurization (FGD) scrubbers reduce the
sulfur oxide emissions to acceptable levels. A problem, though, is created
-- waste disposal. With a large coal-burning power plant producing more than
1000 tons of fly ash and FGD sludge each day, coastal metropolitan areas and
areas with shallow ground water tables may lack sufficient land to safely
9-37
-------
dispose of these wastes. Ocean disposal could become an option if the
materials can be stabilized to prevent any harmful effects to the marine
envi ronment.
In September, 1980, 500 tons of stabilized FGD sludge and fly ash, in
the form of 15,000 hardened blocks (20 x 20 x 40 cm), were placed in the
Atlantic Ocean as artificial reef substrate. Two years of prior laboratory
and small-scale field investigations sought to detect any possible adverse
effects to the marine community as well as any physical deterioration of the
blocks themselves. With the manufacture of the 15,000 blocks and their
placement in the sea, a three-year monitoring program continued to
investigate the environmental acceptability of constructing artificial reefs
with coal wastes.
Several types of biological surveys were conducted to compare the
success of the"coal waste reef with that of an existing reef of rock and
building rubble. The fish community was sampled with traps to estimate
population densities. Test organisms were placed at the reef to detect
possible uptake of block components. The structural integrity of the blocks
was monitored using ultrasonic and compressive testing. To understand any
chemical changes occurring and to allow estimation of the lifetime of the
blocks in the sea, the chemical composition of the blocks was determined
throughout the study.
As these investigations continue to provide positive results, the
engineering and, ultimately, the economic feasibility of disposing of coal
wastes in the ocean as artificial reefs is now being determined.
BLOCK PRODUCTION AND REEF CONSTRUCTION
The 500 tons of FGD sludge and fly ash were obtained from two power
plants, the Conesville station of Columbus and Southern Ohio Electric (CSOE)
company and the Petersburg station of the Indiana Power and Light Company
(IPALCO). The Conesville material was produced with a ratio of fly ash to
FGD sludge of approximately 3:1 while the IPALCO material had a fly
ash:sludge ratio of about 1.5:1. The chemical compositions of the components
of each material as well as the stabilized blocks were determined and
reported in Parker _et _al_. , (1981).
The actual block production utilized standard concrete block industry
equipment at a commercial plant in Pennsylvania (Figure 1). Very few
modifications had to be made to accommodate the mixture of fly ash, sludge,
lime and cement which was somewhat wetter then conventional concrete
materials. The 15,000 blocks were all cured for 24 hours in a 'wet steam1
kiln at about 65°C and finally trucked to a dock in New Jersey. A bottom
opening barge transported the blocks to the project site (Figure 2) and
released them to settle on the bottom in 20 m depth of water.
BIOLOGICAL INVESTIGATIONS
To determine the success of the coal waste blocks as an artificial reef
substrate, it was necessary to measure colonization of the block surfaces by
9-38
-------
F6D
SLUDGE
Figure 1. Schematic of block factory.
Figure 2. C-WARP project site.
epibenthic fauna and to determine the population of fish taking up residence
in the "nooks and crannies" provided by the reef. Most importantly, analyses
were made on reef organisms to detect possible uptake of toxic components
from the blocks into the food chain. Special test bricks of coal waste were
set out and returned to the laboratory periodically for identification of
species and measurement of coverage (Figure 3 and Table 1). Close-up
underwater photographs were also taken throughout the entire study period of
the same areas on tagged blocks to document changes in the epibenthic
community.
9-39
-------
lOOr
TOPS
SONDJFMAMJJASONDJ FMAMJJ ASOND
.,981 1 1982
-1980-
o CONESVILLE, D CONCRETE, A IPALCO
Figure 3. Coverage by epifaunal colonizers.
TABLE 1. EPIBENTHIC SPECIES GROWING ON SURFACES ON REEF BLOCKS, 1980-81,
Porifera
Sponqe, unident.
Colenterata
Metridium senile
Clytia sp.
Obel ia dichotoma
Sertularia cupressina
Astranqia danae
Bryozoa
Aetea .sp.
Bugula turrita**
Callopora aurita
Cribulina jui n c t a t a
Electra hastingsae
Electra jrijosa
Schizoporel 1 a unicornis**
Microporella sp.
Mol 1 usca.
Anomia aculeata
Anomia simplex
Mytilus edulis**
Zirphaea crispata
c.pisula sp.
Onchidoris sp.
Acanthodoris pilosa
Eubranchus exiquus
Annel ida
Harmothoe extenuata,
Phyllodoce arenae
Pol vdora social i s *";"
Syllis sp.
Nereis grayi
Sabellaria vulqaris**
Asabellides oculata
Arthropoda
Balanus crenatus**
Caprella linearis
Pokoqeneia inermis
Edotea triloba
Unciola irrorata
Cancer irroratus
harpacticoid copepods
Echinodermata
Asterias rubens
Tunicata
tunicate unident.
9-40
-------
Fish traps were used to capture, tag, and
coal waste reef and a long established fishing
rubble (Fire Island Reef) nearby. Because the
was similar to that of the part of Fire Island
comparison of fish populations was conducted.
coal waste reef increased from approximately 2
reef had been established for one year, to 8.5
1981 and 1982 the cunner population densities
release fish from both the
reef of rock and building
area of the coal waste reef
Reef being studied, a
The density of cunner on the
6 fish per m2 in 1981 when the
fish per m2 in 1982. During
at the Fire Island Reef
remained steady at approximately 7.5 fish per m2. These results indicate
that, in less than 2 years the coal waste reef had achieved "carrying
capacities" for fish comparable to that of the existing artificial fishing
reef.
Initial toxicity experiments in the laboratory tested sensitive marine
organisms, a diatom and winter flounder eggs and larvae (Parker et al.,
1981), for adverse effects due to exposure to various concentrations of an
elutriate made from powdered coal waste block material. The lack of
significant effects was further investigated by field tests at the reef.
Mussels, Mytil us edulis, widely used indicator organisms for toxic uptake
studies, were placed at both the coal waste reef and at a separate control
site. Again, analysis of mussel tissue showed no significant increases in
potentially toxic trace elements after 9 months exposure on the reef. A
laboratory experiment using mussels exposed to elevated concentrations of
suspended coal waste material yielded similar results (Table 2). A spurious
increase in copper was not seen at higher concentrations of coal waste. A
doubling of iron concentrations in the mussels would not be expected to
create any toxic effects. Fish were tagged during the three-year study
period and recapture data indicate that these fish remained resident on the
same reef throughout the complete study. Fish samples from the reef have
been taken for tissue analysis of trace elements and should represent the
most realistic case to detect possible uptake of block components.
TABLE 2. TRACE METAL CONTENT OF SOFT TISSUES FROM MUSSELS EXPOSED TO
SUSPENSIONS OF POWDERED CONESVILLE COAL WASTE
CONCENTRATION OF
COAL WASTE
(mg/A)
Cd
TRACE ELEMENTS (ppm, dry weight)
Cu Fe Mn Ni Pb
Zn
0
(control)
89
178
267
1.28
1.42
1.42
1.24
8.47
8.98
13.45
7.87
56.4
102.4
101.5
106.4
6.48
4.43
4.03
5.60
1.46
2.29
1.70
5.32
2.31
2.58
2.93
2.76
107.1
101.2
116.5
97.4
Values are averages of values for 5 replicates.
Underlined values are significantly different than controls at p <_ 0.5 level.
9-41
-------
PHYSICAL AND CHEMICAL STABILITY
Although the blocks' surfaces successfully support a wide diversity of
epibenthic colonizers, it is important to determine how quickly the blocks
might deteriorate in the sea. Previous laboratory experiments indicated a
gradual increase in compressive strength and density over time. After the
placement of the reef, blocks were tagged for repeated ultrasonic
(non-destructive) testing in the sea. Results have provided a good
correlation between the series of ultrasonic data and compressive strength
measurements when the blocks were finally returned to the laboratory after
extended exposure at sea. Both sets of results indicate a gradual increase
in strength seen in Figure 4 (Parker _et aj_., 1982).
10
1800
1500
1200
900
600
300
100
200
SOAK DAYS
300
400
Figure 4. Compressive strength of Conesville (A) and
IPALCO (0) blocks in the ocean.
The coal waste blocks undergo cementitious processes similar to those
occurring in concrete. These reactions normally continue long after initial
formation and it is important to understand what changes might occur due to
protracted exposure to seawater at 3 atmospheres of pressure. Blocks
returned from the sea were analyzed for any changes in mineralogy. As
exposure time in the sea increased, mineralogical changes were found to occur
in the surface layer of the blocks. In Figure 5, mineralogical results
indicate that calcium sulfite hemihydrate and gypsum are major and minor
components, respectively, in the core of the blocks. In the surface (1-2 cm)
layer, however, a conversion has occurred making gypsum a major component.
Although this reaction results in an overall positive volume change, the
inherent porosity of the material accommodates any internal pressures that
would otherwise develop.
Little, if any, change has occurred in the composition of minor and
trace elements. However, significant surface losses of calcium by
dissolution have been accompanied by an enrichment in magnesium. Laboratory
9-42
-------
CORE
SURFACE
HEMIHYDRATE
CaS03 --j-HjC
GYPSUM
CaS04 -2H20
HEMIHYDRATE
GYPSUM
m
m
m
CONESVILLE
m M t M M
t -
m
M
mm M
m
M
14 32 38 49 61
WEEKS OF SUBMERSION
80
CORE
HEMIHYDRATE
CaS03 --^-HjO
GYPSUM
CaS04 -2H20
M
m
SURFACE
HEMIHYDRATE M
IPALCO
M Mm
t t t
M Mm
M
GYPSUM
M
m
t m M M
m
i—
0
14 32 38 49 61
WEEKS OF SUBMERSION
80
Figure 5. Mineralogical changes occurring in the blocks.
experiments have been conducted to measure the dissolution rate of calcium
and, using these results in a model, to estimate the overall lifetime of the
blocks. Preliminary results indicate that, after 30 years, the diffusion
zone would only penetrate 2-3 cm into the blocks.
ENGINEERING ASPECTS
After preliminary tests making blocks with machines at research
facilities of the Besser Company, Alpena, Michigan, the 500 tons of coal
wastes were processed at a commercial concrete block plant utilizing standard
"off-the-shelf" equipment. To move a step closer to the full-time disposal
of coal wastes from a typical power plant using stabilized waste block
production techniques for reef construction, the capability to handle as much
as 1000 tons of fly ash and FGD sludge per day must be demonstrated. Figure 6
illustrates a simplified system for the ocean disposal of coal waste blocks.
The process flow is indicated in Figure 7. Calculations were made for a
hypothetical power station equipped with two 500 MW units, and a 45% load
factor over a 30 year life. The details were taken from the EPRI FGD Sludge
Disposal Manual, 2nd Ed. (1980). The fuel was an eastern coal with 3.5%
9-43
-------
TEMPORARY
SLUDGE FILTERCAKE
HOLDING BASIN
POWER PLANT
HrORATED LIME BIN
BLOCK PRODUCTION
STEAM KILNS
BLOCK PLANT
|TUG| BARGE] BARGE LOADING FACILITY
OCEAN
DISPOSAL SITE
Figure 6. Block production design for ocean disposal.
sulfur and 12% ash content. Table 3 presents the estimations of the weights
of coal waste requiring disposal daily.
For estimation of the capacity of machines to make blocks from sludge
and fly ash, it has been assumed that 4% lime and 3% cement would be added to
stabilize the waste mix and that the material would be consolidated to at
least 1.76 g/cm3 (110 lb/ft3) as delivered from a block machine. The block
size was that of a standard construction block, nominally 8 x 8 x 16 in
(20 x 20 x 41 cm), but with an actual volume of 0.017 m3 (0.59 ft3) and
weighing 29.4 kg (65.2 Ib). Larger block molds are also readily available
for use in conventional block machines.
The calculations of materials processing capacity (Table 4) indicate
that a single V6-12, or two V3-12 block machines, operating at a rate of
eight block-forming cycles per minute and with an efficiency of 90%, have the
potential to fabricate the entire daily waste production of the power plant
into blocks for placement in the ocean.
9-44
-------
PROCESS SCHEME
V6 SYSTEM
FLY ASH
SLUDGE •
BOTTOM
to BAT
CEMENT
"uiMr:_-. u r\
^^ LJM 1 ^ i i i i i« w^^ 1 Ip W
1 LIME
BLENDING
MOL
1
DING ^<
i
ACCUMULATION
x *- LOA
CONV
i
DING
EYING
, PALLETS
RACKS CURING '
/
CONVEYING
1
^ UNLOADING
HFDAI
i I
1
ITTIMr V
1
CUBING
YAR
DELI\
1
1
DING
/ERY
USE
Figure 7. Process flow for coal waste
block production.
Although the location and layout of an existing power plant would
control how the block facility was constructed, a design that minimizes
handling of the blocks in transfer to the ocean-going barge would greatly
reduce costs.
REEF PLACEMENT ASPECTS
For a power plant producing 1000 tons of coal wastes per day, as many as
31,000 blocks could require placement in the ocean. Suitable sites must be
chosen for the placement of the blocks if the benefits of the artificial
reefs are to be fully realized. Economics factors will dictate the maximum
distance from the loading facility but bottom type and accessibility to
recreational fishermen will also be considered when regulatory agencies are
asked to approve the designation of disposal/construction sites. The hard,
sandy bottom and number of inlets off Long Island, New York make this area
very suitable for artificial reef construction. With several hundred square
miles of ocean bottom between the depths of 40 and 90 feet, sites can be
chosen which would minimize interference with commercial fishing operations,
9-45
-------
Table 3 ESTIMATION OF AVERAGE WASTE QUANTITIES FOR 3.5% SULFUR EASTERN COAL
(FROM EPRI REPORT CS-2009 VOL. 2, NOV. 1981).
ASSUMPTIONS (for a two 500 MW unit generating station)
COAL
Btu/lb
Ash content, %
Fly ash/bottom ash
Sulfur content, %
Annual consumption @ 45% capacity factor, metric ton
AIR QUALITY CONTROL SYSTEM
Upstream fly ash removal efficiency, %
Overall fly ash removal efficiency, %
S0~ scrubbing reagent
S0? removal, %
Stoichiometry (CaO/SO? removed), %
CaS04.2H20/CaS03.±H2(T
Thickener underflow solids, %
Dewatered sludge solids, %
ASH PRODUCTION DAILY
Total ash produced
Total fly ash produced
Fly ash collected in precipitator
SLUDGE PRODUCTION DAILY
Fly ash collected in scrubber
CaSO,.2HnO
Weignt of excess and unreacted lime reagent
Weight of dry sludge solids
Weight of water, at 60% solids content
Total weight of wet sludge, at 60% solids
STABILIZATION ADDITIVES
Lime, 4%
Cement, 3%
WASTES TO BE PROCESSED DAILY
Fly ash + sludge solids + lime + cement
Total dry weight
Water in sludge at 60% solids
Total wet weight
12,500
12
80/20
3.5
1,430,000
99
99.8
Lime
90
110
50/50
35
60
METRIC TONS
470
376
373
3
132
398
108
642
427
1069
41
30
METRIC TONS
373 + 642 +41+30
1086
427
1513
9-46
-------
TABLE 4. CALCULATION OF COAL WASTE BLOCK PRODUCTION RATES BY ONE V6-12
SUPERPAC OR TWO V3-12 VIBRAPAC MACHINES
(FOLLOWING ASSUMPTIONS IN TABLE 1)
Block unit size, nominal:20.3 x 20.3 x 40.6 cm (8 x 8 x 16 in)
Block unit volume:
Block forming cycle:
Block wet density:
Weight of 6 blocks:
Block forming rate:
Hourly block production:
0.017 m3 (0.59 ft3)
One 6-12 block machine
or two V3-12 block machines
1.76 g/cm3 (110 lb/ft3)
176.7 kg (391 Ib)
8 cycles per minute
176.6 x 8 x 60
At 90% efficiency
Ash & sludge (dwb) process rate
1 pallet of 6 units
0.100 m3 (3.56 ft3)
2 pallets of 3 units
0.100 m3 (3.56 ft3)
84,805 kg/hr
84.8 metric/hr
(935 short ton/hr)
76.3 metric ton/hr
(84.2 short ton/hr)
76.3 x 71.7% solids
54.7 metric ton/hr
(60.3 short ton/hr)
From Table 1,
Coal wastes to be processed daily (dwb) = 1086 metric ton (1197 short ton)
Block production time per day = 1086 metric ton/day = 19.8 hours
54.7 metric ton/hr
prevent damage to shellfish areas, and would not "fill in the ocean." In
fact, assuming the production of 31,000 blocks each day from a power plant
and allowing a doubling of volume when released on the bottom due to random
stacking, over 30 years of blocks could be contained in a reef covering 1
square mile and rising only 10 feet above the bottom. Water depth would
control how high the reef could be constructed, due to potential hazard to
navigation. Other potential uses of the blocks include rehabilitation of old
sand mining areas and enhancement of fishing in areas that, historically,
have had soft bottoms.
9-47
-------
THE ECONOMIC VIABILITY OF COAL WASTE REEFS
This program has demonstrated the technical feasibility of processing
power plant coal combustion by-products into blocks by using conventional
concrete block making equipment. The environmental acceptability has been
demonstrated in the laboratory and field investigations. To answer the big
question on whether this concept is commercially viable, the Electric Power
Research Institute commissioned Michael Baker, Jr., Inc. to do an evaluation
of this technology from a hypothetical coal plant located in the New York
Metropolitan area. The reef construction system would consist of four parts:
(1) The block product plant with storage facilities; (2) a barge loading
facility; (3) transportation to the reef site for placement, and (4)
monitoring the site. The engineering-economic assessment indicated that a
hypothetical 600-MW power plant would produce nearly 300,000 tons of
by-products each year. The reef construction system would cost about $45/ton
to build and operate. This cost is double the amount that most power plants
spend on by-product disposal when land is available nearby. In highly
urbanized areas where land is not available, the disposal costs are nearly
the same. Therefore the potential for reef construction as a viable
alternative to traditional waste disposal technologies looks promising and
can be considered by coastal utilities in waste management planning.
The comparative economics of construction of coal waste artificial
fishing reefs versus conventional land disposal practices would be
drastically changed if a bill now in the 98th Congress becomes law. The
proposed legislation (H.R. 3474) is called the National Fishing Enhancement
Act of 1983. The legislation would establish national standards for the
construction of siting of artificial reefs in the waters of the United States
in order to enhance fishery resources and fishing opportunities. To
stimulate the construction of artificial fishing reefs, the proposed law
provides a tax credit equal to the excess costs of reef construction over the
usual disposal costs. This bill if enacted would put artificial reefs on an
equal footing with conventional disposal costs anywhere. The reason for the
congressional interest in artificial reefs is that fishery products provide
an important source of protein and industrial products for U.S. consumption,
yet U.S. production annually falls $3.2 Billion short of satisfying demand.
REFERENCES
Knight, R. G., E. H. Rothfuss, and K. 0. Yard, 1980. FGD Sludge Disposal
Manual, Second Edition, Electric Power Research Institute, CS-1515,
Palo Alto, CA.
Parker, J. H., P. M. J. Woodhead, I. W. Duedall, and H. R. Carleton, 1981.
Coal Waste Artificial Reef Program, Phase 3, Volume 2. Electric Power
Research Institute, CS-2009, Palo Alto, CA.
Parker, J. H., P. M. J, Woodhead, I. W. Duedall, and H. R. Carleton, 1982.
Coal Waste Artificial Reef Program, Phase 4A. Electric Power Research
Institute, CS-2574, Palo Alto, CA.
9-48
-------
SOLID WASTE ENVIRONMENTAL STUDIES AT ELECTRIC
POWER RESEARCH INSTITUTE
I. P. Murarka
Presented by Karen Summers
-------
SOLID WASTE ENVIRONMENTAL STUDIES
AT ELECTRIC POWER RESEARCH INSTITUTE
by: Ishwar P. Murarka, Ph.D.
Environmental Physics and Chemistry Program
Electric Power Research Institute
Palo Alto, CA 94303
ABSTRACT
Solid Waste Environmental Studies (SWES) at Electric Power Research
Institute (EPRI) is a comprehensive research project aimed at generating
predictive methods and the essential data bases to evaluate the effect of
disposal and reuse of solid wastes produced from fossil fuel combustion and
flue gas cleanup operations on groundwater quality. EPRI has developed
detailed research plans and has initiated research in leaching chemistry,
chemical attenuation mechanisms, groundwater transport processes, and the
evaluation of existing geohydrochemical models. For the next three or four
years, fundamental research in geochemistry and geohydrology is expected to
yield quantitative data on release rates, transformation characteristics, and
subsurface transport of inorganic solutes leached from waste. Results of the
research will be integrated by improving or developing new predictive methods
and by validating the results with data from operating facilities.
INTRODUCTION
Do constituents released from land disposal and from reuse of solid
residues by electric utilities influence groundwater quality in the surrounding
area? What factors control leachability of chemicals? What solid waste types
are of concern? What is the environmental fate of leached solutes? When and
to what extent should technologies be applied to contain leachates? EPRI's
SWES project was initiated in 1982 to develop useful answers to these and other
questions.
SWES is a comprehensive research project dealing with fundamental studies
in geochemistry and subsurface hydrology to produce methods and associated data
9-49
-------
sets for use in evaluating the effects of disposal and the reuse of solid
residues on groundwater quality. The solid residues most emphasized in the
SWES research are coal fly ash, coal bottom ash, mixtures of the two, flue gas
desulfurization (FGD) sludges with fly and bottom ash, and oil ash.
GOALS AND OBJECTIVES
The long-term or ultimate goal of the SWES project is to develop and vali-
date methods (namely geohydrochemical models) for predicting the fate of inor-
ganic solutes released to the subsurface environments from utility industry
operations.
This long-term goal will be met by first conducting fundamental research to
define the cause-and-effect relationships of greatest importance and of highest
uncertainty to the development of predictive methods. Once the fundamentals in
the research areas are well understood and data sets are properly developed,
the results are to be integrated into predictive methods (models) and validated
with field data.
The near-term concurrent research is to accomplish the following:
(a) Evaluate existing geohydrochemical models and computer codes and
assemble an interim usable geohydrochemical model(s)
(b) Develop quantitative data on the leaching chemistry of solid
residues
(c) Develop quantitative data on the chemical attenuation in the
geologic environment of inorganic constituents
(d) Develop quantitative data on groundwater transport of solutes
(e) Develop mathematical descriptions of important leaching process-
es, chemical attenuation mechanisms, and hydrological transport
factors in the migration of solutes in groundwaters.
RESEARCH PLANNING
EPRI's initial research into solid waste environmental studies began in
1977 when projects were initiated on (1) chemical characterization of coal fly
ash (RP1061), (2) evaluation of reproducibility of EPA's extraction procedure
as applied to utility solid wastes (RP1487), (3) comparison of the physical and
chemical properties of solid wastes from coal combustion and coal gasification
processes (RP1486), and (4) providing a statistical evaluation of the variabil-
ity of a representative coal ash sample from an individual power plant (RP1620).
Until late 1980, the degree of emphasis that should be placed on solid waste dis-
posal and groundwater pollution potential was undetermined.
9-50
-------
In 1981 a planning study was conducted to define research for the SWES proj-
ect. This work was carried out through four regional meetings, simultaneous
use of questionnaires addressed to about 260 individuals followed by a research
planning workshop, and continuing discussions in SWES Advisory Committee meet-
ings. As a result, research plans for individual SWES contracts have been de-
veloped to accomplish its objectives. The bases for this research effort are
described in detail in two EPRI reports (Murarka 1982 and SAI 1982). Figure 1
provides a compartmentalized sketch defining the overall problem of estimating
the release, transport, transformation, and fate of leached constituents. To
date EPRI has initiated research contracts in the high-priority research areas
of waste leaching chemistry, chemical attenuation of solutes in the geological
environment, transport of solutes to groundwaters, evaluation of existing geo-
hydrochemical models, and evaluation of groundwater sampling methods and relat-
ed field measurements. Present plans call for an expenditure of about $21 mil-
lion dollars over the next six years.
RESEARCH METHODOLOGY AND MANAGEMENT
SWES is divided into the following six parallel research efforts:
1) Identifying and quantifying the solutes present in the utility
industry solid residues
2) Developing data on the leaching chemistry of these solutes from
the solids
3) Developing data on the chemical attenuation of the solutes in
the geological environments
4) Identifying and quantifying groundwater transport parameters for
solute migration
5) Evaluating, improving, and validating methods (primarily geohy-
drochemical model[s]) that can predict the fate of solutes in
land, surface, and groundwaters
6) Evaluating and testing groundwater sampling and related
measurement methods for use in the field environments.
The implementation of research in the six categories employs a general
strategy of evaluating the existing state of knowledge through critical litera-
ture reviews, a detailed specification of experimental designs, and laboratory
and field experiments to develop the quantitative results. An integrated ana-
lysis of the data would then provide mathematical descriptions of cause-and-
effect relationships for use in predictive methods. This analysis would also
provide data values for input parameters to the predictive methods. While the
fundamental research continues toward achievement of the long-term goal, an
interim assemblage of geohydrochemical model(s) will provide usable predictive
methods for the near-term.
9-51
-------
I
Ui
r-o
ESTIMATING THE TRANSPORT, TRANSFORMATION, & FATE OF CONSTITUENTS
INPUTS
WASTE CHARACTERISTICS
• Type
• Mass (Volume)
« Physical Properties
phase(s)
bulk density
permeability (T)
• Chemical Properties
ele composition
contaminant speciation
solubility, teachability
volatility, (KH, PC)
gov. chem. variables
" speaation Cu-'1'* OH - — CuOH -
ion-exchange ^^-^ complexation
R COOH - Me - — R - COOMe - H -
Microbially Mediated Rxns. Other
oxidation-reduction precipitation-dissolution
hydrolysis (chemical weathering)
conjugation pH
alkylation r5"1"r'"'~"~
ring fission, etc. t VN.
nw 0 - lOQ C r S\*V ^
ry^rr -^-(TcS \>C
•^ k^-OH V^-C-o- \ /V^
r
I , /'
-]/
OUTPUTS
HYDROLOGIC
• Peizometric surface
h(x,y,t)
• Flow Velocities
v(x,y.z.t)
CHEMICAL
« Solute Concentration
C(x,y,z,t)
• Solute Mass Flux
F(M/L2T)
INPUT TO
MGT. DECISIONS
• Evaluation of
Alternatives
• Estimation of
Acceptable Loadings
• Assimilative Capacities
• Monitoring Guidance
EJS
Evaluating Potential
for Ground Water
Contamination
INPUT TO REGULATORY
RULE-MAKING
PROCESSES
Figure 1. Estimating the Transport, Transformation, and Fate of Waste Constituents: Inputs,
Outputs, and Example Processes*
*This figure is taken from EPRI Report EA-2415, Figure 1-2.
-------
The management and technical direction of SWES contracts are accomplished
primarily through the EPRI project manager. However, because of the nature and
magnitude of this project, a technical management contractor is added to assist
the EPRI project manager. In addition, an SWES Advisory Committee has been
formed to advise the EPRI project manager on the SWES research. This advisory
committee is composed of seven utility industry scientists, six scientists from
academia and government agencies, two scientists from other EPRI divisions, and
the EPRI project manager as the chairman.
RESEARCH CONTRACTS
SWES now has seven contracts to conduct the parallel research (fig. 2). A
short description of the scope of work for these contracts is provided in the
following sections.
Geohydrochemical Models Evaluation and Development
This was the first SWES contract awarded. The research in this effort has
been divided into three phases:
• Phase I: The contractor is evaluating several models (computer
codes) on geochemistry, geohydrology, and microbiological pro-
cesses for their ability to predict the fate of constituents in
groundwater- This phase of research is to be completed by March
1984.
• Phase II: The contractor will assemble interim usable geohydro-
chemical model(s) by collating the best methods identified dur-
ing the Phase I research. The interim models connecting geo-
chemical and geohydrological processes for modeling the fate of
selected inorganic constituents will be tailored for near-term
use by the industry. The Phase II research is to be completed
by the end of 1986.
• Phase III: The results from research in chemical attenuation,
leaching chemistry, and groundwater transport processes will be
integrated in the improvement and field validation of geohydro-
chemical models. Present plans call for Phase III research to
be completed by 1990.
Attenuation Rates, Coefficients, and Constants in Leachate Migration
Chemical interaction of solutes in solution with the surrounding soils and
geological materials has profound effect on the migration of the leached con-
stituents to groundwater. Therefore, the study of chemical attenuation has
9-53
-------
VD
I
Ln
-P-
Pre 1982
1982
1983
1984
1985
1986
1987
Post 1987
Project Development and Management
by EPRI Staff
Technical Management Through Contract and EPRI Staff,
Research Results Transfer 2485-1
Physical Chemical Characterization
and Regulatory Leachate Testing
1061, 1486, 1487
Environmental Setting and
Disposal Systems 1487-13, 2198-7
Initial Evaluation of
Geohydrochemical Models
1619-1
Interim Models Development
and Testing
2485-2
Models Improvement,
Development and Validation
2485-2
Short- and Long-Term Chemical Attenuation Studies
(2198-1), 2485-3
Short- and Long-Term Leaching Studies
(2198-2), 2485-4
J L
Saturated and Unsaturated Groundwater Transport Studies
(2280-1), 2485-5, 2485-6
Groundwater Sampling and Measurement Methods
(2283-1), 2485-7
t I
Groundwater Data Analysis & Evaluations
(2283-2), 2485
Sampling and Temporal Variations Statistics
1620-1
Geochemical and Geohydrological Field
Measurements 2485-9
Figure 2. Solid Waste Environmental Studies (SWES) Project Contracts & Schedule
Environmental Physics & Chemistry Program, EPRI, September 1983
-------
been considered as the area that can produce the largest incremental improve-
ments in methods for predicting leachate migration to groundwaters. Thus, this
research is aimed at fundamental processes at the onset.
The research has been divided in two successive phases. In Phase I the re-
searchers have compiled and collated the quantitative data from existing
literature on chemical attenuation rates, constants, and coefficients for 21
inorganic elements (Table 1).
The critical evaluation of the literature in Phase I has included data
derived from laboratory and field research. Emphasis has been placed on the
acquisition of existing data on adsorption-desorption, precipitation-
dissolution, and thermodynamic data on stability constants for the 21 inorganic
elements. Thermodynamic data have been used to evaluate the coefficients and
constants for chemical attenuation mechanisms applicable to leachates from
utility industry solid residues. A critical review of the experimental or
observational procedure used behind the literature data has been done prior to
accepting the data values for this research. This review of literature and
assembly of data was completed in July 1983.
TABLE 1. INORGANIC ELEMENTS FOR CHEMICAL ATTENUATION STUDIES
Aluminum
Barium
Cadmium
Fluoride
Manganese
Nickel
Sulfate
Antimony
Beryllium
Chromium
Iron
Mercury
Selenium
Vanadium
Arsenic
Boron
Copper
Lead
Molybdenum
Sodium
Zinc
In Phase II the contractor designed and started laboratory experiments (on
solubilities, solubility products, and kinetics of precipitation-dissolution of
solid phases) to produce data on the solubility-limited concentrations of se-
lected elements in the natural environment. Additional experiments will be
conducted to measure intrinsic adsorption constants, quantify the effects of
competing ions and complexing ligands on the adsorption-desorption characteris-
tics, and to quantify adsorption-desorption on mineral mixtures and natural
soils for single and multiple electrolytes. These experiments will employ
batch, column, and controlled field set-ups to simulate the complexity of the
environment. Attenuation rate constants will be established by summarizing the
results. This contract research is expected to be complete by the end of 198?.
9-55
-------
Solid Waste Leaching Studies
What factors (e.g., composition of waste and environmental conditions) con-
trol leachability? What constituents can be leached from wastes? Under what
conditions? In which chemical form? How does one determine the rate of
leaching of a constituent for field conditions? Answers to most of these
questions lie in understanding the chemical mechanisms. Earlier research by
Hulett et al. (1981) and by Turner et al. (1982) has identified chemical forms
of elements and distribution of trace elements found on the different solid
phases of fly ash samples.
This work has provided a firm basis for research on chemical mechanisms for
waste leaching. Therefore, the objectives of the planned leaching studies are
to collate available data, critically evaluate the methods used in producing
the data, and develop quantitative results on the leaching chemistry of solid
residues. The research on leaching studies is divided into two phases. In
Phase I the researchers are compiling quantitative data from existing litera-
ture, conducting feasibility experiments, and designing experiments for Phase
II. All batch, column, sequential, and other leaching methods used to produce
quantitative data on the chemical composition of leachates are being critically
examined to identify relevant variables that are particularly important in con-
trolling the leaching behavior of the fossil fuel combustion wastes. These
variables could include factors such as occluded liquors, pH and quality of
natural waters that such wastes come in contact with, solid-liquid ratio,
permeability of wastes, age of wastes, contact time between waste solids and
liquids, and the solid phases present. This phase of research is to be com-
pleted by April 1984.
In Phase II the researchers will conduct laboratory and controlled field
experiments as a means of developing quantitative data on the leaching rates of
constituents from utility solid wastes. These experiments will be conducted to
include a broad range of waste materials and environmental conditions such that
the results can be useful to a large segment of the utility industry. Phase II
research is expected to be complete in 1986.
Groundwater Transport Studies
Solutes generally migrate by groundwater flow. The groundwater flow and
transport calculations are based on mathematical equations using Darcy's formu-
lation during the past 50 years. But the calculations on solute transport by
groundwater are relatively new, perhaps less than 10 years old. The solute
transport problem has two major processes: a chemical interaction phenomenon
and a physical dispersion phenomenon. The groundwater transport studies relate
to the physical dispersion phenomenon. Two contracts have been let out to con-
duct field experiments on groundwater transport of solutes in the porous media.
The research is divided in two phases. In Phase I the researchers are
compiling and evaluating the existing data on physical-hydrological transport
factors used to predict the fate of solutes in groundwaters. Based on this
9-56
-------
critical review, tracer experiments will be designed for several field sites to
measure the transport parameters in the development of plumes. The Phase I
research is to be completed by June 1984. In Phase II experiments are to be
conducted for at least three sites for a period of three or more years to esti-
mate the hydrodynamic dispersion parameters as a function of environmental var-
iables and the solutes of interest. The groundwater transport research is
expected to be complete by the end of 1987.
Groundwater Sampling Methods and Related Field Measurements
Because the ultimate goal of SWES is to produce methods for predicting the
fate of solutes in the environment, reliable input data (e.g., soil and waste
permeability, groundwater flow velocity, and predictive groundwater flow direc-
tions) are needed. The research in this contract establishes the availability
and performance of field methods for sampling geochemical and groundwater envi-
ronments specifically.
This research is divided into two successive phases. In Phase I the
contractor is to compile and critically evaluate existing geochemical and
hydrologic field sampling procedures. This evaluation will focus on at least
the following three categories of sampling methods:
1) Groundwater sample collection, handling procedures, and instru-
ments
2) Soil and waste sample collection, handling procedures, and in-
struments specifically for quantifying hydrologic and geochem-
ical properties
3) Geophysical methods for geologic and subsurface hydrogeologic
measurements.
The contractor is to prepare a manual containing details on field sampling
procedures to define the strengths, weaknesses, and applicability to the util-
ity industry. The Phase I research is to be completed by August 1984.
In Phase II the contractor will develop and field test procedures for
groundwater sampling problems identified in Phase I. These tests will be re-
peated several times to establish reproducibility, accuracy, and ease of use
for utility industry applications. The Phase II research is expected to be
complete in 1987.
Groundwater Data Evaluation
To properly direct the applied research in SWES, a contract has been issued
to analyze data from a full-scale field monitoring project conducted for the
U.S. EPA and individual utility companies. The information from these studies
9-57
-------
includes data on groundwater quality, hydrology, soils and related geological
measurements, and waste disposal systems. The contractor will analyze these
data to identify hypotheses that should be tested in the SWES research on chem-
ical attenuation, leaching chemistry, and groundwater transport of solutes.
This data analysis is to be completed by July 1984.
Geochemical, Geohydrological, and Groundwater Quality Measurements
This research is planned for the later part of SWES. Intensive measure-
ments at a selected number of sites will be made to develop a complete data
base for the testing and validation of predictive methods developed in the SWES
project. This contract is expected to begin in 1985 at the earliest and may
continue for approximately four years. Detailed sampling plans for this con-
tract will be specified prior to the beginning of the measurements.
RESEARCH RESULTS
Since this research began in 1977 and the initial phase of the current SWES
research started in 1982, reports are now available that offer usable results.
In this section a summary of the technical results from several EPRI reports
has been prepared. For complete details, readers are referred to the reports
sited in this paper.
EXTRACTION PROCEDURE AND UTILITY INDUSTRY WASTES
Two EPRI reports (Rose et al, 1981 and Eynon et al. 1983) deal with the
topic of reproducibility of results (concentration of eight elements: As, Ba,
Cd, Cr, Pb, Ag, Se, and Hg) in the extracts from using the extraction procedure
(EP) specified by the U.S. EPA as applied to fly ash, bottom ash, and scrubber
sludge. The concentration of silver in the EP extract was found to be below
detection limit most of the time; therefore, no statistical analysis was per-
formed on the data for silver.
The solid samples used in this research were: (1) dry fly ash—alkaline
(DFA--ALK), (2) dry fly ash—acidic (DFA--AC), (3) wet bottom ash—alkaline
(WA--ALK), (4) wet bottom ash—acidic (WA—AC), and (5) scrubber sludge (SS).
Four different laboratories were used to extract four samples for each of the
five waste types using the EPA's extraction procedure of December 1978. Each
extract liquid was split into eight aliquots by each laboratory. Each aliquot
sample was analyzed for the eight elements using two different analytic proce-
dures. Numerical results obtained from this research are summarized in Table 2.
Based on this research the following conclusions can be drawn:
• Results on the concentration of the eight elements were found to
be hundreds of percent apart between different laboratories.
9-58
-------
TABLE 2: SUMMARY STATISTICS AND APPORTIONED VARIATION IN LEACHATE COMPOSITION*
I
i_n
'*£>
Element
Arsenic
Barium
Cadmium
Chromium
Mercury
Lead
Selenium
Analysis
FURNACE AA
FLAME AA
FURNACE AA
FLAME AA
FURNACE AA
FLAME AA
FURNACE AA
FLAME AA
FLAME AA
FURNACE AA
FURNACE AA
FLAME AA
Prep
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
EP
Waste
DFA-AC
DFA-ALK
SS
DFA-AC
DFA-ALK
SS
DFA-AC
DFA-ALK
WA-AC
WA-ALK
SS
WA-ALK
DFA-AC
DFA-ALK
WA-AC
WA-ALK
SS
DFA-AC
DFA-AC
DFA-ALK
WA-AC
WA-ALK
SS
DFA-ALK
SS
DFA-AC
SS
DFA-AC
DFA-ALK
SS
DFA-AC
DFA-ALK
SS
N
Samp
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
126
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
Used
128
128
128
128
128
128
128
128
128
128
128
127
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
128
Geom
Mean
(ppb)
7.14
7.07
77.22
6.21
4.58
42.98
69.96
129.32
43.23
181.79
83.44
890.54
95.81
0.19
0.37
1.56
7.09
126.33
9.01
536.74
1 .09
2.72
28.32
590.10
0.73
1.04
0.88
24.11
51.88
39.08
5.25
20.37
18.60
Minimum
(ppb)
0.63
0.63
10.00
1.00
0.17
6.55
1.43
0.63
1 .43
0.49
1.26
24.86
23.00
0.01
0.03
0.28
0.58
38.03
2.50
331.57
0.11
0.16
2.61
195.38
0.05
0.14
0.18
4.63
23.53
15.69
0.05
0.30
0.60
Maximum
(ppb)
46.00
110.00
220.00
140.00
147.23
288.74
650.00
3400.00
1300.00
1600.00
730.00
32000.00
240.00
48.00
11 .00
22.00
62.31
240.00
31.69
830.00
412.37
290.00
110.00
1000.00
59.00
9.00
5.44
70.00
121 .62
73.00
62.00
89.00
68.00
Pei
Inter
Lab
Extr
32
4
47
35
10
29
0
0
0
2
1
6
28
0
12
7
13
29
0
11
0
42
70
3
53
12
6
23
16
0
4
6
2
-cent o
Intra
Lab
Extr
2
44
12
2
33
1
0
3
7
0
0
0
1
29
12
6
5
0
28
39
71
18
24
65
4
0
0
2
34
15
0
15
0
f Total
Inter
Lab
Anal-I
51
31
37
14
4
10
82
73
65
88
90
63
45
41
56
65
70
32
4
13
9
2
3
15
0
54
32
34
11
13
86
58
72
Variance
Inter
Lab
Anal-II
9
7
2
17
36
23
5
18
12
0
0
16
7
2
2
11
4
13
19
13
10
22
0
13
24
7
25
6
22
23
3
14
4
Intra
Lab
Anal
6
13
2
32
17
38
12
6
15
11
10
15
20
28
17
10
8
26
50
24
10
16
3
4
19
27
38
36
17
49
6
7
21
* The data in this table are taken from EPRI Report EA-3181, Tables 3-2 and 3-5.
-------
• Reproduciblity differed from element to element, waste type, and
the analytic technique used.
• The interlaboratory variation in results was generally higher
than the intralaboratory variation.
TRACE ELEMENTS AND THEIR CHEMICAL FORMS
Two EPRI reports (Hulett et al. 1981 and Turner et al. 1982) deal with the
general subject of ash chemistry. The objective for Part I in this research
was to establish the distribution of trace elements in the various solid phases
present in fly ash. By using physical and chemical methods, researchers first
separated the fly ash samples into three phases: (1) glass (approximate compo-
sition Al-Si-0), (2) mullite-quartz (approximate composition 3(Al2Oo).2(Si02)),
and (3) magnetic spinel (approximate composition Fe2A10{j). The distribution of
trace elements in each of the three phases was then defined. The research
results show:
• Glass is typically the most abundant solid phase of the non-
magnetic alluminosilicates.
• Trace elements distribution analyses show that a distinct pro-
pensity occurs in specific phases:
— The glass phase seems to contain almost all of the alkali,
alkaline earth, and rare earth elements, as well as the
majority of As, Pb and Se.
— Small amounts of trace elements are associated with the
mullite-quartz phase with the exception of V, Cr, Ga, Zr, Fe,
and Ti—all of which seemingly occupy the +3 or +4 valence
sites in the crystalline structure.
— The magnetic-spinel phase seems to be largely enriched in the
first row transition elements (V, Cr, Mn, Fe, Co, Ni, Cu, and
Zn).
The objective of research in Part II was to identify patterns in the coal
ash composition and its leaching behavior in actual and laboratory-simulated
disposal experiments. Arsenic was selected for a more complete study on
leachability and chemical speciation in leachates. The important results
reported by the researchers (Turner et al. 1982) are summarized below:
• Ash sluicing-ponding systems that are neutral to alkaline in pH
appear to show appreciable solubilization of elements (e.g., S,
As, Se, Cr, and Mo) with the anionic aqueous speciation. In
9-60
-------
contrast, the sluicing-ponding systems with pH in the acidic
range appear to result in high solubilization of most cationic
elements (particularly Fe, Mn, Cu, Zn, and Cd).
• The study of As indicated that in controlled pH experiments the
As (v) varied systematically with pH variations and the As (III)
concentration remained nearly constant over a wide range of pH
(3 to 12).
• A considerable amount of variablility among fly ashes with
respect to quantity and to the oxidation state of arsenic
released was observed in the laboratory leaching experiments.
COMPARISON OF SOLID WASTES FROM COAL COMBUSTION AND PILOT COAL GASIFICATION
PLANTS
The research results from this work are given in an EPRI report prepared by
Turner and Lowry (1983). The researchers principally focused on developing
data on the physical and chemical characteristics of solid wastes from conven-
tional coal-burning power plants and solid wastes from coal gasification in the
CE-PDU plant. Bulk composition results of the waste samples analyzed are shown
in Table 3 and the EP extract composition results are shown in Table 4. The
important results from this contract research are summarized below:
• Slag from the gasifier and slag from the cyclone-fired combus-
tion unit were morphologically similar and predominantly con-
sisted of a dense shards of specific surface area of less than
1 m /g. In contrast, the bottom ash from a pulverized-coal com-
bustion plant predominantly contained round vesicular cinders of
specific surface area between 1 and 3 m /g.
• The two slags did not show any crystalline structures, but
mullite crystals were found in the fly ash from the pulverized
coal combustion power plant.
• Based on aqueous extraction and inorganic elements analyses
according to the EPA extraction procedure, none of the com-
bustion or the gasification solid wastes studied would be
classified as hazardous on the basis of the toxicity criteria
of the Resource Conservation and Recovery Act regulations.
• The gasification slag showed a more acidic reaction in
continuous-flow column extractions and thus may leach some
constituents at a higher rate than combustion slag and bottom
ash.
• Although some differences exist between the wastes examined, the
gasification slag appears to be nearly as inert chemically as
slag from the cyclone furnace.
9-61
-------
TABLE 3. ELEMENTAL COMPOSITION OF COMBUSTION AND GASIFICATION SOLID
WASTES PRODUCED FROM PITTSBURGH COAL*
Si
Al
Fe
Ca
Na
Mg
K
Ti
S
C
Ag
As
Ba
Cd
Cr
Cu
Hg
Mn
Ni
Pb
Se
Zn
Mitchell
Fly Ash
20.5
11.6
10.2
3.09
0.56
0.45
1.19
0.62
0.80
7.04
<1
93.
743.
0.82
113.
68.
0.10
188.
129.
50.
12.
92.
Station**
Bottom Ash
21.6
11.9
14.9
2.38
0.55
0.46
1.10
0.56
0.18
0.58
<1
3.2
551.
<0.1
87.
44.
0.006
176.
122.
8.
N.D. ******
32.
##*
CE-PDU
Slag
«
19.2
14.9
14.4
1.32
0.87
0.31
1.32
0.59
0.31
0.80
MB' 6
<2
3.6
1470.
0.18
4750.
68.
0.011
273.
228.
4.9
1.0
28.
Merrimack
Slag
21.4
12.6
12.2
5.52
0.94
1.62
1.03
0.60
0.07
0.16
<2
<1.8
1253.
0.64
126.
69.
0.008
205.
98.
8.1
2.2
32.
Station****
. , ******
Fly Ash
18.0
11.2
13.1
3-72
1.34
0.73
1.58
0.58
0.91
4.6
<2
393.
1400.
4.2
298.
431.
0.04
182.
158.
309.
N.D.
460.
* This table is taken from EPRI Report EA-2867 Table 3.4.
** Pulverized coal furnace.
*** Entrained-flow gasifier (low Btu), collected June 3, 1980.
**** Wet-bottom slagging furnace.
***** Recycled to furnace at this power plant.
****** Not detected.
9-62
-------
TABLE 4. RESULTS OF EPA'S EXTRACTION PROCEDURE APPLIED TO SOLID WASTES
PRODUCED FROM PITTSBURGH COAL* (concentrations are in
yg/L except for SOJj which is in mg/L)
Mitchell Station** CE-PDU****
Ag
As
Ba
Cd
Cr
Cu
Fe
Hg
Mn
Ni
Pb
Se
Zn
sojj
Acid,
mL/g
Final
pH
Fly Ash
0.06
9.2
145.
12.
87.
106.
362.
0.014
1700.
188.
4.35
2.8
160.
940.
1.2
5.0
Bottom Ash
<0.01
0.88
85.
0.12
1.3
1.4
144.
0.011
244.
11.
<0.06
<0.05
5.7
16.2
0.15
5.0
Slag
<0.05
0.55
58.
0.06
1.45
5.1
326.
0.058
4.45
180.
0.53
<1
5.2
6.24
0
4.6
Merrimack
Slag
<0.05
0.09
2.5
0.11
<0.1
0.74
275.
0.004
1.3
0.28
0.14
<0.5
0.61
--
0.01
4.8
Station
Drinking
Water
Fly Ash****** Criteria
0.12
63.
100.
69.
1.5
290.
22.
0.04
580.
210.
3.1
1 1 .
1500.
--
1 .2
5.1
50
50
1000
10
50
1000
300
2
50
—
50
10
5000
250
* This table is taken from EPRI Report EA-2867 Table 3.8.
** Pulverized coal furnace.
*** Wet-bottom slagging furnace.
**** Entrained flow gasifier (Low Btu), collected June 3, 1980.
***** Recycled to furnace at this power plant.
9-63
-------
STUDY OF TIME-VARIABILITY OF ELEMENT CONCENTRATIONS IN POWER PLANT ASH
This research contract was initiated to study the uncertainty in ash
composition due to representativeness of the ash samples subjected to the
extraction test. This study is critical to answering: "What constitutes
a representative waste sample, and how variable is the waste stream over
time?" The research has been conducted in two parts. Results from the
Part I pilot study are given in an EPRI report prepared by Switzer,
Eynon, and Holcombe (1983). Part II research has also been completed and
the report is in preparation. The Part I research consisted of obtaining
samples of coal and coal ashes from a southwestern power plant for a 30-
day period. The samples were chemically analyzed for 14 inorganic ele-
ments. The elemental concentration data have been subjected to statisti-
cal analyses to estimate the time variability in chemical composition of
coal ashes. Results from this research are summarized below:
• Time, sampling, sample preparation, and sample analysis are the
four components providing sizable contributions to the total
variability in ash composition.
• The contribution of the four components to total variability
differed in degree depending upon the chemical element and
method of analysis.
• The correlation between plant process variables (feed coal com-
position, boiler temperature, boiler oxygen, and electrostatic
precipitator air temperature) and the ash chemical composition
showed very little association. This lack of correlation may be
due to the relatively constant chemical composition of feed coal
during this research.
PHYSICAL-CHEMICAL CHARACTERISTICS OF UTILITY SOLID WASTES
The primary purpose of this research was to compile existing data on waste
physical properties, chemical composition, and leachate composition for esti-
mating the physical-chemical composition of solid wates. The EPRI report by
Summers, Rupp, and Gherini (1983) provides detail on the results. A large
quantity of the data compiled are on coal ash and its leachates. These data
are for 19 minor and trace elements. Relatively small amounts of data are on
physical properties of ash and flue gas desulfurization (FGD) sludge, chemical
composition of FGD sludge and leachates, and chemical composition of oil ash
and geothermal wastes. The results of this research are summarized in Figures
3 and 4.
9-64
-------
PLANT COAU ASH
PftOLXJCTION £ PI5PO5AU
HICM, I—» pLV
CDU.tcrof>. ASH
BOTTOM ASH
RELATIVE AMOUNTS AMMOM- 615 A5H
I
ON
Ln
TOTAL-'. 68 million
tons per year
: 16 million
•tons per year
PDFD5EP
| j5oF*et
l-AHPFIU-
FmWfj'ft'tttr'ftp
fonp
Prrr51CAU-CH£MICAU
PHYSICAL. CHARACTERISTICS
FLY ASH 0.5-IOO^ L PW55
35 )
-x 03)
Co. V
Quorti
GlaM
Carbon
10 j M^-Ti^Zr
10 Al
35
IO
LEACHING
EXTRACTION R£56ILT5
AVG. CQNC H9/t
SILVER
AR5EHIC
BAKlOt-1
CAPMICJM
CHROMIUM
MtRCdJRV
LEAP
.6
12
200
4,7
J6
O4
5
10
O.I — 2OO
04 - I6OO
3 - 76OO
O.I — I4OO
I - 700
0-—
(i-apid hyarolysis of 5tJrface mctoV
5ubs«)cient pH anop dtrf to wash oat of aikolinrty.
Pnpp apon air eqailibratlon: tV+^OH'+ZCO, (oir)
— fV'+2HCO,-
Acidio Ashes: Innd dissolution of strong odd onions
(e.g 5O^-, C|- ) exceeds tti« dissola+ion of base
cations («jj.Co*+,No')
Figure 3. Summary of Physical-Chemical Characteristics of Power Plant Coal Ash*
*This figure is taken from EPRI Report EA-3236, Figure 7-1.
-------
PLANT PGP
PROPdCf ION ^ PI5PO5AU
VTJ
1
Co-bostd Met 80
FWj«r*rnbl< R-CXKCO S
Pry FVoozsses 2
BA5I5:®IOOO M«4z Plant
- 15% 5 Coal
LZTtAsh
AVet Ume 5crabt*r
'
©3?OOO MWe
Scrubbed CN»)
OS PRORJCtlOn~l
P15FO5AL-
VOLaM& P15PO5EP
PhffSICAL-CHEMICAU
PHYSICAL CHARACTERI5riC5
SHAPES'. 5dlUFrT6R)6H-Platey 5dJUPATeR-&bcJ<,5
SEE. RAHGt: l-60/u«qaiv diam. (80% by mass)
i-IOO-IIS^i
„ . io-*toiO-Jgj-
fclmilar to sandy clay)
rixed ^fajdan IO"5 to I0'r£g-
(simitar roclay)
POROSITY: Ra« audoe 20-60%
Fixed Sludge <5 %
CH&hlCAL COMP05ITlOn
MAJOR COMPOrl&rrr5 (Wet U F&P Sludge 1
• Calcitim salfate. Ca5O4'Ci-2)H2O
• Calcium salflte CaSOj
• Calcium cartxpnate CaCOa
• 5odiam Sijlfate MaaSG\-
• "Incrts^CAsh and Impurrlies)
AVG. -TKAC& ElfcriEJ _ .
C5asi5. 7 investigations —OVet FQP 5ladg
&UeMfirfT COttC IM ,5OUP5 COriC IM
mg/hg ^9A
Arsenu: IO IO
Boron 134 I4.OOO
Chromium 13 10
Copper 27 25
Flaorine 53O 3,OOO
Mercery O.3 ' O.8
Urad 6 2O
5?leni£Jm 6 IOO
• Umited Pata
• Similar to wet systiem sludges
wh^n asinq +h« same, absortuznf
• Mo trace clement" data available
UACHING CHAr\ACTEf\l5TlC5
EP EAtCh EXfKACTIOH
5IL-VEP,
ARSENIC
BARIUM
CAPMKJM
CONCEMTRATIOMS fig/f
WSf F6P 5U-IPOC PRY fcrP 5UJPQC
<6O -
6O4-4O HbOO
<-)00-|oOO —
<25O H5OO
-------
GEOHYDROCHEMICAL MODELS FOR SOLUTE MIGRATION—AN EVALUATION
An initial evaluation of existing geohydrochemical models and codes is be-
ing carried out by the researchers in Phase I. The first volume of the reports
is in the final stages of preparation (Kincaid, Morrey, and Rogers, 1983). This
volume contains descriptions of processes deemed important in solute migration
and a selection, for initial evaluation, of 21 existing codes (Table 5) which
deal with these processes. These codes employ one or more of the three compon-
ents in geohydrochemical model(s): (1) hydrological flow and transport, (2)
geochemical interactions influencing the availability of a constituent for mi-
gration, and (3) microbial transformation which may alter the constituent dur-
ing migration. The second and third volumes are to be prepared by early 1984
and will report on the findings of simulations on the applicability of a subset
of these selected codes for predicting solute migration around the utility
industry disposal sites.
CHEMICAL ATTENUATION RATES, COEFFICIENTS, AND CONSTANTS IN LEACHATE MIGRATION
A two-volume report prepared under this Phase I research contract is being
finalized for publication. In Volume I of the report (Rai and Zachara, 1983)
data obtained from the available literature have been summarized for each of
the elements. The quantitative results given allow therrnodynamic data-based
predictions of the relative stability of solid and aqueous species, observed or
hypothesized solubility controls in the geochemical environments, and the rates
of reactions for adsorption-desorption and precipitation-dissolution processes
for chemical attenuation. Most of the data compiled pertains to single element
and pure minerals.
Volume II of this report (Rai, Zachara, Schmidt, and Schwab, 1983) is an
annotated bibliography of over 350 publications which appeared in existing
literature prior to February 1983 and deals with research on chemical (precipi-
tation-dissolution and adsorption-desorption) and biological (methylation and
alkylation) attenuation mechanisms for Al, As, B, Ba, Be, Cd, Cl, Cr, Cu, F,
Fe, Mn, Mo, Na, Ni, Pb, Sb, Se, SO^, V, and Zn.
SUMMARY AND CONCLUSIONS
In this paper a chronological summary of the SWES project at EPRI is devel-
oped. Although the research is in an early stage of implementation, it is
expected that thorough and deliberate planning has given the research a very
high chance for successful completion. In the short-term (the next 3 to 5
years), the research results will be in the form of data which can be used in
assessing the potential for release, transformation, and transport of inorganic
solutes from utility industry solid residues. In the long term (the next 6 to
8 years), the research results are to provide predictive methods (mostly geohy-
drochemical model [s]) to quantitatively estimate the fate of solutes in the
groundwaters.
9-67
-------
TABLE 5. COMPUTER CODES SELECTED FOR EVALUATION IN SWES
I. UNSATURATED FLOW AND TRANSPORT CODES:
Code Designation
(1) SESOIL
(2) NRC-SLB
(3) OR-NATURE
(4) UNSAT1D
(5) FEMWATER/FEMWASTE
(6) TRUST/MLTRAN
(7) SATURN
II. SATURATED FLOW AND TRANSPORT CODES:
(8) PATHS
(9) SWIP2/SWENT
(10) TRANS
(11) VTT
(12) FE3DGW
(13) USGS MOC
(14) AT123D
III. EQUILIBRIUM GEOCHEMICAL CODES:
(15) GEOCHEM
(16) MINTEQ
(17) PHREEQE
(18) EQUILIB
(19) EQ3/EQ6
IV. MICROBIOLOGICAL CODE:
(20) BIOFILM
V. HYDROLOGICAL AND GEOCHEMICAL CODE:
(21) FIESTA
Description/Function
Compartmental
One-dimensional
One-dimensional
One-dimensional
Multi-dimensional
Multi-dimensional
Multi-dimensional
Multi-dimensional-analytical
Multi-dimensional
One or two dimensional
One or two dimensional
Two or three dimensional
Two dimensional
Multi-dimensional-analytical
Equilibrium, geochemistry,
includes adsorption
Equilibrium, geochemistry,
includes adsorption
Precipitation-dissolution
Precipitation
Precipitation
Primarily organic compounds
One-dimensional flow and
transport coupled with
equilibrium geochemistry
9-68
-------
REFERENCES
Eynon, B. and Switzer, P., 1983. A Statistical Comparison of Two Studies on
Trace Element Composition of Coal Ash Leachates. Palo Alto, CA: Electric
Power Research Institute. EPRI EA-3181.
Hulett, L.D.; Weinberger, A.J.; Ferguson, N.M.; Northcutt, K.J.; and Lyon, W.S.
1981. Trace Element and Phase Relations in Fly Ash. Palo Alto, CA:
Electric Power Research Institute. EPRI EA-1822.
Kincaid, C.T.; Morrey, J.R.; and Rogers, J.E., 1983. Geohydrochemical Models
for Solute Migration: The Selection of Computer Codes and Description of
Solute Migration Processes. Volume I. Palo Alto, CA: Electric Power
Research Institute.
Murarka, I.P., 1982. Solid-Waste Environmental Studies; The Needs and the
Priorities. Palo Alto, CA: Electric Power Research Institute. EPRI EA-
2538-SR.
Rai, D. and Zachara, J.M., 1983. Chemical Attenuation Rates, Coefficients, and
Constants in Leachate Migration: A Critical Review. Volume I. Palo Alto,
CA: Electric Power Research Institute.
Rai, D.; Zachara, J.M.; Schmidt, R.A.; and Schwab, A.P., 1983. Chemical
Attenuation Rates, Coefficients, and Constants in Leachate Migration: An
Annotated Bibliography. Volume II. Palo Alto, CA: Electric Power Research
Institute.
Rose, S.J.; Dane, J.; Eynon, B.; and Switzer, P., 1981. Extraction Procedure
and Utility Industry Solid Waste. Palo Alto, CA: Electric Power Research
Institute. EPRI EA-1667-
Science Applications, Inc., 1982. Planning Workshop on Solute Migration From
Utility Solid Wastes. Palo Alto, CA: Electric Power Research Institute.
EPRI EA-2415.
Summers, K.V.; Rupp, G.; and Gherini, S., 1983. Physical-Chemical Character-
istics of Utility Solid Wastes. Palo Alto, CA: Electric Power Research
Institute. EPRI EA-3236.
Switzer, P.; Eynon, B.; and Holcombe, L.J., 1983. Pilot Study of Time Vari-
ability of Elemental Concentrations in Power Plant Ash. Palo Alto, CA:
Electric Power Research Institute. EPRI EA-2959.
Turner, R.R. and Lowry, P.D., 1983. Comparison of Solid Wastes From Coal
Combustion and Pilot Coal Gasification Plants. Palo Alto, CA: Electric
Power Research Institute. EPRI EA-2867.
Turner, R.R.; Lowry, P.; Levin, M.; Lindberg, S.E.; and Tamura, T., 1982.
Leachability and Aqueous Speciation of Selected Trace Constituents of Coal
Fly Ash. Palo Alto, CA: Electric Power Research Institute. EPRI EA-2588.
9-69
-------
SESSION 10, PART I: DRY FGD: PILOT PLANT TEST RESULTS
Chairman: Theodore G. Brna
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
CURRENT STATUS OF DRY S02 CONTROL SYSTEMS
M. A. Palazzolo, M. E. Kelly. T. G. Brna
-------
CURRENT STATUS OF DRY SO CONTROL SYSTEMS
by: Michael A. Palazzolo Mary E. Kelly
Radian Corporation Radian Corporation
Durham, NC 27705 Austin, TX 78766
Theodore G. Brna
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
This paper provides an update on commercial applications and research
and development (R&D) activities involving three dry SO control techno-
logies: spray drying, dry injection, and electron-beam irradiation.
Spray drying continues to be the only commercially applied dry flue gas
desulfurization (FGD) process, with two additional spray drying systems
having been sold since mid-1982. To date, 17 commercial utility spray
drying systems have been sold, totalling over 6,800 MWe. Six of these
systems are currently operational with two of the systems having been turned
over to the utility. In addition to the utility applications, there are
21 industrial spray drying units, 7 of which are operational. Several
additional utility and industrial systems are expected to start up in the
next 2 years. Demonstration- and pilot-scale testing of the spray dryer
process is continuing with emphasis on high sulfur applications.
The first planned commercial application of dry injection technology
has been announced for a 500 MWe utility. The recent demonstration-scale
testing on a 22 MWe unit has been completed.
The electron-beam process is also in an early developmental state.
Pilot-scale testing of the electron-beam/lime spray drying version of the
process is scheduled to begin this fall.
INTRODUCTION
This paper updates the commercial and research and development (R&D)
status of dry SO. control processes for utility and industrial boilers.
Four EPA reports on the status of dry S0» control systems have been
published to date (1, 2, 3, 4), the most recent in August 1983. A fifth
update is currently being prepared, with completion scheduled for early
1984. The primary sources of information have been contacts with dry SO.
control system vendors and purchasers and government and industry agencies
10-1
-------
involved with dry S09 control in the United States. Three types of dry SO
control (nonregenerable) processes'•will be covered in this paper, the latter
one briefly because of coverage elsewhere during this symposium:
• Spray drying.
• Dry injection.
• Electron-beam irradiation.
Spray drying, which includes a fabric filter or an electrostatic
precipitator (ESP) for collection of fly ash and desulfurization products,
is the only one of the three processes to be commercially applied. Dry
injection, a process involving injection of dry alkaline material directly
into the flue gas and subsequent collection of the desulfurization products
and fly ash in a fabric filter, is nearing commercial application. The
electron-beam process, which involves reagent injection into the flue gas
followed by electron-beam irradiation, is in an early developmental state.
A more detailed description of these processes, including some of the
important process variables and design parameters, is presented below. The
technology descriptions are followed by a discussion of the current
commercial applications and R&D activities for each of the three processes.
TECHNOLOGY DESCRIPTIONS
SPRAY DRYING
Figure 1 is a schematic diagram of the spray drying FGD process. The
gas containing fly ash and S0~ enters the spray dryer and is contacted with
a finely atomized alkaline slurry or solution. During the approximately
10-second residence time in the dryer, the flue gas is adiabatically
humidified as the water in the slurry or solution is evaporated. Simul-
taneously, flue gas SO reacts with the alkaline species to form solid
sulfite and sulfate salts. The solids formed are dried to generally less
than 1 percent free moisture. The flue gas, which has usually been
humidified to within 11 to 28°C (20 to 50°F) of its adiabatic saturation
temperature, passes through the dryer and into a downstream high efficiency
particulate matter control device. In some designs a portion of the solids
drop out of the dryer, but the bulk of the desulfurization products are
collected with fly ash in a fabric filter or an ESP. The reaction between
the alkaline material and flue gas SO continues as the gas passes through
the ductwork and the fabric filter or ESP.
Table 1 lists the major design and operating factors for spray drying
FGD. These considerations are discussed in more detail below.
Using a spray dryer as a flue gas contactor involves adiabatically
humidifying the flue gas to within a specified number of degrees above
saturation. With set conditions for inlet flue gas temperature and humidity
and for a specified approach to saturation temperature, the amount of water
which can be evaporated into the flue gas is set by energy balance consid-
erations. Liquid-to-gas ratios are generally in the range of 0.27 to
10-2
-------
o
I
(-0
_
BOILER
COMBUSTION />
,
i
^ FLUE G/i
IR
HOT
s
HOT OR WAR
I
/ \
/ \ 1
H
PREHEATER
Ai
M G/
T.
i
R
CLEAN GAS TO
^s BYPASS ATMOSPHERE
'ARM / \
1 (~l FAN GA Q / \
1 1 /? ^-^v '^ / \
1 ff n / \
t I 1 \~^J 1 '
I 1 r
/K ' I~AN STACK
SPRAY ._
DRYER w
\ / * >r
\/ Fi HP RA-; BAGHOUSE
& SOLIDS OR ESP
SOLIDS
PARTIAL RECYCLE OF SOLIDS f
yL
\( *
r PRODUCT SOLIDS &
SoRBENT FLY ASH DISPOSAL
SLURRY
TANK
X
SORBENT STORAGE
Figure 1. Typical spray dryer/particulate collection flow diagr
am.
-------
TABLE 1. IMPORTANT DESIGN AND OPERATIONAL FACTORS
FOR SPRAY DRYING FGD SYSTEMS
Operational Factors
• Approach to saturation at dryer outlet
• Sorbent stoichiometry
• Inlet SO concentration
• Temperature drop over the spray dryer
• Fly ash alkalinity
• Gas residence time
Design Factors
• Spray dryer design (gas distribution, gas exit configuration)
• Atomizer type
• Slaker type
• Use of solids recycle
• Use of flue gas bypass
10-4
-------
o
0.40 liter/m (0.2 to 0.3 gal./kef). The sorbent stoichiometry is varied by
raising or lowering the concentration of a solution or weight percent solids
of a slurry containing this set amount of water. While holding other
parameters such as temperature constant, the obvious way to increase SO
removal is to increase sorbent stoichiometry. However, as sorbent stoichio-
metry is increased to raise the level of SO removal, two limiting factors
are approached:
• Sorbent utilization decreases, raising sorbent and disposal costs.
• An upper limit is reached for the solubility of the sorbent in the
solution, or for the weight percent of sorbent solids in a slurry.
There are at least two methods of circumventing these limitations. One
method is sorbent recycle, using the solids that have either dropped out in
the spray dryer or collected in the particulate emission control device.
Recycle increases sorbent utilization and can also increase utilization of
any alkalinity in the fly ash.
The second method of avoiding the above limitations on S0_ removal is
to operate the spray dryer at a lower outlet temperature; that is, a closer
approach to saturation. Operating the spray dryer at a closer approach to
saturation has the effect of increasing both the residence time of the
liquid droplets and the residual moisture level in the dried solids. As the
approach to saturation decreases, S0_ removal rates and sorbent utilization
generally increase dramatically.
The approach to saturation at the spray dryer outlet is set by either
the requirement for a margin of safety to avoid condensation in downstream
equipment or restrictions on stack temperature. The spray dryer outlet can
be operated at temperatures lower than these restrictions would otherwise
allow if some warm or hot gas is bypassed around the spray dryer to reheat
the dryer outlet gas. Warm gas (from downstream of the boiler air heater)
entails no energy penalty, but bypassing untreated gas for reheating can
limit overall S0? removal efficiencies. Significantly less hot gas
(upstream of the air heater) is required for heating, but using hot gas
decreases the energy available for air preheating. Figure 1 illustrates
these two reheat options.
Several similarities exist between the designs of the commercial spray
drying FGD systems sold to date. All of the systems except one use lime as
the sorbent; the exception uses commercial soda ash. In addition, all but
two of the purchased systems include a fabric filter for collection of fly
ash and waste products. Two utility systems will have an ESP instead of a
fabric filter. Reverse-air fabric filters have been selected for utility
systems, while pulse-jet units have been the choice for most industrial
systems.
Most commercial system designs use rotary atomizers, although nozzle
atomizers are used in some designs. Solids recycle and flue gas bypass for
limited reheat are site-specific design options. Recycle, however, is being
included more often than not, primarily because it improves sorbent
10-5
-------
utilization and lowers reagent costs. The use of recycle and flue gas
bypass depends primarily on the SO removal requirements.
Several differences between the commercialized designs of spray dryer
FGD systems are evident. The differences can be categorized into four major
areas:
• Spray dryer design and operation.
• Atomization.
• Sorbent preparation (lime slaking)-
• Fabric filter design and operation.
Variations in spray dryer design include the shape of the dryer, gas
disperser configurations, multiple or single rotary (or nozzle) atomizers
per dryer, one- versus two-point flue gas admission to the spray dryer, and
solids collection from the spray dryer. Each vendor claims his design to be
suitable for FGD and to have certain advantages. Some important
considerations in evaluating the various designs include:
• Liquid/gas contact.
• Turndown capability.
• Potential for plugging or carryover of wet solids to baghouse
during upset conditions.
Differences in spray dryer operation include gas residence time and the
approach to saturation at the dryer outlet. Gas residence times range from
7 to 12 seconds in most commercial designs; most operate in the 10- to
12-second range. Many commercial system designs call for a relatively close
approach to saturation at the dryer outlet [10 - 14°C (18 - 25°F)]. These
close approaches are common where S0_ removal efficiency requirements
approach 85 to 90 percent, but they may also be used to decrease reagent use
in lower removal efficiency applications. Some designs, however, are based
on a wider approach to saturation [17 - 28°C (30 - 50°F)].
Atomization techniques vary with regard to the use of rotary or nozzle
atomizers, atomizer wheel speed in rotary atomizers, atomizing fluid (air or
steam) for nozzles, internal or external mixing of fluids in nozzles, and
the number of atomizers. Rotary atomizers may be used on a single or
multiple atomizer per dryer basis. Several commercial utility systems use
three atomizers per dryer, but most sold to date have a single atomizer per
dryer. Rotary atomizers are generally thought to have certain advantages
over nozzle atomizers. Specifically, unlike nozzle atomization, the size of
the droplets produced by rotary atomizers is independent of slurry flow
rate, providing for good turndown capability. Some plugging problems have
occurred in nozzles using steam for atomization. One potential advantage of
nozzle atomizers lies in the fact that they are mechanically simpler than
rotary atomizers.
10-6
-------
The method used to slake the lime also varies for commercial system
designs. Ball mill, paste, and detention slakers are all used, but ball
mill and paste slakers are more common. Each method is perceived to have
certain advantages. Ball mill slakers generally produce a more finely
ground and possibly more reactive slurry and, unlike paste slakers, no grit
removal and disposal are required. Paste slakers, on the other hand, have
lower capital costs, lower noise levels, and lower power requirements.
Also, they generally produce a less abrasive slurry than do ball mills,
which could reduce the potential for wear and maintenance of system pumps
and piping.
As stated above, fabric filters have been chosen over ESPs in nearly
all the commercial spray drying applications to date. Fabric filters may
have an advantage over ESPs in that unreacted alkalinity in the solids and
fly ash collected on the fabric surface can react with remaining SO in the
flue gas as the gas passes through the fabric filter. Some process
developers have reported that SCL removal across the fabric filter can
account for at least 10 percent of the total system SO removal. In some
applications, however, an ESP might be more desirable than a fabric filter.
The factors that are important in making the choice between ESPs and fabric
filters include:
• Use of recycle (increases dust loading to ESP, thus increasing ESP
size and cost).
• Fly ash resistivity (high ash resistivity often requires larger,
more expensive ESPs).
• Pressure drop (an ESP has a lower pressure drop and related costs
than a fabric filter).
Fabric filter designs for spray dryer FGD systems vary primarily with
regard to bag fabric, cleaning frequency, and cleaning mode. Other more
subtle variations, too numerous to mention, between the well-established
fabric filter design practices of the different vendors also exist.
Teflon-coated fiberglass bags have been selected for many systems, although
acrylic bags may be a lower cost alternative for low gas temperatures.
DRY INJECTION
A generalized flow diagram of the dry injection process is shown in
Figure 2. The process involves pneumatic injection of a dry, powdery sodium
compound into the flue gas with subsequent particulate collection in a
fabric filter. The point of alkali injection may vary from upstream of the
air preheater to the inlet of the fabric filter. Reaction between the
reagent and SO occurs both in the duct and on the filter bag surface (5,6).
Although other alkaline reagents (e.g., lime and limestone) have been
tested, only certain sodium compounds have shown the capability for high SO
removal from the flue gas. Nahcolite and trona ores, which contain
naturally occuring sodium compounds, appear to be the most promising
reagents for dry injection in terms of reactivity and cost. Nahcolite,
which is usually associated with western oil shale reserves, contains about
10-7
-------
CLEAN GAS TO
ATMOSPHERE
FLUE GAS
AIR PREHEATER
REAGENT FROM
BULK STORAGE
o
i
oo
REAGENT
HOLDiNG
BIN
FLUE GAS AND
REAGENT
STACK
REAGENT PULVERIZER
PRODUCT SOLIDS AND
FLY ASH DISPOSAL
INJECTION
FAN
Figure 2. Dry alkali injection flow diagram.
-------
80 percent sodium bicarbonate (NaHCO»). Trona ore, which is found in large
deposits in Wyoming and California, contains sodium carbonate (Na CO ) and
sodium sesquicarbonate (Na CO • NaHCO • 2H 0). Nahcolite has been shown to
be somewhat more reactive with SO in the flue gas than trona (5,6).
Major factors that affect SO removal by nahcolite and trona injection
include stoichiometric ratio and the flue gas temperature at the point of
injection. Higher S0» removals are obtained at higher normalized stoichio-
metric ratios (equivalent moles Na 0 per mole of inlet SO ) . However,
higher stoichiometric ratios also result in lower reagent utilization (5).
Injection of the reagent at too low a temperature will reduce the initial
S02 reaction rate and may limit overall SO removal. For nahcolite it
appears that SO removal drops off dramatically at injection temperatures
below 135°C (275°F) (6,7).
Other parameters that are important in the dry injection process
include reagent particle size, mode of injection (batch, semi-batch, or
continuous), baghouse air-to-cloth ratio, and bag cleaning frequency.
ELECTRON-BEAM PROCESS
The electron-beam (E-beam) process involves the irradiation of flue gas
containing a reactant, such as ammonia or lime. The process removes both
SO,, and NO from the flue gas and results in a dry waste product.
L. X
A schematic diagram of the E-bearn/ammonia process is shown in Figure 3.
In this process, flue gas from a fly ash collection device is cooled and
humidified in a quench tower. The resulting gas moisture content is about
10 percent. Ammonia is injected into the cooled gas and the gas is passed
through an E-beam reactor. In the reactor, oxygen and water are excited and
ionized to form the radicals [OH], [0], and [HO ] by the application of
electrons at a dose of 1 to 3 Mrads (1 Mrad is equivalent to 10 joules/g of
flue gas). These radicals react with SO- and NO to form sulfuric acid
(H?SO.) and nitric acid (HNO.,). The acids are neutralized by ammonia and
water in the flue gas to form solid ammonium sulfate (NH.SO.) and ammonium
sulfate nitrate [(NH.) • SO,* NH.NO ]. The reaction time for formation of
the sulfate and nitrate salts is less than 1 second. Product solids are
collected in a hopper below the E-beam reactor or in a downstream parti-
culate collector.
In another version of the E-beam process, the water quench tower is
replaced with a lime-based spray dryer. S0~ is removed from the flue gas in
the spray dryer and NO and additional SO,., are removed in the E-beam
reactor. Reactions inXthe E-beam reactor to form sulfuric and nitric acids
occur in the same manner as in the ammonia process. The acids are
neutralized by Ca(OH) to form calcium salts (CaSO , Ca(NO ) , and CaSO ).
Bench- and pilot-scale studies of E-beam processes have shown that
variables important to performance include gas moisture content, gas
temperature, oxygen content, reagent ratio, and electron dosage. In
addition, efficient penetration of the gas stream by the beam requires a
unique discharge pattern and other special design considerations.
10-9
-------
o
I
TT
QUENCH WATER
TT
T T T T T
- AMMONIA
FLUE
GAS
E-GUN
E-BEAM
REACTOR
PARTICU-
LATE
COLLECTOR,
PRODUCT SOLIDS
DRAIN
Figure 3. E-beam/anunonia process flow diagram.
-------
DEVELOPMENT STATUS
Of the three dry SO control technologies considered here, only spray
drying has been commercially applied. However, dry injection technology is
nearing commercialization with plans for a 500 MWe system. Research and
development activities are being conducted for all three technologies by
vendors, government agencies, and industry organizations.
SPRAY DRYING
Through October 1983 there have been 17 commercial utility systems
sales, totaling over 6,800 MWe. The general characteristics of the utility
systems, including vendor, size, location, design criteria, and projected
start-up date or current operational status, are shown in Table 2. The
commercial utility spray drying systems have been selected for units firing
relatively low sulfur western coal or lignite (generally less than
1.5 percent sulfur fuel). S0_ removal guarantees range from 60 to
91 percent, with design controlled emission levels ranging from 77 to
516 ng/J (0.18 to 1.2 lb/10 Btu). Some of the guarantees stem from the
need to meet the most recent New Source Performance Standards (NSPS) for
utility boilers (8).
Two commercial utility spray drying systems have been sold since
mid-1982, and a third system was awarded but is currently on hold. These
awards were all to Joy Industrial Equipment Company and Niro Atomizer, Inc.
The recently awarded utility systems are for units firing coal/lignite with
average sulfur contents of less than 1 percent.
According to the PEDCo survey prepared for EPA, 24 utilities
considering FGD for new coal-fired boilers have not yet made awards (9). Of
these utilities, two are considering only dry FGD, while nine utilities did
not specify whether they were considering only wet, only dry, or both.
Table 3 shows the utilities, the planned units, and reported coal sulfur
contents for which FGD is being considered. The data in Table 3 indicate
that dry FGD is not yet being widely considered for utility applications
involving moderate or high sulfur coal.
The six commercial utility spray drying systems now operational are
listed in Table 2. Of these the sodium-based system at Coyote Station was
turned over to Montana-Dakota Utilities for operation following performance
tests conducted in August 1982 (10) . Earlier test results reported for this
system (11) showed 70 percent S0? removal at a soda ash utilization of
75 percent and an approach to saturation of 36°C (65°F).
The 110 MWe unit at Northern States Power Company's Riverside Station,
initially a demonstration spray dryer coupled with existing ESPs and now
coupled to a fabric filter treating flue gas from two boilers, has also been
turned over to the utility. Process testing of this system was performed
during the summer of 1983 under EPA and EPRI funding. Results from these
tests are scheduled to be presented later at this symposium.
10-11
-------
TABLE 2. COMMERCIAL UTILITY SPRAY DRYING SYSTEMS
System Purchaser
Otter Tall Power Co.
(Operator: Montana-
Dakota Utilities)
United Power Assoc.
Marquette Board of
Light and Power
Basin Electric Power
Coop .
M
O
1 i Colorado Ute
K> Electric Assoc.
Basin Electric Power
Coop.
Basin Electric Power
Coop.
Station/Location
Coyote, Unit 1
(Beulah, ND)
Stanton, Unit 1A
(Stanton, ND)
Shiras, Unit 3
(Marquette, MI)
Laramie River,
Unit 3
(Wheatland, WY)
Craig, Unit 3
(Craig, CO)
Antelope Valley,
Unit 1
(Beulah, ND)
Antelope Valley,
Unit 2
(Beulah, ND)
Size3
MWe Start-up Date/Status
440 Operational. Turned
over to utility.
60 Operational. Not
turned over to
utility.
44 Operational. Not
turned over to
utility.
575 Operational. Not
turned over to
utility.
447 Late 1983. Under
construction.
440 Operational. Not
turned over to
utility.
440 April 1985.
Outlet SO
ng£J
Coal SO Removal Guarantee (lb/10 Btu)
North Dakota lignite; 70% for all coal.
0.78% S average.
North Dakota lignite; 91% for max. S coal.
0.77% S average;
1.94% S maximum.
Western subbitumi- 80% design efficiency.
nous; 1.5% S maximum.
Wyoming subbituminous 82% for avg. S coal;
0.54% S average; 90% for max. S coal;
0.81% S maximum.
Western subbituminous 87% for design coal.
0.7% S design coal.
North Dakota lignite; 62% for avg. S coal;
0.68% S average; 78% for max. S coal.
1.22% S maximum.
North Dakota lignite; 89% for max. S coal.
0.68% S average;
1.22% S maximum.
516
(1.2)
258
(0.6)
404
(0.94)
86
(0.20)
86
(0.20)
335
(0.78)
168
(0.39)
b ,, . c
Vendor
Rockwell/
Wheelabrator-Frye
Cottrell Environmental/
Komline-Sanderson
G.E. Environmental
Services
Babcock and Wilcox
Babcock and Wilcox
Joy/Niro
Joy/Niro
Tucson Electric Power
Tucson Electric Power
Platte River Power
Authority
Springerville,
Unit 1
(Springerville, AZ)
Springerville,
Unit 2
(Springerville, AZ)
Rawhide, Unit 1
(Fort Collins, CO)
370 June 1985.
370 June 1987.
280 December 1983.
Under construction.
New Mexico subbltu- 61%.
minous; 0.69% S
average.
New Mexico subbitu- 61%.
minous; 0.69% S
average.
Western subbituminous 80%.
0.3% S average;
0.44% S maximum.
258 Joy/Niro
(0.6)
258 Joy/Niro
(0.6)
77 Joy/Niro
(0.18)
(continued)
-------
TABLE 2. (CONTINUED)
o
i
System Purchaser
Sunflower Electric
Sierra Pacific Power
Grand River Dam
Authority-State
of Oklahoma
Northern States
Power Company
Cajun Electric
Central and South-
western Services
Northern States
Power Company
Pacific Power and
Light
Station/Location
Holcomb, Unit 1
(Holcomb, K.S)
North Valmy
(Valmy, NV)
Pryor, Unit 2
(Pryor, OK)
Sherburne County
Unit 3
(Becker, UN)
Oxbow, Unit 1
(Coushatta, LA)
Coleto Creek
Unit 2
(Corpus Christ! , TX)
Riverside, Units 6
and 7 (Minneapolis,
MN)
Jim Bridger, Unit 2,
(Rock Springs, WY)
Size3
MWe Start-up Date/Status
319 In start-up
270 1984.
520 March 1985.
860 1990.
563 Projected for 1989
but need for system
being re-evaluated
710 1988.
110 Operational. Turned
over to utility.
100 Shut down. Demon-
stration testing by
vendor completed.
No plans for restart.
Coal SO- Removal Guarantee
Wyoming subbituminous; 80%.
0.34% S average.
Western subbituminous; 76% for all coal.
0.4 to 1.0% S.
Western subbituminous; 85% for all coal.
0.4 to 1.5% S.
Sarpy Creek subbitu- 70% minimum.
minous; 0.9% S average;
0.4 - 2.3% S range.
Louisiana lignite; 80%.
0.93% S average.
Subbituminous; 0.4% S 70%.
average.
Varies with test Varies with test
series. program.
Western subbituminous; Lime: 62-75%
0.56% S average; Soda ash: 74-86%
0.8% S max imum .
Outlet SO,
ng/J .
(lb/10 Btu) Vendor0
116 Joy/Niro
(0.27)
258 Rockwell
(0.6)
NR Flakt
258 Joy/Niro
(0.6)
258 Joy/Niro
(0.6)
NR Joy/Niro
516 Joy/Niro
(1.2)
129 Flakt
(0.3)
Gross MWe output.
bSource for SO outlet emissions: Ireland, P. A. (Sterns-Roger) Status of Spray Dryer Flue Gas Desulfurization. EPRI CS-2209. Final Report. EPRI,
Palo Alto, CA?, January 1982, p. 2-3.
CRockwell/Wheelabrator-Frye is no longer a joint venture. Both are, however, selling spray drying systems. Joy/Niro: Joy takes the lead in utility sales
while Niro takes the lead in industrial sales. G.E. Environmental Services was formerly Buell Emission Control Division, Envirotech Corp. Spray dryer
supplied by Anhydro A/S and fabric filter supplied by G.E.E.S. Cottrell Environmental Sciences supplies fabric filter and Komline-Sanderson supplies
spray dryer.
-------
TABLE 3. UTILITIES PLANNING OR CONSIDERING FGD SYSTEM THAT
HAVE NOT AWARDED BIDSS (as of August 3, 1983)
Type of FGD
System Projected Unit
Utility Considered Coal (%S) Start-Up
Atlantic City Electric
Buckeye Power
Central Illinois Light
Central Maine Power
Cincinnati G&E
Colorado Ute Electric
Delmarva Power & Light
Desert Gen & Trans
Florida Power & Light
General Public Utilities
Indianapolis Power & Light
Kentucky Utilities
Louisville G&E
Nebraska Public Power
Nevada Power
Orlando Utilities Commission
Pacific Power & Light
Salt River Project
Seminole Electric
Southwestern Electric Power
Texas Utilities
WA Water & Power
West Texas Utilities
Wet
Wet
Wet
Wet
NR
NR
Wet
Wet
NR
NR
Wet
Wet
Wet
Dry
NR
Wet
Wet (retrofit)
Dry
Wet
NR
NR
NR
Wet
NR
3. £5
NR
3.3
2.23
4.0
0.5
2.5
0.5
NR
3.5
3.5
3.5
4.0
0.36
NR
NR
0.56
NR
0.6
NR
1.5
0.8 (lignite)
NR
0.34
1988
1990
1989
1989
1988
1988
1987
1992
2 units;
1992, 1993
2 units;
1993, 1994
3 units;
1990
1994
2 units;
1986, 1990
1987
4 units;
1984-87
1987
3 units;
1986-90
1986
1989
2 units;
1988, 1990
2 units;
1990
1989
4 units;
1990-94
1996
^Source: Reference 9.
NR - Not reported.
10-14
-------
Three of the other four operational spray drying systems (Stanton,
Shiras, and Antelope Valley) have all recently met performance guarantees.
However, no test results have been reported yet for any of these systems.
One other utility system is in the initial start-up stage, and two
systems are scheduled for start-up before the end of 1983.
In addition to the utility systems, a number of industrial coal-fired
boilers are, or will be, equipped with spray drying FGD systems. Table 4
gives a summary of the 7 operational and 14 planned industrial spray drying
units. Coal sulfur contents for the industrial applications range from 0.6
to 3.5 percent, with seven systems designed for coal sulfur contents of
3 percent or greater. Removal guarantees for the industrial systems range
from 70 to 90 percent.
Performance test results have been reported for five of the operational
industrial systems. A continuous monitoring program conducted by EPA on the
Celanese system showed 70 percent average S0? removal over a 23-day test
period (12). Removal averaged 75 percent on days when the daily inlet S0?
concentrations were 1720 ng/J (4.0 lb/10 Btu) or greater. It should be
noted that the tests were conducted during a period of fluctuating boiler
operation.
Performance tests of the Strathmore Paper Company system showed an
average SCL removal of 92.4 percent (90.1 to 96.7 percent range) at an inlet
SO concentration of about 2000 ppm (13). Data on lime stoichiometric ratio
or other operating conditions required to achieve these removals were not
included in the performance test results. This system and the Celanese
unit, which do not employ recycling of solids, have now been turned over to
the system purchasers.
Compliance tests conducted on the Argonne system showed SO removals
ranging from 79.7 to 95.6 percent at lime stoichiometries ranging from 0.8
to 2.0, respectively (14). These tests were conducted on a 3-percent sulfur
coal at an approach to saturation of approximately 13°C (23°F). In this
system, receiving flue gas from a stoker-fired boiler and described in a
presentation to be made later in this session, solids from the flue gas
cleaning system are recirculated to the spray dryer.
This system has now been turned over to the Argonne National
Laboratory, and EPA-funded process testing of the system is on-going.
Preliminary results from a 100-hour test run on a 4.6 percent sulfur coal
showed 90 percent SO removal at a lime stoichiometry of 1.4.
The Container Corporation and Austell Box Board systems have both met
compliance requirements, and the former system has been turned over to the
system purchaser. Compliance tests on the Container Corporation system
showed average outlet SO emissions of 14.3 ng/J (0.03 lb/10 Btu) on a
0.6 percent sulfur coal fl5). Tests on the Austell Box Board system (16)
showed an average outlet SO emission rate of 366 ng/J (0.85 lb/10 Btu).
No data on reagent ratios have been reported for these two systems. The
system at GM-Buick Motors and one unit at the University of Minnesota have
only recently become operational, and no performance data have yet been
released.
10-15
-------
TABLE 4. COMMERCIAL INDUSTRIAL BOILER SPRAY DRYING SYSTEMS
System
Purchaser
Celanese Fibers
Co.
Strathmore Paper
Co.
University of
Minnesota
Argonne National
Lab
Container
Corporation
GM-Buick Motors
Fairchild AFB
Austell
Box Board Co.
Puget Sound
Naval Shipyard
Maelstrom AFB
Griffis AFB
Location
Cumberland, MD
Woronco, MA
Minneapolis ,
MN
Argonne , IL
Philadelphia,
PA
Flint, MI
Units 1,2,3
Spokane , WA
Austell, GA
Units 1,2,3
Bremerton, WA
Units 1,2,3
Great Falls, MT
Units 1,2,3,4
Rome, NY
Size, Mg/hr
(Ib/hr)
Steam
50
(110,000)
39
(85,000)
Two units @
40 MW each
77
(170,000)
77
(170,000)
204
(450,000)
3 @ 50
(110,000)
114
(250,000)
3 @ 64
(140,000)
3 @ 41
(90,000)
hot water
4 (3 41
(90,000)
Start-up Date/Status
Operational. Turned over
to purchaser.
Operational. Turned over
to purchaser.
One unit operational.
Second start-up in
September 1983. Not turned
over to purchaser.
Operational. Turned over
to purchaser.
Operational. Turned over
to purchaser.
Operational. Not turned
over to purchaser.
Initial start-up
stages .
Operational. Not turned
over to purchaser.
Late 1987.
Spring 1985.
Late 1984
Coal
1.5% S to 2.5% S eastern
coals .
2.3 to 3% S eastern coal.
Subbituminous coal; 0.6
to 0.7% S.
Bituminous coal; 3.5% S.
1.0% S eastern coal.
Indiana bituminous coal;
1 to 3% S.
1.0% S western coal.
Bituminous coal; 1.0
to 2.5% S.
1.6% S maximum.
Western subbitiuminous
coal; 1.0% S.
Bituminous; 3% S
S02 Removal
Guarantee
70% for 1.5% S coal.
87% for 2.5% S coal.
75%.
70%.
78.8% (516 ng/J or
1.2 lb/10 Btu outlet
so2).
90%.
70 to 90%.
85%.
516 ng SO /J (1.2 Ib
S02/10 Ecu outlet).
84%.
85%.
85% (0,71 Ib
S02/10 Btu outlet).
Vendor
Rockwell/
Wheelabrator-Frye
Mlkropul
Flakt
Niro/Joy
Ecolaire
Niro/Joy
Niro/Joy
Wheelabrator-Frye
G.E. Environmental
Services
Niro/Joy
Ecolaire
Rockwell/Wheelabrator-Frye is no longer a joint
industrial system sales, and Joy takes the lead
venture. Both are, however, selling spray drying systems. Niro of Joy/Niro takes the lead in
in utility system sales.
-------
In addition to the commercial sales, there are a number of spray drying
research and development activities currently underway. Table 5 lists the
major current activities. These programs, which are being funded by EPA,
EPRI, and DOE, are designed to address information needs in the areas of
high sulfur coal applications, alternate spray drying sorbents, and
sustained operation at a close approach to saturation. Results from several
of the programs are scheduled to be presented at this symposium. On-going
research by several vendors stresses the further improvement of atomization
techniques with goals of longer service life, lower energy consumption, and
improved slurry droplet size distribution. Rotary and nozzle atomizers are
being compared on the basis of cost and performance.
DRY INJECTION
The first planned commercial application of dry injection technology
has been announced by Public Service Company of Colorado for a 500 MWe unit
scheduled for start-up in 1990 (17). The system, which will use trona ore
as the sorbent, will be designed for 70 percent SO removal on a 0.4 percent
sulfur western coal. The trona will be injected just upstream of the fabric
filter into flue gas at approximately 132 to 138°C (270 to 280°F). A clay-
and plastic-lined landfill will be used for solids disposal.
The 22 MWe demonstration testing at Public Service of Colorado's Cameo
Station has been completed. Final reports on the EPRI-sponsored dry
injection program have been prepared (18,19). Results of the testing showed
70 percent SO- removals for trona and nahcolite at stoichiometric ratios of
approximately 1.3 and 0.8, respectively. These tests were conducted on a
low sulfur western coal yielding flue gas at 450 ppm S0_.
A recent cost study based on the Cameo results indicates that dry
injection may be less expensive than spray drying, depending on reagent
cost, coal sulfur content, and S0_ removal requirements (7). Estimated
costs for nahcolite and trona dry injection systems and a lime spray drying
system are shown in Table 6 for two-500 MWe units firing 0.48 percent sulfur
coal with 70 percent S0_ removal. These estimates assume nahcolite
(70 percent NaHCO ) and trona (85 percent Na2C03) costs of $100 and $75 per
ton, respectively, and sodium-based waste disposal costs of $7.40/ton.
The application of trona dry injection had been previously constrained
by questions regarding S09 removal limitations and cost. However, the
recent demonstration-scale studies have shown that S02 removal efficiencies
of 70 to 80 percent can be achieved with trona on low sulfur coals at
reasonable stoichiometric ratios. Trona ore is currently mined in large
quantities for conversion to sodium carbonate.
The commercial application of nahcolite dry injection hinges on the ,
economic availability of nahcolite. Nahcolite is not currently mined in the
United States, but at least one firm has announced intentions to develop a
nahcolite mining operation and several other companies are investigating the
possibility of supplying nahcolite through solution mining techniques (20,21)
10-17
-------
TABLE 5. MAJOR CURRENT SPRAY DRYING R&D ACTIVITIES
o
h-'
00
Vendor, Agency, or
Industry Group
Electric Power Research
Institute (EPRI)3
Size
255 m /min
(9000 cfm)
Location
Arapahoe pilot plant;
Denver, CO.
Comments
Test program began in March 1982.
Testing planned through December
1983.
EPRI/EPAC
EPA
100 MWe
77 Mg/hr
(170,000 Ib/hr)
steam
Riverside Station,
Northern States Power.
Minneapolis, MN.
Argonne National Lab.
Argonne, IL.
Ten weeks of testing beginning in
July 1983, including 3.4% S coal.
Performance test (90-day), start-
ing in November 1983, on 3.4% S
coal.
DOE/Babcock and
•3
Wilcox
25 m /min
(900 scfm)
B&W Alliance
Laboratory.
Alliance, OH.
Two-year test program began in
mid-1982 on eastern high sulfur
coals.
Papers on these studies will be presented at this symposium.
Actual flue gas conditions.
-------
TABLE 6. COMPARISON OF COSTS FOR NAHCOLITE AND TRONA DRY INJECTION
AND LIME SPRAY DRYING ON TOO NEW 500-MW COAL-FIRED
POWER-GENERATING UNITS
FGD Process
Capital
Cost ($/kW)
Levelized Costs
(mills/kWh)
Trona injection
89.1
Nahcolite injection
90.4
9.6
Lime spray drying
175.9
10.!
Notes: All costs include baghouse for particulate control. Upper midwest
plant firing 0.48 percent sulfur coal at 8,020 Btu/lb. 1983
revenue requirements,
Trona system: 70 percent S0? removal, $75/ton for trona
(85 percent Na0CO ), $7.40/ton for waste disposal.
Nahcolite system: 70 percent SO.., removal, $100/ton for nahcolite
(70 percent NaHCO ), $7.40/ton for waste disposal.
Lime system: 70 percent S00 removal, $60/ton for lime, $4.60 for
waste disposal.
Source:
Reference /.
10-19
-------
Another important consideration for dry injection applications is the
high solubility and leachability of the sodium-based wastes. Application of
dry injection may be limited in some locations by the costs of disposing of
the wastes in an environmentally safe manner.
ELECTRON-BEAM PROCESS
The electron-beam process is in an early developmental state. The DOE
has signed cost-sharing agreements with Research-Cottrell and EBARA/Avco-
Everett to conduct pilot-scale demonstrations of E-Beam processes.
Research-Cottrell will be developing the E-beam/lime process and EBARA/Avco-
Everett will be developing the E-bearn/ammonia process.
Testing of the E-beam/lime process is scheduled to begin in late 1983
on a 283 to 425 m /min (10,000 to 15,000 acfm) pilot plant at TVA's Shawnee
power plant. The tests will be conducted on flue gas from firing a 2.5 to
3 percent sulfur coal (22). Test plans for a 570 Nm /min (20,000 scfm)
pilot E-bearn/ammonia system are being finalized with plans calling for
construction to begin in October 1983 (23) - This unit will also be tested
on flue gas from firing a 2.5 to 3 percent sulfur coal.
SUMMARY
Currently over 6,800 MWe of electrical utility generating capacity has
been committed to spray drying SCL control. Utilities are operating two
systems (550 MWe), and four systems (1,120 MWe) are operational but not yet
turned over for utility operation. Recent sales of industrial spray drying
have been mainly to Federal government installations, with the industrial
systems now sold totaling 21 units. Seven of these units are operational,
with four units now being operated by their owners. While spray drying
currently remains as the only commercially applied dry FGD technology, plans
for a new 500 MWe electrical generating unit specify the dry injection of
trona for S0» control at a Colorado site.
Research and development (R&D) on spray drying and the E-beam process
is continuing. Both EPA and EPRI have active spray drying R&D programs.
DOE is supporting the E-beam studies which are currently at the pilot level.
REFERENCES
1. Blythe, G. M., J. C. Dickerman, and M. E. Kelly (Radian Corporation).
Survey of Dry S02 Control Systems. EPA-600/7-80-030 (NTIS PB 80-
166853), U.S. Environmental Protection Agency, Industrial Environmental
Research Laboratory, Research Triangle Park, N.C., February 1980.
2. Kelly, M. E. and S. A. Shareef (Radian Corporation). Second Survey of
Dry^S02 Control Systems. EPA-600/7-81-018 (NTIS PB 81-157919), U.S.
Environmental Protection Agency, Industrial Environmental Research
Laboratory, Research Triangle Park, N.C., February 1981.
3. Kelly, M. E. and S. A. Shareef (Radian Corporation). Third Survey of
Dry^S02 Control Systems. EPA-600/7-81-097 (NTIS PB 81-218976), U.S.
Environmental Protection Agency, Industrial Environmental Research
Laboratory, Research Triangle Park, N.C., June 1981.
10-20
-------
4. Kelly, M. E. and M. A. Palazzolo (Radian Corporation). Status of Dry
SO Control Systems: Fall 1982. EPA-600/7-83-041 (NTIS PB 83-247585),
U.S. Environmental Protection Agency, Industrial Environmental Research
Laboratory, Research Triangle Park, N.C., August 1983.
5. Apple, C. and M. E. Kelly (Radian Corporation). Mechanisms of Dry SO
Control Processes. EPA-600/7-82-026 (NTIS PB 82-196924), U.S.
Environmental Protection Agency, Industrial Environmental Research
Laboratory, Research Triangle Park, N.C., April 1982, pp. 97-98.
6. Muzio, L. J., et al. (KVB, Inc. and Electric Power Research Institute).
22-MW Demonstration of Dry S0? Scrubbing With Sodium Sorbent Compounds.
Paper #83-38.5, presented at the 76th Annual Meeting of the Air
Pollution Control Association, Atlanta, Georgia, June 1983.
7. Naulty, D. J., et al. (Stearns-Roger Engineering). Economics of Dry
FGD by Dry Sorbent Injection. Paper #83-38.6, presented at the 76th
Annual Meeting of the Air Pollution Control Association, Atlanta,
Georgia, June 1983.
8. U.S. Environmental Protection Agency. New Stationary Sources
Performance Standards; Electric Utility Steam Generating Units. In:
Federal Register, Vol. 44, No, 113, June 11, 1979, pp. 33580-33624.
9. EPA Utility FGD Survey; On-line Information System. U.S. Environ-
mental Protection Agency, Industrial Environmental Research Laboratory,
Research Triangle Park, N.C., August 3, 1983.
10. Gehri, Dennis (Rockwell International). Telephone conversation with
M. A. Palazzolo, Radian Corporation, August 3, 1983.
11. Lewis, M. F. and D. C. Gehri (Montana-Dakota Utilities and Rockwell
International). Atomization - The Key to Dry Scrubbing at the
Coyote Station. In Proceedings: Symposium on Flue Gas Desulfuri-
zation - Volume 2, Ayer, F. A. (ed.). Electric Power Research
Institute, Palo Alto, CA, March 1983, pp. 673-688.
12. Mostardi-Platt Associates, Inc. Particulate and Gaseous Emission Study
Performance for Strathmore Paper Company at the Woronco Mill Power
Plant, Woronco, MA. Bensenville, IL, May 7, 1981.
13. Reinauer, T. V., et al. (Mikro Pul Corporation). Dry FGD on an
Industrial Boiler. Chemical Engineering Progress, _79_(4) , March 1983.
14. Farber, P. S. (Argonne National Laboratory). Startup and Performance
of a High Sulfur Dry Scrubber System. Paper #82-40.5, presented at the
75th Annual Meeting of the Air Pollution Control Association,
June 1982.
15. Air Techniques, Inc. Compliance Particulate, SO,., and NO Emission
Testing on Coal-Fired Boiler at Anstell Box Board Corporation, Austell,
GA. Marietta, GA, April 1983.
10-21
-------
16. Glazer, Norman (City of Philadelphia Air Management). Letter and
attachment to M. A. Palazzolo, Radian Corporation, August 19, 1983.
17. Green, George (Public Service Company of Colorado). Telephone
conversation with M. A. Palazzolo, Radian Corporation, July 27, 1983.
18. Muzio, Larry (KVB, Inc.). Telephone conversation with M. A. Palazzolo,
Radian Corporation, August 2, 1983.
19. Muzio, L. J., et al. Dry S0?-Particulate Removal for Coal-Fired
Boilers - Volume 1; Demonstration of SO,., Removal on a 22-MW Coal-
Fired Utility Boiler by Dry Injection of Nahcolite. EPRI CS-2894,
Electric Power Research Institute, Palo Alto, CA, March 1983.
20. Shah, N. D. (Multi-Mineral Corporation). Dry Scrubbing of SO .
Chemical Engineering Progress, ^7_8_(6), June 1982.
21. Solution Mining of Nahcolite (Natural Sodium Bicarbonate) May Begin
by 1986. Chemical Engineering, _90_(12):20, June 13, 1983.
22. Williams, John (Department of Energy, Pittsburgh, PA). Telephone
conversation with M. A. Palazzolo, Radian Corporation, August 3, 1983.
23. Frank, Norman (EBARA). Telephone conversation with M. A. Palazzolo,
Radian Corporation, August 15, 1983.
10-22
-------
ACID RAIN PREVENTION THRU NEW SOX/NOX DRY
SCRUBBING PROCESS
K. S. Felsvang, P. Morsing, P. L. Veltman
-------
ACID RAIN PREVENTION THRU NEW SOx/NOy DRY SCRUBBING PROCESS
by: Karsten Felsvang and Per Morsing
A/S NIRO ATOMIZER, Soeborg,
Denmark
Preston Veltman
Niro Atomizer Inc., Columbia,
Maryland, U.S.A.
ABSTRACT
The issue of acid rain has attracted much attention both in the U.S. and in Europe.
To cope with problems associated with acid rain Niro Atomizer is developing a dry
scrubbing process for simultaneous removal of SOX and NOX.
A description of the process is given in this paper. The equipment used is
essentially the same as used in the over 5000 MW utility dry scrubbers currently in
operation, start-up or under construction.
Pilot plant test results from the Copenhagen dry FGD facility and results achieved
during a full-scale demonstration of the process at one of JOY/NIRO's operating dry
scrubbers are presented.
Waste product characteristics are shown and compared with EPA standards.
Finally, the process economics are analyzed and compared with other existing
processes for SOX/NOX removal.
INTRODUCTION
In recent years the issue of acid rain has attracted much attention both in the
United States and in Europe.
In the U.S.A. signs of acidification of the environment have become obvious in the
North-East. Discussions focus on how big a reduction of SOX and NOX should be required
(1). The U.S. EPA is evaluating which technique or combination of techniques would be
the overall most cost effective in meeting the possible requirements (2, 3).
In Europe the 1982 Stockholm conference "Acidification Today and Tomorrow" (4)
likewise concluded that dispersion of SOX and NOX arising from combustion of fossil
fuels has a dramatic impact on the environment. Early this year a 30% reduction of SOX
10-23
-------
was suggested by Scandinavian countries, and recently the German conference "Acid
Deposition - A Challenge for Europe" (5) again focused on the effects of acid deposition
and abatement technologies available.
A number of techniques has been developed for reduction of SOX and NOX
separately. Recent developments on low NOX burner technology look promising. Con-
siderable progress has been made in this area and achievable emission levels are
approaching the 0.1 - 0.2 Ib/MBtu range (6).
To encourage development of processes for simultaneous removal of SOX and NOX
the U.S. Department of Energy and the U.S. EPA have funded a number of research
programs that should lead to more cost effective methods for SOX/NOX control (7).
Dry scrubbing has in the recent years gained acceptance as an attractive FGD
technique for many utility companies. Niro Atomizer has developed a new dry scrubbing
process for simultaneous removal of SOX and NOX (8). The results of this development
are presented in this paper.
DESCRIPTION OF NEW SOX/NOX PROCESS
The new dry scrubber process for simultaneous SOX and NOX removal (9) utilizes
the same equipment as currently supplied for the JOY/NIRO dry FGD system. Normal
operating conditions in these systems are low approach temperatures such as 18°F to
achieve the lowest lime consumption and the highest SOX scrubbing efficiency. Under
these conditions virtually no NOX absorption is achieved.
The new SOX/NOX process is schematically shown in Figure 1.
FLUE GAS
DRYER j
Tout
195*F
(
v_>
(3-1
U V
FEED TANK
Figure 1. Niro Atomizer Dry Scrubber SOX/NOX Process
10-24
-------
A key feature in the new process is the existence of a so-called thermal window
where simultaneous SOX and NOX removal can be obtained. The process is operated with
an outlet temperature of the spray dryer absorber of 195 - 215°F. As in the normal dry
scrubbing process lime is the main reagent. The process is operated in a recycle mode,
i.e. recirculated dry waste product and slaked lime are mixed in the feed preparation
system to a slurry with a solids content of 30-40%, before being atomized in the spray
dryer absorber module. An inexpensive additive is required to make the process work in
the mentioned thermal window. An additive, such as sodium hydroxide, is simply added
into the feed preparation system in a quantity of approximately 10% of the lime added
In the Niro Atomizer spray dryer absorber (SDA) the efficient atomization and intimate
mixing with the flue gas create a uniform recycle product which under close control is
impregnated with a minimal amount of sodium sulfite formed by the selective reaction
of sodium hydroxide with SO2. Hereby a effective solid sorbent is created for the
simultaneous removal of SOX and NOX. A major portion of the SOX is removed in the
SDA, but only minor amounts of NOX is absorbed. The conditioned flue gas and activated
particulates then enter the fabric filter where simultaneous removal of NOY and SOY is
accomplished.
RESULTS OF BENCH SCALE AND PILOT PLANT INVESTIGATIONS
BENCH SCALE INVESTIGATIONS
As NO has very little affinity to water the SOX/NOX research program was
directed towards developing active solid materials that would promote an SOX/NOX
gas/solids reaction.
Bench scale tests were performed using a fixed bed reactor with a layer of reactive
powder through which the flue gas was passed (Figure 2). The powder used was the
normal dry scrubber waste solids including calcium sulfite, calcium sulfate and calcium
hydroxide. The thickness of the powder layer and the gas velocity through the bed was
adjusted to simulate as closely as possible the conditions in a baghouse filtercake.
REACTOR
TEMPERATURE &
HUMIDITY CONDITIONING
TTTT
AIR N2 NO SO2
Figure 2. Bench Scale Set Up For SOX/NOX Tests
RECORDER
10-25
-------
Effect of Additive
Various additives including sodium sulfite, Fe++-compounds and Ethylene-
diaminetetra-acetic acid, disodium salt (EDTA) were used to impregnate the bed of
powder. Sodium sulfite was found, for technical and economical reasons, to be the
preferred reagent.
Tests with a soda ash based dry scrubber waste, which has the sodium sulfite in
situ, showed no substantial NOX absorption.
Effect of Reaction Temperature
Normal lime based dry scrubber powder was impregnated with small amounts of
sodium suifite and was subsequently exposed in the fixed bed reactor to various dry
scrubber operating temperatures. The results shown in Figure 3 clearly indicate the
existence of the so-called thermal window where SOX and NOX react simultaneously.
% Removal (1 h average)
90
80 •
70 •-
60 •-
50
40 ••
30
20
10
THERMAL
WINDOW
SO,
NO,
Legend:
Inlet SOX = 300 PPM
Inlet NOX- 300 PPM
Temp. °F
158 176
194
212
230
Figure 3. Influence of Reaction Temperature on SOX/NOX Reaction
Effect of SOY/NOV Ratio
The process is a simultaneous removal of SOX and NOX. If no SOX is present no NOX
removal is achieved. The ratio between SOX and NOX should preferably be 1 (one) or
greater. Table 1 shows the influence of SOX/NOX ratio.
10-26
-------
TABLE 1.
INFLUENCE OF SOX/NOX RATIO ON NOX REMOVAL
sox
PPM
0
100
300
NOX
PPM
300
300
300
SOX/NOX
RATIO
0
1/3
1
NOX
REMOVAL
(1st hour)
0
34
70
SOX
REMOVAL
(1st hour)
_
92
91
Effect of Other Parameters
Other parameters such as the flue gas oxygen content and the moisture content of
the powder also affect the SOX/NOX reaction. Table 2 & 3 show the results. Under
normal power plant operating conditions the flue gas will contain 5% ©2 or more, so no
negative effect on the SOX/NOX reaction is expected.
TABLE 2.
INFLUENCE OF FLUE GAS OXYGEN ON SOX/NOX REACTION
O2 %
FLUE GAS
0.15
0.80
1.5
5.0
NOX
(1st hour)
0
>47
56
70
SOX
(1st hour)
80
92
92
91
Table 3 shows that a certain residual moisture content has to be present in the
powder to achive a substantial SOX/NOX removal.
TABLE 3.
INFLUENCE OF POWDER MOISTURE ON SQXNQX REACTION
MOISTURE
MOIST POWDER
DRY POWDER
NOX
70
30
SOX
91
34
PILOT PLANT TEST RESULTS
Following the bench scale investigation a comprehensive pilot plant test program
was carried out at the Niro Atomizer in-house FGD facility in Copenhagen. This pilot
plant is routinely used to develop dry FGD system design parameters and provides very
accurate scale-up information.
10-27
-------
Table 4 (10) shows the correlation between pilot plant and full scale test results.
TABLE 4.
COMPARISON OF PILOT AND FULL SCALE (RIVERSIDE) TEST RESULTS***
PLANT
PILOT
RIVERSIDE
PILOT
RIVERSIDE
TEST
No.
6110-6115
20189
6120-6123
20186
OPERATIONAL PARAMETERS
TEMPERATURES (°F)
PPM
S02
700
710
768
758
SDA
IN
31*
315
3*3
3*0
SDA'
OUT
1**
1*3
1*6
1*6
BH**
OUT
136
1*2
1*0
1*6
T
Adsat.
18.5
18.7
18.2
18.7
CRT
Sec.
12.9
13.2
10.2
10.5
FEED
SOLIDS
(%)
33.5
31.1
32.0
32.6
AIR
TO
CLOTH
RATIO
1.25
1.15
1.58
1.**
CLEAN-
ING
CYCLE
1 hr
1 hr
1 hr
1 hr
RESULTS
% SO2
REMOVAL
TOTAL
96.3
96.8
97.7
96.*
SR
1.02
1.09
1.15
1.20
* SDA: Spray Dryer Absorber ** BH: Baghouse
*** Pilot Plant: 5,000 ACFM - Riverside Full-Scale Plant: 500,000 ACFM
The existence of the thermal window was confirmed during the pilot tests. Table 5
shows results from parametric tests that were not performed under optimum conditions
with respect to other parameters.
TABLE 5.
PILOT PLANT DEMONSTRATION OF THERMAL WINDOW IN SOX/NOX PROCESS
SPRAY DRYER ABSO
INLET
TEMP.
oF
330
sox
PPM
1500
NOX
PPM
500
RBER
OUTLET
TEMP.
oF
151
194
214
230
266
% NOX REMOVAL
SDA
0
6
8
5
5
SDA+BH
0
21
35
37
30
% SOX REMOVAL
SDA
72
54
35
25
15
SDA+BH
100
83
67
55
34
10-28
-------
These tests with the spray dryer absorber / baghouse combination aiso reveal that
the major part of the SOX/NOX reaction takes place in the filter cake on the baghouse
bags. Only minor amounts of NOX are removed in the spray dryer absorber.
Pilot plant results confirmed the influence of the ratio between SOX and NOX
removal. Table 6 shows typical results.
TABLE 6.
EFFECT OF SOX/NOX RATIO
SO2 LEVEL
INLET SDA
PPM
NOX LEVEL
INLET SDA
PPM
% NOX
REMOVAL
1000
300
40
1600
300
50
2500
300
65
As can be seen from the table the highest NOX removal efficiency is - at the
present state of development - achieved with higher sulfur coals . Pilot plant test results
for typical high sulfur conditions are shown in Table 7. These tests were performed with
addition of sodium hydroxide to the feed tank. A quantity of 10 weight percent in
relation to the lime flow was used.
TABLE 7.
PILOT PLANT RESULTS WITH SOX/NOX PROCESS APPLIED TO HIGH SULFUR COAL
% Solids in Feed Slurry
T . + Temp. °F
Inlet v
SOX PPM dry
SDA NOX PPM dry
Inlet
BH TemP' °F
% NOX Removal
% SOX Removal
Lime Stoichiometry
TEST
No. 12138
34
371
2814
310
188
65.5
98.8
1.44
TEST
No. 12149
30
336
2762
309
222
62.5
86.2
1.10
TEST
No. 12150
36
334
2638
346
204
57.2
93.3
1.20
10-29
-------
REACTION MECHANISM OF NEW SOX/NOX PROCESS
Based on the absorption results and chemical analyses of the SOX/NOX dry scrubber
product, the reaction mechanism is illustrated in Figure 4.
SPRAY DRYER ABSORBER
SOX NO,
SO, NOX
H2O Monolayers
LIQUID DROPLET
HIGHLY REACTIVE
SOLID PARTICULATE
FABRIC FILTER
SOX +• O2
NOX + O2
H2O
-a(N03)2
Catalysis
at active site
SOLID PARTICULATE
Figure 4. SOX/NOX REACTION MECHANISM
The added sodium hydroxide is converted to water soluble sodium sulfite during
the absorption process in the spray dryer absorber. The sodium sulfite is partly
precipitated as amorphous solids and partly kept in solution in the residual moisture, thus
impregnating the spray dried solid particle of recycle solids and calcium hydroxide
reaction products. Figure 5 shows a scanning electron microscope (SEM) photograph of
such a particle.
Figure 5. SEM-PHOTO OF ACTIVE SITES
10-30
-------
The solid participate with the active sites enters the fabric filter and forms a
highly porous filter cake. According to Klingspor (11) between one and two monolayers of
water will still exist on the solids surface in the thermal window where the process is
operated. No NOX absorption could be accomplished if too many monolayers were
present. If on the other hand no water was present virtually no SC>2 absorption could be
achieved.
At the active site SOX and NOX are oxidized producing calcium sulfate, calcium
nitrate and calcium hydrogen sulfate, nitrate.
The proposed reaction mechanism is based on the following observations:
Oxygen and water are necessary to obtain NOX removal.
The amount of NOX absorbed agrees with the amount of nitrate found in the
reaction product.
The amount of sulfite in the reaction product is increased when compared
with normal SOX dry scrubber products.
RESULTS OF FULL-SCALE DEMONSTRATIONS
ARGONNE DEMONSTRATION
Since November 1981, i.e. for 2 years, Argonne National Laboratory has operated
the first commercial JOY/NIRO dry scrubber designed for high (3.5%) sulfur coal (12,
13). The scrubber treats flue gas from a Spreader-Stoker boiler rated at 170.000 Ib/hr of
steam. The coal is from the Illinois basin. A typical analysis is shown in Table 8. System
characteristics with respect to normal operation of the spray dryer absorber and the
baghouse are given in Table 9.
TABLE 8
COAL CHARACTERISTICS FOR ARGONNE HIGH SULFUR COAL
COAL PARAMETER
Heating value Btu/lb
Moisture %
Ash %
Carbon %
Hydrogen %
Nitrogen %
Sulfur %
Chlorine %
Oxygen %
VALUE
12,027
9.59
7.40
66.98
4.65
1.48
3.32
0.06
6.52
10-31
-------
TABLE 9
ARGONNE DRY SCRUBBER SYSTEM CHARACTERISTICS
SPRAY DRYER ABSORBER
Niro Atomizer 2^.5 ft diameter spray dryer
with compound gas disperser
35% feed slurry
20°F approach to saturation in exit gas
80 % SC>2 removal with lime stoichiometries of 1.0
BAGHOUSE
JOY four compartment pulse flow baghouse
12 ft long woven fiber glass bags
Air/cloth ratio 3:1
A P: 3 inch W.G.
Argonne National Laboratories permitted Niro Atomizer to use their dry scrubbing
system for a two week demonstration of the new SOX/NOX process. The demonstration
was accomplished in April 1983 just before the boiler was brought down for its annual
overhaul.
For the demonstration period Niro Atomizer brought in a continuous NOx-monitor
consecutively taking samples from spray dryer inlet, spray dryer outlet and baghouse
outlet. During the tests back-up wet sampling of NOX and SOX were was Additive sodium
hydroxide solution was stored in an intermediate 4000 gallon tank. A small centrifugal
pump, a flow meter and some piping were all what were required to provide the
capability of a controlled addition of sodium hydroxide to the recycle mixing tank.
Some mechanical problems caused by the handling characteristics of the powder
during the test period showed the minor changes that have to be done on full-scale
installations operating in SOX/NOX mode.
The test period was commenced with a test under normal operating conditions for
the dry scrubber. No sodium hydroxide was added. This test demonstrated that under
normal dry scrubber operating conditions virtually no NOX is removed in the system.
Then the system was brought into the SOX/NOX scrubbing mode by increasing the
spray dryer outlet temperature from l^op to 195OF and starting sodium hydroxide
addition to the recycle mixing tank. The result was very impressive and encouraging in as
10-32
-------
much as substantial NOX removal was observed. Table 10 shows the overall performance
with and without NOX removal. In the SOX/NOX operating mode more than 50% NOX
removal was demonstrated while the scrubber was still removing 90%+ SC^. The
demonstration shows full agreement between pilot plant data and full-scale scrubber
performance.
TABLE 10
DRY SCRUBBER SOX/NOX PERFORMANCE IN FULL-SCALE
MODE
OF
OPERATION
NORMAL 5O2
REMOVAL ONLY
50X/NOX
MODE
SCRUBBER
INLET
PPM
S02
1800
1500
PPM
NOX
290
280
TEMP.
°F
335
330
GAS-
FLOW
Ib/h
200,000
1*0,000
OUTLET
TEMP.
°F
1*8
195
BACFILTER OUTLET
PPM
S02
90
60
PPM
NOX
281
126
RESULTS
% REMOVAL
SOX
95
95+
NOX
3
55
STOICHIOMETRY
FOR
LIME
l.lf
1.3
RIVERSIDE DEMONSTRATION
A 3 week test period is planned at JOY/NIRO's Riverside demonstration plant this
fall. These tests will furnish additional full-scale data for process optimization. Results
will be published as soon as available.
CHARACTERIZATION OF SOX/NOX DRY SCRUBBER WASTE MATERIAL
The waste product from dry scrubbing systems has been thoroughly analyzed and
characterized over the last 3-4 years (14, 15, 16, 17, 18 and 19). These studies have
shown that the dry scrubber waste product is an excellent material for landfill due to its
pozzolanic properties and low permeability. The leachate quality can meet EPA limits
and therefore no liners would be required in properly designed landfills.
The waste product from the SOX/NOX process has the same appearance as other
dry scrubber poducts, i.e. a fine-grained, dry free-flowing powder.
The chemical composition is of course somewhat different. Table 11 shows the
chemical analyses of the waste product from the Argonne tests with and without NOX
removal. The Argonne product has low ash content as the boiler is Stoker fired and multi
clones are installed up-stream of the scrubber.
10-33
-------
TABLE 11
CHEMICAL COMPOSITION (wt%) OF WASTE PRODUCT
WITHOUT AND WITH NOX REMOVAL
Sulfite as CaSO3 • ^H2O
Sulfate as CaSO^ -2H2O
Nitrate as Ca(NO3)2
Excess alkalinity
as Ca(OH)2
Ash & Inerts
Moisture
Na2SO3
SOX
REMOVAL
59
16
-0
15
9
1
— 0
SOX/NOX
REMOVAL
16
43
3
21
8
1
9
As can be seen from the analyses, a major part of the sulfite has been oxidized to
sulfate when the scrubber operates in SOX/NOX mode. The absorbed NOX is present in
the product as nitrate. No nitrite has been found.
To evaluate the suitability of the waste product for landfill the following important
parameters were investigated:
Density
Compressive strength
Permeability
Leachate quality
Figure 6 shows the moisture-density curve for the two products. The SOX/NOX
product has a higher dry density and about the same optimum moisture content as the
normal dry FGD product.
10-34
-------
kg/m1 Dry Density
1000--
900-'
800
40
SOX/NOX
Removal
50
PureSOx
Removal
-I-
60
70 % Water Added
Figure 6. Moisture-density data of waste product from pure SOX and
combined SOx-NOx-removal
Table 12 shows a comparison of compressive strengths and permeabilities. The
strength development is the same as for the normal product, and the permeability is very
low for both products.
TABLE 12
COMPRESSIVE STRENGTH AND PERMEABILITY OF
PURE SOX AND COMBINED SOX/NOX PRODUCTS *
COMPRESSIVE STRENGTH (kPa)
1 DAY
7 DAYS
30 DAYS
56 DAYS
PERMEABILITY (ICT6 cm/s)
1-2 DAYS
28 - 30 DAYS
PURE SO2
PRODUCT
120
180
480
600
2.5
SOX/NOX
PRODUCT
110
200
340
580
6
-
All samples are compacted to 960-980 kg/m^ in dry density
10-35
-------
The ieachate quality obtained by the EPA Toxic Extraction Procedure is shown in
Table 13. All of the Ieachate values shown are less than 100 times the Primary Drinking
Water Standards (PDWS), therefore the product would most likely be classified as non-
hazardous.
TABLE 13
LEACHATE QUALITY OF WASTE PRODUCTS
FROM PURE SOX AND COMBINED SOX/NOX REMOVAL*
ARSENIC
BARIUM
CADMIUM
CHROMIUM
LEAD
MERCURY
NITRATE (as N)
SELENIUM
SILVER
EPA-
PDWR
mg/1
0.05
1.0
0.010
0.05
0.05
0.002
10
0.01
0.05
PURE S02
REMOVAL
mg/1
0.001
< 0.5
< 0.005
0.010
< 0.05
< 0.002
2
0.004
< 0.01
SOX/NOX
REMOVAL
mg/1
0.001
< 0.5
< 0.005
0.012
< 0.05
< 0.002
260
0.004
< 0.01
test carried out according to EPA toxic extraction procedure
The waste product from the Argonne tests contains little ash. The dry scrubber
product from a typical high sulfur application would have a fly ash content of 50 60 96.
This would change the characteristics in a positive direction.
It is expected that the SOX/NOX waste product will be regulated as non-hazardous.
However, due to somewhat elevated values for nitrate and total dissolved solids (TDS), it
may be necessary to install a liner before disposal.
Currently Niro Atomizer is working on methods to improve Ieachate quality and to
reduce Ieachate quantity.
COMPARISON OF ECONOMICS WITH STATE OF THE ART SOX/NOX PROCESSES
A cost comparison has been made between the Niro Atomizer dry SOX/NOX process
and a conventional limestone wet scrubber / selective catalytic de-NOx (SCR) system.
The reason for using SCR for comparison and not the cheaper selective non-catalytic de-
NOX system (SNCR) is that the latter is able to achieve only 30 - 50% NOX removal.
Furthermore SNCR has not been demonstrated on high sulfur coal. Excessive ammonia
10-36
-------
carryover is likely to result in severe air preheater plugging problems as well as problems
with waste characteristics. This excludes the use of SNCR for high sulfur applications
(22, 23). FF
The basis for the cost comparison is a 500 MW power plant burning 3% sulfur coal
and emitting 300 ppm NOX. A 65% NOX removal and a 90% SO2 removal are assumed for
both systems. The operating cost for the SCR system is critically dependent on the
catalyst lifetime. A two year lifetime has been assumed for this cost comparison.
The capital requirement and the operating/maintenance cost for the two systems
are shown in Table 14. It is seen that the capital requirement is considerably less for the
Niro Atomizer dry SOX/NOX system. Operating and maintenance costs are very similar
for the two systems.
TABLE 1*
OPERATING AND CAPITAL COST ESTIMATE FOR
COMBINED SOX AND NOX REMOVAL SYSTEMS *
COST ITEM
CAPITAL REQUIREMENT
OPERATING AND MAINTENANCE
COSTS (1st Year) (O&M)
LEVELIZED CAPITAL REQUIREMENT
LEVELIZED O&M
TOTAL LEVELIZED COST
WET LIMESTONE
+
SCR
105 M$ FGD
30 M$ SCR
135 M$
17.2 M$ FGD
4.6 M$ SCR
21.8 M$
7.4 mills/kWh
12.7 mills/kWh
20.1 mills/kWh
NIRO
DRY
SOX/NOX
85 M$
19.5 M$
4.8 mills/kWh
11.3 mills/kWh
16.1 mills/kWh
* References: (3, 20, 21 and 22)
A single annual levelized cost value has been computed based on a 30 year
economic plant life to represent the revenue requirement for the operating and
maintenance (O&M) cost and the capital requirement . For the O&M cost a levelization
factor of 1.9 is used. For the capital requirement a levelization factor of 0.18 is used.
The total levelized cost for the Niro Atomizer dry SOX/NOX system is 16.1 mills/kWh
compared to 20.1 mills/kWh for the limestone wet scrubber/SCR system.
10-37
-------
Besides having favourable economics when compared to a limestone wet
scrubber/SCR system the Niro Atomizer dry SOX/NOX process has the additional
advantage of being one integral air pollution control system. It will be easier to operate
and will not interfere with the boiler operation since no equipment is installed upstream
the air preheaters.
CONCLUSION
• Dry scrubbing has gained much acceptance within the utility industry due to its
low cost and ease of operation.
• Dry scrubbing with simultaneous remoVal of SOX and NOX is a new process -
developed by Niro Atomizer - which uses the same equipment as provided for dry
scrubbers presently in operation or under construction.
« The process is well suited for high sulfur coal applications where low NOX burners
have reduced NOX to the 0.25 - 0.35 Ib/MBtu level. The process can further reduce
NOX emissions down to the 0.1 Ib/MBtu level together with 90 - 95% SOX removal.
• Process performance has been tested in pilot-scale and has been confirmed in full-
scale at Argonne National Laboratories.
• Waste products from the process will most likely be classified as non-hazardous.
• Economics of the new process compare favourably with limestone wet scrubbing
plus selective catalytic reduction (SCR).
10-38
-------
REFERENCES
(1) Inside E.P.A. Weekly Report, by US EPA, Washington.
(2) Drehmel, D.C. et al
"Low NOX Combution Systems with SC>2 Control Using Limestone"
Industrial Environmental Research Laboratory, US EPA - paper no. 83-38.7-
(3) Slack, A. V.
"Technology for Power Plant Emission Control"
Report prepared for Niro Atomizer, June 30, 1983.
(!+) Swedish study by the Ministry of Agriculture
"Acidification Today and Tomorrow"
Prepared for the June 1982 Stockholm Conference on The Acidification of the
Environment.
(5) EEC Symposium
"Acid Deposition - A Challenge for Europe"
Karlsruhe, Germany - September 19-21, 1983
(6) 1982 Joint Symposium on Stationary Combustion NOX Control
Dallas, Texas, November 1982.
(7) EPA Environmental News.
The October 22, 1982 issue.
(8) Felsvang, K.S. et al
"Status of the JOY/NIRO Dry FGD System and Its Future Application for the
Removal of High Sulfur, High Chloride and NOX from Flue Gases"
1983 Joint Power Generation Conference, Indianapolis, Indiana.
September 25-29, 1983.
(9) Donnelly, J.R. et al
"Process for Removal of Nitrogen Oxides and Sulfur Oxides from Waste Gases"
U.S. Patent Application S.N. 382,968
(10) Felsvang, K.S.
"Results from Operation of the Riverside Dry Scrubber"
JOY/NIRO Seminar, Minneapolis, Minnesota - June 1981.
(11) Klingspor, J.
"Kinetics and Engineering Aspects on the Wet-Dry FGD Process"
Department of Chemical Engineering II, Lund's Institute of Technology, Sweden.
June 1983
(12) Farber, P.S.
"Start-Up and Performance of a High Sulfur Dry Scrubber System"
APCA 75th Annual Meeting, New Orleans, LA
June 1982
10-39
-------
(13) Farber, P.S.
"The Argonne High Sulfur Dry Scrubber"
JOY/NIRO Seminar on Dry Scrubbing of High Sulfur Coals, Minneapolis, Minnesota,
June 25-27, 1982
(14) Radian Corporation
"Characteristics of Waste Products from Dry Scrubbing Systems"
EPRI Report CS-2766, 1982
(15) Thompson, Carol M.
"Characteristics of Waste Products from Dry Scrubbing Systems"
EPA/EPRI FGD-Symposium, Hollywood, Florida
May 1982
(16) Donnelly, J.R.
"Disposal and Utilization of Spray Dryer FGD End Products"
Canadian Electrical Association Seminar, Ottowa, Canada
October 1981
(17) Buschman, J.C. et al
"Disposal of Wastes from Dry SC>2 Removal Processes"
Joint Power Generating Conference, Phoenix, Arizona
September, 1980
18) Donnelly, J.R. et al
"Dry Flue gas Desulfurization End-Product Disposal Riverside Demonstration
Facility Experience"
EPA/EPRI FGD-Symposium, Hollywood, Florida
May, 1982
(19) Farber, P.S.
"Leachate of Dry Scrubber Wastes"
70th APCA Meeting, Atlanta, Georgia,
June 19-24, 1983
(20) Stearns-Roger Eng. Corp. and Radian Corporation
"Technical and Economic Feasibility of Ammonia-Based Postcombustion
NOX Control"
EPRI Report CS-2713.
November, 1982
(21) "Preliminary Economic Analysis of NOX Flue Gas Treatment Processes"
EPRI Report FP-1253,
February, 1980
(22) "NOX Reduction Alternatives Study"
Belridge Field Cogeneration Project. Shell California Production Inc.
Submitted to the California Energy Commission, July 1983
(23) Jumpei, Ando
"NOX Abatement for Stationary Sources in Japan"
EPA-600/7-83-027
May, 1983
10-40
-------
PROCESS CHARACTERIZATION OF S02 REMOVAL IN
SPRAY ABSORBER/BAGHOUSE SYSTEMS
E.A. Samuel, T. W. Lugar, D. E. Lapp,
K. R. Murphy, 0. F. Fortune, T. G. Brna,
R. L. Ostop
-------
ABSTRACT
A new equation for correlating the SC>2 removal with stoichiometric ratio,
approach to saturation, inlet S02 concentration, inlet temperature, and
inlet moisture content is proposed. It reveals essential information about
the kinetics of the SC>2 removal process. The chemical reaction between the
S02 and lime appears to limit the rate of S02 removal in the spray
absorber. The effect of recycle enhancement of SC>2 removal may be
understood in terms of this model. The model also permits a classification
of flyash suitable for designing dry flue gas desulfurization (DFGD)
systems.
DFGD pilot plant results are reviewed from theoretical and mechanistic
viewpoints. The SC>2 removal as a function of the gas flow rate (or average
residence time) and the angle of the secondary swirl vanes in the absorber
gas disperser is discussed relative to the velocity flow field in the
absorber. The SC>2 removal efficiency is also discussed as a function of
the atomizer angular speed.
When the gas flow rate and approach to saturation are fixed, the S02
removal efficiency in the spray absorber increases with increasing
stoichiometric ratio and inlet temperature and decreasing inlet moisture
content and S02 concentration. An explanation of these effects is given
by their impact on the droplet diameter integrated over the lifetime of
the droplets. Another effect that increases the time-integrated droplet
diameter is the decrease in the volume of the core of solids (lime) for
smaller slurry droplets.
Partial recycle of the spent products, in the absence of externally
injected flyash, results in S02 removal efficiency being nearly independent
of S02 concentration. Thus, the enhancement from recycle is greater for
the higher S02 concentrations. S02 removal performance with recycle was
improved by the injection of flyash derived from a low sulfur Texas lignite
and from high sulfur eastern coals. The enhancement with recycle in the
latter case is dramatic and supports DFGD applicability to high sulfur
coal.
NOMENCLATURE
£ = 502 removal efficiency across spray absorber
a = Ys Cp/(NNuK)
Ys = Product deposition reaction rate constant
Cp = Flue gas specific heat at constant pressure
N = Nusselt number for liquid droplets in gas stream
Nu
K = Thermal conductivity of flue gas
10-41
-------
Xc = Correction factor for effect of the solids fraction on
the S02 removal rate
= i - Djj /D|
D-L = Initial mean droplet diameter
Df = Final mean particle diameter of evaporated droplet
S = Stoichiometric ratio
Tsi = Spray absorber inlet temperature
XSf = Spray absorber outlet temperature
Tas = Adiabatic saturation temperature of flue gas
entering spray absorber
A = Time-integrated droplet diameter
T
= r / D(t)dt
0
r = Concentration of droplets
T = Droplet evaporation time
D(t) = Droplet diameter at time, t
u = Empirical constant characteristic of a given flyash
F = Mass fraction of solids in slurry droplet or atomizer
slurry flow
Reduced penetration
= - £n(l - ec) / {X0 S
Tsi ~ Tas
c
sf as
INTRODUCTION
The New Source Performance Standards (NSPS) for S02 require cleanup of
flue gases from utility boilers burning even low-sulfur coals. The
parallel increase in application of fabric filters to control particulate
emissions from coal-fired boilers has generated considerable technical and
commercial interest in dry flue gas desulfurization (DFGD) processes. The
most commercially promising DFGD process is spray absorption. In this
process, an alkali sorbent liquid is atomized in a spray dryer using flue
10-42
-------
gas as the drying medium. The DFGD reactions occur between the gas and
droplets in intimate contact during the various stages of drying.
Eventually the reaction products are reduced to a suspended solid par-
ticulate in the gas stream, some solids dropping out in the absorber
hopper. When a fabric filter is used for removal of remaining spent sor-
bent and flyash from the gas stream, further DFGD reactions occur in the
filter cake by virtue of unreacted alkali and residual moisture. While a
variety of calcium- and sodium-based alkalis have been evaluated, lime
currently has the greatest commercial importance.
In 1978, the Buell Emission Control Division of the Envirotech Corporation
(now part of General Electric Environmental Services, Inc.) and Anhydro
A/S constructed a spray dryer FGD prototype facility at the Martin Drake
Station of the City of Colorado Springs. A 4.01-m3/s (8,500-cfm) spray
absorber was installed in slipstream configuration in the inlet duct to a
189 m3/s (400,000 cfra) fabric filter serving an 85 MW pulverized-coal-fired
steam generator burning northwestern Colorado coal. An extensive program
of parametric performance and process demonstration tests was undertaken at
that facility jointly with EPA and the City of Colorado Springs. As a
result improved models for spray dryer FGD performance were developed.
This paper presents these models and experimental data which support their
validity.
THEORY
Two models for predicting the S02 removal in the spray absorber/baghouse
system are presented in this paper. The first is an analytical model
restricted to straight-through operating conditions only. The second is an
empirical model applicable under wet recycle of solids collected from the
baghouse and/or the spray absorber hoppers. The analytical model provides
a theoretical basis for understanding the fundamental processes which
remove S02 by spray absorption. The empirical model extends the usefulness
of the analytical model by including the effect of flyash on S02 removal by
spray absorption.
Analytical Model for Straight-Through S02 Removal (No Recycle)
An analytical expression for the S02 removal rate (or efficiency) in a
spray absorber may be derived for the following idealized conditions within
the spray absorber.
1) The gas velocity is uniform within the spray absorber (the presence of
vortices, eddies, and recirculation zones is ignored).
2) The droplet/gas mixing is uniform, with the droplets following gas
streamlines.
3) Most S02 removal takes place during the period of water evaporation
from the droplets (often called Phase I), when surfaces are maintained
at the adiabatic saturation temperature.
10-43
-------
4) The rate of S02 removal is determined by the rate of the chemical reac-
tion between the 862 and lime to form sulfites or sulfates.
Under these assumptions, the simultaneous occurrence of S02 removal and
water evaporation is controlled by the rate of S02 removal, the rate of
droplet size changes, and the conservation of thermal energy in the spray
absorber (1,2). Consistent with these relationships the SC>2 removal effi-
ciency, under straight-through conditions, can be expressed by Equation 1.
T • - T
1si 1as
- £n(l - £s) = a Xc S [in j-^— ] (1)
sf as
The correction factor, XC) arises from the proportionality between the
penetration [i.e., - £n(l - es)] and the time-integrated droplet diameter.
For pure water droplets, the final diameter, Df, = 0. When solids are pre-
sent in the slurry droplet, its initial diameter is increased due to volume
displacement by the solids, and its final diameter, at the end of evapora-
tion, is non-zero due to the volume of the solids. As the solids fraction
increases, the departure of the correction factor, Xc, from unity becomes
progressively more significant. The correction factor, Xc, depends on both
the mean diameter, D^, of droplets produced in the atomizer and the final
mean diameter, Df( of particles produced at the end of evaporation. Its
accurate evaluation is complicated by the dependence of the initial droplet
diameter on the solids fraction in the slurry being atomized and by the
dependence of the final particle diameter on the extent to which the dried
particle retains some moisture. A reasonable expectation is for Xc ^ 1 for
solids fractions <0.1. At solids fraction of 0.5, the correction factor,
Xc, appears to be about 0.5 to 0.7. Table 1 discusses the dependence of
the model prediction for the spray absorber SC>2 removal efficiency on
selected spray absorber and flue gas parameters.
Empirical Model for SC>2 Removal with Wet Recycle
The S02 removal efficiency in the presence of wet recycle of offproduct is
invariably better than that under straight-through conditions. In addition
to unspent lime, the recycled offproduct contains flyash. The flyash
appears to have a synergistic effect on the utilization of lime from both
fresh and recycled sources. Different types of flyash in the recycle
slurry provide different levels of enhancement of the SC>2 removal. This
recycle enhancement appears to be related to the coal from which the flyash
originates and on flyash characteristics such as pH of hydrolysis, par-
ticle size, and chemical composition. Since the flyash effect on
SC>2 removal through wet recycle is not well understood, a fundamental rela-
tionship for the efficiency in the presence of recycle cannot be derived.
The following empirical relationship is suggested, based on Equation 1,
to be consistent with the observation of improved S02 removal in the pre-
sence of recycle which increases the solids fraction, F , in a newly pro-
duced slurry droplet (or in the atomizer feed):
10-44
-------
TABLE 1.
MDDEL PREDICTIONS FOR OPERATIONAL EFFECTS ON
STRAIGHT-THROUGH REMOVAL IN SPRAY ABSORBER
Parameter
How Parameter Enters
Equation 1
Relationship to
S02 Removal
Comment
Absorber Residence Time
Atomizer Angular Speed
Gas Inlet Temperature
O
I
-P-
Ui
Does not enter equation
explicitly.
Does not enter equation
explicitly.
Enters equation explicitly
as Tsl.
Gas Inlet Moisture Content
Enters equation indirectly
through the adiabatic sat-
uration temperature, Tas.
None, provided residence
time is long enough to
allow complete evaporation.
None, provided droplets are
of such size that they evap-
orate completely.
S02 removal efficiency
increases with increasing
inlet temperature for
fixed approach to
saturation temperature.
S02 removal efficiency
decreases with increasing
inlet moisture content
for fixed inlet gas temp-
erature and approach to
saturation.
Water flow rate and the inlet and
outlet gas temperatures are held
at conditions which give complete
evaporation and maintain Xc - 1.
For a fixed atomizer water flow
rate, the time-integrated
droplet diameter is independent of
the initial droplet diameter,
provided complete evaporation
occurs and X - 1.
The water required to reach the
same approach to saturation
temperature increases with
increasing inlet gas temperatures.
The time—integrated droplet
diameter increases with increasing
inlet gas temperature.
For the same inlet gas temp-
erature, the adiabatic satura-
tion temperature increases with
increasing inlet moisture content.
Consequently, the water required
to reach the same approach to
saturation at fixed inlet gas
temperature decreases with
increasing inlet moisture content.
Inlet S02 Concentration
Enters equation indirectly
through the correction
factor, Xc.
S02 removal efficiency is
independent of inlet S02
concentration for low con-
centrations «1000 ppm).
S02 removal efficiency
decreases with inlet S02
concentration at high con-
centrations O1000 ppm) at
fixed stoichiometry.
Water flow rate and the inlet and
outlet gas temperatures are held
at conditions which give complete
evaporation. The droplet evapora-
tion time decreases with increasing
inlet S02 concentration. For in-
let S02 concentrations less than
about 1000 ppm, the lime rates are
low such that Xc = 1. At
higher inlet S02 concentrations,
Xc decreases with increasing
concentration.
-------
or
T • - T
si as
- es) = Xc S [£n —r~ ] (u F + a) (2)
sf as
P = u F + a (3)
where u is an empirical constant which varies with flyash type. The
constant, a, comes from Equation 1. Implications of this empirical equation
will be discussed later.
EXPERIMENT
TEST OBJECTIVES AND EQUIPMENT
The overall objective of the dry FGD (DFGD) pilot program at Colorado
Springs was to demonstrate and characterize S02 removal in a spray
absorber/baghouse system for low sulfur and high sulfur applications.
Although lime was the main reagent tested, other reagents, such as trona
and dolomitic lime, which may be preferable for certain power plant sites,
were also evaluated. Limestone was also tested because its effective use
in a DFGD system will dramatically shift FGD economic evaluation in its
favor. Other secondary objectives of the Colorado Springs program were to
study the effects on S02 removal of baghouse reheat, adipic acid addition
to lime, and the use of wastewater for lime slaking. Details of the tests
conducted in the pilot program are reported elsewhere (3,4,5,6). Only
those tests which are relevant to the proposed models are discussed in
this paper.
The primary elements of the Colorado Springs prototype plant are a 3.8-m
(12.5-ft) diameter spray absorber equipped with a top-entry vaned scroll-
type gas disperser; a centrifugal atomizer with a 0.25-m (10-in.) diameter
wheel; a full-scale commercial-type lime storage, handling, and slaking
system; and a reverse-air baghouse fitted with full-size (0.30-m or 12-in.
diameter, 9.14-m or 30-ft length) fiberglass filter bags. The plant is
also equipped with a four-element cyclone collector in parallel with the
baghouse at the absorber outlet to provide the capability to independently
vary the absorber flow volume and baghouse air/cloth ratio. An S02 vapor-
izing and injection system enabled increasing the S02 concentration at the
spray dryer inlet from 200 to 400 ppm (the value typical of the Colorado
coal) to values in the range from 1,000 to 4,000 ppm. The higher values in
this range are characteristic of the flue gas derived from burning high
sulfur eastern coals. Flue gas characteristic of a given coal other than
the Colorado coal burned at the Martin Drake Station was simulated by the
external injection of S02 as needed and flyash derived from a boiler firing
the coal type of interest.
Major measurement systems included venturi meters for gas flow measurement
at the absorber and baghouse inlets; an extractive S02 monitoring system
(DuPont Model 460) with probes at the absorber inlet, at the absorber
10-46
-------
outlet, and at the baghouse outlet; a dew point hygrometer with a probe at
the baghouse outlet; magnetic flow meters for the measurement of slurry
flows; and extensive pressure and temperature instrumentation. All equip-
ment and materials required for instrument calibration were available on-
site, as were chemical laboratory facilities for measurement of slurry
reactant concentration, slurry total solids by evaporation, and offproduct
moisture content.
The Colorado Springs prototype plant was sized to accommodate a wide range
of operating parameters. Following the plant start-up, shakedown, and ini-
tial optimization tests, operation was established within the range of
parameters summarized in Table 2.
RESULTS
OPTIMIZATION TESTS
Average Gas Residence Time in Spray Absorber
Figure l(a) summarizes the results of a series of tests to determine the
variation in the S02 removal efficiency across the spray absorber,
baghouse, and the spray absorber/baghouse system, due to variation in the
residence time (from 7 to 12 seconds) for fixed atomizer angular speed
(12,500 rpm), stoichiometric ratio (1.0), and approach to saturation
(17°C, 30°F). The spray absorber S02 removal efficiency increases slowly
with decreasing residence time (increasing flow rate). On the basis of the
analytical model, a constant S02 removal efficiency with respect to resi-
dence time would be expected. The increase shown by test results appears
to be due to improved gas flow distribution and gas/droplet mixing with
increasing gas flow rate.
The S02 removal in the baghouse was more sensitive to the average gas resi-
dence time in the spray absorber than was S02 removal in the spray absorber
itself. The baghouse removal shows a maximum at a residence time of about
10 seconds. The increase in the baghouse efficiency when the residence
time increases from 7 to 10 seconds (decreasing gas flow rate) seems due
to the decreased cyclonic action in the spray absorber and the consequent
increase in the mass loading of the spent product into the baghouse. The
decreasing baghouse performance with increasing spray absorber residence
times beyond 10 seconds seems to be due to poorer gas flow velocity distri-
bution and the accompanying reduced gas/particle mixing in the spray
absorber. This baghouse behavior is also consistent with the improved
cyclonic activity and the accompanying decrease in the mass loading to the
baghouse when the angular momentum of the slurry becomes progressively
greater at the low gas flow rates corresponding to residence times greater
than 10 seconds.
Atomizer Wheel Angular Speed
Figure l(b) summarizes the observed variation in the spray absorber, baghouse,
and spray absorber/baghouse system S02 removal efficiencies with atomizer
angular speed. The gas flow rate was fixed at 4.01 rn^/s (8500 cfm)
10-47
-------
TABLE 2. COLORADO SPRINGS SPRAY ABSORPTION PROTOTYPE PLANT SUMMARY
OF OPERATING PARAMETERS
Parameter3
Typical
Value
Minimum
Maximum
Maximum
Minimum
Variation
during test,
percent
o
i
00
ABSORBER
Inlet flow volume, cfm"
Inlet temperature, °F
Inlet S02 concentration, ppm
Inlet saturation temperature,
Outlet temperature, °F
Tp, approach temperature, °F^
BAGHOUSE
Inlet flow volume, cfm
ATOMIZER
Wheel angular speed, rpm
SLURRY SYSTEM
8,500
350
1,000
120
140
20
3,000
12,500
5,000
250
250
110
122
12
1 ,000
6,300
10,600
400
4,000
140
180
40
.,000
14,000
2.1
1.6
16.0
1.3
1.5
3.3
6.0
2.2
Lime slurry flow, gpme
Total atomizer feed, gpm
Lime slurry total solids, percent
a Units shown apply only to columns
b Multiply cfm value by 4.72 x 10 ~L
c T. °n. = 5fT. °F - 321/9.
1.0 0.5 3.0 6.0
2.0 1.0 6.0 6.0
15 8 50 6.3
under "Value.1
h to convert to m-Vs.
2
2
2
Tp, °C = 5(Tp, ..
Multiply gpm value by 6.3 x 10 5 to convert to mj/s.
-------
- SPRflY flBSORBER
- BflGHOUSE
- SPRftY fiBSORBER/BflGHOUSE
SYSTEM
-\ 1 1 1 1 p
10 11 12
BVERflGE RESIDENCE TIME, SECONDS
O
I
1(a). Observed dependence of
straight-through 862
removal efficiency on
average gas residence
time in the spray absor-
ber. A residence time
of 10 seconds appears
to optimize the system
SC>2 removal efficiency.
- SPRflY fiBSOSEER
o - BflGHOUSE
- SPROT flBSORBER.BflGHOUSE
SYSTEH
- SPRflY flBSORBER
- BfiGHOUSE
- SPRflY flBSORBER/BflGHOUSE
SYSTEM
9000 11000 13000
RTOMIZER flNGULflR SPEED, RPfl
l(b). Observed dependence of
straight-through S02
removal efficiency on
the atomizer angular
speed. The system S02
removal efficiency
appears to be optimized
for angular speeds _>_
10,500 rpm. ~
Kc)
IB 20 30
VflNE flNGLE, DEGREES
Observed dependence of
straight-through S02
removal efficiency on
the inlet gas disperser
vane angle (with the
vertical). A vane
angle of 15 to 20°
appears to optimize the
system S02 removal
efficiency.
Figure 1. Optimization test results for pilot lime spray absorber/baghouse system without recycle
of offproduct. [For these tests, the following parameters were fixed: inlet S02
concentration (1000 ppm), stoichiometric ratio (1.0), inlet gas temperature (177°C,
350°F), and approach to saturation (17°C, 30°F).]
-------
(10-second average residence time), the stoichiometric ratio at 0.95,
approach to saturation at 17°C (30°F) , and the inlet temperature at 111°C
(350°F) for these tests. S02 removal in the spray absorber was relatively
independent of wheel speed in the range tested and, consequently, indepen-
dent of droplet diameter (6,7,8). This independence is consistent with the
theoretical expression summarized in Equation 1 for the case of low solids
concentration in the atomizer feed. In this case, Df - 0 and the correc-
tion factor, Xc,~ 1. Hence the time-integrated diameter of the droplets to
which the S02 removal is proportional is independent of initial droplet
diameter.
Figure l(b) also shows a stepwise increase in the baghouse SC>2 removal effi-
ciency at an atomizer angular speed of about 10,500 rpm. The system
S02 removal efficiency reflects the sensitivity of the baghouse S02 removal
efficiency to variations in atomizer angular speed and appears to be opti-
mized at atomizer angular speeds above 10,500 rpm.
Atomizer Disperser Vane Angle
By tilting the secondary swirl vanes of the gas disperser from the ver-
tical (i.e., parallel to the atomizer wheel axis), it is possible to impart
a tangential component to the gas velocity as the gas enters the spray
chamber and thereby induce a swirl in the gas velocity flow field. An ini-
tial swirl in the gas velocity flow field is advantageous in extending the
gas residence time by inducing a helical motion to the gas molecules in the
spray absorber and in improving gas/particle mixing by providing a centri-
petal acceleration to the particles. Figure l(c) summarizes the observed
S02 removal efficiencies with lime as a function of the vane angle
(measured from the vertical) for fixed stoichiometric ratio (0.92), inlet
gas temperature (177°C, 350°F), approach to saturation (17°C, 30°F) , and
atomizer angular speed (12,500 rpm).
The vane angle appears to improve the SC>2 removal in the spray absorber
noticeably in the range from 0 to 20°. Increasing the vane angle beyond
20° does not lead to noticeable improvement in the performance of the spray
absorber. This suggests that a higher level of turbulence at the higher
vane angle does not allow the swirl to be completely characterized by the
initial angle of projection and that, as the flow progresses into the
chamber, a spectrum of swirl angles is produced. In this way, the tur-
bulence in the flow field removes any identity of the flow with the initial
swirl angle for vane angles above 20° in the range from 0 to 45°. The
contribution of the angular momentum of the slurry to the swirl of the gas
will also tend to lessen the effect of increasing the vane angle.
The baghouse SC>2 removal efficiency appears to be independent of the vane
angle. This suggests that the size distribution of particles produced at
the end of evaporation of the slurry droplets created in the rotary atom-
izer are in such a range that their carryover to the baghouse is unaf-
fected by changes in vane angle.
10-50
-------
Spray Absorber Inlet Temperature at Fixed Inlet Flue Gas Moisture Content
The effect of varying the temperature of the flue gas at the spray absorber
inlet without changing its moisture content was studied in a series of
tests. In these optimization tests, the atomizer disperser vane angle was
set at 15° and the atomizer wheel speed was maintained at 12,500 rpm. As
in earlier tests, the gas flow rate was 4.01 m-Vs (8,500 cfm) and
corresponded to an average residence time of 10 seconds in the spray absorber.
The spray absorber inlet temperature was varied by using a surface heat
exchanger with water as the cooling fluid upstream of the spray absorber
inlet.
Figures 2(a) and 2(b) display the experimental results for the S02 removal
efficiencies under straight-through conditions for approach temperatures of
11°C (20°F) and 17°C (30°F). The spray absorber S02 removal efficiency
increased with increasing inlet temperature, while the baghouse efficiency
displayed the opposite trend. Thus the system SC>2 removal efficiency
increased only weakly with increasing inlet temperature. The rate of
increase of the spray absorber efficiency and the rate of decrease of the
baghouse efficiency with increasing inlet temperature at a 17°C (30°F)
approach to saturation are more than double the rates at an 11°C (20°F)
approach.
Since the thermal energy of the inlet gas decreases with decreasing tem-
perature, the water required to reach a fixed approach to saturation
decreases with decreasing inlet temperature. When the atomizer is operated
at a fixed angular speed and the lime feed rate is low enough such that the
correction factor, Xc, ~1, the mean size of droplets produced in the atom-
izer may be regarded as being independent of the inlet gas temperature.
The decrease of the atomizer water feed rate with decreasing inlet gas tem-
perature then translates to a decrease in the initial total surface area of
droplets produced in the atomizer because fewer droplets are formed.
The decrease of the spray absorber efficiency with a decrease in the inlet
temperature can be related to this lower total initial surface area of
droplets resulting from a lower water feed rate to the atomizer to reach
the same approach to saturation. The corresponding increase in the
baghouse removal suggests that the slower evaporation rate induced by the
lower inlet temperature leaves particles at the spray absorber outlet with
a higher moisture content, allowing more effective S02 removal in the
baghouse. For the same gas flow rate at the same inlet temperature, the
atomizer water requirement to reach the 17°C (30°F) approach is less than
that for the 11°C (20°F) approach. Thus, the higher rate of decrease of
spray absorber efficiency with decreasing inlet temperature for the 17°C
(30°F) approach may arise from the difference in the water requirements and
the associated difference in the initial droplet surface area.
Spray Absorber Inlet Temperature Variation with Prequench
The previous section investigated the effect on S02 removal of varying the
inlet temperature while holding the absolute humidity fixed. That set of
10-51
-------
» P
cv:
UJ
UJ
t—t
LJ
100-
75—
50-
25-
o
I
EFFECT OF HEAT EXCHANGE (STRAIGHT-THROUGH)
INLET S02 CONG. = 1000 PPM
TP
D - SPRAY ABSORBER
x - BAGHOUSE
+ - SYSTEM
I—I—T
n—i—i—r
i—i—r
i—i r
100 •
*_J
LU
I—I
t! 50—|
l_l_
U_
LU
OJ
d>
LTb
25—
EFFECT OF HEAT EXCHANGE (STRAIGHT-THROUGH)
INLET S02 CONC. = 1000 PPM
Tp - 17°C (30°F); 5 = 1.0
D - SPRAY ABSORBER
x - BAGHOUSE
+ - SYSTEM
200 250 300 350 400
SPRAY ABSORBER INLET TEMPERATURE, DEC. F
n—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i i i r
200 250 300 350 400
SPRAY ABSORBER INLET TEMPERATURE, DEC. F
2 (a). Dependence of the straight-through SC>2
removal efficiency on the spray absor-
ber inlet gas temperature varied by
cooling the inlet gas with a surface
heat exchanger for fixed approach to
saturation (11°C, 20°F).
2(b). Dependence of the straight-through SC>2
removal efficiency on the spray absor-
ber inlet gas temperature varied by
cooling the inlet gas with a surface
heat exchanger for fixed approach to
saturation (17°C, 30°F).
Figure 2. Effect of the spray absorber inlet gas temperature on the SOo removal efficiency in
pilot lime spray absorber/baghouse system without recycle of offproduct. [For these
tests, the following parameters were fixed: inlet S02 concentration (1000 ppm),
stoichiometric ratio (1.0), average gas residence time in spray absorber (10 seconds),
atomizer angular speed (12,500 rpm), and inlet gas disperser configuration (20° vane
angle with vertical and small vane inserts).]
-------
tests was useful in understanding the effect of inlet temperature on spray
absorber/baghouse performance for bituminous and subbituminous coals. It
is also of interest to determine the effect of inlet temperature with
moisture content in the range representative of flue gas derived from
burning lignite. Increased flue gas moisture content and decreased spray
absorber inlet temperature were simultaneously accomplished by cooling the
flue gas entering the spray absorber by adiabatic humidification using
spray nozzles. Figure 3 summarizes the test results. Flue gas originally
at 177°C (350°F) and a moisture content of 10 percent by volume will, after
adiabatic humidification to 149°C (300°F), contain moisture at 11.5 percent
by volume. As in the case of the heat exchanged gas, the water required in
the absorber to reach the same approach to saturation will be lower when the
inlet gas has already been partially prequenched. The lower water require-
ments and the corresponding decrease in the initial droplet surface area
are reflected in the decreasing S02 removal efficiency with decreasing tem-
perature (increasing extent of humidification) seen in Figure 3. When a
gas is cooled at constant absolute humidity (by surface heat exchange), its
adiabatic saturation temperature decreases. When it is cooled by adiabatic
humidification (by prequench), its adiabatic saturation temperature remains
the same, but its absolute humidity increases. As a result, for the same
inlet gas temperature, more water is required by the flue gas cooled by
surface heat exchange than by the prequenched (humidified) flue gas to
reach the same approach to saturation. Based on this difference in water
requirement, the SC>2 removal efficiency for the heat exchanged gas would be
expected to be higher than that for the prequenched gas for the same inlet
gas temperature. The experimental results shown in Figures 2(a) and 3(a)
confirm this theoretical expectation. Moreover, the rate of decrease of the
spray absorber efficiency with inlet temperature for the prequenched gas is
about triple that for the precooled gas. At the same time, the SC>2 removal
efficiency of the baghouse is correspondingly greater for the prequenched
gas than for the precooled gas at the same inlet temperature.
STRAIGHT-THROUGH AND WET RECYCLE TESTS
Comparison with Theoretical Model
Parametric tests were performed to characterize S02 removal in the spray
absorber/baghouse system with respect to variations in stoichiometric ratio
(0.7 to 1.5), approach to saturation (7 to 22°C, 12 to 40°F) , inlet S02
concentration (500 to 4,000 ppm), and recycle flyash (mixtures of flyash
derived from eastern, western, and lignitic coals with the Martin Drake
flyash). Other spray absorber operational parameters [such as atomizer
angular speed (12,500 rpm), inlet gas flow rate (4.01 nrVs, 8,500 cfm), and
inlet gas temperature (177°C, 350°F)] were fixed for these parametric
tests. Figure 4 presents the results of the parametric tests. When the
chemical reaction between the lime and S02 is the rate limiting step, the
theoretical model (Equation 1) predicts that all points belonging to a
fixed inlet concentration should correlate on the same curve. Figure 4
shows good agreement (as supported by correlation coefficients better than
0.9) between measurements and the theoretical model predictions. The
measurements and the fitted curves show a concentration-dependent perfor-
mance in the spray absorber, with the lowest concentration (500 ppm) giving
10-53
-------
100-
75 —
OJ
Q
U1
25—
o
i
Ln
-P-
EFFECT OF PREQUENCH (STRAIGHT-THROUGH)
INLET £02 CONC. = 1000 PPM
Tp = 11°C (20°F); S = 1.0
D - SPRAY ABSORBER
x - BAGHOUSE
+ - SYSTEM
i I I I | I T I I I I I I 1 I I T
200 250 300 350
SPRAY ABSORBER INLET TEMPERATURE, DEC. F
100-
75-
CVJ
CD
in
25 —
EFFECT OF PREOUENCH <50 PERCENT WET RECYCLE)
INLET S02 CONC. = 1000 PPM
TD = 11°C (20°F); S = 1.0
a - SPRAY ABSORBER
x - BAGHOUSE
+ - SYSTEM
\ \ T T
I I II F 1 I
350
200 250 300 350 400
SPRAY ABSORBER INLET TEMPERATURE, DEC. F
3(a). Dependence of straight-through SC>2
removal efficiency on the spray absor-
ber inlet gas temperature varied by
incomplete adiabatic humidification
of the inlet gas (prequench).
3(b). Dependence of the S02 removal efficiency
under wet recycle conditions on the spray
absorber inlet gas temperature varied by
incomplete adiabatic humidification of
the inlet gas (prequench).
Figure 3. S02 removal efficiency in pilot lime spray absorber/baghouse system for both straight-
through and wet recycle of offproduct operation with water prequenching of the flue gas
ahead of the spray absorber. [For these tests, the following parameters were fixed:
inlet SC>2 concentration (1000 ppm), stoichiometric ratio (1.0), average gas residence
time in spray absorber (10 seconds), atomizer angular speed (12,500 rpm), inlet gas
disperser configuration (20° vane angle with vertical and small vane inserts), and
approach to saturation (11°C, 20°F).]
-------
100 •
90 —
OJ
CD
in
CD
in
GO
70—
60 —
50-
o
I
40-
INLET S02
CONC., PPM
(20°F)
- Tp = 17°C (30°F)
o - Tp = 22°C (40°F)
1.0
\^ | I I I I | I I I I | I I I I | I I I I
2.0 3.0 4.0 5.0 6.0
S X LN«TINLET - TflS>/(TEX1T -
100-
90
n 80 —
70 —
CM
CD
in
60-
50 —
INLET S02
CONC., PPfl
s - Tp =
(20°F)
- TD = 17°C (30°F)
o - TD = 22°C (40°F)
i i i i i i i i i i i i i i i i i i i i i i i r
1.0 2.0 S.0 4.0 5.0 S.0
S X LN«TINLET - Tfls>/2 removal efficiency
data for operation with wet recycle of
offproduct in the absence of external
flyash injection compared with
Equation 1.
Figure 4. Comparison of spray absorber S02 removal efficiency data (points) from pilot lime spray
absorber/baghouse system with predictions of Equation 1 (curves) for straight-through
and wet recycle operation.
-------
the best performance. Successive curves in Figure 4 are separated by a
change in the inlet concentration of 500 ppm. For a fixed inlet gas tem-
perature and inlet gas moisture content, the rate of water flow in the ato-
mizer (rate of water evaporated in the spray absorber) is fixed by the
approach to saturation at the spray absorber outlet. At the same time, the
lime feed rate to the atomizer required to maintain a fixed stoichiometric
ratio increases with increasing inlet S02 concentration. Consequently, the
fraction of solids (lime) in the atomizer feed (or in a newly produced
slurry droplet) increases with increasing inlet SC>2 concentration for a
fixed stoichiometric ratio and approach to saturation. The observed
deterioration of the spray absorber S02 removal efficiency under straight-
through conditions with increasing inlet S02 concentration in the range of
1000 to 2500 ppm may be explained by the decreasing evaporation time of a
slurry droplet with increasing S02 concentration. Figure 4 shows the
straight-through S02 removal efficiency in the spray absorber to be nearly
independent of inlet 862 concentration in the range 0-1000 ppm. This
concentration-independent performance in the spray absorber for low inlet
S02 concentrations is also predictable by the analytical model. In this
case, the solids fraction in the atomizer is low enough that it essentially
has no effect on the droplet evaporation time. Figure 4 also shows that
the introduction of partial wet recycle of the offproduct results in the
near restoration of concentration-independent spray absorber S02 removal.
It appears that this recycle enhancement results from improved reaction
sites and a greater lime availability at the same stoichiometric ratio.
Characterization of the Flyash Effect
The correlation based on Equation 2 for different types of flyash in the
slurry droplets is illustrated in Figure 5. The pertinent ash charac-
teristics are presented in Table 3. The notable feature of the correlation
is the inclusion of straight-through as well as wet recycle data for all
combinations of S02 concentration, stoichiometric ratio, and approach to
saturation on a single curve. It is equally noteworthy that, despite phy-
sical and chemical differences, the data involving all three types of
eastern flyash correlate on a single curve.
The recycle enhancement of 862 removal efficiency is the lowest for the
Martin Drake flyash (no external flyash injection) which gives a nearly
neutral hydrolysis (pH - 7). The recycle enhancement is higher in the pre-
sence of the flyash derived from the Texas lignite which leads to an alka-
line hydrolysis (pH ~ 8). The performance enhancement from wet recycle is
dramatic in the presence of flyash derived from eastern coal which yields
acidic hydrolysis (pH - 4). This extremely high recycle enhancement sup-
ports application of the DFGD technique to high sulfur coal.
Under straight-through conditions, the effect of an increased solids frac-
tion was to decrease S02 removal efficiency in the spray absorber due to
the decrease in the evaporation time of the slurry droplets. In contrast,
under wet recycle conditions, the effect of an increased solids fraction
appears to be an increased spray absorber S02 removal efficiency. This
increased efficiency suggests that the beneficial effects of flyash in
10-56
-------
o
I
Ln
NO INJECTED FLYRSH
• - STRfllGHT-THROUGH
« - HET RECYCLE
SPRflY flBSORBER
n 1 1 ] i i i ] i i r
0.2 0.4 0.6
MflSS FRflCTION OF SOLIDS IN flTOMIZER FEED
TEXflS LIGNITIC FLYRSH INJECTED
o - STRfllGHT-THROUGH
O - WET RECYCLE
SPRflY flBSORBER
T 1 i 1 1 1 1 1 1 1 T
0.2 0.4 0.6
MflSS FRACTION OF SOLIDS IN flTOMIZER FEED
EflSTERN COflL FLYflSHES INJECTED
BOILER STRftlGHT WET
TYPE -THROUGH RECYCLE
CYCLONE x
TflNGENT. +
STOKER >•
SPRflY flBSORBER
T 1 1 1 1 1 1 1 I I r
B.2 0.4
BflSS FRflCTION OF SOLIDS IN RTOHIZER FEED
5(a). Reduced penetration in
the spray absorber with
no injected flyash com-
pared with Equation 2.
5(b). Reduced penetration in
the spray absorber with
injection of Texas lig-
nitic flyash compared
with Equation 2.
5(c). Reduced penetration in
the spray absorber with
injection of various
eastern coal flyashes
compared with
Equation 2.
Figure 5. Comparison of spray absorber reduced penetration data (points) from pilot lime spray
absorber/baghouse system with predictions of Equation 2 (straight lines) for straight-
through and wet recycle operation with and without external flyash injection.
-------
Table 3. CHEMICAL AND PHYSICAL CHARACTERISTICS OF FLYASH EVALUATED FOR THEIR
EFFECT ON S02 REMOVAL IN PILOT SPRAY DRYER
Composition, percent by mass
Constituent
or
Characteristic
Si02
A1203
CaO
F6203
Na20
K20
Li20
MgO
P205
S03
LOIf
Initial pH?
Alkalinity,
meq/gh
Mass median
diameter, (am
Log-normal
standard deviation*
Volume-to-surf ace
diameter, yml
Specific surface area,
">2/kg
Typical surface area
at spray dryer inlet
m2/m3 (ft2/ft3) gas
Typical concentration
at spray dryer inlet
1CT3 kg/m3 (gr/ft3)
Stoker a
Eastern
Coal
34.00
14.34
2.19
23.64
0.49
1.90
0.02
1.17
0.38
1.11
19.40
4.7
-0.062
67
4.1
24.8
106
0.62(0.19)
5.7(2.5)
Tangential
Eastern
Coal
45.77
19.26
2.05
25.96
0.41
1.98
0.02
0.90
0.16
0.45
1.42
4.4
-0.038
28.5
4.3
9.8
251
1.15(0.35)
4.6(2.0)
Cyclone c Front-fired d
Eastern
Coal
27.99
13.47
1.49
21.01
0.31
1.75
0.02
0.69
0.24
0.53
31.34
3.2
-0.307
14.3
2.6
9.1
253
0.59(0.18)
2.3(1.0)
Western
Coal
62.44
23.17
3.15
3.70
0.21
1.20
0.02
1.27
1.22
0.45
1.85
>7 .0
0.180
17.3
2.8
2.8
10.2
280
1.15(0.35)
4.1(1.8)
Lignite6
flyash
60.05
21.91
8.87
3.09
0.46
0.89
0.02
2.04
0.10
0.31
0.10
>7 .0
5.12
—
—
—
—
—
8.0(3.5)
a Stoker-fired boiler at the University of Iowa
° Tangentially fired boiler in Units 5 and 6 of the Conesville Station operated by Columbus and Southern Ohio
Electric
c Cyclone-fired boilers in Unit 2 of the Conesville Station operated by Columbus and Southern Ohio Electric (inlet
hopper sample)
d Front-fired boiler in Unit 6 of the Martin Drake Station operated by the City of Colorado Springs
e Flyash derived from burning Texas lignite at the Monticello Station operated by Texas Utilities Services
f LOI stands for loss on ignition at 800°C (1472°F).
S Initial pH of 0.01 kg fly ash in 0.1 i water at 52°C (126°F)
h Negative alkalinity indicates acidity being titrated against base to a final pH of 7.5.
1 Log-normal standard deviation, a, is obtained from a - 0.5 (zj/T + z7z2) where 15.9, 50.0, and 84.1 percent of
the mass of the particle distribution originate from particles whose diameters are greater than zj, z", and z2.
J Equal to djj exp [-1/2(C n o )2 ] where dm is the mass median diameter and a is the log-normal standard deviation
10-58
-------
enhancing the rate of chemical reaction between the S02 and lime override
the limitation in efficiency due to reduced evaporation time.
CONCLUSION
The data obtained from the Colorado Springs DFGD test program strongly
support the models given by Equations 1 and 2. As a result, S02 removal
performance can now be accurately predicted for a conservatively designed
system when the stoichiometric ratio, approach to saturation temperature,
and slurry solids content are defined. The predictions can be extended to
systems using ash recycle when an additional empirical constant derived
from the flyash type is available.
An important DFGD feature which was identified is the manner in which the
spray absorber and baghouse work in unison toward maximizing the system
S02 removal efficiency. This trend is evident in the variation of the
S02 removal efficiencies with inlet temperature. The decrease in spray
absorber performance with decreasing inlet temperature is seen to be com-
pensated by an increase in the baghouse performance to result in a nearly
constant system removal. Conditions of high humidity and low inlet tem-
perature are favorable to good S02 removal in the baghouse. Conversely,
conditions of low humidity and high inlet temperature appear to be
favorable to good S02 removal by the spray absorber. Significantly,
however, the system performance appears to be only weakly dependent on
either the inlet temperature or the humidity which may occur either as a
result of variations in coal moisture content or variations in the ambient
(combustion) air. The spray absorber/baghouse system appears capable of
effecting good performance even in the presence of these variations when no
baghouse reheat is provided. Baghouse reheat removes the compensating
character of baghouse operation and allows the spray absorber sensitivity
to inlet humidity and inlet temperature to predominate.
10-59
-------
REFERENCES
1. Downs, W., W.J. Sanders, and C.E. Miller. "Control of S02 Emissions by
Dry Scrubbing." (Presented at the American Power Conference, Chicago,
Illinois. April 21-23, 1980.)
2. Apple, C. and M.E. Kelly. "Mechanisms of Dry S02 Control Processes."
EPA-600/7-82-026 (NTIS No. PB 82-196924), April 1982.
3. Parsons, E.L., Jr., L.F. Hemenway, O.T. Kragh, T.G. Brna, and R.L.
Ostop. "S02 Removal by Dry FGD." In Proceedings: Symposium on Flue
Gas Desulfurization - Houston, October 1980, Volume 2,
EPA-600/9-81-019b (NTIS No. PB81-243164), April 1981, pp. 801-852.
4. Parsons, E.L., Jr., V. Boscak, T.G. Brna, and R.L. Ostop. "S02 Removal
by Dry Injection and Spray Absorption Techniques." in Third Symposium
on the Transfer and Utilization of Particulate Control Technology,
Volume I, EPA-600/9-82-005a (NTIS No. PB83-149583), April 1982, pp.
303-312.
5. Samuel, E.A., T.W. Lugar, D.E. Lapp, O.F. Fortune, T.G. Brna, and R.L.
Ostop. "Dry FGD Pilot Plant Results: Lime Spray Absorption for High
Sulfur Coal and Dry Injection of Sodium Compounds for Low Sulfur Coal."
In Proceedings: Symposium on Flue Gas Desulfurization, Hollywood, May
1982, Volume II, EPRI CS-2897, March 1983, pp. 574-594.
6. Samuel, E.A., K.R. Murphy, T.W. Lugar, E.L. Parsons, Jr., and D.E.
Lapp. "Evaluation of Spray Absorption FGD," Draft Report submitted to
EPA under Contract No. 68-02-3119.
7. Marshall, W.E. and W.R. Marshall, Jr. "Evaporation from Drops (Parts 1
and 2)." Chemical Engineering Progress 48, 141 and 173. 1952.
8. Masters, K. "Spray Drying Handbook." George Goodwin Ltd., London,
1979.
10-60
-------
DRY SCRUBBER, FLUE GAS DESULFURIZATION ON HIGH SULFUR,
COAL-FIRED STEAM GENERATORS: PILOT-SCALE EVALUATION
B. J. Jankura, J. B. Doyle, T. J. Flynn
-------
DRY SCRUBBER, FLUE GAS DESULFURIZATION ON HIGH-SULFUR,
COAL-FIRED STEAM GENERATORS: PILOT-SCALE EVALUATION
By:
B. J. Jankura, J. B. Doyle,
and T. J. Flynn
Babcock & Wilcox Company
Research and Development Division
Alliance Research Center
Alliance, Ohio 44601
ABSTRACT
This paper describes the pilot-scale investigation of methods for adapt-
ing dry flue gas desulfurization (FGD) to utility steam generators burning
high-sulfur coal. Development of the dry scrubber for FGD was initially
directed toward reducing SO in steam generators burning low-sulfur, western
coals. The reason for limiting dry scrubbing to low-sulfur coals was two-fold:
First, federal New Source Performance Standards (70% reduction) were less
stringent. Second, western coals generally contain less sulfur and large
amounts of ash alkali. This significantly contributes to the effectiveness of
the dry scrubbing process. Several potential drawbacks (both technical and
economic) limiting dry scrubber technology to western coals are discussed.
The Department of Energy (DOE) conducted dry scrubber FGD tests in 1981.
Preliminary results indicated dry scrubbing could remove more than 90% of the
sulfur released from the combustion of eastern coals at less-than-anticipated
levels of lime consumption. The Babcock & Wilcox Company (B&W), under DOE
contract, has tested the dry scrubbing process to evaluate the variables that
have a major effect on SO. capture.
There are potential ways of increasing SO capture via dry scrubbing. One
method requires limestone injection into the boiler furnace, reducing SO lev-
els to the scrubber, calcines limestone to more reactive lime, and simulates a
high-alkali-ash, high-sulfur coal. Other methods covered include recycle and
dry scrubber approach to saturation at temperatures less than 20°F to obtain
high SO capture.
Prepared for presentation at the Environmental Protection Agency/Electric
Power Research Institute Symposium on Flue Gas Desulfurization, New Orleans,
La., November 1-4, 1983.
10-61
-------
INTRODUCTION
The use of dry scrubbers for controlling SO emissions from power plants
burning low-sulfur coal has gained wide industrial acceptance. A natural out-
growth of the success of this technology is to establish its limits of appli-
cation in regard to higher-sulfur coals. The interest in developing dry scrub-
ber technology has been accelerated due to growing concerns about controlling
acid precipitation.
BACKGROUND
Using dry scrubbers for flue gas desulfurization began in the late 1970s.
The first commercial utility units are operating on low-sulfur coal. B&W1 s
development of dry scrubber technology has been a four step process:
• A small, field, pilot plant (8000 ACFM) was constructed and put
into operation in 1978 at the Basin Electric Neal Station, Velva,
N. D. This pilot provided basic process information but limited
information for scale-up and system sensitivity to fuel changes.
• To satisfy the need for additional design data, B&W proceeded
with construction and operation of a 120,000-ACFM field demon-
stration unit to demonstrate mechanical scale-up of the system.
• B&W then constructed a smaller 1500-ACFM pilot plant at the its
Alliance (Ohio) Research Center (ARC). This unit was designed to
develop basic process data.
• The net result of B&W's development efforts has been construction
and start-up a dry scrubbing system consisting of four units of
800,000 ACFM each (600 MW total) at Basin Electric's Laramie
River Station, Wheatland, Wyoming. (Another 450-MW unit at Colo-
rado Ute's Craig Station is scheduled for start-up in 1984.)
The data obtained from the B&W pilot plants and demonstration unit has been
used to establish a basis for predicting performance for dry scrubber systems.
PROGRAM OBJECTIVES
The objective of the B&W-DOE program was to evaluate dry scrubber tech-
nology when applied to high-sulfur coals. The first step was to develop para-
metric baseline data using standard dry scrubber practices. This would be fol-
lowed by using innovated methods for improving performance. This program was
funded writ hi n the DOE Advanced Environmental Control Technology (AECT) plan
initiated in 1979. The AECT program will develop the technology base for
controlling contaminants produced during coal conversion [1],
To accomplish the objectives, a wide range of process data were obtained.
We decided that the B&W pilot plant at (ARC) would be most suitable to meet
these objectives. This unit was selected because it was small enough
(1500 ACFM) to keep operating costs reasonable. Also, the unit was character-
ized against a larger (120,000-ACFM) field demonstration unit that provided a
sound basis for scaling up the data.
10-62
-------
DESCRIPTION OF TEST FACILITIES
B&W's 1500-ACFM, pilot dry scrubber (Figure 1) is highly flexible and de-
signed so a wide range of variables could be investigated. The system starts
with a combustion chamber that has a rating of 5 x 10 Btu/Hr heat input. The
combustor is also capable of firing oil or gas. The exhaust gas from the com-
bustor can be directly vented to the atmosphere or diverted to the dry scrub-
ber pilot facility. The gas stream going to the scrubber can be conditioned
with fly ash and/or SO . This capability provides a way of simulating flue
gas pollutants without burning actual fuel. The flue gas can travel directly
to the dry scrubber or pass through a heat exchanger that controls the flue
gas temperature at the scrubber inlet.
The first compartment of the dry scrubber is shown in Figure 2. A per-
forated plate is located at the inlet for distribution purposes. The gas then
passes through a TurboDiffuser™, where swirl energy is introduced to the gas
stream. The gas is then intimately mixed with a finely atomized reagent
slurry as it leaves the TurboDiffuser™. Here, the S09 reacts with the spray's
alkali as moisture in the spray is evaporated into the flue gas. A portion of
dry, spent product consisting of fly ash, calcium sulfates, calcium sulfites,
and unreacted lime deposits in the hoppers at the bottom of the scrubber
(Figure 3). The remainder is carried to a cyclone collector and/or baghouse,
where the majority of the particulate by-product is removed. The gas stream
is then exhausted through an induced-draft fan to the atmosphere.
The lime slurry reagent preparation system shown in Figure 1 is composed
of a reagent hopper feeding dry bulk reagent through a rotary valve to a com-
mercial paste slaker equipped with an automated grit separator. The lime
putty then discharges by gravity through a screen to a heated slurry tank. The
tank is continuously agitated and the slurry is continuously pumped around a
recycle loop to avoid potential settling out of the reagent and associated
plugging problems. The slurry feed line to the dry scrubber is equipped with
a strainer to protect the spray nozzle from plugging. Feed slurry is intro-
duced to the scrubber through a pneumatic spray nozzle (Figure 4).
TEST PLAN
A test plan was developed based on dry scrubber data obtained during
earlier development by B&W on low-sulfur coals. The starting point was to
establish a list of process and operating variables that were known to
influence process performance. (These variables are listed in order of
significance on Table 1; the test plan is shown on Table 2.) Included in the
test plan were methods for improving the overall process performance for dry
scrubbers when used in high-sulfur coal applications; these methods are:
• Use of recycle.
• Use of low temperature at the scrubber outlet.
• Use of furnace limestone injection in combination with the dry
scrubber.
10-63
-------
o
I
PULVERIZED
LIMESTONE
HOPPER
PREHEATED
SECONDARY
COMBUSTION AIR
WATER
GRIT
RECYCLE
STRAINER
TEMPERATURE
CONTROL
WATER
BCTU
STACK
n
DRY
ASH
INJECTOR
STEAM
HEATER
LIQUID SO2
WATER
COOLED
HEAT
EXCHANGER
BAG!
ASH
ASH
Figure 1. Babcock & Wilcox 1500-ACFM dry scrubber pilot.
-------
FLUE GAS
DISTANCE PIECE
SLURRY/ATOMIZING
AIR OR STEAM
DISTRIBUTION PLATE
GAS FLOW
GAS FLOW
*VENT TUBE CLOSED DURING ARC TESTING
Figure 2. Dry scrubber plenum with distributor
plate and throat configuration.
10-65
-------
?
FLUE
GAS
PLENUM
REGISTER ARRANGEMENT
DSR REACTOR
COMPARTMENT NO. 1 HOPPER
COMPARTMENT NO. 2 HOPPER
COMPARTMENT NO. 3 HOPPER
DUPONT S02 ANALYZER
Figure 3. Dry scrubber dimensions.
EXIT HOLE
—SLURRY-r-* J-H-P_
INSERT
3 HOLES, 120° APART
AIR OR STEAM HOLES
Figure 4. Advanced sprayer plate for pneumatic
atomization of slurries.
10-66
-------
Table 1
PROCESS AND OPERATING VARIABLES INFLUENCING
DRY SCRUBBER SYSTEM PERFORMANCE
Process
• Reagent type
• Dryer inlet
temperature
• Dryer Inlet S
concentration
• Fly ash
composition
Operating
• Stoichiometry
• Dryer outlet temperature
• Solid recycle rate
• Dryer residence time
Table 2
DESCRIPTION OF DRY SCRUBBER
TEST VARIABLES
Baseline tests :
Stoichiometric ratio
Dryer Inlet temperature
Dryer outlet temperature
Baghouse outlet temperature
Baghouse air-to-cloth
Baghouse pressure drop
Dryer residence time
Reagent variation
Coal variation
Recycle tests
Low-temperature outlet tests
Limestone injection
10-67
-------
PILOT PLANT TEST RESULTS
The major emphasis of dry scrubber testing was to evaluate utilizing lime
slurry reagent at high concentrations of SO at the inlet under a variety of
process and operating conditions. Holding other variables constant, testing
began by generating a nonrecycle data base that included a variation of both
reagent and coal. The first series of tests were designed to show the effect
of stoichiometry (over a wide range) on S02 capture. Subsequent test series
looked at recycle, low temperature at the scrubber outlet, and limestone
injection as an alternative means of improving lime utilization. Baghouse,
dryer residence time, and coal variation test results are not yet available.
Lime slurry reagent was prepared on site with a paste slaker. The Miss-
issippi pebble lime specifications are listed in Table 3. Typical chemical
analysis of pebble line samples taken during slaking operation are shown in
Table 4. The lime slurry has approximately 50% wt/wt below 7-micron-diameter
particles.
The test furnace was fired with a Western Kentucky bituminous coal; chem-
ical analyses of the coal and ash are shown in Tables 5 and 6. The sulfur
content of this coal ranged from 2.6% to 3.0%; furnace excess air ranged from
15% to 20%.
STOICHIOMETRIC RATIO AND OUTLET TEMPERATURE VARIATION
Once-through dry scrubber plus baghouse SO absorption values are shown
in Figure 5; typical inlet conditions were:
Inlet SO level — 1900 - 2300 ppm
Inlet Gas Temperature — 290° - 311°F
Dryer Gas Residence Time — 9.6 to 10.8 sec
Atomizing Air Consumption — 0.17 to 0.21 Ibs air/lbs slurry
Baghouse inlet temperature was held at least 20°F above the flue gas
adiabatic saturation temperature to prevent bag blinding. The approach to
saturation temperature (AST) is calculated as the difference between the flue
gas bulk temperature and its adiabatic saturation temperature.
The significance the dry scrubber's outlet temperature has on S09 capture
is evident in Figure 5. At a stoichiometry of 1.15, approximately 90% SO
capture was measured at a 10°F AST. ^
Dry scrubber absorbers have been recently tested on high-sulfur coals by
several researchers. At a lime feed rate corresponding to a stoichiometric
ratio of 1.2, and inlet S02 concentration of 2000 PPM, Samuel, et al. [2],
obtained 84% S02 capture with an AST of 20°F. Yeh, et al. [3], using commer-
cial hydrated lime, obtained approximately 65% SO capture at a 1.0 stoich-
iometry and inlet SO concentration of 2100 ppm and a 30°F AST. These results
are consistent with the baseline data in Figure 5.
10-68
-------
Table 3
TYPICAL MISSISSIPPI PEBBLE LIME SPECIFICATIONS
Mississippi or Peerless Rotary Plant
Pebble Quickline
Size Designation 1/2 - 1 - 2 inches
CaO — available
CaO — Total . .
CaCO
CaSO,
4
S — equivalent •
SiO
2
Al-O.,
23
Fe o
23
MgO
LOI
Chemical Analysis
93.0%
96.0
03
0.04
0.01
0.75
0.15
0.08
0.53
0.30
to
to
to
to
to
to
to
to
to
to
97.0%
98.0
4.0
0.21
0.05
1.10
0.35
0.12
0.75
2.00
Table 4
CHEMICAL ANALYSIS OF MISSISSIPPI PEBBLE LIME
022682 042382
Negligible Negligible
0.1 0.06
1.0 0.17
93.3 96.5
0.6 not available
Date
Moisture (%)
S as S03 (%)
C03 as C02 (%
Ca as CaO (%)
Mg as Mg (%)
10-69
-------
Table 5
Dry Scrubber
ACG-S3-4030-31
April 5, 1982
Sample No.
Description
TYPICAL CHEMICAL ANALYSIS:
WESTERN KENTUCKY COAL
C-16459
C-16460
Basis
Total Moisture, %
Proximate Analysis, %
Moisture
Volatile Matter
Fixed Carbon
Ash
Gross Heating Value,
Btu per Ib.
Btu per Ib. (M&A-Free)
Ultimate Analysis, %
Moisture
Carbon
Hydrogen
Nitrogen
Sulfur
Ash
Oxygen (Difference)
Total
Pulverized
Run 0011,
0700
As Received
3.8
3.8
40.9
45.1
10.2
12450
—
3.8
69.1
4.9
1.33
2.92
10.2
7.75
100.00
Coal
031082
Dry
42.5
46.9
10.6
12940
14470
—
71.9
5.1
1.38
3.04
10.6
7.98
100.00
Pulverized Coal
031282, 0830
As Received
3.8
3.8
41.6
44.4
10.2
12420
—
3.8
68.8
4.8
1.38
3.05
10.2
7.97
100.00
Dry
—
43.3
46.1
10.6
12910
14450
71.5
5.0
1.43
3.17
10.6
8.30
100.00
10-70
-------
Table 6
TYPICAL CHEMICAL ANALYSIS: WESTERN KENTUCKY COAL ASH
Dry Scrubber
ACG-83-4030-31
April 5, 1982
Sample No.
Description
Ash Analysis (Spectrographlc), %
Silicon as SiO
Aluminum as Al 0
Iron as Fe20,
Titanium as TiO
Calcium as CaO
Magnesium as MgO
Sodium as Na 0*
Potassium as K 0*
Sulfur as SO
Phosphorus as P90.
C-16459
Pulverized Coal
Run 0011, 031082
0700
48.61
18.75
19.26
0.99
3.04
0.92
0.51
2.83
2.72
0.42
C-16460
Pulverized Coal
031282, 0830
48.86
19.26
19.92
0.99
3.12
0.90
0.53
2.71
3.05
0.44
100
IT
t
<
O 50
f\i
O
O
APPROACH TO SATURATION
TEMPERATURE
35°F (19.4rC)
15°F (8.3 = C)
10°F (5.6°C)
6°F (3.3°C)
20°F (11.1°C)
RESULTS OF
SAMUEL, ET AL.
30°F (16.7-C)
RESULTS OF
YEH. ET AL.
1.0
MOLES CALCIUM/MOLE SULFUR
2.0
Figure 5. Stoichiometric ratio testing.
10-71
-------
In general, eastern coals contain relatively small amounts of low alka
line ash when compared to Western coals. Previous B&W work on several western
coals indicated dry scrubber S09 capture of 25% - 35% at 0.0 stoichiometry.
For the 2000 ppm SO flue gas treated at a 35°F AST, total S02 capture was
less than 10%, as expected, at a 0.0 stoichiometry.
TEMPERATURE VARIATION AT SCRUBBER INLET
Evaporation time of slurry droplets containing relatively insoluble
solids is characterized by the slurries' physical properties (solids content,
density, initial particle diameter, and temperature) and the dryer inlet and
outlet temperature. Typical dry scrubbers in FGD applications reduce inlet
temperature by 100° - 150°F (38° - 66°C). The effect of this temperature
difference on SO capture for western coals was shown to be very small by
Parsons, et al. [4]. Figure 6 shows the effect of increasing dry scrubber
inlet temperature for a 35°F AST and a 1.0 stoichiometry. This indirectly
shows how scrubber temperature drop affects SO capture and clearly indicates
a significant influence of inlet temperature on SO capture. Previous B&W
tests with high-sulfur coal also implied inlet temperature can play a more
dominant role on SO removal for high-sulfur coals in dry scrubber
applications [5].
REAGENT VARIATION
Commercial and industrial dry scrubber FGD systems have relied on lime
slurries as the principal reagent. Limestone slurries have shown acceptable
S0? absorption properties in low-sulfur, wet scrubber FGD. But in dry scrub-
bing FGD, these slurries have unacceptable reactivity. Research in various
lime and, most recently, sodium-based reagents (such as Trona [4] and soda
ash [3]) has characterized relative reactivities.
Johnson [6] performed tests with fly ash alkali and concluded certain
high-calcium ash, western coals contained enough inherent alkali to obtain the
required SO capture. The majority of commercial dry scrubber systems today
use lime reagents; one exception is Rockwell's Coyote Station dry scrubber,
which uses soda ash.
Table 7 shows chemical analysis of two reagents tested for SO capture.
The soda ash was a granular, high-quality (99+%), commercial-grade sodium car-
bonate (Na^CO^). This material is a much stronger base than lime and forms
hygroscopic solids, both properties enhancing SO capture in the aqueous- and
dry-phase chemistry. A thiosorbic lime containing 4.8% Mg as MgO was also
tested. Figure 7 indicates total SO capture for soda ash, thiosorbic lime,
and the base Mississippi lime. As expected, the soda ash outperformed both
lime reagents. The thiosorbic lime tests indicate poorer SO capture relative
to the base lime. As noted, the thiosorbic lime supplier reported excess
carbonate core. This was due to low kiln temperatures. While the reported
slaking time is well within the 3-minute-limit slaking time for high-reactiv-
ity lime, these tests should be repeated.
10-72
-------
DC
I
CM
O
(A
100
80
60
40
20
35°F |19°C) AST
1.0 STOICHIOMETRY
O
O
J_
200
250 300 350
SPRAY DRYER INLET TEMPERATURE (°F)
400
Figure 6. Effect of gas inlet temperature on SO2 capture.
Table 7
REAGENT VARIATION CHEMICAL ANALYSIS
Sample No.
Description
Basic
Calcium as CaO, %
Magnesium as MgO, %
Carbonate as CO , %
Total Sulfur as SO %
Total Insolubles, %
(including Si02)
Carbonate as Na CO.,, %
(as-recvd.( calculated)
M-38236
Soda Ash
7-6-83
1500
As-Received
41.46
99.84
M-38237
Thiosorbic
Lime
7-5-83
As-Received
88.35
4.85
0.39
0.34
3.00
Note: Vendor reported thiosorbic shipment contained
3.1% excessive core, 91.9% available lime index.
and 1-minute slaking time.
10-73
-------
100
cc
z>
0.
o
o
(fl
<
o
50
CONDITIONS
> 300DF (149°C)
INLET TEMPERATURE
» 35°F (19.4°C) AST
» 10 SEC GAS
RESIDENCE TIME
O
o
A
REAGENT
SODA ASH
HIGH CALCIUM
LIME
THIOSORBIC
LIME
i.o
2.0
MOLES ALKALI/MOLE SULFUR
Figure 7. Reagent variation testing.
RECYCLE TESTS
The degree of reagent utilization in a dry scrubber system has a signifi-
cant effect on system economics. One significant difference between wet and
dry scrubbing systems is the lower reagent utilization in dry scrubbers. This
is mainly due to type and rate of gas-liquid contact. Wet scrubbers use rela-
tively large liquid-tq-gas ratios, on the order of 40 - 90 gpm/1000 ft gas,
treated (53 - 119 L/M ) to achieve intimate gas-liquid contact with counter-
current absorption trays. Dry scrubbers in FGD have average gas-liquid
ratios of 0.2 - 0.3 gpm/1000 ACFM (0.267 - 0.40 L/M ) and generally use co-
current gas-liquid contact via a gas diffuser and atomizer.
Recycle of solid material collected in the dry scrubber and particulate
collector can improve reagent utilization by reusing unreacted calcium hydrox-
ide and other fly ash alkali. The second source of alkaline material — fly
ash — originates from calcined carbonates of calcium, sodium, and potassium
formed during coal combustion. Figure 8 shows the increase in dry scrubber S09
capture when the feed slurry ash content increases for a 33°F (18°C) and 19°F
(10.6°C) AST. Approximately 7% more sulfur was captured at a 10% wt/wt wet
recycle rate. Samuel, et al. [2] conducted similar tests at substantially
higher recycle rates. Note that his work was at a 1.2 stoichiometry and a
utilization at no recycle of 63%.
10-74
-------
CAPTURE
CM
O
CO
CC
LU
OC
Q
EC
Q.
A
O
D
100
90
80
—
^**
&r-
"t
A-^
60^
50
40
30
20
10
0
0
fA
—
-
-
-
TEST
RUNS
BASE,
175. 176
BASE,
183, 184
-
^^<> r
^1. —
.
A
1 1 1 I
]
10 20 30 40 50
RECYCLE ASH CONTENT IN FEED SLURRY (% WT/WT)
AST INLET INLET
Ca/S (F/C) SO2 (PPM| TEMP. (F/C| SOURCE
1.6 33/18 2200 303/151 B&W
1.6 19/10.6 2150 305/152 B&W
1.2 20/11.1 2000 - REF4
Figure 8. Wet recycle spray of dry scrubber system solids.
Description
Table 8
CHEMICAL ANALYSIS OF RECYCLE TEST SLURRYS
Calcium as Carbonate as Sulfur as
Solids , % CaO, % Dry CO.,
Dry SOV % Dry
Base Lime
Tests 175 & 176
Test 175
Test 176
Base Lime
Tests 183 & 184
Test 183
Test 184
23.9
25.9
24.2
19.6
23.0
20.2
71.1
67.2
69.0
72.4
62.6
65.6
_£
0.29
3.42
2.07
0.46
6.52
4.90
0.04
4.05
2.36
0.05
9.96
4.46
10-75
-------
Test Runs 183 and 184 were at the same AST but, due to the higher
stoichiometry, attained only 47% utilization. The difference in marginal
increase of SO capture based on increases in wet recycle between these two
test series appears to be a function of reagent utilization with no recycle.
Also, at the same SO capture (no recycle), as stoichiometry increases from
1.2 to 1.6, so does Che marginal effect of wet recycle on S02 capture. For
dry scrubber FGD on high-sulfur coals, wet recycle will be essential to
maintain acceptable reagent and fly ash utilization.
Table 8 lists chemical analyses of the base and recycle tests shown in
Figure 8 from this program. Addition of ash to the base slurry is evident by
the increase of total sulfur in forms of calcium sulfates and sulfites and
carbonates, principally calcium carbonate.
LIMESTONE INJECTION
Limestone injection in a multistage burner (LIMB) uses limestone (CaCO )
as reagent introduced into high-thermal-combustion zones, where calcination to
lime (CaO) occurs. The solid lime then adsorbs gaseous S02, forming sulfites
and sulfates of calcium (CaSO and CaSO ). The CaC03 dissociation reaction is
principally a function of temperature and the flue gas C02 partial pressure.
At atmospheric pressure, limestone will begin dissociation under normal coal
combustion conditions at approximately 1450°F (788°C) [7]. Once the C02 gas
has evolved, sulfation may proceed.
Obtaining a high-quality calcined limestone depends primarily on the
temperature profile existing in the combustor and time profiles from the point
of reagent injection. Murray [8] studied calcination efficiency on 43 commer-
cial limestones and concluded each particular limestone has a unique time-
temperature profile for optimum calcination. Dogu [9] investigated initial
limestone pore structure effect on S09 adsorption. The average lime particle
pore radius and overall porosity were measured at various calcining tempera-
tures, with both having an optimum value at 1740°F (950°C). Dogu's studies
also found diffusion resistance from pore closure by CaSO the controlling
mechanism.
Recent work to develop low-NO coal burners has helped limestone injec-
tion studies, since their characteristics of low flame temperature and long
burn time are quite suitable for limestone injection. Dremmel, et al. [10]
studied three types of coal-fired furnaces as part of EPA's simultaneous SO ,
N0x study. Wall-, tangential-, and stoker-fired combustors were used to X
evaluate primarily coal firing rate, cooling rate, and sorbent injection
velocity. A wall-fired furnace simulating the normal time-temperature profile
seen by coal and sorbent particles was fired in a low NO mode. At sorbent-
sulfur ratios of 2.0, 40% to 60% SO capture was measured. Quantitative
effects of time, temperature, and calcination rates were still undetermined.
Figure 9 lists furnace SO capture versus stoichiometry for primary and
secondary limestone injection (also see Table 9). At the 2000°F injection
temperature, primary injection (limestone interspersed with coal) captured
slightly less S02 than secondary injection (13.6% to 13.3%). Within the
accuracy of measuring S02, however, these removals are essentially identical.
10-76
-------
ou
UJ
IT
< 20
O
FURNACE SO2
O
0
D
O
0- xs
/J \/ PRIMARY INJECTION
/AT 2000°F (1093°C)
D SECONDARY INJECTION
AT 1800°F (982°C)
O SECONDARY INJECTION
/v AT 2000°F
I I I I
1234
STOICHIOMETRY
(1093°C)
Figure 9. Limestone injection results.
Table 9
TEST RESULTS: LIMESTONE INJECTION
Injection
Method
Primary
Secondary
DOE Test
Number
D221
D222
D223
D223
D223
D212
D213
D214***
D215
D216
D217
Load
MBtu/hr
4.0
4.0
4.0
4.0
4.0
3.1
2.0
4.0
4.0
4.0
* 0,*
2.9
2.75
2.75
3.0
3.13
3.75
3.73
3.17
3.78
3.79
SR
3.0
2.0
1.0
1.0
1.0
3.0
2.0
3.0
2.0
1.0
Injection
Temp.
2000°F**
2000°F
2000°F
2000°F
2000°F
1800°F
1800°F
2000°F
2000°F
2000°F
% Removal
in Furnace
20.0
13.6
6.7
4.1
6.3
26.0
13.6
22.4
13.3
7.3
*As measured at reactor inlet
**Furnace exit temperature
***Insufficient turndown on screw feeder
10-77
-------
Primary reagent injection exposes the limestone particles to the highest flame
temperatures and longest residence time possible. Sintering of the newly
formed CaO particle pores is a major concern with this type of injection mode.
For secondary injection at varying furnace outlet temperatures, a small in-
crease was also measured in SO capture as the furnace temperature was
lowered. This assumes that an equilibrium between decarbonation and recarbon-
ation had been established. Some partial pore closure by sintering is also
suspected. In general, with a low-NO burner producing 250 - 400 ppm N0x>
approximately 25% SO was captured atXa 3.0 stoichiometry. The inserted lime-
stone was pulverized to approximately 70% through 200 mesh — equivalent in
size to the pulverized coal. A finer limestone grind would have increased
furnace SO capture by producing more solid surface area and minimizing the
pore blinding effects of CaSO formation.
DRY SCRUBBING TECHNOLOGY FOR FGD OF EASTERN COAL
Successful adaption of dry scrubber FGD systems to high-sulfur-coal
applications requires first an understanding of SO capture at various
operating conditions meeting federal New Source Performance Standards (NSPS).
Based on nonrecycle operation, these tests indicated 90% SO capture can be
obtained at a dry scrubber stoichiometry of 1.20 (75% utilization) and an
approximate 15°F AST. At a 1.6 stoichiometry and 20°F AST, 20% solids recycle
increased SO capture from approximately 75% to 90%. Maintaining the higher
20°F AST required recycle and increased lime consumption to meet NSPS.
However, the higher AST will reduce the possibility of a system upset. A dry
scrubber treating 3500 ppm SO flue gas at a 20°F AST and a 1.6 stoichiometry
requires approximately 25% lime slurry. Recycle will increase slurry solids
and is primarily limited by the maximum atomizer solids for acceptable
atomization. For an industrial dry scrubber system, adjustments between AST,
recycle rate, and stoichiometry will define actual process conditions required
to meet NSPS and high availability.
10-78
-------
REFERENCES
1. DOE/METC/SP-194, Topical Report, Advanced Environmental Control
Technology.
2. E. A. Samuel, et al., "Dry FGD Pilot Plant Results: Lime Spray Absorption
for High-Sulfur Coal and Dry Injection of Sodium Compounds for Low-Sulfur
Coal," EPRI-CS-2897, May 1982.
3. J. T. Yeh, et al., "Experimental Evaluation of Spray Dryer Flue Gas
Desulfurization for use with Eastern U.S. Coals," EPRI-CS-2897, May 1982.
4. E. L. Parsons, et al., SO Removal by Dry FGD, Buell Emission Control
Division, Environtech Corp.
5. J. B. Doyle and B. J. Jankura, "Furnace Limestone Injection with Dry
Scrubbing of Exhaust Gases, Babcock & Wilcox," Spring 1982, Central
States Section of the Combustion Institute Technical Meeting.
6. C. A. Johnson, Flyash Alkali Technology-Low Cost Flue Gas FGD, Peabody
Process Systems Inc.
7. R. S. Boynton, Chemistry and Technology of Lime and Limestone, John Wiley
and Sons, 1980.
8. J. A. Murray, et al., Journal American Ceramics Society. 37, No. 7,
323-328, 1954.
9. T. Dogu, "The Importance of Pore Structure and Diffusion in the Kinetics
of Gas-Solid Noncatalytic Reactions — Reaction of Calcined Limestone
with SO , Chemical Engineering Journal, 21, 213-222, 1981.
10. D. C. Dremmel, et al., "SO Control with Limestone in Low-NO Systems:
Development Status," EPRI-CS-2897, May 1982. X
10-79
-------
EPRI SPRAY DRYER/BAGHOUSE PILOT PLANT STATUS
AND RESULTS
G. M. Blythe, R. G. Rhudy
-------
EPRI SPRAY DRYER/BAGHOUSE PILOT PLANT STATUS AND RESULTS
Gary M. Blythe
Radian Corporation
Austin, Texas 78766
Richard G. Rhudy
Electric Power Research Institute
Palo Alto, California 94303
ABSTRACT
In February 1982, the Electric Power Research Institute (EPRI) initiated
a 2-1/2 MW spray dryer/baghouse FGD pilot plant program at their Arapahoe test
facility. The objective of the pilot plant program is to confirm the capabil-
ities of the FGD process and to provide the electric utility industry with
reliable design and operating information for spray dryer/baghouse FGD sys-
tems. The pilot unit was described and initial results for sodium carbonate
and once-through lime operation were presented at the May 1982 FGD symposium
in Hollywood, Florida. This paper presents the results of test work conducted
from May 1982 through August 1983.
The majority of the test work has been conducted with lime reagent in the
recycle, rather than once-through mode. Effects of a number of variables have
been studied. Spray dryer inlet SC>2 concentrations have been varied from a
nominal 350 ppm up to 2000 ppm. Other variables examined have included re-
agent ratio, recycle rate, system flue gas flow rate, atomizer feed slurry
preparation and feeding configurations, and approach to adiabatic saturation
at the dryer outlet. A significant result has been the observation that re-
cycle operation greatly improves spray dryer operation in addition to im-
proving S02 removal performance.
The fabric filter has been shown to contribute significantly to overall
system S02 removal, particularly at higher system removal levels (80 percent
and greater). No bag/fabric-related problems have been observed. However,
corrosion of mild steel baghouse walls and mild steel caps on bags near the
walls has in some instances been severe. The corrosion has been largely
attributed to insufficient insulation of baghouse surfaces, and to the fact
that the pilot-scale compartment shares no common walls with other compart-
ments. As a result of several bag cap failures, the fabric filter compartment
was re-bagged in April 1983. The new bags were brought on-line with no condi-
tioning by fly ash-only operation, and after 4 months continue to operate at a
very low pressure drop.
10-81
-------
INTRODUCTION
Currently, spray dryer based flue gas desulfurization (FGD) systems for
16 utility boilers with a total capacity of over 6000 MW have been installed
or are under contract. Six of the systems are in start-up or commercial
operation. However, of the two systems that have been in operation longest,
one is unique in using soda ash reagent, while the other has been operated by
the vendor until only recently. Thus, in spite of the large number of units
sold, very little information on the design and operation of these systems is
available to the utility industry.
Because of this lack of information, the Electric Power Research Insti-
tute has chosen to pilot a spray dryer/fabric filter system at their test
facility adjacent to the Public Service of Colorado Arapahoe station. Numer-
ous pilot-scale programs have been conducted by process vendors, but detailed
results from these programs have generally not been published. The goals of
the EPRI program are to independently assess the operability of a spray dryer
and fabric filter system in FGD service, and to provide design data applicable
to typical utility installation. Emphasis has been placed on low to medium
sulfur operation, with S02 removal efficiencies from 70 to 90 percent. Most
test work has been in the operating mode where sorbent and fly ash have been
recycled from the fabric filter back to the spray dryer, as is typical of many
utility systems.
The pilot unit began operation in March 1982. This paper summarizes
results from May 1982 through August 1983. Work during this period has in-
cluded both once-through and recycle operation, inlet S02 concentrations of
400 to 2000 ppm, various modes of recycle, normal (270°F*) and low (210°F)
inlet flue gas temperature, and varied makeup water compositions. In the fol-
lowing sections, the pilot unit is described, results are presented regarding
the operability of the spray dryer and fabric filter, S02 removal results are
presented, and the conclusions available as of August 1983 are summarized.
DESCRIPTION OF PILOT UNIT
The spray dryer unit at Arapahoe consists of a spray dryer vessel, one
compartment of a fabric filter, a reagent preparation system, and reagent
feed/recycle systems, as well as a comprehensive instrumentation system.
Figure 1 is a simplified process flow diagram for the pilot unit as set up for
rotary atomization.
^ spray dryer vessel has a 10-foot diameter, 5.5 foot straight side and
a 50° cone bottom. The vessel has interchangeable flue gas inlet vane rings
(one for normal flow rates and one for low flow rates), can accommodate rotary
or nozzle atomization, and can be rearranged for either bottom or side gas
exit. An additional straight side section which doubles the volume of the
dryer vessel can be installed to facilitate nozzle atomization tests. At
^British Engineering Units rather than SI units are used in this paper because
of customary usage in the electric power industry. An appendix provides
appropriate conversion factors.
10-82
-------
normal flue gas rates (6500 to 9000 acfm dryer outlet gas rate), the vessel
flue gas residence time without the added section is approximately 5 to 7
seconds. Thus, the vessel is quite versatile, allowing a number of potential
configurations. This versatility was desired to provide the capability to
include aspects of various vendors' designs in the test program.
TEMPERATURE
CONTROL
WATER I
FLUE GAS
FROM BOILER ^T
FLUE GAS BYPASS AROUND DRYER
CLEAN
FLUE GAS
DETENTION
SLAKER
NOT SHOWN
LIME DILUTION
TANK
Figure 1.
Simplified Process Flow Diagram for Spray Dryer
Pilot Plant (Rotary Atomizer Configuration)
The fabric filter used for this spray dryer pilot plant is one compart-
ment of a 10-MW equivalent four-compartment pilot unit which had previously
been undergoing characterization for over a year. The fabric filter is also
quite versatile in design, allowing a choice of bag diameters and bag lengths.
The compartment is currently equipped with 36 utility-size (12 in. by 34 ft)
glass fiber bags (tri-coat finish) in a 6 x 6 array. At the normal 6500 to
9000 acfm fabric filter module inlet flue gas rate, the air-to-cloth ratio
varies from 1.7 to 2.3 ft/min. The unit has provisions for either reverse
gas, shake/deflate, or combined cleaning. Each module, including the one used
10-83
-------
for the spray dryer pilot unit, is controlled by a microprocessor. The micro
processor enables full variation in cleaning cycle duration, cleaning intens-
ity, null period durations, damper opening and closing rate, etc. So far in
the program, the fabric filter has been operated on a 3-hour cycle, using
reverse gas cleaning.
A recent modification to the pilot unit allows bypassing warm gas around
the spray dryer to reheat the gas entering the fabric filter. The controls on
this bypass line will allow either a constant reheat temperature or a constant
bypass flow.
The reagent preparation area contains ball mill, paste-type, and deten-
tion slakers. Additionally, provisions are available to mix simulated cooling
tower blowdown waters for use as slaking and/or dilution water. Lime slurry
used by the spray dryer has primarily been slaked in the paste slaker, al-
though some ball mill- and detention-slaked lime has been used.
Slurry is fed to the atomizer using progressing cavity pumps from one of
two covered, well agitated tanks. Temperature control water is added at the
atomizer. The system has the capability of recycling fly ash/spent sorbent
from the spray dryer bottoms catch (during side flue gas exit operation) and/
or from the fabric filter hopper each. Solids from each location can be pneu-
matically transferred to separate recycle bins. Recycle material can either
be slurried and fed from a separate recycle tank, or can be slurried in the
main slurry feed tank along with fresh lime. The feed rates of recycle mate-
rial and makeup water to either tank are ratioed to the rate at which fresh
lime slurry is added to the slurry feed tank. A separate pump and controller
are available to meter recycle material to the atomizer separate from the lime
slurry if the two are not slurried together in the same tank.
The instrumentation on the pilot unit was discussed in detail in a pre-
vious paper, "EPRI Spray Dryer Pilot Plant Status and Results,"1 given at the
EPA/EPRI FGD Symposium in Hollywood, Florida, in May 1982.
OPERATION RESULTS
The EPRI pilot unit has provided insight into the operation of the equip-
ment used in spray dryer/baghouse systems. Many improvements in pumps, tanks,
agitators, etc., have been made since start-up to improve the operation of the
pilot unit. However, most of these changes were necessary because of the
small scale of this pilot plant and are not directly applicable to utility
systems. However, observations about the operation and maintenance of the
spray dryer and fabric filter are more generic in nature and may be applicable
to most spray dryer/baghouse FGD systems. The spray dryer and fabric filter
operation results are discussed in the following section.
SPRAY DRYER OPERATION
Operation of the system in the once-through mode, with lime reagent, for
low inlet S02 levels (300 to 400 ppm) has been the most troublesome configura-
tion tested to date. After an initial check-out of the system with soda ash
reagent and a spinning disc atomizer wheel, the first lime slurry tests were
10-84
-------
conducted in the once-through mode. The results of soda ash testing and early
once-through lime tests were reported in the previously referenced FGD sym-
posium paper. The problems encountered after switching to lime slurry were
multiple. First, the pilot spray dryer had been designed for a flue gas flow
rate that resulted in a residence time of approximately 5 seconds. When try-
ing to operate at a 20°F approach to adiabatic saturation at the dryer outlet,
as many utility systems sold to date will, a continued buildup of wet solids
on the dryer walls was encountered. This buildup caused several problems.
First, wet solids from the walls would intermittently fall to the bottom of
the dryer. The dryer is equipped for pneumatic conveying of solids from the
bottom of the dryer, but these wet solids (10 percent moisture and greater)
tended to plug the pneumatic lines. Once the pneumatic lines plugged, in
spite of efforts to quickly clear the pluggage, wet solids would begin to
bridge across the cone bottom on the dryer. This bridge of wet solids would
continue to grow until the flue gas exit duct from the dryer was almost
plugged. At this point, the dryer would have to be shut down, the rotary
valve and pneumatic line removed from the bottom of the cone, and the vessel
would have to be washed out. The problems of drop out of wet solids on the
walls of the dryer were substantially reduced, however, when operating a lower
flue gas flow rate (7 to 8 second residence time) or at a 30°F or higher
approach to saturation at the dryer outlet.
The second major problem during the once-through lime tests was related
to the atomizer rather than the vessel itself. For lime reagent tests, a
nozzle insert-type wheel normally used for abrasive slurry service has been
used. During the first several weeks of operation, the spray machine was
taken off-line several times due to high vibration. Invariably, when the unit
was removed for inspection, it was found that several of the twelve nozzles
around the periphery of the wheel were plugged with lime solids. The pluggage
problem was lessened but not solved by increasing the wheel speed from 12,000
to 15,000 rpm. It became apparent that the pluggage of nozzles with lime
solids was caused by flue gas pumping through the wheel. The 9-inch diameter
wheel with twelve 3/8-inch openings had a hydraulic capacity much greater than
the nominal 2 gallons per minute of slurry normally fed at these flue gas
rates. Because it had a greatly increased hydraulic capacity, it would educt
hot flue gas across the top of the wheel to be pumped through the nozzles.
Because of this flue gas pumping, lime and/or fly ash solids were deposited on
the top side of the wheel, and lime solids tended to dry within the wheel,
particularly within nozzle openings. An attempt was made to fit a seal
between the spray machine and the wheel in order to restrict this pumping
effect. However, the pumping continued, although perhaps at a reduced rate.
As flue gas was pumped through this low clearance area, lime and/or fly ash
deposited on the top of the wheel caused the seal on the bottom of the spray
machine to rub the wheel. The resulting imbalances led to early failure of
the lower bearing on the machine. An alternate approach, then, was to reduce
the number of active nozzles in the wheel from twelve to six. The other six
nozzles were replaced with solid inserts. Although the tendency for the
nozzles to plug was reduced, on numerous occasions, three of the six nozzles
would plug. If any two adjacent nozzles plugged, an imbalance would lead to
high vibration measurements and cause the operators to shut the spray machine
down. However, if every other of the six nozzles plugged, vibration problems
were not encountered. this led to a further reduction in active nozzles to
10-85
-------
only three, where now nine of the original twelve nozzles were replaced with
solid inserts. While this last modification solved the nozzle pluggage prob-
lem, it tended to aggrevate the previously discussed problem with wet product
on the dryer walls. Empirical correlations in Masters2 for predicting the
droplet size produced by a vaned atomizer wheel suggest that for a nozzle
wheel with a given nozzle diameter, increasing the number of nozzles around
the periphery of the wheel will produce a finer droplet size. Based on this
correlation, it would seem that in reducing the number of active nozzles in
our atomizer wheel, we produce a coarser droplet size which is not as readily
dried. This no doubt contributes to the buildup of wet product on the spray
dryer chamber walls.
After once-through operation, lime reagent recycle tests were begun. In
most recycle tests conducted thus far in the program, recycle material has
been taken from the fabric filter hopper, and mixed with fresh lime and makeup
water in the main slurry feed tank. This fresh lime/recycle slurry is nor-
mally fed to the atomizer by a single pump. Recycle ratios are expressed in
this paper as pounds of recycle material per pound of hydrated lime (CaCOH^)
solids in the fresh lime feed. The lowest recycle ratios tested were around
2:1, but even at this level of recycle an immediate improvement in spray dryer
operation resulted. With recycle, the number of active nozzles in the atom-
izer wheel could be increased back to six from the three required for once-
through operation. This increase in active nozzles should result in somewhat
finer atomization, further promoting improved drying.
Upon the implementation of recycle, problems with buildup of solids
within the atomizer wheel, buildup of wet solids on the spray dryer chamber
walls, and plugging of the pneumatic transfer line from the spray dryer bottom
have also been virtually eliminated. The dryer can be successfully operated
over a long term at a 20°F approach, at a flue gas flow rate resulting in a
7-second dryer residence time. At recycle ratios near 12:1, resulting in a
solids level at the atomizer wheel of 35 to 45 percent, the dryer has been
successfully operated for days at a time at a 20°F approach at residence times
as low as 5 seconds.
It appears that much of the improvement in spray dryer operability in
recycle operation is related to having an increased weight percent solids
level in the slurry fed to the atomizer. This effect has become apparent in
recent once-through lime tests conducted at 1000 ppm inlet SC>2 levels. At
1000 ppm inlet levels, roughly three times the amount of lime is fed compared
to low sulfur operation, at equivalent reagent ratios. This solids level,
then, is equivalent to the solids level at the same reagent ratio for 350 ppm
inlet SC>2 and 2:1 recycle ratio operation.
In the baseline 2:1 recycle tests, the weight percent solids in the
slurry atomized ranged from 6 to 16 percent. Similarly for the 1000 ppm inlet
S02 once-through lime tests, the weight percent solids level to the wheel
varied from 12 to 16 percent. In these once-through tests at a nominal
7-second dryer residence time, problems of wheel pluggage with the six nozzle
atomizer wheel, and pluggage of the drying chamber with wet solids were not
encountered. These successful once-through lime tests at a 1000 ppm inlet S02
10-86
-------
level seem to indicate that at least part of the improved spray dryer oper-
ability during recycle operation comes from the increased solids level in the
slurry atomized rather than from the presence of the recycled material per se.
Further evidence of this weight percent solids effect on operability is
seen for the high recycle rate tests. For the nominal 12:1 recycle tests the
percent solids to the wheel ranged from 31 to 37 percent. Operability of the
spray dryer was improved over that even for nominal 6:1 recycle ratios. At an
inlet flue gas flow rate of 5800 scfm (5.0-second residence time in dryer)
when recycling to a nominal 35 weight percent solids level to the wheel, no
tendency to drop out wet solids in the dryer or plugging of pneumatic transfer
lines was noted. However, some tendencies for wet product formation had been
seen when operating at this flue gas rate at lower recycle ratios (hence lower
weight percent solids to the wheel).
The effects of operating condition on the weight percent solids in the
slurry fed to the atomizer are summarized in Table 1. The table shows that
once-through operation at low inlet SC>2 concentration results in the lowest
solids content in the slurry fed to the atomizer. It was at these conditions
that problems with wet spray dryer products persisted. Other operating modes
resulting in higher solids content in the slurry feed (recycle, and/or higher
inlet S02 concentration) greatly improve the spray dryer operation.
TABLE 1. SUMMARY OF WEIGHT PERCENT SOLIDS IN SLURRY FED TO ATOMIZER WHEEL
AT VARIOUS OPERATING CONDITIONS
Operating Mode
Once-through
Recycle (2:1)
Recycle (6:1 to £
Recycle (12:1)
Once-through
Recycle (2:1)
Recycle (2.5:1)*
Nominal
Inlet
S02 (ppm)
350
350
5:1) 350
350
1000
1000
1000
Weight
Percent Solids
Atomizer Wheel
3 to 7
6 to 16
17 to 24
31 to 37
12 to 16
20 to 34
25*
Vessel
Residence
Time (sec)
7
7
7
5
7
7
5
Nature
of S.D.
Product
Wet
Dry
Dry
Dry
Dry
Dry
Dry
*0ne test.
Theoretical considerations presented by Masters^ would tend to support
these observations of improved operability at higher feed solids contents.
Masters describes the first two phases of drying of droplets in a spray dryer
as the "constant rate" period (the first phase of drying) and the "first
falling rate" period (the second phase of drying). In the first phase, or
constant rate period, the surface of the droplet behaves as though no solids
are contained, and water evaporates freely from the surface. In the first
falling rate period (second phase of drying), solids protrude from the surface
of the droplet, and the rate of evaporation of moisture from the droplet is
slowed by the rate at which water diffuses through the network of solid par-
ticles.
10-87
-------
It is important that droplets atomized are beyond the constant rate, or
first drying period before they come into contact with vessel walls. The
length of the first drying period is labeled as the critical drying time. In
general, problems with wet product drop out in the dryer can be avoided by
decreasing this critical drying time for the droplets atomized. Empirical
correlations presented in Masters allow one to predict the impact of variables
such as solids content on critical drying time. The results of example cal-
culations are summarized in Table 2. These example results show that for a 50-
micron droplet, the critical drying time is reduced by one-half if the solids
content is raised from 5 to 30 percent. Calculations for 100-micron droplets
show that critical drying times are 3 to 4 times those for the 50-micron drop-
lets, but that the higher solids level again nearly reduces the critical dry-
ing time by one half. Of course a higher solids content in the slurry feed to
an atomizer will also impact the atomized slurry particle size distribution to
a minor degree, but it is clear from these example calculations that the net
effect of increased solids content in the slurry feed is reduced critical dry-
ing time, hence potentially improved dryer operation. Also in Table 2 are
falling rate drying times and total drying times for each of the four cases.
The falling rate drying times and total drying times actually increase at
higher weight percent solids levels. This increase may account for some of
the improved SC>2 removal performance seen in recycle operation. Please note
that the values presented in Table 2 were calculated for example only, and do
not represent an attempt to accurately predict the critical and/or total dry-
ing times in the Arapahoe spray dryer.
TABLE 2. EXAMPLE DRYING TIMES FOR ATOMIZED DROPLETS
Case 1 Case 2 Case 3 Case 4
Droplet Size, y
Weight Percent Solids
Critical Drying Time, sec
Falling Rate Drying Time, sec
Total Drying Time, sec
50
5
0.4
0.1
0.5
50
30
0. 2
0.8
1.0
100
5
1.4
0.4
1.8
100
30
0.8
2.0
2.8
BAGHOUSE OPERATION
Throughout over 6000 hours of spray dryer operation, no bag fabric prob-
lems have been observed. This result is significant, because the operating
time includes both sodium carbonate and lime reagent, once-through and recycle
operation, low sulfur to medium sulfur inlet S02 levels, a great number of
start-ups and shutdowns, and several spray dryer upset conditions. Spray
dryer upsets which one would have expected to have a detrimental impact on bag
performance did not seem to affect the bags at all. In one incident, a new
operator flushed a slurry feed line into the spray dryer through the atomizer,
driving the dryer outlet temperature to the adiabatic saturation temperature.
In spite of such abuse, no bag failures have occurred, outlet opacity remains
quite low, and tube sheet pressure drop values have been quite acceptable.
Although there have been no bag fabric problems due to operation down-
stream of the spray dryer per se, there has been a problem with corrosion on
10-88
-------
the baghouse walls. The problem did not become apparent until after the spray
dryer had been in operation for approximately 8 months, in November 1982.
During an outage, the baghouse module was opened and large flakes of rust
scale were noted to be forming along the inside walls. The corrosion was
attributed to moisture condensation on compartment walls. Several factors may
be contributing to this condensation and corrosion problem. First, the com-
partment is part of a pilot unit, so it goes through numerous start-ups and
shutdowns, passing through the dewpoint of the flue gas each time. In addi-
tion, since the compartment shares no common walls with other compartments,
all four walls of the compartment are exposed to ambient temperatures. Also,
the baghouse was originally designed to clean hot gas (250°F to 300°F) rather
than gas which is approaching adiabatic saturation. Consequently, the bag-
house walls are insulated with only a 2-inch thickness of fiberglass batts.
At the spray dryer outlet, the flue gas is generally 17 to 22°F from adiabatic
saturation. This corresponds to roughly 20 to 30°F from the dewpoint. As the
flue gas is cooled in the outlet ductwork and in the baghouse, the bulk gas
temperature may come within 10 to 15° of its dewpoint. Any signficant cooling
of the baghouse walls, then, may put the skin temperature at or below the
dewpoint of the flue gas. This effect would particularly be noticeable in
very cold weather, approaching 0°F, where the driving force for cooling the
walls would exceed 100°F. The explanation that the corrosion is caused by
localized cold spots is supported by the observation that corrosion is worst
near an entry door to the compartment, where not only is the wall insulation
sparse, but where cold air can leak into the compartment.
Sulfuric acid condensation is not thought to directly contribute to the
corrosion problem, because no measurable 803 content has been found in the
flue gas downstream of the filter bags. However, S02 present in the outlet
flue gas would be sorbed into any condensation which forms. Subsequent liquid
phase oxidation reactions would tend to form sulfuric acid in the condensed
moisture. Sulfuric acid-laden moisture then would not readily evaporate be-
cause of the azeotropic nature of the sulfuric acid-water mixtures.
A later outage, after approximately 10 months of operation, showed con-
tinued corrosion. Additionally, during this outage the top side of the tube
sheet was found to be covered to a depth of several inches with wet ash and
flakes of corrosion. The moisture on the tube sheet was attributed to the
severe spray dryer upsets mentioned previously, where an inexperienced opera-
tor repeatedly overloaded the atomizer with water while flushing a line. The
ash on the tube sheet had been observed during a previous outage. It is not
clear when this dry ash was deposited. The top of the tube sheet was cleaned
and the unit operators were reminded of proper line flushing procedures which
call for blocking out flow to the atomizer.
During this outage, thermocouples were installed on two wall surfaces
within the fabric filter compartment and continuously monitored. In colder
weather, it was observed that the temperatures were generally several degrees
below the bulk gas temperature. In April 1983, the ambient temperature over-
night in Denver dropped to 9°F. One of the two wall temperatures dropped to
96°F for over an hour. This is several degrees below the dewpoint of the bulk
gas at the fabric filter outlet. The pilot unit was shut down and the com-
partment was opened. As expected, moisture was found on compartment walls and
10-89
-------
on top of the tube sheet. More startling was the discovery of three bags
collapsed onto the tube sheet. The bags had fallen due to corrosion failure
of the mild steel bag caps. These bag cap failures were not thought to be the
result of this one low temperature excursion, but a cumulative effect of 13
months of operation downstream of the spray dryer and a particularly cold
winter.
The bags were all removed and replaced, and the removed bags were exam-
ined. A map of bag cap corrosion was prepared, and is presented in Figure 2.
In general, the caps showing the worst corrosion were around the edges. Of
the three failed bags, two were in corners and one was near the compartment
doors. Top and bottom portions from twenty of the bags were submitted to
Albany International for routine performance evaluation. No failures or areas
of potential failure were observed on any of the submitted samples except a
top sample from one of the fallen bags. This top sample contained two small
holes in an area which was heavily stained with rust. The damage likely
occurred when the bag fell.
FLUE GAS
INLET ""^
• O
® O
® o
(X O
0 0
o
o
0
o
o
® ft <»
O O
-------
FLUE GAS
OUT
FLUE
GAS
IN
Figure 3. Cutaway View of Thermocouple Locations
in Baghouse Compartment
10-91
-------
filter enclosure, while the west and north walls face the inside of the en-
closure. Table 3 summarizes selected temperature measurements since the pilot
unit was put back in service. Case 3 in the table includes the two wall mea-
surements previously recorded for a 9°F ambient temperature. Note that in
Case 1 and Case 2, no temperatures are actually below the bulk gas dewpoint,
but for Case 2 (36°F ambient) one of the wall temperatures is approaching that
value. Also note that the 15 temperature which reached 96°F in Case 3 is not
even the coolest temperature in the other two cases. Monitoring of all ten
locations during the upcoming winter is expected to identify a number of ex-
cursions of wall and cap temperatures below the bulk gas dewpoint.
TABLE 3. AMBIENT TEMPERATURE EFFECTS ON BAGHOUSE TEMPERATURES
Temperature (°
Ambient
S.D. Outlet
Baghouse Outlet
Approximate Dewpoint
(F.F. Outlet)
TI (Top Interior Wall)
15 (Bottom Interior
Tg (Bottom Exterior
Tg (Bag 16-Cap)
TIQ (Bag 36-Cap)
Wall)
Wall)
Case 1
82
138
132
108
117
125
125
136
133
Case 2
36
135
124
104
107
113
117
130
124
F)
Case 3
9
137
125
104
-
96
109
-
—
These results certainly indicate that full-scale baghouses downstream of
spray dryers will have to be designed with ample insulation on exterior walls,
and should be insulated between compartments to avoid cold walls when adjacent
compartments are off-line. Also, compartment doors must be well sealed, and
penetrations of supports through the insulation and welding directly to com-
partment walls should be avoided.
Replacement of the bags in the compartment downstream of the spray dryer
has allowed a number of "before" and "after" comparisons. Since this compart-
ment is one of four in the same baghouse, a control compartment which treats
the same flue gas without the spray dryer upstream is also available for com-
parison.
The first comparison (Figure 4) shows the pressure drop across Compart-
ment A (downstream of the spray dryer) compared to that for Compartment C
(control, fly ash only) during the month of January 1983. In this plot, the
original bags were still in service downstream of the spray dryer. The lower
pressure drop across Compartment A is expected considering Compartment A was
operated at a lower filtering air-to-cloth ratio of 1.7 ft/minute compared to
2.0 ft/minute in Compartment C. This result is somewhat surprising though,
considering that Compartment A is removing 2 to 3 times the quantity of partic-
ulate matter downstream of the spray dryer than Compartment C collects from
the untreated flue gas.
10-92
-------
* 10
i 9
i' •
a
ui 7
oc
6
4
3
2
1
0
I I I I I I I I I I I I I I
COMPARTMENT C
COMPARTMENT Ax
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
JANUARY 1B83
Figure 4. Average Tube Sheet Pressure Drop History -
Compartments A and C - January 1983
The bags in Compartment A were originally filtering fly ash only for
approximately 5 to 6 months prior to start-up of the spray dryer (July 1981 -
January 1982), thus establishing a residual fly ash cake on the bags. When
new bags were placed in service after the corrosion problem in April 1983, the
spray dryer was started up immediately; no conditioning on fly ash-only was
conducted. In Figure 5 a pressure drop comparison of Compartments A and C, 4
months after the new bags were installed in Compartment A, for July 1983,
shows a substantially reduced pressure drop across Compartment A relative to
that of Compartment C. When the difference in A/C ratios is accounted for,
the Compartment A Ap is still significantly lower than that for Compartment C.
The difference is explained by comparing the period of operation between the
two compartments. The Compartment C bags have been in operation for over two
years and have reached a steady-state pressure drop and residual dust cake
weight (50-60 Ibs). The Compartment A bags, after operating only 4 months,
have a very low pressure drop and only about 15 Ib of residual dust cake.
Whether the new Compartment A bags will eventually reach a Ap and residual
cake weight similar to Compartment C remains to be seen. However, short-term
operation without fly ash conditioning has not proven detrimental. In fact,
the AP across Compartment A has risen very slowly and does not appear to be
substantially increasing with time. This effect is shown in Figure 6.
i—i—i—i—i—i—i—i—i—i—i—r~ir
1C
D
Uj a)
X "g
W c
UJ ~.
co a.
D O
I- or
uj O
EC
UJ
COMPARTMENTC
COMPARTMENT A
I I I I I I I I I I I I
I
I I I I
J_
I I I I
1 2 34 5 67 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
JULY, 1983
Figure 5.
Average Tube Sheet Pressure Drop History
Compartments A and C - July 1983
10-93
-------
1.7:1 A/C Ratio, April through July 1983
Q O
4-
_
O o
M- C
o-
c
(0
CD
2.3:1 A/C Ratio
D
a a
QDDD°
DaaOQD D
DDDaD 0
QDU '-'--'-' Q
D aO
D
a
a
30 60
Days since Startup
90
Figure 6.
Pressure Drop Profile of a Recent Start-Up of Compartment A
Downstream of Spray Dryer
Several important findings have resulted from the Arapahoe pilot baghouse
operation downstream of the spray dryer. These include:
1) Operation downstream of a spray dryer does not appear to have
an adverse affect on the bag fabric, even when subjected to
occasional upset conditions at or very near adiabatic satura-
tion;
2) A baghouse downstream of a spray dryer operated at close
approach to adiabatic saturation must be well insulated, with
close attention paid to installation details in order to avoid
cold spots which lead to localized corrosion; and
3) A baghouse downstream of a spray dryer may operate at substan-
tially lower pressure drop than the same baghouse treating
normal fly ash-laden flue gas, particularly if the bags are not
conditioned by any lengthy period of fly ash-only operation.
S02 REMOVAL RESULTS
The S02 removal results presented here concentrate on the recycle lime
reagent operating mode. This mode is typical of the majority of utility spray
dryer/fabric filter dry FGD systems sold to date, and represents the most
operable configuration for this pilot unit. The recycle data presented here
all represent recycling solids from the fabric filter.
10-94
-------
In most of the data plots presented in this section, S02 removals across
the overall system, across the spray dryer alone, and across the fabric filter
are plotted for each test. S02 removal across the fabric filter in these
plots is calculated by the following equation:
S02 Removal in = (SC>2 Spray dryer outlet) - (SC>2 Fabric filter outlet) x IQQ%
Fabric Filter (SC>2 Spray dryer inlet)
By calculating SC>2 removal across the fabric filter in this manner, the SC>2
removal across the spray dryer and across the fabric filter can be directly
summed to get the overall system removal. The SC>2 removal values for the
fabric filter presented here represent time averaged values, which have a
number of 3-hour cleaning cycles included in the averaging time. This time
averaged value is thought to coincide well with the instantaneous values one
would encounter across a multicompartment baghouse, where perhaps ten compart-
ments would be evenly distributed with respect to time throughout their clean-
ing cycles.
Each data point for the recycle tests represents the mean value for
approximately 24 to 48 hours of steady-state operation at those conditions.
Steady-state operation is normally achieved in approximately 24 hours. For
the system to be at steady state, the bags in Compartment A must have been
cleaned, the Compartment A hopper emptied, the recycle material holding bin
refilled, and the combined lime/recycle slurry feed tank refilled at least two
to three times each. When steady state is achieved, the chemical composition
of the slurry fed to the atomizer and of the material collected on the bags in
Compartment A no longer change significantly from batch to batch.
Baseline SC>2 removal results for recycle operation at normal Arapahoe
inlet S02 concentrations around 350 to 400 ppm are presented in Figure 7.
These results represent an approach to adiabatic saturation at the dryer out-
let within 20°F, and a recycle ratio of 2:1. The recycle ratio was defined
earlier as:
Ib of material recycle (dry basis)
Recycle ratio = ^ of fresh Ca(OH)2 makeup (dry basis)
The results in Figure 7 show that at low reagent ratio values (below 0.8)
fabric filter SC-2 removal contributes little to overall SC>2 removal, because
lime reagent utilization in the spray dryer approaches 100 percent. At higher
reagent ratio values (1.3 and greater) fabric filter SC>2 removal contributes
greatly to system S(>2 removal, because S02 removal in the spray dryer does not
increase with any significance as the reagent ratio is increased. In this
paper, unless otherwise noted, reagent ratio is defined as:
Ib-mol fresh lime fed to spray dryer
Reagent Ratio = ib-mol S02 in inlet flue gas to spray dryer
10-95
-------
80-
^ 60-
03
O
E
03
oc
C\J
0 40-
C/J
20-
0-
/ .-o--°
I/
Cp __ 3
i n x id
^
! LEGEND
/
1 O Overall SO2 Removal
/ n Spray Dryer Removal
/ A Fabric Filter Removal
/
/
/
/
/ f__
/ ^A'"'"
/ ^^
1 A'
/
i i
0 0.5 1.0 1.5 2.
Reagent Ratio
Figure 7. Baseline S02 Removal for 400 ppm Inlet
20° Approach, 2:1 Recycle Ratio
Once baseline performance curves for moderate recycle rates were deter-
mined a series of tests at higher recycle rates was conducted. These results
are plotted in Figure 8. Recycle at ratios of approximately 6:1 to 8:1 did
not appear to demonstrate any marked improvement in SC>2 removal over that for
recycle ratios around 2:1. Recycle ratios of approximately 12:1 did show a
marked improvement, primarily in SC>2 removal by the spray dryer. These
highest recycle ratios correspond to solid contents of 30 percent or better in
the slurry fed to the atomizer. The fabric filter contribution to overall
removal does not appear to be affected by the increased recycle ratio. The
upper curves in Figure 8 represent a visual fit of the SC>2 removal data across
the spray dryer during 12:1 recycle operation. These upper curves show virt-
ually complete lime utilization up to a reagent ratio of 0.9.
10-96
-------
100-
80-
- 60-
co
o
E
01
cc
20-
Nominal
Recycle Ratio
6:1 12:1
O •
D •
A A
Overall SO2 Removal
Spray Dryer Removal
Fabric Filter Removal
0.5
i i
1.0 1.5
Reagent Ratio
2.0
Figure 8.
S0? Removal Results for Recycle Tests
20 F Approach Temperature
Mention was made earlier in this paper that operation of the spray
dryer/fabric filter FGD system is improved by recycle versus once-through lime
operation. Both spray dryer operability and S0« removal improvements have
been noted. The results of once-through lime tests are compared to the 2:1
recycle, 20°F approach, 350 to 400 ppm inlet S0_ level baseline test results
in Figure 9. Because of the repeated operability problems during once-through
operation, many once-through S0_ removal test results have been discarded.
The results in Figure 9 represent a number of successful short-term once-
through tests. The open symbols represent once-through results at a 20°F
approach to saturation at normal Arapahoe inlet S0« levels. The dashed curves
are the visual best fit of the 2:1 recycle, 20°F approach baseline results.
The figure shows that at low reagent ratio (below 1.0), the difference in S0«
removal performance between once-through and recycle operation is difficult to
quantify because of scatter in the results of only two successful short-term
tests. The results at higher reagent ratios (nominally 1.5) show that once-
through operation gives approximately 10 to 15 percent lower S0_ removal than
for recycle operation. This difference is primarily due to a decrease in
spray dryer rather than fabric filter SO- removal performance during once-
through operation. It may have been that more successful once-through tests
were completed at higher reagent ratios because of the beneficial impacts on
spray dryer operability of higher weight percent solids in the lime feed for
these tests.
10-97
-------
100-
80-
60-
0
E
CD
o
CO
40-
20-
LEGEND
•Overall SO2 Removal
• Spray Dryer Removal
A. Fabric Filter Removal
0.5 1.0 1.5
Reagent Ratio
2.0
Figure 9. Comparison of Once-Through S0~ Removal Points
with Best Fit Curves from Recycle Results
for 400 ppm Inlet SO Levels
With the performance of the spray dryer/fabric filter system well charac-
terized for a nominal 350 to 400 ppm inlet S0« level, an extensive series of
tests at a 1000 ppm inlet SO level was begun. Testing at this inlet SO-
level was of interest because it simulates performance of the system on medium
sulfur coals (1 to 2 percent sulfur). Also, it was felt that operation at
higher inlet SO levels would make subtle variable effects on SO- removal per-
formance more pronounced.
Figure 10 summarizes the results of a number of tests conducted at a 1000
ppm inlet SO level, with a 2:1 to 4:1 recycle ratio and a 20°F approach to
adiabatic saturation. Also plotted is a "best fit" line for the 400 ppm
results at similar conditions. The majority of the test results are in the 80
to 90 percent S02 removal range, which is what current utility boiler New
Source Performance Standards would require for most medium sulfur content
coals. The figure shows that at moderate overall SO removal levels (70 to 85
percent), somewhat lower S02 removals result compared to what would result at
an equivalent reagent ratio in the 400 ppm tests. At around 90 percent S0?
removal, the overall SO removal curves tend to intersect, with a reagent
ratio of approximately 1.3 required to achieve 90 percent removal in either
case. The shape of the overall performance curves in each case appears to be
set by spray dryer removal performance. The fabric filter contribution data
10-98
-------
appear to plot on one line regardless of inlet SC>2 concentration. The inter-
section of the two spray dryer SCL removal curves appears to be the result of
approaching an observed maximum for SO removal across this particular spray
dryer. It is not obvious whether this observed maximum, at around 78 to 80
percent, is the result of limitations on the effectiveness of flue gas SO^ and
slurry contact in this particular vessel, or is perhaps dependent on some
other parameter, such as atomized slurry particle size distribution.
100-
80-
o
E
0)
O
05
40-
20-
LEGEND
O Overall SO2 Removal
D Spray Dryer Removal
A Fabric Filter Removal
0.5 1.0 1.5
Reagent Ratio
2.0
Figure 10.
Comparison of 1000 ppm Inlet S02 Recycle
Results with Best Fit of 400 ppm
Inlet S09 2:1 Recycle Test Results
Once-through tests at 1000 ppm inlet S02 levels were also conducted.
These results are compared to the 1000 ppm recycle results in Figure 11. Con-
trary to the 400 ppm inlet S02 test results, the difference between once-
through and recycle tests at TOGO ppm inlet S02 levels is less dramatic. For
a given reagent ratio, SO removal in once-through operation is generally
within 10 percent of removal in recycle operation. A reason for this obser-
vation may be the greatly increased weight percent solids to the wheel in once-
through operation when the inlet S02 levels are increased by a factor of 2-1/2
to 3. For example, once-through operation at 400 ppm or lower inlet S02 re-
sults in a lime weight percent of only 3 to 7 percent at the atomizer. Simi-
lar reagent ratios for once-through operation at 1000 ppm inlet S02 levels
result in weight percent solids levels of 10 to 16 percent. These solids
levels are the same as solids levels at 400 ppm inlet S02 for 2:1 recycle
operation.
10-99
-------
100-
80-
(0 60-
o
E
o>
QC
Q' 40-
crt
20-
LEGEND
Once-
Through
• Overall SOz Removal
• Spray Dryer Removal
A Fabric Filter Removal
0.5
I
1.0
i
1.5
2.0
Reagent Ratio
Figure 11. Comparison of Recycle Versus Once-Through
Operation Results for 1000 ppm Inlet S02
Improved once-through S02 removal performance at higher inlet SO levels
is further seen in Figure 12, where once-through SO removal results at both
inlet S02 levels are plotted. In this figure, it can be seen that except for
one 400 ppm test at low reagent ratio, all of the 1000 ppm test results for
overall and spray dryer S02 removal plot on a curve above that for the 400 ppm
test results. The one 400 ppm inlet S0? test result at close to 100 percent
utilization at an overall removal level"of 75 to 80 percent is suspected to be
.invalid. Current heat balance analyses of all test results should serve to
either substantiate the results of this test or indicate that the point should
be repeated.
A number of tests were conducted at low inlet temperatures of 210°F to
220°F. At Arapahoe barometric pressure conditions, this results in a spray
down temperature of only 100°F to 110°F. For normal inlet temperatures of
260 F to 280°F, the spray down temperature is on the order of 145°F to 165°F.
The 400 Ppm inlet S02 tests in recycle operation are of particular interest
at reduced inlet temperature. It is for low sulfur content coal, resulting
in FGD inlet S02 levels of 400 ppm and less, that low inlet temperatures are
^ ^™be encountered- F°ur tests were conducted at low inlet temperatures
in the 400 ppm, recycle mode. These test results are plotted with the 400
ppm, 2:1 recycle, normal inlet temperature baseline results in Figure 13. The
four low inlet temperature tests are at a wide range of reagent ratios and
show a good bit of scatter. However, they show that SO, removal performance
is not substantially reduced at low inlet temperatures/ In fact, one point
10-100
-------
demonstrates overall SO^ removal in excess of 90 percent at a reagent ratio
between 1.2 and 1.3.
100-
o
E
-------
100-
80-
:£ BO-
'S
o
E
cc
0 40-
20
/ ^
i''
9^ n^---13
tif "
®/
, LEGEND
• Normal 215°F
/ Inlet Inlet
/ Temp. Temp.
1 O 9 Overall SO2 Removal
/ D • Spray Dryer Removal
1 A A Fabric Filter Removal
' ^
0 0.5 1.0 1.5 2.
Reagent Ratio
Figure 13. Effects of Reduced Inlet Temperature on SO^ Removal
for 400 ppm Inlet SO., Recycle Operation
filter after the new bags were installed. However, the results in Figure 14
do not support this speculation. These results compare 1000 ppm inlet SO-,
recycle test results with the old bags to similar tests after the new bags
were put in service. The open symbols indicate the previous results, while
the darkened symbols represent results with the new bags. The tests were con-
ducted over a time period from immediately upon startup of the new bags to
several months afterward. No dependence of overall or fabric filter S0_
removal on old versus new bags or on operating hours on the new bags is
apparent in these results.
Crucial to the use of spray drying technology for many western plants is
the ability to employ makeup waters that are blowdown streams from other plant
water systems. Water balance considerations make cooling tower blowdown a
particularly desirable spray dryer makeup water source. A series of tests was
conducted on the Arapahoe pilot unit to determine the suitability of two simu-
lated cooling tower blowdown (CTB) streams as makeup to the system. The com-
positions of the two simulated CTB streams are summarized in Table 4.
10-102
-------
100-
80-
'. 60-
o
E
o>
IT
CNJ
O
C/3
40-
20-
LEGEND
Overall SO2 Removal
Spray Dryer Removal
Fabric Filter Removal
0.5 1.0 1.5
Reagent Ratio
i
2.0
Figure 14.
Comparison of 1000 ppm Inlet SO Recycle Results
with Old and New Bags in Fabric Filter
TABLE 4. NOMINAL COMPOSITIONS OF SIMULATED COOLING TOWER SLOWDOWN WATERS
(mg/L)
Component
CTB No. 1
CTB No. 2*
Ca++
Na+
Mg++
S04
ci-
1160
480
-
1000
2050
410
1830
53
5080
170
*Includes 10 ppm AMP, HEDP, or polyacrylate scale inhibitors for some runs.
CTB No. 1 was chosen because it represents the sulfate levels of a gypsum
limited, acid treated system. Originally of interest in this water, then, was
the sulfate content of 1000 ppm. This sulfate content has generally been
accepted as being too high for use as slaking water. Poor slaked lime quality
has been reported to result from the use of such waters in slaking. In order
to evaluate these impacts for a spray dryer based FGD system, tests were con-
ducted with CTB No. 1 used as slaking water, makeup to the recycle/lime slurry
mix tank, and temperature control water. In one test, CTB No. 1 was not used
for slaking, but still used in the mix tank and for temperature control. The
10-103
-------
results of these tests are compared to baseline 1000 ppm inlet SO,, recycle
tests in Figure 15. The results in Figure 15 are on the surface quite sur-
prising. All of the tests using CTB No. 1 show elevated S0? removal perfor-
mance both in the fabric filter and overall.
100-
80-
ra 60-
o
E
CD
rr
040H
to
20-
LEGEND
CTB
Base- CTB Slaking
line M/U & M/U
0
a
A
Overall SO2 Removal
Spray Dryer Removal
Fabric Filter Removal
0.5
i
1.0
T
1.5
r
2.0
Reagent Ratio
Figure 15.
1000 ppm Inlet S02 Recycle Results for Baseline
Conditions When Using Simulated Cooling Tower
Slowdown No. 1 for Slaking and Makeup Water
A further examination of simulated CTB No. 1 shows that the nominal
chloride level is quite high, at 2050 ppm. Although not yet confirmed by
chemical analyses, material balance calculations indicate that for the three
tests where CTB No. 1 was used as slaking, mixing, and temperature control
water, steady-state atomizer feed chloride levels were approximately 5000 ppm,
and the fabric filter solids had approximately a 1 percent chloride content.
These levels are very near those reported for substantial S00 removal en-
hancement by others.4,5,6 The enhancement effect may be cauied by the deli-
quescent properties of CaCl2, resulting in a lengthened second drying phase
time period in the spray dryer, and significantly increased residual moisture
levels in the solids collected on the fabric filter. Both impacts would tend
to improve SO removal performance.
The results in Figure 15 appear to substantiate a chloride enhancement
effect. The three tests involving CTB No. 1 for slaking, mixing, and tempera-
ture control show dramatically increased fabric filter SO, removal contribu-
tion. The one test employing CTB No. 1 as mixing and temperature control
water only shows a somewhat reduced effect. This is expected because the use
10-104
-------
of service water for slaking tends to dilute the steady-state chloride level.
While the presence of this chloride enhancement effect may mask the impact of
high sulfate levels on lime slaking, it does appear that any detrimental efect
of 1000 ppm sulfate levels is small compared to the potential benefits of
elevated chloride levels. One consideration in evaluating these results,
though, is that the chloride level in CTB No. 1 is artificially high, signif-
icantly higher than might be seen in a typical CTB water. The high chloride
content in CTB No. 1 is a result of adding CaCl2 to produce a gypsum relative
saturation of near 1.0 for the 1000 ppm sulfate level simulated blowdown
water.
The second simulated cooling tower blowdown water in Table 4, CTB No. 2,
accurately represents a side-stream softened cooling water system. This water
contains a significant amount of sodium as the result of softening, and con-
sequently low calcium and high sulfate levels. Four tests were conducted with
this simulated CTB; one with the constituents in Table 4 alone, and one each
with 10 ppm as active ingredient of HEDP, AMP, and polyacrylate scale inhib-
itor, respectively, added. These tests were conducted because most cooling
towers are treated with scale inhibitors. Scale inhibitors are intended to
delay the precipitation of limited solubility species on heat exchanger sur-
faces. However, precipitation of these species is the very operation occur-
ring as water evaporates from the droplets in a spray dryer. Although the
dosage of scale inhibitors encountered in most cooling tower blowdowns is
relatively low (<10 ppm), the net effect of such inhibitors on spray dryer S02
removal or solids drying performance was unknown. The results of these tests
are presented in Figure 16. The results show that, in general, SC>2 removal
across the spray dryer and for the overall system is enhanced by the use of
CTB No. 2 as mixing and temperature control water (slaking water use was not
tested). The individual tests are not identified because there is some
scatter in the four tests conducted, and individual points are not sufficient
to rank the magnitude of enhancement resulting from the use of the three
generic scale inhibitors. What is important in Figure 16 is that for all of
the four tests, overall system performance is enhanced by the use of the simu-
lated CTB as makeup.
Based on the encouraging test results presented in Figures 15 and 16 for
the use of simulated cooling tower blowdowns as system makeup water, a number
of additional tests using other simulated as well as actual CTB waters are
planned for the coming months.
SUMMARY AND CONCLUSIONS
A number of significant conclusions can be made at this point in the
pilot test program. These include:
• Spray dryer operability and S02 removal performance both are
improved by operation in the recycle rather than once-through
mode. The improvements appear to be related to increased
solids content in the slurry fed to the atomizer as well as to
the other benefits of recycling in improving sorbent utiliza-
tion.
10-105
-------
100-
80-
o
E
o>
o:
CO
40-
20-
LEGEND
Overall SO2 Removal
Spray Dryer Removal
FabricFilter Removal
I
0.5
1
1.0
1.5
2.0
Reagent Ratio
Figure 16.
1000 ppm Inlet SO Recycle Results for
Baseline Conditions and When Using
Simulated Cooling Tower Slowdown No. 2
for Makeup Water
Operation of a fabric filter downstream of a spray dryer
appears to be acceptable in terms of bag fabric impacts, and
may actually be beneficial with respect to pressure drop per-
formance. The one area of concern is that of corrosion of
compartment walls and mild steel bag caps in cold spots within
the compartments. Corrosion problems can most likely be
avoided through the use of adequate insulation thicknesses and
careful design and installation practices to avoid localized
cold spots.
For low spray dryer inlet S02 concentrations, SO removal per-
formance and lime reagent utilization remains quite acceptable
in spite of very low dryer inlet temperatures (as low as
210°F).
Cooling tower blowdown waste streams available in most power
plants appear to be an acceptable and in some cases beneficial
source of makeup water to a spray dryer based FGD system. Typ-
ical scale inhibitors present in these cooling tower blowdown
streams do not appear to adversely affect SO,, removal or solids
drying performance. ^
10-106
-------
REFERENCES
1. Blythe, Gary M. , et al. "EPRI Spray Drying Pilot Plant Status and
Results," presented at the Joint EPA/EPRI Symposium on Flue Gas Desul-
furization, Hollywood, FL, May 17-20, 1982.
2. Masters, Keith. Spray Drying Handbook. Third Edition. George Godwin,
Ltd., London, 1979.
3. Baker, Robert J. and Richard W. Jordan. "Effect of Dissolved Solids in
SC>2 Scrubbed Water Used for Lime Slaking," presented at and part of the
proceedings of the 3rd WWEMA Industrial Pollution Conference, Chicago,
IL, April 1975.
4. Karlsson, Hans T. , et al. "Activated Wet-Dry Scrubbing of SC>2. " Journal
of the Air Pollution Control Association. Volume 33, No. 1, January
1983. pp. 23-28.
5. Hansen, Svend Keis, et al. "Status of the Joy/Niro Dry FGD System and
Its Future Application for the Removal of High Sulfur, High Chloride and
NOX from Flue Gas." Publication No. 83-JPGC-APC-8, The American Society
of Mechanical Engineers, New York, NY.
6. Blythe, Gary M. and Richard G. Rhudy. "Field Evaluation of a Utility Dry
FGD System." To be presented at the joint EPA/EPRI Symposium on Flue Gas
Desulfurization, New Orleans, LA, November 4, 1983.
APPENDIX
CONVERSION OF BRITISH ENGINEERING (ENGLISH) UNITS TO SI UNITS
To Obtain From Multiply by
ng/J lb/106 Btu 430
m ft 0.3048
m2 ft2 0.0929
0.0283
mg/nm^
°C
gr/dscf
°F
1.83
tc = 5/9(tf-32)
10-107
-------
SESSION 10, PART II: DRY FGD: FULL SCALE INSTALLATIONS
Chairman: Richard G. Rhudy
Electric Power Research Institute
Palo Alto, CA
-------
FIELD EVALUATION OF A UTILITY DRY SCRUBBING SYSTEM
G. M. Blythe, J. M. Burke, T. G. Brna, R. G. Rhudy
-------
FIELD EVALUATION OF A UTILITY DRY SCRUBBING SYSTEM
Gary M. Blythe and Jack M. Burke
Radian Corporation
Austin, TX 78766
Theodore G. Brna
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Richard G. Rhudy
Electric Power Research Institute
Palo Alto, CA 94303
ABSTRACT
This program, co-funded by the U.S. Environmental Protection Agency and
the Electric Power Research Institute, has resulted in an evaluation of a full-
scale utility spray dryer/baghouse dry FGD system. The system is installed at
the Northern States Power Company's Riverside Station and treats flue gas from
a nominal 100 MW of coal-fired power generation. This has been the first
independent evaluation of a full-scale spray dryer/baghouse system.
For the test program, two different coals were used as boiler fuels. One
coal was a subbituminous coal and coke mixture with a nominal 1.2 percent sul-
fur content. The second was a 3.4 percent sulfur Illinois bituminous coal.
During the test program, S02 removal, particulate emissions, sulfuric
acid removal, and extensive process data were recorded. The test program was
conducted from July to October 1983, so only preliminary results are pre-
sented. Low sulfur coal tests indicated up to 90 percent S02 removal was
achievable in the short term with slightly sub-stoichiometric amounts of lime
addition. A similar removal was achieved in short term tests with high sulfur
coal at reagent ratios of 1.3 to 1.4. Calcium chloride addition was found to
reduce the lime addition requirements for high sulfur tests by approximately
25 percent.
10-109
-------
INTRODUCTION
This paper presents preliminary results from a program entitled "Field
Evaluation of a Utility Dry Scrubbing System." The objective of the program
has been to acquire performance data on an operating, utility-scale, spray-
dryer-based, dry FGD system. The system was evaluated primarily to determine
S02 and particulate removal performance and lime reagent consumption. The
system chosen for evaluation is the Joy/Niro Demonstration Unit located at the
Northern States Power Company (NSP) Riverside Station in Minneapolis, Minne-
sota. The Riverside system was chosen for this program because it is the
first lime-based system in operation using a full-size (46 ft* diameter) spray
dryer module. Testing was conducted with both low and high sulfur boiler
fuels. The program is being conducted for the Environmental Protection
Agency, Industrial Environmental Research Laboratory in Research Triangle
Park, and for the Electric Power Research Institute under a cooperative fund-
ing arrangement.
This test program was needed for several reasons. First, a significant
number of spray-dryer-based dry FGD systems have been sold to the utility
industry. At least 17 systems representing approximately 6,800 MW of electric
generating capacity have been sold to the utility industry. No independent
evaluation of the technology is available to aid utilities purchasing dry FGD
systems, since virtually all of the performance data have come from pilot- or
demonstration-scale units operated by system vendors. Several full-scale sys-
tems are coming on-line, but little performance data on these systems have
been published. A second consideration is that most of the new systems and
all of the future systems must meet the 1978 New Source Performance Standards
for coal-fired utility boilers, which call for 70 to 90 percent S02 removal on
a 30-day rolling average basis and particulate emissions of 0.03 lb/10" Btu
heat input or less. Few data exist to confirm the ability of full size spray
dryer-based dry FGD systems to achieve these levels of performance, particu-
larly for high sulfur coal requiring 90 percent S02 removal. Finally, the
spray dryer-based systems sold to date have been justified using economics
based on vendor guarantees for lime consumption. However, no independent eval-
uation of lime consumption on an operating full-scale system has been con-
ducted. The program described in this paper was designed to collect the type
of data that can help determine the suitability of this technology for utility
application.
""British engineering units rather than SI units are used in this paper because
of customary usage in the electric power industry. An appendix provides
appropriate conversion factors.
10-110
-------
The on-site test program was conducted from July 11 to October 8, 1983.
Because little time has been available for data evaluation, this paper does
not include a complete evaluation of results. Instead, the project is des-
cribed, system operation is discussed, and preliminary performance results are
presented.
PROJECT DESCRIPTION
The project description includes a description of the Riverside Station
and FGD system, a summary of the test program, and a discussion of the limita-
tions of the Riverside system and how they have affected this evaluation.
SITE DESCRIPTION
The Riverside Generating Station, operated by Northern States Power
Company, is located on the east bank of the Mississippi River in northeast
Minneapolis. Some parts of the facility are well over 50 years old, and the
two units of interest on this project, No. 6 and No. 7, began operation in
1949. The combined generating power of Units 6 and 7 is rated at 98 MW. How-
ever; the pulverized coal, wall-fired units were originally designed to fire
an eastern bituminous coal. Recently, the units have fired a western (Sarpy
Creek) subbituminous coal. A small amount (10 to 15 percent) of high sulfur
coke is added to the subbituminous coal to improve its firing properties. The
units can still be fired with high sulfur bituminous coal and, in fact, were
so fueled for five weeks of this test program.
Units 6 and 7 are used as peaking units in the NSP system. During the
cooler months of the year, the units are rarely operated. During the summer
months, the units are operated at up to 90 MW during daytime hours and at a
minimum load during the night. The units are normally banked over the weekend
during the summer.
In 1980, a full-scale, Joy/Niro, spray dryer/fabric filter FGD system was
installed to treat the combined flue gas from the two units. The fabric fil-
ter was actually purchased by NSP because of the inability of existing ESP
collectors to efficiently collect the ash from the western coal. The spray
dryer system was installed by the Joy/Niro joint venture under a cooperative
agreement with the utility to serve as a full-scale demonstration of the capa-
bilities of their dry FGD system. Figure 1 is a simplified flow diagram for
the system.
The spray dryer is a 46-ft diameter vessel, with flue gas introduced both
above the atomized spray in a roof gas disperser and below the atomized spray
in a central gas disperser. A rotary atomizer is used, currently employing a
700 hp drive motor. The spray dryer was sized to treat flue gas corresponding
to a 70 MW boiler load. This reduced sizing allowed the capability to test
the system at greater than design flow rates. It should be noted that, be-
cause Units 6 and 7 are over 30 years old, the flue gas flow rate at 70 MW is
equivalent to the flue gas rate from approximately a 100 MW new unit. A new
unit would experience much less air inleakage and operate at a much lower net
plant heat rate than these units. The downstream fabric filter contains 12
compartments, in 2 rows of 6 compartments each. Because the fabric filter was
10-111
-------
sized to treat hot flue gas, it is actually oversized when the spray dryer is
in operation because of the flue gas volume shrinkage which results from the
reduced spray dryer outlet temperature.
Pebble lime reagent is slaked in a Joy/Denver attrition slaker. A Joy/
Denver ball mill is also available for lime slaking. Milk of lime, dilution
water, and recycle solids are added to a mix tank at rates determined by a
Honeywell process control computer. The mix tank effluent is pumped to a sep-
arate atomizer feed tank. From the atomizer feed tank, slurry is pumped to a
head tank at the top of the spray dryer. A pinch-type control valve regulates
the flow of slurry to the atomizer to maintain either a constant spray dryer
outlet temperature or a constant approach to adiabatic saturation. When the
system is operated in an S02 removal control mode, the Honeywell process con-
trol computer calculates the amount of lime which must be added upstream at
the mix tank in order to achieve the desired S02 removal. Recycle material is
added at the mix tank at a rate required to bring the mix tank solids level up
to a set point, normally 35 weight percent solids. The recycle solids are col-
lected from the spray dryer bottom dropout and largely supplemented by a por-
tion of the fabric filter catch.
Unit?
Stack
Flue Gas
From Unit 7
Lime
Slurry
Storage
Trough
Figure 1. Flow Diagram for Riverside Dry FGD System
10-112
-------
The atomizer feed scheme tested is now an obsolete design. The mix tank
and atomizer feed tanks used at Riverside each have about 45 minutes of resi-
dence time. Thus the system is relatively slow to respond to changes in spray
dryer inlet flue gas conditions (i.e., changing SC>2 concentration, inlet tem-
perature, or wet bulb temperature)- Current Joy/Niro system designs call for
preparing milk of lime slurry and recycle slurry in separate tanks and then
mixing them in a tank with a relatively short residence time. This should
result in a quicker response to process changes.
TESTING APPROACH
As described in the introduction, a primary objective of the program has
been to quantify SC>2 removal by the system. A continuous emission monitoring
system (CEMS) has been installed to quantify S02 removal. The CEMS includes a
DuPont Model 460 two-point extractive 862 monitor and a Thermox 02 monitor
sampling the spray dryer inlet and spray dryer outlet ducts, and a Lear
Siegler SM810 in-situ point-type S02 analyzer installed in a short run of duct
at the fabric filter outlet. A second Thermox 02 analyzer is mounted on the
duct exactly opposite the Lear Siegler monitor.
Other than S02 removal data, lime consumption and other important process
parameters were recorded as hourly averages for each test day- Lime consump-
tion was measured primarily by determining the lime content of the milk of
lime slurry introduced to the atomizer feed mix tank. The flow rate of this
slurry was continuously measured with a magnetic flow meter and recorded by
the Honeywell process control system computer. Other methods of lime consump-
tion measurement have included continuous quicklime weigh belt rate measure-
ments, recording of daily quicklime truckload deliveries, and determination of
the lime content and flow rate of the actual atomizer feed slurry- Energy bal-
ances have been used to confirm agreement between slurry feed rate and flue
gas flow measurements.
Other than quantifying S02 removal and lime consumption performance for
the spray dryer/baghouse system, determination of particulate removal
performance for the system was also a primary objective. This performance was
determined by manual sampling of flue gas streams for particulate
concentrations, using EPA methods. Particulate loadings were measured at the
spray dryer inlet, spray dryer outlet, and fabric filter outlet locations.
TEST PLAN
The test schedule is summarized in Table 1. The schedule shows four dif-
ferent sets of conditions with low sulfur, Sarpy Creek coal/coke blend, and
three different sets with a high sulfur Peabody Illinois coal. The Sarpy
Creek coal/coke blend has a nominal sulfur content of 1.1 to 1.2 percent and a
heating value of 9300 Btu/lb. New Source Performance Standards for utility
boilers would require 75 to 80 percent S02 removal for boilers firing a fuel
with this sulfur and heating value. In some localities, state or local regula-
tions might require as high as 90 percent S02 removal with this fuel. Conse-
quently, low sulfur tests were conducted at both 75 percent and 90 percent
target S02 removal levels. Additionally, some tests were scheduled at a fab-
ric filter air-to-cloth ratio higher than the normal value of 2:1 and with the
10-113
-------
TABLE 1. TEST SCHEDULE AND TARGET SYSTEM OPERATING CONDITIONS3
Fuel
Sarpy Creek Coal/Coke
Sarpy Creek Coal/Coke
Sarpy Creek Coal/Coke
Sarpy Creek Coal/Coke
Illinois Coal
Illinois Coal
Illinois Coalb
S02
Removal
Level,
%
75
75
90
90
90
90
90
Fabric
Filter
A/C Ratio,
cfm/ft2
2:1
2.2:1
2:1
2:1
2:1
2:1
2:1
Slaker
Type
Attrition
Attrition
Attrition
Ball Mill
Ball Mill
Attrition
Attrition
aAll tests planned to be conducted at an 18°F approach to adiabatic
saturation, 35 weight percent solids in the atomizer feed slurry, a 70 MW
daytime boiler load, and with a once-per-hour baghouse cleaning frequency,
^Calcium Chloride Addition Tests
10-114
-------
ball mill slaker rather than attrition slaker being used to prepare the lime
slurries.
The test at the higher air-to-cloth ratio was originally intended to be
conducted at an air-to-cloth ratio near 2.7:1. However, the air-to-cloth
ratio is increased at a constant boiler load by taking individual compartments
off-line. After taking compartments off-line to raise the gross air-to-cloth
ratio to near 2.7:1, only 6 of the 12 compartments remained on-line at a 70 MW
boiler load. The combined effects of taking an operating compartment off-line
for cleaning and reverse gas flow during cleaning momentarily sends the effec-
tive air-to-cloth ratio to 3.5:1 to 4:1. In a unit with more compartments on-
line, the effects of cleaning one compartment in increasing the effective air-
to-cloth ratio are much smaller. Additionally, in the event of a temporary
loss of slurry flow to the atomizer while a compartment is cleaning, the
increased flue gas volume due to the much higher spray dryer outlet tempera-
ture would result in a further increase in the effective air-to-cloth ratio to
well over 4:1. When operated at such a high effective air-to-cloth ratio, the
pressure drop across the fabric filter is increased to several times its nor-
mal value. Such a combination of events did occur on the first day of the
high air-to-cloth ratio tests, and the resulting momentary high pressure drop
caused the control system to bypass the fabric filter. In order to provide a
margin of safety, one more compartment was put on-line. This lowered the
gross air-to-cloth ratio to 2.2:1 to 2.3:1, which is only 10 to 15 percent
higher than baseline conditions.
The fourth set of conditions for the low sulfur tests (see Table 1) was
to have included ball mill rather than attrition slaked lime. However, ball
mill operation during this test program was generally not successful. The
reasons for the ball mill operating problems are further discussed under Oper-
ational Results. Because of these problems, less than one day of system
operation was observed using ball mill slaked lime.
Two sets of conditions were tested with the high sulfur coal. The coal
was an Illinois No. 6 coal with a nominal 3.4 percent sulfur content and
10,800 Btu/lb heating value. Current New Source Performance Standards for
utility boilers require 90 percent S02 removal when coal of this sulfur con-
tent and heating value is burned. Consequently, only a 90 percent target S02
removal was tested with this high sulfur coal. The tests included baseline
conditions of 90 percent removal with attrition-slaked lime and a second test
run employing calcium chloride addition for lime utilization enhancement.
Chloride addition has been reported previously to enhance lime utilization in
spray dryer/baghouse FGD systems^ ' , but this is the first test of chloride
addition at a full-scale utility installation. The enhancement effect is
thought to occur because of the deliquescent properties of calcium chloride.
This effect is further discussed in the System Performance portion of this
paper.
Originally, ball mill slaker tests and high air-to-cloth ratio tests were
planned for the high sulfur coal portion of the test program. However, due to
10-115
-------
the ball mill slaker operating problems previously mentioned and the previous-
ly discussed constraints on gross air-to-cloth ratio, only the last two tests
listed in Table 1 were conducted during the high sulfur test period.
SYSTEM LIMITATIONS
Several system limitations combined to restrict the amount and the type
of data that could be collected. First, as mentioned earlier, Riverside Units
6 and 7 are peaking units. As such, they are rarely operated in the winter
and only operate part-time during July to October. This part-time operation
involves unit loads of 70 to 90 MW during weekday daylight hours, minimum load
(30 to 50 MW) overnight during the week, and banking the boilers over the
weekend. Figure 2 is an illustration of the combined Units 6 and 7 loads dur-
ing a typical week. Therefore, although the FGD system was operated at
desired S02 removal levels 24 hours per day, only about 12 hours per weekday
of near full-load operation were available for evaluating FGD system perfor-
mance. At the beginning of each 12-hour full-load period, the FGD system gen-
erally goes through a transient period due to a large increase in boiler load.
On Mondays the unit must undergo a cold start-up. Although this cycling pro-
vides a severe test of the capabilities of the system, it reduces the period
of steady state operation at the desired S02 removal level over which lime
reagent consumption can be measured.
100-
80-
60-
_
~ 40 H
c
20^
Sun Mon Tues Weds Thurs Fri Sat
Figure 2. Illustration of Typical Unit Load During Test Program
Additionally, the Riverside system was the first utility-scale system
designed and built by the Joy/Niro joint venture and was built as a demonstra-
tion unit. As the first unit built, the Riverside system has provided the
opportunity to refine and modify design features for subsequent systems.
10-116
-------
Thus, some specifics of the Riverside system are different from what will be
found in later designs. An example of this is the slurry feed system design,
which has been modified for subsequent system designs to provide quicker
response to transients such as load changes. At Riverside, lime slurry, make-
up water, and recycle solids are mixed in a single tank. The lime slurry
addition rate is based on SC>2 removal (for the SC>2 removal control mode of
operation), the makeup water rate is controlled to maintain level in the mix
tank, and the recycle solids rate is controlled to achieve a set weight per-
cent solids (normally 35 percent solids). The mix tank has a residence time
of approximately 45 minutes. This slurry mixture then overflows to an atom-
izer feed tank with a similar residence time. When the inputs to the mix tank
change composition abruptly (such as when the unit load increases), approxi-
mately 3 or 4 hours are required before the feed system stabilizes near steady
state for the new conditions. Recently designed Joy/Niro systems have a
faster response slurry feed system. In these systems, recycle solids and
makeup water are mixed in a recycle slurry tank to a consistency of approxi-
mately 50 to 60 weight percent solids. The lime and recycle slurries are then
mixed in a feed tank with a relatively short residence time of one-half hour
or less. Because only this feed tank residence time impacts the system
response time, this feed system might stabilize after an abrupt change of
operating conditions in one hour or less.
The impacts of the slurry feed preparation system design are particularly
important at the Riverside Station, as the normal station operation causes
significant load changes at least twice per day. In fact, on some days during
this test program the NSP dispatcher called for the unit load to be varied
between 70 MW and 90 MW throughout the day. On these days, the unit never
operated at one load long enough for the feed system to stabilize. On most
other days, even if the load was steady all day, only 8 or 9 hours of the 12
hours of full load operation actually represented steady state conditions.
Other aspects of the Riverside system's status as a demonstration unit
affected the results of this program. For example, the system contains only
one spray dryer module, while most utility systems will be multiple module
systems. In a multiple-module system, equipment problems which affect one
individual module have a smaller impact on overall system performance. Being
a one-module system, equipment problems tended to cause the entire system to
have to be shut down, or operated at conditions other than those desired.
Even considering that the smaller Riverside system only warrants a single
dryer vessel, some individual components of the system have not been installed
with the redundancy that the vendor would likely install in a commercial sys-
tem. These considerations have had a detrimental effect on both the amount of
system downtime and the number of process-equipment-related upsets during the
test program.
A final consideration which has affected the test program involves the
recent history of the Riverside station. For nearly the first two years of
operation, the FGD system was used as a full-scale demonstration and testing
unit by the process vendors. During this time the Joy/Niro joint venture had
responsibility for the operation of the system, even though NSP actually pro-
vided operating personnel. Within the year prior to the test program, NSP had
assumed responsibility for the operation of the FGD system. Immediately prior
10-117
-------
to the test program, the Units 6 and 7 boilers and FGD system were off-line
for much of the winter and spring, as NSP does not need power production from
these units during this time. During this long period of downtime, normal
personnel turnover (promotions, retirement, transfers, etc.) resulted in a
number of new operators rotating into the FGD system operating staff. Early
in the test program, then, the FGD system was being operated for the first
time in several months with a staff of operators having little previous exper-
ience with the FGD system. Consequently, a lot of test days early in the pro-
gram were not productive as far as representing steady state operation at test
conditions. As the test program continued, successful test days represented a
higher fraction of potential days of operation.
Another consideration here is that, after the Joy/Niro joint venture
turned over the dry FGD system to the utility, and prior to conducting this
test program, NSP only operated the system at 50 to 60 percent S02 removal
levels. For such moderate S02 removal levels, they had previously operated
the system under very conservative conditions (i.e., higher spray dryer outlet
temperatures). Thus, operation at 75 or 90 percent S02 removal, with a 20°F
approach to adiabatic saturation, was not a normal operating mode immediately
prior to this test program. Early in the program, the operators tended to
revert to conservative operation (i.e., higher spray dryer outlet temperature)
during any minor upset, such as soot blowing in the boilers. This would move
the operation away from the desired conditions and would preclude acquiring
desired steady state operating data. As the program continued though, this
occurred much less frequently as the operators became more comfortable with
operating at test conditions.
Considering the previous discussions, it was not realistic to report
availability of the system, as the availability of the Riverside system would
tell little about that of a commercial utility, multi-module, dry FGD system
on a new base-loaded boiler. The combined effects of weekly cold startups,
frequent load changes, little redundancy, and a somewhat undertrained operat-
ing staff at Riverside do little to promote a fair assessment of the potential
availability of a commercial system.
However, the general operation of the system was closely observed during
the test program. Much of the downtime or off-condition operating time was
due to problems specific to the Riverside system. Others appear to be more
generic to dry FGD systems. These more generic problems are discussed in this
paper, as they are more likely to occur in other systems.
RESULTS
The results of the program are divided into two areas: Operational
Results, which includes a qualitative discussion of the operation of the sys-
tem during the test program; and System Performance, which includes prelimi-
nary S02 removal, lime consumption, and particulate and sulfuric acid removal
data.
10-118
-------
OPERATIONAL RESULTS
In general, the equipment that comprises the basis of the dry FGD system,
the spray dryer and baghouse, were relatively trouble-free throughout the
program. At the conditions tested, the spray dryer did not show evidence of
potential problems such as wheel nozzle pluggage, excessive buildup of solids
on the walls, or formation of wet solids within the dryer. Some atomizer
problems were observed, but most of these appeared to be caused by circum-
stances specific to the situation at Riverside rather than being generic to
the Joy/Niro system. These problems will be discussed further later in this
section.
The baghouse also operated well, with no significant bag/fabric related
problems being observed. In this short-term test though, long-term effects
such as bag life or compartment wall corrosion rates could not be evaluated.
Some problems were observed in four specific areas—the slurry feed sys-
tem, the ash handling system, the ball mill slaker, and in atomizer protec-
tion. The system vendors may have addressed these problems in system designs
subsequent to Riverside, but the problems could be encountered in virtually
any spray-dryer-based, dry FGD system. Each of these areas is discussed
be 1 ow.
Slurry Feed System—In a recycle lime system, lime slurries containing
up to 25 percent solids and recycle/lime slurries of 30 to 40 percent solids
are commonly encountered. When dealing with slurries with a high solids con-
tent and high viscosity, problems such as plugging of pump suction lines,
solids buildup on tank walls, plugging of in-line screens used to remove over-
size material, and loss of flow when switching pumps are commonly encountered.
Such problems were encountered often at the Riverside system. Years of opera-
tion of wet lime/limestone FGD systems have established means of dealing with
such problems: always keep slurry in tanks well agitated, never allow slurry
levels to drop to the agitator blade level, never allow slurry to stand in
filled lines, allow for on-line changing of plugged or partially plugged
screens, etc. These same solutions will need to be applied in spray drying
systems.
The quantities of these slurries that must be dealt with in a spray dryer
system are much smaller than what would be encountered in a wet system. At
Riverside, typical atomizer feed slurry rates are 150 to 200 gpm. In a simi-
larly sized limestone wet FGD system, the slurry recirculation rate could be
as high as 40,000 gpm. The point to be made is that while some of the slurry
handling problems of a wet FGD system may still be encountered in a spray
dryer system, they will occur on a greatly reduced size of equipment. This
should make both problem solving and routine maintenance easier. Another
important point to be noted is that no chemical scaling tendencies were
observed at Riverside.
Ash Handling System—In comparing the wet versus dry FGD systems the
spray dryer system has a slurry feed system that deals with much lower flow
rates for a given unit capacity, but the quantities of dry ash and FGD
10-119
-------
by-products that must be moved around the system are substantially greater
than for a comparably sized particulate collection device/wet scrubber system.
For example, at Riverside the amount of solids entrained in the flue gas
at the fabric filter inlet (spray dryer outlet) is 3 to 5 times that in the
spray dryer inlet gas. For low sulfur coal operation only a small portion of
this increase is attributable to the lime reagent introduced and S02 removed
in the spray dryer. The majority is material being recycled from the fabric
filter catch to the spray dryer. Thus, the rate at which solids are removed
from the fabric filter hoppers and transported to recycle storage bins can be
several times greater than the net rate of waste product actually leaving the
system.
Specifically, for low sulfur coal, 75 percent S0£ removal, and the unit
operating at 80 MW, the steady state rate of fly ash and spray dryer by-pro-
ducts leaving the system amounted to 7 to 8 tons/hr- However, due to the
effects of recycle, approximately 20 tons/hr were actually being collected by
the baghouse. The excess between that going to the disposal silo and the 20
tons/hr being collected, or 12 to 13 tons/hr, was being recycled to the slurry
feed system. In this example, roughly three times the quantity of ash and
spray dryer solids leaving the system must be continually removed from the
baghouse hoppers, and roughly twice the quantity leaving the system must be
continually transported to the recycle storage bin.
At Riverside, solids collected in the baghouse hoppers and at the bottom
of the spray dryer are intermittently dumped onto mechanical conveyors through
motor-driven tipping valves. One conveyor each handles ash from the six com-
partments on the north and south sides of the baghouse, respectively, while a
third conveyor handles spray dryer bottom solids. These conveyors empty into
surge bins mounted above air conveying blow pots. The blow pots cycle through
the following five consecutive processes: filling from the surge bin, isola-
tion from the surge bin, pressurization, pneumatic conveying to either the
recycle or disposal storage silos, and then depressurization. The recycle
silo empties through a rotary valve onto a weigh belt which meters recycle
material into the mix tank.
The problems with solids handling which most frequently occurred at
Riverside involved the baghouse mechanical conveyors, blow pots, and the
recycle bin rotary valve. The problems with the mechanical conveyors appeared
to be related to marginal capacity. The north or south conveyors frequently
tripped out because their drive motors were overloaded. Momentary overloading
of the conveyors is aggravated by the nature of the operation of the baghouse,
where compartments clean intermittently throughout each hour. That is, the
average rate at which the baghouse collects material might be 20 tons/hr, as
in the example above. Each conveyor would see an average of 10 tons/hr if the
same number of compartments were on-line on both sides. However, this average
includes short periods of high solids flow, immediately after a cleaning com-
partment begins reverse gas flow, and periods of low solids flow rate between
compartment cleaning periods. Thus, one conveyor may see instantaneous rates
of several times 10 tons/hr. This problem was apparently corrected after this
test program was completed by a modification which meters the rate at which
ash/spray dryer by-products are emptied from the hoppers onto the conveyor.
10-120
-------
In addition, the conveyors require occasional maintenance, such as
replacement of failed drive motors and bearings. If one conveyor must be
off-line for a very long time, then all six of the compartments emptied by
that conveyor must be taken off-line once the hoppers below those compartments
are full of ash. At Riverside, failure of one conveyor can cause the boiler
load to be limited to that which can be handled by the six compartments
remaining on-line. Because the rate at which solids are collected in a fabric
filter downstream of a spray dryer is greatly increased, unless the hoppers
are increased in size compared to conventional particulate control configura-
tions, such a conveyor failure can very quickly limit unit load.
Failure of a blow pot can have a similar effect, as only a small surge
bin separates each conveyor from a blow pot. NSP has attempted to provide
some redundancy in blow pots, but in an instance where only one blow pot is
available to accept solids from a conveyor, a blow pot outage can cause the
conveyor to be shut down as well. Common blow pot failures observed included
plugging, leaking isolation valves, and loss of air pressure to the
pneumatically operated valves.
A third piece of solids handling equipment which rapidly affected the
operation of the system on failure was the rotary valve at the bottom of the
recycle solids bin. When this valve failed, often due to a piece of tramp
material jamming the blades, the solids content in the atomizer feed slurry
immediately began to drop, and the system began to approach once-through lime
operation. The large slurry mix tank and atomizer feed tank could be consi-
dered a 1.5 hour reservoir of feed slurry, but in order to avoid splashing
slurry on tank walls and a resulting buildup of dry solids, the tanks must be
continually kept near full. Consequently, these tanks begin to dilute immedi-
ately when recycle solids flow is interrupted because makeup water must be
added to maintain level.
The solids handling equipment appears to be a key to successful operation
of the spray dryer/baghouse system. The individual pieces of equipment must
be sized to handle solids rates much higher than is typical for conventional
particulate removal alone.
Ball Mill Slaker—As mentioned previously, during this program the ball
mill slaker did not operate successfully for any extended period. The feed
end of the slaker tended to plug with wet lime solids. While there are sever-
al possible reasons why the plugging continually occurred, the actual cause
was not identified. The ball mill operation is discussed below.
The ball mill slaker is fed by an independently driven screw feeder
inserted directly through the trunion bearing of the mill. Product lime flows
out the opposite end, through a cylindrical screen, and into a product
collection trough. Figure 3 is a simplified cross section of the ball mill
depicting the feed and product ends of the mill. Two things are clear from
this simplified illustration: first, vapors produced by the heat of the
slaking reaction can only leave the slaker with the product or flow back
through the feed end. Secondly, little hydraulic head is available to promote
the flow of the viscous slaked lime slurry out the product end of the slaker
without immersing the feed screw in slurry.
10-121
-------
Product
End
Figure 3. Simplified Cross Section of Ball Mill Slaker
Either of these considerations may have contributed to feed screw plug-
gage. The ball mill slaker was operated at a very low water-to-lime ratio,
resulting in 30 to 35 percent solids in the mill product slurry and a product
temperature of 97°C (207°F) or greater. At these conditions, the product
slurry is very viscous, having a paste-like consistency. Also, with the prod-
uct slurry very near or perhaps actually boiling, vapor production rates are
high. Both of these effects of low water-to-lime ratio can promote feed screw
pluggage. Another consideration is that the ball mill slaker had not been
operated for some time prior to this test program, as NSP prefers to operate
the attrition slaker at Riverside. The mill had not been completely cleaned
out after its last use, and solids buildup around the discharge end undoubt-
edly contributed to the slurry level in the mill backing up into the screw
feeder -
A final problem specific to the Riverside system may also have contrib-
uted to feed screw pluggage. The SC>2 removal software in the process control
computer tended to adjust the lime slurry makeup rate to the mix tank after a
major change in the process (e.g., unit load change) in a classic dampened
controller response. That is, after an abrupt process change the lime slurry
makeup rate to the mix tank would tend to overshoot, then undershoot, for
several cycles, before stabilizing at the required value. During the under-
shoot portion of the response, the lime slurry feed rates might be completely
stopped. As the Riverside system has little lime slurry storage capacity, the
slaking rate tended to directly follow the slurry makeup rate. Thus, periods
of high slaking rates would be followed abruptly by periods where the screw
feeder would shut down. Shutting down the screw feeder momentarily several
times in succession undoubtably contributed to the plugging problems.
In
The ball mill slaker problems at Riverside appear to be somewhat
site-specific as ball mill slakers have operated successfully elsewhere,
retrospect, the slaker might have run more successfully at a higher
water-to-lime ratio, resulting in a less viscous product slurry and less vapor
10-122
-------
release. However, in the attempts to run the ball mill slaker during the high
sulfur tests in particular, the slaking water piping size did not allow opera-
tion at higher water-to-lime ratio at the lime slaking rates required. During
the high sulfur tests, lime slaking rates averaged approximately four times
that required for the normal low sulfur fuel at similar unit load and percent
SC>2 removal conditions.
Atomizer Problems—Although the atomizer motor, gearbox, and nozzle
wheel were generally trouble-free, on several occasions a problem which could
result in atomizer damage was observed.
Two scenarios for potential damage were observed. One occurred when the
unit was forced to run from the basic, or less sophisticated, control station
of the computer control system while the more sophisticated supervisory sta-
tion was undergoing repair. The control software at Riverside does not have
full interlock protection for the atomizer when running from the basic sta-
tion. Interlocks are software which automatically shut down the atomizer when
given inputs that are indicators of problems which might result in damage to
the atomizer. While running in this mode, a minor problem involving loose
wires to the atomizer oil circulating pump occurred, intermittently shutting
the pump off. However, the basic station only gave the control operator an
alarm rather than shutting the atomizer down automatically. The atomizer con-
tinued to run for several minutes without oil circulation and sustained gear-
box damage.
This incident may be very site-specific to the Riverside system due to
the limitations discussed earlier. First, Joy/Niro has reported that their
commercial systems have full atomizer interlock protection in both the super-
visory and basic computer control stations. Second, an operating staff with
more experience on a spray dryer system might have shut down the atomizer more
quickly when given an indication of a loss of oil flow and would not have
attempted to restart the machine prior to remedial action.
The second scenario which could result in atomizer damage was observed on
more than one occasion. This incident involved feeding slurry to the atomizer
wheel when it was not rotating. Since the non-rotating wheel has a much lower
hydraulic capacity than a rotating wheel, slurry fed to the standing wheel
tends to overflow the wheel and can flow up the spindle to which the wheel is
attached and enter the atomizer oil system. In such instances the oil sump
can be immediately emptied and flushed to avoid damage, but if the atomizer is
operated before cleaning, the slurry in the oil can eventually cause gearbox
damage.
Such an instance of feeding slurry to a standing wheel can occur from
operator error in establishing flow to the atomizer when it is not in opera-
tion and/or starting the atomizer from the basic station without full inter-
lock protection. As for the previous discussion, such an occurrence is much
less likely for a commercial Joy/Niro system, as even the basic station would
have full interlock protection. The interlock system would normally close a
slurry feed block valve before the atomizer shut down and would not allow this
valve to open until the atomizer had returned to operation at full speed.
10-123
-------
However, even with a properly operating interlock system, the previously
described problem could occur. One such event occurred during this test pro-
gram. In this instance, during an atomizer trip, valve seat damage prevented
the atomizer feed slurry block valve from closing. In this case, all of the
interlocks operated properly, but a mechanical problem allowed slurry flow to
the standing wheel. An alert operator could catch such a problem very
quickly, but it is still likely that slurry would get into the atomizer lube
oil unless the flow was stopped immediately.
It is interesting to note that in the example above, the spray dryer was
hardly upset by full slurry flow to a standing wheel. Several thousand gal-
lons of slurry were fed in one instance, and the only repercussions were a
mess under the dryer vessel where the slurry flowed out of the dryer onto the
ground, and a minor buildup of dried solids on the vanes of the central gas
disperser just below the wheel.
What is clear from these observations of potential atomizer damage are
two conclusions. First, the rotary atomizer on a spray dryer FGD system
should always be operated with the full protection of interlock controls.
Second, the control operators should be well trained to respect the potential
damage which can result to the atomizer during upset conditions.
SYSTEM PERFORMANCE
The results of S02 removal and lime consumption measurements during this
test program are summarized in Tables 2 and 3. Table 2 summarizes the low
sulfur S02 removal results, while Table 3 summarizes those for the high sulfur
test period.
For several reasons, the S02 removal results from the program cannot be
expressed as 30-day rolling averages. First, the unit was never operated at
one set of conditions for that long a period during the test program. Also,
due to the operating characteristics of the peaking boilers, only 8 to 9 hours
per day typically represent full load, steady state operation. Thus, the S02
removal and lime consumption results represent values measured during steady
state unit operation on only a portion of a number of successive days. The
fact that these results are the average of a number of short term tests cannot
be used to determine whether or not the dry FGD system could sustain these
removal levels over a long period of time; instead, they only reflect the fact
that the boiler rarely operated at or near full load for a long period of
time.
Lime consumption in Tables 2 and 3 is expressed as a reagent ratio. This
term is defined as:
Reagent Ratio = Moles Calcium in Fresh Lime Fed to System m
Mole S02 in Inlet Flue Gas
(This definition corresponds to that of the term "stoichiometric ratio" in
many other dry FGD papers.)
10-124
-------
It should be noted that the values in Tables 2 and 3 should be considered
preliminary values which are based on a number of manual calculations for each
set of conditions. Where possible, each value is supported by alternate cal-
culations. For example, lime slurry feed rates are compared to lime weigh
belt readings: essentially a calcium balance on the lime slaker. Also, weigh
belt readings have been compared against lime truckload delivery inventories.
Flue gas flow rates are checked against slurry feed rates by energy balance
calculations. In general, the values in Tables 2 and 3 represent the most
accurate values available at this time. As more data reduction is completed,
these values will be refined but are not expected to be changed substantially.
Table 2 shows that the desired SC>2 removal levels of 75 percent and 90
percent were achieved for the lower sulfur fuel with lower than or right at
stoichiometric amounts of lime. This may be attributable to two factors.
First, at 35 percent feed slurry solids in low sulfur operation very high
recycle rates are possible. This is seen in the recycle ratio values in Table
2. Recycle ratio is defined in this paper as:
P i P j. • _ Lb Recycle Material in Atomizer Feed Slurry
Kecycie Katio - [Ca(OH)2]
(2)
TABLE 2. PRELIMINARY S02 REMOVAL RESULTS, LOW SULFUR
Nominal
S02
Removal ,
%
75
90
Spray Fabric
Dryer Filter
Removal, Contribution, Reagent Recycle
% % Ratio Ratio
67-69 7-9 0.6 - 0.7 11:1 - 14:1
80-81 9-10 0.7 - 0.8 9:1 - 13:1
TABLE 3. PRELIMINARY S02 REMOVAL RESULTS, HIGH SULFUR
Nominal Spray
S02 Dryer
Removal, Removal,
% %
90 75 - 77
90a 67 - 69
Fabric
Filter
Contribution, Reagent Recycle
% Ratio Ratio
13 - 15 1.3 - 1.4 2:1 - 3:1
22 - 24 0.9 - 1.1 3:1 - 4:1
aHigh Chloride Concentration Tests
10-125
-------
Since a large fraction of the baghouse catch and all of the spray dryer bottom
solids are recycled, from pilot plant results it would be anticipated that
high sorbent utilization would be promoted. Additionally, analyses of the
Sarpy Creek coal ash are not yet available, but the ash is known to have a
high alkaline earth content (CaO, MgO, Na20, and K20). These alkaline com-
ponents are believed to have contributed to SC>2 removal.
The results in Table 2 also show the range of SC>2 removal in the spray
dryer (S.D.) and fabric filter (F.F.). SC>2 removal in the fabric filter is
calculated relative to the spray dryer inlet S02 concentration as follows:
[SO into F.F. - SC>2 out of F.F.]
S02 RemovalF>F> = [S02 into S.D.](3)
With this definition, spray dryer removal and fabric filter removal can be
summed directly to yield overall removal values. The results show that at 75
percent overall SC>2 removal, fabric filter removal contributes little to the
overall removal. This occurs because sorbent utilization is very high in the
spray dryer itself. For 90 percent removal, the fabric filter contribution
increases somewhat, but in an amount roughly proportional to the increase in
overall removal level.
Table 3 summarizes the high sulfur coal test results. For the first set
of data, corresponding to normal low chloride operation, an excess of fresh
lime is required. Several factors probably contribute to this effect. One
may be that, because of the increased lime addition rates, the recycle ratio
is greatly reduced relative to low sulfur values in order to maintain the
total solids in the slurry at the desired weight percent value. Also, the
Illinois coal ash being nonalkaline contributes no alkalinity to the S02
removal reactions. For the first set of high sulfur data, the fabric filter
contribution to overall S02 removal appears to be more important than for low
sulfur operation.
Chloride addition has been reported by others to promote increased sor-
bent utilization in spray-dryer-based FGD sy stems-'- >2. The benefits are
thought to result from the deliquescent properties of calcium chloride, which
delay complete drying of the droplets in the spray dryer and result in higher
residual moisture levels in the fabric filter solids. The second set of S02
removal data in Table 3 corresponds to the addition of calcium chloride at
levels which result in a chloride content of 1 percent in the fabric filter
solids collected. This chloride level in the fabric filter solids was recom-
mended by the system vendor- Niro Atomizer, as being an optimum value for lime
utilization enhancement, based on pilot-scale studies conducted at their
Copenhagen, Denmark, test facility. At Riverside, this solids chloride level
required a liquid phase chloride concentration of approximately 7,500 ppm in
the atomizer feed slurry. For the recycle rates at Riverside, fresh makeup of
calcium chloride accounted for about half of this liquid phase content, and
the remainder dissolved from the recycle material. The results in Table 3
show that chloride addition significantly reduced the lime reagent ratio
requirements to achieve 90 percent removal; the lime requirement was reduced
by about 25 percent.
10-126
-------
The SC>2 removal results for these high chloride tests indicate increased
SC>2 removal across the fabric filter for 90 percent removal overall. This is
an indicator that the benefits of high chloride level on residual moisture
level in the fabric filter have a greater impact on SC>2 removal than impacts
within the spray dryer.
At Riverside, with chloride levels in the fabric filter solids at 1 per-
cent, residual moisture levels increased from just below 1 percent to nearly
2 percent moisture. These moisture levels are still low enough to avoid
problems which result from handling wet solids. Also, no buildup of wet
solids on the spray dryer walls occurred during testing, and solids collected
at the bottom of the spray dryer contained moisture levels below 4 percent.
These tests were conducted at an 18°F approach to adiabatic saturation at the
dryer outlet, just as were all previous tests.
Material balance calculations indicate that, for a coal such as the Illi-
nois coal fired in the high sulfur tests at Riverside, a chloride content
of around 0.3 percent would provide the chloride levels of this test. This
would be an uncharacteristically high chloride level for a typical 3.5 percent
sulfur coal. However, using published bulk prices, it appears that calcium
chloride could be delivered to a typical plant site for approximately $225/ton
on a 100 percent CaCl2 basis^. For $70/ton lime, delivered, a net savings
would result from chloride addition at Riverside if only a 15 percent
reduction in lime consumption resulted. The actual saving observed was well
above 15 percent. It appears to be economic in this case, disregarding
capital cost considerations, to operate at high chloride levels even if virtu-
ally all of the chloride must be added as calcium chloride. For a 3.5 percent
sulfur coal with a higher chloride content (0.1 percent or better), the
economics would likely be improved. Additionally, a makeup water source with
a significant chloride content, such as some cooling tower blowdowns, would
further improve these economics.
MASS LOADING MEASUREMENT RESULTS
Table 4 presents mass loading results for both the high sulfur and low
sulfur test periods. The results show that the spray dryer increases the
grain loading at the fabric filter inlet to 3 to 4 times that of the spray
dryer inlet value. The data also show that particulate removal levels
remained high throughout the test program. Removal efficiencies across the
fabric filter varied from 99.95 to over 99.99 percent. The emission levels in
Table 4 are expressed as grains per dry standard cubic foot. Calculation of
emission levels in pounds per million Btu cannot be completed until the
results of coal ultimate analyses are available, but the pounds per million
Btu values should be approximately 1.5 times that in grains per dry standard
cubic foot. Therefore, emission rates varied between approximately 0.002
lb/106 Btu and 0.012 lb/106 Btu, with most values below 0.006 lb/106 Btu. The
few higher values were related to minor bag problems in one compartment. In
all cases, particulate emission rates measured were well below the current
NSPS level for utility boilers of 0.03 lb/106 Btu.
10-127
-------
TABLE 4. FLUE GAS MASS LOADING SUMMARY
Sampling Mass Loading, gr/dscf
Location Low Sulfur Tests High Sulfur Tests
Spray Dryer Inlet 3.2 to 4.1 2.8 to 4.1
Spray Dryer Outlet 11.0 to 13.8 14.9 to 17.2
Fabric Filter Outlet 0.001 to 0.003 0.001 to 0.008
SDLFURIC ACID MEASUREMENT RESULTS
Flue gas sulfuric acid concentration measurements were also conducted.
For the low sulfur tests, no measurable sulfuric acid levels were detected at
either the spray dryer inlet or fabric filter outlet. The inability to
measure sulfuric acid at the dryer inlet is apparently related to the alkaline
nature of the Sarpy Creek coal ash. It is not clear whether the alkaline ash
removes all sulfuric acid upstream of the spray dryer, or whether any sulfuric
acid present is removed on the ash collected on the heated filter in the
sampling train. At the outlet of the fabric filter though, it is clear that
no sulfuric acid is present. Measurements could not be made at the spray
dryer outlet due to the high grain loading at that point, as the upstream
filters tended to plug before an appreciable amount of flue gas could be
sampled.
Measureable levels of sulfuric acid were found during the high sulfur
test periods. Spray dryer inlet values were measured at 2 to 6 ppm 803.
Fabric filter outlet values varied from 0.1 to 0.5 ppm. On a limited number
of days where the spray dryer inlet and fabric filter outlet 803 concentra-
tions were measured simultaneously, removal efficiencies of 90 percent or
better across the system were indicated.
SUMMARY AND CONCLUSIONS
Based on preliminary data reduction for the 3-month test program on the
NSP Riverside dry FGD system, the following conclusions are apparent:
• In general, the Riverside system ran quite well. None of
the problems anticipated for spray dryer systems, such as
rotary atomizer wheel pluggage, buildup of wet solids on
dryer vessel walls, or wetting of fabric filter bag sur-
faces during upset conditions, were observed.
10-128
-------
• Some problem areas at Riverside appear to be potential
sources of problems on other similar dry FGD systems.
These include typical problems with mixing and pumping
slurries with a high solids content, solids handling
equipment which requires continual maintenance, and some-
times inadequate atomizer protection during upset condi-
tions.
• At sulfur levels up to a nominal 3.5 percent, high SC>2
removal efficiencies (90 percent) were readily achievable
in the relatively short-term periods of this program. For
the low sulfur Sarpy Creek coal/coke mixture, substoichio-
metric amounts of lime were required even at 90 percent
S02 removal. This was attributed to the alkaline nature
of the Sarpy Creek coal ash. For the high sulfur Illinois
coal, 90 percent SC>2 removal required reagent ratios of
approximately 1.3 to 1.4 moles lime per mole of inlet S02.
• Calcium chloride addition to the atomizer feed slurry to
achieve chloride levels of approximately 1 percent in the
fabric filter solids catch appeared to be successful in
promoting lime utilization. For the high sulfur tests,
the lime reagent ratio to achieve 90 percent SC>2 removal
was reduced from 1.3 to 1.4 down to a range of 0.9 to 1.1
moles lime per mole of inlet SC>2. This chloride level
would correspond to 0.3 percent chloride in a nominal 3.5
percent sulfur coal. Even for a low chloride, high sulfur
coal, high chloride levels achieved through calcium chlor-
ide addition appear to be cost effective for reducing lime
consumption.
• Particulate control efficiencies were high throughout the
test program, maintaining outlet grain loadings well below
required levels. In spite of baghouse operation within
18°F of the adiabatic saturation temperature and very high
baghouse inlet grain loadings, no bag-fabric-related
problems were observed and flange-to-flange pressure drop
remained acceptably low.
• Based on a limited number of simultaneous measurements
during the high sulfur test periods, sulfuric acid removal
levels of 90 percent or greater were observed.
REFERENCES
1. Karlsson, Hans T., et al., "Activated Wet-Dry Scrubbing of S02." Jour-
nal of the Air Pollution Control Association. Volume 33, No. 1, Jan-
uary 1983. pp. 23-28.
10-129
-------
Hansen, Svend Keis, et al. "Status of the Joy/Niro Dry FGD System and
Its Future Application for the Removal of High Sulfur, High Chloride and
NOX from Flue Gas." Publication No. 83-JPGC-APC-8, The American Society
of Mechanical Engineers, New York, NY. October, 1983.
Chemical Marketing Reporter. Volume 224, No. 15, October 10, 1983.
Schnell Publishing Company, Inc., New York, NY.
APPENDIX
CONVERSION OF BRITISH ENGINEERING (ENGLISH) UNITS TO SI UNITS
To Obtain
ng/J
m
m2
m3
mg / Nm
°C
KW
joules /kg
m / sec/m
tonnes
kg
m / sec
nr / s e c
From
lb/106 Btu
ft
ft2
ft3
gr/dscf
°F
hp
Btu/lb
cfm/ft2
tons (short)
Ib
cfm
gpm
Multiply by
(or use equation)
430
0.3048
9.29
2.83
1.83
tc (°C) = 5/9[tf(°
0.746
1.33
5.08 x
0.907
0.454
4.72 x
6.31 x
x 10~2
x 10~2
F) - 32]
x 10~4
io-3
10~4
10-5
10-130
-------
OVERVIEW AND EVALUATION OF TWO YEARS OF OPERATION
AND TESTING OF THE RIVERSIDE SPRAY DRYER SYSTEM
J. M. Gustke, W. E. Morgan, S. H. Wolf
-------
OVERVIEW AND EVALUATION OF TWO YEARS OF OPERATION
OF THE RIVERSIDE SPRAY DRYER SYSTEM
by: John M. Gustke and Wayne E. Morgan, Ph.D.
Black & Veatch, Engineers-Architects
Steven H. Wolf
Northern States Power Company
ABSTRACT
Initial operation of the 100 MW spray dryer system at Northern States Power Company's
Riverside Generating Station began in November 1980. At that time, a comprehensive test program
was initiated to demonstrate the suitability of dry flue gas desulfurization (FGD) for utility coal
fueled boilers. Since initial operation of the Riverside spray dryer began, other publications have
described individual aspects of this system's performance. This paper provides a comprehensive
analysis and overview of the performance test data collected during the initial two-year period
of operation of the Riverside spray dryer.
Overall data correlations describing important variables in spray dryer operation and per-
formance are established from test results obtained during operation of the Riverside spray dryer
system under a wide range of conditions. Correlations between total and fabric filter sulfur dioxide
(802) removal and parameters such as lime stoichiometric ratio, total alkalinity, and approach
temperature are presented for several different coals. Variations in moisture content of the solids
collected in the spray dryer and fabric filter are evaluated to establish the sensitivity to a wide range
of operating variables. In addition, system operation and control experiences are described to
illustrate the interaction between flue gas flow, feed slurry flow, absorber outlet temperature,
and SO2 emissions during normal operation, as well as during transient conditions such as start-up,
shutdown, and load swings. The effects of boiler soot blowing on the flue gas saturation tempera-
ture and system control are also discussed.
INTRODUCTION
Joy Manufacturing Company, Niro Atomizer Incorporated, and Northern States Power
Company (NSP) jointly funded the 100 MW spray dryer demonstration program at the Riverside
Generating Station. Design support for the spray dryer installation was provided by Black & Veatch,
Engineers-Architects. The purpose of the Riverside demonstration program was to confirm that a
full size spray dryer installed on a utility coal fueled power station could reliably and economically
meet current SOo and particulate emission regulations. A single, full-size spray absorber module
followed by a fabric filter was installed and began initial operation at Riverside Units 6 and 7 in
November 1980. Testing of the Riverside system began in March 1981 and remains in progress.
10-131
-------
This paper presents an overview of all performance test results obtained during 1981 and
1982. Overall data correlations are developed and relationships are established describing the sulfur
dioxide removal efficiency and waste solids drying performance of the system. Fairly wide varia-
tions in some of the performance test results were observed over the extended test period. The
observed range of performance appropriately reflects the actual performance of a full scale utility
dry scrubber system subject to normal variations in operating conditions. Operational performance
and control characteristics of the system are also presented. Major system parameters which are
monitored and controlled during start-up, shutdown, and rapid load changes are discussed. The im-
pacts of steam generator soot blowing on operation of the Riverside spray dryer are also described.
SYSTEM DESCRIPTION
Riverside Units 6 and 7 have identical Babcock & Wilcox steam generators which began
operation in 1949 and 1950, respectively. The combined generating capacity of these units is
approximately 100 MW. The design fuel for the steam generators is a 10,800 Btu/lb Illinois coal
with a sulfur content of 3.5 per cent. Prior to the installation of the spray dryer and fabric filter
system, each unit had individual electrostatic precipitators, ID fans, and chimneys.
The retrofit spray dryer system at Riverside treats the combined flue gas flow from Units 6
and 7. Figure 1 presents a simplified flow diagram of the spray dryer system. Flue gas at the air
heater outlet from both units is combined and drawn through the single spray dryer module. Design
features of the 46-foot diameter spray dryer module include a compound gas disperser and a single
rotary atomizer. After flowing through the spray dryer, flue gas is directed to either the fabric filter
or the original weighted wire precipitators. The fabric filter uses reverse gas cleaning and has twelve
compartments. The gas flow at the discharge of the particulate collectors is divided and flows to the
original Unit 6 and 7 chimneys.
Both a ball mill and an attrition slaker are installed for lime slurry preparation. Lime slurry is
fed to a mix tank where it is mixed with recycled material from the spray dryer and particulate
collector before being pumped to the atomizer. The recycle material is collected from the spray
dryer and particulate collector (fabric filter or electrostatic precipitator) hoppers and consists of
flyash, unreacted lime, and flue gas desulfurization reaction products. Material collected in the
spray dryer and particulate collector hoppers which is not recycled is stored in silos prior to dry
landfill disposal.
PERFORMANCE TEST RESULTS
During 1981 and 1982, 164 performance tests were conducted at Riverside. Most of the
testing during this time consisted of short term parametric tests which were conducted to determine
the influence of various individual parameters on system performance. Longer term system per-
formance was evaluated by conducting five sets of demonstration tests. During the demonstration
test periods, the system was maintained at relatively stable operating conditions for 5 to 10 days.
Figure 2 summarizes the schedule for the performance tests which comprise the data presented
herein.
During the two years of testing at Riverside, the operating conditions varied widely. Five
different fuels were tested: Colstrip coal, Colstrip/Coke blend, Sarpy Creek coal, Sarpy Creek/Coke
blend, and high sulfur Illinois coal. The Colstrip and Sarpy Creek coals are low sulfur coals from
10-132
-------
Flue Gas
From Unit 7
Boiler
Flue Gas
From Unit 6
Boiler
o
I
LO
OJ
Water
Fabric
Filter
1 1
Flue Gas to
Chimneys
Solids to
Disposal
—111— Water
Mix
Tank
Classifier
K7-
Slurry
Transfer
Pumps
Feed
Tank
K7-
Feed
Pumps
Lime Slurry
Pumps
FIGURE 1 RIVERSIDE SPRAY DRYER SYSTEM
-------
o
Spray Dryer
System Slart-Up
Parametric Tests,
Colstrip Coal
Demonstration Tests,
Colstrip Coal
Electrostatic Precipitator
Tests, Colstrip Coal
Parametric and
Demonstration Tests,
Illinois Coal
Demonstration Tests,
Sarpy Coal
Miscellaneous Tests,
Various Operating
Conditions
Extended Units 6 and 7
Outage
Parametric and
Demonstration Tests,
Sarpy Coal
1980
O , N , D
1981
F ,M| A |M| 1 , 1 , A , S
D
1982
1IF|M|A|M|||||AISI0,N|D
FIGURE 2 RIVERSIDE SPRAY DRYER PERFORMANCE TEST SCHEDULE
-------
Montana. Table 1 presents some of the typical properties for each coal. Tests were performed with
gas flows at the absorber inlet ranging from 34 to 114 per cent of design flow (170,000 acfm to
570,000 acfm). Approach temperatures ranged from 16 to 78 F. Feed slurry was atomized at fine,
intermediate, coarse, and very coarse levels.
This paper focuses on the effects of a range of variable operating conditions on SOo removal
performance and waste moisture content (dryness of the solid waste products). The test results are
primarily viewed on the whole rather than as isolated individual groups of data or periods of testing
to establish specific correlations. The results presented in this analysis should reflect the effects on
performance of a wide range of operating conditions judged representative of those that full scale
units will routinely experience.
SO2 REMOVAL CORRELATIONS
LIME STOICHIOMETRIC RATIO
The ability to economically achieve SO2 removal efficiencies required by New Source Per-
formance Standards (NSPS) is a requirement for any flue gas desulfurization (FGD) system. As a
measure of the Riverside dry scrubber performance, total SOo removal versus lime stoichiometric
ratio (moles of lime added per mole of SO2 removed) was plotted for all of the 1 64 performance
tests. A multiple linear regression analysis was used to establish correlations between SOo removal,
lime stoichiometric ratio, and approach temperature. Figure 3 presents the results of this analysis,
applied to all 164 tests. The multiple linear regression technique of data correlation utilizes test data
to generate a correlation between related parameters. Multiple linear regression is not a curve fitting
technique. The curves presented on Figure 3 were generated using the regression equation estab-
lished for total SO2 removal and lime stoichiometric ratio. Curves were calculated for approach
temperatures of 18 F and 40 F. These approach temperatures are typical of low and high approach
temperatures frequently selected as design points for utility spray dryer installations. Since the tests
were performed at a variety of approach temperatures, the distribution of test results around the
regression curves is shown for comparative purposes only.
As shown on Figure 3, both the lime stoichiometric ratio and approach temperature affect
the SOo removal. For a specific lime stoichiometric ratio, decreases in the approach temperature
resulted in increased SOo removal. The form of the regression equation used to generate correla-
tions between SO2 removal and lime stoichiometry is as follows:
SRE=
where
SRE = SO2 removal efficiency, per cent
A, B = Regression coefficients
LSR = Lime stoichiometric ratio
AA = Ash Alkalinity constant
AT = Approach temperature, degrees F
K = Regression constant
Table 2 lists the coefficients and constants for the regression equation shown above, as well
as the number of tests conducted using each fuel and the average SO2 removal efficiency. The value
10-135
-------
TABLE 1. TYPICAL PROPERTIES OF COALS AND FUEL BLENDS BURNED
AT RIVERSIDE DURING 1981 AND 1982
Fuel
Colstrip
Colstrip/Coke
Sarpy Creek
Sarpy Creek/Coke
Illinois
Heating
Value
Ib/MBtu
8,700
9,200
8,300
9,200
10,900
Ash
Content
per cent
8.3
6.7
10.8
8.9
10.4
Sulfur
Content
per cent
0.8
1.2
0.8
1.2
3.2
TABLE 2. REGRESSION EQUATION COEFFICIENTS FOR SO2 REMOVAL
AS A FUNCTION OF LIME STOICHIOMETRIC RATIO AND APPROACH
TEMPERATURE FOR ALL RIVERSIDE TESTS
Fuel
Colstrip/Coke
Sarpy Creek/Coke
Illinois**
Sarpy Creek
Sarpy Creek (no lime)
Colstrip
All Fuels
Tests
Average SO2
Removal A
94
24
22
17
4
3
164
85.9
88.0
92.0
75.4
30.6
69.3
84.3
Coefficients/Constants
-16.9
-79.3
-29.2
B
AA
K
Standard
Error of
Estimate
-0.65 0.42 114.9 9.1
-0.38 0.42 148.5 6.1
-0.36 99.9 3.9
-0.21 0.42 105.0 6.9
-0.66 0.42 124.8 8.6
*The value for the ash alkalinity constant (AA) was established by optimizing the "all fuels" regression equation. The
addition of such a constant to the other equations also improved the data correlations, and the same value was assumed for each case.
"For the limited range of data available, no significant correlation with lime stoichiometric ratio was found.
10-136
-------
o
I
to
100-
80-
c
V
U
O.
_* 60
ro
>
O
E
o»
at
6 40
20-
Approach Temp. = 18 F
Approach Temp. = 40 F
The Regression Equation Used to Develop the Specific
Curves on This Figure Accounts for Variations in Test
Conditions for Each Displayed Data Point.
- - Approach Temperature Below 25 F
- Approach Temperature Above 25 F
0.0
0.2
0.4 0.6 0.8 1.0 1.2 1.4
Lime Stoichiometric Ratio (Based on SO2 Removed)
1.6
1.8
2.0
FIGURE 3 SO2 REMOVAL AS A FUNCTION OF STOICHIOMETRIC RATIO
AND APPROACH TEMPERATURE (ALL FUELS)
-------
of the ash alkalinity constant (AA) was selected by optimizing the regression equation using all of
the test data. Inclusion of this same constant in the individual fuel regression equations improved
these correlations as well. Insufficient data were available to optimize this value for each case.
Based on an analysis of the entire range of data, factors such as degree of atomization, gas
flow, SO-, inlet concentration, gas retention time, and amount of recycle could not be directly
correlated" to SCh removal for the limited variation in these values tested at Riverside. However,
pilot testing has demonstrated that factors such as these are significant in spray dryer design and
performance. The results show that for the full scale Riverside spray dryer system operating over a
range of conditions, stoichiometric ratio, approach temperature, and ash alkalinity are the dominant
process parameters which control SC>2 removal.
A separate regression analysis was performed on each fuel to determine individual contribu-
tions to the overall regression equation. Figures 4 through 6 show the SC>2 removal efficiency
predicted by regression equations obtained for each fuel, plotted along with the respective data
points. The Colstrip/Coke test data shown on Figure 4 correspond primarily with the mid-range of
the SO-> removal and lime stoichiometries for the overall correlation presented on Figure 3. The
Sarpy Creek coal test data shown on Figure 5 generally correspond to the lower SC>2 removals and
stoichiometries in the data base. The Illinois coal test data shown on Figure 6 comprise most of the
high SO-) removals and corresponding high lime stoichiometries in the data base. The regression
equations obtained from the analysis of test results from each fuel, indicate varying sensitivities to
approach temperature and lime stoichiometric ratio. The Sarpy Creek coal is the most sensitive to
lime stoichiometry but the least sensitive to apporach temperature. The Illinois coal did not exhibit
much sensitivity to stoichiometry but was sensitive to approach temperature. The Colstrip/Coke
blend was the most sensitive to approach temperature and it also was moderately sensitive to
stoichiometric ratio.
TOTAL ALKALINITY RATIO
The alkalinity ratio is a measure of the total alkaline material which is fed to the atomizer. In
addition to the fresh lime, the alkaline material includes alkalinity in the ash and unreacted lime in
the recycled solids. Since the lime stoichiometric ratio provided a good correlation with total S02
removal, the alkalinity ratio was expected to provide a better correlation. Total SOn removal and
the corresponding alkalinity ratio for all 164 tests are shown on Figure 1'. The overall trend of data
on this figure indicates that there is a correlation between SO2 removal and alkalinity ratio. How-
ever, a significant number of tests have alkalinity ratios of less than 1.0. Since alkalinity ratio is
defined in this case as equivalent moles of CaO per mole of SO2 removed, the minimum theoretical
alkalinity ratio is 1.0. During the Riverside test program two different techniques were used to
measure the total alkalinity ratio. The results obtained using each technique differed and neither
technique appeared to provide consistent and reasonable results. Thus, no correlations were made
using this parameter.
S02 REMOVAL IN THE PARTICULATE COLLECTOR
The total SO2 removal which occurs across a dry scrubber system consists of SO2 removed in
the spray dryer and SO2 removed in the particulate collector. SO2 removal in the particulate col-
lector results from the additional contact of the alkaline solids with the remaining SO2 in the flue
gas. SO2 removal in the particulate collector can be affected by process parameters in a different
10-138
-------
o
I
100-
«- 80
c
-------
100
o
i
e
lit
U
k.
01
Q.
O
E
UJ
at
80
60-
40-
20-
0.0
Approach Temp. = 18 F
Approach Temp. = 40 F
The Regression Equation Used to Develop the Specific
Curves on This Figure Accounts for Variations in Test
Conditions for Each Displayed Data Point.
- Approach Temperature Below 25 F
- Approach Temperature Above 25 F
0.2
0.4 0.6 0.8 1.0 1.2 1.4
Lime StoichSometric Ratio (Based on SO2 Removed)
1.6
1.8
FIGURE 5 SO2 REMOVAL AS A FUNCTION OF STOICHIOMETRIC RATIO
AND APPROACH TEMPERATURE (SARPY CREEK COAL)
-------
o
I
c
01
U
O
E
O
1/5
100-
80 -J
601
40-
20-
0.0
Approach Temp. = 18 F
Approach Temp. = 40 F
The Regression Equation Used to Develop the Specific
Curves on This Figure Accounts for Variations in Test
Conditions for Each Displayed Data Point.
- Approach Temperature Below 25 F
- Approach Temperature Above 25 F
0.2
0.4 0.6 0.8 1.0 1.2 1.4
Lime Stoichiometric Ratio (Based on SOz Removed)
1.6
1.8
FIGURE 6 SO2 REMOVAL AS A FUNCTION OF STOICHIOMETRIC RATIO
AND APPROACH TEMPERATURE (ILLINOIS COAL)
-------
o
I
100-
c
Ol
U
>
o
E
OJ
(V
*E^
O
80-
60-
40-
20-
0.0
0.2
C CD
3 ,
<»^
- Approach Temperature Below 25 F
- Approach Temperature Above 25 F
0.4 0.6 0.8 1.0 1.2
Alkalinity Ratio (Based on SO2 Removed)
1.4
1.6
1.8
2.0
FIGURE 7 SO2 REMOVAL AS A FUNCTION OF ALKALINITY RATIO
AND APPROACH TEMPERATURE (ALL FUELS)
-------
manner than SC>2 removal in the spray dryer. To determine the sensitivity of SOo removal in the
particulate collector to various process parameters, a separate analysis was performed.
Figure 8 presents total and fabric filter SC^ removal efficiency as a function of lime stoichi-
ometric ratio. As shown in this figure, fabric filter SC>2 removal efficiencies generally range between
10 to 20 per cent and average approximately 15 per cent. Figure 9 presents total and electrostatic
precipitator SOo removal as a function of lime stoichiometric ratio. As shown in this figure, electro-
static precipitator (ESP) SC>2 removal ranges from 0 to 20 per cent and averages approximately 6
per cent. Higher average SC>2 removal efficiencies in the fabric filter than in the ESP are attributed
to more effective interaction between the flue gas and the particulate in the filter cake on the filter
bag surfaces. Although measurements were not consistently available from the performance test
data, factors such as fabric filter gas-to-cloth ratio, pressure drop, and frequency of bag cleaning are
expected to have a significant effect on fabric filter SC>2 removal.
An analysis was performed to determine the sensitivity of fabric filter SO-, removal to the
inlet SC>2 concentration. Figure 10 presents a comparison of these parameters for all the fuels. No
significant correlation could be found in the Riverside data between fabric filter SC>2 removal and
inlet 862 concentration. A similar analysis performed on each fuel individually did not yield an
improved correlation. Figure 1 1 presents a comparison of the Riverside fabric filter SC>2 removal as
a function of approach temperature for all the fuels. Flue gas approach temperature to the dew
point establishes the relative moisture content of the flue gas. High flue gas moisture content should
result in improved SC>2 removal in the fabric filter. Contrary to what was expected, there appeared
to be no correlation between fabric filter SC>2 removal and approach temperature. A similar analysis
was performed on each of the fuels; however, no significant correlations were obtained. Correlations
between fabric filter SC>2 removal and approach temperature have been shown to exist in laboratory
and pilot scale units. However, on a full scale operating unit, variations in other process parameters
such as fabric filter gas-to-cloth ratio and frequency of cleaning may diminish the influence of
approach temperature on SC>2 removal.
An interesting correlation which was apparent from the test data was the correlation between
fabric filter SC>2 removal and spray down temperature. Spray down temperature is the difference
between the spray dryer inlet and outlet temperatures. Spray down temperature determines the
total moisture content of the flue gas but is independent of the dew point temperature. Figure 12
shows the regression equation curve established for these parameters. As shown on Figure 12, fabric
filter SOo removal appears to increase with increased spray down temperatures. The composite
regression equation developed for all fuels (all tests with the fabric filter) for fabric filter SC>2
removal and spray down temperature is as follows:
A
FFSRE = + K,
where
FFSRE = Fabric filter SO-» removal efficiency, per cent
A = Regression coefficient (-2204.6)
SDT = Spray down temperature, per cent
K = Regression constant (28.2)
Analysis of fabric filter SC>2 removal versus spray down temperature for each individual fuel
did not produce improved correlations. The Riverside spray dryer is operated to maintain a fixed
10-143
-------
O
I
100
c
0)
U
I 80
O
E
a;
^ 60 H
.2
3
"O
c
40-
• 20 ^F
J3
fO
0.0
dC
cS-'c
;§;
CE
Total SOi Removal Efficiency
Fabric Filter SO? Removal Efficiency
A± A
*
0.2
0.4 0.6 0.8 1.0 1.2 1.4
Lime Stoichiometric Ratio (Based on SO2 Removed)
1.6
1.8
2.0
FIGURE 8 TOTAL AND FABRIC FILTER SO2 REMOVAL AS A FUNCTION OF
LIME STOICHIOMETRIC RATIO (ALL FUELS)
-------
h-1
O
I
I—"
-P-
L/i
c
U 100
I 80
n
QC
rs
O
60
13
40-
.
'w
0)
« 20H
o
^^
u
0.0
- Total SO2 Removal Efficiency
- Electrostatic Precipitator SOa
Removal Efficiency
CD
0.2
0.4 0.6 0.8 1.0 1.2 1.4
Lime Stoichiometric Ratio (Based on SO? Removed)
1.6
FIGURE 9 TOTAL AND ELECTROSTATIC PRECIPITATOR SO2 REMOVAL
AS A FUNCTION OF LIME STOICHIOMETRIC RATIO (COLSTRIP/COKE FUEL)
1.8
2.0
-------
O
I
c
0*
U
O
E
24-i
20-
16-
12-
8-
4-
O
O
«>
CD
a + c c
1 T>
O ^O CD
C CD O<
CDS
CDC"-
c - Colstrip/Coke
A- Illinois
-I-- Sarpy Creek
«- Sarpy Creek/Coke
x- Colstrip
80
160 240 320 400 480 560
Fabric Filter Inlet SCh Concentration, PPM
640
720
FIGURE 10 FABRIC FILTER SO2 REMOVAL AS A FUNCTION OF
FABRIC FILTER INLET SO2
-------
O
I
c
U
O
E
- Colstrip/Coke
•- Illinois
- Sarpy Creek
- Sarpy Creek/Coke
- Colstrip
45
50
55
-------
24
o
i
CO
c
-------
proportion of solids and water in the atomizer feed. Thus, increasing the spray down temperature
also increases the amount of particulate matter in the flue gas at the fabric filter inlet. From the
Riverside data it appears that the greater the amount of particulate matter entering the fabric filter,
the greater the amount of SC>2 removed in the fabric filter. Higher particulate concentrations in the
fabric filter may provide a higher ratio of alkaline material on the bags which improves SO-> removal
efficiency. Other parameters such as gas-to-cloth ratio, pressure drop, and cleaning frequency may
also contribute to this correlation.
Due to the small size of the data base, no attempts were made to correlate precipitator SOo
removal with any of the process parameters. As previously discussed, the average SO^ removal was
less in the ESP than in the fabric filter. However, the Riverside precipitators are relatively small and
were not designed for spray dryer service. New dry scrubber installations using precipitators may
achieve slightly higher SO2 removal efficiencies in the particulate collector than those achieved
using the precipitators at Riverside.
WASTE MOISTURE CORRELATIONS
The ability of the spray dryer to produce a waste product which is relatively dry is a critical
requirement of the system. As a measure of the ability of the Riverside dry scrubber to achieve a
dry product, the moisture content of the solids discharged from the spray dryer cone and fabric
filter hoppers was periodically measured during each performance test.
The average and peak moisture contents of the solids recorded during the performance tests
are as follows.
Average Moisture Peak Moisture
Content Content
(per cent) (per cent)
Spray Dryer Solids 1.5 4.0
Fabric Filter Solids 1.0 2.3
Generally, the solids were sufficiently dry to be conveyed by conventional ash handling
equipment.
To determine the effect of major process parameters on the waste moisture content, a com-
parison of spray dryer waste moisture and approach temperature was made. Figure 13 shows the
corresponding approach temperature for each waste moisture content. Although there is a general
trend for decreasing waste moisture with increasing approach temperature, a strong correlation was
not apparent from the Riverside test data. Since approach temperature is a relative indication of the
amount of moisture in the flue gas, it was expected to correlate with spray dryer waste moisture.
Additional comparisons of waste moisture with spray down temperature, feed solids content,
atomization, and mass flow rate were made. In each case, no strong correlation with waste moisture
was apparent. A similar analysis was performed on each fuel, but no significant correlations were
found. The range of recorded spray dryer waste moisture at Riverside is very narrow. Correlations
may exist between these parameters; however, on a full scale operating unit, the range of operating
conditions would not normally fluctuate enough for these correlations to be apparent.
10-149
-------
3.0
o
i
Ui
o
c
CD © ©
CD CD
A
CD CD
C- •* CD
O CD
O
a o
O
CD
CD
O
1 o
CD CD
CD
CD
CD
CD - Colstrip/Coke
A - Illinois
+ - Sarpy Creek
«•- Sarpy Creek/Coke
x - Colstrip
15
20
25 30 35 40
Approach Temperature, F
45
50
55
FIGURE 13 SPRAY ABSORBER WASTE MOISTURE AS A FUNCTION OF
APPROACH TEMPERATURE
-------
One correlation which was apparent from the test data was the correlation between spray
dryer and fabric filter waste moisture. Figure 14 shows this correlation. As presented on Figure 14,
fabric filter waste moisture content increases linearly with spray dryer waste moisture. The com-
posite equation developed for all fuels (all tests) is as follows:
FFWM = A * SDWM,
where
FFWM = Fabric filter waste moisture, per cent
A = Regression coefficient (0.67)
SDWM = Spray dryer waste moisture, per cent
This correlation reflects the longer drying time of the solids in the fabric filter. Flue gas
passing through the filter cake on the fabric filter bags provides additional drying of the collected
solids.
OPERATIONAL PERFORMANCE AND CONTROL
SYSTEM CONTROL
A discussion of the primary control parameters is necessary before discussing the effects of
start-up, shutdown, and rapid load swings on the performance of the spray dryer. Control of the
Riverside spray dryer is based on maintaining a constant approach temperature at the spray dryer
outlet. The approach temperature is held constant by maintaining the required evaporation rate in
the spray dryer to quench the flue gas to the desired temperature above the adiabatic saturation
temperature. Moisture is added to the gas stream through the atomizer with the additive feed slurry,
while the fraction of solids in the atomizer feed is maintained at a constant level of approximately
35 per cent. To decrease the approach temperature, the total flow to the atomizer is increased.
To control the SO2 removal efficiency of the dry scrubber, the ratio of lime and recycled
solids in the feed slurry is changed, but the total amount of solids and water in the atomizer feed
remains constant. To increase SOo removal, the ratio of the lime to feed solids is increased and the
amount of recycle material is decreased. Conversely, to decrease SO-> removal, the amount of lime
in the atomizer feed is decreased and the amount of recycle material is increased. Due to the resi-
dence time in the slurry preparation system, changes in SO2 removal efficiency can not be made
instantaneously unless a change in approach temperature is also desired. Since the approach tem-
perature has a significant effect on SO2 removal, it is difficult to stabilize the absorber outlet
temperature if approach temperature is used to control SO-> removal. However, since only an
average SO2 removal rate is required, periods of low SO2 can be offset by periods of higher SO2
removal.
SYSTEM START-UP
The start-up of the Riverside spray dryer can be accomplished whenever the flue gas tempera-
ture and flue gas flow rate are high enough to allow proper drying in the chamber. At Riverside, a
minimum gas flow of 125,000 ft^/min (25 per cent gas flow) at 200 F is necessary to meet the
minimum requirements for spray dryer operation. Start-up of the spray dryer is performed by an
automatic control system which requires only limited operator interface once feed slurry flow is
10-151
-------
o
I
c
01
U
^
01
a.
'5
0>
o
E
The Regression Equation Used to Develop the Specific
Curve on This Figure Accounts for Variations in Test
Conditions for Each Displayed Data Point.
- Colstrip/Coke
- Illinois
- Sarpy Creek
- Sarpy Creek/Coke
- Colstrip
1.0 1.5 2.0 2.5 3.0 3.5
Spray Absorber Waste Moisture, Per Cent
FIGURE 14 PARTICULATE REMOVAL DEVICE WASTE MOISTURE AS A FUNCTION OF
SPRAY ABSORBER WASTE MOSITURE
-------
established to the atomizer. Figure 15 shows a typical warm start-up of the spray dryer at Riverside.
As shown on Figure 15, the gas flow is increased as the boilers come on line, and 807 is generated
when coal firing begins. The feed slurry flow to the atomizer is initiated when the minimum gas
flow and temperature are reached. SC>2 removal begins immediately when feed slurry flow is estab-
lished. As the boilers continue to gain load, the slurry flow to the atomizer increases accordingly to
maintain the desired absorber outlet temperature. When the absorber outlet temperature has stabi-
lized, the automatic SC>2 removal control system adjusts the amount of lime in the feed to achieve
the desired SC>2 removal. The time required for complete stabilization at the desired SC>2 removal is
dependent primarily on the composition of the feed stock prior to start-up. The time to reach stable
SC>2 removal also varies for hot, warm, or cold steam generator start-ups, because of differing initial
flue gas temperatures.
SYSTEM SHUTDOWN
The flue gas flow and temperature requirements used for the shutdown of the spray dryer are
the same as for the start-up, but it is easier to reduce transient 862 peaks during a controlled boiler
shutdown because sufficient gas flow and temperature are present to allow spray drying until after
the steam generator fire is extinguished. The feed slurry to the atomizer is stopped when the gas
flow is no longer sufficient to ensure proper drying and the atomizer is flushed and taken out of
service. As shown in Figure 16, the spray dryer follows load as the boilers are ramping down and
comes off line without SC>2 emissions increasing.
RAPID LOAD SWINGS
The Riverside spray dryer is often subject to rapid load swings. As shown in Figure 17, the
system automatically responds to the increased and decreased gas flow. The feed slurry flow rate
follows closely to the gas flow, resulting in an almost constant outlet temperature. The load change
illustrated on this figure was accomplished by bringing one of the boilers on line after the other was
operating at steady load. This results in some dilution of the SO2 concentration into the spray dryer
and accounts for the reduced emissions during this period.
WET BULB TEMPERATURE FLUCTUATIONS
To control a spray dryer at a constant approach temperature, it is necessary to measure the
adiabatic saturation (wet bulb) temperature of the flue gas. At Riverside, wick-and-water reservoir
devices are installed at the fabric filter outlet to measure and transmit wet bulb temperature to the
control system. It is necessary to locate these devices downstream of the fabric filter to minimize
the fouling of the wick by particulate matter. Although the wet bulb measurement devices provide
an accurate and reasonably reliable indication of the wet bulb temperature downstream of the
baghouse, they do not provide precise information on the wet bulb temperature in the spray dryer
chamber. Fluctuations in wet bulb temperature at the spray dryer inlet are not transmitted to the
control system until 30 to 90 seconds after they occur. In addition, by the time the gas passes
through the fabric filter and reaches the outlet duct, the heat transfer and air leakage reduce the wet
bulb temperature by 1 to 3 F from that at the spray dryer inlet.
To determine the type of wet bulb fluctuations which may occur at the spray dryer inlet, a
temporary device was installed and continuously monitored. Figure 18 presents the wet bulb
temperature fluctuations at the inlet of the spray dryer during a typical soot blowing sequence using
10-153
-------
350,000
o
i
140
120
100
Q.
U
80
_
1/1 60
-o
OJ
40
300,000 -
250,000 -
200,000 -
£
O
•5 150,000 -
100,000 -
20 - 50,000 -
0
Gas Flow
Feed Slurry Flow
Spray Dryer
Outlet Temperature
SO2 at Chimney
J L
275
250
225
200 E
175
150
125
3
O
£
k.
Q
O.
(SI
700
600
500
400 c
E
IE
U
300 ™
200
100
0 .5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
Time (hr)
FIGURE 15 RIVERSIDE SPRAY DRYER SYSTEM RESPONSE
DURING START-UP
100
-------
350,000 -
o
I
140
120
100
S 80
o
J 60
-c
£ 40
20
_, 250,000
Feed Slurry Flow
10 15 20
25 30 35
Time (min.)
40
Spray Dryer
Outlet
Temperature
» T
275
250 ~
£
2
225 |
Q.
E
0)
200 T,
O
175 £
150 "
125
700
600
500
400
300
200
100
U
100
50 55 60
FIGURE 16 RIVERSIDE SPRAY DRYER RESPONSE
DURING SHUTDOWN
-------
O
I
Ul
ON
120
100
O 80
_o
"• 60
_3
•o 40
20
_o
u.
t/i
fS
300,000
250,000
200,000
150,000
100,000
50,000
Gas Flow
Spray Dryer Outlet Temperature
I I
I I I I
I I
10 15 20
25 30 35
Time (min.)
250
225 £
OJ
5
m
200 5
Q-
E
600
175
150
O
Q
>%
m
400 >p
c
_E
300 u
IS
rs
O
200 ">
100
100
-JO
40 45 50 55 60
FIGURE 17 RIVERSIDE SPRAY DRYER RESPONSE TO RAPID
LOAD FLUCTUATIONS
-------
a
E
3
as
a!
150
140
130
120
110
100
0
12
15
18
Time, Minutes
A. Wall Blowers
(Steam Flow = 6,000 to 8,000 Ib/hr)
«J
3
01
a
E
£
05
a;
150
140
130
120
110
100
a
E
3
as
150
140
130
120
110
100
0
Time, Minutes
B. Economizer Blowers
(Steam Flow = 7,000 to 9,000 Ib/hr)
12
Time, Minutes
C. Retractable Blowers
(Steam Flow = 10,000 to 12,000 Ib/hr)
FIGURE 18 EFFECT OF STEAM SOOTBLOWINC ON SPRAY DRYER
INLET WET BULB TEMPERATURE
10-157
-------
the steam soot blowers. The economizers and furnace wall blowers raised the wet bulb temperatures
approximately 5 F The retractable furnace soot blowers increased the wet bulb temperature by as
much as 8 F At low loads, soot blowing increased the wet bulb temperature by as much as 12 F at
the spray dryer inlet.
Another phenomenon which occurs at Riverside is the stratification of dry bulb and wet bulb
temperatures in the inlet duct to the spray dryer. This is caused by differing operation of the Unit 6
and 7 boilers. Figure 19 presents examples of stratified wet bulb temperatures at the spray dryer
inlet during soot blowing with one unit off line. As shown on Figure 19, a 3 F stratification across
the inlet duct and an 11 F increase over the initial wet bulb temperature occurs during peak soot
blowing. With both units on line - one blowing the furnace walls and the other blowing the air
heater - a wet bulb stratification of 5 F can occur across the duct. It was initially believed that
the stratification would not be significant when the FGD system was treating flue gas from one
boiler. Similar tests were performed at NSP's Sherburne County Unit 1 (750 MW). Measurements in
the two air heater outlet ducts on this unit demonstrated that wet bulb stratification occurred
across each duct and from one duct to the other. The stratification measured at Sherburne County
Unit 1 corresponds with the number and location of the soot blowers in operation. The magnitude
of the stratification at Sherburne Unit 1 was similar to that experienced at Riverside. It is important
for successful low temperature operation of spray dryer systems that the fluctuations and stratifica-
tions in wet bulb temperature be quantified and monitored whenever possible.
CONCLUSIONS
The Riverside spray dryer system has demonstrated the capability to provide compliance with
New Source Performance Standards for SOo emissions. The average removal efficiency was 84.3 per
cent for the 164 tests, including low lime and no lime additive tests. SOo removal efficiencies
exceeded 95 per cent in 30 of the 164 tests. SO7 removal ranging from 70 to 90 per cent removal
was demonstrated for a variety of fuels.
The spray dryer system is capable of drying the waste product to a powderlike consistency.
The moisture content parameters in the data base indicate that the spray dryer system produced a
relatively dry waste product in all of the tests. While operating the spray dryer during the two-year
test period, upset conditions did occur occasionally which raised the moisture content of the waste
solids to relatively high levels. The solids moisture content recorded in the data base represents
typical steady-state utility operating conditions.
From an analysis of the performance test data collected during 1981 and 1982 at Riverside,
certain overall correlations were apparent. A strong relationship between SO2 removal, approach
temperature, ash alkalinity, and lime stoichiometric ratio was observed. Other correlations between
fabric filter SOo removal and spraydown temperature, and between spray absorber and fabric filter
waste moisture content were apparent from the performance test data. Many other correlations
which have been demonstrated on pilot size units were not observed from an evaluation of the
overall data base.
The Riverside dry scrubber system responds well to changes in operating conditions. It was
possible to maintain absorber approach temperature at an almost constant condition during changes
in boiler operations. SO2 removal efficiency typically changed with changing operating conditions;
however, it was usually greater than required, such that it was possible to maintain compliance
10-158
-------
Q.
E
K
"5
OB
5
V
1
a
E
U
»r
5 140
75
S. 130
E
u
>- 120
JD
3
= 110
S 100
^
/ ^ V
, /v /\ / v A^rNiV ^ _
-
i i i i
Boiler 7
Side of
Inlet Duct
to
Spray Dryer
0 3 6 9 12
Time, Minutes
A. Boiler 6 Off Line; Boiler 7 at Low Load
I3U
140
130
120
110
tnn
>W^
^^^X^^^^X^ /s ^_^^X^-^/ ^^
-
-
1 I 1 1 1 1
Boiler 6
Side of
Inlet Duct
to
Spray Dryer
03 6 9 12 15 18
15U
140
130
120
110
inn
_
c~
_
-
1 1 1 1 1 1
Boiler 7
Side of
Inlet Duct
to
Spray Dryer
B.
0 3 6 9 12 15
Time, Minutes
Boiler 6 Wall and Retractable Sootblowing; Boiler 7 Air Heater Sootblowing
18
FIGURE 19 STRATIFICATION OF FLUE CAS WET BLUB
TEMPERATURE AT SPRAY DRYER INLET
10-159
-------
almost continually. Variations in the absorber inlet wet bulb temperatures were observed during
soot blowing operations. The maximum recorded wet bulb temperature variation during soot
blowing was 8 F at full load and 12 F at low load. Sufficient margin must be maintained in the
approach temperature to account for such operational variations.
The Riverside FGD system exhibits normal variations in performance consistent with a full
scale operating system and, based on a broad range of data, is not as sensitive to variations in some
parameters as are laboratory or pilot systems.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the contributions of Niro Atomizer Incorporated and
Joy Manufacturing Company. The performance data presented in this paper were collected as a part
of their testing efforts and were released to Northern States Power Company and Black & Veatch
for independent evaluation. In addition, Northern States Power Company's Riverside operating staff
is to be commended for their ability to endure the extensive testing program conducted at their
installation.
10-160
-------
DESIGN AND INITIAL OPERATION OF THE SPRAY DRYER
FGD SYSTEM AT THE MARQUETTE, MICHIGAN, BOARD OF
LIGHT AND POWER - SHIRAS #3 PLANT
0. Fortune, T. F. Bechtel, E. Puska, J. Arello
-------
ABSTRACT
This paper discusses the design issues, design decisions, start-up, and
early operation of the Spray Dry Flue Gas Desulfurization (SDFGD) system
which went into operation at the Marquette Michigan Board of Light and
Power Shiras #3 in May 1983. This forty-four (44) megawatt unit consisting
of a rotary atomizer reactor, reverse air fabric filter, lime preparation,
and reagent recycle system was engineered in the 1980-82 time period uti-
lizing pilot plant and prototype industrial system results as a design
basis.
The initial operation of
scaleup from pilot plant
the unit is discussed, as is the success of the
to Commercial size boiler.
I. BACKGROUND
The development of the process and equipment design technology on which
the Shiras #3 Spray Dry Flue Gas Desulfurization (SDFGD) System is based
began in 1978 under a joint venture agreement with Anhydro A/S of Denmark.
It merged the FGD process/equipment and particulate collection system
knowledge of General Electric Environmental Services, Inc. with the spray
atomization/drying knowledge of Anhydro into a single
which was bench tested in Copenhagen, pilot tested at
Springs Martin Drake Station (CCS), and characterized
testing (2, 4, 5) partially funded by the USEPA.
integrated system
the City of Col orado
in two years of
In the late 70's, the Board of Light and Power of the City of Marquette
Michigan commissioned Lutz, Daily and Brain to design the totally new 44 MW
Shiras #3 unit. The plant could have complied with Ib/MBTU emission
requirements for S02 via the compliance coal approach but was forced by the
% removal clause of the 1977 Clean Air Act to include FGD in the plant
design. For this size installation the emerging SDFGD technology provided
the best balanced solution.
II. APPLICATION
Shiras Unit #3 is a 44 MW installation. Plant equipment includes a
415,000 pounds (188,244 kg) of steam per hour Combustion Engineering (CE)
pulverized coal fired boiler fed by four CE coal mills using #2 fuel oil in
the igniters. The boiler utilizes steam soot blowing and is operated as a
balanced draft system with single FD and ID fans and a CE regenerative air
heater. The ID fan system, consisting of a Westinghouse Sturdivant radial
tip centrifugal fan, American Standard fluid coupling, and 1250 HP (932 KW)
/900 RPM (94 rad/sec) General Electric motor, was provided as part of the
air pollution control system contract.
Design conditions utilized for the Air Pollution Control System defini-
tion are summarized in Table 1.
10-161
-------
TABLE 1 DESIGN CONDITIONS
Air Heater Inlet Temperature
Air Heater Outlet:
Temperature
Air Flow
Pressure
Flyash
852°F (728°K)
265°F (402°K) continuous
550°F (506°K) maximum
205°F (370°K) @ 25% MCR
226,000 ACFM (107 ACFM/sec)
723,000 #/Hr (91 kg/sec)
1500 PPMV S02
12% Vol. H20
-4" W.6. (-1 kPa G)
9075 #/Hr (1.14 kg/sec)
The basic SDFGD process is defined in the process flow diagram (Figure
1). In most cases, the flue gas goes directly form the air heater outlet
to the spray atomizer/reactor. However there are retorfit applications
where some flyash removal occurs as a result of existing cyclones or preci-
pitators.
GRITS
SLAKING WATER
LIME
STORAGE UNLOADING
LJ VWYYY
TEMPERATURE
MONITOR
"1 '
URE
R
*
ASH
SILOS
I. D. FAN(S)
STACK
PUMP
RECYCLE
FLOODED
LOOP
TANK
Y
LANDFILL
FIGURE 1. DRY FLUE GAS DESULFURIZATION SYSTEM
10-162
-------
Aerodynamic devices in the chamber (inlet scroll, pre-swirl vanes,
inlet annulus) impart appropriate axial velocities and angular momentum to
the gas stream to insure uniform mixing with the reagent stream and cyclo-
nic action that retains 40-55% of the dried product and flyash in the
reactor vessel hopper. This total vessel cyclonic effect is unique to
single atomizer/vessel systems, although other reactors (spray nozzle,
multiple rotary atomizer/chamber) exhibit some collection of particulate in
the vessel due to its "drop-out box" characteristics. The reagent slurry
streams ("milk of lime" and recycle material) are fed to the centrifugal
atomizer wheel with rates controlled respectively by system outlet
S02 level and reactor vessel outlet temperature. The separate streams par-
tially mix on the wheel and chemically react, but not to physical or chemi-
cal equilibrium because of the short (millisecond) residence times on the
wheel. Equilibrium of these mixtures takes on the order of 5 to 20 minutes
in controlled bench scale experiments. The design intent is to minimize
the reaction between the two reagent streams to the greatest- degree
possible. The reagent stream passes from the central feed point on the
wheel (Figure 2) to the tip as a thin Coriolis-force-induced liquid film in
radial holes distributed about the circumference of the wheel. When the
film reaches the tip, it is sheared off by the gas to form a droplet
stream. The initial velocity of these droplets has a radi al /tangential
velocity ratio that depends on mass flow rate per hole and liquid viscosity
but is always less than one. The droplet size distribution is affected
weakly by film thickness and strongly by gas shear force (7, 8); hence ato-
mizer wheel tip speed has the strongest influence on droplet size.
FEED 2
FIGURE 2. DUAL FEED ATOMIZER
10-163
-------
The droplet streams form what looks like an umbrella to the stationary
observer. The interaction of the inlet gas with this umbrella of droplets
is a critical factor in controlling the uniform mixing of reagent and
S02 laden gas. The drying of these droplets of highly alkaline (pH 8-12)
liquid with suspended particles of (.5 - 4 micrometer) calcium hydroxide
and (.5 - 30 micrometer) flyash into (10 - 50 micrometer) porous particle
clusters proceeds through a complex series of surface chemistry and dif-
fusion mechanisms (6, 9) to produce a relatively dry (1 - 3% moisture) par-
ticulate stream at the reactor outlet. The size of the resultant porous
clusters depends on both the initial droplet size and the % solids in the
feed slurry. For very high inlet temperature situations without recycle
where the % solids in the slurry varies, this porous cluster size could be
a significant variable.
The spray absorber outlet stream mixes with reheat gas of a quantity
dictated by the desired particulate collector inlet temperature and pro-
ceeds to the particulate collector. Some further reaction occurs between
the remaining gas stream SO? and the particulate material. In the case of
a fabric filter, this additional reaction can capture 20-50% of the
remaining SO?; in a precipitator, the additional capture is in the 5-15%
range.
SPRAY ATOMIZER/
REACTOR
LIME PREP SYSTEM
FABRIC
FILTER
REVERSE
AIR DUCT
1DAND
RAFAN
NCLOSUR
FIGURE3.SHIRAS#3
10-164
-------
The particle clusters normally have a central core of unreacted calcium
hydorxide and a substantial amount of calcium carbonate produced by
absorption/reaction of gas stream C02. While S02 is more reactive than
C02, the high C02/S02 ratio in the gas stream counter-balances the reac-
tivity difference and results in substantial calcium carbonate formation
(4).
In addition, certain flyashes contain substantial quantities of alka-
line material. Reslurrying the collected solids will get the residual
alkalinity of the calcium salt/flyash clusters back into solutions while
also providing relatively large nucleation sites for the fresh calcium
hydorxide particles.
The physical equipment which turns the process flow diagram into
reality is shown in the site photograph of the Shiras #3 unit (figure 3).
III. SDFGD SYSTEM DESIGN ISSUES AND INITIAL OPERATING EXPERIENCE
AT SHIRAS NO. 3
There are many sub-system design issues which have to be resolved in
the design of the SDFGD system. In this section, we will briefly define
the issues and discuss how the designs have functioned during initial
operation.
A. REAGENT PREPARATION
References 1, 4, 11 and 16 deal in detail with the effect of various
reagent choices on the performance of the systems. Materials such as lime,
limestone, sodium carbonate, Nahcolite, Trona, ammonia compounds, etc. were
considered. For a variety of reasons, e. g., cost, performance, waste sta-
bility, and availability, soft-burned pebble lime has become the favored
material for SDFGD systems. In the case of Shiras #3 the lime supply is
high calcium with specified available calcium oxide content of 88%.
Lime is a generic term for a range of materials which vary in their CaO
content, may contain significant levels of MgO, have varying amounts of
"grits"*, and may have "hard burned" constituents (caused by surface
melting in the kiln). This variation in composition causes a significant
diversity of opinion in how the lime should be slaked to "milk to lime", a
water slurry of Ca(OH)2.
The slaking process is described chemically as follows:
CaO(s)+H20(l) Ca (OH)2 (1)+15,300 K - Cal
Kg - Mole
*"Grits" are composed of uncalcined limestone, kiln brick fragments,
silica, alumina, ferric oxide, etc.
10-165
-------
When this reaction is carried out with water of almost potable qualtiy
and high calcium, soft-burned, pebble lime at a water/lime ratio of 3 to 4,
its exothermic nature causes an almost explosive disintegration of the
pebbles and a 70°-80°F (40-45°K) temperature rise in the water in four
minutes resulting in a slurry of extremely fine (.5-4 micrometer) particles
suspended in water. This slurry quality level is highly desirable for the
SDFGD process.
The well known paste or detention slakers, which are simply mixing
tanks with agitation, depend on this exothermic reaction to achieve the
desired product. The larger "grits" are allowed to settle and are physi-
cally removed from the system via a combination of screw classifier,
vibrating screen, and dilution water wash lowering the specific gravity of
the lime slurry from > 1.3 to > 1.2. These commonly used slakers are
simple, reliable, and relatively inexpensive.
When poor quality slaking water, hard burned contents, or MgO contents
(almost always hard burned) are a significant factor, different slaker
types such as attrition or ball mill are usually considered. Poor quality
slaking water has dissolved solids which react with the surface of the lime
pebbles, "blinding" their pores and preventing water infiltration, thus
stopping the slaking process. Hard burning causes the same problem via a
surface glaze effect, i.e., the pores left by the escaping C02 in the
calcining process collapse due to overheat resulting in "blinding" pebbles.
These more aggressive slaking methods use mechanical means to break up the
surface and allow water contact with the internals of the pebbles to
complete the slaking. Even with these techniques the quality of the
reagent is not as good as that produced under the more ideal conditions.
Even though these slaking systems can grind up the grits, they still should
be rejected from the system. Since they are abrasive and require water for
the transport of inert material, their retention results in an ineffective
use of a key resource in this water limited process.
Shiras #3 uses a pre-assembled Clow-Coffman lime preparation plant. It
consists of a pebble lime storage silo with an integral vent filter and
dual vibrating bin discharges, redundant lime feeders, redundant lime sla-
kers with agitators and grit removal screws, dual vibrating screen grit
removers with associated screw conveyors, a lime slurry tank with redundant
agitators, redundant slurry pumps, and associated flushing, drain and
control systems. A major factor in the selection of this design was the
fact that a similar system (without the operational redundancy) designed
and built by the same vendor performed very well in the CCS pilot program
(2). The lime slurry design range is 15-25% solids controlled by agitator
current and mass/thermal balances.
In actual around-the-clock automatic operation we found that several
small practical improvements could be made to the slaking system, and that
one major issue had to be dealt with. Minor adjustments were made, such as
the installation of a lime feed cutoff device that is activated if slaker
water pressure drops below a given level. In a small power plant
10-166
-------
water supply line pressures have occasional wide fluctuations. It
is better to stop slaking for a few minutes when this occurs, rather
than depart from the optimal lime slaking temperature ranges (180°F
to 190°F).
A more serious problem is that we found that one lime truckload out of
a dozen had grit contents over 20%, most of which was unburnt limestone.
This problem was later traced by the lime supplier to learning curve
quality control problems caused by the change from a gas-fired to a coal -
fired kiln. A large percentage of the limestone grit would leave the
slaking system in the form of a fine suspension, and cause pluggage
problems in the lime distribution system. It was found that a simple
hydroclone was adequate to filter out the calcium carbonate, provided that
one took into account the highly viscous, non-Newtonian nature of the lime
slurry and over-sized the hydroclone accordingly.
B. SPRAY ABSORBER
The chamber, where the reagent and recycle slurries are atomized into
fine droplets, where the reagent droplets mix with the incoming flue gas,
where the absorption/chemical reaction of S0£ takes place, and where a
significant amount of the dried particulate is collected, is the most
critical piece of process equipment. A cutaway schematic of a typical
spray absorber is shown in Figure 4. Operational experience with hundreds
of industrial spray drying applications provide the user with a range of
cost-effective reliable alternatives. The major concerns that utility
S02 service causes for this proven equipment is the abrasive and corrosive
constituents in the inlet flue gas; abrasive chemically reactive slurries
are not new to the spray dryer industry (13).
Reference 14 gives a credible review of the choice between two fluid
spray nozzle systems and spinning disk rotary atomizer systems. The main
factors in that choice are cost, power consumption, droplet size and
variability control, reliability, and gas/droplet mixing. Those vendors
who supply high horsepower rotary atomizers use one centrally located in
each reactor of a multiple reactor system, treating up to 150 megawatts
equivalent gas flow per reactor. This symmetry allows use of the vessel as
a moderately efficient cyclonic collector. This can be particularly useful
in protecting the downstream particulate collector from wet material in
case of upset conditions. Our emphasis in this paper is on the single
rotary atomizer system supplied by Anhydro A/S to GE for the Shiras #3
installation. The spray absorber/reactor consists of a low pressure drop
gas disperser, belt-driven atomizer, and absorber chamber. The spray
absorber must be carefully optimized so that the reagent slurry is ato-
mized, contacted with the flue gas, and dried in a manner that promotes
maximum capture of S02, minimum reagnt consumption, and low energy use,
while maintaining stable and reliable plant operation.
10-167
-------
Atomizer
Removal
Monorail
Penthouse
Gas
Distributor
Axial
Entry Vanes
Atomizer
High Gas
Exit Duct
FIGURE 4. SPRAY ABSORBER
For the dry SDFGD application, we use a standard 400 mm tip diameter
centrifugal atomizer designed to generate a uniform spray of fine (10-80
micrometer) droplets over a wide range of feed rates. For larger gas
flows, the number of radial holes in the atomizer wheel is increased. The
droplet size model used in our system design has been confirmed in detailed
experiments at GE's Corporate Research and Development Center. Typical
results over a range of tip speeds are shown in figure 5.
For the Shiras #3 atomizer, this means that the mass mean particle size
i s about 30 mi crons.
The two slurry feeds are piped independently (see figure 2) to a co-
annular liquid distributor. In this way the two chemically reactive
calcium hydroxide and recycle slurries are kept separate until fed on to
the atomi zer wheel.
10-168
-------
ATOMIZER OUTPUT FOR
CIRCULAR SLOT AT 2.5 GPM
DECREASING co,
Acu = 2500 RPM
eu = 30000 RPM
20 40 60 80 100 120 140 160
DROPLET DIAMETER (MICRONS)
FIGURE 5. FOUR INCH WHEEL DATA
The slurries are introduced to a central cavity on the rapidly rotating
atomizer wheel, and are induced by inertial forces to flow outward through
radial passages in the wheel and then break off to form spherical droplets
whose size range is chiefly governed by the viscosity and surface tension
of the liquid and the atomizer wheel tip speed. As the droplets move away
from the wheel and disperse into the gas stream exiting the gas dsiperser
vanes, they form an umbrella-shaped spray pattern that is symmetric about
the chamber axis and serves as the zone of initial contact between the
reagent and flue gas. Figure 6 shows a stroboscopic photograph of one
ligament of that umbrella.
The size of the droplets must be controlled to assure optimum reaction;
small droplets are desirable because they provide a large surface area for
mass transfer, but they must not be so small that they do not penetrate to
the outer diameter of the inlet gas annulus.
A belt drive system, which makes it possible to change atomizer wheel
speed by simply changing the belts and sheaves, is used on Shiras #3.
Because of the abrasive character of the slurries atomized in the spray
absorption process, a wheel design is used which features silicon carbide
inserts in the slurry passages. The inserts may be rotated as local wear
spots appear to extend their useful life.
10-169
-------
FIGURE 6. DROPLET STREAM STOP-ACTION
The Shiras vessel is 36 feet (11 meters) diameter, 71 feet (21.6
meters) tall and uses a 200 HP (150 KW) motor to drive the atomizer wheel
at 7800 RPM (817 rad/sec) thru a custom designed belt/sheave system.
The absorber gas dispenser is designed for high S02 removal without high
pressure drop or flyash abrasion problems. Anhydro selected a top-entry
vaned scroll-type gas dispenser, which dischanges an annulan vontex flow of
flue gas down into the chamben on all sides of the atomizer wheel. The gas
dispensen is equipped with vanes whose angle can be adjusted to obtain
mixing of the spnay and the flue gas. The initial gas notation fnom the
inlet scnoll and vanes is in the same dinection as wheel notation which
pnovides an incneased cyclonic effect but slowen deceleration of the
dnoplet stneam and langen dnoplets.
The absonben chamben must be sized in nelation to the gas flow volume
to assune that the slunny dnoplets will have adequate nesidence time (in
most applictnions 8 to 12 seconds based on absonben outlet volume) in the
chamben fon the vanious stages of neaction with SO? and dnying to occur.
The chamber design also affects the degnee of dnopout in the absonben,
which should be maximized to neduce the panticulate loading to the fabnic
filten. The absonben is designed as an axial-entny cyclone to achieve
appnoximately 50% dust dnopout in the chamben.
The GE CR&D Centen has done extensive coupled modeling of the gas flow,
particle trajectory, drying, absorption and chemical reaction to provide a
tool for connelating openating data and guiding futune design optimization.
10-170
-------
The model results indicate that the key phenomena occurring in the gas
inlet/droplet injection region of the vessel is mixing of the gas and
reagent; the bulk of the absorption process occurs after this initial
gas/slurry mixing phase.
If the droplet/gas momentum ratio is too low, the gas at the outer
diameter of the inlet annulus will see no reagent, i.e., bypass occurs. If
the ratio is too high, a similar bypass core occurs at the inner diameter
and the larger droplets get picked up by a recirculation vortex and get
deposited on the wall of the atomizer. Since these ratios are constantly
changing as boiler load changes, the Shiras #3 design divides the inlet
into three dampered co-annular rings so that high gas momentum can be main-
tained as gas flow volume changes. The variable speed drive concept men-
tioned earlier can also be used to advantage in this turndown management
system, i.e., smaller droplets at low gas velocities, bigger ones at high
velocities.
A practical illustration of the importance of the droplet/gas momentum
ratio was provided early in the Shiras #3 startup. Persistent boiler soot
blowing system problems resulted in weeks of steady-state gas temperatures
into the spray absorber of between 350°F and 410°F; far above the design
value of 265°F. The GE/Anhydro atomizer had the reserve power and pumping
capacity to increase slurry flow and maintain constant approach tem-
perature. However, this resulted in a significantly higher droplet/gas
momentum ratio, and a much broader reaction umbrella, which caused minor
solids deposits on the vessel walls and roof to be built up over a one
month period. As soon as this was seen, the setting on the gas distribu-
tion vanes (the "wing vanes") was altered to permit a more vertical gas
flow into the vessel, thus reducing droplet/gas momentum ratio. After
another month of operation, the vessel was inspected, and the roof and
walls were free of deposits.
C. PARTICIPATE COLLECTOR
In a SDFGD system the spray atomizer/reactor is primarily an S02
capture device but doubles as a particulate collector. Not to be outdone,
the fabric filter (or precipitator in some cases) particulate collector
also doubles as an S02 absorber. Effective operation in this secondary
mode requires operation of the fabric filter, within 20-50°F (11-28°K) of
the gas dewpoint. Although many utility baghouses operate that close to
the sulfuric acid dewpoint, the quantities of liquid potentially produced
are so different that special design attention is rquired in the SDFGD
case, particularly in the areas of reheat control, internal/external insu-
lation, and leakage control.
References 10 and 15 thoroughly discusses the pros and cons of precipi-
tators and fabric filters and the special design considerations for each.
10-171
-------
In summary, the superior FGD absorption contribution of the fabric
filter makes it the logical choice in high efficiency applications or in
any situation where lime costs are high and operating cost savings are
seriously considered.
The Shiras #2 unit is a conventional eight (8) compartment custom GE
reverse air design. It uses 168, 35 foot (10.7 m) long, 12 inch (30.5 cm)
diameter, acid-resistant-coated, fiberglass bags per compartment. All
fabric filter dampers are of the low leakage poppet type except the bypass
which is a pressurized double louver selected by the LD&B engineers because
of their excellent experience at the City of Colorado Springs Martin Drake
Plant. The fabric filter system is a conventional design with redundant
reverse air fans, programmable controller control of the cleaning cycle in
a variety of operator selected modes (manual, time, P batch, P distri-
buted, etc.), and a centrally manifolded inlet, outlet, R/A duct system.
With the atomizer in operation, the fabric filter design conditions are
based on gas at an inlet temperature of 175°F (325°K) and a gross air to
cloth (A/C) ratio of 1.49. With the atomizer off line the inlet tem-
perature is 265°F (402°K) and the gross A/C is 1.62. Note that the design
dust loading changes substantially between these two situations going from
6.0 to 4.7 grains/ACF. In actual operation this inlet gas tempreature to
the system has been as high as 360°F (455°K).
Baghouse operation at Marquette has been exceptionally trouble free.
Operation has been at approach temperatures between 25 and 60°F. There
have not been any occasions of condensation on the baghouse walls or the
bag filter cakes (all baghouse compartments are kept on line regardless of
load level). There has not been any build up in average baghouse pressure
drop in over four months of operation.
Most critical and interesting of all, the frequency of the baghouse
batch cleaning cycle actually drops when the spray absorber is brought on-
line. Typically, if the baghouse has been cleaning 50% of the time while
the atomizer is being inspected, it will drop to cleaning only 33% of the
time once the atomizer is restarted.
Temperature drop across the baghouse is minimal (under 5°F) when the
spray absorber is operating.
D. REHEAT
The evidence is conclusive that operation with reactor discharge tem-
peratures as close to adiabatic saturation as possible optimizes the lime
use - collection efficiency relation. However, temperature variation
within the reactor due to non-uniform mixing of reagent and flue gas, wall
deposits, reactor hopper pluggage and ductwork condensation, etc., constrain
this optimization. Recognizing that reactor and particulate collector
10-172
-------
constraints on approach temperature are different, most designers have made
provisions for reheating the gas downstream of the reactor. Several
methods have been considered for accomplishing this reheat, for example:
a. Bypass of air heater inlet gas.
b. Bypass of air heater outlet gas.
c. Reci rculation of fabric filter outlet gas through a water/steam - gas
heat exchanger.
d. Injection of ambient air through a water/steam-air heater exchanger.
The first two cause similar problems. Because they contain particulate
and S02, clean-up of the reheat system is required. If the reheat injec-
tion is upstream of the fabric filter, this clean-up is automatic.
However, since the fabric filter is a low efficiency S02 device (the preci-
pitator is even lower) which degrades in efficiency as its inlet tem-
perature is increased, these types of bypass have a "double kicker" effect
on required reactor efficiency.
Of the two, airheater inlet gas is theoretically more effective since
smaller quantities are required for the same amount of reheat. However,
the system heat rate effect is worse. The pilot work at Colorado Springs
(12) raised some serious doubts about the validity of this theoretical
conclusion which warrants further evaluation. The issue may be related to
heterogenous mixing but even that is not clear.
Methods c and d use a scrubbed clean gas for reheat. They not only
eliminate the "double kicker" (only the temperature degradation of baghouse
S02 efficiency remains) but also provide the option for reheat injection at
the baghouse outlet (eliminating the temperature degradation effect). This
option assumes that reheat was not for baghouse protection purposes but for
ductwork, stack, and plume buoyancy purposes. The negative factors in
these approaches involve recirculation fan horsepower and water/steam BTU
requi rements.
In spite of the concerns raised about the theoretical conclusions by
pilot results, Shiras #3 uses the air heater inlet bypass approach which
requires a reactor efficiency of 80% to provide a system efficiency of 83%
with 20°F (11°K) reheat. Air heater outlet reheat would require the reac-
tor efficiency to be higher than the system efficiency as indicated in
figure 7. For example a system efficiency of 80% would require a reactor
efficiency of 100% if 22°F (12°K) reheat was required. This dramatically
shows the impact of the selected reheat approach, particularly with the
system design point on the high stoichi ometry/ efficiency portion of the
performance curve where lime savings more than counter balances heat rate
increases.
Operation of the reheat system at Shiras #3 has been routine.
10-173
-------
>- 100
o
z
LU
o
LL
80 -
cc
LLJ
CQ
X 60
O
sr
UJ
>
IT
Q
Q
LLJ
DC
40
O
LLJ
CC
20
EFFECT OF AIRHEATER BYPASS REHEAT
ON SYSTEM EFFICIENCY
25°F REHEAT
(AIRHEATER OUTLET)
0°FREHEAT
AIRHEATER
INLET
20 40 60 80
SYSTEM SO, REMOVAL EFFICIENCY
100
FIGURE 7. REHEAT EFFECTS
Baghouse approach temperatures have been varied between 25°F and 60°F,
and reheat has been used for the upper end of the temperature range.
E. CONTROLS
The important issues in DFGD system controls are:
a. Reagent feed control
b. Temperature control
c. Turn down control
d. System response to transients
The other control aspects such as lime slaker control, recycle slurry
generator control, and baghouse control are not unique to SDFGD and are not
covered in this paper.
Each design deals with the specifics of these control issues in its own
way. Therefore, our focus will be generic with the specifics relating to
Shiras #3.
It is clear that the reagent feed control strategy should be to maxi-
mize solids concentration to the limits permitted by solids drop-out and
abrasion consideration and to maximize the ratio of recycle/fresh "milk of
10-174
-------
lime". In our system design we use the GE patented (3) Two Loop Control
system which automatically optimizes system efficiency/stoichiometry. The
lime slurry is prepared and stored at a solids concentration substantially
higher than would normally be required to achieve the required absorber
S02 removal level. For an optimum outlet temperature to be reached, the
lime slurry is diluted with recycle slurry on the atomizer wheel just prior
to atomizati on.
The first of the two automatic loops regulates the flow of lime slurry
to the absorber, based on the fabric filter outlet SOg level, by modulating
the position of the lime slurry control valve. SOg levels are also moni-
tored at the absorber inlet and outlet using a Dupont extractive system.
The second loop controls recycle slurry flow based on approach to
adiabatic saturation temperature at the reactor outlet.
Changes in the adiabatic saturation temperature due to steam soot
blowing, coal moisture changes, ambient moisture changes, etc., are pro-
vided to the system via feed forward controls and/or operator input.
Automatic determination of adiabatic saturation temperature is conceptually
attractive but has been inhibited by the unavailability of proven reliable
sensors. Although not part of the system design contract, we will evaluate
three candidate sensors for this service at Shiras #3. The sensors are all
extractive and subject to particle contamination if used at spray dryer
inlet or outlet. It is conceptually possible to use these devices at the
baghouse outlet but the increased time delay substantially reduces its
value. Also, use at the outlet would be questionable because of the
variable moisture absorption characteristics of the bag cake.
In operation at Shiras #3, it has been found the Two Loop Control
system regulates the temperature leaving the spray absorber vessel to with
+ 3°F during normal operation, and to within +_ 7°F during rapid boiler
excursi ons.
For multiple (3 or 4) vessel systems, this range is enough to permit
system turndowns of 40-50% and excess flows to 20% via vessel dampering and
individual vessel shutdown. In the single vessel Shiras #3 design, the
turndown is accomplished via variable inlet geometry in the vessel. The
geometry variation involves dampering of flow in three co-annular inlet
rings and variable swirl vane angle. Variable atomizer speed will also be
tested, although not part of the basic Shiras design. The system allows
variation of the gas/slurry momentum relations in the inlet area which dra-
matically affects system performance.
Turndown control is complicated for the single vessel Shiras design.
Pilot data indicates that operation of a reactor vessel in the + 20% gas
flow range is practical and has no more than a 5% - 1% effect on
stoi chi ometry for a given efficiency.
System control dynamics issues are related primarily to the time delays
in sensing the effect of slurry feed on S02 levels at the fabric filter
10-175
-------
outlet and temperatures at the reactor outlet and fabric filter inlet and
the slow read rate of the S02 monitor. If not considered properly, system
interaction with these loops could lead to excessive hunting, overshoot,
and offset in the control system. To better understand the dynamics of the
system, GE performed a computer simultion of the Marquette design. (17)
Operating experience has proven out that with 20 second effective
response times on the recycle slurry loop and 15 minute response times on
the lime slurry loop, that hunting does not occur.
F. MATERIAL HANDLING
The major issues in this area are the handling of potentially wet pro-
duct from the absorber hopper and the sourcing of recycle material. Based
on pilot experience (1,5), GE favors mechanical conveying (screw, drag
chain, etc.) of reactor discharge material in combination with multiple
assists for hopper clearage. These include vibrating bit bottom, poke
holes, and an auxiliary side removal system from the hopper. In order to
ensure sufficient recycle solids, reactor hopper and/or fabric filter
hopper discharge should be available.
Because the Shi ras #3 ash collections system was combined with upgrades
on units #1 and #2, the City of Marquette purchased a pneumatic system
which combined pressure and vacuum portions.
Operating experience has been that the pneumatic system needs a
grinding system to break up occasional loose clumps, and such a device is
being added to the ash removal system.
Another practical consideration is that although the solids hopper
catch handles like a dry; western flyash, the boiler gas is brought to
within 20 to 40°F of its dewpoint. Thus care must be taken to thoroughly
heat trace and insulate pneumatic vacuum systems, since they will entrain
some boiler gas which must not be allowd to cool and condense. Another
solution is to use heated (250 to 300°F) ambient air in the pneumatic
system. The former was done at Shiras #3 for the baghouse hoppers, and the
latter for the spray absorber hopper.
G. MATERIALS OF CONSTRUCTION
Because of the non-corrosive nature of the gas stream down-stream of
the reactor inlet, assuming condensation is avoided, carbon steel is used
throughout the reactor, ductwork, and baghouse. Since the centrifugal ato-
mizer outer shell is colder than the inlet gas and subject to untreated
inlet gas, 300 series stainless is utilized in this protective cover.
Slurry piping is unlined heavy wall carbon steel; valves are rubber
lined. Acid resistant coating was used on the fiber glass bags because its
10-176
-------
superior "fiber wetting" characteristic will protect the bag material from
wet alkali attack; there is no acid attack concern.
So far the only sign of erosion in the entire system has been the
expected wear at the atomizer wheel ceramic ports.
H. INSULATION/LEAKAGE/ T
Because of the low approach temperature/minimum reheat sytem design,
there is a real potential for localized gas condensation on duct surfaces,
reactor or fabric filter walls, etc. To minimize the risk, extra thick
insulation (8 inch (20.3 cm) on reactor, 6 inch (15.2) plus air gap on
fabric filter) was employed, special leak tests were run on the reactor and
fabric filter, and purge dampers were not included on the fabric filter.
One feature that might have been desirable was internal wall insulation in
the fabric filter to avoid compartment off-line cold wall effects. There
are concerns about condensation under the insulation in this approach, but
it could have design merit if impervious insulation is used. Marquette
doesn't have internal wall insulation. The fact that all compartments will
be on line at all times (except for on-line maintenance) reduces the cold
wall condensation concern. Units that have had this problem were running
continuously with off-line compartments in an attempt to minimize fabric
filter gas temperature drop.
Operating exprience has been that (with the exception of the vicinity
of one leaky door gasket) wall condensation is not occuring, and that tem-
perature loss across the system is under 5°F.
I. PERFORMANCE
Design performance requirements for the Shiras #3 system are:
1. 80% S02 removal at all boiler loads from 20 to 100%.
2. Outlet particulate of .005 grains per ACF or 99.5% weight efficiency
which ever is less stringent.
3. Lime use not to exceed 2874 #/Hr (.36 kg/sec) at design conditions.
4. System pressure drop of 12.4" W.G. (3.1 kPa G)
5. Gas temperature at stack inlet of 165°F.
6. Bypass reheat of 3,600,000 BTU/Hr (1055 KW) based on 852°F (729°K) gas.
In September 1983 compliance test of all the above objectives were
met, even though the system was not yet optimized, e.g., air preheater exit
gas temperatures were from 70 to 80°F above the design condition.
10-177
-------
A comparison of the compliance test data with 1980 data from the 2 MW
demonstration facility at the City of Colorado Springs, Martin Drake No. 6
plant, shows that for the same stoichiometries, S02 removal efficiency was
10 to 20% higher in the 44 MW system than in the 2 MW system. Thus, scale-
up has been properly applied in the Marquette system.
The overall conclusion from the first four months of spray absorber
performance is that it is possible to operate at S02 removal efficiencies
between 70 and 90% simply by varying lime stiochi ometry. This performance,
as well as the lack of scale-up problems, is attributed to the Marquette
design which was supported by the R & D effort at General Electric's
Schenectady laboratories.
J. ON GOING PROGRAMS AT MARQUETTE
For DFGD vessels handling from 100MW to 150MW of boiler gas each, 500
HP to 800 HP atomizer motors are required. Radial bearing loads for either
belted or spur-pinion gear systems become excessive for these high hor-
sepower drive trains, and bearings B-10 lives become unacceptibly short
(under 10,000 hours). As a result, General Electric has developed a Direct
Drive System which uses a voltage inverter to operate a vertical AC motor
at power frequencies other than 60 Hertz, so that the motor rotates at the
atomizer speed. A flexible metal coupling is used to transmit the torque
from the rotor to the atomizer.
A Direct Drive system (shown in Figure 8) has been successfully
operated for several months at the General Electric Turbine Technology
Laboratory in Schenectady, New York, and is currently being installed at
Shiras #3 for field testing.
INDUCTION MOTOR
200 TO 800 H.P. AT
8000 RPM
COOLING AIR
FEEDPIPES
5 HP. FAN DRIVE
MOTOR LIFTING LUG
INSULATION
MOTOR SUPPORT
BRACKET
CONTOURED FLEXIBLE
DIAPHRAGM COUPLING
ATOMIZER BODY
ATOMIZER WHEEL
FIGURE 8. DIRECT DRIVE ATOMIZER ASSEMBLY
10-178
-------
REFERENCES
1. Rhudy, R. G. and Blythe, 6. M.; "EPRI Spray Dryer/Baghouse Pilot Plant
Status and Results"; EPRI Symposium on Particulate; April, 1983.
2. Samuel, E. A. et al; "Dry FGD Pilot Plan Results: Lime Spray Absorption
for High Sulfur Coal and Dry Injection of Sodium Compounds for Low Sul-
fur Coal"; EPA/EPRI S02 Sumposium; May 1982.
3. Roth, A.; U.S. Patent #4, 322, 224.
4. Parsons, E. L., Boscak, V., Brna, T. C., and Ostop, R. L.; "S02 Removal by
Dry Injection and Spray Absorption Techniques"; USEPA Third Symposium on
Particulate Technology; March 1981.
5. Parsons, E. L., Hemenway, L. F., Kragh, 0. T., Brna, T. G., and Ostop, R.
L.; "S02 Removal by Dry FGD"; USEPA Sixth FGD Symposium, October 1980.
6. Getler, J. L., Shelton, H. L., and Furlong, D. A.; "Modeling the Spray
Absorption Process for S0? Removal"; Journal of APCA, Vol. 29, No. 12;
December 1979.
7. Snaddon, R. W. L.; "Rotary Atomization Studies for Dry Flue Gas
Desulfurization"; GE CR&D internal report; March 1983.
8. Master, K.; "Spray Drying Handbook"; John Wiley; 1979.
9. Downs, W. et al; "Control of S02 Emissions by Dry Scrubbing"; American
Power Conference; April, 1980.
10. Fortune, 0. F. and Miller, R. L.; "Design Considerations for Baghouse -
Dry S02 Scrubber Systems"; EPA Particulate Symposium, November, 1982.
11. Kelly, M. B. and Shareef, S. A.; "Second Survey of Dry S02 Control
Systems"; US EPA-600/7-81-018; November 1980.
12. Samuel, E. A. and Brna T. C. "Final Report on CCS"; June 1983.
13. Tuttle, J. N. et al; "Neutralization During Atomization in Spray Dryer
Yeilds 100% Dry Powder"; Chemical Processing; November, 1979.
14. Maurin, P. G. et al; "Two Fluid Nozzles vs. Rotary Atomization for Dry -
Scrubbing Systems": Chemical Engineering Progress; April 1983.
15. Campbell, K. S. et al; "Economics of Fabric Filters Versus Precipitators";
EPRI FP-775 June, 1978.
16. Yeh, J. T., Demski, R. J. et al; Experimental Evaluation of Spray Dryer
Flue Gas Sulfurization for use with Eastern U.S. Coals"; EPA-EPRI FGD
Symposium; May 1982.
17. Forsberg, C. H. and P. S. MacDonald; "Flue Gas Desulfurization System
Model"; GEESI-ADAPCO Report 18-03-001; May 1983.
10-179
-------
START-UP AND INITIAL OPERATING EXPERIENCE OF THE
ANTELOPE VALLEY UNIT 1 DRY SCRUBBER
R. L. Eriksen, F. R. Stern, R. P. Gleiser, S. J. Shilinski
-------
START-UP AND INITIAL OPERATING EXPERIENCE
Robert L. Er ikson
Environmental Control Supervisor
Basin Electric Power Cooperative
FrederIck R. Stern
PI ant Eng i neer
Basin Electric Power Cooperative
Richard P. Gleiser
Field Service Engineer
Joy Manufacturing Company
Stan ley J . Sh I I i nsk i
Field Service Engineer
N i ro Atom i zer Inc.
AESIBACI
The first competitively bid and awarded utility dry scrub-
bing system utilizing lime as the scrubbing reagent was for
Basin Electric's Antelope Valley Station Unit No. 1.
Awarded In 1978 to Joy Manufacturing with Niro Atomizer as
the major subcontractor, the system was scheduled to start-
up in 1981, however, due to reduced load growth, was
delayed until this year.
The dry scrubbing system treats flue gases from a 435 MW
lignite fired boiler and consists of reagent preparation
equipment, five spray dryer absorbers, and two fabric
f I Iters.
Initial operation on coal began in May, 1983 and commercial
operation of the system Is scheduled for July, 1984. This
paper will review the start-up procedure, any problems
which have developed thus far and how they have been
handled, as well as the results of the operation of the
system to date.
10-181
-------
In September, 1978, Basin Electric awarded a contract for
a pollution control system to the Western Precipitation
Divlson of Joy Manufacturing with Niro Atomizer as a major
subcontractor. The contract was for the supply and erec-
tion of an SO dry scrubbing system and baghouse for Unit
#1 at Basin Electric Power Cooperative's Antelope Valley
Station near Beulah, North Dakota. Major equipment in-
cluded In this contract were the reagent unloading and
storage system, reagent preparation equipment, five spray
dryers, mechanical conveyors for a recycle system, two
baghouses for particulate collection, and a computer
control system for the operation of the dry scrubber.
Engineering commenced in November, 1978, with field mobi-
lization and the first equipment arriving on site In
July, 1979. Field erection activities lasted from then
until May, 1982. The plant was originally scheduled for
commercial operation in 1981; however, due to lower than
anticipated regional power demands, the actual plant
start-up did not occur until May, 1983.
Joy and Niro start-up personnel arrived on site in
November, 1982, and began pre-start check-out of the spray
dryers and baghouses, Including checking motor rotation,
programming of the computer control system, etc. The
actual system start-up began In May 1983, when the
Combustion Engineering 435MW lignite boiler was first
fired with coal, although the first flue gas to the system
occurred on January 13, 1983, when the boiler was fired on
oil. Start-up and testing lasted until the first week of
October, 1983, when the performance testing to verify
operating parameters as guaranteed (lime consumption,
pressure drop, power consumption, etc.) was successfully
compIeted.
The Antelope Valley Station dry scrubber consists of five
46-foot diameter spray dryer absorbers. An inlet manifold
equally distributes the boiler flue gas to each of the
five spray dryers (Figure 1). Gas is distributed between
the roof and central gas dispersers of each spray dryer.
Each dryer is equipped with a single, F-800 model rotary
atomizer, which is powered by a flange mounted, vertical
shaft 700 hp motor.
10-182
-------
ROOF GAS
DISPERSER
EXHAUST DUCT
CENTRAL
GAS
DISPERSER
INLET
DUCT
FIGURE 1 - ABSORPTION CHAMBER
10-183
-------
The full design gas flow of 2,055,000 ACFM (at 307°F) can
be handled by any four of the five spray dryers with the
number of modules selected for service based on boiler
load. The inlet and outlet ducts of each absorber have
guillotine isolation dampers to isolate individual modules
when they are not in service. The scrubber computer
control system signals the operator when spray dryers need
to be brought on or off line (due to changes In gas flow),
and the operator then initiates the appropriate sequence.
The slurry feed to the atomizers Is prepared In a con-
tinuous feed preparation system with lime, water, and
recycle materials controlled to match the system demands.
Two complete lime slaking and feed preparation trains
provide redundant equipment, maximizing system
availability. Each slaking train Is sized for 100$ of the
required lime consumption at the average performance
conditions. The mixing trains are each sized for 100$
capacity under maximum operating conditions.
Figure 2 shows the overall process scheme. Pebble quick
lime is pneumatically transferred from the railcar/truck
unloading facility to the 3,000 ton capacity silo and from
there, conveyed pneumatically to a 100 ton capacity lime
bin. Weigh belt feeders meter the lime from the lime bin
into the two Denver Equipment ball mills, each sized to
slake 6 tons/hr of pebble lime. The ratio of the treated
water and pebble lime fed Into the mills is controlled to
maintain a 35$ solids slurry within the mill. Slaked lime
slurry leaving the ball mills Is diluted to 20$ solids,
and classified In spiral classifiers to 100 mesh. The
grits separated by the classifier are recycled back into
the ball mill. The classified lime slurry Is pumped by
rubber lined pumps to the mixing tank, where baghouse and
spray dryer recycle material along with dilution water are
added. The feed slurry in the mix tank is continuously
pumped to the feed tank, which overflows back to the mix
tank .
From the feed tank, the feed slurry Is pumped through
individual supply lines to head tanks located above each
spray dryer module. There Is a continuous overflow from
the head tanks back to the feed tank. The purpose of the
head tank is to supply slurry to each atomizer feed con-
trol valve at a constant pressure. The control system
regulates the amount of feed delivered through the control
valve to the rotary atomizer to maintain a constant outlet
temperature at each spray dryer exit.
Lear Siegler SM810 S02 monitors, located at the scrubber
10-184
-------
OO
Ul
FIGURE 2 - FLOW DIAGRAM
-------
Inlet duct and in the stack, feed signals Into the scrub-
ber controls (Honeywell TDC-2000 computer system). Boiler
flue gas flow, temperature, and S0_ content are con-
tinuously monitored and the feed preparation system inputs
are varied continuously to maintain a 35$ total solids
content while obtaining the desired S02 removal at the
required stack temperature.
At the outlet of each spray dryer is an inclined mechani-
cal drag link conveyor to transport the dried powder and
ash collected in each absorber. Gas exits from each spray
dryer via a side discharge duct along with entrained
flyash and spray dried material. A portion of the par-
ticulate in the gas stream drops to the bottom of the
spray dryer and on to the drag conveyor for reuse in the
feed system. The major benefit of this two point collec-
tion concept Is that if wet material should exist in the
absorber due to an upset condition, It will fall to the
bottom of the chamber and be col lected by the drag con-
veyor, while the flue gas will exit from the side
discharge duct in the conical section of the chamber.
Therefore, a clear flue gas path will always exist.
From the five spray dryer conveyors, the collected recycle
material is deposited onto two horizontal collecting
conveyors, and from the collecting conveyors, the material
dumps Into a single vertical lifting conveyor which dis-
charges into the recycle bin. A portion of the ash
collected In the baghouses is pneumatically conveyed to
the recycle bin to insure that an adequate supply of
recycle powder is available to support process
requirements. The required amount of recycle powder is
metered by a variable speed drag link conveyor into the
mix tank along with lime slurry and dilution water to make
up the final feed slurry. The use of recycled powder in
the process is to reduce the amount of fresh lime needed
for SO scrubbing, and to enhance the drying aspects of
the overall spray absorption process.
sie
The two particulate collectors installed at Antelope
Valley for collection of flyash and spray dried product
are Western Precipitation reverse air baghouses. Each
baghouse consists of two parallel rows of seven compart-
ments for a total of twenty-eight compartments. The
compartments are connected to the common inlet, outlet,
and reverse air manifolds which are located on the cen-
terline of each baghouse (Figure 3).
10-186
-------
FIGURE 3 - BAGHOUSE DRAWING
10-187
-------
Each compartment contains 288 filter bags. The bags have
a 12-!nch diameter and are 35 feet long. The bag type is
Menardi-Southern 601T Teflon coated fiberglass. At the
maximum design gas volume of 1,894,000 ACFM, the effective
filter ratio is 2.19. The design operating temperature is
192°F. The bags are suspended under spring tension from a
hanger bar at the top of the compartment and retained at
the bottom by a flared thimble. The thimble floor is
"stepped" down towards the baghouse outer wal Is. Beneath
each compartment are two pyramidal hoppers, each equipped
with an access door, poke holes, a hopper heating system,
and a nuclear dust level detector. Each baghouse compart-
ment has one blade type Inlet damper, two outlet poppet
type valves, and one reverse air poppet valve. All valves
and dampers are pneumatically actuated by solenoids
operated from the control panel.
The operation of the two baghouses is controlled Independ-
ently by one Texas Instruments STI Programmable Controller
utilizing a Control Junctions, Inc. multiplexer system.
Reverse air cleaning can be Initiated by baghouse dif-
ferential pressure or be set to clean continuously with a
variable time delay between cleaning cycles. One compart-
ment per baghouse is cleaned at a time. One to three
reverse air cycles can be programmed for each cleaning
period. Reverse air is provided by drawing cleaned flue
gas from the baghouse outlet via three (two operating, one
stand-by) 200 hp reverse air fans.
For start-up and upset conditions, the baghouses can be
bypassed through double louver dampers which connect the
inlet and outlet flues. These dampers will open for
emergency bypass on high or low inlet gas temperature or
high differential pressure. During baghouse operation,
reverse air is utilized to purge the area between the two
sets of louvers.
Process and mechanical checkout of the various scrubber
sub-systems and the baghouses began in November, 1982, and
continued through May, 1983. On May 24, 1983, the first
coaI firing occur red.
Throughout initial firing of the boiler on oil and the
first few days of Initial coal fire, the baghouses were
operated on bypass. During this bypass period, the com-
partment access doors were cracked to vent fresh air into
the compartments and the hopper heaters were turned on to
help preheat the baghouse.
10-188
-------
On May 27, 1983, it was felt that coal firing was stable.
The baghouse access doors were sealed and the bypass
dampers closed, bringing the baghouse on line for the
Initial coating of flyash. The boiler was operating at
approximately 60% load and the flue gas entering the
baghouse was approximately 320°F. Unfortunately, within a
few hours after the initial bag coating process had begun,
the boiler control system suffered a momentary power loss,
causing a unit trip. The baghouses were put back on
bypass for relgnlting of the boiler.
Approximately 24 hours later, the baghouses were brought
back on line. The reIntroductIon of flue gas to the
baghouses was at essential ly the same load and temperature
as the previous day. The initial pressure differential
across both baghouse inlet to outlet flues was two Inches
water. After about six hours, the pressure differential
had risen to four inches water, and the first reverse air
cleaning cycle was Initiated.
On June 8, 1983, after two weeks of bag coating with
flyash only, S0? removal with single pass operation (lime
only) In the scrubber commenced. The first scrubber
operation with recycled material being added to the slurry
occurred on June 11, 1983. On June 12, the boiler was
taken off-line for a one month shutdown for turbine fine
screen removal and boiler chemical cleaning and was back
on-line on July 10. The scrubber operation was gradual ly
brought to design operating conditions (eg. 35% total
solids and 170°F spray dryer outlet temperature) during
the next two months by Increasing the solids content of
the slurry with recycle material and decreasing the outlet
temperature with higher slurry feed flow rates.
PROBLEMS_ENCOUNIERED_AND_IH£iR_CURR£NI_5IAIilS_
The Initial operation of the scrubber and baghouses fol-
lowed the anticipated schedule and at no time delayed the
boiler start-up or operation. The overall success of the
start-up effort was probably the result of Joy/Niro's
experience at the Riverside full scale demonstration
plant. There were, however, initially some mechanical
problems with various components of the absorber product
conveying system, atomizer wear plates, and the atomizer
head tanks .
10-189
-------
The recycle lifting conveyor has horizontal sections at
the upper and lower extremes. During the first weeks of
recycle operation, a portion of the paddle type flights of
the conveyor experienced fatigue failure due to the wear-
ing of the paddle tips on the center plate in the lower
90° bend of the conveyor casing. To remedy this, a guide
rail was added supporting the chain in the conveyor bend,
preventing the paddle flights from riding on the center
p I ate .
Dusting from the conveyors made it necessary to redesign
shaft and dump gate seals, and to increase the venting of
the recycle bin. Operation of the variable speed drag
link conveyor which feeds ash from the recycle bin Into
the mix tank indicated that the 3 hp motor originally
supplied had insufficient power at the lower range of the
operating speed. A 5 hp drive will be Installed.
The interior of the atomizer wheel Is equipped with a
ceramic bottom wear plate. During the early operation,
several of these plates cracked, possibly due to thermal
shock as the wheel came Into contact with the flue gas and
then the feed slurry. A procedural change, initiating
atomizer wheel protection water flow simultaneously with
initiation of chamber gas flow, has solved this problem.
The atomizer head tanks (1 foot diameter by 6 feet taI I )
are equipped with a vertically mounted, internal,
cylindrical screen. The feed slurry pumped from the feed
tank enters the bottom of the head tank and into the
center of the screen. The portion of the slurry which Is
del ivered to the rotary atomizer must pass through the
cylindrical screen wall In order to enter the atomizer
feed line. The remainder of the slurry continues through
the center of the head tank and out the open top of the
screen to the overflow pipe, which returns the slurry to
the feed tank. The screen self-cleans as this return flow
through the center of the screen washes It's inner
surface. During the early scrubber operation at the
design spraydown rate and full gas flow, the head tank
screens periodically restricted flow to the atomizers.
The problem was found to be insufficient flow through the
interior of the screen to provide the se I f-c I ean ing
effect. Using the adjustable pitch sheaves provided on
the feed pumps, the pump speeds were increased, thereby
providing additional head tank flow.
One of the prime factors for successful drying of the feed
slurry is maintaining sufficiently high feed solids. A
basic concept of general spray drying technology is that
10-190
-------
It Is easier to dry a thick, or higher solids slurry, than
a thin, or low solids one. During the start-up operation
of the scrubber, periodic tripping of the recycle feeder
to the mix tank caused low feed solids. This, along with
a period of operation without the Implementation of the
gas flow feed forward signal to the atomizer feed control-
lers, caused some Intermittent spray dryer wall deposits.
The deposits did not interrupt scrubber operation,
however, pluggage of the spray dryer bottom discharge did
occur as dry lumps fell from the dryer walls. An Inter-
lock of outlet temperature to the feed solids has been
programmed Into the controls to automatical ly raise the
outlet temperature in the event of a low feed solids
content, thereby reducing the possibility that deposits
will develop within the dryers.
BaghoiLse
The first and perhaps most significant problem encountered
with the baghouse operation was high differential
pressure. From the time of the Initial reverse air clean-
Ing at four inches pressure differential, it became
apparent that the reverse air was not effectively removing
the dust cake on the bags. During the first two weeks of
operation, baghouse pressure drops ranged from four Inches
to over nine inches water gage, depending on boi ler load.
For full load boiler operation, the baghouses had to be
bypassed to assure that the I.D. fans would not enter a
staI I condIt I on.
Although it was realized that the higher gas temperatures
through the baghouses (since the spray dryers were still
off line) were resulting In higher gas volumes and a lower
reverse air fan efficiency, these factors could not ex-
plain the excessive pressure loss. Inspection of the
fabric dust layers showed heavy build-ups of caked dust on
the bags. Laboratory analysts of fabric samples confirmed
that the bags had experienced a dew point encounter caus-
ing soluble sulfate formations resulting in agglomerations
and nodula formation. The level of agglomeration and cake
pluggage was not considered severe and the dust cake was
easily removed by manual agitation of each individual bag.
Current baghouse pressure differential has stabilized at
approximately five and one-half inches water gage at
desIgn cond111ons .
During the initial two weeks of operation, stack opacities
of 10-20$ were observed. The reason for visible emissions
10-191
-------
from the baghouse was apparent after the June 12, 1983,
shutdown, when It was discovered that several of the
originally supplied bypass damper blade seals had torn
loose, possibly due to high boiler load bypass operation.
Velocities through the louver damper were in excess of 100
ft/sec. The design of the blade seals was modified and
new seals were Installed.
lie merits.
The dry scrubber system is designed to remove 62% of the
S0? when the average sulfur coal (0.68?) is burned and 78$
for the maximum sulfur coal (1.22?). Unit 2 of the
Antelope Valley Station is scheduled to begin commercial
operation in July, 1986. At that time. Unit 1 SO- removal
requirements will increase to 81? for average sulfur coal
and 89? for maximum sulfur coal. The S02 emissions from
Unit 1 and Unit 2 combined are limited to 3,845 pounds per
hour (0.39 pounds per million Btu), as a requirement of
the Permit to Construct issued by the North Dakota State
Department of Health.
ia n.ce_Ies±i
Stack emission tests were conducted on August 23, 1983, to
determine compliance with the Permit to Construct. The
spray dryers had only operated for a total of about five
weeks at the time of the emission compliance tests. The
scrubber system had not reached design operating condi-
tions, but was determined to be capable of meeting
emission compliance requirements, so the tests proceeded.
The results of the compliance tests are summarized in
Tab I e 1 .
Pe_£±or_ma_n.ce_J_ests.
Performance tests to determine guarantee compliance with
the Joy contract were conducted September 26-30, 1983.
During the tests, the SO and particulate removal system
operated at the conditions specified in the contract for
normal operating conditions with the average sulfur coal.
EPA reference methods were used to determine SO and
particulate concentrations and gas flows at the inlet to
10-192
-------
Table 1. Emission Compliance Test Results
Parameter
Test Result
Permit Limitation
Gross Generation, MW
Stack Gas Flow, ACFM
Stack Temperature, °F
Total Sol ids In Slurry, %
Inlet S02, PPM
Stack S02, PPM
Stack S02, #/hr
Stack S02, #/MKB
SO Remova 1 , %
Stack NO , #/hr
Stack NO , #/MKB
X
Stack Particulates, #/hr
Stack Particulates, #/MKB
468
1,880,000
222
24
668
171
2,308
.46
74
2,222
.45
72.7
.015
3,845
2,465
210
10-193
-------
and in the stack, and S02 and gas flow of the bypass
reheat duct. Lime consumption was determined from the
pebble lime weigh belt feeder, with the lime being
analyzed for available CaO. The lime consumption was
verified with the slaked lime slurry analysis and flow
rate. Pressure drop was confirmed from static pressure
readings taken at several locations to account for the
pressure drop from the outlet of the air heater to the
inlet of the ID fan, and from the outlet of the ID fan to
the entrance of the stack.
An enthalpy balance was used to confirm the gas flows and
the pressure drop used In the evaluation. Electrical
energy consumption was measured with a watt transducer.
Samples of coal, raw lime, slaked lime, feed slurry,
treated water, dilution water (cooling tower blowdown),
scrubber Inlet flyash, and recycle products were collected
and analyzed. Temperatures were taken of feed slurry, and
spray dryer inlets, outlets, bypass reheat, and stack
gases .
L§-S.±_Re-S_u.lts.
Preliminary results of the guarantee performance tests on
the S0? and particulate collection system are listed In
Tablet. As of this writing, sample analyses have not
been completed to determine the actual sto I chIometr I c
ratio. An estimate is used here based on correlation of
aval I abIe data.
During the performance tests, the ash conveying system
from the baghouse to the recycle bin was not conveying
adequate quantities of ash. To replace that ash, a truck
was used to transport ash from the ash storage silo to the
recycle bin. Preliminary results Indicate that this ash
contained higher alkalinity than the ash collected from
the baghouse. This Issue Is currently under investigation
to determine its impact on stoIchIometrIc ratio, if any.
Testing was also conducted under conditions without reheat
by-pass to assess the guarantees at 62%, 81$, and 89%
removal. The preliminary results Indicate a slight In-
crease in stoichiometric ratio when the reheat by-pass was
not used for 62% removal. A stoIchIometrIc ratio of about
0.9 was used for 8]% removal, which was well below the
guarantee of 1.1.
A test at the high removal rate (89$ guarantee) was
10-194
-------
Table 2. Performance Test Results
Parameter Test Result Guarantee Conditions
Spray Dryer Inlet Gas Flow, ACFM
Spray Dryer Inlet, °F
Spray Dryer Outlet, °F
Approach to Saturation, °F
Total Sol Ids In Slurry, %
Baghouse Outlet, °F
Inlet S0_, ppm
Outlet SO , ppm
Outlet S02, Ib/hr
SO Remova 1 , %
Stolen lometr Ic Ratio
(for 63.9$ removal )
Stolen lometr Ic Ratio
(corrected for 62% removal)
Inlet Partlculate, gr/dscf
Outlet Particulate, gr/dscf
Outlet Partlculate, Ib/hr
Partlculate Removal, %
Total Pressure Drop, In. W.G.
Power Consumption, KW
Reheat By-pass, 10 Btu/hr
2,039,641
314
166
29
41
189
655
220
2,973
63.9
.47
.45
2.8496
.0025
24
99.91
14.96
2,341
7.1
185
800
304
3,844
62
.54
.012
210
13.4
2,757
7.2
10-195
-------
conducted using data obtained from the plant monitoring
equipment, and correlated to the previous EPA reference
method tests. A s t o I ch Iometr Ic ratio of 1.45 Is
guaranteed for 89$ removal with the maximum design sulfur
content of 1,380 ppm S02 Into the system. The SO^ inlet
during the test was only 442 ppm. Re'moval of 92$ was
obtained with a stoIchIometric ratio of about 1.0.
Although some preliminary optimization testing has been
conducted on the scrubbing system, additional testing will
resume when the unit returns on-line In December. The
testing will focus on maximizing lime utilization and
minimizing differential pressure across the baghouse and
spray dryers. The parameters to be studied during this
testing will be sto I chIometrIc ratio, approach tempera-
ture, and reheat usage.
Further testing of this system will also be conducted at
81$ and 89$ S0? removal conditions. The testing at these
conditions will be used to confirm operation for the
stringent removal requirements when Unit 2 begins opera-
tion and when the maximum sulfur content coal is burned.
Other tests will be performed to determine the actual
turn-down capabilities of the spray dryers.
Engineering studies are currently planned to Investigate
the possibility of decreasing the system differential
pressure by modifying existing duct work. Replacement of
the existing slurry feed pumps and piping is also being
considered In order to operate at lower approach
temperatures. Lower approach temperature operation should
Improve lime utilization and assist in maintaining the
plant water balance.
10-196
-------
REEEBENCES
1) Janssen, Kent E. and Erfksen, Robert L . , "Basin
Electric's Involvement with Dry Flue Gas
Desulfurization". Presented at the EPA Symposium on
Flue Gas Desulfurization, Las Vegas, Nevada;
March, 1979.
2) Davis, R.A., Meyler, J.A., Gude, K.E., "Dry SO-
Scrubbing at Antelope Valley Station". Presented ar
the American Power Conference, Chicago, I I I Inols;
Apr I I , 1979
3) Eriksen, Robert L., "The Development of Dry Flue Gas
Desulfurization at Basin Electric Power Cooperative".
Presented at the Second Conference on Air Qual ity
Management in the Electric Power Industry, Austin,
Texas; January, 1980.
10-197
-------
CHARACTERIZATION OF AN INDUSTRIAL SPRAY DRYER AT
ARGONNE NATIONAL LABORATORY
P. S. Farber, C. D. Livengood
-------
CHARACTERIZATION OF AN INDUSTRIAL SPRAY DRYER
AT ARGONNE NATIONAL LABORATORY
by: P.S. Farber and C.D. Livengood
Energy and Environmental Systems Division
Argonne National Laboratory
Argonne, 111. 60439
ABSTRACT
Argonne National Laboratory (ANL) is operating an industrial-scale,
coal-fired boiler with a flue gas cleaning (FGC) system consisting of a
spray dryer and fabric filter. This paper presents a description of the FGC
system together with a status report for an EPA-sponsored project being
carried out by ANL to characterize the operation of the system. This
project involves a design and economic analysis of the FGC system,
determination of waste characteristics, and analysis of system operation
through monitoring of inlet/outlet gas streams and sampling of various
process streams. Preliminary data and material balances are presented in
the paper, as well as a proposed performance model based on an analysis of
key operating parameters.
INTRODUCTION
"Dry" scrubbers consisting of spray dryers and fabric filters or elec-
trostatic precipitators have recently emerged as a viable pollution control
option for coal-firing electric utilities and industries. Such integrated
flue gas cleaning (FGC) systems have several potential advantages over con-
ventional wet scrubbers, including higher reliability, lower cost (for com-
bined sulfur oxides and particulate matter control), lower consumption of
energy and water, and greater ease of waste disposal. The latter point is
due primarily to the production of a dry powdery waste that is readily
handled, transported, and disposed of in landfills.
Most existing or planned dry scrubbers have been designed for systems
firing low- to medium-sulfur coals. However, Argonne National Laboratory
(ANL) has been operating the first commercial spray-dryer FGC system
designed for high (3.5%) sulfur coal since November 1981. This system
treats the flue gas from a spreader-stoker boiler rated at 170,000 Ib/h of
10-199
-------
steam. Sulfur dioxide (S02) removals of about 80% are achieved routinely to
meet the emission limit of 1.2 lb/10" Btu mandated by the State of Illinois.
Because of the unique nature of this system, the extensive provisions
made for monitoring and sampling, and the ready availability of research
facilities at ANL, the U.S. Environmental Protection Agency* (EPA) is spon-
soring a program for characterization of the FGC system. The objective of
this program is to document design and construction experience with the ANL
system and to achieve a thorough understanding of its operation so as to
develop predictive performance correlations. The first part of the program
utilized the information compiled during the specification, bidding, con-
struction, and initial operation phases. The analysis of specifications
revealed items critical to the procurement of a spray dryer FGC system. The
bid analysis reviewed the range of bids received on the basis of both
capital and operating costs. This analysis of equipment and installed costs
will give potential users of spray dryer FGC systems an insight to expected
capital and operating expenditures.
A second part of the program involves characterization of the FGC
system over a 60 to 90 day period. This characterization will entail exten-
sive data acquisition and stream sampling, and will relate system operation
to process conditions.
A third program activity is determination of the properties of the dry
waste produced by the FGC system. Various mixtures of flyash and spent
sorbent have been subjected to the EPA Extraction Procedure (EP) Toxicity
Test in order to determine the extent and composition of any potential
leachate. The results are useful in evaluating disposal characteristics and
determining the partitioning of coal-derived trace elements between ash and
sorbent particles.
This paper presents a brief description of the system, a summary of the
EP test results, and selected operating data obtained in a preliminary test
program. The derivation of material balances around key system components
is demonstrated, and a simple performance model is proposed.
SYSTEM CHARACTERISTICS
ARGONNE BOILER NO. 5
The spray dryer system is installed on Boiler No. 5 at the ANL heating
plant. No. 5 is the largest of the five boilers at the plant, with a name-
plate rating of 170,000 Ib/h of saturated steam at 200 psig. It is a Wickes
*Emissions/Effluent Technology Branch, Industrial Environmental Research
Laboratory, Research Triangle Park, N.C., Theodore G. Brna, Project
Officer.
10-200
-------
(now Combustion Engineering) spreader-stoker unit that was installed in 1965
with the capability to fire either coal or gas. Oil-firing capability was
added as a backup in 1973 when coal use was discontinued at ANL.
In 1980, as part of the DOE effort to maximize coal use and as an econ-
omy measure for the Laboratory, the coal handling equipment at the heating
plant was refurbished and Boiler No. 5 was reconverted to coal firing.
However, it was also necessary to install an FGC system capable of providing
compliance with the applicable State of Illinois emission regulations shown
in Table 1 since existing equipment consisted solely of cyclone separators
for flyash collection. These continue to be used and considerably reduce
the flyash burden on the FGC system.
Several midwestern and eastern coals have been successfully fired in
Boiler No. 5 since the reconversion, ranging in sulfur content from about
1.5% to 4.5%. Table 2 gives a typical analysis for the coal currently being
used, which is from the Illinois basin and is representative of the fuel for
which the FGC system was designed. During acceptance tests using that coal,
S02 emissions of 0.95 lb/106 Btu and total suspended particulate emissions
of 0.007 lb/106 Btu were measured.
FLUE GAS CLEANING SYSTEM
Although spray dryer systems are commonly referred to as "dry," they do
involve the preparation and handling of a slurry. For this discussion, it
is convenient to divide the system into its wet and dry subsystems. The wet
subsystem encompasses pebble lime storage, recycle powder storage, lime
preparation, and slurry preparation. The dry subsystem is basically the gas
train and associated equipment, and includes the spray dryer, fabric filter,
ducts, booster fan, and air compressor. The relationship between major
components is illustrated in Figure 1.
The feed slurry is composed of a mixture of fresh slaked lime and re-
cycled "spent" sorbent. The recycled material contains a significant com-
ponent of unreacted calcium and thus raises the effective Ca/S ratio above
that indicated by the fresh lime alone. Pebble lime from the lime silo is
fed to a weighbelt feeder, which supplies a paste-type slaker where the lime
is converted from calcium oxide (CaO) to calcium hydroxide (CaCOH^). Some
water is added at the outlet of the slaker to dilute the milk of lime to a
concentration of approximately 15% (by weight) solids. This milk of lime
then flows to the milk of lime tank through a rotary screen, which removes
the "grits," or inerts.
From the milk of lime tank, the lime slurry is pumped to the slurry mix
tank. In this tank, milk of lime, recycled spent sorbent powder, and some
additional water are combined into an approximately 35-40% (by weight)
slurry. This slurry is then pumped to another rotary screen (to ensure
10-201
-------
TABLE 1. EMISSION LIMITS FOR BOILER NO. 5
Pollutant Limit
Sulfur Dioxide 1.2 lb/106 Btu
Particulate Matter 0.1 lb/106 Btu
and <20% Opacity
TABLE 2. TYPICAL COAL CHARACTERISTICS
Coal Parameter Value
Heating Value, Btu/lb 12,027
Moisture, % 9.59
Ash, % 7.40
Carbon, % 66.98
Hydrogen, % 4.65
Nitrogen, % 1.48
Sulfur, % 3.32
Chlorine, % 0.06
Oxygen, % 6.52
removal of any large particles which might clog the atomizer nozzles) and
then to the slurry feed tank. Overflow from the slurry feed tank goes back
into the slurry mix tank, so there is a continual circulation between the
tanks. Slurry from the feed tank is pumped, at a constant high flowrate, to
a head tank located above the atomizer. A control valve regulates the
amount of slurry fed to the atomizer, with the excess being returned to the
feed tank. This returned slurry also passes through the rotary screen that
is filtering the stream from the slurry mix tank.
Flue gas, exiting the boiler's induced draft (ID) fan, passes into a
modified breeching at the existing stack. A guillotine damper diverts the
flue gas flow into the FGC system ductwork leading to the spray dryer. This
inlet ductwork splits the flue gas into two streams. One stream, with about
60% of the gas flow, is directed into a roof gas disperser, located on the
top of the spray dryer (Figure 2). The remainder of the gas stream enters a
central gas disperser, located in the middle of the spray dryer. Both gas
10-202
-------
To
Atmosphere
o
I
r-o
o
LO
Slaking
Water
Dust
Suppression
Water
-CM
Waste
Disposal
Pebble
Lime
Silo
Slaker
and Milk
of Lime
System
Figure 1. Argonne National Laboratory FGC System
-------
Sorbent Flow into Rotary Atomizer
Central
Gas
Disperser
Spent
Sorbent
Outlet
Figure 2. Spray Dryer
10-204
-------
streams, upon entering the dryer, are given circular motions with their main
directions of flow being opposed to each other. In the spray dryer the feed
stream of slurry is atomized into fine droplets, through the centrifugal
action of the rotating disk, and introduced into the gas stream. Some of
the dry powder produced in the spray dryer falls to the bottom of unit and
is picked up by a drag link conveyor for storage in the recycle/disposal
silo.
The remainder of the powder is entrained in the gas stream and is car-
ried into the four-compartment pulse-jet fabric filter. In the fabric
filter these particles are separated from the gas stream by fiberglass bags.
The cleaned flue gas exits the baghouse and travels through the outlet duct
to the booster fan. From the booster fan the gas enters a new breeching
section of the stack, and then into the stack itself for discharge to the
atmosphere.
Wet Subsystem
Pebble Lime Silo—
This vessel is constructed of carbon steel, and is 12 ft in diameter
with a 32 ft high straight side. There is a partial-cone lower section on
the vessel, with a "live-bin" vibratory bottom. The silo is designed to
contain up to 110 tons of high-calcium pebble lime, which is approximately a
7-day supply at full boiler load.
Recycle/Disposal Silo—
Standing 46 ft high with a 20 ft diameter, this silo is the larger of
the two dry materials silos in the FGC system. As with the lime silo, this
vessel is constructed of carbon steel. However, it has a flat bottom and is
equipped with an air fluidizing system (air-slide) to assist in powder flow.
Two outlets in the bottom are for recycle of spent sorbent and for sorbent
disposal. The recycle silo has a capacity of 180 tons of dry powder which,
at full boiler load, is sufficient for 7 days' accumulation.
Lime Slaking System—
This system consists of a Wallace and Tiernan paste-type lime slaker
(Series A-758), with a capacity of 2000 Ib/h, and a mechanical weighbelt
feeder (Model 31-120AV). The feeder is equipped with an adjustable-speed
belt drive and a transmitting switch for actuating the totalizing throughput
counter. The slaker, made of heavy gauge steel, consists of a slaking com-
partment containing two sets of counter rotating intermeshing paddles for
mixing, a dilution chamber with rakes for agitation, and a dust and vapor
arrester. In the slaker, water and pebble lime, in a ratio of approximately
2 to 1 by weight, are fed continuously into one end of the slaking compart-
ment. Lime feedrate governs the operating rate of the entire system, and
this feedrate is controlled on the basis of the liquid level in the milk of
10-205
-------
lime tank. Lime-paste consistency controls the addition of water. An in-
crease in torque on the mixing shafts (indicating a thicker paste) opens a
water-control valve to admit additional water.
Vibrating Screens—
These screens are both "Rotex Single Surface Liquatax Separators," 2 ft
by 2 ft with 4 in. inlet nozzles and 3 in. outlet nozzles. One screen, the
classifier screen, has a 40-mesh stainless steel screen and is at the outlet
of the lime slaker. The purpose of this screen is to separate the "grits"
from the slaked lime slurry. The overs from this screen flow through a pipe
into a dumpster for disposal. The slurry passing through the screen dis-
charges to the milk of lime collection tank.
The other screen, called the slurry feed screen, is a 6-mesh stainless
steel screen which processes the slurry from the slurry mix tank as well as
the returned slurry stream from the atomizer head tank. As with the classi-
fier screen, the overs are sent to a dumpster for disposal. The slurry
discharging from this stream falls into the slurry feed tank.
Tanks—
There are four tanks associated with the wet subsystem. These are the:
• Milk of Lime Collection Tank,
• Slurry Mix Tank,
• Slurry Feed Tank, and the
» Head Tank.
The milk of lime collection tank is an agitated, baffled tank with a
600-gal capacity. The tank, constructed from 1/4-in. thick ASTM-A283 plate,
is 60 in. in diameter and 77 in. high. Based on the maximum milk of lime
flow, the tank residence time is approximately 20 minutes.
The slurry mix tank and the slurry feed tank are identical in construc-
tion (920 gal), both being 60 in. in diameter and 101 in. high. As with the
milk of lime collection tank, they are agitated, baffled tanks constructed
of 1/4-in. A283 carbon steel. Their residence times are each approximately
20 minutes based on maximum slurry flows. In a sense these two tanks may be
considered as a single large unit. This is due to the fact that the slurry
mix tank continuously feeds the slurry feed tank at a rate of 45 gallons per
minute (gpm), with the overflow from the slurry feed tank being returned to
the slurry mix tank.
The head tank, located above the atomizer in the spray-dryer penthouse,
is 6 in. in diameter and 4 ft high, and is constructed of carbon steel. It
10-206
-------
supplies a hydrostatic head for the slurry feed to the atomizer. Feed
slurry is pumped to the head tank at a constant rate of 45 gpm. The flow of
slurry to the atomizer from the head tank is modulated by a temperature-
controlled valve in the line between the head tank and the atomizer. This
valve responds to controller signals to ensure a constant gas temperature at
the spray dryer outlet. Any slurry in the head tank that is not sent to the
atomizer overflows into a return line to the slurry feed screen. The con-
stant high flowrate to the head tank is necessary to ensure that, during
periods of low slurry demand by the system, pipeline velocities are kept
high enough (6-8 ft/s) to prevent the feed slurry from settling out in the
long vertical feed line.
Pumps—
All pumps (milk of lime, slurry transfer, and slurry feed) are Warman
1- 1/2 in. x 1 in. Type BM with V-belt drives and overhead motors. These
pumps are split-casing rubber-lined units with rubber-lined open impellors.
The pumping systems are all redundant, with an identical "spare" pump con-
nected with parallel piping and valves to the operating unit. Each pump is
designed to deliver 45 gpm of fluid in order to ensure high velocities in
the piping systems.
Agitators—
These Galigher Model GWO motor-driven agitators are, along with the
pumps, the only rubber covered equipment in the entire sorbent preparation
system. Their purpose is to ensure that all slurries stay in suspension,
although the slurry mix tank agitator (which is fitted with two sets of
impellors) has the additional duty of mixing milk of lime, recycled dry
sorbent, and dilution water into a uniform slurry.
Dry Subsystem
Spray Dryer—
This is a dual-inlet spray dryer of carbon steel construction designed
by Niro Atomizer. The unit is 27 ft 7 in. inside diameter, with a 19 ft
straight section and a 19 ft long cone bottom. This cone bottom has a 3 ft
diameter bottom outlet which is connected to a Niro-designed powder cooler,
no longer utilized in the process. With an interior volume (less the
central gas disperser volume) of 15,000 ft , the spray dryer has a gas
residence time of 12 seconds at the maximum boiler load. This residence
time is based on the inlet gas volume to the spray dryer. Should an average
between the inlet and the outlet volumes be used, the value rises to almost
14 seconds.
The spray dryer was constructed with a dual gas inlet, such that 60% of
the flue gas enters through the roof gas disperser, while the remaining gas
flow enters via the central gas dispenser. These two gas streams are
10-207
-------
projected against each other, forming a turbulent swirl into which the
slurry is injected by the high speed atomizer. This slurry mixes with the
hot flue gas stream, reacting with the sulfur dioxide, and drying into a
fine powder. Some of this powder falls to the bottom of the spray dryer,
where it is picked up by a drag-link mechanical conveyor. The remainder of
the powder is entrained in the gas stream and exits the spray dryer.
The atomizer used in the Argonne FGC system is a Niro Model F-160
equipped with a 200 hp 3570 rpm motor. A gear drive, within the atomizer
body, increases the rotational speed to 14,800 rpm. This high speed spins
the 8-1/4 in. diameter atomizing disk to a tip speed of 533 ft/s. The
atomizing disk is constructed with eight equally spaced 3/8-in. diameter
holes, through which the slurry is injected into the gas stream.
Baghouse Fabric Filter—
The baghouse fabric filter is a Joy Manufacturing, Model 6012, pulse-
jet unit. It consists of four compartments, each having 28 rows of 10
filter bags. The filter compartments are constructed of carbon steel, and
equipped with electrical plate-type heaters on the hopper (lower surface)
walls. These hopper heaters are thermostatically controlled, and ensure
that no moisture condenses in the compartments. Dust-laden gas enters each
filter compartment in the lower plenum. The gas travels upward, through the
bags, into the upper plenum of the filter compartment. From the filter
compartment, the gas travels through ductwork to the booster fan.
The filter bags are fabricated of a woven 16 oz fiberglass fabric with
a Teflon coating. These filter bags are 6 in. in diameter by 12 ft long and
are fitted onto wire mesh cages. Each row of filter bags has a pipe, which
is located over the bags and runs the entire length of the row. These pipes
have holes in them above each bag and are connected through solenoid valves
to a compressed air header. As each solenoid is actuated in turn, it sends
a pulse of compressed air into the pipe. The compressed air exits through
the holes above the bags, entrains some additional gas with the aid of a
venturi nozzle fastened to the top of the bag cage combination, and travels
down the length of the bag. This pulse of air breaks loose some of the cake
collected on the outside of the filter bag. This broken cake falls to the
bottom of the filter compartment, and into another drag-link conveyor.
Booster Fan—
This is a belt-driven induced draft fan manufactured by Chicago Blower
Corporation (Model 5414). The fan wheel is 54-1/4 in. in diameter and is
driven (via the belt drive) by a 250 hp, 1800 rpm motor. The fan is de-
signed for 62,000 ft3/min of gas at 19 in. W.C. static pressure while
running at 1335 rpm. Flue gas from the baghouse fabric filter system is
ducted to this fan and, from there, directly to the stack for discharge to
the atmosphere.
10-208
-------
RESULTS AND CONCLUSIONS
DESIGN AND ECONOMIC ANALYSIS
Five vendors submitted proposals in response to the Request for Pro-
posals (RFP) issued by ANL for installation of a dry FGC system. In the
spray dryer, four vendors proposed to use rotary centrifugal devices to
atomize the lime slurry feed; the fifth proposed two-fluid nozzles to
achieve the atomization. The diameter of the dryer chambers ranged from 23
to 27.6 ft, with straight-wall height ranging from 10 to 22 ft. Solids con-
centration of lime-slurry feed varied from 14 to 35% by weight. Estimated
pressure drop across the dryer, at maximum boiler operating conditions,
varied from 1.5 to 5 in. W.C., with estimated power consumption for operat-
ing the dryer varying from 66 to 111 kW.
For collection of flyash and spent sorbent, reverse-air and pulse-jet
fabric filter baghouses were proposed, with Teflon-coated fiberglass bags
used in four of the proposed systems. The gross air-to-cloth ratio for
these baghouse systems ranged from 1.34 to 4.60, and the resultant net air-
to-cloth ratio ranged from 1.81 to 6.16. The estimated pressure drop across
the baghouse varied from 5.5 to 7.5 in. W.C., resulting in an estimated
power requirement for operation of this subsystem of 17 to 30 hp.
As for overall system chemical and utility consumptions, the estimated
lime (92% pure pebble lime) utilization rate for the proposed FGC systems
ranged from 1,130 to 1,600 Ib/h at maximum boiler capacity. Estimated total
water usage varied from 20 to 40 gpm, and estimated total system pressure
drop varied from 8.75 to 16 in. W.C., with a resultant estimated total power
consumption for operation of these FGC systems ranging from 239 to 545 kW.
Both the total system capital costs and the estimated annual operating
costs varied over a range of more than two to one among the various pro-
posals. The final capital cost for the system as constructed was $3.5
million.
After completion of the vendor selection process, a letter of intent
was awarded to Niro Atomizer, Inc. in the first week of November 1980.
Ground was broken after general facility arrangement drawings were finalized
for construction in the second week of April 1981, and erection of steel and
platforms was completed on August 7, 1981. Delivery of major equipment com-
menced in the first week of June 1981 and by the first week of October, both
the baghouse and spray dryer subsystems had been installed. All mechanical
work, piping, instrumentation, and electrical work were completed in mid-
October, and final painting and insulation were finished in the first week
of November 1981. Construction was about 95% complete at this time (about 1
year after the start of the project), and Argonne and Niro decided to move
into the start-up phase of the project.
10-209
-------
WASTE CHARACTERIZATION
Waste from the Argonne system, which consists of a mixture of spent
sorbent and flyash, is currently landfilled onsite. Spent sorbent is dis-
charged (see Figure 1) from the recycle/waste silo into a dump truck.
Approximately 12% of this material is flyash. The bulk of the flyash is re-
moved from the flue gas by cyclones for reinjection into the boiler. That
ash which is not reinjected is collected in a separate silo and transported
by truck to the landfill site. Various mixtures of ash and spent sorbent
have been subjected to the EP Toxicity Test in order to determine the ex-
tent and composition of any potential leachate. The results indicate that
the leachate composition is a function of both the pH of the leaching solu-
tion and the composition of the waste (ash/sorbent) blend.
Samples of both spent sorbent and flyash were collected under typical
operating conditions while firing the coal specified in Table 2. The S02
removal was approximately 80% at the time. The spent sorbent (which also
contains a small amount of flyash) was obtained from a sampling tap at the
base of the storage silo, while the flyash sample was obtained from the
boiler's cyclone separators.
The test procedures detailed in Ref. 1 were followed to generate liquid
samples for elemental analysis. However, in several cases the pH could not
be adjusted down to the target value of 5.0 ± 0.2 due to the high available
alkali in the spent sorbent. If the pH was below 5.0 (pure flyash), no
adjustments were made.
An Inductively Coupled Argon Plasma Spectrometer (ICAP) was used for
elemental analyses of unblended flyash and sorbent samples. These measure-
ments involved 32 different elements. For the leachate analyses, a Perkin-
Elmer Model 5000 atomic absorption spectrometer was employed. Twenty-two
elements were measured in these tests. In each case, two independent mea-
surements were made for each sample.
One purpose of the analyses was to compare the waste streams from this
system with those from other units in operation. Results for several
elements are shown in Table 3, and are compared to three samples of spent
sorbent from lime-based spray dryers treating flue gas from low-sulfur coal
combustion, as analyzed by Radian for the Electric Power Research Institute
(EPRI).3
Since the ANL boiler employs high efficiency cyclones for flyash col-
lection, it is understandable that ANL's spent sorbent would contain less
flyash than is found in the waste from a utility system. Thus, one would
expect that the utility waste properties would be found between those of
ANL's two waste streams if all other factors were equal. This fact is
illustrated by the amounts of iron and aluminum (common constituents in
10-210
-------
Table 3 ELEMENTAL ANALYSIS OF SAMPLES
Concentration, ppm
Element
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Molybdenum
Nickel
Potassium
Selenium
Silver
Sodium
Strontium
Thallium
Tin
Titanium
Vanadium
Zinc
ANL
Flyash
68,100
<500
<500
130
40
500
70
7100
130
80
150
367,000
150
3000
210
ND
<10
460
2600
ND
<10
2800
250
ND
<30
4000
270
330
Research-
Cottrell
Waste3
41,000
<8
14.2
400
5.1
NDa
<1.0
170,000
50
14
84
22,000
4.4
12,000
200
ND
<0.5
26
3100
6.1
<0.5
12,000
1800
<25
<30
4900
70
43
Rockwell-
Joliet
Station
Waste3
71,000
<8
33
190
8.5
ND
<1
97,000
54
15
81
44,000
17
13,000
110
ND
8.7
39
6300
4.7
0.5
15,000
1800
<30
<30
5200
120
92
Joy/Niro
Riverside
Waste3
58,000
<8
30
350
4.3
ND
<1.0
150,000
52
4.9
16
20,000
<20
15,000
630
ND
16
215
4300
<20
<0.5
2700
1900
<25
<36
3100
580
37
ANL
Spent Sorbent
3200
<500
<500
20
4
500
<5
283,000
20
<10
40
6300
130
5500
100
ND
ND
30
1700
ND
ND
1400
170
ND
<30
230
30
120
aNot Detected.
10-211
-------
flyash) in the three utility sorbents compared to ANL's flyash and sorbent.
Iron comprises over 35% of the total mass in Argonne's flyash, with aluminum
being about 7%. In the spent sorbent, they are found to be 0.6% and 0.3%,
respectively. The utility spent sorbents averaged 2.9% iron and 5.7% alum-
inum. Examining many of the trace elements normally found in flyashes,
similar comparisons could be made between the ANL and utility wastes.
On the other hand, different relationships arise because the utility
sorbents are all the result of combustion of low-sulfur western coal whereas
ANL's flyash is from high-sulfur midwestern coal combustion. For example,
the presence in the utility wastes of large amounts (compared to either the
ANL sorbent or flyash) of potassium and sodium is probably due to basic
differences between eastern and western flyashes. The fact that magnesium
is detected in the utility samples at levels ranging from three to five
times that found in the Argonne samples could be due to two factors. Cer-
tain western flyashes contain large amounts of magnesium compared to eastern
ashes. Another factor could be related to the type or quality of lime used
in the FGC system. Argonne uses a lime which contains approximately 2%
magnesium, and it is possible that the lime used in the utility units had a
higher magnesium content. Coal-related differences in the nature of the
wastes are also seen in the analyses of barium and titanium (more prevalent
in the utility sorbents than in either the flyash or sorbent from ANL) and
boron (undetectable in the utility sorbents, but found at the same (high)
level in both the ANL sorbent and flyash).
Figures 3-5 show typical results from analysis of leachate from the ANL
sorbent/flyash mixtures. A more complete presentation may be found in Ref.
4. In general, three different pH-sensitive trends were observed: 1) metal
concentration increasing with increasing pH, 2) a decrease in concentration
with increasing pH, and 3) specific sensitivity (high concentration) at
approximately pH 5.
The variation of boron concentration with percent sorbent (and hence
pH) is given in Figure 3. In this case, concentration rises with increasing
pH, although a slight dip is noticed at 75% sorbent, or about pH 11.8.
Since wastes from a dry scrubber system will tend to be basic in nature when
initially placed in a landfill or other site, it may be concluded that until
the nature of the waste is modified by soil conditions extensive leaching of
boron and other similarly behaving elements may occur.
Plant growth experiments using soil treated with dry scrubber wastes (0
to 4% by weight) have been conducted at Argonne.^ For both corn and soy-
beans, the biomass decreased as the amount of dry scrubber wastes increased.
Boron concentrations in the leaves of the corn and soybeans grown in soil
treated with scrubber wastes were found to be up to 20 times greater than
those found in the leaves of the control (no dry scrubber wastes).6
10-212
-------
Legend
O PPM
X pH
0.25
0.50
FRACTION SORBENT
0.75
Figure 3. Boron Concentration in Leachate Varies
with Fraction Sorbent and pH
Several elements, including cadmium (Figure 4), were found to have
lower concentrations in leachate from the sorbent and sorbent mixtures than
from the 100% flyash samples. This trend also follows for pH, in that the
concentrations in the leachate decreased as pH increased. This effect is
not unexpected, as it has been seen for these elements in many other
wastes. If leaching of cadmium compounds is actually retarded by high pH,
impoundment of certain chemical wastes, which contain these and like mate-
rials, together with dry scrubber wastes may offer distinct advantages which
should be explored.
Perhaps most interesting are the cases where response was greatest at a
pH of approximately 5. Antimony (Figure 5) showed a concentration of
approximately 60 ppb at pH 5 while levels for the 100% flyash (pH 4) and the
pure waste sorbent (pH 12) were virtually nondetectable. Similar results
were found for arsenic, mercury, and selenium. In the toxicity test pro-
cedure, pH control of the material is carefully specified. The reasons for
the high leaching sensitivity of certain metals at, or about, the pH speci-
fied in the EP test procedure are not clear. Further work to determine the
10-213
-------
1-14
Q
<
O
0.25 0.50
FRACTION SORBENT
Figure 4,
Cadmium Concentration in Leachate Varies
with Fraction Sorbent and pH
Legend
O PPM
X pH
reasons for this occurrence (e.g., the exact chemical form of the metallic
compounds in the waste, and equilibrium conditions during the test) appears
to be warranted.
Under RCRA, if the EP filtrate has a concentration of a substance that
exceeds 100 times the Interim Primary Drinking Water Standard, then the
waste can be classified as hazardous. Presently wastes from flue gas clean-
ing systems are excluded from regulation as hazardous wastes. Although no
element had a concentration greater than 100 times the standard in ANL
tests, it should be noted that the chromium levels were within 2% of the
maximum.
HIGH SULFUR COAL TEST BURN
During April 1983, ANL and the Consolidation Coal Company conducted a
test burn of very high sulfur (4.2-4.5%) coal in Boiler No. 5. This test,
complementary in concept to the EPA-sponsored work, was undertaken to
demonstrate:
10-214
-------
80-,
o
OQ
Q_
Q_
0.25
0.50
FRACTION SORBENT
0.75
Legend
O PPB
X pH
Figure 5. Antimony Concentration in Leachate Varies
with Fraction Sorbent and pH
• Ninety percent or greater
period,
removal over a 100 hour
• Ninety percent SC>2 removal at a stoichiometry (inlet moles
of CaO divided by inlet moles SC^) of 1.4 or less, and
• Seventy percent SC>2 removal across the spray dryer alone
(the balance being attained in the baghouse).
The test consisted of extensive sampling and analysis of process
streams, coupled with acquisition of process data from plant instrumentation
and manual sampling (EPA Method 6) of the flue gas streams (inlet, exit from
the spray dryer, and stack).
Although much of the data still remains to be analyzed, preliminary
findings indicate that all three goals were achieved. In addition, a range
of removals was achieved, which has allowed a comparison of the test results
with data obtained using the nominal coal of 3-3.5% sulfur. A plot of
10-215
-------
percent sulfur removal as a function of external stoichiometry is given in
Figure 6 and shows that, while 90% removal was attained for the 3.5% sulfur
coal at a stoichiometry of about 1.15, a stoichiometry of approximately 1.3
was required for the same removal when treating flue gas from a 4.5% sulfur
coal. It is important to point out that the stoichiometries being quoted
are "external" stoichiometries and reflect the amounts of fresh pebble lime
entering the system. From a more fundamental standpoint, it is important to
examine the internal stoichiometry as a function of system performance.
This is due to the fact that the internal stoichiometry also includes the
available alkali in the recycle powder. It is this combined alkali which
actually drives the S02 removal reaction in the spray dryer, and it should
be taken into account when analyzing system operation. Based on data
analyzed to date, Argonne has determined that the internal stoichiometry is
(depending upon system operation) 15-20% greater than the external
stoichiometry-
Frequent process stream samples were taken and analyzed during the
testing period. For a set of samples taken on April 8, the compositions are
shown in Table 4. At that time, the average S02 removal was 79.6%, with a
temperature drop in the spray dryer of 195°F, and an approach (to dewpoint)
temperature of 22°F. Examination of the results indicates that the ratio of
calcium sulfite to calcium sulfate is on the order of 9 or 10 to 1. This is
significantly different from wet lime scrubbing systems where ratios of
approximately 3 or 4 to 1 are typical. It is felt that this difference
exists due to the relatively short time (2-4 seconds) that a liquid droplet
actually exists in a spray dryer. These short droplet lifetimes, in a dry
scrubber, do not allow the oxidation reaction to proceed to the extent that
it does in a wet scrubber, where droplet lifetimes are an order of magnitude
longer.
The average process flows are shown in Table 5. These flows, together
with the compositions in Table 4, allow acceptable material balances to be
developed for system components. For example, taking a CaO material balance
across the milk of lime tank we have:
a) In: (860.2 Ib/h) (0.946 available alkali) = 813.75 lb Ca°
n
b) Out: (8.7 gpm) (1.09 s.g.) [0.229 fraction Ca(OH)2] (500
lb Ca(OH)?
= 1085.8 - L
= 821.7
h
Ib CaO
10-216
-------
100
D
> 90
E
80-
(S)
~c
0)
o
0)
Q_
70-
60
3.5% Sulfur
4.5% Sulfur
0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6
Stoichiometry (CA/S Ratio)
Figure 6. Percent Sulfur in Coal Affects Performance
TABLE 4. PROCESS STREAM COMPOSITIONS
Percent by Weight
Component
Ca(OH)2
CaS03 . -j H20
1
CaC03
H20
Inerts
Feed
Slurry
14.14
9.60
0.23
1.04
71.70
3.29
100
Recycle
P owde r
11.50
60.87
6.71
5.98
2.46
12.48
100
Spray
Powder
11.89
54.42
5.46
8.46
3.25
16.52
100
Milk
of Lime
22.9
-
-
77.1
-
100
10-217
-------
TABLE 5. AVERAGE PROCESS STREAM FLOWS
Pebble Limea, lb/h 860.2
Milk of Lime to Mix Tank, gpmb 8.7
Slurry to Spray Dryer, gpmc 14.77
Dilution Water to Mix Tank, gpm 4.6
Recycle Powder to Mix Tank, lb/h 1904
Spray Dryer Bottoms Rate, lb/hd 979
Waste to Disposal, lb/he 3012
Stack Gas Flow, dscfmf 38,100
S02 in, lb/h 872
S02 out, lb/h 178
a@ 94.6% available alkali
b+8% level change in milk of lime storage tank
over 24 hour period = 59 gallons
cNo change in slurry tank level
Based on 2.5 minute sample collected
elncludes 1.3 ft drop in silo level = 14,294
Ib or 596 lb/h. Net disposal = 1814 lb/h
based on a settled density of 70 Ib/ft and a
solids moisture in the truck of 25%.
Dry standard cubic feet per minute.
c) Accumulation = -52-Sgl (1.09) (8.34 Ib/gal) (0.229)
Ib Ca(OH)7
= 5.12 _ 2- = 3.9
h
d) In = Out plus Accumulation
813.75 = 821.7 + 3.9 = 825.6
10-218
-------
e) Checking for error in the balance,
825«6~813.75 n m / •> / »
= - , - = 0.014, or 1.4%
. o
Examining the operation of the system, it was determined that the same
fractional removal of S02 is generally achieved across both the spray dryer
and baghouse. For example, if a 50% removal of SC>2 is detected based on
inlet and outlet concentrations at the spray dryer, then a 50% removal of
SC>2 will be seen across the fabric filter, based on the inlet concentration
to the filter.
Therefore, for the overall removal of SC^ of 79.6% on the 8th of April,
approximately 55% removal of S02 occurred in the spray dryer. Using this
value in a balance around the dryer, it was found that for the lime reacting
(9.79 Ib moles /h), the products were:
CaS03 . y H20 = 66%
CaC03 = 24%
CaS04 • y H2° = 10%
which leads to the conclusion that carbon dioxide (C02) pickup does occur in
the spray dryer.
Performing a similar balance over the baghouse show that:
CaS03 . y H20 = 63.98%
CaS04 . y H20 = 7.47%
CaC03 = 7.23%
Ca(OH)2 = 4.66%
H20 = 2.78%
Inerts = 14.17%
This gives the approximate composition of the powder exiting the fabric
filter. It is clear that the system was not at steady state, since the
combined compositions of the spray dryer powder and the baghouse
10-219
-------
(calculated) powder, does not equal the composition of the recycle powder.
The attainment of steady state is complicated by the residence time of waste
product in the recycle silo. This storage volume results in powder not
appearing for recycle until several days following production (assuming plug
flow in the silo).
PERFORMANCE MODEL
Argonne has identified what are believed to be the critical parameters
for S02 removal in a spray dryer. These are:
(Available Alkali] = /gRv
^ V Inlet S°2 /SDA "
This is the stoichiometric ratio (SR) based on the slurry fed to the
spray dryer absorber (SDA).
, v | Mass Available Alkali\ _ (•o-o)
I Total Mass of Solids )
\ / o JJA.
This recycle ratio (RR) of available alkali injected into the spray
dryer to the total mass of injected solids reflects the effect of spent
sorbent recycle on system performance. The available alkali is determined
by a chemical method (titration) that will detect the alkali whether it is
on the surface of the particle or in the interior. If recycling of spent
sorbent results in a partial "blinding" of available alkali by other
compounds in the recycled powder, then the stoichiometry needed for a
particular removal will be higher for a system with recycle than for one in
which the operation is on a "straight-through" basis. An analysis of data
and samples gathered in this program will allow for a determination of the
significance of sorbent recycle on system operation.
c) [1 - exP(-t/t)]
This factor is a reflection of the effects of boiler load (and hence
gas residence time in the spray dryer) as well as the drying phenomenon for
particles in the spray dryer. The variable, t, represents the gas residence
time in the spray dryer, as an average of inlet and outlet gas volumes. The
time constant, T, is the time for drying an average sized particle in the
spray dryer. This average drying time is thought to be a function of
several factors including the percent solids in the slurry, the gas
temperature drop in the spray dryer, and the average droplet size.
T - T \ T
, N I Gas in Gas out 1 1
d) 1 ^ —
Gas in Gas Sat'd
10-220
-------
Many researchers have noted that the closer the exit gas temperature is
to the dew point, or saturation temperature, then the lower the stoichio-
metry for a given removal. If the S02 removal reaction takes place within
the droplets as they dry, then the temperature gradient across the spray
dryer should play a key role. The numerator of the factor is the tempera-
ture drop of the flue gas across the spray dryer due to the evaporation of
the slurry. The denominator is the maximum temperature drop that could be
taken in the spray dryer before a wet product is produced. Another way of
viewing this factor is that it is the ratio between the operation of a dry
scrubber and a wet scrubber, where the gas is quenched to its dew point.
Combining the terms;
Fraction S02 removal = (SR)a (RR) — I (1 - exp (-t/T))
where a, b, and c are exponents which may, or may not, be equal to 1. Data
obtained in the forthcoming performance tests on the FGC system will be used
to check and calibrate this relationship.
REFERENCES
1. Federal Register. 4_5_(98) :33127-33128 (May 19, 1980).
2. Farber, P.S., Startup and Performance of a High Sulfur Dry Scrubber
System, Paper 82-40.5, Proc. 75th Annual Air Pollution Control
Association Meeting, New Orleans (June 1982).
3. Characteristics of Waste Products From Dry Scrubbing Systems, prepared
by Radian Corporation for the Electric Power Research Institute, Report
CS-2766 (Dec. 1982).
4. Farber, P.S., C.D. Livengood, and J.L. Anderson, Leachate of Dry
Scrubber Wastes, Paper 83-29.1, Proc. 76th Annual Air Pollution Control
Association Meeting, Atlanta (June 1983).
5. Bonk, L.A., Argonne National Laboratory, Unpublished Information
(1982).
6. Knight, M., Argonne National Laboratory, Personal Communication (Feb.
1983).
7. Immobilization and Leachability of Hazardous Wastes, Environmental
Science and Technology 16(4): 219a-223a (1982).
10-221
-------
NOTES
1. Company Names and Products.
The mention of company names or products is not to be considered an
endorsement or recommendation for use by the U.S. Environmental
Protection Agency.
2. Units of Measure.
EPA policy is to express all measurements in Agency documents in metric
units. When implementing this practice will result in undue cost or
difficulty in clarity, IERL-RTP provides conversion factors for the non-
metric units. Generally, this paper uses British units of measure.
The following equivalents can be used for conversion to the Metric
system:
British
5/9 (°F-32)
1 Btu/kWh
1 ft 0.3048 m
1 ft2
I ft3
1 ft3/min
1 gallon per minute (gpm)
1 gallon (gal)
1 grain
1 horsepower (hp)
1 in. W.C.
1 Ib (avoir.)
1 lb/106 Btu
1 lb/ft3
1 ton (long)
1 ton (short)
Metric
1055.056 J/kWh
0.0929 m2
0.0283 m3
4.719E-04 m3/s
6.309 E-05 m3/s
0.003785 m3
0.0648 gram
746 watts
249.08 Pascals
0.4536 kg
429.6 ng/J
16.01 Kg/m3
1.0160 m tons
0.9072 m tons
10-222
-------
UNPRESENTED PAPERS
-------
AN ECONOMIC EVALUATION OF LIMESTONE DOUBLE ALKALI
FLUE GAS DESULFURIZATION SYSTEMS
G. A. Hollinden, C. D. Stephenson, J. G. Stensland
-------
AN ECONOMIC EVALUATION OF LIMESTONE DOUBLE
ALKALI FLUE GAS DESULFURIZATION SYSTEMS
by: Gerald A. Hollinden, Ph.D.
Tennessee Valley Authority
Chattanooga, Tennessee
C. David Stephenson
Tennessee Valley Authority
Muscle Shoals, Alabama
John G. Stensland
FMC Corporation
Schaumburg, Illinois
ABSTRACT
Considerable work was done at the EPA Scholz plant facility in defining
the process parameters for limestone double alkali flue gas desulfurization
systems. In general this study work proved the viability of the process but
uncovered several less than optimum operating parameters that needed further
work. FMC and others have continued to work with the process and have
defined operating parameter changes necessary to make the system
commercially viable.
Limestone double alkali is especially appropriate for FGD systems
applied to boilers burning relatively high sulfur (2% and greater) fuel. A
discussion of site-specific design criteria which impact on the selection of
FGD technology is included with a definition of the optimum parameters for
the applications of limestone double alkali.
An in depth economic analysis of the system is included with comparisons
to conventional limestone scrubbing technology. Cost comparisons are made
by subsystems such as absorber system, reagent handling, storage and
preparation system, solids waste production and disposal. The economic data
presented is primarily a result of study work done by TVA under contract
from the Environmental Protection Agency.
11-1
-------
DISCLAIMER
This paper was prepared by the Tennessee Valley Authority (TVA).
Neither TVA nor any person acting on its behalf:
a. makes any warranty or representation, express or implied, with
respect to the use of any information contained in this paper; or
that the use of any information, apparatus, method, or process
disclosed in this paper may not infringe privately owned rights; or
b. assumes any liabilities with respect to the use of, or for damages
resulting from the use of, any information, apparatus, method, or
process disclosed in this paper.
This paper does not necessarily reflect the views and policies of TVA.
11-2
-------
I. INTRODUCTION
For the past five years there have been extensive efforts to develop a
sodium-based double alkali process in which limestone, rather than lime, is
used to regenerate the absorbent, thus retaining the advantages of using a
solution as the absorbent liquid while avoiding the higher operating costs
inherent with the use of lime. FMC Corporation, Combustion Equipment
Associates, Inc. (now Thyssen - CEA Environmental Systems, Inc.—TESl),
A. D. Little, Inc., Accurex Corporation, and the U.S. Environmental
Protection Agency have been active in the development of a limestone double
alkali process. These efforts have advanced the technology of the process
to the point that a commercially acceptable limestone double alkali process
seems assured.
Among the studies of limestone double alkali processes in the past few
years, there has been an EPA-sponsored prototype scale evaluation at the
Gulf Power Company's Scholz steam plant, an economic evaluation by the
Tennessee Valley Authority, and an economic evaluation by the Stearns-Roger
Engineering Corporation. This paper summarizes these three evaluations and
discusses studies and test work by the FMC Corporation to develop a
limestone double alkali process.
II. TESI PLANT SCHOLZ WORK *
Thyssen-CEA Environmental Systems, in conjunction with Arthur D. Little,
conducted an EPA-sponsored test program at Gulf Power Company's Scholz steam
plant near Sneads, Florida. This involved the conversion of a 20-megawatt
lime based dual alkali system to operation with limestone regeneration.
Budgetary problems caused a reduction in the test program from six
months to two months. However, even during the shortened program the
process proved to be technically feasible.
A. SYSTEM DESIGN BASIS
The limestone dual alkali system at Scholz was based on firing the
boiler with coal containing 2.9 to 3.4 weight percent sulfur. These sulfur
levels corresponded to approximately 1900 to 2300 ppm S02 in the boiler
flue gas. Since it was desirable to run tests at the high end of this inlet
loading and to include even higher S02 inlet concentrations, an S02
injection system was provided. This allowed operation with both a normal
and maximum load condition with regard to S02» Normal system operation
was based on a gas flow of 40,000 dscfm with an inlet S02 concentration of
2200 ppm dry. The maximum condition was based on a 45,000 dscfm gas flow
with an S02 concentration of 2650 ppm dry. Table 1 and Table 2 give
additional details for the normal and maximum design cases.
Valencia, J.A., et. al., Evaluation of the Limestone Dual Alkali
Prototype System at Plant Scholz, Report to EPA, Contract 68-02-3128,
August 1981. (The majority of the data contained in the section is
taken from this report.)
11-3
-------
TABLE 1. DESIGN BASIS - NORMAL OPERATION
Inlet Gas:
Flow Rate
S02
Removal Efficiency:
S02
Absorber/Scrubber Feed:
Absorber Tray Feed
Scrubber Recycle
Soda Ash and Limestone Feed:
Soda Ash Feed Rate
Limestone Purity
Limestone Feed Rate
Waste Solids:
Wash Ratio
Insoluble Solids
40,000 scfm (dry)
14.9 Ibs/min (2,200 ppm dry)
95%
2.4 gallons/1000 acf
16 gallons/1000 acf
0.07 mole Na+/mole S02
95 wt.%
0.986 mole available CaC03/mole S02
4 displacement washes
55 wt.%
TABLE 2. DESIGN BASIS - MAXIMUM LOAD OPERATION
Inlet Gas:
Flow Rate
S02
Removal Efficiency:
S02
Absorber/Scrubber Feed:
Absorber Tray Feed
Scrubber Recycle
Soda Ash and Limestone Feed:
Soda Ash Feed Rate
Limestone Purity
Limestone Feed Rate
Waste Solids:
Wash Ratio
Insoluble Solids
45,000 scfm (dry)
20.2 Ibs/min (2,650 ppm dry)
95%
2.6 gallons/1000 acf
15 gallons/1000 acf
0.081 mole Na+/mole S02
90 wt.%
1.010 mole available CaC03/mole S02
4 displacement washes
55 wt.%
11-4
-------
B. DESCRIPTION OF THE SCHOLZ SYSTEM
The test plant facilities at Scholz consisted of three separate areas:
the boiler, the limestone dual alkali scrubbing system, and the waste
disposal facilities, that were provided as part of a parallel EPRI sponsored
program to evaluate landfill disposal of the waste cake generated. The
waste disposal facilities will not be discussed in this paper.
1. Boiler
The dual alkali scrubbing system was installed on Unit No. 1 at Scholz
steam plant, a 40 MW capacity (47 MW peak capacity) B&W pulverized coal
fired boiler. An electrostatic precipitator designed for 99.5% particulate
removal follows the boiler. A portion of the flue gas, equivalent to a 20-
MW boiler load, was directed to the limestone dual alkali system. The
remaining gas was exhausted to the main stack.
2. Limestone Dual Alkali System
During 1975 and 1976 the Southern Company and EPA sponsored a jointly
funded test program, conducted by TESI, utilizing a lime dual alkali system
installed at Plant Scholz. The limestone dual alkali system utilized during
the TESI testwork was a modification of that plant. The modified dual
alkali system consisted of four basic sections: absorption; regeneration;
waste solids dewatering; and reagent storage and preparation. A process
flow diagram is shown in Figure 1.
a. Absorption—
The absorption system consisted of a plumb-bob type venturi scrubber
followed by an absorption tower. The variable throat venturi was designed
for both particulate removal and S02 absorption. Typically, a utility
type dual alkali system would take flue gas from a particulate removal
device (electrostatic precipitator or fabric filter) and would not require a
venturi scrubber. However, rather than removing the venturi, it was used to
quench and saturate the flue gases.
The absorption tower was designed to be used either as a spray tower or
a tray tower. The limestone dual alkali testing was conducted utilizing the
tower with two trays.
b. Regeneration Section—
Spent recirculated scrubbing solution was bled from the absorber to a
multi-stage reactor system. The reactor system consisted of five separate
tanks—an existing 750 gallon primary reactor followed by four new 3400
gallon tanks in series to allow a total of approximately 100 minutes
retention time. All tanks were cylindrical, baffled vessels with center
mounted agitators. Tests were run with all five reactors in series with dry
pulverized limestone fed to the first reactor.
11-5
-------
CLEANED GAS
LIMESTONE
SILO
WATER
NORMAL FLOW-
ALTERNATE FLOW
Figure 1 Plant Scholz Limestone Dual Alkali System Flow Diagram
-------
c. Waste Solids Dewatering—
An existing 93,000 gallon thickener was utilized for solids settling.
Slurry from the reactor system was pumped to the tank with clear liquor
overflow collected in an existing thickener hold tank. This tank provided
surge capacity for the regenerated return liquor feed to the scrubbing
system. Slurry underflow was recirculated in a loop around the thickener to
prevent solids settling and plugging of these lines. A bleed from this
recirculation loop was fed to the existing rotary drum vacuum filter.
Filtrate was returned to the thickener. Filter cake was washed on the drum
using two wash water spray banks. Solids were discharged to a weigh belt
conveyor for handling in the waste processing system.
d. Reagent Storage and Preparation—
Ground limestone (97-99.9% less than 325 mesh) was utilized as the
regenerative reagent. It was received, stored and fed to the system in a
dry form even though the capability existed for feeding it as a slurry.
Soda ash solution was utlized to make up for sodium losses in the system.
It was normally added to the thickener hold tank.
C. PROCESS PERFORMANCE
As stated earlier, the original intent of the test program was to
evaluate several aspects of limestone dual alkali process performance. As
the test program had to be reduced to only a two month period, a thorough
and complete evaluation of the technology was not possible. However,
process performance was evaluated for each of the following conditions:
* SO2 removal capability
* Reagent consumption
* Power usage
* Waste product properties
* Reliability and ease of operation
1. SO? Removal
During the test period, inlet S02 concentrations ranged from 1460 to
3240 ppm (dry). Average removal efficiencies ranged from 93.5% in January
to 96.7% in March.
Scrubbing liquor pH was the most important variable affecting S02
removal efficiency. At pH's in excess of 5.7, outlet S02 concentrations
below 100 ppm were obtained. When operating at a pH of 6.0, outlet
concentrations of less than 50 ppm were obtained. It was determined that
operating the absorber at a bleed pH of 5.7 to 6.0 provided the best
compromise between high S02 removal efficiencies and adequate regeneration
of the scrubbing liquor in the reactors.
11-7
-------
It was also determined that the ratio of bisulfite to total oxidizable
sulfur compounds (TOS) concentrations in the liquor could be utilized as a
guideline for S09 removal. The ratio of bisulfite to TOS during the test
period typically ranged between 0.60 and 0.75, and this ratio produced
efficiencies between 94% and 98%.
Other variables such as pressure drop, inlet S02, and active sodium
concentration had a much smaller effect on performance than pH. Their
effect appeared to be secondary in nature.
2. Limestone Utilization
Limestone utilization was very good during the test period. Utilization
in the reactor train effluent was between 85% and 95% of available Ca(X>3
in the raw limestone. Final system utilizations ranged from 93% to 100%
averaging 97.5%. Fredonia limestone from Kentucky was used during the
start-up and initial break-in period and Sylacauga limestone from Alabama
was used during the end of the break-in period and the entire testing
period. Table 3 gives the characteristics of these limestones. An attempt
was made to determine the progress of limestone utilization through the five
vessel reactor train. Utilization in the first reactor ranged from 23% with
a level of 104 ppm solids carryover from the thickener to 63% with 1910 ppm
carryover. The test report hypothesized that this was possibly due to the
solids entering the first reactor acting as crystals to facilitate an
increase in the rate of precipitation of calcium sulfur salts. This, it was
felt, would promote an increase in the rate at which calcium was dissolved
and reacted with sodium bisulfite. It was recommended that further studies
be done to verify this hypothesis. Testing indicated that regardless of the
extent of reaction achieved in the first reactor, efficient limestone
consumption in the remainder of the train and dewatering system would occur.
3. Waste Cake Properties
a. Settling Characteristics—
The most significant process limitation encountered at Scholz was the
generation of solids with good settling characteristics. During periods of
the test program, solids with excellent settling characteristics were
generated. However, in other periods inconsistent and poor quality solids
also occured. In the months of December and February, solids that settled
out to 10% of the initial slurry volume in six to eight minutes were
generated. In January poor solids were generated, which took hours to
settle to the same consistancy. It was determined that a significant
contributor to poor solids quality was the carryover of fine solids in the
thickener overflow. These solids, which were ultimately fed to the reactor
after passing through the absorber, promoted the formation of finer and more
difficult to settle solids in the reactor train. Laboratory work seemed to
confirm this hypothesis indicating that poor settling behavior was due in
large part to a high percentage of fine needle shaped solids present during
operation with high solids carryover from the thickener. It was determined
that solids carryover of 1000 ppm in the thickener overflow could easily be
handled by the system without detrimental effects to the settling
properties. However, carryover in excess of 5000 ppm resulted in rapid
deterioration of solids quality.
-------
TABLE 3. LIMESTONE PROPERTIES
Fredonia
Mine Location
Type
Particle Size Distribution
Wet Sieve Analysis (wt.%)
60 mesh (250 )
100 mesh (149 )
325 mesh (44 )
400 mesh (38 )
Bulk Density (lbs/ft3)
Chemical Analysis
Alkalinity (wt.% as
Ca (wt.%) as CaC03>
Mg (wt.% as MgC03)
Fe (wt.% as Fe203>
Sy lacauga
Fredonia, Kentucky Sylacauga, Alabama
High-Calcium Calcite High-Calcium Calcite
99.7
99.4
96.4
94.4
55-60
97.9
93.6
4.2
0.05
99.9
99.8
97.7
93.9
60-65
97.8
95.7
1.3
0.04
It was further determined that waste solids deteriorated in their
settling properties with time. On various occasions during filter downtime,
solids in the thickener began resuspending as underflow was continuously
recycled to the feed well. It was determined that it was possible that
agglomerated, quick settling solids had begun to dissolve and recrystalize
into poor settling solids.
b. Filter Cake Characteristics—
Limitations in equipment size and poor mechanical performance of the
filter contributed significantly to two major shortcomings in filtering
operations. These were the inability to produce filter cake with 55% or
higher insoluble solids and the inability to wash the cake to reduce sodium
losses to 4% of the insoluble solids content of the cake.
Leaks in the internal piping of the filter drum caused substantial
losses of vacuum. This, coupled with the inability of the thickener
underflow pumps to handle slurries containing more than 15% solids and the
attendant dilution requirement, resulted in the filter cake having insoluble
solids contents ranging between 35% and 45%.
High sodium losses in the cake were encountered throughout the test
program. The system was designed with two spray banks to wash the cake
utilizing up to 40 gpm of wash water. This represented the equivalent of
displacing four times the final volume of liquid in a 55% solids filter
cake. Corresponding sodium losses were anticipated to be approximately 4%
of the insoluble solids content of the cake (equivalent to a Na/Ca ratio of
0.08 in the final cake). However, physical limitations in the handling of
the filtrate, limitations in the wash water supply, and operating require-
11-9
-------
ments of the waste disposal system precluded the use of more than two
displacement washes. In February the average number of displacement washes
was 1.5 resulting in a high Na/Ca ratio of 0.2, and in March the average
number of displacement washes was only 0.8 with a correspondingly higher
Na/Ca ratio of 0.4. However, there were times when up to three displacement
washes were utilized resulting in sodium losses of as low as 0.03 to 0.04.
It was concluded that the problems of low insoluble solids content cake
and high sodium losses in the cake were not inherent to the limestone dual
alkali technology but were rather caused by mechanical conditions of the
equipment used and operating requirements specific to the test facility.
4. Soda Ash Consumption
Actual sodium losses from the system at Plant Scholz were much higher
than those anticipated. This was due not only to excessive loss in the
filter cake as described above but also to leaks, spills and liquor purges
needed to maintain volume balances. During the test period soda ash feed
amounted to 0.29 moles of Na2C03/mole of SC>2 removed. As much as half
of this amount was needed for cake losses with the remainder associated with
other liquor losses. It was felt that sodium consumption could be held
within design limits of 0.04 moles of Na2CC>3/mole of SC>2 removed on a
production system with little difficulty.
5. Power Consumption
Most of the power consumption at Scholz was associated with the system
forced draft fan. The venturi required pressure drop 2-3 times that of the
absorber. Thus, in a typical limestone dual alkali scrubbing system where
particulate removal is performed by either an electrostatic precipitator or
a fabric filter, the expected power consumption would be much lower than
that encountered at Scholz. Power consumption ranged from 2.5% (0.53 MW) at
flue gas flow rates equivalent to a 21 MW boiler load, to 5.3% (0.42 MW) at
an equivalent load of 8 MW. If the power requirement associated with the
venturi is not included, a requirement of 1 to 1-1/2% of power generated at
full boiler load was anticipated.
6. Process Operability
The system at Scholz was reported to have good process operability after
the first two to three days of stable operation. The major problems with
process operability were encountered during the initial days following any
restart of the system. This was principally due to the inconsistent
settling characteristics of solids generated. In some cases the solids
settling rate would level off and stable operation would be achieved. At
other times settling rates would continue to deteriorate resulting in
significant solids carryover in the thickener overflow which further
exacerbated the problem. It was hypothesized that the restart problems were
caused by redissolution of solids that had been left in the thickener at the
time of the outage and subsequently recrystalized into fine crystals with
poorer settling characteristics.
11-10
-------
Once the system reached stable operaton, the process operability was
considered to be very good. Variations of inlet SC>2 concentrations of as
much as 500 ppm were handled easily by adjustment of the feed forward rate
of regenerated solution to the absorber. Variations in boiler load were
handled in the same manner. In addition, upsets including carryover of fly
ash in the flue gas due to precipitator malfunction, overfeeding of
limestone due to operator oversite, and occasional limestone and soda ash
feed outages were handled well by the system.
Some solids deposition was encountered in various portions of the
regeneration system. Some scale buildup was encountered on the walls of the
reactor and in the overflow pipe connecting the first and second reactors.
Scale buildup in the other reactors was minimal, and it was concluded that a
semi-annual cleaning operation as part of a regular maintenance program
might be adequate to control any reactor scaling.
Conclusions reached as a result of the Scholz plant testing are as
follows:
" A limestone double alkali process is technically feasible.
" Additional refinement and further testing of the system was
necessary to develop process information required for commercial
operation.
* Laboratory or small pilot plant tests should be conducted to better
understand the generation of solids with good settling
characteristic s.
Ill FMC TESTWORK
Since 1977 FMC has pursued the use of limestone for regeneration in its
double alkali process. They felt that the performance claims made for the
process had been validated by the actual operation of three utility double
alkali systems furnished by three different suppliers. The operating
benefits of the double alkali technology have been demonstrated but overall
evaluated costs continued to be a concern.
In 1981 FMC determined that the growing differential in the price of
lime and limestone caused many evaluators to judge lime double alkali as
excessively expensive when compared with conventional limestone scrubbing
systems. This caused a renewed emphasis on developmental work for a
limestone regenerated double alkali process. FMC's goal for a limestone
double alkali process extended beyond technical feasibility. It was felt
that any system developed must realize the cost benefit of using limestone
while maintaining the advantages associated with lime double alkali
scrubbing, and without increasing other cost components significantly. A
program was developed to test limestone double alkali, with emphasis on
three general operating parameters: reaction rates, solids quality, and
absorber performance. Beginning in September 1981, bench scale experimental
work was initiated on limestone kinetics at FMC's Central Engineering
Laboratories in San Jose, California, and at Illinois Institute of
11-11
-------
Technology Research Institute. In early 1982 the next phase of the pilot
plant work was started at their Princeton Research Center in Princeton, New
Jersey. The intent of the program was to define system operating ranges
that allowed high utilization of limestone and commercially feasible
reaction times while maintaining good solids quality and high levels of
SO2 removal.
FMC determined that the primary technical issues concerned the overall
regeneration reaction. Reaction rate characteristics and the parameters
that affect it were well understood from previous work done by FMC and
others. It was determined that reaction rate characteristics put
constraints on the process in the sense that high utilization of limestone
and commercially feasible reaction times could only be accomplished with
certain combinations of conditions.
It was also found that solids quality is a function of the size and
shape of individual solid particles. As was seen at the Scholz test plant,
these parameters were in turn a function of the chemical and physical
conditions under which the particles were formed.
FMC also noted that optimum absorption of sulfur dioxide placed certain
constraints on the composition of the absorbing solution, but that these
conditions are not necessariliy compatible with optimum results in the
regeneration reaction and the solids quality. A plan was developed to
optimize the various competing factors.
A. REACTION RATES
The rate of limestone reaction and the degree to which it proceeds to
completion were found to be dependent upon several parameters. First, the
dissolution of limestone creates carbonate and bicarbonate ions which place
an upper limit on the pH in the regeneration reaction. It is not possible
to regenerate all of the sodium bisulfite without using excessive amounts of
limestone. This general constraint of limited capacity for bisulfite
regeneration lead to the definition of several variables that were used for
defining the feasible ranges of process chemistry.
The first of these is the regeneration fraction (shown as "Y" in Figure
2). In general it was found to be difficult to achieve limestone
utilizations in excess of 90% at regeneration fractions in excess of 0.6.
That is, if more than 60% of the bisulfite was regenerated limestone
consumption became excessive.
Another parameter defined was the ratio of the active sodium
concentration to the change in bisulfite concentration in the regeneration
portion of the process (shown as "Z" in Figure 2). For a given absolute
amount of regeneration, the ratio is an expression of the active sodium
level in the system. A high ratio corresponds to a very concentrated system
and is^a practical limitation based on the total dissolved solids in the
scrubbing solution. A low ratio corresponds to relatively dilute solutions
which lack the buffering capacity FMC normally attempts to achieve in their
lime double alkali process. It was determined that scrubber pH,
11-12
-------
regeneration pH, change in bisulfite concentration, fractional regeneration,
and ratio of active sodium concentration to the change in bisulfite
concentration were all interrelated such that classifying any three of these
parameters fixed the other two. These constraints, plus additional ones
discussed below, set the optimal operating range defined by FMC for their
limestone double alkali process.
Several other factors which also affect the regeneration reaction rate
and limestone utilization were identified. These factors and their effects
were as follows:
1. Limestone Particle Size
It was found that, as with direct limestone scrubbing, limestone
particle size is directly related to utilization. Larger particles tended
to become coated with the calcium sulfite reaction product rendering a
portion of the limestone unavailable for reaction. Grinding the limestone
to minus 200 mesh was determined to be essential for good utilization, and
the relative small incremental costs for grinding to minus 325 mesh was
deemed generally cost justified.
2. Limestone Reactivity
It was found that limestone varied considerably in its microstructure
with a wide variety of both amorphous and crystalline characteristics.
Essentially, reactivity is a surface area phenomenon with a limestone having
a high surface to mass ratio being the most reactive. Therefore, it was
determined that reactivity of the limestone was not necessarily defined by
the mesh size.
The principal effect of limestone reactivity was determined to be on the
regeneration reaction residence time requirements.
B. SOLIDS QUALITY
FMC found that solids quality is a function of particle size and shape,
which are in turn functions of the chemical and physical conditions under
which the particles are generated. As a practical matter, the chemical
conditions are dictated largely by the kinetic and material balance
constraints. Poor solids quality were generally the result of excessive
nucleation, or formation of new particles of reaction products. In this
sense, the goal of a fast regeneration reaction rate is in direct conflict
with a goal of crystal growth. A reasonable compromise in this regard was
accomplished with the proper design strategy in the reactor area. First,
multiple reactors in series were utilized with split limestone feed so that
the degree of reaction was controlled. Also, since the first reaction step
is the most likely point to experience excessive nucleation, some solids
from the thickener underflow were recycled to the first reactor, and the
scrubber bleed pH kept relatively high. The former furnished seed crystals
which promote particle growth instead of nucleation and the latter minimizes
the driving force which also favors crystal growth.
11-13
-------
C. ABSORBER PERFORMANCE
FMC found that the overall system material balance put constraints on
the system chemistry. First, sulfur dioxide absorption began to decline
dramatically for absorber pH's below 6.0. Very high collection efficiency
typically requires a pH at the top of the absorber of 6.3 or higher. In
contrast to FMC's lime double alkali system, however, a low pH scrubber
bleed stream was necessary in order to achieve good limestone utilization.
This meant that the regenerated scrubbing solution had to be fed to the top
of the absorber rather than blended with the overall recirculating scrubbing
liquor as is done in their lime based process. The result of this
constraint was that a spray tower or packed tower proved to be the most
effective absorber for a limestone double alkali process. The packed tower
was deemed the most effective device because it can operate at very low
liquid to gas ratios and the regenerated solution flow rate is typically
equal to the total absorber flow rate requirement thus eliminating all
recirculation requirements to the absorber. It was determined that a spray
tower could also be utilized with regenerated solution used directly in the
top stage with lower stage(s) supplied with recirculated liquor from an
absorber sump.
The other important material balance constraint was found to be the
practical limitation on the absolute regeneration level. The total flow
rate required to regeneration is inversely proportional to the changes in
bisulfite concentration achieved. For example, if the bisulfite change is
0.1 M, the flow rate required to regeneration is twice that required for a
typical lime based system using a bisulfite concentration change of 0.2 M.
This means that the limestone double alkali process must operate at higher
active sodium concentrations in order to have regeneration flow rates equal
to lime double alkali.
D. FMC PROCESS DESIGN
From their test and process design work, FMC developed an operating
envelope that allows simultaneous achievement of all of the desired
objectives. This envelope is summarized in Figure 2. The boundaries of the
operating envelope are sumarized as follows:
1. Scrubber bleed pH equal to 5.8 - 6.6. It was found that if the absorber
bleed pH was too low, the reaction rate would improve; but this was offset
by an increased tendency to dissolve magnesium and a requirement for
two-loop absorption. If the bleed pH was too high, only limited bisulfite
regeneration per pass was achieved and a higher active sodium level was
required.
2. Regeneration return pH equal to 6.3 - 7.0. It was determined that
regeneration return pH must be high enough to allow good S02 absorption
and adequate bisulfite regeneration per pass but not so high as to
significantly inhibit limestone utilization.
11-14
-------
I
Q.
I
a
z
DC
D
h-
LLI
DC
•R
CC
LU
1 °R
LU
DC
Y = BISULFITE REGEN
FRACTION
Z RELATIVE SYSTEM
CONCENTRATION
D
B
BLEED pH (pH)
B
Figure 2 Limestone Double Alkali Design Envelope
11-15
-------
3. Regeneration fraction (the percentage of the bisulfite in the scrubbing
bleed solution that is regenerated in the regeneration section): FMC feels
that this fraction should be in the range of 35 to 70% and preferably
between 35 and 60%. Higher regeneration fractions led to poor limestone
utilization, and lower regeneration fractions required excessive flow rates
to regeneration.
E. DEMONSTRATION PLANT
FMC is currently in the process of installing a limestone double alkali
demonstration plant at Northern Indiana Public Service Company Schahfer
Station. The plant will be a 3 MWe (9,000 acfm) system installed on a slip
stream from the existing Schahfer Unit 17 lime double alkali FGD system.
The Schahfer Unit is designed to burn coal to 3.6% sulfur. The objective of
the demonstration plant program is to confirm the limestone double alkali
test results that have been obtained in the past by FMC and at the Scholz
Plant.
The program schedule calls for the completion of erection by early
January 1984. Testing is divided into three parts: 1) shakedown testing,
2) baseline testing, and 3) stress testing.
Shakedown testing will involve the characterization of the
responsiveness of the system. The results of this period will determine
those variables which most influence system operation and the appropriate
values to be selected for each of these variables to attain optimal system
operation. This data will be utilized as the basis for baseline testing
work. The aim of the baseline test period is to demonstrate the ability of
the limestone double alkali system to sustain performance over long periods
of time while responding to normal changes in boiler load and flue gas
conditions. Stress testing will involve establishing conditions outside the
normal expected to determine the response of the system to those conditions
and to determine corrective actions necessary to compensate for any adverse
effects.
FMC plans to have their demonstration plant program essentially complete
by the end of 1984.
IV ECONOMIC ANALYSIS
TVA recently completed an economic analysis of the limestone double
alkali system for the EPA. In addition, Steams-Roger has just finished a
study for EPRI which also reviews limestone double alkali economics. The
design bases and the economic procedures used in the studies differ
considerably and detailed point-by-point comparisons are difficult. The
major conditions and criteria used in each study are shown in Tables 15 and
16. In addition, the TVA limestone double alkali process design is based on
information provided by EPA derived primarily from the tests at the Scholz
steam plant described in this paper and on in-house information. The
Steams-Roger design is based largely on the process developed by FMC. In
general, both studies show that the limestone double alkali process is less
expensive in capital investment and operating costs than those of a forced-
11-16
-------
oxidation limestone FGD process under the conditions used in the evaluation.
These results are discussed below. In addition, an ecomonic comparision of
the TVA design and the FMC design, both based on TVA premises is discussed.
A. TVA/EPA STUDY*
The base case for this study utilized a new single 500 megawatt
pulverized coal fired boiler firing 3.5% sulfur eastern bituminous coal.
The heat rate utilized for the boiler was 9,500 Btu/kWh. The unit was
assumed to be installed during the construction of the power plant and
utilized a thirty year life with full load operation for 5,500 hours a year.
The system design included four 125 megawatt operating absorber trains.
A spare absorber train was included to allow utilization of an emergency
bypass. Hardware spares consisted of spare crushing and grinding equipment
in the limestone preparation area, a spare filter, and spare process pumps
in addition to the spare absorber train.
It was assumed that construction would begin in early 1981 and cover a
three-year span to late 1983. Capital investment requirements were based on
1982 costs, and annual revenue requirements were based on mid-1984 costs.
In the TVA study the economics of the limestone double alkali FGD system
were compared with the economics for a forced oxidation limestone scrubbing
process. The major absorber process design conditions utilized in the study
are shown in Table 4.
Capital Investment
The total capital investment for the limestone double alkali process was
$95 million ($190/kW). The comparable forced oxidized limestone system had
capital investment requirements of $103 million ($206/kW). The total direct
investment, representing installed equipment costs, was $47 million for
limestone double alkali and $57 million for the limestone process. A
breakdown of the capital investment requirements for the two systems are
shown in Table 5.
As can be seen, the largest cost differential between the systems is in
the SC>2 absorption area. Here the double alkali process has a cost
advantage of $9 million. While the absorbers in both processes are similar
in size, the higher L/G ratio required for the limestone system resulted in
a large cost penalty. The double alkali system required an L/G ratio of 3
with an L/G ratio of 106 required by the limestone process. This resulted
in a recirculation pump cost comparison of $1.7 million for limestone versus
$80,000 for the double alkali process. In the areas of raw material
handling and feed preparation, the systems showed similar cost requirements
and the regeneration area required for the limestone double alkali process
did not materially add to the process cost.
Stephenson, C. D., Burnett, T. A., and Torstrick, R. L., Economic
Evaluation of a Sodium/Limestone Double Alkali FGD Process, to be
published.
11-17
-------
TABLE 4. PROCESS DESIGN CONDITIONS
Process
Sodium/Lime stone Limestone
Double Alkali Scrubbing
Absorber type Sieve tray tower Spray Tower
Superficial Gas Velocity, ft/sec 9 10
L/G, gal/kaft3
Presaturator/Underspray 2/1 4/10
Absorber 5 106
Stoichiometry, mole Ca/mole (S02 + 2HC1) 1.0 1.4
Absorbed
Sulfite Oxidation, % 10 95
Thickener Feed Solids, % 1-4 8
Thickener Underflow Solids, % 25 40
Filter Cake Solids, % 55 85
The dewatering system required for limestone double alkali was evaluated
at 50% higher than that required for the limestone process. This was based
on the high-sulfite waste produced in the double alkali system. However,
when the costs of forced oxidation were included in the limestone process,
the total cost of preparing waste for disposal were similar for both
systems. When a credit was applied to the limestone double alkali process
for flyash disposal, the landfill disposal costs for that system were about
10% less.
The other difference in capital investment between the two processes
consisted of indirect investment and other capital investment, both of which
are based for the most part on direct capital investment. Exceptions to
limestone double alkali's generally lower investment cost were the
contingency fee and start-up and modification allowance assigned by TVA due
to the lower level of technical development of the process.
Annual Revenue Requirements
Tables 6 and 7 depict the annual revenue requirements assigned to the
double alkali and limestone processes respectively in the TVA study. The
limestone double alkali process showed a first year annual revenue
requirement of $26 million (9.28 mills/kWh) compared with $29 million (10.58
mills/kWh) for the conventional limestone process. Annual direct costs for
the double alkali process are shown as $9.2 million (constituting 36% of the
total first year annual revenue requirement) and for the limestone process
are shown at $10.6 million (also 36% of the total annual revenue
requirement s) .
11-18
-------
TABLE 5. CAPITAL INVESTMENT
Direct Investment
Materials Handling
Feed Preparation
Gas handling
SC>2 Absorption
Reheat
Regeneration
Oxidation
Solids Separation
Fixation
Total Process Capital
Services, Utilities and Miscellaneous
Total Direct Investment Excluding Landfill
Landfill Construction
Landfill Equipment
Landfill Credit (fly ash disposal)
Total Direct Investment
Indirect Investment
Engineering Design and Supervision
Architect and Engineering Contractor
Construction Expense
Contractor Fees
Contingency
Waste Disposal Indirect Investment
Total Fixed Investment
Other Capital Investment
Allowance For Startup and Modification
Interest During Construction
Royalties
Land
Working Capital
Total Capital Investment
Dollars per kW of Generation Capacity
Capital Investment, $
Double Alkali Limestone
2,426
4,506
10,800
11,348
3,630
1,506
0
,493
906
40,615
2,437
43,052
5,247
1,454
(2,312)
47,441
3,014
861
6,888
2,153
11,194
1,720
73,271
6,716
11,430
406
554
2,775
95,152
190.30
2,528
4,715
11,281
20,288
3,634
0
2,670
3,679
0
48,795
2,928
51,723
3,781
1,123
0
56,627
3,621
1,034
8,276
2,586
6,724
1.707
80,575
5,917
12,570
0
614
3,392
103,068
206.14
Basis: 500-MW new coal-fired power unit; 3.5% in coal; 89% SC>2 removal;
fixation and landfill solids disposal; north-central, new power unit
with 30-yr life at 5,500 hr/yr full-load operation; 1 spare absorber
train, and provisions for emergency bypass of 50% of the scrubbed
gas; reheat to 175°F as required. Mid-1982 cost basis.
11-19
-------
TABLE 6. SODIUM/LIMESTONE DOUBLE ALKALI PROCESS ANNUAL REVENUE REQUIREMENTS
Annual
Quantity
107,700 tons
6,190 tons
10,580 tons
Direct Costs - First Year
Raw Materials
Limestone
Soda Ash
Lime
Total Raw Materials Cost
Conversion Costs
Operating Labor
and Supervision
FGD
Solids Disposal
Utilities
S team
Process Water
Electricity
Diesel Fuel
Landfill Fly Ash Credit
Maintenance
Labor and Material
Analysis
Total Conversion Costs
Total Direct Costs
Indirect Costs - First Year
Overheads
Plant and Administrative
(60% of conversion costs
less utilities)
Total First-Year Operating and Maintenance Costs
Levelized Capital Charges
(14.7% of total capital
investment)
Total First-Year Annual Revenue Requirements
Levelized First-Year Operating and
Maintenance Costs (1.886
First-Year 0 and M)
Levelized Capital Charges (14.7%
of Total Capital Investment)
Levelized Annual Revenue Requirements
Unit
Cost,
8.50/ton
160.00/ton
75.00/ton
First-Year Annual Revenue Requirements
Levelized Annual Revenue Requirements
25.52
35.74
Mills/kWh
9,28
13.00
Total Annual
Cost, k$
915
990
794
2,699
43,860 man-hr
33,280 man-hr
528,500 klb
262,100 kgal
29,618,000 kWh
123,200 gal
4,990 man-hr
15. 00/man-hr
21. 00/man-hr
2.50/klb
0, 14/kgal
0. 037/kWh
1. 60/gal
21, 00/man-hr
658
699
1,321
37
1,096
197
(312)
2,715
105
6,516
9,215
13,987
25,521
21,753
13,987
35,740
Basis: One-year, 5,500-hr operation of system as described in capital investment table. Mid-1984
cost basis.
11-20
-------
TABLE 7. LIMESTONE SCRUBBING PROCESS ANNUAL REVENUE REQUIREMENTS
Annual
Quantity
142,900 tons
Direct Costs - First Year
Raw Materials
Limestone
Total Raw Materials Cost
Conversion Costs
Operating Labor
and Supervision
FGD
Waste Disposal
Utilities
Steam
Process Water
Electricity
Diesel Fuel
Maintenance
Labor and Material
Analysis
Total Conversion Costs
Total Direct Costs
Indirect Costs - First Year
Overheads
Plant and Administrative
(60% of conversion costs
less utilities)
Total First-Year Operating and Maintenance Costs
Levelized Capital Charges
(14.7% of total capital investment)
Total First-Year Annual Revenue Requirements
Levelized First-Year Operating
and Maintenance Costs (1.886
first-year 0 and M)
Levelized Capital Charges (14.7%
of total capital investment
Levelized Annual Revenue Requirements
Unit
Cost,
First-Year Annual Revenue Requirements
Levelized Annual Revenue Requirements
29 .10
41,46
.50/ton
Mills/kWh
10-58
15-08
Total Annual
Cost, k$
1,214
1,214
43,860 man-hr
29,120 man-hr
542,200 klb
193,400 kgal
57,657,000 kWh
103,200 gal
4,990 man-hr
15.00/man-hr
21.00/raan-hr
2 .50/klb
0 .14/kgal
0 ,037/kWh
1.60/gal
21 .00/man-hr
658
611
1,356
27
2,133
165
4,285
105
9,340
10,554
3,395
13,949
15,151
29,100
Basis: One-year, 5,500-hr operation of system as described in capital investment table. Mid-1984
cost basis.
11-21
-------
Levelized annual revenue requirements are $36 million (13.00 mills/kWh)
and $41 million (15.08 mills/kWh) for the double alkali and limestone
processes respectively. TVA utilized a projected value of money at 10% and
an inflation rate of 6% to determine these figures.
A comparison of the first year annual revenue requirements was made and
the major components are shown in Table 8. The major differences in direct
costs between the systems are in the area of raw materials, electricity, and
maintenance requirements. The major direct costs for the limestone double
alkali process in order of significance are maintenance, raw materials,
labor, steam, and electricity. For the conventional limestone process they
are maintenance, electricity, labor, steam, and raw materials.
TABLE 8. COMPARISON OF ANNUAL REVENUE REQUIREMENT COMPONENTS
Annual Revenue Requirements, k$
Sodium/Limestone Limestone
Double Alkali Scrubbing
Raw Materials
Limestone 915 1,214
Soda Ash 990
Lime 794 —
Total Labor 2,699 1,214
Utilities
Steam 1,321 1,356
Electricity 1S096 2,133
Other 234 192
Maintenance 2,715 4,285
Landfill Credit (312)
Direct Costs 9,215 10,554
Overheads 2 319 3 395
Operating and Maintenance Costs 11,534 13 949
Levelized Capital Charges 13,987 15 151
Total First-Year Annual Revenue Requirements 25,523 29,100
In general, the TVA study concluded that the limestone double alkali
process has direct costs about 13% lower than those of a conventional
limestone process primarily because of lower maintenance and power costs.
Along with lower overheads which are based on direct costs—and lower
capital charges for the double alkali process, the first year annual revenue
requirements of the double alkali process are 12% lower than those of the
limestone slurry process.
11-22
-------
B. STEARNS-ROGER/EPRI STUDY REVIEW*
Steams-Roger conducted a study under contract to EPRI similar to the
TVA/EPA study discussed previously. The Stearns-Roger study also compared a
limestone dual alkali system with a conventional limestone scrubbing
system. Different design and economic criteria were utilized for the
Stearns-Roger and TVA comparisons but similar conclusions were reached.
The Stearns-Roger study was based on two 500 megawatt PC fired boilers
firing 4.0% sulfur coal. Table 9 outlines the specific process design
criteria utilized in the cost evaluation. Estimates of raw material and
utility consumptions that were utilized in the study are shown in Table 10.
Stearns-Roger also conducted a relatively subjective technical
evaluation of the advantages and disadvantages of the limestone double
alkali process compared with conventional limestone scrubbing. This
comparison is shown in Table 11. Credit was given to the limestone double
alkali process for significant advantages in maintenance as a result of
reduced scaling potential, reagent reactivity, and the ability to follow
changes in flue gas composition and condition.
Tables 12, 13, and 14 review the base case economic evaluation of the
limestone double alkali process done in the Stearns-Roger/EPRI report.
Table 12 shows the total system capital cost including the particulate
removal system. Once the particulate removal system was subtracted from the
total, the capital cost of the limestone double alkali FGD system was
$162/kW. When compared with Stearns-Roger evaluation of conventional
limestone scrubbing systems, the double alkali process requires 5% less
initial investment. This is true even though a higher process contingency
is included in the double alkali cost due to its lack of commercial
development.
Table 14 depicts the levelized busbar cost for the limestone double
alkali system. These cost are calculated by the "present worth" method (as
defined in EPRI Economic Premises) and assume a thirty-year plant operating
life. This levelized bus bar cost is the sum of the fixed and varible
operating costs (Table 13) plus the cost of capital over the thirty-year
plant life. The fixed operating costs shown for limestone double alkali are
approximately 33% lower than those Stearns-Roger calculated for conventional
limestone systems. This is principly due to the elimination of the high
costs high maintenance of slurry recirculation pumps and the attendent lower
overall maintenance requirement. Stearns-Roger also found that the total
varible and capital operating costs are approximately equal for both
systems.
The costs developed by Stearns-Roger in the same study for a limestone
forced-oxidation process are shown in Tables 15, 16, and 17.
* Reisdorf, J. B., Keeth, R. J., Scheck, R. W., Miranda, J. E., Economics
of FGD Systems, to be published 11/83.
11-23
-------
TABLE 9. LIMESTONE DUAL ALKALI SYSTEM
PROCESS SPECIFIC DESIGN CRITERIA*
Flue Gas Handling Area Criteria
Flue Gas Flow Rate (105% design load)
Pressure Drop (flange to flange)
ESP Fly Ash Removal Efficiency
Spray Tower Outlet Gas Temperature
SO? Removal Area Criteria
S02 Removal, High Sulfur Coal
Scrubber Design
Scrubber Modules
L/G Ratio
Makeup Ratio
Absorber Tower Superficial Velocity
Reaction Mix Tank Retention Time
Oxidation Rate of Sulfite to Sulfate
Solution pH
Absorber Pressure Drop
Reagent Feed Area Criteria
Total Soda Ash Storage
Soda Ash 20% Solution Storage Tank
Regenerant Limestone Storage
Limestone Slurry Tank Storage
Limestone Day Bin Storage
Limestone Slurry Solids Concentration
Absorber Regeneration Area Criteria
Limestone Reaction Tank Retention Time
Limestone Reagent Feed Ratio
Waste Handling Area Criteria
Thickener Underflow Percent Solids
Dewatered Sludge Percent Solids
Sulfite/Sulfate Mole Ratio
Thickener Size
Vacuum Filter Design Basis
Water Washes for Sodium Recovery
2,015,000 acfm
9 in. H20
99.8%
127°F
90% (30-day rolling average)
Spray Tower, 3 Levels
4 @ 33-1/3% capacity ea.
(3 opp, 1 spare)
20 gpm/1000 acfm
O.Q425 mole
mole S02 removed
10 fps
6 rain
18% available
5.8 - 6.2
6 in H20
30 days
30 hrs
60 days
12 hrs
30 hrs
25%
120 min (total for 4 tanks)
1.05 Ib mole
Ib mole S02 removed
20%
55%
9:1
22 ft2/tpd day solids
35 Ib/hr dry solids/ft3
3
*For one 500 MW unit
11-24
-------
TABLE 10. LIMESTONE DUAL ALKALI SYSTEM
RAW MATERIAL AND UTILITY CONSUMPTION
FOR TWO 500 MW UNITS*
Item Quantity
Soda Ash @ 0.0425 Stoichiometric Ratio 2.5 t/hr
Limestone @ 1.05 Stoichiometric Ratio 61.6 t/hr
Fixative Lime @ 3% of Dry Sludge and Fly Ash 4.7 t/hr
Raw Water 1,250 gpm
Steam @ 572°F and 418 psig 221,000 Ib/hr
Power (Operating Horsepower and Equivalent kW)
Area 10 - Reagent Feed System 2,040 HP (1,520 kW)
Area 20 - S02 Removal System 2,695 HP (2,010 kW)
Area 30 - Flue Gas System 11,010 HP (8,215 kW)
Area 40 - Regeneration System 150 HP (110 kW)
Area 60 - Waste Handling System 2,880 HP (2,150 kW)
Area 70 - General Support Area 400 HP (310 kW)
Area 80 - Particulate Removal System 5,360 HP (4,000 kW)
Total 24,540 HP (18,315 kW)
Fly Ash (direct disposal to landfill fixated
with waste sludge) 63.2 t/hr
*0peratig at 100 percent load
11-25
-------
TABLE 11. TECHNICAL EVALUATION
ADVANTAGES AND DISADVANTAGES
OF THE DUAL ALKALI PROCESS WITH
LIMESTONE REGENERATION*
(Compared to the Conventional
Limestone Process)
Item
1. Process
a. Complexity of Operation
b. Oxidation of Sulfite to Sulfate
c. Turndown Ratio
0)
to
4-J
C
to
CO
• H
o
60
O
4-J
CO
CO
• r-l
Q
0)
CO
X
X
8P
a)
SP
M
C
2
4-1
CO
d. Load Following Capability
e. Adaptability to Flue Gas Temperature Changes
f. Surge Requirements
g. Capability of Using Cooling Tower Slowdown
h. Stability of Process
i. Extreme Vessel Pressure
X
X
X
X
X
2.
j. High Equipment Operating Temperature
k. Use of Liquid Fuel or Natural Gas
1. Material Handling Characteristics
m. Separation/Removal (two phase)
Operation and Maintenance Requirements
a. Labor Requirements
b. Equipment Pluggage, Scaling
c. Equipment Corrosion
X
X
X
X
X
X
X
d. Equipment Erosion
e. Reagent Reactivity/Makeup Rate
3. Effect on Net Plant Heat Rate
a. Power Consumption - Low Absorption
Tower Pressure Drop
b. Steam Usage
4. Disposal
a. Land Requirements
b. Reactivity of Waste
c. Waste Handling Characteristics
5. Use of Exotic Materials of Construction
6. Operational Hazard
a. High Temperature
b. High Pressure
c. Use of Hazardous Chemicals
X
X
X
X
X
X
X
X
11-26
-------
TABLE 12. DUAL ALKALI (LIMESTONE) SYSTEM
TOTAL CAPITAL REQUIREMENT*
Area Description $ / kW
20 S02 Removal System 41.0
40 Regeneration System 4.0
30 Flue Gas Handling System 24.0
10 Reagent Feed System 16.0
60 Waste Handling System 22.0
80 Particulate Removal System 37.0
70 General 3.0
Total Process Capital 147.0
General Facilities 15.0
Engineering and Home Office Fees 18.0
Project Contingency 30.0
Process Contingency 8.0
Total Plant Cost 218.0
Allowance for Funds During Construction (AFDC) 8.0
Total Plant Investment 226.0
Royalty Allowance 0.7
Preproduction Costs 7.2
Inventory Capital 1.4
Initial Catalyst and Chemicals 0.0
Total Capital Requirement 235.0
Less Capital for Particulate Control System 73.0
Total Capital Requirement for FGD System 162.0
^December 1982 dollars, two-500 MW units.
11-27
-------
TABLE 13. DUAL ALKALI (LIMESTONE) SYSTEM
OPERATING COSTS
Fixed Operating Costs
Operating Labor
Maintenance Labor
Maintenance Material
Administration and Support Labor
Total Fixed Operating Costs
Variable Operating Costs
Lime
Limestone
Soda Ash
Raw Water
High Pressure Steam
Power
Sludge (unlined)
Total Variable Operating Costs
Credits for By-products
Total Variable Operating Costs
with By-product Credit
1st Year
$/kW*
1.3
1.9
2.8
0.9
6.9
2.0
0.3
4.8
5.7
5.5
24.9
0.0
24.9
Levelized
mills/kWh
0.5
0.8
1.1
0.4
2.8
0.7
2.0
0.8
0.1
2.0
2.5
2.2
10.3
0.0
10.3
*December 1982 dollars
11-28
-------
TABLE 14. DUAL ALKALI (LIMESTONE) SYSTEM
LEVELIZED BUSBAR COST*
Mills/kWh
Process Capital 3.9
General Facilities 0.4
Engineering and Home Office Fees 0.5
Project Contingency 0.8
Process Contingency 0.2
Total Plant Cost 5.8
AFDC 0.2
Total Plant Investment 6.0
Royalty Allowance 0.02
Preproduction Costs 0.19
Inventory Capital 0.04
Total Capital Requirement 6.3
Fixed Operating Cost 2.8
Variable Operating Cost 10.3
Total Levelized Busbar Cost 19.4
Less By-product Credit 0.0
Total Levelized Busbar Costs
with By-product Credit 19.4
Reference Case Levelized Busbar Cost 3.8
Levelized Busbar Cost Apportioned to FGD 15.6
*Two-500 MW units.
11-29
-------
TABLE 15. FORCED OXIDATION SYSTEM
TOTAL CAPITAL REQUIREMENT*
Area Description $/kW
20 S02 Removal System 67.0
30 Flue Gas Handling System 24.0
10 Reagent Feed System 14.0
60 Waste Handling System 10.0
80 Particulate Removal System 37.0
70 General 7.0
Total Process Capital 159.0
General Facilities 16.0
Engineering and Home Office Fees 20.0
Project Contingency 34.0
Process Contingency 3.0
Total Plant Cost 232.0
Allowance for Funds During Construction (AFDC) 8.0
Total Plant Investment 240.0
Royalty Allowance 0.8
Preproduction Costs 7.4
Inventory Capital 0.9
Initial Catalyst and Chemicals 0.0
Total Capital Requirement 250.0
Less Capital for Particulate Removal System 73.0
Total Capital Requirement for FGD System 177.0
*December 1982 dollars, two-500 MW units.
11-30
-------
TABLE 16. DUAL ALKALI (LIMESTONE) SYSTEM
OPERATING COSTS
Fixed Operating Costs
Operating Labor
Maintenance Labor
Maintenance Material
Administration and Support Labor
Total Fixed Operating Costs
Variable Operating Costs
Limestone
High Pressure Steam
Power
Dry Solids (unlined)
Gypsum
Total Variable Operating Costs
Credits for By-products
Total Variable Operating Costs
with By-product Credit
1st Year
$/kW
1.0
3.0
4.5
1.2
9.7
5.4
4.8
7.8
1.7
1.8
21.5
0.0
21.5
Levelized
mills/kWh
0.4
1.2
1.8
0.5
3.9
2.2
2.1
3.3
0.7
0.7
9.0
0.0
9.0
*December 1982 dollars.
11-31
-------
TABLE 17. FORCED OXIDATION SYSTEM
LEVELIZED BUSBAR COST*
Process Capital
General Facilities
Engineering and Home Office Fees
Project Contingency
Process Contingency
Total Plant Cost
AFDC
Total Plant Investment
Mills/kWh
4.3
0.4
0.6
0.9
0.1
6.3
0.2
6.5
Royalty Allowance
Preproduction Costs
Inventory Capital
Total Capital Requirement
Fixed Operating Cost
Variable Operating Cost
Total Levelized Busbar Cost
Less By-product Credit
Total Levelized Busbar Costs
with By-product Credit
Reference Case Levelized Busbar Cost
Levelized Busbar Cost Apportioned to FGD
0.02
0.20
0.02
6.7
3.9
9.0
19.6
0.0
19.6
3.8
15.8
*Two-500 MW units.
11-32
-------
Stearns-Roger concluded from their study that the levelized bus bar cost
apportioned to the limestone double alkali process is approximately 9% less
than that of a conventional limestone process.
Discussion of the Economic Evaluations
As the foregoing summary of the TVA and Stearns-Roger economic
evaluations shows, the evaluations are similar in general approach and
procedure but differ appreciably in both the design and economic assumptions
and specific methodology. The differences occur in three categories: the
power plant size and operating conditions, the FGD system design and
operating conditions, and the economic criteria and costing procedures.
These differences preclude comparisons of the two evaluations in terms of
absolute values or specific details. Both evaluations, however, produce the
same general conclusions and illustrate the same economic relationships of
the limestone double-alkali process and the limestone forced-oxidation
process. This is best shown by the direct capital investment and operating
cost summaries shown in the preceding discussion, selected values of which
are summarized in Table 18. The absolute values of the TVA and
Stearns-Roger results are not comparable because of the differences
discussed, but the differences between the costs for the two processes in
each evaluation can be compared.
TABLE 18. MAJOR DIRECT COSTS FROM THE
TVA AND STEARNS-ROGER ECONOMIC EVALUATIONS
TVA Evaluation
Capital Investment
SO2 Removal
Flue Gas Handling
Reagent Feed
Waste Handling
Regeneration
General
Operating Costs
Labor
Maintenance
Limestone
Soda Ash
Lime
Steam
Electricity
Limestone
Double
Alkali
Limestone
Forced
Oxidation
$/kW
23
29
14
22
3
5
46
30
14
17
0
6
Levelized, mills/kWh
0.9
1.8
0.6
0.
0.
0.9
0.7
0.9
2.9
0.8
0
0
0.9
1.4
Stearns-Roger Evaluation
Limestone Limestone
Double Forced
Alkali Oxidation
$/kW
41
24
16
22
4
3
67
24
14
10
0
7
Levelized, mills/kWh
0.5
1.9
2.0
0.8
0.7
2.0
2.5
0.4
3.0
2.2
0
0
2.1
3.3
11-33
-------
In direct capital investment, the limestone double alkali process has a
much lower cost for SC>2 absorption in both evaluations, while the
additional costs for absorbent regeneration are relatively small. The
difference in absorber costs is almost wholly responsible for the lower
capital investment of the limestone double alkali process. (The Stearns-
Roger evaluation also has a lower waste disposal cost for the limestone
forced-oxidation process that reduces the cost difference beween the two
processes; this, however, is due to the use of stacking of gypsum disposal,
while the TVA design has a conventional landfill.)
In operating costs, both evaluations also shown that the maintenance
costs and electricity costs are appreciably lower for the limestone double
alkali process, both the result of the lower absorber costs and pumping
requirements. These lower costs more than offset the additional costs for
soda ash and lime.
Since the size of the absorbers is largely determined by the flue gas
volume which is the same for both applications, the lower SC>2 absorption
costs for the limestone double alkali process are largely a result of the
smaller and less complicated absorbent recirculating system and lower
pumping costs associated with the lower L/G ratio. The double alkali
process is inherently flexible in the L/G ratio because the highly reactive
absorbent solution—which also has a low propensity for scale formation—
allows a wide choice of absorber designs. It is possible that the absorber
design could range from a spray tower to a design similar to a packed tower,
with corresponding high to low L/G ratios, thus allowing an optimization of
absorber design and L/G ratio.
The different absorber design philosophy is seen in the designs used in
the TVA and Stearns-Roger economic evaluations. The TVA design is based on
information provided by EPA, in part from the limestone double alkali test
at the Scholz steam plant. A sieve tray absorber is used and the L/G ratio
is 8 gal/acf, 5 gal/acf of which is regenerated absorbent from the
regeneration system. The Stearns-Roger design is based on information from
FMC Corporation. A spray tower absorber is used and the L/G ratio is 20
gal/acf, all of which is unregenerated absorbent from the absorber hold
tank.
The effects of these differences are impossible to quantify by comparing
the two economic evaluations because of the differences in the design and
economic premises. To compare the economics of the two designs, TVA
incorporated the FMC design and operating conditions (which includes a
higher limestone stoichiometry, a longer regeneration system hold time, and
a lower oxidation rate, as well as, spray tower absorbers and a higher L/G
ratio) into the TVA process and repeated the economic evaluation. All other
conditions and economic procedures and criteria remained the same. The
major conditions are shown in Table 19; other conditions do not differ from
those summarized in the preceding discussion. The capital investments and
annual revenue requirements are shown in Tables 20 and 21. The economics of
the limestone forced-oxidation process are included for comparison.
11-34
-------
The overall effect of the changes is small; the total capital investment
is increased 4% and the annual revenue requirements are increased 4%, as
compared to the original design. The use of a spray tower absorber with a
higher recirculation rate increased the capital investment for SC>2
absorption by 14%, but electricity costs were decreased about 3% because of
the lower flue gas pressure drop in the system. The only other capital
investment components affected to a significant extent were the regeneration
area costs, which increased 13% because of the longer hold time, and the
fixation costs, which increased 10% because of the higher stoichiometry.
The only annual revenue requirement components affected to a significant
extent in addition to the electricity costs were the limestone costs, which
increased 4% and the maintenance costs, which increased slightly because of
the higher absorber costs.
Based on this comparison, the effects of the different design criteria
on the economics of the process are minimal. The use of an absorber such as
a sieve tray that allows a lower L/G ratio reduces the capital cost of the
SC>2 absorption area to some degree but it has little effect on electricity
costs because of the higher fan costs associated with the higher flue gas
pressure drop. The changes in limestone stoichiometry; regeneration area
hold time, and the oxidation rate have little effect on the economics of the
process.
CONCLUSIONS
As a result of the work done by TESI at Plant Scholz and by FMC, it can
be concluded that a limestone double alkali process is technically
feasible. The economic evaluations by TVA and Stearns-Roger indicate that
the cost of owning and operating a limestone double alkali system is less
than those of conventional limestone scrubbing. On this basis, it appears
that consideration of the technology is warranted by those faced with the
installation of flue gas desulfurization systems, especially on plants
burning relatively high sulfur fuels.
TABLE 19. MAJOR OPERATING CONDITIONS FOR THE ECONOMIC COMPARISON
OF SIEVE TRAY AND SPRAY TOWER ABSORBERS
Absorber Type
System Pressure Drop, in. H20
Superficial Velocity, ft/sec
Unregenerated L/G, gal/acf
Regenerated L/G, gal/acf
Lb Na2C03/lb S02 absorbed
Mole CaC03/mole S02 absorbed
Oxidation, %
Regeneration Hold Time, Min
Limestone
Double
Alkali
Sieve Tray
10
9
3
5
0.01
1.01
10
100
Limestone
Double
Alkali
Forced
Oxidation
Limestone
Spray Tower Spray Tower
10
20
0
0.01
05
1
5
120
10
110
0
0
1.4
95
11-35
-------
TABLE 20. CAPITAL INVESTMENT COMPARISON FOR SIEVE AND SPRAY TOWER ABSORBERS
Capital Investment, k$
i
UJ
ON
Direct Investment
Materials Handling
Feed Preparation
Gas Handling
S02 Absorption
Reheat
Regeneration
Oxidation
Solids Separation
Fixation
Total Process Capital
Services, Utilities, and Miscellaneous
Total Direct Investment Excluding Landfill
Landfill Construction
Landfill Equipment
Landfill Credit (Fly Ash Disposal)
Total Direct Investment
Indirect Investment
Total Fixed Investment
Other Capital Investment
Total Capital Investment
Dollars per kW of generation capacity
Limestone
Double Alkali
Process
Sieve Tray
2,426
4,506
10,800
11,348
3,630
1,506
0
5,493
906
40,615
2,437
43,052
5,247
1,454
(2,312)
47,441
73,271
21,881
95,152
190
Spray Tower
2,458
4,565
10,615
12,979
3,630
1,700
0
5,533
1,000
42,480
2,548
45,028
5,247
1,454
(2,312)
49,417
76,352
22,747
99,099
198
Forced-Oxidation
Limestone
Process
2,528
4,715
11,281
20,288
3,634
0
2,670
3,679
0
48,795
2,928
51,723
3,781
1,123
0
56,627
80,575
22,493
103,068
206
-------
TABLE 21. ANNUAL REVENUE REQUIREMENT COMPARISON
FOR SIEVE AND SPRAY TOWER ABSORBERS
Annual Revenue Requirements
Raw Materials
Limestone
Soda Ash
Lime
Total Labor
Utilities
Steam
Electricity
Other
Maintenance
Landfill Credit
Analysis
Direct Costs
Overheads
Operating and Maintenance Costs
Levelized Capital Charges
Total First-Year Annual Revenue
Requirements
Levelized Operating and
Maintenance Costs
Levelized Capital Charges
Levelized Annual Revenue
Requirements
First-Year, mills/kWh
Levelized, mills/kWh
Limestone
Double Alkali
Sieve Tray
915
990
794
2,699
1,357
1,321
1,096
234
2,715
(312)
105
9,215
2,319
11,534
13,987
25,521
21,753
13,987
35,740
9.3
13.0
Spray Tower
951
990
794
2,735
1,357
1,321
1,067
236
2,834
(314)
105
9,341
2,391
11,732
14,578
26,310
22,126
14,578
36,704
9.6
13.3
Limestone
Scrubbing
1,214
1,214
1,269
1,356
2,133
192
4,285
105
10,554
3,395
13,949
15,151
29,100
26,309
15,151
41,460
10.6
15.1
11-37
-------
DEVELOPMENTS AND EXPERIENCE IN FGD MIST ELIMINATOR
APPLICATION
R. T. Egan, W. Ellison
-------
DEVELOPMENTS AND EXPERIENCE
IN FGD MIST ELIMINATOR APPLICATION
By:
Richard T. Egan, PE
The Munters Corporation
Ft. Myers, Florida 33901
William Ellison, PE
Ellison Consultants
Monrovia, Maryland 21170
ABSTRACT
The purpose of this paper is to detail available mist-eliminator
technology and to assess U.S. practices and experience in utilization of
these sub-systems of flue gas desulfurization (FGD) processes. Pertinent
trends in FGD system design and operation tied to the criticality of mist
eliminator performance are identified and discussed and advancements in
eliminator selection and application are reviewed. Case histories of a
number of significant existing mist eliminator facilities are described and
the importance of selection of internals design and arrangement permitting
use of elevated mist-eliminator inlet-face gas velocity to enhance droplet
separation forces is emphasized.
INTRODUCTION
The mist eliminator is a critical part of the wet flue gas desulfuriza-
tion system and should be selected and effectively designed to avoid in-
crustation and corrosion of downstream system components, as well as to help
limit the amount of carryover to the stack of suspended solids, dissolved
salts and liquid (1).
FGD OPERATIONAL AND MAINTENANCE PROBLEMS RELATED TO MIST ELIMINATOR DESIGN
Results of the operation of flue gas desulfurization systems in utility
plants indicate that a mist eliminator facility may be a major operational
and maintenance problem, detracting from gas-cleaning system reliability.
Principal difficulties have included inefficient performance,
plugging/scaling, erosion and corrosion of mist eliminators, and deterioria-
tion of downstream components including stack liners.
11-39
-------
Inadequate Eliminator Performance
Instances of poor droplet collection are the result of system design
deficiencies. Under ideal service conditions, such as with provision for
uniform gas flow distribution, effective means of keeping elements clean,
and adequate drainoff of collected liquid, eliminators perform well. In
tests conducted with mist eliminator surfaces clean and gas-flow distri-
bution nearly uniform, extensive measurements of outlet carryover solids
loadings in flue-gas discharges from 10-20 MW demonstration FGD systems with
vertical-flow mist eliminators indicated that eliminator outlet solid
particulate concentration was consistently lower than 0.01 Ib/million Btu,
and thus very much lower than the current New Source Performance Standard of
0.03 Ib total solid particulate/million Btu. In all cases, scubber outlet
solids concentrations were shown to be less than inlet concentrations, even
for inlet solids loadings as low as 0.004 grains/SCF. Available data on
commercial systems strongly suggest that poor emission-control performance
by FGD mist eliminators is in some cases the direct result of fouling of the
elements that changes blade profile geometry causing deterioration of
efficiency.
Fouling/Corrosion
Provisions for chemical scale control designed into the scrubber
slurry-recirculating system best achieve their purpose in the absorber
section where dispersal of liquid flow at high rates continuously irrigates
the exposed internal surfaces. Because of its lack of such irrigation, and
since carryover liquid collected on and wetting the mist eliminator internal
surface will absorb residual SO and 0 from the scrubbed gas, the mist
eliminator is the most likely component to be fouled. Thus the surfaces of
the mist eliminator must be effectively washed to prevent gypsum scaling and
deposition of solids contained in sulfite/sulfate-laden droplets.
Impact on FGD Reliability
Nationwide FGD reliability data indicate (2) that the scrubbed-gas
handling facilities downstream of the FGD absorbers are the principal
contributors to reduced unit reliability. Critical corrosion and fouling in
this low-pH wet section have been found to be strongly influenced by the
extent of droplet-carryover flow at the exit of mist eliminators.
CRITICAL DESIGN ASPECTS
As noted, scrubbing systems can operate without significant deposition
of solids on mist eliminator collecting surfaces. Furthermore, if the
surfaces are clean and the system is properly designed, entrainment of FGD
liquor does not increase outlet solid particulate emission.
Gas Flow Distribution
Uniformity of gas-flow .distribution is of major significance in elimi-
nator performance and, thus, use of design and flow-modeling techniques to
equalize gas velocity across the cross-sectional flow area is essential.
11-40
-------
One report (3) shows that local vertical gas-flow velocities entering the
first level of spray banks in a spray tower ranged from approximately 120%
to 20% of average velocity after the addition of integral gas-distribution
vaning, as compared to a range of 250% to minus 20% without vaning. While
other adjustments were called for to achieve a more satisfactory distribu-
tion to prevent poor mist-eliminator operation and liquid entrainment, it
can be seen that vaned internal distribution means can be an essential first
step.
In-Situ Washing Means
Although proper washing of eliminator internals is best done with fresh
water, the material-balance constraints of closed-loop recirculating
scrubbing-slurry systems require close limits on fresh-water input. Thus,
recycled process liquid must be used in many existing scrubber systems,
particularly those of the vertical gas-flow mist-eliminator type, to achieve
the required mist-eliminator wash-spraying intensity. By the mechanical
action of wash sprays upstream and downstream of the mist eliminator signif-
icant deposition of soft, sulfite-solids can be prevented. To control
deposition of CaSO from a super-saturated scrubber liquid the designer
relies on absorber pH control and the use of CaSO,-unsaturated mist-
eliminator wash liquid.
Prevention of Reentrainment by Controlled Drainage
High carryover-liquid loading at the mist eliminator inlet and/or
localized, high gas velocity can lead to adverse reentrainment of collected
droplets when the design fails to provide adequate means of removing the
liquid-catch from the eliminator elements. Through advancements in develop-
ment and application of mist eliminators, improved liquid drainage provi-
sions are available to the FGD designer.
APPLICABILITY OF ALTERNATIVE DESIGN CONFIGURATIONS
Vertical Gas Flow
Vertical-flow eliminators may be conveniently fitted into the upper
section of the scrubber tower and have been preferred for this reason. They
have generally used a chevron-type design with horizontal zig-zag baffling
of two to six passes. Due to the tendency for liquid reentrainment the
chevrons may be positioned on a slant to obtain better drainage.
Horizontal Gas Flow
The horizontal-flow mist eliminator design, most common in Japan and
Germany, has only recently been applied in utility design in the United
States. Baffles in a horizontal-flow eliminator are generally oriented in a
vertical position. In high gas-velocity designs J-shaped phase separation
chambers or disengaging columns extending from the surface of the elements
counter to the gas flow afford drainage paths external to the gas-flow
regime to avoid reentrainment of liquid by the gas stream. This configura-
tion lends itself to use of higher design gas velocities with resulting
11-41
-------
increased droplet-removal efficiency due to increased inertial separating
forces and better drainage. It also permits segregation of the eliminator
liquid-catch from the recirculating absorber slurry stream, which is advan-
tageous in eliminator washing-system design, permitting use of a recirculat-
ing wash water circuit.
AVAILABILITY OF GENERIC MIST ELIMINATOR TYPES FOR FGD SERVICE
INERTIAL METHODS
The most commonly applied FGD mist eliminators use gravitational (dyna-
mic) forces caused by change in gas direction to accomplish separation of
liquid droplets from scrubbed gas. A vertical-flow arrangement has been
regularly used in utility service in U.S.A. Horizontal-flow design, which
is more common in Japan and Germany, has only recently been introduced in
utility applications here.
Chevron/Baffle Type
Vertical-flow eliminators have historically been of the rudimentary
chevron, zig-zag baffling type. Per Figure 1, the sweep of the gas flow
Figure 1. Reentrainment in Chevron Mist Eliminators
tends to reentrain collected liquid. An improvement in this design through
slanting of baffles (Figure 2) provides a vertical directional component for
liquid flow along the length of the baffle to give better drainage of wash
water and collected mist downstream of a bulk entrainment separator of
horizontal orientation designed to collect most of the inlet liquid (3).
Inclining of mist eliminator baffles in this manner at 30 degrees from the
horizontal (4) increases inlet liquid loading capability free of significant
reentrainment by at least 100% at high vertical gas velocity. A general
11-42
-------
review (5) of commercial FGD design in the utility industry reports the
following U.S. practices in eliminator application:
Chevron Vinos
Washer Lance
Bulk Entninment Separator (BES)
Figure 2. Slanted-Baffle Mist Eliminator
Principal use of mist eliminators of chevron multi-pass (continuous-
vane) construction in a vertical-gas flow configuration is favored
from strength and cost considerations. Baffle-type units that
substitute noncontinuous-slat construction are also commonly used
and, like the chevron design, offer comparatively simple, open
geometry with low gas-pressure drop.
Baffle spacing is normally 1^ to 3 inches except in the second stage
of two-stage designs which generally use 7/8 to 1 inch spacing.
Plastic construction is most common due to reduced weight, cost and
corrosion tendencies.
Precollection and pre-washing stages are commonly used to improve
mist eliminator operation.
Eliminator wash systems typically operate intermittently and use a
mixture of clear scrubbing liquid and fresh water makeup.
11-43
-------
A unique and highly efficient vertical-flow baffled arrangement that
utilizes exceptionally high design velocity without reentrainment is illus-
trated in Figure 3. This unit has advantageous 45 deg. slanting in two
Gas out
Liquid and
gas in
Liquid drain area
against flow
Figure 3. Mist Eliminator Drainage by Multi-Directional Slanting
directions in that the total gas stream is deflected from its straight
course by a series of zig-zag shaped channel walls. Slanted chevron-shaped
airfoils on the eliminator walls collect and drain liquid droplets removed
from the gas stream. Installations of this type operating in U.S. for over
nine years have aptly field-demonstrated the high efficiency of this slant
design. In some instances, a bulk entrainment separator of 45 degree
11-44
-------
orientation (Figure 4) is placed below (upstream of) the zig-zag channel-
wall eliminator.
Figure 4. Slanted-Baffle Bulk Entrainment Separator
Vane Type
Elements in the horizontal flow eliminators are typically of vaned
construction and oriented in a generally vertical position. As noted
earlier, phase separation chambers (6) extending counter to the gas flow
direction afford drainage paths external to the gas-flow regime for removal
of liquid to avoid reentrainment by the high-velocity gas stream. (Figure
5).
PACKED TYPE
Neither loose nor rigid types of packing have been used for mist
elimination in FGD service as they are generally prone to excessive fouling
and plugging when used in services in which the flue gas or recirculating
scrubbing medium have significant concentrations of suspended solids. Based
on its efficient performance on small micron droplets and reliable service
in difficult industrial applications, a unique layered-mesh dry-bed (7) mist
11-45
-------
Figure 5. Function of Phase Separation Chamber at Elevated Gas Velocity
eliminator construction is now being considered for trial application in the
utility industry. It is understood that the packing, an interlaced struc-
ture of thermoplastic monofilaments with essentially all of the filaments
oriented perpendicular to gas flow, may generally be cleaned in-situ by
intermittent spray washing. In very fouling-prone industrial services the
layered-mesh beds have been found to conveniently lend themselves to perio-
dic mechanical cleaning when temporarily removed from the unit. It is
contemplated that a utility FGD scrubber module will be backfitted with a
substitute triple-layer mesh-bed mist eliminator of this type to compare its
reliability and performance with that of the original common horizontal-
chevron type.
WET ELECTROSTATIC PRECIPITATION
Wet electrostatic precipitators, extensively used in the chemical
process industries, offer a highly efficient and effective means of col-
lected carryover droplets from FGD systems and, when present in significant
concentrations, reducing emission of sub-micron sulfuric acid mist parti-
cles. Engineering and costing studies indicate (8) that the wet electro-
static type offers savings in overall stack gas cleaning system design and
operation when also used simultaneously to control emissions of fine fly-ash
particulate matter in fly-ash scrubbing type FGD applications that use
neither dry fly-ash collectors nor venturi scrubbers. A 1 mW pilot plant
test of this concept has been completed in 1983 on high-sulfur-coal utility
plant service, and a 50-100 mW size demonstration is now contemplated. A
unique new 150 mW cogeneration-type powerplant that is to be located at the
Houston Ship Channel and fire petroleum coke has selected a wet-
electrostatic-precipitator type mist eliminator for its wet FGD system in
order to prevent an exceptionally high sulfuric acid mist emission tied to
very high sulfur content of the fuel.
11-46
-------
ADVANCED DESIGN TECHNIQUES FOR IMPROVED FGD RELIABILITY AND ECONOMICS
The design of the mist elimination facility is critical to FGD system
dependability and to performance and economics of the utility plant itself.
Available FGD mist eliminator technology has advanced broadly since earliest
designs that used primitive structural angle irons and later chevrons in the
scrubber outlet to reduce the liquid escape. Thus, while many U.S. system
designers are content to use the simple low-velocity, low-efficiency generic
eliminator configurations, advancements in inertial separation by empiri-
cally designed profiles continue to contribute significantly, worldwide, to
the performance and reliability of many installations. Technologically
advanced dynamic mist eliminator designs of German origin, (Figure 6 and
Figure 3), utilized only sparingly in the U.S.A. to date, greatly reduce
carryover that can aggravate solid particulate emissions and downstream
corrosion and fouling on low-pH duct walls, flue-gas reheat devices and
Figure 6. High-Performance Horizontal-Flow German Mist Eliminator Design
stack-liner walls. As noted earlier, these eliminators provide unique
liquid drainage columns incorporated in aerodynamically designed eliminator
vanes or passages. This permits use of elevated design superficial gas-
velocity so as to more effectively trap out small micron droplets without
reentraining the liquid that collects on the internal eliminator surfaces.
RELIABILITY EXPERIENCE AND CONCERN
FGD system reliability experience continues to be generally fair to
poor and it appears, clearly, that greatest utilization of advancements in
mist eliminator technology will be of significant benefit in improving the
dependability of the entire wet scrubbing operation.
11-47
-------
FGD Reliability in U.S.A.
FGD reliability, the accumulative amount of time the system was oper-
ated during any calendar period divided by the duration of time periods
during which the system was called upon to operate, has averaged only 81% in
the U.S.A. during the period 1974-1982 for a large number of FGD systems
tracked by EPA (9). For the most recent years the average reliability has
been 83.3% (1980), 82.0% (1981), and 81.3% (1982). No major improvement
trend is discernable and, thus, there is substantial room and possible
motivation for upgrading mist eliminator design practice and performance to
yield significant benefits in increased system reliability. Moreover, in
high-sulfur coal service, very low reliability, often less than 70%, has
been experienced, due in part to the impact of carryover slurry contributing
to outages through downstream fouling and low-pH corrosion. Selection of a
high-performance mist eliminator is particularly vital in such service.
Mist Eliminator Reliability/Performance and Their Influence on FGD and Unit
Reliability
Reliability of the mist eliminator depends on its being maintained in
clean condition, which is directly tied to the nature of the absorber
emission and the chemistry of the recirculating scrubbing medium, the
quality of wash water used and the adequacy and timing of washing flows. A
reliable eliminator installation that achieves high droplet removal effic-
iency will contribute to high FGD and unit reliability. Conversely, fouling
of the mist eliminator will ultimately require shutdown of the absorber
module, placing a spare module into operation if available. Additionally,
FGD failure data by cause code collected from 1978 to the present by the
National Electric Reliability Council (NERC) indicates (9) that FGD unreli-
ability caused only a 2-3% forced outage rate of power-generating capacity.
Thus, since as noted earlier, FGD reliability is typically only 81%, it is
apparent that bypassing of FGD during FGD-shutdown, which is possible in
most existing installations, has been very common. The NERC data also shows
that mist eliminators were deemed directly responsible for only 9% of all
plant outages caused by FGD. However, other components that are not spared,
but can be significantly impacted by unremoved slurry carryover droplets,
caused a substantial amount of the lost plant time. For example of all
plant outages attributed to FGD failures, 28% was caused by dampers, 19% by
ductwork and 17% by fans. Thus it appears that a combination of poor
reliability and poor performance in many mist eliminator installations may
be a major cause of FGD unreliability experienced in many FGD installations.
Moreover, as discussed later, through an understanding of of the adverse
chemistry and nature of emissions from U.S. supersaturated-CaSO mode FGD
installations (e.g., natural-oxidation limestone slurry scrubbing) and of
critical difficulties in achieving effective washing in this mode, one can
readily and clearly discern why long-term U.S. reliability experience for
high-sulfur FGD installations of this mode/type has been so very inferior
(10) (11) to that of unsaturated-mode installations, (i.e., magnesia-
buffered calcium-based slurry scrubbing and sodium-based dual alkali scrub-
bing.)
11-48
-------
Field Investigative Program by Electric Utility Industry
Due to widespread mist eliminator problems that have been reported to
it, and a feeling that FGD system suppliers and architect-engineers may have
been neglectful, EPRI has begun a field study program to increase under-
standing of how to size mist eliminators and avoid corrosion and plugging
(12).
Concerns of EPRI — Principal difficulties reported by EPRI that are to be
addressed are as follows:
- Half of the existing FGD systems are probably experiencing mist
elimination system problems.
- Malfunction of mist elimination systems has resulted in plugging and
corrosion, causing significant FGD absorber down-time.
- The impact on mist eliminator operations of gas distribution, wash
water quality and water droplet size distribution needs to be
quantified.
- The effects on wet stacks of mist-eliminator operational inade-
quacies need to be quantified.
There is a need for comprehensive information on mist eliminator
systems so that design and operating guidelines can be established.
Scope of EPRI Work — The EPRI project plan includes:
Execution of field tests and data collection at specific FGD sites,
including flue gas measurements, sample collection/analysis and
historical data collection.
Evaluation of mist eliminator pressure drop and velocity profile,
and analysis of the source and type of plugging or carryover.
Recommendation of improved wash sequences, water sources, mainte-
nance procedures and design modifications.
Development of guidelines for testing, evaluation and troubleshoot-
ing.
TRENDS IN ADVANCEMENT OF FGD SYSTEM DESIGN AND OPERATION
Broader application of high-performance eliminators is being made in
conjunction with continuing advancements in FGD system design. Increasing
use of spray tower type FGD absorbers has substantially raised mist elimina-
tor liquid loadings requiring improved performance and liquid handling
capacity. Droplet removal for protection against downstream fouling becomes
particularly critical in systems that employ wet flue gas heat exchangers of
the direct and regenerative reheat type. Current general practice in
selection and specifying of mist eliminators calls for limiting droplet
emission to no more than about 0.1 pounds liquid per million Btu boiler heat
11-49
-------
input, (approximately 0.05 grains/SCF, dry), thus limiting the solid emis-
sion contribution of carryover to about 0.01 pounds/MM Btu (0.005
grains/SCF, dry), which is a small fraction of the current 0.03 pounds/MM
Btu New Source Performance Standard for total solid particulate emissions
from coal-fired utility boilers. Sustained performance at such high effic-
iency is critically tied to operating conditions in the FGD system, and thus
it is essential that the absorber vessel designer be cognizant of the design
criteria and performance limitations of the mist eliminator. Moreover, the
successful FGD installation requires closest coordination between the
responsible system designer and the mist eliminator manufacturer.
Regenerative Flue Gas Reheat
The very high plant lifetime cost of energy required for adequate
reheat of flue gas has led to increasing use of regenerative reheat exchang-
ers downstream of the mist eliminator. Ljungstrom rotary heaters using hot
flue gas are being employed in Japan and West Germany while tubular heat
exchangers supplied with hot water are operating in U.S.A. In all such
installations mist eliminator design has been a special concern, and in such
applications it is particularly essential that outlet droplet loading
specifications be met reliably.
Spray-Tower Type Absorbers
Because of simplified-internals and good load-turndown capability spray
tower type absorbers have come into broad use in FGD service. The rela-
tively uniform distribution of spray water across the cross-sectional
gas-flow area of vertical towers served by multiple levels of sprays gener-
ally enhances gas flow distribution at the mist eliminator inlet. However,
the liquid loading is generally high, perhaps higher than 1 to 2 gpm per
square foot in some vertical tower designs, and efficiency potential and
liquid handling capacity of alternative mist eliminator types have become
significant considerations in establishing the design and staging of mist
eliminator facilities.
Mist Eliminator Inlet Loadings
Inlet loadings are particularly high in limestone scrubbing systems
designed for high SO removal efficiencies in high sulfur service with
high liquid to gas flow ratios. With high inlet liquid loading eliminator
performance will generally require the inertial design to be based on use of
elevated superficial gas velocity in order to achieve high-efficiency liquid
removal. Advanced vertical-gas flow mist eliminator designs are adequate
when meeting the outlet loading specification requires limit drop size no
lower than 35 microns. Corresponding horizontal-gas-flow designs can be
employed to use gas velocity sufficient for limit drop size as low as 15 to
20 microns. Moreover, generous provision is made to give adequate drainage
of the liquid catch so as to avoid reentrainment.
11-50
-------
Influence of Gas Velocity on Mist Eliminator Performance
The mist eliminator removes droplets by use of dynamically-induced
inertial forces that move the liquid out of the conveying gas flow channels.
Thus it becomes more efficient at increasing inlet face velocities. This is
expressed quantitatively by the expression for the removal efficiency of an
inertial mist eliminator:
E=l-e-Z
...,, , x 0.5
Where z = cd v
and where:
d = mass median droplet size of the inlet distribution
v = gas velocity
c and x are empirically determined values.
Typical Design — The generic chevron arrangements that are extensively used
in FGD service (Figures 1 and 2) are typically designed for nominal inlet
face velocities of about 500 feet per minute yielding a removal efficiency
no higher than about 98.5% at d =150 microns. The high-performance
vertical flow designs (Figure 7J, use design gas velocities up to 1100 feet
per minute, and at the same value of d achieve removal efficiencies
exceeding 99.8%. The newest horizontal* flow designs (Figure 6), use design
Figure 1. High-Performance Vertical-Flow German Mist Eliminator Design
velocities up to 1300 feet per minute, and comparable removal efficiencies
are as high as 99.99%.
Maximum Feasible Velocities — At a given inlet liquid loading, increasing
inlet face velocity requires increased capability for drainage of the liquid
catch in order to avoid significant reentrainment. With chevrons reentrain-
ment takes place at 600 to 800 feet per minute velocity. Because of high-
11-51
-------
velocity pickup of liquid (rather than drainage limitations) the high-
performance, aerodynamically designed elements using drainage columns will
generally encounter reentrainment above 1200 to 1400 feet per minute in
vertical-gas-flow design and above 2000 feet per minute in horizontal-gas-
flow design.
Gas Flow Distribution
Uniformity of gas flow distribution across the inlet face of the
eliminator is critical to overall droplet collection capability. Case
histories are replete with instances of poor performance coupled with
mal-distribution, e.g., side-entry absorber towers with insufficient in-
ternal baffling primarily due to a lack of or inadequate gas flow modeling
at the equipment design state. Inlet gas distribution is critical because
at the comparatively low gas pressure-drop, e.g. generally less than 0.5
inches W.G., at which many mist eliminators operate, the eliminator is,
itself, not an adequate gas-flow distribution means and cannot compensate
for extreme gas-velocity gradients. Unlike other particulate collection
devices, its performance is adversely affected by both very reduced gas
velocity (e.g., less than 400 feet per minute) and very elevated velocity
(e.g., more than 1200 feet per minute in vertical-gas-flow designs.) For
optimum design to ensure that the inherent capability of the mist eliminator
is utilized, the RMS (root mean square) deviation of the inlet velocity
profile should ideally be in the 20% to 30% (maximum) range. Note that
every zone in the mist eliminator face area at which the local velocity
exceeds the level at which reentrainment starts provides a "window" through
which excessive liquid discharge is emitted. Current stringent standards
for total solid particulate emissions make it necessary for the conceptual
design and geometry of the in-tandem absorber and mist eliminator to be
established in a coordinated manner so that acceptable system performance
may be realistically predicted and achieved.
Washing Technique
The key element in ensuring high mist eliminator reliability and sus-
tained performance, and in thereby limiting impact on FGD reliability, is
maintenance of day to day cleanliness of eliminator internals. Eliminator
cleanliness is critically influenced by the chemical analysis of the scrub-
bing slurry carried over to the mist eliminator. Mist eliminator fouling
with calcium sulfate scale is most effectively prevented in unsaturated-
CaSO -mode scrubbing processes, i.e., those with magnesia buffering or
that use sodium liquor scrubbing in a concentrated-active-alkali dual alkali
system, which have dissolved calcium concentration in scrubbing liquor at
the level of 170-200 ppm or less. Maintenance of an adequate level of
suspended solids in scrubbing slurry and good regulation of limestone feed
are important considerations in control of scrubbing slurry chemistry
affecting tendencies for eliminator scaling by calcium sulfate and calcium
sulfate, particularly in supersaturated-CaSO -mode limestone scrubbing
systems. Of special concern in this connection are non-steady state, i.e.
transient, conditions such as during load changes and at startup and shut-
down that lead to elevation of scrubbing slurry pH above the set point. Use
of the minimum pH in the absorber consistent with SO removal requirements
11-52
-------
will help ensure the desired high limestone utilization. In-situ mist
eliminator washing, when carried out effectively, provides an adequate means
of removing deposits that would otherwise occur due to the presence of
collected scrubbing slurry at the surface of the eliminator internals. This
is best achieved when a maximum amount of FGD system makeup water is allo-
cated to the washing activity. Constraints on process liquid discharges and
on acceptable rates of FGD liquid effluent-outfall normally prevent generous
use of fresh water for in-situ mist eliminator washing to ensure good
surface cleanup and call for washing schemes tailored to site-specific
conditions.. Based on the authors' review of available details of existing
installations to be assessed in the pending EPRI investigative program noted
earlier, washing effectiveness (together with gas flow distribution) may
well be the key to troubleshooting success. This is particularly so in
unsaturated-CaSO,-mode FGD systems, which lend themselves to effective
washing due to the less troublesome chemistry of the carryover droplets as
noted earlier and due to the potential for satisfactorily using low-calcium
CaSO,-unsaturated thickener overflow liquor as part of the wash water
supply. Moreover, it is clear that an assessment of the chemistry and
nature of absorber emissions, unsaturated-CaSO -mode FGD vs. supersatu-
ratedmode, performed in conjunction with an evaluation of comparative ease
of mist-eliminator cleaning by in-situ washing, gives an unusually clear
insight as to the principal reason(s) for poor FGD reliability experienced
through 1982 in natural-oxidation supersaturated-mode high-sulfur FGD
installations throughout eastern U.S.A.
Placement of In-Situ Wash Water — Wash water is applied at the inlet face
but not the final exit face of the eliminator facility. Systems incorporat-
ing two mist eliminator stages are preferentially washed on both upstream
and downstream faces of the first stage and upstream, only, on the second
stage.
Eliminator Washing Technique for Supersaturated-Mode Absorbers with Low
Limestone Utilization — Extensive experience has shown that absorber towers
with a high concentration of limestone in the scrubbing slurry, i.e., less
than 85% limestone utilization, should be provided with continuous mist
eliminator wash (13). Vertical gas flow mist eliminators require 0.5 gpm
per square foot, horizontal flow 0.75 gpm per square foot.
Washing When Limestone Utilization is High — With limestone utilization 85%
or greater mist eliminators should be washed intermittently, preferably with
fresh or low total-dissolved-solids makeup water, 1.5 gpm per square foot
for vertical flow, 3.0 gpm per square foot for horizontal flow.
Recycling vs. Once-Through Washing Systems — In mist eliminator systems
that permit segregation of used washwater and the droplet catch from scrub-
bing slurry, such as the horizontal-flow type, recirculated wash systems may
be advantageously used. Makeup water should be comparatively fresh water,
i.e. with total dissolved solids less than 1,000 ppm, and should ideally be
fed at a rate sufficient to limit recirculating-water dissolved calcium
level to 150 ppm and relative calcium sulfate saturation to 20%. In using
once-through systems, particularly if employing recycled thickener overflow
in supersaturated-mode systems, treatment with soda ash is effective in
11-53
-------
removing most of the calcium sulfate saturation in the water. The tolerable
relative saturation in once-through wash liquid supply cannot be explicitly
forecast as it is dependent on the type and amount of total dissolved
solids.
Washing Spray Design — Mist eliminator washing systems normally employ
solid cone sprays spaced so as to provide overlapping spray patterns and
designed for nozzle pressure of 30 to 50 psig. In systems where intermit-
tent spray cycles are specified, sequence timers should provide the capabil-
ity for "on" cycles of 5 to 10 minutes, "off" cycles of 20 to 60 minutes.
Typically the first stage should be washed 5 minutes on, 20 to 30 minutes
off, the second stage 5 minutes on, 60 minutes off. The adjustable timers
should permit the cycles to be lengthened or shortened to suit operating
conditions. Advisedly the wash cycle should be started at the typical rates
and then be adjusted only after the system operation has been stabilized
over several months. This permits use of sufficient operating experience
needed to properly judge the effects of adjusting the cycle. Each adjust-
ment should be followed by stabilized operation for several months duration
to afford an adequate observation period.
Droplet Emission Testing
Measurement of must eliminator performance is of importance both for
the purpose of determining liquid loadings being admitted to flue gas reheat
systems or stacks and to quantify the contribution of liquid-carryover
solids, dissolved and suspended, to the total solid particulate emission
from the stack. In addition, in many instances the determination of the
droplet size distribution is of equal importance. However, the flue gas is
water saturated in the region of the absorber train in which the eliminator
is positioned. Thus, significant heat losses to the ambient cause water
condensation in the FGD system as well as in any gas sample trains used to
withdraw and analyze gas streams. This can complicate field testing to
characterize the liquid loading in gas streams entering and leaving mist
eliminators and requires selected techniques to accomplish such measurements
in an accurate manner.
Candidate Test Methods — For measurement of liquid loading (14) without
need to determine droplet size distribution the tester may possibly use the
throtting calorimeter, heated probe, double calorimeter, condensation method
or cyclone separator. When both loading and distribution measurements are
required the impactor method, light scattering probes, or laser probes may
possibly be used.
Criteria for Choice of Test Method — The testing objectives that are most
critical in selecting a method of carryover measurement are as follows:
- Drop size measurement discriminates among droplet diameters from 1
micron to several hundred microns.
- The method quantifies droplet concentration in the gas.
- Testing can be carried out in water saturated or unsaturated gas.
- Liquid loads in the range from a few milligrams per cubic meter to
several thousand milligrams per cubic meter may be quantified.
- The measuring device must be readily portable at industrial sites.
11-54
-------
It must be possible in a simple way to sample and test gas across
the cross-sectional flow area of large diameter vessels.
Preferred Test Methods — All criteria, above, for selecting mist testing
equipment cannot be met by any single method. Experience has demonstrated
the advantageous use of two specific complementary methods/devices for
measurement of size distribution and loading employing compactors — an
impactor probe of special design and a cascade impactor. The impactor
probe, applicable to the size range of 15 to 2,000 microns, is a plate
coated with a special film such as magnesium oxide that is inserted into the
flowing gas stream via a shield-tube enclosure fashioned with a movable
shutter to permit timed exposure of the plate surface. A cascade impactor,
applicable to the size range of approximately 0.4 to 10 microns, comprises a
group of staged in-series nozzles behind each of which is a removable
thin-glass baffle (collecting) plate and through which nozzles an isokineti-
cally sampled gas stream is drawn by a vacuum pump.
Tracer Type Testing — In the special case of sodium liquor scrubbing
systems such as dual alkali FGD a dry-filtration type collection in a filter
and impingers of an isokinetically sampled gas stream (15) can be used to
accurately quantify process liquor carryover loading by analysis of sodium
content of the catch and relating it to the sodium ion concentration in the
FGD recirculating liquor.
Materials of Construction
Mist eliminator users are increasingly emphasizing the need for capa-
bility to withstand upset or transient temperatures ranging as high as from
350 F. to 700 F. This has brought about rapid development and use of
newer high-temperature plastics as well as high-alloy metals.
Materials with Modest Temperature Capability — The lower temperature
plastics include PVC (polyvinyl chloride), polypropylene and phenyele oxide
based material. The latter contains amounts of polymers such as styrene
added to make it more extrudable or moldable and is made in various grades
identified by heat deflection temperatures as high as 300 F. and more.
However, when subject to stress and heat in actual service, the maximum
service temperature is considerably less than the heat deflection rating.
Materials for Elevated Temperatures — Particularly as a result of proximity
of heat sources such as flue-gas reheat means, positioned close to the mist
eliminator in some applications, high temperature plastic materials such as
FRF (fiberglass reinforced polyester) are increasingly being specified and
used for mist eliminator internals. FRP can typically operate up to approx-
imately 350° F., and in some cases, with some charring, has tolerated
temperatures approaching 500 F. The pultruded blade (profile) forms
typically contain a mixture of premium-grade isophthalic polyester resin,
alumina/silicate filler and approximately 5% antimony trioxide serving as a
fire retardant. The addition of a synthetic cloth surfacing veil provides
protection for the glass reinforcing fibers.
11-55
-------
PRESENT PRACTICES
OUTLOOK FOR FUTURE FGD SYSTEMS APPLICATION
The strong trend in U.S. since 1978 toward application of forced-
oxidation type calcium-based FGD systems in preference to previously common
natural-oxidation supersaturated-mode types may be expected to somewhat
reduce absorber-emission related impacts on mist-eliminator and FGD reli-
ability. Nonetheless the forced-oxidation processes do also, to a degree,
operate in the supersaturated mode. Those being applied in U.S.A. may be
expected to pose significant fouling problems for the eliminator step and
low-pH corrosion and fouling immediately downstream of the mist eliminator
due to many circumstances not encountered in the high-reliability
unsaturated-mode systems previously discussed. These include:
- Presence of unreacted lime/limestone in carryover liquid causing
fouling, particularly when the goal of high system limestone utili-
zation is achieved by use of a precooling or other remote scrubbing
step
Comparatively high carryover rates due to the high absorber
liquid/gas flow ratios required for SO removal and/or gypsum
scale control in the absorber causing fouling/corrosion
- Comparatively high outlet flue-gas SO concentration due to inher-
ently lower SO removal efficiency causing corrosion
- Potentially higher outlet flue-gas HC1, HF, NO , and
SO /H SO concentration (due to a possibly lower absorption
mass-transfer capability) causing corrosion.
CONTINUED MISAPPLICATION OF INERTIAL MIST ELIMINATOR TECHNOLOGY
Notwithstanding the critical need for improved standards of eliminator
performance and reliability, current mist-eliminator specifications of
architect engineers and FGD system suppliers frequently call for use of very
low design gas velocity, e.g., no greater than 480 feet per minute with no
exception allowed, thus sustaining the use of the simple low-efficiency
chevron designs. In some cases it appears that such requirements are set by
structural engineering considerations rather than the needs of the process
design. Further, in some instances where somewhat higher design gas velo-
city is used, chevron-type equipment is employed at design gas velocity
constituting 80 to 85% of the critical (reentrainment threshhold) velocity.
Performance of such an installation is drastically affected by wide devia-
tions in local gas flow rate. Thus, the lack of attention to or the disre-
gard for optimal mist eliminator application will continue to result in
operating FGD installations that fail to meet system performance and reli-
ability objectives. Success of an FGD system is obviously dependent on
close coordination between the Engineer responsible for system design and
the mist eliminator applications-specialist/supplier.
IMPROVED INSTALLATIONS FOR REGENERATIVE AND DIRECT STEAM REHEAT
^Control of scrubber carryover plays a direct role in all FGD systems in
helping to limit fouling and corrosion of downstream components handling wet
11-56
-------
gas. Moreover, mist eliminator performance and reliability is particularly
critical in achieving dependable operation of flue-gas reheat heat-exchange
equipment that is in direct contact with the scrubbed gas. Use of high-
efficiency mist eliminator design to minimize droplet carryover (in con-
junction with employment of in-situ high flow-volume steam or air blowing of
the most exposed portions of the gas-wetted heat exchange surface to period-
ically remove deposits) has been successful in limiting exchanger corrosion
and maximizing heat transfer.
Regenerative Reheat
In current application of FGD in high-fuel-cost nations like Japan and
West Germany growing use is being made of means of utilizing heat in raw
flue gas exitting electrostatic precipitators for reheating of water-
saturated flue gas from wet scrubbers. Ljungstrom rotary heat exchangers
are applied to achieve major plant lifetime-savings by elimination of use of
reheat heat-sources requiring firing of additional fuel. A tubular-
exchanger type regenerative reheat system has been utilized in U.S.A. in an
FGD facility now in initial startup. Like the Ljungstrom units, this
first-of-a-kind facility indirectly cools the raw flue gas below its sulfu-
ric acid dewpoint temperature in extracting "free" heat for flue gas reheat.
TVA, Paradise Steam Plant — The regenerative-reheat-equipped 704 mW Para-
dise Unit No. 1 high-sulfur forced-oxidation FGD system, Drakesboro,
Kentucky, has been undergoing startup and commissioning throughout 1983. A
two-stage high (1500 feet per minute) velocity horizontal-flow mist elimina-
tor operating at a 120° F. water-saturated condition specified for approxi-
mately 0.07 grains/SCF (dry) outlet liquid loading (16 micron limit drop
size) is utilized immediately upstream of a Type 317L tubular stainless
steel reheat exchanger designed to cool recirculating regenerative-reheating
water to 188° F. A downstream tubular carbon steel second-stage reheat
exchanger is designed to receive the hot water supply at only 200°F. and
further reheat the flue gas to a reheater outlet temperature of 170° F.
Although only approximately 50% of regenerative flue-gas reheat capacity has
been realized in the 1983 operation to date, the mist eliminator operation
and use of retractable steam blowers at the reheat exchanger has prevented
any deterioration of the stainless steel tubes. Because of the temporarily
inadequate reheating, the carbon steel reheat exchanger tubes have been
attacked by sulfuric acid deposition for which the system was not designed.
Due to inadvertant overheating of shut-down absorber modules by hot water
allowed to flow through the reheat exchangers, meltdown of mist eliminator
internals has occurred on several occasions, which may require replacement
with FRP.
NWK, Wilhelmshaven Power Plant — The Ljungstrom rotary-regenerative re-
heater equipped 350mW No. 2 FGD module at the low-sulfur Wilhelmshaven, West
Germany, plant has been in sustained high-availability operation since early
1982. A two-stage high-velocity horizontal-flow mist eliminator operating
upstream of the regenerative exchanger at a 115° F. water-saturated condi-
tion was specified for approximately 0.09 grains/SCF (dry) outlet liquid
loading (20 micron limit drop size) and has been discharging only 0.03
grains/SCF (dry) by test. The system design and initial operation provided
11-57
-------
for 10 to 15° F. reheat by flue-gas by-pass upstream of the Ljungstrom,
which heats the flue gas an additional 90° F. The prior reheating is done
so as to help ensure against excessive sticking of mist eliminator carryover
slurry in the internals of the regenerative exchanger. In a current trial
run with by-pass reheat shut off, the FGD system had at last report been
operating for eight weeks without any problems with regenerative reheat.
Direct Reheat
Direct reheat of flue gas is accomplished by use of an in-line tubular
heat exchanger that is installed directly in the scrubbed flue-gas stream
and heats the gas by indirect heat exchange with steam or hot water. The
heat exchanger is particularly vulnerable at its gas inlet end where the
gas-wetted surface is subject to low-pH attack. Crevice corrosion is
especially troublesome if these surfaces become continuously fouled with
solids originating from scrubbing slurry carryover droplets. In a number of
installations improved mist eliminator designs have been successful in
limiting carryover flow and, with periodic tube blowing, avoiding signifi-
cant reheater corrosion.
Arizona Public Service Co., Cholla Station — No. 2 FGD system, low-sulfur
coal service, utilizes a modern-design vertical flow mist eliminator as in
Figure 7 to protect a steam heated flue-gas reheat exchanger that contains
Incoloy alloy 825 tubes. In conjunction with intensive periodic steam
blowing of the inlet face of the reheater bundle, the mist eliminator has
been fully effective in preventing fouling and corrosion of the reheater.
Montana Power Company, Colstrip Station — Utilizing a wash tray with a very
high (3" to 5" W.G.) gas-pressure drop ahead of chevron mist eliminators the
low-sulfur FGD systems at Colstrip have operated without significant fouling
or corrosion of the steam-heated direct-reheat exchangers, which are plate
coil type constructed of Inconel alloy 625 (first stage) and Hastelloy alloy
G (second stage). As in the case of Cholla Station, intensive periodic
steam blowing of reheater internals are essential to sustained reheater
reliability.
RETROFIT UPGRADING OF COMMON LOW-EFFICIENCY MIST ELIMINATORS
In numerous instances, and in order to upgrade performance, reliability
and temperature rating the advanced vertical-gas-flow mist eliminator design
(Figure 7) has been retrofitted in FRP to existing modules utilizing space
originally occupied by chevron type installations.
Big Rivers Electric Cooperative, Green Station
The original chevron elements suffered from meltdown during brief shut-
down periods due to proximity to the flue-gas reheat heat source. In retro-
fitting the advanced vertical-flow design, area-by-area replacement was
required by the Engineer and, hence, no increase in superficial gas velocity
nor advantageous reduction in limit drop size was accomplished. However,
field observations indicated that the appearance (opacity) of the stack-
discharge trailoff was improved.
11-58
-------
Duquesne Lighting Company, Phillips Station
A similar replacement was made at Duquesne. Operation is understood to
be satisfactory.
Louisville Gas and Electric Company, Cane Run Station
Retrofitting of the same advanced design together with a compatible
bulk-removal mist eliminator (Figure 4) upstream of it was motivated, in
part, by incidents in which prior chevron elements melted down during
shutdown periods due to proximity to the reheat heat source. Through an
eliminator washing improvement program a substantial increase in sustained
performance and reliability has been achieved.
CONCLUSIONS
1. Mist eliminator selection and design as well as and washing tech-
nique are major factors influencing the performance and reliability of FGD
systems.
2. Mist eliminator technology advancements in design and forming of
aerodynamically shaped, extruded profile elements for use in either a
horizontal or vertical-gas-flow regime offer means of gaining potential
performance substantially greater than that of chevrons.
3. Absorber design provisions for gas flow distribution are of criti-
cal importance in realizing the performance capability of an eliminator
system.
4. Satisfactory mist eliminator performance and reliability will be
best achieved when the designs of the absorber system and of the mist-
eliminator system are coordinated through the integrated efforts of the wet
scrubber designer and mist eliminator supplier/specialist.
5. Traditional use of low-velocity chevron mist eliminators continues
and is judged to be an important contributing factor in many low-pH-end FGD
reliability problems and in instances of poor or marginal gas cleaning
performance.
6. Conservative, high-performance mist eliminator design and effective
eliminator washing operation are of critical importance in supersaturated-
CaSO -mode FGD systems, both natural and forced oxidation, in view of
serious impact on FGD system reliability of the absorber-emission species
characteristic of this type of process.
REFERENCES
1. Ellison, W. Scrubber Demister Technology for Control of Solids Emis-
sions from SO Absorbers. (Presented at EPA Symposium on Transfer
and Utilization of Particulate Control Technology. Denver, Colorado.
July 24-28, 1978.)
11-59
-------
2. Ellison, W., and S. A. Lefton. FGDS Reliability: What's Being Done to
Achieve it. POWER, May 1982.
3. Moen, D. A. et al, Coal Creek Station Air Quality Control System.
(Presented at 29th Annual Conference of the Association of Rural
Electric Generating Cooperatives. Vail, Colorado. June 11-14, 1978.)
P 6,7.
4. Calvert, S. Guidelines for Selecting Mist Eliminators. Chemical Engi-
neering, February 27, 1978. p 110.
5. Laseke, B. A., and T. W. Devitt. Status of Flue Gas Desulfurization
Systems in the United States. (Presented at 29th Annual Conference of
the Rural Electric Generating Cooperatives. Vail, Colorado. June
11-14, 1978.) p 33-35.
6. Tennyson, R. P. Mist Eliminator Design and Application. (Presented at
70th Annual Meeting of the Air Pollution Control Association. Toronto,
Ontario, Canada. June 20-24, 1977.) p 77-25.4.
7. Pedersen, G. C. Experiences with Control Systems Using a Unique
Patented Structure. Kimre, Inc. May 1983.
8. Bakke, E., and H. P. Willett. The Application of a Tubular Wet Elec-
trostatic Precipitator for Fine Particulate Control and Demisting in an
Integrated Fly Ash and SO Removal System on Coal Fired Boilers.
(Presented at E.P.A. Symposium on Particulate Emission Control. 1982.)
9. Reference 1.
10. Henzel, D. S., and D. H. Stowe. A Proven Reagent for High Sulfur Coal
Flue Gas Desulfurization. (Prepared for the EPA/EPRI Flue Gas Desul-
furization Symposium, Hollywood, Florida. May 17-20, 1982.)
11. Electric Light and Power, Technical Publishing Division, Dun-Donnelley
Publishing Corporation. (Overview of EPA/EPRI FGD Symposium) Table 5
Availability FGD System Records for Plants Burning 3.0% or Greater
Sulfur. June 1982.
12. Tampa Electric Company, letter of September 2, 1983.
13. Henzel, D. S. et al, Limestone FGD Scrubbers: Users Handbook. PEDCo
Enviromental Inc. and Black and Veatch. EPA-600/8-81-017. December
1981. p 3-37 to 3-41.
14. Private Communication from E. Reinhard, Euroform GmbH, March, 1983.
15. Johnson, L. D., and R. M. Statnick. Measurement of Entrained Liquid
Levels in Effluent Gases from Scrubber Demisters. EPA-650/2-74-050.
June 1974.
11-60
-------
FGD GYPSUM: UTILIZATION VS. DISPOSAL
By:
William Ellison, PE
Ellison Consultants
Monrovia, Maryland 21770
ABSTRACT
The purpose of this paper is to give a technical and economic evalua-
tion of alternative means of managing wastes collected in forced-oxidation
(FO) type flue gas desulfurization (FGD) systems. Worldwide application of
FO-FGD is described and environmental considerations in selection of waste
management alternatives - gypsum utilization vs. disposal - are reviewed.
Feasibility of and barriers to commercial gypsum from FGD in North America
during the 1980s are analyzed. The impetus, method of implementation, and
potential for gypsum use are addressed for major by-product gypsum FGD
installations now in operation, under construction or in the design phase.
It is concluded that in many instances, the production and sale of usable
gypsum from FGD would be a benefit and a source of direct profit to the
utility plant owner, but that in most cases the gypsum depletion allowance
under IRS Code 613B will continue to be a major disincentive for purchase of
by-product gypsum by vertically integrated gypsum companies in U.S.A.
INTRODUCTION
PRODUCTION OF GYPSUM BY FORCED OXIDATION
Gypsum, double-hydrated calcium sulfate, is a naturally occurring non-
metallic mineral used as a raw material in the manufacture of gypsum board,
Portland cement, plaster products and in agriculture. Forced oxidation as
applied to wet lime/limestone flue gas desulfurization (FGD) consists of
forcing air into the spent scrubbing slurry to produce gypsum, the oxidized
form of calcium sulfite. The gypsum can then be easily dewatered to a cake
containing greater than 80% solids by weight. This operating technique is
commonly used worldwide. North America is the last major FGD-using area to
adopt FO operation beginning in 1978, bringing about a major departure from
the natural-oxidation design and operation of FGD facilities. Since that
time, approximately 20 power plants have purchased provision for FO in FGD
operations constituting 16,000 mW of scrubbed generating capacity and almost
half of all limestone FGD committed to date.
11-61
-------
ADVANTAGES OF FORCED OXIDATION
This SO removal method is superior to natural-oxidation FGD commonly
applied in U.S.A. since it provides improved control of FGD-process operat-
ing conditions, which may result in higher FGD availability in limestone
based SO absorption. It results in improved control of internal scaling,
better utilization of limestone chemical reagent and enhanced SO removal
efficiency.
Moreover, the solid waste produced from the SO -catch is in the form
of comparatively coarse-grained calcium sulfate solids that may be more
easily dewatered, transported and stored and which require less than 60
percent of the solid waste disposal area required for direct ponding of
sludge without forced oxidation. Additionally, gypsum formed by FO-FGD has
potential for use in gypsum board manufacture, cement manufacture and
agriculture.
MANNER OF WORLDWIDE IMPLEMENTATION
On a worldwide basis, wet scrubbing using FO of the S0»-catch is the
principal means of desulfurization of flue gas (FGD) from electric utility
boilers. FGD waste solids, initially generated as a slurry, are subse-
quently dewatered to a sludge or cake containing both solid and liquid waste
forms. The decanted liquid and filtrate is generally returned to the
process. However, build-up of dissolved solids in the scrubbing slurry may
preclude closed-loop operation resulting in a wastewater effluent requiring
disposal. Production and use of salable gypsum avoids discarding of FGD
solid waste at a permanent disposal site.
The production and utilization of a usable grade of gypsum from FO-FGD
is routinely done in Japan and West Germany, the two principal countries
that have applied flue gas desulfurization overseas. However, in view of
the limited number of usable-gypsum type limestone-scrubbing systems that
have been purchased in North America (see Table 1), we continue to be
principally oriented to throwaway-solid-waste FGD operation even after the
transition in the past five years to use of forced oxidation instead of
natural oxidation.
DISPOSAL OF FGD GYPSUM
The disposal of FGD wastes, both dewatered solids and liquid effluents,
has the potential for pollution of groundwater and surface water bodies
impacted by waste disposal activities.
WASTE SOLIDS
Land-based disposal of scrubber solids from solids-precipitation-type
FGD systems is a problem principally because of the very large waste quant-
ity generated. At the same time, EPA regulatory activity is concerned with
the possible long-term effects leading to pollution of surface water and
groundwater bodies due to the leaching of constituents such as trace-metals
11-62
-------
TABLE 1. COMPARISON
OF COMMITTED FO-FGD INSTALLATIONS YIELDING
SALABLE GYPSUM VS. THROWAWAY SOLID WASTE
Power Plant mW % of Total
AESC, Cogen. 1,
Houston, Texas 150 (Approx.) 1.0
JEA, SJRPP 1 and 2,
East Port, Florida 1,200 7.7
MPW, Muscatine 9,
Muscatine, Iowa 166 1.1
Tampa Electric, Big Bend 4,
Tampa, Florida 475 3.0
Sub-Total, Salable Gypsum Producing: 2,991 12.8
Aggregate of FO-FGD Throwaway
Waste Installations: 13,517 87.2
Total FO-FGD Committed in U.S.A. 16,508 100.0
from these waste solids. However, hydrogeology of typical sites that may be
used for storage or ultimate disposal of wastes from such high-volume
sources can be complex. Thus, for an individual site there may be uncer-
tainties as to the extent of adverse environmental impact due to permeation
of leachate through the discarded waste mass and the underlying strata. The
two main disposal options for discarding solid waste generated by FO-FGD are
landfilling and wet stacking.
Landfilling
FO-FGD operation yields typical gypsum crystal size of from 40 to 100
microns, producing an easily dewatered and handled coarse-grained cake that
can, if necessary, be readily transported as an 80-90% solids mass to a
remote disposal site. However, an implication of this low moisture content,
(and the inability of the discarded coarse-grained waste to retain any
significant amount of FGD liquor over an extended period), is that little of
the scrubber liquor can be disposed of by occlusion in the solid waste mass.
Consequently, the operation of closed loop FO-FGD may not be possible
because of an increase in system corrosion resulting from a greater than
10-fold increase in FGDliquor dissolved solids concentration e.g., to in
excess of 100,000 ppm chloride in the case of a medium/high sulfur coal
application (1).
Environmental concerns associated with landfilling of FO-FGD waste
solids include:
11-63
-------
- Water-soluble material (dissolved metals and salts including calcium
sulfate) is present in seepage and runoff.
After the lifetime of the FGD operation, seepage and runoff of
leachate will continue unabated with no means of assimilating it by
recycle operation. (The substantial solubility-in-water of the
gypsum solids will yield an endless source of leachate saturated
with calcium sulfate and containing other components such as sulf-
ite, dissolved metals and salts.)
Wet Stacking
Based on field tests (2), FO-FGD solid waste may instead be disposed of
by wet stacking. This is a technique utilized extensively in the phosphoric
acid industry in Florida for land-based disposal of waste-gypsum, wherein a
10 to 15% solids aqueous-slurry is discharged at a disposal site so as to
permit formation of a sharply-pitched stack over the plant lifetime. In
this method to be applied at a number of U.S. power plants, reduced acreage
use is achieved through such stacked disposal without need for dewatering by
thickeners and filters or centrifuges. Overall capital costs and annual
revenue requirements are understood to compare favorable with even simple
ponded-waste natural oxidation FGD systems. In the so-called "upstream
method" of construction, the wet stacking is begun by first constructing an
earthen starter-dike forming a sedimentation pond. FO-FGD solid waste is
pumped to the pond in a slurry form, usually at 10 to 20 percent solids, and
allowed to settle and drain (3). Process water is decanted from the pond
and returned to the plant. When sufficient gypsum is deposited within the
pond, gypsum is excavated with a dragline to form and step-wise build up a
perimeter dike immediately inside the starter dike. Decanted liquor is
piped under the starter dike to an annular perimeter ditch and surge pond
formed by the starter dike and a clay exterior dike. The process of
sedimentation, excavation, and raising of the perimeter dikes continues on a
regular basis during the active life of the stack. Using the upstream
method of construction, gypsum stacks in the phosphate industry have reached
heights exceeding 100 feet (30.5 m) with slopes of 1.5 horizontal to 1.0
vertical, which is approximately the a.r.gle of repose of some gypsums. These
steep slopes result from casting the gypsum with a dragline and allowing
some gypsum to roll down the outside of the stack to eliminate need for
shaping. Therefore, the gypsum perimeter dikes of some stacks have a factor
of safety very close to unity and from a conventional geotechnical engineer-
ing point of view, failures of gypsum stacks sometimes occur. Fortunately,
gypsum is a very forgiving material, and, unlike most mine tailings, gypsum
does not readily flow. Therefore, the consequence of a failure is the loss
of process water stored on the gypsum stack. If process water escapes the
plant property, liabilities from pollution and environmental damage may also
result. On a day-in, day-out basis, seepage of process liquid to ground-
water and surface water should be controlled by use of an impervious founda-
tion as well as of clay-type exterior and starter dikes, and by collecting
and treating or recycling all seepage and runoff. Although a greater
occlusion of process liquor in solid waste occurs in wet stacking than in
landfilling, in situ washing of the stacked solids will occur (due to the
high permeability of the stack). This leachate when subsequently collected
11-64
-------
and recycled would be expected to lead to a salt concentration build-up even
greater than that which would occur using in-plant dewatering and landfill
disposal. This would most certainly preclude closed loop operation of the
FGD system in the case of wet stacking.
Environmental concerns in use of wet stacking for a FO-FGD waste
include:
- The very large working area of the stack facility will cause the
liquor recycle to FGD (returning from the stack-pond section) to
also return large quantities of liquid running off from the entire
stacking facility. Resulting water-balance problems in rainy
weather may lead to intermittent outboarding of substantial dis-
charges of process-liquor.
- All pile run-off will be saturated with calcium sulfate and contain
gypsum surface-moisture constituents (dissolved metals and salts).
- Over a twenty-year lifetime, a large stack in excess of 100 feet
high may be created, e.g., an increment of 40 mW, 3 percent sulfur
bituminous coal, will generate a simple truncated conical stack, 400
feet in base diameter, 100 feet in diameter at the top, 100 feet
high, with side slopes of 1.5 horizontal to 1.0 vertical. In no
country are regulations or experience available as to how to satis-
factorily decommission a lifetime gypsum stack or how to reclaim it.
Stack slopes are too great to hold topsoil and vegetation. Thus,
normal concepts for "covering" of the waste in this manner do not
apply.
- After the FGD lifetime, seepage of leachate will continue unabated
with no user to which to recycle for re-use. The substantial
solubility-in-water of the gypsum solids will yield an endless
source of leachate saturated with calcium sulfate plus other compo-
nents such as sulfite (chemical oxygen demand — COD) plus any
remaining surface-moisture constituents.
- Cold winters may severely restrict dragline operations during cold-
weather months and jeopardize continuity of FGD operation. In addi-
tion, freeze-thaw action is of concern so far as maintaining the
integrity of dikes.
LIQUID EFFLUENT
In view of the low surface-moisture content of coarse-grained solid
waste from FO-FGD systems, provisions for management of saline FGD liquor
blowdown are essential for pollution control. Thus, the FO-FGD system
designer would be expected to seek all practical design means to manage
waste effluent such that acceptably low scrubber-liquor salinity is main-
tained (4) .
11-65
-------
Significance of Discharge
Although FO-FGD blowdown may, by volume, be only a small proportion of
the total power-plant wastewater, it contains dissolved-salts concentrations
approaching those in sea water and thus may exceed the assimilative capacity
of available receiving streams. Such outfall, untreated, can also contain
objectionable concentrations of trace metals such as arsenic, mercury,
cadmium and lead that may originate from the fired coal (5).
It should be noted that liquid purge outfall such as required from
FO-FGD is not always necessary in a natural oxidation mode because these
latter systems yield a solid waste (primarily calcium sulfite) that is
fine-grained, thixotropic and highly water retentive (containing approxi-
mately 50% moisture). The encapsulation of a substantial amount of FGD
liquor in the discarded waste mass may provide the liquid purge necessary to
maintain the scrubber dissolved salts at an acceptable concentration.
However, natural oxidation sludges have many disadvantages as compared to
those produced in a forced oxidation mode. These disadvantages include both
physical properties — the completed disposal site cannot be reclaimed for
uses requiring the soil to support structural loads — and chemical proper-
ties. The principal contaminants in the sulfite sludge are dissolved salts
(principally sulfates), chemical oxygen demand due to sulfites, and soluble
heavy-metal trace elements originating in the coal. Thus leaching and
runoff from sites used for disposal of such solids have also been of sig-
nificant environmental concern. An important remedy is sludge fixation
(stabilization), which consists of chemical processing of scrubber sludge to
facilitate handling, transportation, placement, and consolidation at the
ultimate disposal site. One of the methods used is pug-mill blending and
aging of mixtures of FGD filter cake, dry fly ash and other dry additives,
followed by landfill compaction. An alternative procedure involves the
addition of silicate-containing materials to thickened FGD slurry before
discharge to a pond site. Extensive testing indicates that these methods of
chemical treatment significantly improve the load-bearing characteristics of
FGD sludge, reduce the solubility of the major chemical species by a factor
of 2 to 4 and reduce sludge permeability by an order of magnitude or more
(6). Further, as outlined previously, the encapsulation of a substantial
amount of FGD liquor in the consolidated waste mass from natural oxidation
FGD provides the option for all-in-one, combined disposal of all solid and
liquid waste without discharge of a liquid outfall.
Alternatives for Treatment/Ultimate-Disposal
The liquid purge may be treated to precipitate trace-metal constituents
and reduce their concentration to an acceptable level so that the effluent
stream may be discharged to a suitable receiving stream or water body. Such
treatment followed by outfall discharge is the economically preferred method
of FGD liquor disposal when it can be permitted.
Candidate wastewater treatment methods to reduce trace metal concen-
trations include:
11-66
-------
Chemical precipitation
Ion exchange (exchange of complex anionic species)
Sulfide precipitation
Chelation/absorption.
As indicated above, because of the large dissolved solids concentration
of purge-liquor, available surface water receiving streams at inland sites
may be unable to tolerate the environmental impact of saline outfall dis-
charge of this type. Blowdown liquor after treatment can, in this case, be
concentrated utilizing commercially available equipment to crystallize
dissolved salts to a comparatively small mass of salt cake that can be
discarded at a secure landfill disposal site or used as a rock salt by-
product.
UTILIZATION OF FGD GYPSUM
INTRODUCTION
Annual consumption of crude gypsum in the United States is estimated to
be approximately 20 million tons, with 30 - 40 percent being imported,
mainly from Canada and Mexico. The average value of crude gypsum was
estimated to be $9.00 per ton (F.O.B. mine) in 1982 (7).
The sale and utilization of gypsum from FGD would help to offset a
small portion of the operating cost of the FGD system and would minimize the
environmental impact of and use of land for ultimate disposal of FGD solid
wastes. There are several factors to be considered in determining the fea-
sibility of FGD gypsum sale and utilization. These include technical and
economic aspects as well as incentives, problems and barriers which are
specific to the circumstances in the United States.
TECHNICAL FEASIBILITY
The gypsum formed by FO-FGD is unlike natural gypsum, the major differ-
ences being: bulk handling properties (moisture content, grain size), quan-
tity of impurities, and nature of impurities. Natural gypsum is mined as a
coarse, low-moisture rock at approximately 70 - 90 percent purity, with
limestone and/or anhydrite (non-hydrated calcium sulfate) comprising the
majority of impurities. FGD gypsum is produced as a higher moisture filter
cake of approximately 90-95 percent purity, primarily coarse-grained crys-
tals along with ash, lime or limestone components and dissolved salts
comprising the majority of impurities. The technology to make and use
board-grade and cement-grade FGD gypsum exists and is currently in use in
Japan and Europe. Depending on end use (cement, prefabricated products or
agriculture), the FGD gypsum may require processing (beneficiation) before
it is suitable for use. For example, gypsum is used in cement manufacture
both as a grinding aid and as a source of calcium sulfate to retard setting.
To be acceptable for this use, FGD gypsum cake would need only minimal
washing (to ensure chloride content of less than 1 percent), but requires
11-67
-------
drying/agglomeration to achieve suitable grinding properties. To be used in
gypsum board manufacture, FGD gypsum cake requires washing, (the amount
depending on chloride content, unwashed, and board manufacturers' specifi-
cation) and may require drying/agglomeration. Additionally, suitable FGD
process design and control is vital to ensure an adequate degree of oxida-
tion of sulfite.
ECONOMIC FEASIBILITY
There will be costs incurred by the utility in converting the system
design and ultimate operation from disposal to utilization of FGD gypsum.
The net cost to the utility will be the beneficiation costs less the hypo-
thetical disposal cost if the gypsum was not to be used. However, the
utility should be able to sell the gypsum, which may either offset the
production cost or result in a net credit to the utility. As described
earlier, natural gypsum has an average value of $9.00 per ton (F.O.B mine).
Thus the apparent value of gypsum to the user is typically $9.00 per ton
plus transportation costs. Transportation costs for gypsum imported from
Nova Scotia and delivered to the southeastern U.S.A. were estimated to be
$15.00 per ton in 1982 (8). Thus the value of gypsum to a user can range
from $9.00 to $20 - 30 per ton at site of use, the F.O.B. value in Nova
Scotia being as high as $15.00 per ton. However, in the United States, a
depletion allowance of 14 percent of the price of gypsum is granted to
mining companies as a tax benefit to compensate for the depletion of mine
reserves (9). This depletion allowance is applied to the price of gypsum
F.O.B. point of sale, not minesite, and therefore includes transportation
costs (plus other costs for handling and preparation). Thus, the tax credit
to the gypsum company effectively reduces the cost of mined gypsum. Al-
though it is beyond the scope of this study to quantify the benefits of the
depletion allowance, it is suggested that the actual cost of gypsum to a
gypsum mining company/board-manufacturer is substantially less than this
apparent value of $9.00 to $20 - 30 per ton at site of use.
The costs to the utility of producing usable gypsum are estimated to
range from $5 to $18 per ton for board-grade gypsum and from $8 to $11 per
ton for cement-grade gypsum (8). The wide range of costs is due to the
diversity of possible methods of processing the gypsum cake (10). This is
in turn due to the wide variance in board manufacturers' specifications (see
Table 2) and the various options available for dewatering (rotary drum
filter, horizontal belt filter, etc.), agglomeration (pelletization, extru-
sion, briquetting) and drying (fuel-fired vs regenerative). However, the
cost of producing usable gypsum may be partially or completely offset by
hypothetical disposal costs. The comprehensive analysis of the costs
associated with landfilling a filtered gypsum waste is a complex procedure.
A preliminary assessment of these costs yields disposal costs ranging from
$10.00 to $25.00 per ton (11). Costs associated with such disposal include
land costs, site preparation costs, dewatering costs, transport to disposal
site, wastewater treatment and reclamation. The wide range of costs is due
to variance in land availability and suitability, and uncertainty in waste-
water treatment costs. The estimate of wastewater treatment cost is uncer-
tain as it depends on the quantification of the flow volume of liquid purge
required to limit the dissolved solids in the scrubber slurry to an accept-
11-68
-------
able level. This limit and volume depend on coal chloride content, washed-
cake quality criteria, materials of construction, etc.
Table 2. WALLBOARD
MANUFACTURERS' SPECIFICATIONS
FOR FGD GYPSUM FOR USE IN BOARD
National
Gypsum
Georgia
Pacific
U.S.
Gypsum
Gypsum Content, %, Min.
Calcium Sulfite
Hemihydrate, %, Max.
Na, ppm, Max.
Cl, ppm, Max.
Mg, ppm, Max.
Free Water, %, Max.
PH
Particle size, Min.
94
0.5
500
800
500
1
6-8
xy=
2000 sq.
micron
90
N.S.
200
200
N.S.
12
3-9
N.S.
95
2
75
120
50
10
6.5-8
20
(mean)
micron
Note: N.S. = Not specified
In summary, the cost of FGD gypsum beneficiation is offset by the dis-
posal costs which would have been incurred in a non-utilization landfill
disposal option. If lower-cost wet stacking is the nominal disposal option,
gypsum beneficiation may be seen to cost more than disposal. However,
because the gypsum can be sold for use in gypsum board or cement, a net
credit may be realized if arrangements are made for a suitable user. If
this is so, the economics of production and sale of by-product gypsum can be
expected to be favorable.
PROBLEMS AND BARRIERS
Many influences within existing gypsum markets affect the potential for
recovery of gypsum from by-product sources. In each case, many factors
(location, capacity, quality, etc.) will significantly affect the
possibility of such by-product gypsum utilization. Assessment of these
factors is complex since there has been no extensive by-product gypsum
utilization in North America and because of the vertical integration of
major firms in the gypsum industry (i.e. gypsum mines, transportation
systems and board manufacturing plants are generally all owned and operated
by gypsum companies).
Principal obstacles to gypsum use by board and cement manufacturers
are
11-69
-------
as follows:
- Natural gypsum is a plentiful, inexpensive, raw material. (Ease of
transportation of by-product to the potential user(s) as contrasted
with the conventional source is a crucial factor in its economic
use. As with natural gypsum, the transportation cost will generally
be the most significant part of the delivered cost.)
- Gypsum users are reluctant to convert to by-product gypsum unless
both quality and quantity of the gypsum are assured. Given the fact
that electric utilities exist to produce power and are not
accustomed to serving markets outside the energy sector, warranting
of by-product quality and quantity may be a significant problem.
- When ample land areas for disposal are available close to the power
plant, the impetus for producing salable gypsum may not be present.
- Historically, the gypsum and utilities industries have each been
very conservative and insular. The production and use of by-product
gypsum by these industries will require policy changes that they may
find difficult to implement.
ACTIVE BY-PRODUCT FGD PROJECTS IN NORTH AMERICA
The four plants (Table 1) that have committed to date to the production
of salable gypsum vary so far as geographic location, system size and type
of utility company, but all are similar so far as:
The basis for use of forced oxidation including the impetus for
production of usable gypsum
- System design and operation
The prospect of gaining use of the by-product.
KEY MOTIVATING FACTORS
Aside from major process benefits of forced-oxidation mode operation,
including improved control of internal scaling and increased reagent
utilization, by-product gypsum applications to date have incorporated
forced-oxidation with production of a salable product because of substantial
constraints preventing use of an extensive area of nearby land for throwaway
type disposal. Also, apart from potential sale/use of the gypsum, this
manner of applying forced-oxidation ensures that the SO -catch is rendered
in a cake form that may readily be handled and temporarily stockpiled, and
thereafter ultimately transported any necessary distance to an off-the-site
point or points of delivery. Under such circumstances, and because of the
comparative ease and limited cost to adjust the design and quality of the
solid waste to yield a usable product, each of the projects has with
adequate incentive (but without agreements or contracts for product sales),
committed to a design compatible with by-product utilization. The resulting
gypsum yield is a benign and highly acceptable solid waste except for its
11-70
-------
significant solubility in water, and can possible be sold to existing users
or suppliers of natural gypsum located reasonably close to the power plant.
DESIGN AND OPERATION
In the FGD systems sold for usable gypsum production, the bulk of SO
removal is achieved in counterflow absorbers that discharge slurry effluent
to gas quenchers (precoolers) that yield an exit scrubbing slurry having a
suitably low percent calcium sulfite in the solid phase. No isolation of
the gypsum product from hydrogen chloride gas or other non-SO components
of the raw flue gas removed in wet collectors is achieved in the gas
scrubbing section. After dewatering of this final slurry, a wet cake
product containing as little as 8 percent surface moisture is discharged
from rotary vacuum filters equipped with cake washing accessories. The
concentration of objectionable water-soluble constituents in the gypsum cake
product is limited both by limiting of concentrations in the scrubber liquor
through discharge of liquid purge from the slurry dewatering system to an
acceptable receiver, and by the washing of the cake after it forms on the
filter drum. The concentration of objectionable solid-phase constituents
other than calcium sulfite is limited by use of a high-efficiency
electrostatic precipitator upstream of FGD, and of limestone reagent of
adequate quality/purity.
POTENTIAL FOR USE OF GYPSUM PRODUCED
These FGD systems are capable of continuous production of gypsum that
meets specifications for intended use, wallboard in the case of three plants
in Table 1 and other less critical uses in the case of the relatively small
output of Muscatine Power and Water. Market demand for natural gypsum in
the regions in which the plants are located is sufficient to assimilate the
FGD by-product yields, but each of the FGD systems has been purchased
without benefit of a gypsum sales agreement with a potential user(s), and no
such contracts have been announced to date.
Advantages of By-Product Substitution
The principal advantages to potential buyers of use of these FGD gypsum
outputs are:
The material may be expected to be supplied at a uniformly
acceptable level of quality.
The quantity of supply is substantial in relation to the
contemplated end use(s).
Particularly in Florida where natural gypsum is presently shipped
long-distance from Nova Scotia to supply wallboard manufacturers, or
principally from Spain in the case of gypsum rock for cement set
retarder, substantial savings in transportation cost may be
realized.
11-71
-------
Means for Meeting User Requirements
Provisions for drying and/or agglomerating of the FGD gypsum may in
some instances be necessary since prospective users may incur excessive
costs or inconvenience using raw washed FGD cake instead of gypsum rock.
Presence of surface moisture may also make it difficult to handle and store
the material in existing facilities designed for natural gypsum rock.
Moreover, water adds to shipping weight and unit transportation cost and
requires added energy use by the buyer since drying will generally be needed
for purposes of the end use. In specific instances, for example supply of
gypsum for use as cement set retarder if relevant, agglomeration of the
product by briquetting, extruding or other suitable means may be essential
for successful marketing of plant output.
Principal Obstacles to Use of By-Product
Lack of success in sales contract negotiations to date points to the
need for an equitable, generally acceptable, and broadly applicable basis
for pricing of the product. This may be essential in establishing long-term
by-product gypsum sales contracts between power plants and the major gypsum
users best able to benefit from use of the new source of supply. The deple-
tion allowance on natural gypsum authorized by the U.S. Internal Revenue
Service Code-613B significantly reduces the financial benefit to the verti-
cally integrated, North American gypsum industry of substitution of by-
product. If the allowance is to continue to apply in an era of newly and
plentifully available by-product gypsum suitable for principal market uses,
the by-product producer must realistically evaluate the supply economics and
competitive pricing of the product taking into account the purchasing disin-
centive created by the IRS tax allowance.
CONCLUSIONS
1. Forced oxidation type flue gas desulfurization (FO-FGD) technology
offers a design means for forming a usable gypsum by-product
instead of a throwaway solid waste.
2. Widespread application of FO-FGD in U.S.A. since 1978 has been
almost entirely in a throwaway waste mode with little apparent
incentive for usable gypsum production. Only in limited instances
in which the throwaway-disposal option is very constrained, i.e.,
13 percent of all U.S. FO-FGD capacity purchased to date, is the
FGD design predicated on manufacture of usable gypsum, and in none
of these cases, four in number, has a formal commitment to
purchase or use the by-product been obtained to date.
3. Cost advantages point to possible future use of "wet stacking"
type disposal in FO-FGD design and to significant environmental
impacts that are largely circumvented in earlier natural oxidation
type FGD designs that employ sludge fixation by use of fly ash and
other additives.
11-72
-------
4. Available technology and the successful widespread utilization of
FGD gypsum in Japan and West Germany indicate that its use is
technically feasible.
5. It is likely that, with or without beneficiation, the sale of FGD
gypsum to the board and cement manufacturers would be economically
preferable for the utility company in lieu of the throwaway-
disposal alternative.
6. Institutional barriers to the use of FGD gypsum throughout North
America include the high degree of vertical integration of the
gypsum wallboard manufacturing industry, the disincentive for
by-product gypsum utilization created by the gypsum depletion
allowance under U.S. Internal Revenue Service income tax codes,
and the abundance at diverse locations of natural gypsum of
adequate quality.
7. Additional major factors influencing the profitability of
production and sale of usable FGD gypsum include the magnitude of
alternative hypothetical throwaway-disposal costs, transportation
costs to user(s), and costs of transportation of user's nominal
source(s) of natural gypsum supply.
REFERENCES
1. Jenkins, S.D., and W. Ellison. Uitilization of FGD By-Product Gypsum.
(Presented at EPA/EPRI FGD Symposium, Hollywood, Florida, May 17-20,
1982.)
2. Pruce, L. M. Evaluating the Newest Disposal Options for Scrubber
Sludge. POWER, May 1981. p. 64.
3. EPRI Report CS-1579 by Ardaman & Associates, Inc., Orlando, Florida,
Evaluation of Chiyoda Thoroughbred 121 FGD Process and Gypsum Stacking,
Volume 3: Testing the Feasibility of Stacking FGD Gypsum. November
1980.
4. Ellison, W., and P.M. Kutemeyer. New Developments Advance Forced
Oxidation FGD. POWER, February 1983. p. 43-45.
5. Disposal of By-Products from Nonregenerable Flue Gas Desulfurization
Systems: Second Progress Report. US Environmental Protection Agency,
Report No. EPA-600/7-77-052, May 1977-
6. Control of Waste and Water Pollution from Power Plant Flue Gas Cleaning
Systems: First Annual R and D Report. EPA-600/7-76-018, October 1976.
7. Mineral Industry Surveys. U.S. Dept. of Interior, Bureau of Mines.
Gypsum in 1982, January 1983.
11-73
-------
8. Luckevich, L.M., Ontario Research Foundation, Mississauga, Ontario,
letter of September 22, 1983.
9. Byron, B.B., Member of House of Representatives, Washington, B.C.,
letter of September 6, 1983.
10. Economics of Disposal of Limestone Scrubbing Wastes: Sludge/Flyash
Blending. EPA-600/7-79-069, February 1979.
11. FGD By-Product Disposal Manual, Third Edition. Michael Baker Jr. Inc.,
Chapter 16, January 1983.
11-74
-------
OPERATING EXPERIENCE WITH THE CHIYODA THOROUGHBRED 121
FLUE GAS DESULFURIZATION SYSTEM
S. Kaneda, M. Nishimura, H. Wakui, I. Kuwahara, D. D. Clasen
-------
OPERATING EXPERIENCE WITH THE
CHIYODA THOROUGHBRED 121 FLUE GAS DESULFURIZATION SYSTEM
Seiichi Kaneda and Mitsuhiro Nishimura
Mitsubishi Petrochemical Company
Yokkaichi, Japan
Hitoshi Wakui and Ikuro Kuwahara
Chiyoda Chemical Engineering and Construction Company, Ltd.
Yokohama, Japan
Donald D. Clasen
Chiyoda International Corporation
Seattle, Washington
ABSTRACT
This paper reviews the design and operating experience of the Chiyoda
THOROUGHBRED 121 Flue Gas Desulfurization System installed and operated by
Mitsubishi Petrochemical Company at their Yokkaichi, Japan, complex.
3
The plant consists of a single scrubber and treats 260,000 Nm /h
(162,000 scfm) of flue gas from a 280 T/h (87 MW equivalent) boiler burning
high sulfur oil. Plant operation, since startup in May 1982, has been smooth
and trouble free over a wide range of operating conditions. Plant reliability
has been 100%. The plant is operated at a S02 removal efficiency of 97 to 99
percent for inlet S02 concentrations ranging between 1000 and 2000 ppm. Lime-
stone utilization is greater than 99 percent and the dry, gypsum by-product is
sold to a wallboard manufacturer. Operating and maintenance functions for the
system are minimal and completely absorbed by normal boiler plant operations.
INTRODUCTION
In March of 1981, the Mitsubishi Petrochemical Company awarded Chiyoda
Chemical Engineering and Construction Company, Ltd. a contract to install its
patented, second generation Chiyoda THOROUGHBRED 121 (CT-121) Limestone/Gypsum
System. The System was selected due to its low space requirement, simple
operation, and low plant cost. In addition, Chiyoda's engineering capability
and expertise were highly regarded based on Mitsubishi's experience with two
first generation CT-101 Systems.
The CT-121 System was to be designed to remove 97% of the S02 contained
in the flue gas from a new boiler being constructed by Mitsubishi Petro-
chemical Company at its Yokkaichi complex. The boiler provides steam for
11-75
-------
generating 40 MW of electricity and 47 MW equivalent of process steam. Fuel
oil containing 3 to 4% sulfur is burned, producing a flue gas S02 concentra-
tion of 1000 to 2000 ppm.
In addition to removal of SC>2, the System was also designed to remove
particulate matter (mostly carbon dust) from the boiler, thereby taking advan-
tage of the high particulate removal capability of the CT-121 Process and
eliminating the need for installing an ESP. The CT-121 System described
herein, therefore, incorporates equipment such as a fly ash thickener and
filter press that ordinarily would not be required.
The CT-121 System, including particulate removal equipment, was routinely
started up on May 11, 1982. The System has met or exceeded all performance
guarantees and operation has been smooth and trouble free over a wide range of
boiler loads and inlet SO concentrations.
GENERAL SYSTEM DESCRIPTION
Table 1 is a summary of the major design conditions for the CT-121 FGD
and particulate removal system and Table 2 is an equipment list including size
and materials of construction.
TABLE 1. SUMMARY OF DESIGN CONDITIONS OF THE
MITSUBISHI PETROCHEMICAL CT-121 PLANT
Flue Gas Source
Flue Gas Flow Rate
Flue Gas Temperature
S02 Inlet Concentration
Particulate Loading
S02 Removal Efficiency
Particulate Emission
Reagent
By-Product
Total Plot Area*
Schedule
Start Construction
Start Operation
280,000 kg/h (617,000 Ib/h)
oil-fired boiler (87 MW equivalent)
260,000 Nm3/h (162,000 scfm)
140-160°C (284-320°F)
1500-2000 ppm
200 mg/dNm3 (0.09 gr/dscf)
97%
< 50 mg/dNm3 (0.02 gr/dscf)
Powdered limestone
Saleable gypsum
1540 m2 (0.4 acres)
June 1981
May 1982
Includes gypsum by-product storage, particulate removal equipment and 120
meter solid FRP wet stack.
11-76
-------
TABLE 2. EQUIPMENT LIST
Equipment
Type
Size
Material
Remarks
CT-121 FGD EQUIPMENT
Flue Gas Fan
Precooler
Jet Bubbling Reactor
Air Blower
Overflow Tank
Limestone Silo
Limestone Slurry Tank
Limestone Feeder
Mother Liquor Tank
Centrifuge
Mist Eliminator
Gas Cooling Pump
Gypsum Slurry Pump
Overflow Pump
Limestone Slurry Pump
Mother Liquor Pump
Fly Ash Thickener
Neutralization Tank
Filter Press
Filtrate Tank
Thickener Overflow Pump
Thickener Underflow Pump
Filter Feed Pump
Filtrate Pump
Blow Down Tank
Blow Down Pump
Turbo
Cylindrical
Cylindrical
Turbo
Cylindrical
Cylindrical
Cylindrical
Screw Feeder
Cylindrical
Screw Decanter
Chevron
Centrifugal
Centrifugal
Centrifugal
Centrifugal
Centrifugal
FLY
Cylindrical
Cylindrical
Filter Press
Cylindrical
Centrifugal
Centrifugal
Centrifugal
Centrifugal
Cylindrical
Centrifugal
260,000 Nm3/h
(162,000 scfm)
9.8 m0xlO mH
(32'0x33'H)
5,000 Nm3/h
(3100 scfm)
6.8 m0x!9.6 mH
(23'0x65'H)
2,300 kg/h
(5100 Ib/hr)
3 x 1,330 kg/h
(3000 Ib/hr)
ASH REMOVAL EQUIPMENT
c
c
.s .
.s .
316L
cast
c
c
c
c
c
.s.
.s.
. s .
.s .
.s.
316L
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
. s .
.s.
.s .
.s.
.s .
.s.
.s .
.s .
.s .
.s.
.s .
.s .
.s .
. s .
.s .
.s .
with
s . s.
F1RP lining
FRP
with Agitator
aluminum
with
with
with
s . s.
with
with
with
with
with
with
with
with
with
with
with
with
with
with
with
with
F1RP lining
F1RP lining
F1RP lining
F1RP lining, P
Rubber lining
Rubber lining
Rubber lining
Rubber lining
Rubber lining
F1RP lining
F1RP lining
Rubber lining,
F1RP lining
Rubber lining
Rubber lining
Rubber lining
Rubber lining
F1RP lining
Rubber lining
with Agitator
with Agitator
.P.
with Rake
with Agitator
316L s.s.
-------
Figure 1 is a plot plan of the entire system showing the arrangement of
major equipment including the stack. Figure 2 is an overview of the FGD
portion of the System.
Flue gas to the System fluctuates daily between 50-100% of design
capacity due to changes in boiler load. The S02 concentration also fluctu-
ates widely due to the variety of fuel oils burned.
The control of the power plant, particulate and SC>2 removal equipment is
centralized and computer controlled. Both the boiler and CT-121 System can
be started up, operated, and shutdown from the central control room.
The heart of the CT-121 Process is the patented absorber which is called
a Jet Bubbling Reactor (JBR). This device, illustrated in Figure 3, allows
S02 to be absorbed, neutralized, oxidized, and crystallized in a single
vessel. The overall reaction equation for the process is:
S02 + CaC03 + %02 + 2H20 -*- CaSO^I^O + CC>2
PROCESS DESCRIPTION
Figure 4 shows the process flow for the CT-121 FGD System including
particulate removal equipment. Flue gas from the air preheater is pressur-
ized by the Flue Gas Fan and introduced into the Precooler. The gas is con-
tacted with a fine spray of recirculated water to humidify and cool the gas
to its adiabatic saturation temperature. The Precooler is also designed to
remove particulates and other impurities contained in the flue gas. The
humidified gas flows to the JBR where it is bubbled into a shallow layer of
absorbent. S02 is absorbed and precipitated as calcium sulfate (gypsum) by
the addition of limestone slurry and oxidizing air.
The desulfurized gas leaving the JBR flows through a two stage mist
eliminator to remove carryover mist. The mist eliminator is washed once per
day with mother liquor to remove accumulated fine gypsum crystals.
The level in the JBR is held constant by maintaining a liquid flow over
a weir to the overflow tank. Overflow liquid is recycled to the JBR.
Gypsum slurry is withdrawn from the JBR and is mechanically dewatered
using a solid bowl decanter centrifuge. The by-product gypsum is essentially
a dry material and is collected in a temporary storage area directly below
the centrifuge. Mother liquor from the centrifuge flows to the mother liquor
tank and recycled to the System.
Powdered limestone from the silo is fed to the limestone slurry tank via
a screw feeder and mixed with mother liquor to obtain a 20 wt.% slurry. The
slurry is pumped to the JBR on flow control based on a feedforward boiler
load signal and a feedback JBR pH signal.
A portion of the Precooler cooling water flows to a fly ash thickener to
concentrate captured particulates. The overflow liquid is returned to the
Precooler and the underflow is pumped to a neutralization tank. A small amount
11-78
-------
60.5 m (199 FT.)
FROM BOILER
1 JET BUBBLING REACTOR
2 PRECOOLER
3 FLUE GAS FAN
4 MIST ELIMINATOR
5 LIMESTONE SILO
6 FILTRATE TANK
7 FLYASH THICKENER
8 SLOWDOWN TANK
9 CENTRIFUGE (SOLID BOWL DECANTER) AND
GYPSUM STORAGE HOUSE
10 FILTER PRESS AND FLYASH STORAGE
11 STACK
12 BYPASS DUCT
13 FOUNDATION OF STACK SUPPORT
Figure 1. ARRANGEMENT OF THE CT-121 FGD PLANT FOR MITSUBISHI PETROCHEMICAL CO.
-------
I
CD
o
Figure 2. OVERVIEW OF THE MITSUBISHI PETROCHEMICAL CT-121 PLANT
-------
Flue Gas
Limestone
Slurry
Gypsum
Slurry
Figure 3. JET BUBBLING REACTOR
11-81
-------
. BOILER PLANT! FQD PLANT
i
CO
ro
STACK PRECOOLER MIST JET BUBBLING OVERFLOW
ELIMINATOR REACTOR TANK
(JBR)
LIMESTONE SILO CENTRIFUGE MOTHER LIQUOR
LIMESTONE SLURRY TANK
TANK
Figure 4. PROCESS FLOW SHEET OF THE CT-121 FGD PLANT FOR MITSUBISHI PETROCHEMICAL CO.
-------
of mother liquor is also added to this tank. The combined stream is neu-
tralized with limestone slurry. The neutralized liquor, containing approxi-
mately 2 wt.% solids, is pumped to a filter press where the solids are
removed. The filtrate is collected in a filtrate tank and recycled to the
system via the mother liquor tank. A small amount of the filtrate (1 to
3 m3/hr (4 to 13 gpm)) is bled from the filtrate tank to maintain system
water balance.
OPERATING RESULTS
PLANT OPERATION
The CT-121 System was smoothly and routinely started up on May 11, 1982.
Tables 3a and 3b are summaries of the operating history of both the boiler
and the CT-121 System. As shown in Table 3b, System reliability has been
100%. This is particularly impressive considering the System consists of only
a single scrubbing module and is operated under severe conditions.
DESULFURIZATION EFFICIENCY
The S02 removal efficiency, as shown in Table 3a, has averaged 97 - 99%
for inlet S02 loadings ranging between 1000 - 2000 ppm.
Figures 5 and 6 show the effect of boiler load and S02 inlet concentra-
tion changes on desulfurization efficiency. These two Figures illustrate
the CT-121 Process' capability to maintain stable desulfurization under
widely fluctuating boiler loads and inlet S02 concentrations.
OPERABILITY AND CONTROLLABILITY
The boiler load changes daily from 50 - 100% of design capacity. The
System responds well and automatically to such load fluctuations. Figure 7
illustrates the CT-121 System response to a typical change in load.
The CT-121 System is extremely simple to operate and requires little
operator attention. System operation is basically automatic, being controlled
by a digital computer control system.
The flow of flue gas to the System is set by boiler load. Flow is auto-
matically adjusted by a pressure controller installed on the suction side of
the Flue Gas Fan which controls the position of the fan inlet vanes. A by-
pass damper is installed to protect the boiler against sudden changes in
draft.
The limestone slurry concentration is maintained constant by adding
limestone in proportion to mother liquor feed by using a proportional control
loop. Limestone slurry is monitored using an in-line density meter.
The gypsum concentration in the JBR is maintained at about 15 wt.% by
pumping a small slip stream of slurry to the centrifuge. The flow rate is
controlled in proportion to the S02 load signal which is the product of the
boiler load signal and the S02 inlet concentration.
11-83
-------
TABLE 3a. OPERATING HISTORY OF THE MITSUBISHI PETROCHEMICAL CT-121 FGD PLANT
Operating Conditions
Flue gas source
Boiler capacity
Startup date
Gas flow rate
Gas temperature
S02 concentration
Desulfurization efficiency
oil-fired boiler
280 T/h steam (87 MW equivalent)
May 11, 1982
100,000-246,000 Nm3/h (62,000-153,000 scfm)
140-160°C (284-320°F)
700-2,300 ppm
96-99%
I
oo
-0-
Operating Summary
1982
1983
MAY
l,500b
JUN
96
1,500
JUL
98
1,000
AUG
99
1,000
SEP
98.5
1,000
OCT
98.5
1,500
NOV
97.5
2,000
DEC
97.5
2,100
JAN
98
2,000
FEE
98
1,800
MAR
98
1,800
APR
97.5
1,600
MAY
98
2,000
JUN
98
2,000
Startup
5/11/82
Start of commercial operation
Cause of Shutdowns
Operating
Shutdown
^Scheduled shutdown for boiler/turbine inspection
A*Scheduled shutdown for first annual inspection of steam
generating plant.
Average monthly S02 removal, %
Average monthly inlet S02 cone., ppm
-------
TABLE 3b. CT-121 RELIABILITY PARAMETERS (MAY 11, 1982 THROUGH JUNE 30, 1983)
FOR THE MITSUBISHI PETROCHEMICAL FGD PLANT
Parameter Value (%) Total Time (hours/hours)
Reliability j* 100.0 9,249/9,249
Availability 98.1 9,816/10,008
Operabilityc 99.9 9,249/9,255
Utilization Factor 92.4 9,249/10,008
!_, . . _____ . . , , , .
!_, , . . ,—
I
oo
Reliability - hours the FGD System was operated divided by the hours the FGD system
was called upon to operate.
Availability - hours the FGD System was available for operation (whether operated
or not), divided by the hours in the period.
Q
Operability - hours the FGD System was operated divided by the boiler operating
hours in the period.
Utilization Factor - hours that the FGD System was operated divided by the hours
in the period.
-------
o
c
CD
'o
c
o
'•£
03
N
3
CO
100
0
20
40
60
80
100
Boiler Load. %
Figure 5. EFFECT OF BOILER LOAD ON DESULFURIZATION EFFICIENCY
^\
o
c
CD
O
LU
C
o
03
N
"5
w
CD
Q
100
99
98
97
96
- — ._
-
,
i —
^^^^
i
-
-
1000 pnnn
S02 Concentration In The Flue Gas, ppm
Figure 6. EFFECT OF S02 CONCENTRATION ON DESULFURIZATION EFFICIENCY
11-86
-------
CO
o
—
"o
CD
100
50
0
a 1600
a
"cvj 1500
O
w 1400
0
10
20
0
10
20
0
10
20
0
30
40
30
40
E 98
01 97
C\J
O
CO
/
^
^"X,
v/
30
40
3.7
a
£T 3.5
m
^ 3.3
t
r
\j
\
V
r
\
10 20 30 40
Elapsed Time, Min.
Figure 7. TYPICAL SYSTEM RESPONSE TO CHANGE IN BOILER LOAD
11-87
-------
The pH in the JBR is controlled by the addition of limestone. The flow
of limestone slurry to the JBR is controlled by the feedforward signal of the
S02 load, compensated by the feedback signal of the JBR pH. This control loop
results in only a ± 0.2% perturbation in JBR pH for maximum load changes.
POWER CONSUMPTION
The power consumption is 1,100 kW for operation at the following condi-
tions:
Boiler Load 100% (87 MW equivalent)
Inlet S02 2,000 ppm
Desulfurization Efficiency 98%
Power consumption is less than 1.3% of the boiler equivalent generated
MW. This consumption includes particulate removal as well as SOz removal.
CT-121's low power consumption is due to the elimination of slurry recycle
pumps characteristic of first generation limestone systems.
LIMESTONE UTILIZATION
Limestone utilization has averaged greater than 99%. Table 4 shows a
typical analysis of the limestone used.
TABLE 4. TYPICAL LIMESTONE ANALYSIS
Chemical Composition
CaCO 97.98 wt.%
CaO 54.87 wt.%
CO 43.80 wt.%
SO 0.050 wt.%
MgO 0.549 wt.%
Al 0 0.059 wt.%
Fe203 0.023 wt.%
SiO + Insol. 0.86 wt.%
Particle Size
325 mesh 90% pass
BY-PRODUCT GYPSUM
A typical analysis of the gypsum by-product is shown in Table 5. The
gypsum is sold to cement and wallboard manufacturers.
11-J
-------
TABLE 5. TYPICAL GYPSUM BY-PRODUCT ANALYSIS
CaS04«2H20 99.2 wt.%
S03 46.1 wt.%
CaCO 0.1-0.6 wt.%
Combined water 20.6 wt.%
Moisture 11-12 wt.%
pH 6.5-7.5
Crystal size 49 microns
(average)
WASTEWATER
A typical analysis of the wastewater, after neutralization with lime-
stone, is shown in Table 6. The water is discharged to the ocean.
TABLE 6. TYPICAL WASTEWATER PROPERTIES
pH 6.0
Suspended solid 0.5 ppm
COD less than 7 ppm
OPERATING AND MAINTENANCE REQUIREMENTS
Operating and maintenance functions for the CT-121 System have been
minimal and are completely absorbed by normal boiler plant operations. No
scaling, plugging, or other maintenance intensive problems have been encoun-
tered. The total average personnel requirement, including maintenance, has
averaged less than one-half man per shift.
RESULTS OF THE FIRST ANNUAL INSPECTION
The steam generating plant was shutdown March 1983 for the first sche-
duled annual inspection. The CT-121 System was also inspected at this time.
Inspection of the CT-121 System, including the JBR, revealed no scaling,
corrosion, or other problems. As such, maintenance was limited to routine
servicing of equipment and flushing off some soft deposits of carbon dust and
gypsum crystals that had accumulated on the upper and lower decks of the JBR.
11-89
-------
SUMMARY
The CT-121 System has met or exceeded all performance guarantees and
expectations. The System has routinely achieved greater than 99% desulfuri-
zation using essentially stoichiometric amounts of limestone. System relia-
bility has been 100%.
The CT-121 System has also proven to be an exceedingly simple process to
operate and maintain. Maintenance has been minimal and limited to routine
servicing of equipment. Total operating and maintenance manpower requirements
have averaged less than one-half man per shift.
11-90
-------
OPERATION EXPERIENCE WITH FGD PLANT II AT
WILHELMSHAVEN POWER PLANT, WEST GERMANY
B. Stellbrink, H. Weissert, P. Kutemeyer
-------
by:
OPERATION EXPERIENCE WITH FGD PLANT II AT
WILHELMSHAVEN POWER PLANT, WEST GERMANY
(Figure 1)
B. Stellbrink
General Manager
Power Plant Wilhelmshaven
West Germany
H. Weissert
Director of Research
Bischoff GmbH
Essen, West Germany
& D evelopment
P. Kutemeyer
General Manager
Bischoff Environmental Systems
Pittsburgh, PA
Figure 1. Wilhelmshaven Power Plant
11-91
-------
ABSTRACT
The Nordwestdeutsche Kraftwerke Aktiengesellschaft (NWK) has been
operating a 720 MW coal-fired power plant in Wilhelmshaven, West Germany,
since 1976. In March, 1982, the second flue gas desulfurization plant, called
REA 2, the German acronym for flue gas desulfurization plant 2, was put into
operation after a two-year construction period. Startup and checkout proceed-
ed without major difficulties and was completed within approximately
three months. Beginning in June, 1982, NWK accepted the plant. Since then
REA 2 has been in operation without significant interruptions, effectively
reducing SO emissions of the power plant,
(Figure 2)
Figure 2. FGD Plants 1 & 2 at Wilhelmshaven
11-92
-------
FGD treatment at Wilhelmshaven is accomplished by two plants. REA 1 is
a co-current scrubber in operation since 1977. Together REA 1 and 2 treat
about 80 percent of the flue gas generated. The FGD plants are located in
the by-pass to the main flue gas duct behind the electrofilters, and are equ-
ipped with their own blowers. REA 1 has a throughput of 500,000 nm3/h
(294,287 SCFM) and REA 2; 1,500,000 nm3/h (882,862 SCFM) . (Figure 3)
Instrumentation for measuring dust content, SO and NO concentrations,
and exhaust temperature of the cleaned flue gas is located at the 27 m (88
ft.) platform elevation of the 275 m (902 ft.) high exhaust stack.
Steam Generator
ESP
REA 1
V =500,000 nnWh
(294,287 SCFM)
REA 2
V =1,500,000 nrrvVh
(882,862 SCFM)
Stack
V = 2,500,000 nrrWh
(1,471,436 SCFM)
tg =100°C (212°F)
Figure 3. Present Flue Gas Flow at Wilhelmshaven
Both REA 1 and 2 use a Bischoff wet absorption process. REA 2, however,
is a counter-cur rent scrubber with regenerative re-heating of the flue gas.
A portion of the flue gas generated by the power plant is injected into the
REA 2 by means of an axial blower having variable vanes. From the blower,
the flue gas enters the hot side of regenerative heat exchanger (REX), and
from there enters from the side into the counter-cur rent scrubber. In the
scrubber, flue gas passes upward through various levels of lime slurry sprays
and is cooled and desulfurized. At the top of the scrubber tower the flue
11-93
-------
gas passes through a mist separator, reverses direction, and enters a venturi
mixer/preheater in which the remaining water droplets in the flue gas are
evaporated by mixing hot, untreated gas with the desulfurized flue gas. From
the venturi mixer, the flue gas enters the cold side of the REX and is reheat-
ed to a temperature of 100°C (212°F) . From there the flue gas exits through
the exhaust stack. (Figure 4)
Scrubber
Residue Sludge
Water
Flow Diagram REA 2
Figure 4. Schematic of 415 MW Single-Module Bischoff System
OPERATING EXPERIENCES, PROBLEMS, AND SOLUTIONS
Operation and control functions for the two FGD plants are effected
from the power plant control room. To date, no power plant interruptions
have been caused by either REA 1 or REA 2. Both FGD plants can be shut down
or re-started on short notice, and desulfurization occurs immediately after
start-up of the blowers. In the event longer shut-down periods are antici-
pated, such as when a low sulfur fuel is used, for example, REA 2 can be
separated from the main flue gas ducts by means of double valves, and pre-
served by blowing dried air into the system through both the raw and clean gas
ducts.
11-94
-------
SCRUBBER
The scrubber exhibits good operating characteristics and quickly adjusts
to changes in the flue gas volume, maintaining a high degree of desulfuriza-
tion. Despite a high concentration of solids in the scrubber slurry, (100
g/1) , slurry liquid circulation can be readily re-started, even after being
shut down over the weekend. The system does not require the use of recircu—
lating pumps or agitators during the time it is shut down. During shut-down,
about 150 tons of gypsum settle out in slurry lines, pumps and the sump. At
start-up, the slurry pumps re-suspend the gypsum by backflushing for five
minutes using a specially-designed system of flushing lines, after which the
pumps are switched to the spray nozzles. The flushing process operates auto-
matically.
Due to settled gypsum behind the valves, difficulties were experienced
initially, in that torque overload sensors shut the automatic flushing pro-
cess down when higher than permitted torques where required to open the valvas.
Increasing the size of the valve drive shafts and motors solved this problem.
Lime is added below the oxidation zone into the scrubber sump. The
amount of lime added is controlled by means of an rpm-regulated pump. Ini-
tially, a small amount of scrubber slurry was used as the medium to add the
lime. After about 2,000 hours of operation, however, the supply linesfor
adding the lime clogged with gypsum. Since then, lime is added without the
use of slurry fluid. The 1,000-hour operating life of the pump stators is
not considered acceptable. It is hoped that this problem can be solved by
installation of a bypass line directly to the scrubber sump and thus avoid-
ing the upper rpm range of the pumps, which seems to be the cause of the
problem.
During the acceptance test period, a large amount of water was found
after the mist separator. The reasons were improperly installed separator
segments and reintrainment of water due to open collection troughs. The
former problem was corrected; the later was solved by installing covers
over the collection troughs. Makeup water for the scrubber is added by using
it to flush the mist separator.
To date, no scaling has been observed in the slurry systems, the scrub-
ber, or mist separator. Nozzles have shown no wear after 4,000 hours of
operation.
LIMESTONE
At the present time there are several FGD plants under construction in
the Federal Republic of Germany (FRG) , which, due to their proximity to lime-
stone quarries, intend to use limestone instead of lime. In order to deter-
mine the effects in an operating plant, REA 2 was tested using limestone.
The results were as follows:
The use of limestone is possible; however, a lower cegree of desulfuriz-
ation is achieved than with lime. This is a result of the slower reaction
11-95
-------
rate of limestone. In order to achieve the same degree of desulfurization a
considerably larger quantity of scrubbing slurry must be used. This, of
course, means a higher energy requirement and greater wear of the equipment.
Also, the quantity of limestone required is 1.8 times greater than that of
lime due to the lower amount of CaO in limestone. However, byproduct gypsum
can also be produced without difficulty.
For the Wilhelmshaven power plant, the use of limestone is not an econ-
omically viable alternative, due to the high power requirements and the
greater transportation costs.
FLUE GAS DUCTING
The flue gas ducting is divided into sections. Each section is free to
expand or contract. Corrosion-resistant fabric expansion joints are used
between the sections. During startup and testing, flue gas escaped through
leaky bolt connections on the expansion joints and caused damage to the insul-
ation. The bolted connections were improved. Since, in an FGD plant using
regenerative reheat, sections of the flue gas ducting may be operated at
temperatures below the dew point of the acids in gas, special precautions
must be taken with respect to corrosion protection of those sections. Affect-
ed are especially the REX, the flue gas ducts between the REX and the scrub-
ber , and the venturi mixer/dryer. Also, the flue gas ducts and especially
the "floors " must be even with a slight slope and have a gutter to allow
condensates, which usually have a pH of less than 1, to be drained off. The
flue aas ducts between the REX and the scrubber should be sloped toward the
scrubber to allow the condensate to drain into the scrubber. Other conden-
sates flow over acid-resistant gutters to a rubber-lined collection basin
from which they are pumped to the scrubber. This solution prevents an efflu-
ent problem. In case the REX must be flushed, the arrangement also allows the
drainage of the flushing water to the scrubber sump.
Coating of the various components of the FGD plant presented a problem,
since no experience with coating materials was available involving applica-
tions of highly concentrated sulphuric acid at temperatures greater than
80 C (176 F) . Various specialty firms in this area were consulted. Only a
few were willing to guarantee their coating for a minimum of two years.
Several different materials were selected and applied to different sections
of components of the system. After 4,000 operating hours, no final evalu-
ation of the various coating materials is possible. Results, however, appear
encouraging. It has been determined that all coatings used are satisfactory
on vertical walls and on the "roof" of the flue gas ducting.
The following results are noteworthy:
The flue gas ducting should ba fabricated and assembled as free of all
stress as possible and should remain stress-free during operations. This is
important because movements of the walls will cause the coating to crack or
flake off, resulting in corrosion. The flue gas ducting should be installed
such as to prevent dents, and should have sufficient slope to prevent acid
puddles from forming during operations and unnecessarily stressing the ccatLng.
11-96
-------
The coating material must remain flexible, even after long periods of
use. It is also necessary to apply the coating a sufficient number of times,
as well as during warm weather. (Figure 5)
Scrubber
Clean Gas
Rubberized
Kreiwatherm
Keraflake
Plastics
Enameled
Scrubbing Slurry
A Test Coating 3,398 hrs
Colebrand
CXL 2000 Coating
S =1 mm/10 Coats
Clouth
Rubber Application Durabilit 1565
S = 1.6mm
Ceilcote
871 x Lining (Viton)
S =1.4mm/4 Coats
Ceilcote
Flakeline 282
S =1.6mm/4 Coats
Materials
Figure 5. View of FGD Plant
INSULATION
The insulation must be thick enough and must be carefully installed.
Improper insul ation will result in cold spots, and thus, condensation. This
in turn will result in corrosion and scaling, which can result in the dis-
charge of solid particles from the stack. Special care must be taken in
insulating support structures. It is important that in the initial request
for bids, the flue gas conditions and the demands placed on the insulation
be clearly spelled out.
11-97
-------
REGENERATIVE HEAT EXCHANGER (REX)
TEMPERATUR 1°C
Figure 6. Temperature Profile of REX for Cold Start
The concept of reheating the desulfurized flue gas has been viable
todate. Startup of the REX, from a cold condition, has not presented
any problems. Gas discharge temperatures, measured in the stack, always
exceed 100 C (212 F) , even during periods of partial loads. The heat-stor-
age chambers are made of plastic material on the cold side and of enamel-
coated steel on the hot side. On the cold side, deposits form, caused by
water droplets entrained in the clean gas. These evaporate at the inlet of
the REX. On the hot side, deposits form due to entrained dust particles
containing acidic deposits in the raw gas. (Figure 6)
In case the electrostatic precipitator does not function properly, the
resulting increase in ash content of the raw gas quickly increases the pres-
sure drop across the REX.
The hot surfaces are cleaned by means of pressurized air, which is
blown over the surfaces by special nozzles. This is performed after every
12 hours of operation. During normal operations, i.e. without malfunctions
in the ESP or the scrubber, this cleaning interval has been found to be
11-98
-------
sufficient. Should larger deposits form more quickly due to malfunctions in
either the EXP or the scrubber, cleaning by means of high pressure water is
required. The same nozzles used for the high pressure air are also used for
the high pressure water spray. For this purpose, they are switched to a
high pressure water pump. On the hot side, the water pressure before the
nozzles is 80 bar (1160 psi) and 40 bar (580 psi) on the cold side. The REX
continues to turn during the cleaning process with a reduced velocity. The
amount of water used is held to a minimum. This water is added to the scrub-
ber as make-up water. The cleaning method has been used during operation of
the REA 2, as well as during shutdowns.
Pressure Loss of REX During Operating Period
Figure 7. Pressure Drop Diagram For REX
Measurements to determine pressure losses across the REX from start-up
until the 29th of March, 1983, show that the pressure loss increase has been
smaller than anticipated. The deposits on the REX have been maintained in
acceptable limits during the first 4,000 hours of operation by cleaning regu-
larly with high pressure air or water. The installed reserve capabilities
of the blower, layed out for a pressure drop increase up from 10 mbar (0.145
psi) to 13 mbar (0.216 psi), has been sufficient. (Figure 7)
11-99
-------
Tests will soon be conducted to determine whether deposits on the hot
side can be effectively removed using steam instead of water. To date, no
noticeable corrosion has been determined in the REX.
No definite statement can be made regarding the lifetime of the heat-
storage chambers after this short period of operations.
PRESSURE LOSSES
Measurements of pressure losses across various components of the REA 2
system at start-up showed that the pressure drops were below those given by
the various suppliers. Pressure losses across the scrubber and the flue gas
ducts do not change with time.
ENERGY REQUIREMENTS
Testshave shown that in those cases when the SC>2 content of the f^Lue gas
is in the lower end of the operational spectrum (S02 le^ss than 2 g/mn );
(0.8 gr/scf) or flue gas volume less than 1,000,000 nm /h (588,574 scfm) ,
the desulfurization efficiency does not increase significantly if two slurry
pumps are operated. Using only one slurry pump, the desulfurization effici-
ency will only decrease by 2 percent; however the power saved is 1.2 MW.
(Figure 8)
Operation with Two Pumps
SO?>22g/nm3
Operation with One Pump
SO_, ' 2 2g/nm3
400
600 800 1000 1200
Flue Gas Volume 103 nnWhr (589 SCFM)
1400
1600
1800
Energy Requirements of REA 2
Figure 8. Energy Requirements for REA 2
11-100
-------
INSTRUMENTATION AND CONTROLS
Control
For controlling the REA 2, it was decided for the first time in this
power plant to use a programmable controller. The controller used is of the
Siemens type 5 and is used in conjunction with a more conventional control
system, ISKAMATIC B. The S 5 system is used to control the functional groups,
including the safety interlocking systems. The great advantage offered by
the programmable S 5 system is the ease with which the automatic controls can
be adjusted to the operational process. This was shown to be of great advan-
tage during the start-up phase of the operation.
The following group controls are used:
- flue gas blowers with oil supply and valves
- slurry pumps with flushing program
- flushing of mist separator
- lime addition
Modifications are presently under way to upgrade the automatic control
system to also handle clarifying and gypsum dewatering. It is intended to
control the entire FGD plant from the power plant control room and includes
starting the plant and shutting it down.
Performing the various controlling functions is accomplished with the
process control system TELEPERM C. Control is accomplished using the fol-
lowing control loops:
- raw gas volume
- drying of the clean gas
- controlling amount of lime added
- solids removal, slurry discharge
- control of various fluid levels
Since the REA 2 is a counter-cur rent scrubber, it is, theoretically at
least, possible to adjust the pH value at the slurry discharge to such a
value as to optimize the oxidation of the sulphite to gypsum, without adver-
sely affecting the desulfurization.
This results in the following very important advantages. First of all,
the lime usage is stoichionB trie, and readily adjustable as operating para-
meters change. Second, it is possible to integrate the oxidation stage in
the scrubbing tower. This, of course, does increase the demands placed on
the instrumentation and control system, since, unlike the co-current system,
11-101
-------
the scrubbing fluid is practically unbuffered.
The pH control loop was modified during start-up. Instead of continu-
ously changing lime dosage as a function of pH, lime dosage is now adjusted
in discreet steps. However, more recent tests indicated that, by also
measuring density, a more accurate mass flow rate of the lime to be added
is now possible. Originally, a rough dosage and a fine dosage were used
for the injection of lime required. The fine dosage was found to be not
required and was disassembled.
T OUS BKB • Km
A. mm •cJiMEK
VJ1 hi' mUHW
RAW GAS VOLUME
LIME ADDITION
Control Diagram of REA 2
Figure 9. Control Diagram
INSTRUMENTATION
In order to monitor and automate the operation of the FGD plant, as well
as to optimize the dosage of additives and the desulfurization, reliable and
maintenance-free instruments must be used. This is especially true for
determining the amounts of additives supplied to the system by measuring the
solids content in the slurry and the amount of lime added, the pH value of
the wash fluid and the S02 content in the raw and clean gas.
11-102
-------
The following systems were selected:
- solids and density measurements: radiometric analysis using the
absorption principle
— SO concentration measurement spectral photometer using the
absorption principle
- pH value measurement electrometric analysis, potentio-
meter
- flow measurements induction measurement
- scrubber fluid measurement ultrasonic measurement
- flue gas volume measurement pito tube (annular tube)
Most of these instruments have proven to be very reliable. Determina-
tion of SOj concentration using the absorption principle causes some problems,
however. (Figure 9)
EFFICIENCY OF S02 REMOVAL
The degree of SO removal in the scrubber is influenced by various
factors. These are: the relationship between scrubbing fluid voJnroe and gas
volume;gas velocity; pH value; and SO concentration of the raw gas. During
acceptance tests, it was found that 95 percent of the SO in the gas was
scrubbed out. Under operational conditions, with S0? concentrations of be-
tween 1 g/nm3 and 3 g/nm3 (0.4 gr/scf and 1. gr/scf), 91 percent to 95 percent
of the SO. was removed. Because the regenerative gas reheater was selected
in the REA 2, the overall SO removal efficiency is lower than in the scrub-
ber, due to the fact that raw gas is added to the clean gas after the mist
separator to evaporate any water droplets still in the gas, and also due to
the fact that some raw gas leakage occurs in the regenerative gas reheater.
Drying of the clean gas, using raw gas, was selected to prevent
deposits from forming on the regenerative gas preheater.
To evaporate the rest moisture in the clean gas, the clean gas temp-
erature must be raised. The rate of evaporation, due to the increase in
temperature, is determined by the available transit time of the gas through
the dryer. It must also be considered that the salt content of the evapo-
rated moisture will increase the salt concentration, resulting in decrease
of the water vapor pressure and thus increasing the dew point.
The amount of increase in the dew point depends on the type of salt
desolved. Table salt, magnesium chloride and calcium chloride result in
dew point increases of about 5°C (9°F), 30°C (54°F), and 45°C (81°F) resp-
ectively, when present in saturated solutions, Drying, by using raw gas,
is thus only viable when readily soluble solutions such as magnesium of
calcium chloride are not present. At Wilhelmshaven, where the REA 2 is
operated with sea water, this does not appear to be a problem.
11-103
-------
A delta T of 8 C (14.A F) is achieved in the gas preheater, resulting
in a decrease in the SO removal efficiency of about 10 percent. This dis-
advantage of clean gas drying can only be avoided if a highly efficient mist
separator can be used, decreasing the need for such a dryer by removing a
sufficient amount of the residual moisture and salt.
A T (Dryer)
Decrease of Scrubber
Desulphurization
Efficiency
Scrubber
Stack
Power Plant Wilhelmshaven
8°C (46.4°F)
2°C (35.6°F)
0°C (32°F)
12.5%
5 %
2.5%
Recirculatmg 1 M^L ^1
Blower V ^^P •
, '—I 6 w
Drying with Recirculating Blower
2-5 °C
(35.6-41 °F)
2.5%
Blower Using Hot Clean Gas
2-5°C
(35.6-41 °F)
2.5%
Blower Using Cold Raw Gas
1%
Blower Using Cold Clean Gas
~3°C (37.4°F)
Due to Blower
1%
Figure 10. Total Desulfurization Efficiency of REA 2 with REX
By decreasing the delta T from 8°C (14.4°F) to 5°C (9°F) , no effects on
the REX have been determined. During the winter of 1982, a delta T of 2 C
(3.6 F) was measured across the gas dryer. In this case, deposits began to
build up on the surfaces of the cold side of the REX. These deposits could
not be removed by means of high pressure air; they were, however, removed
with high pressure water.
a 2 C (3.6 F) delta T in the dryer is possible if cleaning of the
„„—4.-T--I - Increased mist separator effi-
Thus, , . „. _.._ ^ _y
REA with high pressure water is acceptable. a_iiuj.fc:asea mist separator
ciencies would increase operating times between flushing considerably.
Without the use of the venturi dryer, no noticeable difference in the
increase in pressure drop across the REX resulted, as compared to operation
with the dryer. These tests encompassed a time period of two months of
11-104
-------
effective operation. In this case, „.._,, „„...„„. „..-. „.„„.
intervals was employed. Tests are presently continuing.
i, only normal air cleaning at 12-hour
' /"I 1-V -V- /-I O i-lfl f-T IT /-«|—tT-»f-T1-IT1-1T^ r>
It is not possible, after the relatively short time since this system
went into operation, to determine if flushing with high pressure water will
cause damage to the REX due to increased corrosion or loosening of the
regenerative heater packages.
By reducing the delta T from 8°C (14.4°F) to 5°C (9°F) , SO removal
efficiencies increase from 82 percent to 89 percent. By not using the
dryer, the efficiency could be raised to 92 percent. If this is not possi-
ble the gas would have to be dried using the blower on the clean gas side.
This would not affect the efficiency of the scrubber.
In such a case, gas leakage in the REX does not decrease the S0? removal
efficiency of the system significantly. The resulting increase in the clean
gas temperature caused by the blower minimizes the need for a gas dryer.
Such a solution is presently being contemplated. (Figure 10)
Figure 11. Magnified Photograph of the Crystals
11-105
-------
GYPSUM PREPARATION
The byproduct generated by the Bischoff FGD plant is high quality
gypsum containing a high proportion of large gypsum crystals. Because of
this, NWK and a wallboard manufacturer have signed a contract to utilize
the FGD byproduct gypsum for wallboard production. The contract specifies
the quality of the gypsum. To meet these specifications, the byproduct
gypsum must be washed and dewatered. (Figure 11)
In purchasing the equipment necessary to wash and dewater the gypsum,
cost was an important factor, since the profit resulting from the sale of
the gypsum should result in a reasonable payback period. The equipment
presently being installed or planned is complicated and expensive. In
general, such a system consists of various intermittently-operating cent-
rifuges, or continously-operating vacuum filters with an additional system
to transport the gypsum to a silo or other storage facility without addi-
tional handling.
To find a suitable system, extensive tests were conducted with the fol-
lowing systems:
- vacuum-band filter
- vacuum-drum filter
- two-stage centrifuge
- four-stage centrifuge
- screw centrifuge
All of these systems are, under certain conditions, suitable for use in
washing and dewatering byproduct gypsum. The following residual moisture con-
tents are measured.
vacuum-filter 8 - 10 % (without free water)
centrifuge 6-8% (without free water)
11-106
-------
Transmission
Oil
Reservoir
Rubber Mounts
Fill Tube
Screw Drum
Screw
Wash Tubes
Filter
Discharge
Solids
Discharge
Screw Centrifuge
Figure 12. Screw Centrifuge
At Wilhelmshaven the screw centrifuge was selected due to the fact that
it was less expensive, required less space and energy, could be automated
without much difficulty, and finally, could be expected to have low-wear
characteristics. (Figure 12)
11-107
-------
/ / / / /////////////// // 7
Figure 13. Gypsum Silo with Centrifuge
The centrifuge is installed over a silo, which has a capacity of
2% days of gypsum production. Centrifuge, silo, and loading facility are
one structure. Loading is accomplished by means of a screw conveyor which
is activated by the truck driver. Before the gypsum slurry reaches the screw
centrifuge, fine particles are removed by a hydrocyclone. A portion of the
separated material flows to a thickener. The main stream is returned to the
scrubber. The gypsum preparation plant is started up, or shut down, depend-
ing on the measured solid content in the scrubber slurry. Boundary values
are 60 to 120 g/1. (Figure 13)
11-108
-------
Mist
Separator
Cyclone Water Discharge
Thickener
Figure 14. Schematic of Gypsum Preparation Plant
The gypsum generated in 24 hours of desulfurization can be withdrawn
from the scrubber within eight hours. Gypsum is withdrawn when the solids
content in the slurry reaches 123 g/1. Withdrawal ends when the solids
content reaches 60 g/1. To insure that only large gypsum crystals are re-
moved, the centrifuge filters out undersized crystals and returns them to
the scrubber. In the time required for the solids content to again reach
120 g/1, the gypsum crystals have ample time to grow. Thus, an adequate
supply of large crystals is always assured. (Figure 14)
The cost for the REA 2, including the gypsum preparation plant, was
about $25,000,000,00. The various components, such as washer, REX, blowers,
gas ducting, instrumentation and controls, as well as installation, were
selected and ordered by the power plant. Engineering of the plant was
accomplished by an engineering company.
11-109
-------
SUMMARY
The availability of the REA 2 meets our expectations. Scaling the
REA 2 from REA 1, which is 1/3 as large, presented no problems. As a result,
a F(D plant capable of serving an 800 MW power block was developed.
Regenerative reheat has been proven to be possible, and it saves
energy. Definitive conclusions regarding corrosion can not as yet be made.
REFERENCES
W. Bosselmann, H. Weiler
Erste Betriebserfhrunge mit der Rauchgasentschwefelungsanlage
im Kraftwerk Wilhelmshaven
(VGB 1978)
W. Bosselmann, K-R. Hegemann, J. Leimkuhler
Weitere Betriebserfahrungen mit einer Rauchgasentschwefelungsanlage
in Wilhelmshaven und deren Obertragung auf die Anlagenerweiterung
(VGB 1981)
M. Frauenfeld
Ljunstrom - Gasvorwarmer zur Wiederaufheizung nassentschwefelter
Reingase
Technische Losungsmoglichkeiten zur Verminderung von Leckage,
Verschmutzung und Korrosion
(1982)
J. Leimkuhler, H. Weissert
Erst Betriebserfahrung mit der Rauchgasentschwefelungsanlage 2
im Kraftwerk Wilhelmshaven und Entsorgung der REA-Ruckstandsprodukte
(1982)
11-110
-------
THE SULF-X PROCESS
E. Shapiro, W. Ellison
-------
THE 5ULF-X PROCESS
By:
Edward Shapiro, PE
Pittsburgh Environmental Systems, Inc.
Pittsburgh, Pennsylvania 15236
William Ellison, PE
Ellison Consultants
Monrovia, Maryland 21770
ABSTRACT
The purpose of this paper is to report the status of development of SULF-X
technology for flue gas desulfurization and to review the design and applicability of the
Process in boiler service firing diverse coal fuels. Details are given of process
chemistry, design and operation, economics and past and future system demonstration
programs. The flexibility and attractiveness of the technology is shown in its use either
for SO?-removal-only or for simultaneous SO9/NO removal.
^ £ X
INTRODUCTION
The SULF-X Process is a wet absorption process that utilizes a slurry of
regenerated ferrous sulfide solids to achieve removal of 90 to 99% of sulfur dioxide
from boiler flue gases by wet scrubbing. It is technically feasible for use with all fossil-
fuel types. A coal-fired calciner regenerates both spent and oxidized forms of the
ferrous sulfide reagent, thereby converting collected sulfur dioxide into salable
elemental sulfur while minimizing formation of non-regenerable waste solids and liquids
requiring disposal.
Current technical evaluation and economic analysis of the SULF-X technology by
the Electric Power Research Institute shows that it is substantially less costly/than the
familiar Wellman-Lord regenerative flue gas desulfurization (FGD) process. The
study also indicates that it may be more attractive than limestone-based throwaway
solid-waste type systems, particularly at sites where waste management options are
limited. The process is primarily applicable in high sulfur coal-burning regions where
marketing of a commerical grade of by-product elemental sulfur is feasible and where
industrial by-product iron compounds such as pyrites (ferrous disulfide) are available for
use as process reagent makeup supply. With the benefit of revenues from sulfur by-
product sales of $150 or more per short ton and availability of inexpensive pyrites or
copperas (ferrous sulfate heptahydrate), SULF-X is highly competitive with limestone-
consuming FGD systems in common use. Pyrites may be derived from coal cleaning or
from mining of non-ferrous metals. Copperas is commonly available from industrial
waste processing including waste pickling liquor neutralization.
11-111
-------
PROCESS CHEMISTRY
The reaction steps in the SULF-X Process are schematically illustrated in the
simplified block diagram (Figure 1).
WET SCRUBBING
SO., Removal
An aqueous slurry containing a complex mixture of iron/sulfur compounds,
including finely divided ferrous sulfide (FeS) reagent in stoichiometric excess, is used to
absorb sulfur dioxide in a wet scrubber. Although the solubility of FeS in water is only 6
mg/liter (at 68°F), it reacts with absorbed SO2 and oxygen converting a portion of the
SO7 to elemental sulfur while being sulfidated and, in part, degraded to soluble ferrous
sulfate. Additionally, while the individual absorption reactions are numerous and
complex, the overall chemical reaction in absorption of SO2 may be expressed as
follows:
(c) FeS + S02 + 02^ (c-1) FeS(m) + Fe++ + SO4= + (1 -fy 5°
The quantity "n" can range from 1 to 10 depending on process conditions. The
quantity "c" is greater than 1 but less than 2, its magnitude varying depending on
equilibrium among tne diverse cnemical reactions taking place. In a typical design, rate
of FeS feed to the aDsorption reaction is based on the assumption that two moles are
required per mole of SO? absorbed (c = 2). The quantity "m", which ranges between 1
and 2, is a measure of th~e degree of sulfidation of the regeneated FeS and reflects the
spectrum of possible reaction products between FeS (fresh/regenerated) and FeS« (fully
spent reagent). The SULF-X Process is typically designed and operated to limit
elemental sulfur formation in the wet-scrubbing step (to maximize "n") so that difficult
to separate finely suspended sulfur will be yielded in the subsequent thermal
regeneration step where it is more easily isolated and recovered.
Intermediate aosorption reaction products include HS™, SO^, HSCC and S.-CC
(thiosultate). Secondary reactions not reflected in the overall absorption reaction
equation include oxidation of ferrous ions to ferric and oxidation of bisulfite and sulfite
ions to sulfate.
NUX Removal
The equation tor tne SCu-removal reaction above illustrates the oxygen absorption
capaoility of the process. Through retrofit modification or by provision in the initial
installation, the SULF-X Process may be designed to use this chemical reducinq
(deoxidizing) cnaracteriscic of its process slurry to achieve simultaneous removal of
NOx from boiler flue gas. Catalysis necessary to accomplish this is provided by
operating tne system so as to maintain approximately 40 grams per liter ferrous ion
(Fe ) concentration in the continuously recirculating scrubbing slurry. The overall
chemical reaction in this simultaneous absorption of SO? and NO may be expressed by
the following simplified single equation: x
(c) FeS + S02 + NO + i DZ ^+ (c-1) FeS(m) + Fe++ + SO= + (1 - JQ) S° + i N2t
11-112
-------
MM 6*5
II
SOLIDS
CLEAI 6AS
OUT
ACID 6AS
•BSORPTIOi
REGENERATED SLURRY
SOLIDS
DEMATERII6
LIQUOR
TRANP-SULFATE
CRYSTALLIZATION
COKE
THERMAL DMYJI6/
REfEIERATIOl
8Y COAL
ELENEITAL SULFUR
It-fRODUCT
LIQUOR
SOLIDS
Figure 1. SULF-X Process Flow Diagram
11-113
-------
Additional intermediate reaction products include Fe(NO)++ formed in initial
capture of NO. To remove most of the NO , substantially greater reaction time is
needed in the absorber than is required in collection of SO_ alone. Thus, for
simultaneous SCL/NO removal, the installation must provide, either through retrofit
modification or oy initial design, additional absorber capacity to increase the gas
residence time.
Note that the absorption-reduction environment in the process slurry chemically
reduces the collected NO to innocuous elemental nitrogen. By contrast, a
simultaneous SO?/NO removal process based on absorption-oxidation chemistry that
utilizes an oxidizing Xagent, converts collected NO to potentially water-polluting
oxidized forms such as nitrites (NO^ and nitrates (NOy.
Absorption-reduction can be distinguished from absorption-oxidation by a compar-
ison among familiar types of calcium-alkali scrubbing systems. An absorption-reduction
environment exists in some of the common scrubbing systems such as those using
magnesia-buffered lime and sodium-liquor scrubbing (dual-alkali PGD) that operate in
an unsaturated-CaSO, -mode with a significant ionic concentration of sulfite/bisulfite
(SO^/HSOp, an oxygen scavenger. On the other hand, lime/limestone FGD systems, as
most commonly applied, function without appreciable dissolved concentrations of such
chemical deoxidants. Their scrubbing slurries typically impose an absorption-oxidation
environment due to the presence of dissolved oxygen and are incapable of chemical
reduction.
SULFIDE/SULFATE REGENERATION
Suspended solids in the slurry bleed-off from the acid gas absorption step (Figure
1) are dewatered prior to roasting in an indirect-fired calciner. Most of the process
liquid from bleed dewatering, principally containing dissolved sodium sulfate (Na^SO.)
and ferrous sulfate (FeSO^), is used as a quench to cool and slurry the hot solids leaving
the calciner. The balance of the liquid is diverted to a crystallizer system to recover
sodium sulfate solids. The spent scrubber solids combined with the crystallizer solids
and a proportion of coke constitute the feed to the calcining step. In the calciner at
1200 to 1400 F, sulfide is regenerated and sulfur by-product is formed in accordance
with the equations as follows:
FeSl + FeS + 50
Na2S04 + 2 C -» Na2S + 2
To avoid oxidation of any of the sulfur yield, excess coke is supplied to the
calciner to provide a protective reducing atmosphere containing carbon monoxide (CO).
During the quenching step noted above, the sodium sulfide in the calcine forms FeS
precipitate:
Na2S + FeSO4 + H2O -> FesJ- + Na2SO4 + H2O
THROWAWAY WASTE GENERATED
The SULF-X Process produces a minimum of non-regenerable wastes and thereby
minimizes the environmental impact of acid gas emission control facilities.
11-114
-------
Solid Waste
Oxidation of scrubbing slurry results primarily in the conversion of sulfides to
sulfate, which is regenerated to sulfide by calcination. Thus, the quantity of throwaway
solid waste is negligible in comparison with the amount of ash and other solid wastes
generated by coal-fired boiler operations. A comparatively small proportion of ferrous
iron, estimated to be equivalent to less than 5% of the stoichiometric SO.-, removal rate
in pulverized coal-fired boiler applications, is oxidized to ferric oxide (Fe^O^) in the
scrubbing step. Ferric oxide is a tramp solid material that may be discarded from the
system through disposal of a small amount of dewatered scrubber-bleed solids.
However, the reducing conditions in the calcination step convert a large portion of the
ferric oxide solids to reactive ferrous compounds, thereby decreasing the net amount of
iron oxide waste production.
Liquid Waste
Chloride Management—
As in other wet scrubbing processes, hydrogen chloride contained in the inlet flue
gas is efficiently absorbed and must be purged in a liquid effluent. To prevent loss of
iron salts in this purge stream and to minimize complexity of wastewater management,
the absorbed chloride may be segregated in an isolated pre-scrubbing loop as is done in
application of other regenerative FGD processes.
Other Waste-ion Formation-
Chloride is the only non-jegenerable dissolved solid that tends to accumulate in
the system. Thiosulfate (S^CO, a common tramp compound that must be purged in
sodium-base regenerative FGD systems, is both formed and reconverted to bisulfite
(HSOQ in the SULF-X scrubbing step. Thus, purging of thiosulfate from the system is
not required. Moreover, unlike sodium sulfite based absorption-reduction processes
utilized for simultaneous SO^/NO removal, the SULF-X Process does not form
detectable quantities of dithionate, imidodisulfonate or other nitrogen/sulfur tramp ion
complexes.
SYSTEM DESIGN AND OPERATION
The general flow diagram for a commercial SULF-X Process installation in coal-
fired service is illustrated in Figure 2.
OVERVIEW OF DESIGN BASIS
Gas Absorption
As in current typical application of other flue gas desulfurization processes, an
electrostatic precipitator (or fabric filter), not shown in Figure 2, is used to control
stack fly-ash emission and to limit the amount of solid particulates entering and
accumulating in the wet scrubbing operation. In addition, a prescrubber loop as shown is
used upstream when required for absorption of other tramp materials, including HC1.
This limits their impact on the operation of the process and the control of process
chemical losses (through closed-loop recycle of SULF-X liquid and solid flows). Design
of large-scale SULF-X absorbers is tied closely to the prior development and continued
test/demonstration activity in the simultaneous removal of SO2 and NO . To ensure
mass transfer capacity sufficient to meet potential NO control objectives, the packed
11-115
-------
Legend
BFW - Boiler Fe*d Water
Equipment:
A - Pre»erubfo«r
B - Absorber
C - Thickener
D - Centrifuge I
E - Heat Exch«nf«e
P - CrysUUIzw
G - Centrifug* B
H - Dryer
1 - Dust Collector
J - Indirect-fired Regenerate*
K - Waste Heet Boiler
L - Sulfur Condenser
M - Incinerator-
N - Quench Tank
O - Ball MIU
P - Reclrculstion Tank
Q - Absorber Feed Tank
R - Flue Gas Reheater
Figure 2. SULF-X Process Flow Diagram
-------
absorber design utilizes fixed cross-fluted Munters packing material in counter-flow
scrubbing operation. In the design of the packed absorber in SO~-only service, a
recirculating slurry to gas flow ratio of approximately 60 U.S. gallons per thousand
actual cubic feet is used with an absorber gas-pressure-drop of 8 in. w.g. at a design
superficial gas velocity of 10 feet per second. In such applications concerned solely
with SO2 removal in the 85-95% efficiency range, a spray-tower-type absorber,
operated at a superficial gas velocity comparable to designs used in lime/limestone
slurry scrubbing systems, may be found to be preferable. Due to provisions in the
process for regeneration of sodium sulfate, the principal product of parasitic oxidation
in the absorber, the SULF-X absorption-reduction system can, uniguely, be applied to
emission sources such as bituminous-coal-fired stoker boilers that may contain oxygen
concentration in the raw flue gas as high as 14%. Thus, with this capability for
processing flue gases with a very high O^/SO^ ratio, the SULF-X Process can be applied
to pulverized coal boilers fueled with coal of all ranks and with either high or low sulfur
content.
Sodium Sulfate Cystallization
An atmospheric pressure crystallizer column is used to preferentially crystallize
the sodium sulfate make for regeneration to sulfide in the calcining operation. Reasons
supporting this method for recovering and reusing sulfate waste product are as follows:
Chemical reduction of sulfate is thereby accomplished without liberation of
SCL that accompanies thermal processing of alternative sulfate solids forms
such as FeSCh.
Iron compounds that are present catalyze the reduction of Na^SO. .
The minimum temperature for reduction does not exceed the design temp-
erature for simultaneous calcination of the spent FeS solids (FeS^ + I-^Q
forms).
Coke intermixed with the calciner feed materials can serve as an effective
reductant.
FeS subsequently precipitated by reaction of Na^S solids with dissolved
FeSO. is more reactive than the thermally regenerated FeS.
Precipitated FeS does not require milling, thereby minimizing the size of
milling equipment required for the total supply of FeS to the absorber.
Thermal Regeneration
After the feed has been dried in a steam-heated co-current rotary dryer, an
indirect coal-fired rotary drum kiln is used to roast and regenerate feed solids at a
temperature as high as 1400°F. Sulfur released during the calcination leaves the kiln as
a vapor and is recovered by condensation. By-product steam is generated in the sulfur
condensation step, as well as in an unfired waste heat boiler fed by the calciner exhaust
combustion gases.
Note that indirect calcination isolates the kiln combustion gas from the CO/CCL
process atmosphere laden with the sulfur product vapor. Thus, unlike FGD processes
such as the MgO process that use elevated temperatures requiring direct-calcination
operation, the SULF-X Process by-product yield and calcined solids do not become
intermixed with or contaminated by the combustion gases from calciner fuel firing.
11-117
-------
This, in turn, affords the opportunity to use the cheapest available fuel for the calcining
operation.
PROCESS CONTROL
Principal provisions for control of the process and for isolation and segregated
shutdown of individual process steps are comparatively simple.
Wet Scrubbing
The scrubbing operation is controlled by maintaining a set-point pH in the
scrubbing slurry through regulation of feed of the regenerated solids slurry entering the
absorber loop. Additionally, feed of this slurry and bleed-off of spent slurry to the
regeneration step are coordinated to maintain scrubbing slurry inventory and density in
an optimum range to ensure scrubbing effectiveness. Safeguards are provided to
prevent slurry-pH excursions that could impair gas cleaning efficiency or cause
objectional decomposition of process chemicals.
Decoupling of Process Steps
To afford means of temporarily shutting down individual process steps for
maintenance without interrupting overall system operation, a process inventory and
freeboard capacity are provided for storage of process materials including:
Dewatered spent scrubber solids in thickener-tank bottom (dryer feed)
Dried-cake (calciner feed)
Calciner quench-tank slurry (scrubber feed)
MATERIALS AND ENERGY UTILIZATION
A current, detailed process economics evaluation by Electric Power Research
Institute (EPRI) includes a preliminary conceptual design of a hypothetical SULF-X
Process installation for removal of SO2 from a 500 MW unit fired by 4.0% sulfur
bituminous coal. The consumptive use of materials and energy and the yield of by-
product elemental sulfur was estimated based on full-load operation at 500 MW.
Assigning applicable dollar values to each item as displayed below, it is seen that the
overall cost of materials and energy can be offset by by-product revenue. Thus, for one
day of 500 MW operation at full load (generating 12 million KWH), the cost summary is
as follows:
Consumption
Pryites, 2.8 tons/hr x 24 x $25/ton $ 1,680
Na-S, 0.3 tons/hr x 24 x $470/ton 3,384
Water, 727 gpm x 1440/1000 x $0.30/Mgal 314
Coke, 5.3 tons/hr x 24 x $51.30/ton 6,525
Coal, 1.3 tons/hr x 24 x $35/ton 1,092
No. 6 oil, 3 gpm x 1440 x $0.82/gal 3,542
11-118
-------
Consumption (Continued)
H.P. steam*, 100,500 Ib/hr x 24/1000 x $4/Mlb 9,648
L.P- steam, 4,500 Ib/hr x 24/1000 x $3 Mlb 324
Electric power, 12,800 KW x 24 x $0.045/KWH 13,824
Total Material and Energy Cost $40,333**
Production
Elemental sulfur, 9 tons/hr x 24 x $200/ton $43,200
Total By-Product Revenue $43,200**
PROCESS ECONOMICS
S02 REMOVAL ONLY
The detailed 1983 EPRI economic study of diverse flue gas desulfurization
processes at the 500 MW scale in 4% sulfur bituminous coal service estimates the cost
for the SULF-X Process as follows:
SULF-X PROCESS
(in December, 1982 dollars)
Levelized Busbar Cost
Capital Cost (Capital and Operating)
(Including contingencies) No By-Product Credit $75/ton Sulfur Credit
$/KW mills/KWH mills/KWH
295 22.8 20.0
FGD process selection is highly site-specific and will depend on items such as the
complexities anticipated in throwaway waste management and the actual revenues
available from by-product sales. The market value of the sulfur by-product
significantly affects the operating cost of the SULF-X Process. For example, with the
sulfur price projected at $200/ton, the levelized busbar cost with by-product credit
decreases to a very favorable (within the guidelines of the study) 15.4 mills/KWH.
SIMULTANEOUS SO2/NOX REMOVAL
1979 in-house studies of similar utility-scale applications for commercially
available combined SO2/NO removal facilities achieving 80% NOx removal confirm
that SULF-X is substantiallyXless costly than other available wet simultaneous SO2/NO
removal processes. Moreover, while capital cost is slightly higher than for combined
removal methods that utilize dry catalytic NO control, the SULF-X Process is
estimated to have lower annual revenue requirements than other available SO2/NO
removal methods including those based on dry NO removal.
/\
The levelized busbar cost of a 500 MW SULF-X installation designed for
ojected to be 4
This increment,
simultaneous SO2/NO removal achieving 80% removal of NO is projected to be 4
mills/KWH higher than that for SO?-only service reviewed above.
* 50°F indirect-type flue gas reheat
** Approximately 3.5 mills/KWH
11-119
-------
representing an increase of 20 to 25% over the cost of simple FGD capability, consists
almost entirely of capital costs for added parallel-operated absorber trains required to
provide increased gas residence time for high-efficiency NOx removal.
INDUSTRIAL BOILER APPLICATIONS
To take advantage of the favorable economics of large-scale sizing, small-
capacity coal-fired boilers in an industrialized region may be equipped with the wet-
scrubbing-system portion of the SULF-X Process while being served as a group by a
single, centrally located regeneration plant. The central plant would supply fresh
calcine to each scrubber and receive the spent sulfides on the return haul. The
attractiveness of the economics of such applications may approach that of utility-plant-
scale installations.
SULF-X PROCESS DEMONSTRATION OPERATION
Development of the SULF-X Process through laboratory and field test work was
completed in 1980. Design and construction of a 1.5 MW stoker-boiler size integrated
process installation for the Commonwealth of Pennsyvlania was completed in 1982.
This demonstration facility has been in round-the-clock operation since early 1983,
providing data for process scale-up to larger coal-fired boiler applications.
INTRODUCTION
The integrated SULF-X Process installation completed in 1982 is now in
demonstration operation at the bituminous coal-fired boiler plant of the Western Center
Hospital of the Commonwealth of Pennsylvania at Canonsburg, Pennsylvania, near
Pittsburgh. The system is designed for the simultaneous control of SO2 and NO
emissions from one of the plant boilers at a flue gas flow rate of approximately 7,000
acfm. Its current test/demonstration operation in 1983, under funding by the
Commonwealth of Pennsylvania, constitutes a critical interim stage in the scale-up and
application of the process through collection of steady-state operating data in
anticipation of its ultimate use in large commercial installations. The continuous,
integrated operation of this demonstration facility at the same time has afforded a
needed opportunity to study the effects on the overall chemistry of any accumulation in
the system inventory of reaction products and tramp materials, and to evaluate the
compatability of the installed equipment and instrumentation.
DESCRIPTION OF WESTERN CENTER FACILITIES
The SULF-X installation at Canonsburg is a $2 million facility dedicated by the
Commonwealth of Pennsylvania to a 12-month demonstration period. The system
advantageously utilizes off-the-shelf process equipment and control instrumentation to
provide a practical replica of the process facilities that are expected to be ultimately
applied in large commercial installations. Thus, the system models a large-scale
installation, using applicable mass transfer devices and mechanical equipment to
duplicate primary unit operations such as the gas absorption, slurry dewatering and
thermal/chemical regeneration steps, all of which are sub-processes commonly used in
industry. See the Flow Diagram, Figure 3.
Flue gas cleaning equipment provided includes a 3 feet diameter radial blade
booster fan, a lined carbon steel gas pre-cooler and a 73 feet high by 5 feet diameter
11-120
-------
4 r *•«.***/*
Equipment
A
a
c
o
E
r
c
H
1
J
K
L
M
N
0 -
Seturetor (Pre-CooO
Abenroer
Thickener
Overflow TV*
Filter Preei
Drrer
CtlclfMT
Sulfur Condeneir
Quench TV*
Attritor
AbKutwr Feed
TV*
Sodium Sulflde
TV*
Preciplmkm
TV*
Precipiutlon
TV*
Sodium Sulfat*
Solution TV*
Figure 3. Flow Diagram - SULF-X System at Western Center
-------
FRP scrubber (absorber) vessel using cross-fluted fixed plastic packing. Slurry-bleed
dewatering equipment consists of a 10 feet diameter lined thickener with overflow tank,
a horizontal plate-and-frame filter containing seventeen 1 meter x 1 meter vertical
filter plates and a steam-heated rotary dryer. The regeneration system consists of an
indirect gas-fired rotary calciner, a calcine quench tank and a steam-cooled vertical
shell-and-tube sulfur condenser. For supply of regenerated reagent to the gas
absorption system, the system includes an agitated-media-type calcine slurry crusher
(attritor) and an absorber fresh slurry feed tank.*
Without adversely affecting the simulation of large commercial system opera-
tions, the overall demonstration program has been simplified by excluding sodium
sulfate crystallization facilities. Instead, dissolved sodium sulfate is purged from the
system by blowdown. Purchased sodium sulfide is used (in lieu of sodium sulfide that
would otherwise be available from calcination of sodium sulfate) to sustain the quench
tank reaction, which recovers dissolved ferrous sulfate by converting it to precipitated
FeS reagent. To further simplify the operation of this small-capacity demonstration
operation, the calciner is fired with LPG (propane) instead of coal. Since calcination is
by indirect heat transfer both in the commercial scale and Western Center designs, use
of alternative fuels does not affect process chemistry or the validity of the test
operations. In addition, pre-cooling and water saturation of the boiler flue gas is
carried out in a simple gas quenching step utilizing recirculating process slurry. In
absence of an isolated pre-scrubber wastewater purge, chloride will be purged by the
discard of the dissolved sodium sulfate blowdown stream. As a result of other
simplifications, calciner combustion gases are exhausted without by-product steam
generation and sulfur condenser off-gas is recycled to the absorber inlet instead of
being treated in an incinerator.
PROBLEMS IN COMMISSIONING OF THE SYSTEM
Principal startup/operating problems encountered with the initial system installa-
tion included difficulties in mechanically transferring the filter cake to the dryer, in
moving the filter cake through the dryer and in oxidation of the dried solids by air that
was entering the process side of the calciner. Because of the small batch-scale nature
of the filtering operation at Western Center, movement of wet filter cake is expected
to continue to require special operator attention during continuous process runs. The
original dryer and calciner were replaced in late 1982 with the more compatible
equipment described above and these operations are now considered to be free of major
problems. Adverse oxidation of wet cake was experienced in early operation of the new
steam-heated rotary dryer, but this problem has been eliminated. Retrofit modifica-
tions have also been made to the sulfur vapor vacuum and condensing systems to avoid
solidification of sulfur before and after the condenser.
OVERVIEW OF SYSTEM TESTING WORK
The Western Center absorber has successfully demonstrated up to 99+% removal
of SO and up to 93% removal of NO in 2%-sulfur bituminous coal operation. At a
superficial velocity no greater than 2 feet per second, NO removal efficiency of 90% is
obtained with essentially 100% SO2 removal. Figure 4 shows the graphical relationship
*For initial ferrous chemical charge and subsequent Fe makeup supply, the Western
Center operation has used copperas produced from waste pickling liquor neutralization.
11-122
-------
I
h-"
U)
100
80
o
lu
QC 60
«0
IU
Q
X
O
lu
O
O
ft
40
20
I I
FERROUS ION CONCENTRATION
35 GRAMS/LITER
IN SCRUBBING SLURRY
100
200
300
400
FLUE GAS VELOCITY (FT/MIN)
Figure 4. Nitrogen Oxides Removal
vs. Superficial Flue
Gas Velocity
-------
between NO removal and absorber superficial gas velocity for the present absorber
configuratior^ functioning as a simple single-stage counter-current contactor. Other
variables affecting NO removal rates are slurry temperature and viscosity and FeS
particle size. Figure £ shows the inlet SO2 flue gas concentrations and the percent
absorbed when the system was operated for simultaneous SO-and NOx removal. At a
superficial gas velocity in the scrubber of approximately ID feet per second and a
liquid/gas ratio of 60 U.S. gallons per thousand cubic feet, typical of operation of
commercial FGD systems, SO2 removal at an efficiency of 95+% is accompanied by only
token NO removal.
The absorption system has operated automatically with the pH controller
regulating fresh slurry addition and spent slurry removal to maintain a constant, pre-set
pK Spent slurry has dewatered well in the thickener and filter, and the FeS
precipitation and chemical makeup systems have performed as designed. Comprehen-
sive data are being collected for a broad range of operating conditions to achieve
parametric testing that will optimize system design for NOx removal levels up to 80%,
and to yield a data base to be used for scale-up.
CONTINUOUS PROCESS OPERATIONS
Chemistry of Steady-State Process Streams
The iron, sulfur and sodium species in the recycled scrubbing slurry were
monitored throughout the project for both daily process control and overall system
performance evaluation. Relatively high concentrations of sodium sulfate (Na^SOj
were measured but were anticipated since the sulfate crystallizer had not been included
as part of this system and a low liquid purge rate had been maintained. This rate of
liquid waste blowdown was adjusted to maintain the dissolved Na^SO^ concentration in
the slurry below its saturation point.
Analyses indicated that steady-state conditions had been achieved by the fourth
operating month. Chemical consumptions and product concentrations have been
consistent with earlier studies and the previously identified chemical reactions.
Approximately 17% more ferrous sulfide was reacted in the absorber than was expected.
This increase is believed to have been caused by the unusually high oxygen content in
the flue gas, ranging from 10 to 14%. A typical analysis of process slurry is presented
in Table 1.
Quantification of Solid and Liquid Waste Discharge
There was no significant solid waste discharged from the system. Approximately
13,600 gallons of sodium sulfate liquid waste were purged from the system. Analysis of
this waste liquor is as follows:
Iron, Fe++ 0.6 g/1 Sulfate (SO
Sodium, Na 34.0
Other cations
(Ca, Mg, Al) 7.6
pH 6.4
Sulfite (SOp
Thiosulfate (S-Op
Chloride (CL~)
76.9 g/1
0.1
9.3
1.0
11-124
-------
ho
Ul
INLET SO CONCENTRATIONS (ppm)
t— •
N) *» 0\ CO 0
O O O O O
O 0 O 0 0 0
-
-
-
so^ cowc.
I
1
1
1
J
1
1
1
EMO
1
ML
1
1
1
1
1
100
99 )?
to
98 i
03
97 m
o
96
6 7 8 9 10 11
OPERATING TIME (WEEKS)
12
13
14
15
16
Fiqure 5. Inlet Flue Gas S0~ Concentrations and
Percent Absorbed by SULF-X System
-------
TABLE 1,
5ULF-X PROCESS SIMULTANEOUS SCU/NO^
REMOVAL
Recycled Scrubber
Slurry Composition
Iron (Fe )
Sodium (Na )
Other Cations*
(Ca, Mg, Al)
Iron, total
Sulfur, total
% Suspended solids
% Dissolved solids
Specific gravity, slurry
pH, slurry
Dissolved Solids,
49.9 g/1
48.0
0.3
6.42%
7.25%
5 . 72%
22.1 %
1.29
5.4
grams/liter
Sulfate (SOp 164.0 g/1
Sulfite (SOp 0.1
Thiosulfate (S9O^) 20.7
£ J
Chloride (Cl~) 1.7
Calcium and magnesium salts are contained as anti-caking agents in the purchased
iron sulfate makeup; aluminum is leached from the scrubber fly ash accumulating
in the slurry.
11-126
-------
Taking into account the change in system chemical inventory, the quantity of the
dissolved sulfur species in excess of that contributed by ferrous sulfate makeup
represents approximately 7% of the sulfur dioxide absorbed during the test period.
Sulfur (sulfate) losses in the blowdown can be expected to significantly decrease and the
ratio fo chloride to dissolved sulfur species correspondingly increase for a SULF-X
system installation that includes a sodium sulfate crystallizer.
SUMMARY OF FINDINGS IN WESTERN CENTER DEMONSTRATION RUNS TO DATE
The installation and operation of the demonstration SULF-X facility for the
Commonwealth of Pennsylvania has been an important and successful step in the
continuing development of the Process. The data and information collected during all
phases of the project are valuable additions to the engineering and operating criteria
needed for building a larger SULF-X FGD installation of optimized design.
During the operation, equipment and instrumentation compatability in the face of
variances in process operation was thoroughly evaluated. The results served to identify
means of overcoming potential problem areas and confirmed the accuracy of previously
established process design criteria:
The packed tower functioned as an effective mass transfer unit and should
be used in future systems requiring substantial NO removal.
X
Use of pH for control of process chemistry and slurry make-up rate was
effective and reliable.
Dithionates and imidodisulfonates did not accumulate in the recirculated
slurry nor in the sodium sulfate solution waste.
Slurry bleed from the absorber was readily dewatered with a pressure filter
after thickening. This final dewatering could probably be performed with a
centrifuge.
Material transfer of the filter cake was troublesome and requires additional
investigation if batch-wise filtration is to be used reliably in routine system
operation.
To ensure efficient regeneration and sulfur recovery, precautions must be
taken to avoid excessive oxidation of the solids during the drying and
calcining steps.
Iron oxide was converted in the calciner to an active ferrous form, thereby
substantially decreasing the anticipated waste solids blowdown rate.
PROJECTED PROCESS DEMONSTRATION AT LARGER SCALE
A new and larger scale demonstration of the process is now planned in conjunction
with future development of process licensees for commercial application of SULF-X
technology in the U.S.A., Canada, Japan, West Germany and other industrialized
countries. In a new system installation for this purpose, it is anticipated that a
pulverized coal-fired boiler facility at the 10 to 60 MW size range would provide a
minimum-risk basis for ultimate scale-up of the process to utility plant applications
larger than 300 MW. At the same time, it should be possible to reliably quantify the
benefits, simplifications and anticipated economies in the process design resulting from
typically low flue gas oxygen content in pulverized-fired service.
11-127
-------
EXPANDED WESTERN CENTER DEMONSTRATION
FOR U.S. DEPARTMENT OF ENERGY
A new test/demonstration program is now being carried out at Western Center
under the sponsorship of DOE to define optimum design/operating conditions for
achievement of 90% NO removal in simultaneous SO2/NOx scrubbing operation. A
systematic sequence of test runs is being conducted in a continuous integrated closed-
loop mode to evaluate performance, identify important operating variables and provide
data for assessing the technical and economic feasibility of high efficiency, simultan-
eous SO-/NO removal. Additionally, process runs are being made to demonstrate
regeneration of sodium sulfate with coke as a reductant. Alternative fuel and reductant
materials are also being investigated.
90% NO REMOVAL
X
The piping of the two in-series absorber beds, originally arranged to operate as a
simple single-stage counter-current scrubber, is being modified to provide two
separately operated scrubbing stages. A series of test runs to evaluate performance and
study important operating variables will be conducted. Parameters to be adjusted and
assessed include liquid/gas flow ratio of the individual absorber stages, superficial gas
velocity and chemistry of inlet slurry to each stage. Parametric testing will take into
account the effect of other variables that may significantly affect the NO absorption
rate.
RUNS WITH MODIFIED CALCINER OPERATION
The calciner test program will model-test typical large-scale SULF-X Process
operation in which the regeneration system is regenerating all tramp sulfate formed in
the scrubbing step to fresh, usable sulfide reagent. This will entail continuous addition
of coke reductant and sodium sulfate solids at the calciner feed inlet. Individual test
runs will be made to investigate alternative use of petroleum coke, metallurgical coke
and boiler-house coal as the reductant.
LONG-TERM RELIABILITY RUN
After establishing an applicable mode and optimum conditions for SULF-X
operation at 90% NO removal, a continuous 30-day system reliability run will be made
to verify the design basis for such commercial operation. Quantities and quality of
waste discharges during sustained operation will be monitored to assess system waste
production characteristics and to identify any specific provisions that would be needed
for treatment and ultimate disposal of the waste.
CONCEPTUAL COMMERCIAL-SCALE DESIGN FOR 90+% SO?, 90% NO REMOVAL
£- A
Based on data generated and conclusions drawn during the field test work, a
conceptual SULF-X Process design will be made for a hypothetical 500 MW power plant
burning high sulfur bituminous coal. This design will serve as a basis for assessment of
the technical and economic feasibility of this mode of operation of the SULF-X Process.
11-128
-------
CONCLUSIONS
SULF-X is an effective new wet-scrubbing-type regenerative flue gas desulfuriza-
tion process applicable to all fossil fuels. It has been demonstrated to advantageously
use and regenerate ferrous sulfide originating from industrial iron by-product wastes to
convert flue gas SO2 into usable elemental sulfur while producing negligible solid and
liquid waste except common chloride liquor purge.
The process has demonstrated an inherent versatility in functioning as either a
system for SO2 removal only, or for simultaneous SOo/NO treatment. Cost projections
indicate that rt is an economically attractive alternative to throwaway-waste-type FGD
system where a regional market for by-product sulfur exists, particularly in applications
where solid waste management costs would be expected to be disproportionately high.
Moreover, the capital investment for small to medium-sized SULF-X units can be
minimized if concentration of industry in a region allows a central regenerating plant to
serve a number of individual gas scrubbing units.
Additionally, the SULF-X system is technically and economically attractive where
nitrogen oxides removal is also necessary at present or in the future. The system lends
itself to being installed and operated as an SO2-only FGD system and then later being
readily modified to also remove nitrogen oxides.
REFERENCES
1. Stearns-Roger Engineering Corporation, Economics of FGD; Electric Power
Research Institute, draft report RP1610-1, March 1983.
2. Pollution Technology Review No. 82, New Developments in Flue Gas Desulfuriza-
tion Technology; edited by M. Satriana, Noyes Data Corporation, Park Ridge, New
Jersey, 1981.
11-129
-------
APPENDIX
Attendees
-------
EPA/EPRI SYMPOSIUM ON FLUE GAS DESULFURIZATION
November 1-4, 1983
Sheraton New Orleans Hotel
New Orleans, Louisiana
List of Attendees
R. S. Abraham
Engineering Design Coordinator
Florida Power & Light Co.
P.O. Box 529100
Miami, Florida 33152
305/552-3834
Robert J. Abrams
Sales Manager
Bishopric Products Co.
4413 Kings Run Drive
Cincinnati, Ohio 45232
413/641-0500
toesim Abuaf
Mechanical Engineer
General Electric Co.
Building 37, Room 647
Schenectady, New York 12301
518/385-3654
Steven J. Achtner
Sales Representative
M. W. Kellogg Co.
433 Hackensack Avenue
Hackensack, New Jersey 07601
201/646-1000 (ext. 2435)
Radford C. Adams
Program Manager
TRW
P.O. Box 13000
Research Triangle Park, NC 27709
D. D. Agarwal
Manager Corrosion Alloys
Cabot Corp.
1U2U W. Park Ave.
Kokomo, Indiana 46901
317/456-6031
Stefan Ahman
Senior Chemical Engineer
Flalt Industri AB
Kvarnvagen, Vaxjo
Sweden S-35187
046/470-87276
0. Ainsworth
Lab Director
Dow Chemical Co.
P.O. Box 150
Plaquemine, Louisiana 70764
504/389-1620
Sy A. Ali
Executive Director, Environmental
Programs
Public Service Indiana
1000 E. Main St.
Plainfield, Indiana 46168
317/838-1229
David L. Almand
Western Regional Sales Manager
Joy Manufacturing Co.
P.O. box 2744, Terminal Annex
Los Angeles, California 90051
213/240-2300
Harold R. Althen
Marketing Manager
Peabody Process Systems
835 Hope Street
Stamford, CT 06907
203/327-7000
Chuck Altin
Project Manager
Ebasco Services
145 Technology Park
Norcross, Georgia 30092
404/449-5800
A-l
-------
Marlin Anderson
Director, Technical Services
Crestmont Assoc.
P.O. Box 770
Central City, Kentucky 42330
502/754-9460
Gary Andes
Air Pollution Control Engineer
Gi1bert-Commonwealth
P.O. Box 1498
Reading, Pennsylvania 19603
215/775-2600, ext. 2160
Jumpei Ando
Faculty of Science & Engineering
Chuo University
1-13-27 Kasuga
Bunkyo-Ku
Tokyo 112
japan
Amjad H. Ansari
Mechanical Engineer
Stone & Webster
16430 Park Ten Place
Houston, Texas 77064
713/492-4148
Roger Antonie
Utility Sales Engineer
Warman International Inc.
2701 S. Stoughton Rd.
Madison, Wisconsin 53716
Bertrans Anz
Vice President
United Engineer & Constructors, Inc.
9111 Cross Park Dr.
Knoxville, Tennessee 37923
615/690-8610
C. William Arrington
President
Crestmont Assoc.
P.O. Box 770
Central City, Kentucky 42330
502/754-9460
Franklin A. Ayer
Symposium Coordinator
Research Triangle Inst.
P.O. Box 12194
Research Triangle Park, North
Carolina 27709
919/541-6260
Bill Babcock
Sales Representative
Marblehead Lime
Salt Lake City, Utah 84103
601/364-7117
Lothar Bachmann
President
Bachman Ind.
29 Lexington St.
Lewiston, Maine 04240
207/784-2338
Ronald J. Bacskai
Vice-President, Marketing
Conversion Systems, Inc.
115 Gibraltar Rd.
Horsham, PA 19044
215/441-5920
Brian Bahor
Application Engineer
Wheelabrator-Frye Inc.
600 Grant St.
Pittsburgh, Pennsylvania 15219
412/288-7519
Even Bakke
Vice President
Peabody
635 Hope St.
Stamford, Connecticut 06907
302/327-7000
Armand A. Balasco
Engineering Consultant
Arthur D. Little, Inc.
20 Acorn Park
Cambridge, Massachusetts 02140
617/864-5770
A-2
-------
h. A. Bambrough
head, Major Combustion Sources
Environment Canada
351 Blvd. St. Joseph, Hull,
Quebec, Canada
819/997-1220
John H. Banks
Market Development
Ashland Chemical Co.
18 Foxhall PI.
Scarsdale, New York 10563
914/472-6953
Bruce P. Bannon
Director, Marketing
RMI Co.
1000 Warren Ave.
Niles, Ohio 44446
216/652-9951
Jim Barlow
Mechanical Engineer
Benham Holway Power Group
5300 S. Yale
Tulsa, Oklahoma 74135
918/492-2411
Yves Barthel
Project Leader
Institut Francais du Petrole
1 et 4, avenue de Bois-Preau
92506 Rueil Malmaison
FRANCE
Daniel C. Beal
Sales Engineer
beneral Electric Co.
2015 Spring Rd
Oak Brook, Illinois 60521
312/986-3021
R. T. Beall
Vice President/Sales Manager,
Aggregates/Cement
Ciifford-Hill & Company, Inc.
P.O. Box 42127
Dallas, Texas 75247
214/258-7321
Joseph L. Beals
Wallace & Tiernan/Pennwalt
2001 Midwest Rd.
Oak Brook, Illinois 60521
312/620-8820
Earl R. Beaver
Business Development Director
Monsanto
800 N. Lindburgh
St. Louois, Missouri 63167
314/694-8620
Steven Becker
Supervisory Engineer
Southwestern Public Service Co.
P.O. box 1261
Amarillo, Texas 79170
806/378-2441
E. B. Beckman
Marketing Engineer
Stone & Webster Engineering
250 W. 34th St.
New York, New York 10119
212/290-6109
Ann Behl
Research Assistant Tech. Writer
Radian Corporation
8501 Mo Pac Blvd.
Austin, Texas 78758
512/454-4797
Ranier Benninghaus
Thyssen Environmental Systems, Inc.
333 Meadow!and Pkwy.
Secaucus, New Jersey 07094
201/330-2600
Keith Benton
Development Engineer
Combustion Engineering
31 Inverness Center Pkwy.
Bham, Alabama 35243
205/967-9100
Jim Berding
Sales, Marketing Manager
CLOW
P.O. Bix 24
Florence, Kentucky 41042
606/283-2121
A-3
-------
Chris Bernabo
Executive Director
Interagency Task Force on Acia Precipitation
722 Jackson PI., N.W.
Washington, D.C. 20006
Albert H. Berst
Engineering Manager, Pollution
Control
Zurn Industries, Inc.
P.O. Box 2206
Birmingham Alabama 35201
205/252-2161
P. A. Bhat
Senior Engineer II
Pennsylvania Electric Co.
1001 Broad St.
Johnstown, Pennsylvania 15907
814/533-8560
Robert T. Bianchi
Sales
Marblehead Lime Co.
300 W. Washington
Chicago, Illinois 60606
312/263-4490
Edward Biedell
Manager, FGD Applications
MikroPul Corp.
10 Chatham Rd.
Summit, New Jersey 07901
201/273-6360
Cal Billings
President
Advanced Energy Systems
400 Cities Service Hwy
Sulphur, Louisiana 70663
318/625-4816
Thomas B. Blair
Senior Program Manager
Radian Corp.
8501 Mo-Pac
P.O. Box 9948
Austin, TX 78766
512/454-4797
Donald W. Blind
Mechanical Engineer
Tennessee Valley Authjority
400 W. Summit Hill
Knoxville, Tennessee 37902
615/632-4626
Julian Blomley
Environmental Analyst
Middle South Services, Inc.
934 Gravier St., P.O. Box 61000
New Orleans, Louisiana 70161
5U4/569-4741
Gary M. Blythe
Senior Chemical Engineer
Radian Corp.
P.O. Box 9948
Austin, Texas 78766
512/454-4797
Michael L. Bolind
Product Manager
United States Gypsum Co.
101 S. Wacker Dr.
Chicago, Illinois 60016
312/321-4371
Michael F. Bellinger
Environmental Chemist
Union Electric Co.
1901 Gratiot St.
St. Louis, Missouri 63103
314/554-3652
Dr. Arthur Boni
Vice President
Physical Sciences, Inc
P.O. Box 3100, Research Park
Andover, Massachusetts 01810
617/475-9030
Dave Bordson
Mechanical Engineer
Minnesota Pollution Control Agency
1935 W. Co. Read B-2
Roseville, Minnesota 55113
612/296-7780
A-A
-------
Ed Boucher
U.S. Sales
ABCO Plastics, Inc.
45 Accord Park
Norwell, Massachusetts
617/878-5068
02061
Williard L. Boward, Jr.
Senior Process Engineer
FMC Corp.
231 N. Martingale Rd
Schaumburg, Illinois 60194-2098
312/843-1700
George H. Bowen
President
Electric Utility Services
163 W. Saddle River Rd.
Saddle River, New Jersey
201/337-1788
Warren Bowman
Senior Development Specialist
E.I. du Pont de Nemours & Co. Inc.
1007 Market St., D-13121
Wilmington, Delaware 19898
302/773-3654
Donald E. Boyd
Director Marketing and Sales
Flakt Environmental
Systems Division
P.O. Box 87
Knoxville, Tennessee 37901
615/693-7550
Paul A. Boyd
Senior Chemical Engineer
U.S. EPA
1200 6th Ave.
Seattle, Washington 98101
26/442-1567
Edward G. Boyer, Jr-
Cromby Station Superintendent
Philadelphia Electric Co.
2301 Market St.
Philadelphia, Pennsylvania 19101
215/933-8995
Richard C. Boynton
Manager, Environmental Systems/Sales
Development
General Electric Co.
P.O. Box 7600
Stamford, Connecticut 06904
203/357-4975
Wallace H. Bradley
Vice President of Engineering
Austell Box Board Corp.
3100 Washington St.
Austell, Georgia 30001
404/948-3100
Jene Bramel
Vice President Sales
Air Clean Damper Co.
Blue Ash Rd.
Cinciannati, Ohio 45236
513/793-1253
Atly Brasher
Permit Program Manager
Louisiana DNR
P.O. Box 44066
Baton Rouge, Louisiana 70804
504/342-8942
Theodore G. Brna
Program Manager for Dry FGD
U.S. EPA
IERL, MD-61
Research Triangle Park, NC 27711
919/541-2683
Klaus L. Bro
Niro Atomizer, Inc.
9165 Rumsey Ra.
Columbia, Maryland 20145
301/997-8700
Norman I. Brody
Technical Sales
Stebbins Engineering and
Manufacturing Co.
4830 North River Rd.
Port Allen, Louisiana 70767-3898
504/343-6671
A-5
-------
Clare E. Brown
General Coordinator of Operations
Pennsylvania Power Co.
1 E. Washington St.
New Castle, Pennsylvania 16103
412/656-5408
F. William Brownell
Lawyer
Hunton & Williams
1919 Pennsylvania Ave., N.W.
Washington, D.C. 20036
202/223-8650
Charles K. Bruhl
Engineering Manager
Chiyoda Int. Corp.
13uu Park Place blag.
Seattle, Washington 98101
2U6/624-9350
C. P- Brundrett
Manager, Market Development
W. R. Grace & Co.
10 E. Baltimore St.
Baltimore, Maryland 21203
301/659-9192
Doug Brusseau
Senior Mechanical Engineer
Salt River Project
P.O. Box 1018
St. Johns, Arizona 85536
602/337-4131
David P. Burford
Research Engineer
Southern Company Services
P.O. Box 2625
Birmingham, Alabama 35202
205/870-6329
Eugene A. Burns, PhD
Vice President & Manager
S-CUBED
3398 Carmel Mountain Rd.
San Diego, California 92121
619/453-0060
Edbert Buter
Project Engineer
N.V. KEMA
Utrechtseweg 310
Arnhem, NETHERLANDS
3185-457057 ext. 3424
Vernon Butler
Senior Environmental Engineer
U.S. EPA
6th and Walnut St.
Philadelphia, Pennsylvania
215/597-3697
Rob Byrne
Product Manager
Research Cottrell
P.O. Box 1500
Somerville, New Jersey 08876
201/685-4244
Ivor E. Campbell
Consultant
Ivor E. Campbell & Assoc.
P.O. Box 153
New Albany, Ohio 43054
614/855-2183
Bernard S. Camponeschi
Industry Manager
FMC Corp.
231 N. Martingale Ra
Schaumburg, Illinois 60194-2098
312/843-1700
W. R. Cares
Senior Research Chemist
M. W. Kellog Co.
16200 Park Row
Houston, Texas 77084
713/492-2500
Teri L. Carter
Process Engineer
Conoco, Inc.
Ponca City, Oklahoma 74603
4U5/767-3131
A-6
-------
Douglas Carter
Engineer
U.S. Department of Energy
10UO Independence Ave.
Washington, D.C. 24585
202/252-4770
Ray Centa
Sales Application Engineer
Pultrusions Corp.
1331 S. Chillicothe Rd.
Aurora, Ohio 44202
216/562-5201
Denise I. Cessna
Program Coordinator
Pennsylvania Electric Co.
1001 Broad St.
Johnstown, Pennsylvania 15907
714/533-6667
Pui Kun Roland Chan
Research Analyst
University of Texas at Austin
University of Texas
Austin, Texas 78712
512/471-4851
John C. S. Chang
Senior Chemical Engineer
Acurex Corp.
P.O. Box 13109
Research Triangle Park, NC 27709
919/549-8915
Richard Chapman
Senior Engineer
Cor Star
2118 Milvia St.
Berkeley, California 94704
415/540-4100
Art Chappie
Technical Sales Representative
ABCO Plastics, Inc.
45 Accord Park
Norwell, Massachusetts 02061
617/878-5068
Gregory H. Cheng
Director/R&D
Ducon Co., The
147 E. 2nd Street
Mineola, New York 11740
516/741-6100
Suhas T. Chitnis
Manager, Gaseous
General Electric
200 N. 7th St.
Lebanon, Pennsylvania
717/274-7231
Collection Products
Environmental Services
17042
Ove B. Christiansen
Niro Atomizer, Inc.
9165 Rumsey Rd.
Columbia, Maryland 20145
301/997-8700
Roger Christman
Consulting Engineer
Christman & Assoc.
11708 Bowman Green Dr.
Reston, Virginia 22090
703/435-3219
M. Yaqub Chughtai
Project Manager
L&C Steinmueller GmbH
P.O. Box 100855
Gummersbach, West Germany
00261-85 2930
Darryl D. Ciliberto
Engineer
Tampa Electric Co.
P.O. Box 111
Tampa, Florida 33601
813/228-4111
Gerald M. Clancy
Vice President, Process Development
Pritchard Corp.
4625 Roanoke Pkwy.
Kansas City, Missouri 64112
816/531-9500
A-7
-------
Curtis W. Clarkson
President
Clarkson Company
3430 W. Bayshore
Palo Alto, California
416/494-1010
94303
Donald D. Clasen
Marketing Manager
Chiyoda International Corp.
1300 Park Place Bldg.
Seattle, Washington 98101
206/624-9350
Larry Cliver
Regional Sales Manager
Feeco International
3913 Algoma Rd.
kreen Bay, Wisconsin 54301
414/468-1000
Robert Cmiel
Power Engineer
San Miguel Electric
Box 280
Jourdanton, Texas 78026
512/784-3411
W. James Cole
Program Director
New York State Energy
Research Authority
2 Rockefeller Plaza
Albany, New York 12223
518/465-6251
Ed Coleman
President
Herbst & Assoc.
P.O. Box 90989
Houston, Texas
713/440-6090
J. David Colley
Radian Corp.
P.O. Box 9948
Austin, Texas 78766
512/454-4797
Marshall F. Conover
Program Manager
Radian Corp.
P.O. Box 9948
Austin, Texas 78766
512/454-4797
Charles B. Cooper
Senior Staff
Arthur D. Little, Inc.
Acorn Park
Cambridge, Massachusetts 02140
617/864-5770
Frederick M. Coppersmith
Manager, Advanced Fossil Fuels
Consolidated Edison Co. of N.Y.
4 Irving Place
New York City, New York 10003
212/460-3098
John Coulston
Sales Engineer
Spraying Systems Co.
North Ave. at Schmale Rd.
Wheaton, Illinois 60188
312/665-5000
Bob Cowen
President
Martek, Inc.
85 Research Rd.
hingham, Massachusetts 02043
617/749-6992
Maxwell E. Cox
Senior Technical Service Representative
Kerr-McGee Chemical Corp.
Kerr-McGee Center
Oklahoma City, Oklahoma 73102
405/270-3373
A. J. Cozza
Manager, Ash Systems Engineering
Combustion Engineering
1000 Prospect Hill Rd.
Windsor, Connecticut 06095
203/688-1911
A-i
-------
Anthony J. Craig
Sales Representative
Anthony J. Craig, Mfg. Rep.
10 Park Place
Butler, New Jersey 07405
201/838-8997
George Cranston
Peaboqy Process Systems
835 Hope St.
Stamford, Connecticut 06907
203/327-7000
William R. Cress
Manager, Engineering Studies
Allegheny Power Service Corp.
800 Cabin hill Dr.
Greensburg, Pennsylvania 15601
412/838-6721
Laird Crocker
Group Supervisor
U.S. Bureau of Mines
729 Arapeen Dr.
Salt Lake City, Utah 84108
801/524-6151
Phillip B. Crommelin, Jr.
Consultant
P.O. Box 38
Stanton, New Jersey 08885
201/236-2324
Richard V. Cross
Supervising Engineer, Mechanical Design
Union Electric Co.
P.O. Box 149
St. Louis, Missouri 63166
314/554-2671
James L. Crowe
Projects Manager
Tennesse Valley Authority
1150 CST2
Chattanooga, Tennessee 37401
615/751-5651
John Cunic
Senior Staff Engineer
Exxon Research & Engineering Co.
P.O. Box 101
Florham Park, New Jersey 07932
201/765-6471
Bob Cunningham
Supt. Production
City Water, Light and Power
7th & Monroe
Springfield, Illinois 62757
217/789-2238
Michael J. Cyran
FMC Corp.
P.O. Box 8
Princeton, New Jersey 08501
609/452-2300
Peter W. Dacey
Team Member, EAS
IEA Coal Research
14115 Lower Grosvenor PI.
London,
England
William C. Daley
Chief Engineer
Virginia Electric & Power Co.
7th & Franklin Streets
Richmond, Virginia 26661
804/771-6269
George Dal ton
Project Designer
B&R Engineering
620 Wilson Avenue
Downsview, Ontario, Canada 1
416/630-7741
Stuart Dal ton
Project Manager
Electric Power Research Inst.
P.O. Box 10412
Palo Alto, California 94303
415/855-2467
Ashok S. Damle
Research Engineer
Research Triangle Inst.
P.O. Box 12194
Research Triangle Park, NC 27709
919/541-5829
A-9
-------
Earl H. Darrough
Senior Sales Representative
Cabot Corp.
1020 West Park Ave.
Kokomo, Indiana 46901
317/456-6000
Steve Davidson
Project Engineer
brown and Caldwell
1501 M. Broadway
Walnut Creek, California 94596
415/937-9010
Lawrence Davidson
FGD Specialist
Stone & Webster
245 Summer St.
Boston, Massachusetts 02107
617/589-2381
R. H. Davis
Professor & Chairman
Florida State University
Tallahassee, Florida 32306
904/644-2867
Richard L. Davis
Manager of Engineering
Koppers Co. Inc.
2701 Koppers Building
Pittsburgh, Pennsylvania
412/227-2610
15219
H. Joel Dean
Sales Manager
Mosser Damper Co.
500 Tilghman St.
Allentown, Pennsylvania
215/395-4900
18104
Ronald N. Deardoff
Senior Engineer
Dayton Power & Light
P.O. Box 1247
Dayton, Ohio 45401
513/224-6374
Co.
Edward D. Deboer
Senior Application Engineer
General Electric Co.
607 Tall an Building
Chattanooga, Tennessee 37402
615/755-5011
R. Dean Delleney
Program Manager
Radian Corp.
3401 La Grande
Sacremento, California 95823
916/421-8700
Christian Demeter
Assistant Director
Reyes Assoc.
1633 16th Street N.W.
Wshington, D.C. 20009
J. Herbert Dempsey
Project Manager
Acurex Corp.
P.O. Box 13109
Research Triangle Park, N.C. 27709
919/549-6915
Charles E. Dene,,
Project Manager
Electric Power Research Inst.
P.O. Box 10412
Palo Alto, California 94303
415/855-2425
Mark S. Dershowitz
Marketing Representative
Shell Oil Co.
200 North Dairy Ashford
Houston, Texas 77079
713/670-2835
Prakash H. Dhargalkar
Manager, Process Engineering
Research-Cottrell
P.O. Box 15uO
Somerville, New Jersey 08876
201/685-4295
A-10
-------
N. N. Dharmarajan
Senior Engineer, Emissions Control
Central & Southwest Services, Inc.
P.O. Box 220164
Dallas, Texas 75222
214/754-1373
Jim Dickerman
Senior Program Manager
Radian Corp.
P.O. Box 9948
Austin, Texas 78766
512/454-4797
William L. Didden
Regional Sales Manager
Conversion Systems, Inc.
115 Gibraltar Rd.
horsham, PA 19044
215/441-5920
David L. Dillehay
Product Line Manager
KVb, Inc.
18006 Skypark
Irvine, California 92714
714/250-6259
Antonio J. DoVale Jr.
Senior Process Engineer
M. W. Kellogg Co.
433 Hackensack Ave.
Hackensack, New Jersey 07601
201/646-1000
Mark L. Doane
Generation Sales Engineer
General Electric Co.
1015 Locust St.
St. Louis, Missouri 63101
314/342-7727
Ir. N. A. Doets
Project Manager
N. V. PNEM
P.O. Box 7
4930 AA Geertruidenberg,
Holland
01621-82542
Temple E. Donaldson
Plant Manager
Central Illinois Light Co.
300 Liberty St.
Peoria, Illinois 61602
309/672-5271
J. R. Donnelly
Assistant Chief Nuclear, Enviornmental
Engineer
Bechtel Power Corp.
P.O. Box 2166
Houston, Texas 77252-2354645
Van Dostveen
Expert
Ministry Environmental Protection
P.O. Box 450
Leidschendam, Netherlands 2260 MB
70/209367
Angelo Dounoloulos
Project Development Manager
General Electric Co.
P.O. Box 8
Schenectady, New York 12301
518/385-9980
Marty Downey
District Marketing Manager
Pullman Power Products
270 McCarty Dr.
Houston, Texas 77001
713/672-2491
William Downs
Research Specialist
Babcock & Wilcox Co.
1562 Beeson
Alliance, Ohio 44601
216/821-9110
Dennis C. Drehmel
Limb. Dev. Branch Chief
U.S. EPA
IERL, MD-61
Research Triangle Park, NC 27711
919/541-7505
A-ll
-------
F. Carter Dreves
Director - Marketing Communications
Wheelabrator Air Pollution Control Div
600 Grant St.
Pittsburgh, PA 15219
412/288-7325
Charles J. Drummond
Chemical Engineer
DOE (PETC)
P.O. Box 10940
Pittsburgh, Pennsylvania 15236
412/676-6011
Michael J. Du Bois
Project Engineer
Commonwealth Edison
P.O. Box 767
Chicago, Illinois 60690
312/294-8483
Thomas Dudick
President
Duaick Corrosion-Proof
576 E. Highland Rd.
Macedonia, Ohio 44056
216/467-1970
Ronald G. Duffy
Peabody Process Systems
835 Hope St.
Stamford, Connecticut 06907
203/327-7000
David M. Dunkle
Manager, Marketing Services
Conversion Systems, Inc.
115 Gibraltar Rd.
Horsham, Pennsylvania 19044
215/441-5920
Antonia Duran-Lopez
Chief Environment Department
Empresa Nacional De
Electricidad, S.A.
Velazquez, 132
f'iadria-6, Spain
(91)2.61.63.00
Roger W. Dutton
Partner
Black & Veatch
P.O. Box 8405
Kansas City, Missouri
913/967-7265
64114
John Dydo
Santa Fe Energy Co.
10737 Shoemaker
Santa Fe Springs, California
Gene H. Dyer
Manager, Process Technology
Department
Bechtel National, Inc.
P.O. Box 3965
San Francisco, California 94119
415/768-4201
John M. Ebrey
Vice President - Marketing & Planning
Lodge-Cottrell Operations
601 Jefferson
Houston, Texas 77002
713/750-2094
Martha Edens
Dow Chemical
P.O. Box 150
Plaquemine, Louisiana
504/389-8000
70764
Richard Egan
Manager, Marketing Opertions
Munters Corp.
P.O. Box 6428
Ft. Meyers, Florida 33911
813/936-1555
Joan Ehzeny
Reporter
BNA
1231 25th Street NW
Washington, D.C.
202/452-4423
Henry W. Elder
Assistant Director
Tennessee Valley Authority
501 Chemical Engineering Bldg.
Muscle Shoals, Alabama 35660
205/386-2514
A-12
-------
M. Fred Ellis
Research Economist
Illinois Energy Resources Commission
3U3 Stratton Office Bldg.
Springfield, Illinois 62706
217/782-8220
William Ellison
Director
Ellison Consultants
4966 Tall Oaks Dr.
Monrovia, Maryland 21770
301/86b-53U2
Jack E. Englick
Sales Manager
FMC Corp.
231 N. Martingale Rd
Schaumburg, Illinois 60194-2098
Donald A. Erdman
Manager, Project Engineering
PEPCO
1900 Pennsylvania Ave., N.U.
Washington, D. C. 20068
202/872-3487
Olav Erga
Professor, Chemical Engineering Dept.
Norway Inst. of Tech.
Sem Saelandsv 4
Trondheim
Norway N-7034-NTH Tr. heim
07/594120
Douglas A. Erickson
Operations Engineer
Getty Oil Co.
Box 197X Rt. 1
Bakersfield, California 93308
805/399-2961
Robert L. Eriksen
Environmental Control Supervisor
Basin Electric Power Coop.
1717 E. Interstate Ave.
Bismarck, North Dakota 58501
701/223-0441, ext. 2144
Michael Esche
President, Dipl. -Ing., -Wittsch-ing,
Saarberg-Holter Umwelttechnik GmbH
Hafenstr. 6
D-6600 Saarbruken, West Germany
01681/32-105-06
Thomas F. Evans
Senior Research Specialist
Niagara Mohawk
300 Erie Blvd. W.
Syracuse, New York 13202
315/474-1511
Brent L. Evans
Division Manager
Stebbins Engineering and
Manufacturing Co.
4831 North River Rd.
Port Allen, Louisiana 70767-3898
504/343-6671
Rita E. Ewing
Environmental Supervisor
Utah International Inc.
550 California St.
San Francisco, California 94560
415/774-2363
Paul S. Farber
Project Manager
Argonne National Lab.
9700 S. Cass Ave.
Argonne, Illinois 60439
312/972-2000
Bobby Faulkner
Associate Director
Allis-Chalmers Corp.
P.O. Box 512
Milwaukee, Wisconsin 53201
414/475-4624
Thomas J. Feeley, III
Research Scientist
DOE, PGh Energy Res. Ctr.
626 Cochran Mill Rd.
Pittsburgh, Pennsylvania 15236
412/675-6079
A-13
-------
Karsten Felsvang
Niro Atomizer Inc.
Oakland Ridge Industrial Cntr.
9165 Rumsey Rd.
Columbia, Maryland 21045
301/997-8700
Bryan W. Ferguson
Chemical Engineer
Texas Power and Light
P.O. Box 226331
Dallas, Texas 75266
214/748-5411
John M. Ferraro
Senior Staff Engineer
Exxon Research & Engineering Co.
P.O. Box 101
Florham Park, New Jersey 07932
201/765-1600
Forjtenlehner
Project Manager
Veest-Alpine
Linz, Austria A-hono
0732/585-8626
John R. Field, Jr.
Senior Contract Administrator
Virginia Electric and Power Co.
P.O. Box 26666
Richmond, Virginia 23261
804/771-3625
Joe Finer
Regional Manager
Joy Manufacturing Co.
11707 E. 51st Ave.
Denver, Colorado 80239
303/371-6140
P. G. Finlay
Head, Electric Power Section
Environment Canada
Ottawa, Ontario. Canada K1A 108
819/997-1220
Gerry Flander
Senior Sales Representative
Cabot Corp.
1800 Place Dunant, St. Bruno
Quebec, Canada J3V 242
514/653-3312
Timothy 0. Flora
Special Assignment
Ohio Edison Co.
76 S. Main St.
Akron, Ohio 44308
216/364-7959
Ray Fmaddalone
Program Manager
TRW
One Space Park
Redondo Beach, California 90278
Gerald F. Foley
Principal Engineer
burns and Roe, Inc.
185 Crosswaus Park Dr.
Woodbury, New Jersey 11797
516/677-2274
Robert Forbus
Project Manager
Central & Southwest Services
P.O. Box 220164
Dallas, Texas 75222
214/754-1373
Jerry Ford
Manager
Custom Pipe Coating
P.O. Box 3274
Houston, Texas 77253
713/675-2324
George C. Ford
Proprietor
Environmental Management Assoc.
607 Rosedale Rd.
Princeton, New Jersey 08540
609/924-0601
Owen F. Fortune
Project Manager DFGD
General Electric Co.
200 N. 7th St.
Lebanon, Pennsylvania 17042
717/274-7223
A-14
-------
Landon D. Fox
head Mechanical Engineer
Tennessee Valley Authority
4UO W. Summit Hill
Knoxville, Tennessee 37902
615/632-362:4
Grant Frame
Manager Scrubber
Flakt Canada, Ltd.
P.O. Box 5060, STNF
Ottawa, Ontario
Canada K2C3p9
613/226-3300
Norman W. Frank
President
Ebara International Corp.
1 Northgate Square
Greensburg, Pennsylvania 15601
412/832-1200
Thomas Frank
Manager of Proposals
GE Environmental Services
200 North 7th St.
Lebanon, Pennsylvania 17042
717/274-7146
George Fraunfelder
Eastern Regional Sales Manager
Komline-Sanderson Engineering Corp.
12 Holland Ave.
Peapack, New Jersey 07977
2U1/234-1000
killiam F. Frazier
Staff Engineer
Virginia Electric & Power Co.
P.O. Box 564
Richmond, Virigina 23204
804/771-6147
Paul E. Fredette
Manager, Technical Center/Program
Marketing
Midland Ross Corp.
900 North Westwood
Toledo, Ohio 43696
419/537-6426
Mark D. Freeman
Supervisory Engineer
Southwestern Public Service Co.
P.O. Box 1261
Annville, Texas 79170
806/378-2184
Donald T. Freese
Research Department
Betz Laboratories, Inc.
Somerton Rd.
Trevose, Pennsylvania 19047
215/355-3300
Steven Frey
Environmental Engineer
U.S. EPA
1860 Lincoln St.
Denver, Colorado 80013
303/837-6131
Dan Froelich
Manager, Air Quality Design
Burns & McDonnell Engineering Co.
P.O. Box 173
Kansas City, MO 64141
Joseph L. Gaines
Staff Engineer
Rust International Corp.
P.O. Box 101
Birmingham, Alabama 35201
205/254-4472
Ignatius J. Gallo
Assistant Manager, Market Research
Texasgulf Chemicals Co.
P.O. Box 30321
Raleigh, NC 27622-0321
919/829-2810
Marsh Galloway
Manager, FGD Marketing
Ceil cote Company
140 Sheldon Rd.
Berea, Ohio 44017
216/243-0700
A-15
-------
P. R. Gambarani
Manager
General Electric Environmental Services
5 Penn Plaza
New York, New York 10001
212/613-3203
Bernie Gardey
Enviornmental Manager
C.F. Braun & Co.
R.D. *1
Bangor, Pennsylvania 18013
717/897-5148
John Gaynor
Research Staff Member
United States Gypsum
700 N. Highway 45
Libertyville, Illinois 60048
312/362-9797
Dennis C. Gehri
Program Manager
Rockwell Int.
8900 De Soto Ave.
Canoga Park, California
213/700-4413
91304
D. B. Geottel
Supervisor, FGD Systems
Winyah Generating Station
Santee Cooper
P.O. Box 1275
Georgetown, South Carolina 29440
803/546-4171
Mario A. Gialanella
Regional Sales Manager
Joy Manufacturing Co.
7 Corporate Park Dr.
White Plains, New York 10604
914/694-1364
Elizabeth D. Gibson
Consultant
Arthur D. Little, Inc.
Acorn Park
Cambridge, Massachusetts 02140
617/664-5770
Carl A. Gilbert
President
Dravo Lime Co.
3600 Neville Rd.
Pittsburgh, Pennsylvania 15225
412/777-5559
Tom Gil Christ
Chemist
Colorado-Ute Electric
P.O. Box 1307
Craig, Colorado 81626
Al Giovanetti
Supervising Engineer
Davy McKee Corp.
0. 0. Drawer 5000
Lakeland, Florida 33803
613/646-7311
Dennis L. Glancy
Superintendent FGD Operations
Southern Indiana Gas & Electric
P.O. Box 569
Evansville, Indiana 47741
812/424-6411
Robert J. Gleason
Director Research & Development
Research-Cottrell, Inc.
Somerville, New Jersey 06876
201/685-4884
Donald E. Glowe
Senior Engineer
Texas Research Inst.
9063 W. Bee Caves Rd.
Austin, Texas 78746
512/263-2101
Robert F. Goecker
Eastern Region Sales Manager
Marcona Ocean Industries, Ltd.
1001 N.W. 62nd St., Suite 200
Ft. Lauaerdale, Florida 33309
305/776-4000
A-16
-------
Klaus Goldschmidt
Environmental Protection Dept.
VEBA Kraftwerke Ruhr AG
Bergmannskluckstr- 41-43
D-465U Gelsenkirchen
Federal Republic of Germany D-4650
0209/601-5851
Richard to. Goodwin
Principal
Environmental Engineering Consultant
261 W. 71 St.
New York, New York 10023
Steve R. Graham
Project Manager
Betz Laboratories
3657 Buckskin Tl. W.
Jacksonville, Florida 32211
904/744-0053
Richard J. Grant
Manager, Environmental Affairs
Central Illinois Public Service
607 East Adams St.
Springfield, Illinois 62701
Co,
Jack Greene
Administrative Officer
U.S. EPA IERL RTP
IERL-RTP (MD-60)
Research Triangle Park, NC 27711
919/541-2903
Donald Gross
Supervisory General Engineer
Cdr, Cml R&D Cen
Attn: DRSMC-CLT-I (A)
APG, Maryland 21010
301/671-4102
K. 0. Groves
Research Manager
Dow Chemical U.S.A.
2020 Dow Center
Midland, MI 48640
517/636-3246
Larry Gruber
Generation Sales Engineer
General Electric Co.
607 Tallan Bldg., 2 Union Square
Chattanooga, Tennessee 37402
615/775-5009
Klaus E. Gude
Niro Atomizer
9165 Rumsey Rd.
Columbia, Maryland
301/997-8700
20145
Dave Guetig
Assistant Plant Superintendent
Inaaianapolis Power & Light Co.
P.O. Box 436
Petersburg, Indiana 47567
812/354-8801
D. Guidetti
Business Development Manager
SOIMI SPA
5100 Westheimer
Houston, Texas 77056
713/961-0873
Connie Guilbeau
Sales Represenative
Xerox Corp.
417-419 Baronne St.
New Orleans, Louisiana 70112
504/568-9015
Navin K. Gupta
Principal FGD Engineer
Stone & Webster Engineering Corp.
3 Executive Campus
Cherry Hill, New Jersey 08034
609/482-4246
C. Richard Hach
Engineer
Tampa Electric Co.
P.O. Box 111
Tampa, Florida 33601
813/677-9141
Albert Hackl
University Professor
Institut fur Verfahrenstechnik
Tecnischen Universitat Wien
A-1060 WIEN, Getreidemarkt 9
Austria
0222/5601-4726
der
A-17
-------
Donald K. Hagar
President
Damper Design, Inc.
P.O. Box 2045
Bethlehem, Pennsylvania 18001
215/861-0111
Qazi Haider
Chemical Engineer
Industrial Generating Co.
P.O. Box 1111
Rockdale, Texas 76567
512/446-5861
F. Allen Hall
Field Salesman
Cabot Corp.
4650 S. Pinemone, Suite 130
Houston, Texas 77041
713/462-2177
J. A. Hall
Manager, Henderson Technical Lab.
TIMET
P.O. Box 2128
Henderson, Nevada 89015
702/564-5831
hark Hal pern
Project Licensing Coordinator
Potomac Electric Power Co.
1900 Pennsylvania Ave., N.W.
Washington, D.C. 20068
202/331-6489
David Ham
Manager
Physical Sciences, Inc.
P.O. box 3100, Research Park
Andover, Massachusetts 01810
617/475-9030
Joseph J. Hammond
Fbb Sales Engineer
Zurn Industries
P.O. Box 2206
Birmingham, Alabama 35201
205/252-2181
Doug Hammontree
Mgr. Air Quality Design Section
Burns & McDonnell Engineering Co.
P.O. Box 173
Kansas City, MO 64141
816/333-4375
Svend Keis Hansen
Vice President
Niro Atomizer, Inc.
9165 Rumsey Rd.
Columbia, Maryland 20145
301/997-8700
Frank M. Harbison
Asst. Environmental Analyst
Louisiana Power & Light
317 Barone
New Orleans, Louisiana 70130
504/595-2308
0. W. Hargrove, Jr.
Senior Staff Engineer
Radian Corp.
805 Mopac Blvd.
Austin, Texas 78759
512/454-4797
John Harkness
Chemical Engineer
Argonne National Lab.
EES Division, Bldg. 362
Argonne, Illinois 60439
312/972-7636
Richard C. Harrington
Senior Engineer
Cleveland Electric Illuminating
P.O. box 5000
Cleveland, Ohio 44101
216/622-9800
Thomas L. Hart
Chemical Engineer
American Electric
1 Riverside Plaza
Columbus, Ohio
614/223-3472
Power Service
A-18
-------
John L. Haslbeck
Engineer
NOXSo Corp.
2625 H. C. Mathis Dr.
Paaucah, Kentucky 42001
502/444-6474
John hattrup
Meteorologist
Baltimore Gas & Electric Co.
116 N. Howard St.
Baltimore, Maryland 21201
301/234-6427
Larry E. Haulter
Sales Representative
Cabot Corp.
1020 W. Park Ave.
Kokomo, Indiana 46901
317/456-6073
John R. Hawksworth
Supervisor, Operations Support
Virginia Electric and Power Co.
P.O. Box 26666
Richmond, Virginia 23261
804/771-4633
Kent D. Hedrick
Environmenal Enginer
Kentucky Utilities Co.
One Quality St.
Lexington, Kentucky 40507
606/255-1461 ext. 542
Carl-Rudolf Hegemann
General Manager
G. Bischoff
baertner Str 44, D-4300
Essen, W. Germany
0207-233037
Glen 0. Hein
Vice President of Sales
Marblehead Lime Co.
300 West Washington St.
Chicago, Illinois 60606
312/263-4490
Michael Heisel
Process Engineer
Linde
6023 Hoellriegelskrenth
F.R. Germany
Rex Helfant
Manager, Generation Programs
Con Edison
4 Irving Place
New York, New York 10003
212/460-3987
Paul D. Hemphill
Scrubber Project Manager
Dresser Industries
601 Jefferson, 27th Floor, Dressor Tower
Houston, Texas 77002
713/972-6011
A. Hennico
Technical Sales Manager
Institut Francais Du Petrole
1-4 Avenue de Bois Preau
Rueil Malmaison 925u6
France
Robert A. Hentges
Senior Engineer
Procter & Gamble Co.
6105 Center Hill Rd.
Cincinnati, Ohio 45224
513/659-5787
David S. Henzel
Technical Services Engineer
Dravo Lime Co.
3600 Neville Rd.
Pittsburgh, Pennsylvania 15225
412/777-5556
William D. Herrin
Supervisor, Air
Alabama Power Co.
P.O. Box 2641
Birmingham, Alabama 35291
205/250-4124
A-19
-------
Robert L. Hershey
Vice President
Science Management Corp.
2101 L St., M.W., Suite 903
Washington, D.C. 20037
202/293-5700
H. Frederick Hess
Operations Manager
FMC Corp.
231 N. Martingale Rd.
Schaumburg, Illinois 60194-2098
312/843-1700
R. A. Hewitt
Senior Engineer
Texas Utilities
2001 Bryan
Dallas, Texas 75201
214/653-4800
Barry R. Hickenbottom
Mechanical Engineer
U.S. Navy-NEESA
Code 111C
Port Hueneme, California 93043
b05/982-5292
Wayne Hickok
Senior Production Engineer
Cooperative Power Assoc.
14615 Lone Oak Rd.
Eden Prarie, Minnesota 55344
612/937-8599
Kent Higgins
Vice President
Koch Process Systems
20 Walkup Dr.
Westboro, Massachusetts 01581
617/366-9111
Richard Hills
Manager of Research
Pittsburgh Des Moines
Neville Island
Pittsburgh, Pennsylvania 15225
412/331-3000, ext. 655
Robert G. Hilton
Director of Programs Management
Monier Resources, Inc.
45 N.E. Loop 410, Suite 700
San Antonio, Texas 78216
Maurice Hixon
Plant Superintendent
Board of Muncipal Utilities
138 N. Prarie
Sikeston, Missouri 63801
314/471-5000
Otto H. Hoegberg
Chief Engineer
MW Kellogg
433 Hackensack Ave.
Hackensack, New Jersey 07601
201/646-1000
Art Hoekstra
President
Sirco
4412 Aicholtz Road
Cincinnati, Ohio 45245
513/752-4700
Jerry Hoffman
Director, Project Development
Burns & McDonnell Engineering Co.
P.O. Box 173
Kansas City, MO 64141
816/333-4375
David C. Hoffman
Vice President, Technical Services
Chemical Lime, Inc.
6000 Western PI., Suite 489
Ft. Worth, Texas 76107
817/732-8164
Kent R. Hoffman
Chemical Engineer
Public Service Co. of New Mexico
Alvarado Square
Albuquerque, New Mexico 87158
505/848-2964
A-20
-------
T. W. Hojnicki
Manager
Argonne National Laboratory
97UO South Cass Ave.
Argonne, Illinois 60439
312/972-5821
Gerald A. Hoi linden
Branch Chief
Tennessee Valley Authority
113U Chestnut St., Towers II
Chattanooga, Tennessee 37401
615/751-3584
Jack Hoi lister
Senior Vice President
Cleveland Cliffs Iron Co.
Huntington Bldg.
Cleveland, Ohio 44115-1448
216/241-2356
Richard G. Hooper
Project Manager
Electric Power Research Inst.
P.O. Box 1U577
Denver, Colorado 80210
303/936-7281
William M. Horton
Chemical Engineer
Lower Colorado River Authority
P.O. Box 220
Austin, Texas 78767
512/473-3544
Robert W. Hospodarec
Environmental Director
Fluor Engineers, Inc.
2310 Kelvin
Irvine, California 92712
714/966-5073
John B. Howard
Manager, Power Production/Construction
Alabama Electric Coop., Inc.
P.O. Box 550
Andalusia, Alabama 36420
205/222-2571
Bill Hughes
Arkansas Lime Co.
P.O. Box 2356
Batesville, Arkansas 72501
501/793-2301
George A. Hugick
Sales Engineer
Kennedy Van Saun Corp.
R.R. St.
Danville, PA 17821
717/275-3050
James Hung
Senior Environmental Analyst
Cajun Electric Power
10719 Airline Hwy.
Baton Rouge, Louisiana 70816
504/291-3060
Joseph Hunt
Market Development Manager
Allegheny Ludlum Steel
Research Center
Brackenridge, Pennsylvania 15014
412/226-2000
James E. Hunt
President
Concord Scientific Corp.
2 Tippett Road
Downsview, Ontario M3H 2V2
416/630-6331
Thomas B. Hurst
Manager Product Development
Babcock & Wilcox
20 S. Van Buren Ave.
Barberton, Ohio 44203
216/860-2674
Howard Hurwitz
Manager, Process Engineering
General Electric Environ Svs.
5 Penn Plaza
New York, New York 10001
212/613-3175
A-21
-------
Robert C. Hyde
F.G.D. Product Manager
Joy Manufacturing Co.
4565 Colorado Blvd.
Los Angeles, California 90039
213/240-2300 ext. 233
Akitaka Ide
Project Manager
Chiyoda International Corp.
1300 Park Place Bldg.
Seattle, Washington 98101
206/624-9350
Mr. hiaeo Idemura
Managing Director
Chiyoda International Corp.
1300 Park Place Bldg.
Seattle, Washington 96101
206/624-9350
W. Ijdeveld
ESMIL International
De Boelelaan 7
Amsterdam, Netherlands
020/541-1054
Paul A. Ireland
Manager, Environmental Control
Stearns-Roger
P.O. Box 5888
Denver, Colorado 80217
303/692-3420
Marty W. Irwin
Research Director
Indiana Department of Commerce
One North Capitol, Suite 700
Indianapolis, Indiana 46204-2248
317/232-8818
Sam Jacobsson
Western Regional Corp.
P.O. Box 6428
Ft. Meyers, Florida 33911
813/936-1555
Bryan J. Jankura
Research Engineer
Babcock & Wilcox
1562 beeson
Alliance, Ohio 44601
216/821-9110 ext. 391
James B. Jarvis
Radian Corp.
P.O. Box 9948
Austin, Texas 78766
512/454-4797
Clarence B. Jeffcoat
Manager, Cross Generating Station
Santee Cooper
P.O. Box 98
Cross, South Carolina 29436
803/351-4586
Frits Jellema
General Manager, APC Division
ESMIL (Intl.)
De Boelelaan 7
Amsterdam, Netherlands
020-5411054
Stephen D. Jenkins
Senior Engineer
Tampa Electric Co.
P.O. Box 111
Tampa, Florida 33601
813/28-4111
Sarah Jenkins
Planning Analyst
Wisconsin Public Service Commission
P.O. Box 7854
Madison, Wisconsin 53707
608/266-5990
Robert M. Jensen
Engineering Specialist
Bechtel Power Corp.
50 Beale St.
San Francisco, California 94105
415/768-1323
A-22
-------
Bob Jewell
District Manager
Research Cottrell
13231 Champion Forest Dr.
Houston, Texas 77459
713/440-0468
Dennis Johnson
Field Process Engineer
Babcock & Wilcox
20 South Van Buren Avenue
Barberton, Ohio 44203
216/860-6324
David I. Johnson
Executive Director
Coal Technology
10703-A Stand iff
Houston, Texas 77099
713/879-8929
Robert M. Johnson
Group Manager, Research
Gold Bond Building Products
1650 Military Rd.
Buffalo, New York 14217
716/873-9750
Lyndon Johnson
Air Quality Supervisor
Muscatine Power and Water
3205 Cedar St.
Muscatine, Iowa 52761
319/263-2631
Howard J. Johnson
Engineer
Ohio EPA
361 E. Broad St.
Columbus, Ohio 43216
614/466-6116
Carl ton A. Johnson
Peabody Process Systems
835 Hope St.
Stamford, Connecticut 06907
203/327-7000
Robert P. Johnson
Research Projects Manager
Pennsylvania Power & Light Co.
2 N. 9th St.
Allentown, Pennsylvania 18101
215/770-5151
C. W. Johnston
Marketing Mgr., Special Products
Badische Corp.
602 Copper Rd.
Freeport, Texas 77541
409/238-6237
Martin A. Jones
Environmental Engineer
Cliffs Engineering, Inc.
P.O. Box 1211
Rifle, Colorado 81650
303/625-2445
Robert L. Jones
Power Research Sales Engineer
General Electric Co.
4410 El Camino Real
Los Altos, California 94022
415/949-1042
J. R. Jones
Vice President
Gitford-Hi 11 &
P.O. Box 47127
Dallas, Texas 75247
214/258-7330
Marketing
Co., Inc.
S. M. Jones
Marketing Manager
MPSI
Box M312
York, Pennsylvania
717/843-8671
17402
Julian W. Jones
Senior Chemical Engineer
U.S. EPA
IERL, MD-61
Research Triangle Park, NC 27711
919/541-2489
A-23
-------
Larry Jones
Environmental Engineer
U.S. EPA, OAQPS, ESED, SDB
MD-13
Research Triangle Park, N.C,
919/541-5624
27711
James I. Joubert
Branch Chief
U.S. Department of Energy
P.O. Box 10940
Pittsburgh, Pennsylvania 15236
412/675-5716
Peter Judersleben
General Manager
Knauf/Research-Cottrel1
Iphofen,
West Germany
John H. Juzwiak
Manager, Utility Operations
Conversion Systems Inc.
115 Gibraltar Rd.
Horsham, Pennsylvania
215/441-5900
Yohi Kameoka
Chiyoda Int.
1300 Park Place
Seattle, Washington 98101
206/624-9350
Joel Y. Kamya
Project Engineer
Boston Edison Co.
800 Boylston St.
Boston, Massachusetts
617/424-3250
02199
Toshio Kanai
Technology and Engineering Division
Chiyoda International Corp.
1300 Park Place Bldg.
Seattle, Washington 98101
206/624-9350
Ann Marie Kanon
Program Secretary
Electric Power Research Institute
P.O. Box 10412
Palo Alto, California
415/855-2466
Ira E. Kanter
Senior Engineer
Westinghouse R&D Center
1310 Beulah Rd.
Pittsburgh, Pennsylvania 15235
412/256-5808
Marilyn Kaplan
Editor
Mcllvaine Co.
2970 Maria Ave.
Northbrook, Illinois 60062
312/272-0010
Steven M. Kaplan
Niro Atomizer
9165 Rumsey Rd.
Columbia, Maryland
301/997-8700
21045
Norman Kaplan
Chemical Engineer
U.S. EPA
IERL, MD-61
Research Triangle Park, NC 27711
919/541-2556
Steve M. Katzberger
Supervisor, Emission Control
Sargent & Lundy Engineers
55 E. Monroe
Chicago, Illinois 60603
312/269-6672
Ronald H. Kaye
Peabody Process Systems
835 Hope St.
Stamford, Connecticut 06907
203/327-7000
T. J. Kayhart
Manager of Sr. Projects
Allied Chemical Co.
P.O. Box 1087 R
Morristown, New Jersey 07960
201/455-4927
Mike Keckritz
Design Engineer
Illinois Power
500 S. 27th St.
Decatur, Illinois 62525
217/424-6962
A-24
-------
James R. Kendle
Regional Manager
FhC Corp.
231 N. Martingale Rd.
Schaumburg, Illinois 60194-2098
312/843-170U
Thomas R. Kendrick, III
President
Pittsburgh Environmental Systems, Inc.
67 Old Clairton Rd.
Pleasant Hills, Pennsylvania 15236
412/653-75UU
Ernest E. Kern
Staff Engineer
Houston Lighting & Power
P.O. Box 1700
Houston, Texas 77001
713/481-7608
Rick Kesler
Senior Applications Engineer
Mine & Smelter
3600 Race
Denver, Colorado 80205
303/296-8700
James R. Kessling
Senior Engineer
Houston Lighting & Power Co.
12301 Kurland Dr.
Houston, Texas 77034
713/481-7921
John W. Kife
(aroup Vice President
Joy Manufacturing Co.
4565 Colorado Blvd.
Los Angeles, California 90068
213/240-2300 ext. 500
Kilgroe
Coal Cleaning
James D.
Manager,
U.S. EPA
IERL, MD-61
Research Triangle Park, NC 27711
919/541-2854
Lawrence P- King
Senior Marketing Specialist
Babcock & Wilcox
1562 Beeson St.
Alliance, Ohio 44601
216/821-9110
Minesh Kinkhabwala
Commercial Manager
Thyssen Environmental Systems, Inc.
333 Meadowland, Pkwy.
Secaucus, New Jersey 07094
201/330-2600
William R. Kins
Supervisor, Air Systems Eng.
Owens Corning Fiberglass
Toledo, Ohio 43659
419/248-6027
Dennis L. Kirchner
Air Quality Engineer
Central 111. Public Service Co.
607 East Adams St.
Springfield, IL 62701
217/523-3600
Noel W. Kirshenbaum
Manager, Mineral Projects
Placer U.S. Inc.
1 California St.
San Francisco, California 94111
415/986-U740
Jonas Klingspor
M. S. Chemical Engineer
Dept. of Chemical Engineering
P.O. Box 240, S-22007
LUND, SWEDEN
OU46-46-108277
Peter Klose
ERM, Inc.
999 W. Chester Pike
West Chester, Pennsylvania 19380
215/696-9110
G. H. Koch
Projects Manager
Battelle
505 King Ave.
Columbus, Ohio 43201
614/424-4480
A-25
-------
Bernard J. Koch
Technical Coordinator
Conoco Coal Research Div.
4000 Brownsville Rd.
Library, Pennsylvania 15129
412/854-6612
Y. Kogawa
Managing Director
Chiyoda International Engineering
1300 Park Place Bldg.
Seattle, Washington 98101
206/624-9350
M. Koike
Chiyoda International Engineering
1300 Park Place Bldg.
Seattle, Washington 98101
206/624-9350
Nubuo Kojima
Process Engineer
Mitsubishi
520 Madison Ave.
New York, New York
212/605-2006
Ken Kondo
Project Engineer
Mitsubishi
520 Madison Ave.
New York, New York
212/605-2006
Dennis Kostick
Physical Scientist
U.S. Bureau of Mines
2401 E. St.
Washington, D.C. 20241
202/634-1177
Karl M. Kozak
Marketing Specialist
Rockwell Intl., Energy Systems Group
8900 De Soto Ave.
Canoga Park, California 91304
213/700-4013
Brandon P. Krogh
Plant Chemical Engineer
Minnesota Power
30 W. Superior St.
Duluth, Minnesota 55802
218/772-2641, ext. 3311
J. Lee Krumme
President and CEO
Vinings Chemical Co.
2555 Cumberland Pkwy., Suite 2000
Atlanta, Georgia 30339
404/436-1542
Tony Ku
Manager of Technical Services
Cadre Environmental Systems
2845 Clearview Place
Doraville, Georgia 30340
404/458-9527
John Kuehl
Utility Sales Supervisor
Warman International Inc.
2701 S. Stoughton Rd.
Madison, Wisconsin 53716
608/221-2261
Peter M. Kutemeyer
General Manager
Bischoff Environmental Systems
135 Cumberland Rd.
Pittsburgh, Pennsylvania 15237
412/364-8860
Charles R. LaMantia
President
Koch Process Systems, Inc.
20 Walkup Dr.
Westborough, Massachusetts 01581
617/366-9111 ext. 541
Dale Ladd
Regional Manager
Martek Inc.
85 Research Rd.
Hingham, Massachusetts 02043
A-26
-------
G. C. Lammers
Senior Research Engineer
Amoco Chemicals Corp.
Box 400 (C-7)
Naperville, Illinois 60566
312/420-5642
Land
Technology Assessment
George W.
Director,
AMAX Coal
105 S. Meridian
Indianapolis, Indiana 46225
317/266-2812
Wes Langeland
Vice President
Martek, Inc.
85 Research Rd.
Hingham, Massachusetts 02043
617/749-6992
William Mike Lankford
Superintendent, Technical Services,
Cross Generating Station
Santee Cooper
P.O. Box 98
Cross, South Carolina 29436
803/351-4586
Ellen E. Lanum
Conference & Travel Supervisor
Electric Power Research Institute
P.O. Box 10412
Palo Alto, California 94303
415/855-2193
Bernard A. Laseke
Group Supervisor
PEDCo Environmental, Inc.
Chester Towers
11499 Chester Rd.
Cincinnati, Ohio 45246
513/782-4700
Dennis Laslo
Senior Development Engineer
Peabody Process Systems
835 Hope St.
Samford, Conn. 06840
203/327-7000
Peter Lawson
Supervising Engineer, Energy and
Studies Development
Ontario Hydro
700 University Ave.
Toronto, Ontario
Canada M4K1A1
416/592-5393
George 0. Layman
Director, Power Supply
Gulf Power Co.
P.O. Box 1151, 75 N. Pace Blvd.
Pensacola, Florida 32520
904/434-8354
John R. Lee
Field Applications Engineer
BF Goodrich
P.O. Box 1010
Tuscaloosa, Alabama 35403
205/752-1521
George C. Y. Lee
Engineering Specialist
Bechtel Group, Inc.
50 Beale St.
San Francisco, California 94119
415/768-3119
Chiun-Chieh Lee
Technical Manager
Union Chemical Laboratories
1021 Kuang Fu Rd.
Hsinchu, Taiwan
Republic of China
035/713131
Yung-Li Lee
Research Assistant
University of Texas at Austin
E.P. Schoch
Austin, Texas 78712
512/471-4851
T. S. Lee III
Research Manager
LaQue Ctr. for Corrosion Techn.,
Auditorium Circle & Hwy. 76
Wrightsville Beach, NC 28480
919/256-2271
Inc.
A-27
-------
Jan Lefers
Dr.
Nv Kema
Utretseweg 310
Arnhem, The Netherlands
L. Karl Legatski
Manager Process Tech.
FMC Corp.
231 N. Martingale Rd.
Schaumburg, Illinois 60194-2098
312/843-1700
Jurgen Leimkuhler
Division Manager
G. Bischoff
Gaertner Str 44, D-4300
Essen, W. Germany
0201-233031
Charles C. Leivo
Advanced Product Planning Manager
Dresser Industries
2408 Timberlock, Blag. C
The Woodlands, Texas 77380
713/367-7355
Arnold L. Leriche, Jr.
Environmental Engineer
U.S. EPA, Region I
JFK Federal Bldg.
Boston, Massachusetts 02203
617/223-5137
Clifford J. Lewis
Director of Environmental Services
National Lime Assoc.
3601 N. Fairfax Dr.
Arlington, Virginia 22201
303/237-2948
Jack A. Li bey
Director of Power
City of Lakeland
1000 E. Parker St.
Lakeland, Florida
813/687-3636
Production
33802
W. A. Liegois
Process Engineer
Stanley Consultants
Stanley Bldg.
Muscatine, Iowa 52761
319/264-6457
Tom Lillestolen
Manager, FGO Systems
Flakt, Inc.
P.O. Box 87
Knoxville, Tennessee 37901
615/693-7550
Nilo Lindgren
R&D Applications Assessment
Electric Power Research Inst.
P.O. Box 10412
Palo Alto, California 94303
415/855-2753
Darol Lindloff
Regional Sales Manager
Wheelabrator APC Div.
12201 Merit Dr.
Dalls, Texas 75251
214/458-8738
Stephen A. Lingle
Chief Technology Branch
U.S. EPA (WH565)
401 M. Street, S.W.
Washington, D.C. 20460
202/382-7917
Michael R. Lintelman
Sales Engineer
Dravo Lime Co.
3600 Neville Rd.
Pittsburgh, Pennsylvania 15225
412/777-5584
C. David Livengood
Environmental Systems Engineer
Argonne National Laboratory
9700 S. Cass Ave.
Argonne, Illinois 60439
312/972-3737
A-28
-------
George C. W. Logan
Energy Applic. Eng.
Middle South Services,
P.O. Box 61000
New Orleans, LA 70161
569-4725
Inc.
Frank Logeat
National Sales Manager
Bachmann Industries, Inc.
29 Lexington St.
Lewiston, Maine 04240
207/784-2336
W. Z. Looney
Project Planning Engineer
Southern Company Services
P.O. Box 2625
Birmingham, Alabama 35202
205/877-7289
D. Antonio Duran Lopez
Empresa Nacional de Electricidad,
Velazquez 132, Madrid
Spain
91/261-63 00
David P. Lovetere
Resident Construction Manager
Burns & McDonnell
Box 1179
Wheatland, Wyoming 82201
307/322-9530
Phillip S. Lowell
Consulting Chemical Engineer
P.S. Lowell & Co., Inc.
4107 Medical Pkwy., #214
Austin, Texas 78756
512/451-3513
Thomas H. Lucy
Northeast Regional Manager
Research - Cottrell
Box 1500
Somerville, New Jersey 08876
201/685-4479
S.A.
John R. Ludwig
Senior Research Engineer
U.S. Steel Corp.
Research Lab
Coleraine, Minnesota 55722
218/245-2200
Louis A. Luedtke
Central Reg. Manager
Niro Atomizer, Inc.
49U1 College Blvd.
Leawood, KS 66211
913/341-3953
Mario Luperi
Sales Manager, Industrial
Products Division
Wlm. Steinen Mfg. Co.
29 E. Halsey Rd.
Parsippany, New Jersey 07054
201/887-6400
Robert D. Lupi
Product Manager, Special Alloys
Jessop Steel Co.
Jessop Place
Washington, Pennsylvania 15301
Bruce MacDonald
Manager, Sales and Marketing
ABCO Plastics, Inc.
45 Accord Park
Norwell, Massachusetts 02061
617/878-5068
Colin MacDonald
Vice President
Filtres Gaudfrin
4950 Highland Dr.
Salt Lake City, Utah 84117
801/278-2851
Gordon C. MacDonald
Vice President
Mitsubishi International Corp.
520 Madison Ave.
New York, New York 10022
212/605-2006
A-29
-------
Richard Madenburg
Technical Director
htorrison-Knudsen Co., Inc.
P.O. Box 7808
Boise, Idaho 83729
2:08/386-6069
Gary A. Maier
Project Manager
Florida Power and Light Co.
P.O. Box 14000
Juno Beach, Florida 33408
305/663-3608
Jason Makansi
Associate Editor
Power Magazine/McGraw-Hill
122:1 Ave. of the Americas
New York City, New York 10020
212/997-4239
John E. Makar
Regional Manager
FMC Corp.
231 N. Martingale Rd.
Schaumburg, Illinois 6U194-2098
312/843-1700
L. Marnewecke
Process Control Engineer
Armco Autometrics
4946 N. 63rd St.
Boulder, Colorado 80301
303/530-1600
Patrick Maroney
Supervisor
Brown and Calowell
1501 N. Broadway
Walnut Creek, California 94596
415/937-9010
Gregory Martin
Product Engineer
Dresser Industries, Inc.
5323 S. Western Blvd.
Chicago, Illinois 60609
312/471-3040
Peter G. Maurin
Director, FGD Systems
Wheelabrator-Frye, Inc.
600 Grant St.
Pittsburgh, Pennsylvania
412/288-7323
15219
Michael A. Maxwell
Chief Emissions/Affluent Tech. Branch
U.S. EPA
IERL, MD-61
Research Triangle Park, NC 27711
919/541-2578
T. J. May
Supervisor, Short Range Planning
Illinois Power
500 S. 27th St.
Decatur, Illinois 62525
217/424-6706
Manville J. Mayfield
Branch Chief
Tennessee Valley Authority
1010 Chestnut St., Towers II
Chattanooga, Tennessee 37401
William N. McCarthy
U.S. EPA
401 M. St., S.W.
Washington, D.C. 20460
202/382-2625
Thomas A. McClellan
Production Engineer
Public Service of Indiana
P.O. Box 1009
Mt. Carmel, Illinois 62863
812/386-8491 ext. 580
Charles J. McCormick
Vice President
Dravo Lime Co.
3600 Neville Rd.
Pittsburgh, Pennsylvania 15225
412/777-5553
Michael W. McElroy
Project Manager
Electric Power Research Inst.
3412 Hi 11 view Ave.
Palo Alto, California 94303
415/855-2471
A-30
-------
Edward T. McHale
Manager, Combustion & Physical
Sciences Dept.
Atlantic Research Corp.
5390 Cherokee Ave.
Alexandria, Virginia 22312
703/642-4088
Robert Mcllvane
President
Mcllvane Co.
2970 Maria Ave.
Northbrook, Illinois
312/272-0010
60062
Marilyn Mcllvane
Editor
Mcllvane Co.
2970 Maria Ave.
Northbrook, Illinois
312/272-U010
60062
John D. McKenna
President
ETS, Inc.
Suite C-103, 3140
Chaparral Dr., SW
Roanoke, Virginia 24018
703/774-8999
D. M. McLane
District Manager
MPSI
Sox M312
York, Pennsylvania 17402
717/843-8671
Brennan McLaughlin
Supervisor Process Engineering
United Engineers & Const.
23 Inverness Way, E.
Englewood, Colorado 80126
303/79-7310
Joseph McNamara
Manager, Field Services
Thyssen Environmental Systems, Inc.
333 Meadow!and Pkwy.
Secauscus, New Jersey 07094
201/330-2600
Michael L. Meadows
Mechanical Engineer
Black & Veatch
P.O. Box 8405
Kansas City, Missouri
913/967-2189
64114
Wayne E. Meadows
Black & Veatch
1500 Meadow Lake Pkwy.
Kansas City, Missouri 64114
913/967-2643
Michael Melia
Environmental Engineer, Project Manager
Peaco Environmental, Inc.
11499 Chester Rd.
Cincinnati, OH 45246
513/782-4877
Heinz Merlet
Dip! .-Ing.
Lurgi Unwelt un Chemotechnik GmbH
Gervinusstr. 17/19, D-600
Frankfurt, Federal Rep. of Germany
611/157228b
Douglas Merrill
Senior Marketing Manager
Church & Dwight Co., Inc.
20 Kingsbridge Rd.
Piscataway, New Jersey 08854
201/885-1220
Frank B. Meserole
Radian Corp.
P.O. Box 9948
Austin, Texas 78766
512/454-4797
Amelia Mesko
Engineer, Generating Engineering
Potomac Electric Power
1900 Pennsylvania Ave.
i.i-v.~Li^»i^i4-x-,k-. n r or if tiv 1
Washington, D.C.
202/872-3575
20068
N.W.
Ronald G. Metz
Manager, Market Development
Mixing Equipment Co.
135 Mt. Read Blvd.
Rochester, New York 14603
716/436-5550
A-31
-------
S. D. Meyer
Senior Engineer
Combustion Engineering
1000 Prospect Hill Rd.
Windsor, Connecticut 06095
2:03/668-1911
Chris F. Meyer
Waste Management Systems Division
Koch Process Systems
20 Walkup Dr.
Westboro, Massachusetts 01581
617/3b6-9111
Sheldon Meyers
Director, Office of Air Quality
and Standards
U.S. EPA
401 M St., S.W.
Washington, D.C. 20460
202/655-4000
John H. Michael
Chief Executive Officer
Princeton Chemical
1810 24th Street N.W.
Washington, D.C. 20008
202/483-0063
Stephen Michel
Project Engineer
PEPCO
1900 Pennsylvania Ave., N.W.
Washington, D.C. 20068
202/872-2436
Samy R. Mikhail
Generating Engineer
PEPCO
1900 Pennsylvania Ave., N.W.
Washington, U.C. 20068
202/872-3487
Robert Miller
Field Sales Representative
Cabot Wrought Products
1020 West Park Ave.
Kokomo, Indiana 46901
317/456-6000
John 0. Milliken
Project Manager
U.S. EPA
IERL, MD-61
Research Triangle Park, NC 27711
919/541-7716
Robert Mingea
Vice President
Lodge-Cottrell
601 Jefferson
Houston, Texas
713/750-2089
- Sales
Operations
77002
John Minnick
Consultant
Box 271
Plymouth Meeting,
Pennsylvania 19462
J. E. Mirabel la
Chiyoda International Engineering
1300 Park Place
Seattle, Washington 98101
206/624-9350
P. R. Misra
Utility Sales Manager
Taulman Sales Co.
3312 Piedmont Rd., N.E.
Atlanta, Georgia 30305
404/261-2535
J. David Mobley
Environmental Engineer
U.S. EPA
IERL, MD-61
Research Triangle Park, NC 27711
919/541-2350
Karsten Moller-Jansen
Niro Atomizer, Inc.
9165 Rumsey R.
Columbia, Maryland 21045
301/997-8700
Virginia M. Moon
Development & Testing Engineer
Combustion Engineering Inc.
31 Inverness Center Pkwy.
Birmingham, Alabama 35209
205/967-9100
A-32
-------
Max A. Moore
Vice President, Equipment Division
KVB, Inc.
Irvine, California 92714
714/250-6310
Keith Moore
Marketing
Rockwell Int.
6900 Desoto
Canoga Park, California 91304
213/700-4016
Thomas Morasky
Project Manager
Electric Power Research Inst.
P.O. Box 10412
Palo Alto, California 94303
415/655-2468
Terri E. Morel and
Manager, Coal Development Programs
Illinois Dept. of Energy &
Natural Resources
325 ]/t. Aaams, Rm. 300
Springfield, Illinois 62706
217/765-2600
David S. Morey
Supervisor, Market Research
Allied Corp.
1411 Broadway
New York, New York 10018
212/391-5149
Robert M. Morford
Vice President, Marekting
Joy Manufacturing Co.
4565 Colorado Blvd.
Los Angeles, California 90039
213/240-2300
Wayne E. Morgan
Manager or Pollution Control
Black & Veatch
P.O. Box 8405
Kansas City, Missouri 64114
913/967-2198
James C. Morgan
Engineer, Mechanical Design
Union Electric Co.
P.O. box 149
St. Louis, Missouri 63166
314/554-2771
Jack Morgenstern
Equipment Engineer
Stone & Webster Engineering
1 Penn Plaza, 250 W. 34th St.
New York, New York 10119
212/290-6668
Michael D. Morris
Generation Equipment Sales Eng.
General Electric Co.
10550 Barkely
Overland Park, Kansas 66212
913/967-6272
Per E. Morsing
Niro Atomizer A/S
9165 Rumsey Rd.
Columbia, Maryland 20145
301/997-8700
Robert E. Moser
Senior Engineer
Bechtel Power Corp.
50 Beale St.
San Francisco, California 94119
415/768-9055
G. A. Mountford
Director of Licensing, Chemicals
Pfizer, Inc.
235 E. 42nd St.
New York, New York 10017
212/573-7405
Hugh Mullen
Director, Government Relations
Conversion Systems, Inc.
115 Gibraltar Rd.
Horsham, Pennsylvania 19044
215/441-5900
A-33
-------
Theo Mullen
Reporter
McGraw Hill
1221 Avenue of the Americas
New York, New York 10020
212/899-9416
Don Muller
Sales Manager
Black River Lime Co.
P.O. Box 1
butler, Kentucky 41006
606/472-7761
Fred Mulloy
Environmental Manager
Phillips Petroleum Co.
7 D 4 Phillips Bldg.
Bartlesville, Oklahoma 74004
916/661-5735
Edward J. Muren
District Manager
Research-Cottrell
85 W. Algonquin Rd.
Arlington Heights,
312/228-6228
Illinois 60005
Andrew J. Murphy
Office Manager
Acurex
3200 Nelson Chapel Hill Hwy.
Research Triangle Park, N.C. 27709
919/549-8915
John Murray
Dow Chemical
P.O. Box 150
Plaquemine, Louisiana 70764
504/389-8000
E. G. Murray
Supervisor, Plant Equipment
ana Systems
Southern Company Services
P.O. Box 2625
Birmingham, Alabama 35202
205/870-6855
John Mycock
Vice President
ETS, Inc.
Suite C-103
3140 Chaparral Dr., Ski
Roanoke, Virginia 24018
703/774-8999
Amir Nassirzadeh
Design Mechanical Engineer
L.A. Dept. of Water & Power
111 N. Hope St., Rm. 661
Los Angeles, California 90012
213/481-4647
Lewis G. Neal
President
Noxso Corp.
2625 H. C. Mathis
Paducah, Kentucky 42001
502/444-6474
Derek Neely
Pittsburgh Des Moines
Neville Island
Pittsurgh, Pennsylvania 05225
412/331-3000
James A. Nelson
Marketing Manager
Effox, Inc.
10921 Reed Harman Hwy.
Cincinnati, Ohio 45242
513/793-1932
Hans Neukam
Sales Manager
Kraftanlagen Heidelberg
Im Breitspiel 7
D6900 Heidelberg,
West Germany
06221/394378
Thomas H. Newhams
Peabody Process Systems
835 Hope St.
Stamford, Connecticut 06907
203/327-7000
A-34
-------
Joseph T. Newman
Engineering Specialist
Bechtel Group, Inc.
bO Beale St.
San Francisco, California
415/768-3114
T. W. Newton
Senior Vice-President
MPSI
Box M312
York, Pennsylvania 17402
717/843-8671
Yen Nguyen
Engineer
Ontario Hydro
800 Kipling Ave.
Toronto, Ontario,
416/231-4111
94105
Canada M8Z 5S4
Arthur F. Nicholson
President
Kentucky Export Resources Authority
Suite 1505, Vine Center
Lexington, Kentucky 40507
606/233-3545
Irvin P- Nielsen
Chairman and C.E.O.
Nathona Resources, Inc.
1600 Broadway, Suite 1120
Denver, Colorado 80202
303/839-1600
Kurt R. Nielsen
Mineral Economist
National Resources Inc.
1600 Broadway, Suite 1120
Denver, Colorado 80202
0303/839-1600
Gary Niles
Proauct Manager
Pathway Bellows, Inc.
P.O. Box 1526
El Cajon, California 92022
619/440-1300
Kimio Nishio
Technology and Engineering Division
Chiyoda Intl. Corp.
1300 Park Place Blag.
Seattle, Washington 98101
206/624-9350
jack Noble
Chief Mechanical
C. T. Main
101 Huntington
Boston, Massachusetts 02199
617/262-3200
Heinrich Novak
Chairman
Grosskraftwerk
Nurnberg,
West Germany
Franken AG
David M. Novick
Product Manager, Utility Env.
Combustion Engineering Inc.
1000 Prospect Hill Ra.
Winsor, Connecticut 06070
205/285-9243
John R. Null
Manager, Environmental Systems
Babcock & Wilcox
P.O. Box 61038
New Orleans, Louisiana 70161
504/587-5636
Jack F. O'Donnell
Chairman & Chief Executive Officer
Advanced Energy Dynamics Inc.
14 Tech Circle
Natick, Massachusetts 01781
617/653-8112
John O'Donnell, Jr.
Director of Market Development
High Performance Tube, Inc.
1460 Morris Ave.
Union, New Jersey 07083
201/964-8520
A-35
-------
Donald O'Hair
Senior Sales Engineer
Joy Manufacturing Co.
1301 k. 22nd Street
Oak Brook, Illinois 60521
312/654-4090
Robert D. O'Hara
Environmental Engineer
Duquesne Light Co.
One Oxford Centre, 27-2
301 Grant St.
Pittsburgh, Pennsylvania 15279
412/393-6098
Andrew T. O'Neill
Peabody Process Systems
835 Hope St.
Stamford, Connecticut 06907
203/327-7000
Gary Ochs
Program Manager
York Research Consultants
938 Quail St.
Denver, Colorado 80215
303/233-1513
Leif Olausson
Swedish State Power Board
S-16287 Vaellingby,
Sweden
David G. Olson
Manager, Utility Marketing
General Electric
200 N. 7th St.
Lebanon, Pennsylvania 17042
717/274-7355
H. Onuma
Manager
Mitsubishi International Corp.
520 Madison Ave.
New York, New York 10022
212/605-2663
Sidney Orem
Executive Director
Industrial Gas Cleaning Inst.
700 North Fairfax St., #304
Alexandria, Virginia 22314
703/836-0480
C. L. Osborne
Santee Cooper
P.O. Box 98
Cross, South Carolina 29436
803/351-4586
Ronald G. Ostendorf
Senior Engineer
Proctor & Gamble Co.
7162 Reading Rd.
Cincinnati, Ohio 45222
513/763-4457
Norman Ostroff
Supervisor, Process Engineering
Peabody Process Systems
835 Hope St.
Stamford, Connecticut 06097
203/327-7000
Delbert M. Ottmers
Assistant Vice President
Radian Corp.
8501 Mopac Blvd.
Austin, Texas 78759
512/454-4797
Buck Oven
Administrator Power Plant Siting
Florida Dept. Env. Regulations
2600 Blair Stone Rd.
Tallahassee, Florida 32301
904/488-0130
Peter Overaick
Establissements Leon Lhoist S.A.
21 Avenue Rogier
Liege,
Belgium B-4000
817/732-8164
David R. Owens
University of Texas at Austin
University of Texas
Department of Chemical Engineering
Austin, Texas 78741
512/471-4851
A-36
-------
Michael A. Ozol
Consultant
Michael A. Ozol Ph.D.
24u3 Ken Oak Ra.
Baltimore, Maryland 21209
301/664-2565
Robert Steven Pace
Bio Environmental Engineer
City of Jacksonville
515 W. 6th St.
Jacksonville, Florida 32206
904/633-3303
Michael A. Palazzolo
Radian Corp.
3024 Pickett Rd.
Durham, North Carolina 277U5
919/493-4574
Knut Papajewski
Manager, Operations
Thyssen Environmental Systems, Inc.
333 Meadow!and Pkwy.
Secaucus, New Jersey 07094
201/330-2600
Rashmi Parekh
Resident Sales Manager, Eastern Region
Dorr Oliver, Inc.
274 Riverside Ave.
Westport, Connecticut 06880
203/358-3800
Jeffrey H. Parker
Marine Science Research Cntr
SUMY
Stony Brook, New York 11794
516/246-5000
Richard W. Patton
President
VFL Technology Corp.
42 Lloyd Avenue
Malvern, Pennsylvania 19355
215/296-2233
R. L. Pearce
Associate Scientist
Dow Chemical Co.
B1605 Bldg.
Freeport, Texas 77541
409/239-1419
Michael Perlsweig
Program Manager
U.S. Department of Energy
FE-23
Washington, D.C. 20545
301/353-4399
Dennis Perrone
Regional Manager
Martek Inc.
1630 Newell Avenue
Walnut Creek, California 94596
415/937-5630
Mogens Petersen
President
Niro Atomizer, Inc.
9165 Rumsey Rd.
Columbia, Maryland 20145
301/997-8700
Vincent Petti
Manager, Application Engr. FkD Sys.
Wheelabrator-Frye Inc.
600 Grant Street
Pittsburgh, Pennsylvania 15219
412/288-7465
Brian R. Phelan
Regional Manager
Joy Manufacturing Co.
4901 College Blvd.
Shawnee Mission, Kansas 66211
913/648-8783
Peter H. Phillips
Manager, I&C
Whellarrator Air Pollution Control
101 Merrit-7
Norwalk, Connecticut 06856
203/852-6846
Allan R. Pike
Consulting Engineer
Self Employed
76 Seir Hill Road
Wilton, Connecticut 06897
203/762-8990
A-37
-------
James E. Pilgrim, Jr.
Section Supervisor
Tennessee Valley Authority
100 IBM Building
Chattanooga, Tennessee 37401
615/751-4455
George Pinheiro
Sales
Research-Cottrell, Inc.
P.O. box 1500
Somerville, New Jersey 08876
201/685-4109
T. Duane Pinson
Southern Region Sales Manager
Marcona Ocean Industries, Ltd.
1001 No. W. 62th Street, Suite 200
Fort Lauderdale, Florida 33309
305/776-4000
Bill Piske
Manager S/W Office
Engineering Science
13740 Midway Drive, Suite 706
Piano, Texas 75075
214/392-0695
Wallace S. Pitts III
Associate
Kilkelly Environmental Assoc.
P.O. Box 31265
Raleigh, North Carolina 27622
919/761-3150
James F. Plappert
Director of Sales
Conversion Systems, Inc.
115 Gibraltar Rd.
horsham, PA 19044
215/441-5920
Fred L. Porter
Section Chief
U.S. EPA, OAQPS, ESED, SDB
MD-13
Research Triangle Park, N.C. 27711
919/541-5624
Tom Potter
Director of Administration
Nat. Lab. Assoc.
3601 N. Fairfax Dr.
Arlington, Virginia 22201
703/243-5463
Geroge Powers
Engi neer-Generati ng Engi neering
Potomac Electric Power Co.
1900 Pennsylvania Ave. - Room 832
Washington, D. C. 2u068
202/331-6250
Paul R. Predick
Acting Head, Mechanical Analytical
Sargent & Lundy Engineers
55 E. Monroe St.
Chicago, Illinois 60603
312/269-6671
Jerry V. Presley
Senior Engineer
Virginia Electric & Power Co.
P.O. Box 564
Richmond, Virginia 23204
804/771-6148
Jack Preston
Senior Engineer
South Carolina Electric & Gas
P.O. Box 764
Columbia, South Carolina 29218
803/478-3849
Tom Priest
Vice President
Ceil cote
140 Sheldon Rd.
Berea, Ohio 44017
216/247-0770
Frank T. Princiotta
Director, Laboratory
U.S. EPA
IERL, MD-60
Research Triangle Park, NC 27711
919/541-2821
Div.
A-38
-------
Edwin J. Puska
Production Systems
Marquette Board of Pwr. & Light
2200 bright St.
toarquette, Michigan 49855
906/226-6900
Irwin A. Raben
President
IAR Technology, Inc.
130 Sandaringham South
Moraga, California 94556
415/376-3951
Alan D. Randolph
Professor
University of Arizona
Tuscon, Arizona 85721
602/621-6051
Richard Rao
Manager, Air Quality Control
Ebasco Services, Inc.
160 Chubb Ave.
Lyndhurst, New Jersey 07071
201/460-1900
Fred Rapp
Sales Representative
Armco Inc.
7000 Roberts Kansas City
Kansas City, Missouri 64125
816/242-5452
Hulic Ratterree
Manager, Technical Services
Blount Energy Resource Corp.
4520 Executive Park Drive
Montgomery, Alabama 36116
205/277-8860
John Reilly
President
Thyssen Environmental Systems, Inc.
333 Meaaowland Pkwy.
Secaucus, New Jersey 07094
201/330-2600
Jack B. Reisdorf
Project Engineer
Stearns-Roger Engineering Corp.
P.O. Box 5888
Denver, Colorado 80217
303/692-3420
Bruce W. Remick
Director of Aglime Marketing
National Crushed Stone Assoc.
1415 Elliot Place, N.W.
Washington, D.C. 20007
202/342-1100
George Rey
Senior Staff Engineer
U.S. EPA
Washington, D.C. 20460
202/382-2626
Richard G. Rhudy
Project Manager
Electric Power Research Inst.
P.O. Box 10412
Palo Alto, California 94303
415/855-2421
David Z. Richards
Kennecty Van Saun Corp.
R. R. St.
Danville, Pennsylvania 17821
Phillip Richardson
Manager, Market Research & Development
Du Pont
1007 Market St.
Wilmington, Delaware 19898
802/774-4930
Ed Riordan
Supv. Waste Treatment
City Water, Light and Power
7th & Monroe
Springfield, Illinois 62757
217/769-2238
Charles J. Rizzo
Vice President/Engineering Service
Air Clean Damper Co.
Kugler Mill & Blue Ash
Cincinnatti, Ohio 45236
513/793-1253
A-39
-------
Russell F. Robards
Chemical Engineer
Tenn. Valley Authority
1160 Chestnut St., Tower II
Chattanooga, Tennessee 37401
615/751-5662
Gary T. Roche!le
Assistant Professor
Dept. of Chemical Engineering
University of Texas
Austin, Texas 78712
512/471-3434
Ron Rohlik
Mechanical Engineering Specialist
Getty Oil Co.
P.O. Box 197X Rt. 1
Bakersfield, California 93308
805/399-2961
Greg A. Rollins
Sales Representative
Timet
726 Avenue R
Grand Prairie, Texas 75050
214/641-4410
Robert N. Roop
Product Manager, Spray Dryers/Fabric
Filters
Research-Cottrell, Inc.
P.O. Box 1500
Somerille, New Jersey 08876
201/685-4451
Kevin C. Rorke
Manager, Projects & Construction
FMC Corp.
1501 Woodfield Rd., Suite 300 E.
Schaumburg, Illinois 60195
312/843-1700
Ruiz-Alsox Rosa
Research Assistant
University of Texas at Austin
E.P- Schoch Lab
Austin, Texas 78712
412/471-4851
Edward C. Rosar
President
Industrial Resources, Inc.
300 Union Blvd., Suite 520
Lakewood, Colorado 80228
303/986-4507
Jean Thomas Rose
Project Supervisor
Betz Laboratories, Inc.
Somerton Rd.
Trevose, Pennsylvania 19047
215/355-3300, ext. 394
Norman A. Rosekrans
Regional Sales Manager
uoy Manufacturing Co.
5775 Peachtree DunwoocHy
Atlanta, Georgia 30319
404/256-2934
Rd., Suite 200E
Harvey S. Rosenberg
Senior Research Engineer
BatteHe-Columbus Lab.
505 King Ave.
Columbus, Ohio 43201
614/424-5010
Clifford A. Rosene
FGD Production Supervisor
No. Indiana Public Service Co.
Route 1 Box 320
Wheatfield, Indiana 46392
219/956-5180
Bud Ross
Industry Manager
Huntington Alloys, Inc.
P.O. Box 1958
Huntington, West Virginia 25720
304/696-3509
Edward S. Rubin
Professor
Carnegie-Mellon University
Schenley Park
Pittsburgh, Pennsylvania 15213
412/578-2491
A-40
-------
Richard A. Runyan
Projects Manager
Tennessee Valley Authority
1160 Chestnut St., Towers II
Chattanooga, Tennessee 37401
615/751-5663
Randall E. Rush
Manager, Flue Gas Treatment
Southern Co. Services
F. Box 2625
Birmingham, Alabama 35202
205/870-6320
James R. Rutledge
Project Coordinator/Operations
Jacksonville Electric Authority
233 U. Duval St.
Jacksonville, Florida 32211
904/633-2220
Stephen B. Ryan
Sales Representative
Marblehead Lime Co.
300 W. Washington
Chicago, Illinois 60606
412/563-1812
George P. Sacco
Environmental Bus. Rep. Dev.
General Electric Env. Services
22 Perimeter Center Ene
Dunwoody, Georgia 30346
404/399-5517
A. Saleem
Vice Presiaent, Int'l. Bus. Dev.
General Electric Environmental
200 N. Seventh St.
Lebanon, Pennsylvania 17042
717/274-7171
Norman C. Samish
Staff Research Engineer
Shell Development Co.
P.O. Box 1380
Houston, Texas 77001
713/493-7944
Eric A. Samuel
Senior Development Engineer
General Electric Co.
200 North Seventh St.
Lebanon, Pennsylvania 17042
717/274-7048
C. J. Santhanam
Manager, Chemical Eng.
A. D. Little, Inc.
20 Acorn Park
Cambridge, Massachusetts 02140
617/864-5770
James S. Sarapata
Director of Process Development
Church & Dwight Co., Inc.
20 Kingsbridge Rd.
Piscataway, New Jersey 08854
201/885-1220
Fred J. Sauereisen
Vice President
Sauereisen Cements Co.
160 Gamma Drive
Pittsburgh, Pennsylvania 15238
412/963-0303
A. E. Saunders
Superintendent, Technical Services
Winyah Generating Station
Santee Cooper
P.O. Box 1275
Georgetown, South Carolina 29440
803/546-4171
Mark Schneider
Project Engineer
Delmarva Power
600 King Street
Wilmington, Delaware 19803
302/429-3616
Donald Schreyer
Peabody Process Systems
835 Hope St.
Stamford, Connecticut 06907
203/327-7000
A-41
-------
Richard Schubert
Sales Manager
General Electric
200 N. 7th St.
Lebanon, Pennsylvania
717/274-7257
17042
David Schulz
Regional Power Industry Expert
U.S. EPA
230 S. Dearborn St.
Chicago, Illinois 60604
312/353-2088
R. W. Schutz
Supervisor, Corrosion Research
and Development
TIMET, Division of Titanium
Metals Corp.
P.O. Box 2128
Henderson, Nevada 89015
702/564-2544 - ext. 215
Richard A. Schwartz
President
D. R. Technology, Inc.
Hidden Pines Dr.
Clarksburg, New Jersey 08510
201/780-4664
Joseph A. Schwartz
Marketing Manager
KVB, Inc.
18006 Skypark
Irvine, California 92714
714/250-6258
Tonny C. Schytte
Niro Atomizer, Inc.
9165 Rumsey Rd.
Columbia, Maryland 20145
301/997-8700
William C. Seale
Mechanical Supervisor
Lower Colorado River Authority
P.O. Box 220
Austin, Texas 78767
512/473-3540
David 0. Seaward
Product Manager, FGD Systems
Wheelarrator Air Pollution Control
101 Merritt 7, P.O. Box 5440
Morwalk, Connecticut 06856
203/852-6884
Roland K. Seward
Process Engineer
Kennedy Van Saun Corporation
R. R. St.
Danville, Pennsylvania 17821
717/275-3050
D. Shack!ey
Mineral Resources Director
British Gypsum Ltd.
Gotham, Nottinghamshire
England
0602/830431
Raymond J. Shaffery
Director of Commercial Development
Church & Dwight Co., Inc.
20 Kingsbridge Rd.
Piscataway, New Jersey 08854
201/885-1220
Navin D. Shah
Consultant
142 Sundance Ct.
Grand Junction, Colorado 81503
303/243-1503
Arvind M. Shah
Vice President
Spraco, Inc.
P.O. Box 3800
Nashua, New Hampshire 03060
603/888-1050
Don Shattuck
Process Engineer
Stearns-Roger Engineering
P.O. Box 5888
Denver, Colorado 80217
303/692-4139
A-42
-------
Charles E. Shelton
Manager, Fossil & Hydro
Operations Support
Viriginia Electric and Power Co.
P.O. Box 26666
Richmond, Virginia 23261
804/771-4441
Robert Shiely
Research Assistant
University of Texas at Austin
Austin, Texas 78712
512/471-4851
Christopher Shih
Senior Project Engineer
TRW
One Space Park 01/2040
Redondo Beach, California 90278
213/536-4105
G. H. Shroff
Engineering Specialist
Bechtel Power Corp.
15740 Shady Grove Rd.
Gaithersburg, Maryland 20877
301/258-3146
William E. Siegfriedt
Senior Mechanical Engineer
Fluor Engineers, Inc.
200 W. Monroe St.
Chicago, Illinois 60606
312/368-3828
Mark S. Siegler
Chief, Technical Support Branch
U.S. EPA
401 M. Street S.W.
Washington, D.C. 20460
202/382-2835
William L. Silence
Associate Engineer
Cabot Corp.
1020 W. Park Ave.
Kokomo, Indiana 46901
317/456-6201
P. L. Simiskey
Research Associate
Dow Chemical Co.
B1605 Bldg.
Freeport, Texas 77541
409/239-1419
A. P. Sandy Simko
Projects Engineering Manager
Arizona Public Service
P.O. Box 21666 (Station 5760)
Phoenix, Arizona 85036
602/271-7261
Michael Skinner
Engineer
Northern States Power
3100 Marshall St.
Minneapolis, Minnesota
612/33U-5991
Co.
55418
John H. Skinner
Director, Office of Solid Waste
U.S. EPA
401 M St., S.W.
Washington, D.C. 20460
202/655-4000
A. V. Slack
President
SAS Corp.
Route 1, Box 69
Sheffield, Alabama 35660
205/383-1627
A. G. Sliger
Assistant Manager Process Design
M. W. Kellogg Co.
3 Greenway Plaza
Houston, Texas 77046
John E. Smigelski
Senior Engineer
New York State Electric & Gas Co.
4500 Vestal Pkwy.
East Binghamton, New York 13903
608/279-2551, ext. 4605
A-43
-------
Earl 0. Smith
Project Manager
Black & Veatch
IbUU Meadow Lake Pkwy,
Kansas City, Missouri
913/967-2643
64114
Ted Smith
Director of Design
Burns & McDonnell Engineering Co,
P.O. Box 173
Kansas City, Missouri 64141
816/333-4375
Jeffrey D. Smith
Plant Reliability Engineer A
Ohio Edison Co.
76 S. Main Street
Akron, Ohio 44308
Peter V. Smith
General Sales Manager
Research-Cottrell, Inc.
P.O. Box 1500
Somerville, Mew Jersey 08876
201/685-4221
Robert H. Smith
President
Rob Smith Assoc., Inc.
5355 Knox St.
Philadelphia, PA 19144
215/843-2209
Roger Smith
Plant Results Engineer
South Mississippi Electric
Power Assoc.
P.O. Box 1589
Hattiesburg, Mississippi 39401
601/268-2083
Norman B. Smith
Chief Chemical Engineer
Stanely Consultants, Inc.
Stanley Bldg.
Muscatine, Iowa 52761
319/264-6299
Robert Smock
Managing Editor
Electric Light & Power Magazine
P.O. box 1030
Harrington, Illinois 60010
312/381-1840
John V. Smolensk!
FGD Specialist
Stone & Webster Engineering
3 Executive Campus
Cherry Hill, New Jersey 08034
609/482-3594
Jeff Snapp
Project Engineer
Public Service Co. of Indiana
1000 E. Main St.
Plainfield, Indiana 46168
317/838-1656
Donald R. Snider
Cincinnati Gas & Electric Co.
P.O. Box 107
Union, Kentucky 41091
606/586-5600
Robert E. Sommerland
Vice President
Foster Wheeler Develop. Corp.
12 Peach Tree Hill Rd.
Livingston, New Jersey 07039
201/533-3650
James P- Spellman
Senior Business Rep.
General Electric Co.
8101 Stemmons Freeway
Dallas, Texas 75248
214/688-6165
H. Michael Spence
President
National Resources, Inc.
1600 Broadway, Suite 1120
Denver, Colorado 80202
303/839-1600
A-44
-------
Herbert W. Spenler, III
Manager, Advanced Technology
Joy Manufacturing Co.
4b6b Colorado Blvd.
Los Angeles, California 90039
213/240-2300 ext. 426
Lawrence J. Stanislow
Assistant Power Industry Manager
Pennwalt Corp.
3 Parkway
Philadelphia, Pennsylvania 19102
215/587-7264
Michael J. Stapf
President
Mosser Damper Co.
5000 Tilghman St.
Allentown, Pennsylvania
215/395-4900
Glenn G. Stauffer
Project Engineer
Pennsylvania Power and Light Co.
Two North Ninth St.
Allentown, Pennsylvania 18101
215/770-6563
Harold Steeves
Vice President/General Manager
ABCO Plastics, Inc.
45 Accord Park
Norwell, Massachusetts 02061
617/878-5068
Edward W. Stenby
Assistant Manager, Engineering
Stearns-Roger Engineering Corp.
P.O. Box 5888
Denver, Colorado
303/758-1122
John G. Stensland
Marketing Manager
FMC Corp.
231 N, Martingale Rd.
Schaumburg, Illinois 60194-2098
312/843-1700
Frederick Stern
Plant Engineer
Basin Electric Power Coop.
P.O. Box 1059
Beulah, North Dakota 58523
701/873-4545
Richard D. Stern
Chief, LIMB Applications
U.S. EPA
IERL-RTP (MD-63)
Research Triangle Park, N.C.
919/541-2547
27711
Nicholas J. Stevens
Chemical Process Development Mgr.
Research-Cottrell
P.O. Box 1500
Sommerville, New Jersey 08876
201/685-4887
Jack F. Stewart
Product Specialist
Bablcock & Wilcox
20 S. Van Buren
Barberton, Ohio 44203
216/860-2118
Dorothy A. Stewart
Project Manager
Electric Power Research Inst.
P.O. Box 10412
Palo Alto, California 94303
415/855-2609
Joseph G. Stites
Vice President
Energy Sciences & Services, Inc.
203 Barley Mill Rd.
Old Hickory, Tennessee 37138
615/847-5031
Jim Stone
Resource Specialist
Louisiana ONR
P.O. Box 44066
Baton Rouge, Louisiana 70804
504/342-1206
A-45
-------
Mark R. Stouffer
Associate Engineer
Conoco Coal Research
4UUU Brownsville Rd.
Library, Pennsylvania
412/854-6639
15234
Don Stowe
Director/Tech Sales & Svc.
Dravo Lime Co.
36UO Neville Rd.
Pittsburgh, Pennsylvania 15225
412/777-5574
Joseph P. Strakey
Director, Process Technology Division
DOE/Pittsburgh Energy Technology Ctr.
P.O. Box 10940
Pittsburgh, Pennsylvania 15236
41Z/675-6125
John Strange
Manager of Mechanical Engineering
bibbs & Hill, Inc.
11 Penn Plaza
New York, New York 100U1
212/760-4162
Harry A. Straw
Product Development Consultant
E. I. Du Pont de Nemours and Co.
1007 Market St.
Wilmington, Delaware 19898
302/744-3673
Joseph J. Stuparich
Senior Business Representative
General Electric Co.
3 Penn Center Plaza
Philadelphia, Pennsylvania 19102
215/241-5240
Ben Y. Su
Environmental Engineer
United Engineers and Constructors
100 Summer St.
Boston, Massachusetts 02110
617/338-6000
Roger P. Summerhays
Product Marketing Engineer
EIMCO
P.O. Box 300
Salt Lake City, Utah 84110
801/526-2367
Karen V. Summers
Senior Hydrogeologist
Tetra Tech, Inc.
3746 Mt. Diablo, Suite 300
Lafayette, California 94549
415/283-3771
Carl E. Swanson
Sr. Metallurgical Engineer
Newmont Exploration Ltd.
44 Briar Ridge Rd.
Danbury. Connecticut 06810
2U3/743-6784
Donald E. Syler, Jr.
Supervisory Engineer
Consumers Power Co.
1945 West Parnall Rd.
Jackson, Michigan 49201
517/788-1946
Barry C. Syrett
Project Manager
Electric Power Research Inst.
P.O. Box 10412
Palo Alto, California 94303
415/855-2956
Atsushi Tatani
Process Engineer
Mitsubishi
520 Madison Ave.
New York, New York
212/605-2006
Pamela D. Taulbee
Editor
Coal Technology Report
1401 Wilson Blvd., Suite 910
Arlington, Virginia 22209
703/528-1244
A-46
-------
Charles E. Taylor
University of Texas at
Department of Chemical
Austin, Texas 78712
512/471-4851
Austin
Engineering
Donald P. Teixeira
Senior Engineer
Pacific Gas & Electric
34uO Crow Canyon Rd.
San Ramon, California 94583
Jeffrey A. Telander
Environmental Engineer
U.S. EPA, OAQPS/ESED/TSB/SDS MD-13
Research Triangle Park, N.C. 27711
919/541-5595
Jack T. Thompson
Chief, Technical Services Branch
Tennessee Valley Authority
705 Edney Bldg.
Chatanooga, Tennessee 37401
615/751-2774
Stepen W. Tippet
Product Manager
CHEMFAB
Water St.
N. Bennington, Vermont 05257
802/447-1131
Ed Tomeo
Generation Engineer
Northeast Utilities
P.O. Box 270
Hartford, Connecticut 06104
203/247-0838
Tasia P- Toombs
Conference & Travel Coordinator
Electric Power Research Institute
P.O. Box 10412
Palo Alto, California 94303
415/855-8973
Robert L. Torstrick
Section Supervisor
Tennessee Valley Authority
501 Chemical Engineering Bldg.
Muscle Shoals, Alabama 35660
205/386-2514
Joseph M. Towarnicky
Research Chemist
United McGill Corp.
One Mission Park
Groveport, Ohio 43125
614/836-9981
Edward C. Trexler
Program Manager
U.S. Department of Energy
Washington, D.C. 20545
301/353-2683
Ronald J. Triscori
Director of Products
Joy Manufacturing Co.
4565 Colorado Blvd.
Los Angeles, California 90039
213/240-2300 ext. 527
Philip C. Tseng
Research Analyst
University of Texas at Austin
E.P. Schoch Lab.
Austin, Texas 78712
J. V. Twork
Engineer, Research Dept.
Bethlehem Steel Corp.
Homer Research Lab
Bldg. A Rm D 332
Bethelehm, Pennsylvania 18016
215/694-6532
Kiyoshi Urakami
Commercial Manager
Mitsubishi
520 Madison Ave.
New York, New York
212/605-2006
Robert P. Van Ness
Louisville Gas & Electric Co.
P.O. Box 32010
Louisville, Kentucky 40232
502/566-4011
Ernest Vanhoose
Program Manager
Kentucky Energy Cabinet
P.O. Box 1188, Iron Works Pike
Lexington, Kentucky 40578
606/252-5535
A-47
-------
Joel Yatsky
Manager, Combustion Technology
Foster Wheeler Energy Corp.
9 Peach Tree Hill Rd.
Livingston, New Jersey 07039
2U1/533-2105
Marl in J. Yeesaert
President
beneral Aggregate Corp.
401 North Lindbergh Blvd.
St. Louis, Missouri 63141
314/997-7777
Kurt Veser
Director
Kraftanlagen Heidelberg
Im Breitspiel 7
D 6900 Heidelberg,
West Germany
06221/394562
Suzanne M. Viale
Marketing Services Coordinator
FMC Corp.
231 N. Martingale Rd.
Schaumburg, Illinois 60194-2098
312/843-1700
Mohr Volker
Process Engineer
Letepro Corp.
Avenue of the Americas
New York, New York
Stephen C. Voss
Noxso Corp.
2625 H. C. Mathis Dr.
Paducah, Kentucky 42001
David A. Wagner
Sales Representative
Marblehead Lime Co.
300 W. Washington
Chicago, Illinois 60606
312/263-4490
J. Peter Wahlman
Senior Project Manager
Cliffs Engineering, Inc.
P.O. Box 1211
Rifle, Colorado 81650
3U3/625-2445
John R. Walenten
Process Development Engineer
Badische Corp.
50 Central Ave.
Kearny, New Jersey 07032
205/589-1600
Shih-Chung Wang
Engineering Specialist
Bechtel Group, Inc.
50 Beale St.
San Francisco, California 94105
415/768-2873
Dale R. Warner
Regional Manager
Air Clean Damper Co.
645 Ridgemont Dr.
Roswell, Georgia 30076
404/587-5293
Earl J. Weber
Mechanical Engineer
Cajun Electric Power Coop.
P.O. Box 15540
Baton Rouge, Louisiana 70816
504/291-3060
Henry C. Weber
President
HAW Management Science Consult.
415 E. 52 St.
New York, New York 10022
212/355-1448
Greg F. Weber
Research Supervisor
Univ. of No. Dakota Energy
Research Ctr.
Box 8213, University Station
Grand Forks, North Dakota 58502
701/795-8222
William C. Webster
American Resources Corp.
P.O. Box 813
Valley Forge Pennsylvania 19482
215/337-7373
A-48
-------
Christopher P. Wedig
Supervisor, Process Engineering
Group Div.
Stone & Webster
P.O. Box 2325
Boston, Massachusetts 02107
617/469-1900
John Weeda
Engineering Superintendent
Cooperative Power Assoc.
Coal Creek Station
P.O. Box 780
Unaerwood, North Dakota 58576-0780
701/442-3211
Maurice W. Wei
Envir. Eng. Manager Proc. Gas
Aluminum Co. of America
1501 Alcoa Bldg.
Pittsburgh, Pennsylvania 15219
412/553-2085
Ing Helmut Weiler
Director
Grobkraftwerk Franken AG
Rudolphstr. 28
Nuerenberg, GERMANY (W) D-8500
911-5397203
Carl Weilert
Air Quality Consultant
Burns and McDonnell Engineering Co.
P.O. Box 173
Kansas City, MO 64141
816/333-4375
Alex Weir, Jr.
Manager, Chemical Systems, R&D
So. Cal. Edison Co.
P.O. Box 800
Rosemead, California 91770
213/572-2785
Don Welch
Staff Engineer
TRE-ASTECH
800 Hillcrest Rd. #5
Mobile, Alabama 36609
R. Murray Wells
Vice President
Radian Corp.
8501 Mo-Pac Blvd.
Austin, TX 78766
512/454-4797
Robert E. Whetstine
Energy Design Supervisor
Middle South Services, Inc.
P.O. Box 61000
New Orleans, Louisiana 70161
504/569-4730
Michael James Widico
Project Manager
Research Cottrell, Inc.
P.O. Box 1500
Somerville, New Jersey 08876
201/685-4213
David S. Wiggins
Supv. Process Engineering
United Engineers
30 S. 17th St.
Philadelphia, Pennsylvania 19101
609/772-0600
Wilbert W. Wiitala
Director
Marquette Board of Light & Power
2200 Wright
Marquette, Michigan 49855
906/228-6900
W. J. Wijdeveld
Manager, APC Division
ESMIL Intl.
De Boelelaan 7
Amsterdam, Netherlands
020-5411054
Dale S. Wiley
Mechanical Engineer
Wisconsin Public Service Corp.
P.O. Box 1200
Green Bay, Wisconsin 54305
414/433-1274
A-49
-------
James H. Wilhelm
President
Codan Assoc.
2394 Charros Rd.
Sandy, Utah 84092
801/571-6974
Don Wilhelm
Staff Chemical Engineering
Morrison Knudsen Co. Inc.
P.O. Box 7808
Boise, Idaho 83729
208/345-5000
John E. Williams
Chemical Engineer, Project Manager
DOE, Pittsburgh Energy Technical Ctr.
P.O. Box 10940
Pittsburgh, Pennsylvania 15236
412/675-5727
D. A. Williams
Sales Manager
Davu McKee Corp.
471b S. Florida Ave.
Lakeland, Florida 33803
813/646-7844
C. Bailey Williams
Vice President, General Manager
Marcona Ocean Ind., Ltd.
1001 N. W. 62nd St., Suite 200
Fort Lauderdale, FL 33309
305/776-4000
R. A. Wilson
Engineering Supervisor
Bechtel Power Corp.
12400 East Imperial Highway
Norwalk, California 90650
213/807-2767
David A. Wilson
Research Leader
Dow Chemical U.S.A.
Blag B-1222
Freeport, Texas 77541
409/236-7737
Gene Winkler
Application Engineer
Munters Corp.
P.O. Box 6428
Ft. Meyers, Florida 33911
813/936-1555
Lloyd Winsor
Asst. Chief Consulting Engineer
Gibbs & Hill Inc.
11 Penn Plaza
New York, New York 10001
212/760-5700
Jozewicz Wojciech
Research Scholar
University of Texas at Austin
E.P. Schoch Lab.
Austin, Texas 78712
512/471-4857
Stanley J. Wojton
Senior Engineer
Cleveland Electric
P.O. Box 97
Perry, Ohio 44081
216/259-3737
Illuminating Co.
Steve Wolf
Engineer
Northern State Power
414 Nicollet Mall
Minneapolis, Minnesota
612/330-5624
55401
Steve Wolsiffer
Assistant Plant Superintendent
Indianapolis Power & Light Co.
P.O. Box 436
Petersburg, Indiana 47567
812/354-b801
John M. Wootten
Director, Environmental Services
Peabody Coal Co.
301 N. Memorial Dr.e
St. Louis, Missouri 63102
314/342-3400
A-50
-------
R. M. Wright
Product Sales Manager
FMC Corp.
2000 Market St.
Philadelphia, Pennsylvania 19103
215/299-6815
Steve Wright
Senior Engineer
Montana Power Co.
40 E. Broadway
Butte, Montana 59701
406/723-5421
Charles R. Wright
Instrument Design Engineer
Tennessee Valley Authority
501 Chemical Engineering Bldg.
Muscle Shoals, Alabama 35660
205/386-2514
Robert J. Wright
President
Zurn Industries, Inc.
P.O. Box 2206
Birmingham, Alabama 35201
205/252-2181
Charles E. Wright, Jr.
Supervisor Applications
Combustion Engineering
31 Inverness Center Pkwy.
Birmingham, Alabama 35213
205/967-9100
Beth A. Wrobel
Engineer
Northern Indiana Public Service Co.
Rt. #1 Box 320
Wheatfield, Indiana 46392
219/956-5251
Jarnes E, Wuchter
Project Leader
Badische Corp.
602 Copper Rd.
Freeport, Texas 77541
409/238-6284
John Yavorsky
Project Leader
ASARCo
901 Oak Tree Rd., S.
Plainfield, New Jersey 07080
201/756-4800
Robert Yeargan
Power Plant Superintendent
Tennessee Valley Authority
510 Edney Bldg.
Chattanooga, TN 37401
615/751-4909
Charles S. Young
Product Manager, New Products
Astro Metallurgical
3225 Lincoln Way West
Wooster, Ohio 44691
216/264-8639
Dean Young
Vice President
Kissick Corp.
108 Benson East
Jenkintown, Pennsylvania 19046
215/885-6650
George J. Ziegenhorn
Senior Environmental Engineer
Arco Petroleum Products Co.
400 E. Sibley Blvd.
Harvey, Illinois 60426
312/333-3000, ext. 359
Denis M. Zielinski
Environmental Scientist
U.S. EPA
6th & Walnut Streets
Philadelphia, Pennsylvania 19101
215/597-0804
Jan T. Zmuda
Senior Development Engineer
Research-Cottrell
P.O. Box 1500
Somerville, New Jersey 08876
201/685-4915
A-51
-------
Joseph Zuzolo
Project Manager
York Research Consultants
938 Quail St.
Denver, Colorado 80215
3U3/233-1513
John C. deRuyter
Senior Engineer
E. I. du Pont de Nemounrs
ana Co., Inc.
Engineering Department, Louviers
Wilmington, Delaware 19898
302/366-6442
Bldg.
A-52
-------
EPRI CS-3706, Volume 2
1. EPRI CS-3706
4. Flue Gas Des
— <
-. c
su 3
o ro
33
<§
6
Cross-Referen
o
CD
w
m
-a
33
T)
"TO
o
05
CO
CD
H
S
o
w
<<
Timely exchan
tion (FGD) sys
standards. Th<
architectural-e
sions about th
CD 13
Q.CO
-*
g3|
£S£
~ W 3.
CQ D °
en — 8L
O W 03
° z ^
| o Q.
^ O n
C 03 O
^ c o
» =3
Proceed!
Novembi
^ 3
(O ^
Contractor
3J
CD
W
CD
to
o
03
CO
CD
5"
CO
c
CD
EPRI CS
Volume :
RP982-3"
"
o
o>
Proceedi
Desulfuri
N 3
•"•• C/3
i'm
CQ
<; ^3"
2. 5
C
M|
O
PROGR
>
2
DESULI
Tl
C
3]
N
H
O
•z.
•u
O
O
m
Below ar
four cros
abstract
is includi
EPRI CS-3706
Volume 2
RP982-31
Proceedings
November 1984
S fflfo £2. £
CO
o
CD
C
-^
jrizatio
D
TJ
0
O
CD
CO
O
(O
a)
3
*r*
EPRI CS-3706, Vol
Flue Gas Desulfui
FT|
sj
o ro
D
K>
3
fo
w
OJ
O
ro
Cfl
c
c"
Q>
o"
13
O
o
CO
CO
CD
CO
?
o
ross-References
m
T)
T)
"CD
o
w
CQ
CD
S
o
£
CO
;*r
•<
li
a§
•n §
O CD
o ~
& ^
's. °
CD" 3
Pa
3
O1 0)
CO c?
•D a
T3 CD
' O
a) cu
rchitectural-engi
ions about the d<
w 3
2. ro
CQ CD
D I?-
D
= Q.
2. W
5 'w
P CD
= ?
O 73
5 ™
ll
^i. o
-^ ^
O CD
D -
<" ^
"5 0
if 3
3 c
«.
3
Ol Q3
CD J
"0 Q.
P CD
c vu E
cF QJ 3
CD q S.
(ft -^ — -
0 C 0
• c 2. =>
0 q 9
111 5
g S. E
c ^.
— . ^. 03
^i' ^^. ^"
^ CD Q-
3 CD
CO
w E.
o' ^
3 N"
W (jj
w i: H
i mely exchanges
on (FGD) system
tandards. These
"5 m o
o g> =!•
» "" 2
CD CO 0
9:^ f
^ CD =-
CQ 13 °
M - e.
o a; a,
° -» ^
S ° °-
= 8 S
C 03 O
fD c O
|||
CTJO
CD fl) 3
CD 3 S.
o S 9
^ CO =1
o q 2
<" 1 =
~" K CD —
O ;4. c
^ 0)
3) < m
-0 0 -0
CD c 3D
00 3 —
^ to ^
O
en
n -n
'roceeding
)esulfuriza
zr. en
o ' '
CQ"
2. 5
Q
ro 3
O
C/3
c'
3
o
Tl
C
CD
o
Q)
O>
n
mm
^W
4
^
mmmm
n
i-r
\
m
CO
1
j
i
4
j
•
T|
1-
m
0
o
m
CO
c
T]
c
DO
N
5
O
z
1
1
J
EPRI CS-3706
Volume 2
RP982-31
Proceedings
November 1984
EPRI CS-3706
Volume 2
RP982-31
Proceedings
November 1984
Below are five index cards that allow for fMing according io the
Terences in addition to the title of the report. A brief
abstract describing the major subject area covered in the report
EPRI
Proceedings: Eighth Symposium on Flue Gas
Desulfurization Volume 2
Contractor: Research Triangle Institute
Timely exchanges of technical and economic information on flue gas desulfuriza-
tion (FGD) systems are essential to coal utilities that must meet strict emissions
standards. These proceedings constitute a valuable resource for utility,
architectural-engineering, and system-supplier personnel who must make deci-
sions about the design, installation, and operation of FGD systems. 578 pp.
EPRI Project Manager: T. M. Morasky
Cross-References:
1. EPRI CS-3706, Volume 2
4. Flue Gas Desulfurization
2. RP982-31
ELECTRIC POWER RESEARCH
Post Office Box 10412, Palo Alto, CA 94303
EPRI CS-3706, VOLUME 2
3. Desulfurization Processes Program
N S T I T U T E
415-855-2000
EPRI
Proceedings: Eighth Symposium on Flue Gas
Desulfurization volume 2
Contractor: Research Triangle Institute
Timely exchanges of technical and economic information on flue gas desulfuriza-
tion (FGD) systems are essential to coal utilities that must meet strict emissions
standards. These proceedings constitute a valuable resource for utility,
architectural-engineering, and system-supplier personnel who must make deci-
sions about the design, installation, and operation of FGD systems. 578 pp.
EPRI Project Manager: T. M. Morasky
Cross-References:
1. EPRI CS-3706, Volume 2
4. Flue Gas Desulfurization
2. RP982-31
3. Desulfurization Processes Program
ELECTRIC POWER RESEARCH INSTITUTE
Post Office Box 10412. Palo Alto, CA 94303 415-855-2000
RP982-31
EPRI
Proceedings: Eighth Symposium on Flue Gas
Desulfurization Volume 2
Contractor: Research Triangle Institute
Timely exchanges of technical and economic information on flue gas desulfuriza-
tion (FGD) systems are essential to coal utilities that must meet strict emissions
standards. These proceedings constitute a valuable resource for utility,
architectural-engineering, and system-supplier personnel who must make deci-
sions about the design, installation, and operation of FGD systems. 578 pp.
EPRI Project Manager- T. M. Morasky
Cross-References:
1. EPRI CS-3706, Volume 2
4. Flue Gas Desulfurization
3 Desullunzation Processes Progran
E L E C T R
Posl Ollice
C POWER RESEARCH INSTITUTE
Box 10412. Palo Alto. CA 94303 415-855 2000
------- |