I2PA-600 / R-9 3-064 e
April 1993
SEPA Research and
Development
PROCEEDINGS:
1991 S02 CONTROL SYMPOSIUM
Volume 5. Session 8
United States
Environmental Protection
Agency
Prepared for
Office of Air Quality Planning and Standards
Prepared by
Air and Energy Engineering Research
Laboratory
Research Triangle Park NC 27711
-------
technical report data
(f'le&e read Jnuruetions on the revrrw before eo"tp!c
i nrronT no. ?
EPA-600/R-93-0G4e
PB93—196137
•4. title ano suotitlc
Proceedings: 1991 SO2 Control Symposium, Volume 5.
Session 8
•_> nCPORT OATE
April 1993
0. PERFORMING ORGANIZATION cooe
7. AUTHOR(S)
Miscellaneous
8. PERFORMING ORGANIZATION REPORT NO.
TR-101054 (1)
9. PERFORMING ORGANIZATION NAME ANO ADOR6SS
See Block 12
lO. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
NA (Inhouse)
12. SPONSORING AGENCY NAME ANO AOORESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT ANO PERIOD COVERED
Proceedings; 1S91
14. SPONSORING AGENCY COOE
EPA/600/13
is.supplementary notes AEERL. project officer is Brian K. Gullett, Mail Drop 4. 919/541-
1534. Cosponsored by EPRI and DOE. Vol. 1 is Opening Session and Sessions 1-3,
Vol. 2 is sessions 4 and 5A, Vol. 3 is Sessions 5B and 6. and Vol. 4 is Session 7.
i6. abstract proceedings document the 1991 S02 Control Symposium, held December
3-6, 1991, in Washington, DC, and jointly sponsored by the Electric Power Research
Institute (EPRI), the U. S. Environmental Protection Agency (EPA), and the U. S. De-
partment of Energy (DOE). The symposium focused attention on recent improve-
ments in conventional S02 control technologies, emerging processes, and strategies
for complying with the Clean Air Act Amendments (CAAA) of 1990. It provided an in-
ternational forum for the exchange of technical and regulatory information on S02
control technology. More than 800 representatives of 20. countries from government,
academia, flue gas desulfurization (FGD) process suppliers, equipment manufac-
turers, engineering firms, and utilities attended. In all, 50 U. S. utilities and 10
utilities in other countries were represented. In 11 technical sessions, speakers
presented 111 technical papers on development, operation, and commercialization of
wet and dry FGD, clean coal technologies, and combined sulfur oxide/nitrogen oxide
(SOx/NOx) processes.
17. KEY WORDS ANO DOCUMENT ANALYSIS
a. DESCRIPTORS !".>.C?rf:TIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Sulfur Dioxide
Nitrogen Oxides
Flue Gases
Desulfurization
Coal
Pollution Control
Stationary Sources
13 B
07B
21B
07A,07D
2 ID
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
314
20.SECURITY CLASS (THISpage]
Unclassified
22. PRICE
EPA Form 2220-1 (9-73) 8B~143
»
-------
ORDERING INFORMATION
Requests tor copies ol this report should be directed to the EPRI Distribution Center. 207 Coggins Drive,
P.O. Box 23205. Pleasant Hill. CA 9<523. (510) 934-4212. There is no charge 1or reports requested by
EPRI member utilities and afliliates. *
(*) Copies of this report are also available to the public, prepaid,
through the National Technical Information Service, 5285 Port
Royal Road, Springfield, VA 22161.
Copyright (c) 1992, EPRI TR-101054. "Proceedings: 1991 SO2 Control
Symposium. Volumes 1, 2. and 3." Since this work was, in part,
funded by the U. S. Government, the Government is vested with a royalty-
free, non-exclusive, and irrevocable license to publish, translate, re-
produce. and deliver that information and to authorize others to do so.
Electric Power Research Institute and EPRI are registered service marts o( Electric Power Research Institute. Inc.
NOTICE
IMs report was prepared by 0>e aecaric Power Research Indue. Inc. (EPH). Neieier EPRL members ot EPft.
nor ary person aaing on flw befofc (a) mates any wararty: express or impfed. wi#» respect 10 «w use of any
iajn.ff.Mini. -n-M.. ~ 1 - —T- —r— — — Ty ^ jrtrinpr
owned rights or (b) assunes any tabUes wiai respect to M use OL. or tor damages Msattng horn the use c(.
any Mormrfon. apparatus, method, or process <£sdosed in Ms report.
-------
]-:P/\-600/R-93-064e
.April 1993
Proceedings: 1991S02 Control Symposium
Volume 5. Session 8
For Sponsors:
Electric Power Research Institute OS-Department of Energy lis. Environmental Protection Agency
B. Toole O'Netl
3412 Hill view Avenue
Palo Alto. CA 94304
Charles J. Drummond
Pittsburgh Energy
Technology Center
P.O. Box 10940
Pittsburgh. PA 15236
Brian K. Guile tt
Air and Energy Engineering
Research Laboratory
Research Triangle Park. NC 27711
-------
ABSTRACT
These are the Proceedings of the 1991 SO2 Control Symposium held December 3-6,
1991, in Washington, D.C. The symposium, jointly sponsored by the Electric Power
Research Institute (EPRI), the U-S. Environmental Protection Agency (EPA), and the
U.S. Department of Energy (DOE), focused attention on recent improvements in
conventional sulfur dioxide (SO2) control technologies, emerging processes, and
strategies for complying with the Clean Air Act Amendments of 1990. This is the
first SO2 Control Symposium co-sponsored by EPRI, EPA and DOE. Its purpose was
to provide a forum for the exchange of technical and regulatory information on SO2
control technology.
Over 850 representatives of 20 countries from government, academia, flue gas
desulfurization (FGD) process suppliers, equipment manufacturers, engineering
firms, and utilities attended. In all, 50 U.S. utilities and 10 utilities in other
countries were represented. A diverse group of speakers presented 112 technical
papers on development, operation, and commercialization of wet and dry FGD,
Clean Coal Technologies, and combined sulfur dioxide/nitrogen oxides (SO2/NOX)
processes. Since the 1990 SO2 Control Symposium, the Clean Air Act Amendments
have been passed. Clean Air Act Compliance issues were discussed in a panel
discussion on emission allowance trading and a session on compliance strategies for
coal-fired boilers.
ii
-------
CONTENTS
PREFACE
AGENDA
VOLUME 1
Opening Session
EPRI Perspective
EPA Perspective
DOE Perspective
Guest Speakers
Shelley Fidler - Assistant, Policy Subcommittee on
Energy and Power, U-S. Congress
Jack S. Siegd - Deputy Assistant Secretary, Office of Coal
Technology, U-S. Department of Energy
Michael Shapiro - Deputy Assistant Administrator, Office
of Air and Radiation, U.S. Environmental Protection Agency
Session 1 - Clean Air Act Compliance Issues/Panel
Session 2 - Clean Air Act Compliance Strategies
Scrubbers: A Popular Phase 1 Compliance Strategy
Scrub Vs. Trade: Enemies or Allies?
Evaluating Compliance Options
Clean Air Technology (CAT) Workstation
Economic Evaluations of 28 FGD Processes
Strategies for Meeting Sulfur Abatement Targets in the
UK Electricity Supply Industry
iii
-------
Compliance Strategy for Future Capacity Additions: The Role of
Organic Acid Additives
A Briefing Paper for the Status of the Flue Gas Desulfurization
System at Indianapolis Power & Light Company
Petersburg Station Units 1 and 2
Evaluation of SO2 Control Compliance Strategies at Virginia Power
Session 3A - Wet FGD Process Improvements
Overview on the Use of Additives in Wet FGD Systems
Results of High SO2 Removal Efficiency Tests at EPRI's High
Sulfur Test Center
Results of Formate Ion Additive Tests at EPRI's High Sulfur
Test Center
FGDPRISM, EPRTs FGD Process Model-Recent Applications
Additive-Enhanced Desulfurization for FGD Scrubbers
Techniques for Evaluating Alternative Reagent Supplies
Factors Involved in the Selection of Limestone Reagents for Use in
Wet FGD Systems
Magnesium-Enhanced Lime FGD Reaction Tank Design Tests
at EPRI's HSTC
Session 3B - Furnace Sorbent Injection
Computer Simulations of Reacting Partide-Laden Jet Mixing
Applied to SO2 Control by Dry Sorbent Injection
Studies of the Initial Stage of the High Temperature
Ca0-S02 Reaction
Status of the Tangentially Fired LIMB Demonstration Program
at Yorktown Unit No. 2: An Update
Results from LIMB Extension Testing at the Ohio Edison
Edgewater Station
iv
-------
VOLUME 2
Session 4A - Wet FGD Design Improvements
Reliability Considerations in the Design of Gypsum Producing
Flue Gas Desulfurisation Plants in the UK
Sparing Analysis for FGD Systems
Increasing Draft Capability for Retrofit Flue Gas Desulfurization
Systems
Development of Advanced Retrofit FGD Designs
Add Rain FGD System Retrofits
Guidelines for FGD Materials Selection and Corrosion Protection
Economic Comparison of Materials of Construction of Wet FGD
Absorbers and Internals
The Intelligence & Economics of FKP in F.G.D. Systems
4A-1
4A-25
4A-41
4A-61
4A-79
4A-99
4A-125
4A-141
Session 4B - Dry FGD Technologies
UFAC Demonstration at Poplar River
\U MW Pilot Results for the Duct Injection FGD Process Using
Hydrated Lime Upstream of an ESP
Scaleup Tests and Supporting Research for the Development
of Duct Injection Technology
A Pilot Demonstration of the Moving Bed limestone Emission
Control (LEQ Prooess
Pilot Plant Support for ADVACATE/MDI Commercialization
Suitability of Available Fly Ashes in ADVACATE Sorbents
Mechanistic Study of Desulfurization by Absorbent Prepared
from Coal Fly Ash
Results of Spray Dryer/Pulse-Jet Fabric Filter Pilot Unit Tests
at EPRI High Sulfur Test Center
4B-1
4B-17
4B-39
4B-61
4B-79
4B-93
4B-113
4B-125
v
-------
Results of Medium- and High-Sulfur Coal Tests on the TVA
10-MW SD/ESP Pilot Plant 4B-151
Evolution of the B&W Durajet™ Atomizer 4B-173
Characterization of the Linear VGA Nozzle for Flue
Gas Humidification 4B-189
High SO2 Removal Dry FGD Systems 4B-205
Session 5A - Wet Full Scale FGD Operations
FGD System Retrofit for Dalhousie Station Units 1 & 2 5A-1
Zinuner FGD System: Design, Construction, Start-Up
and Operation 5A-17
Results of an Investigation to Improve the Performance and
Reliability of HL&Fs Limestone Electric Generating Station
FGD System 5A-37
Full-Scale Demonstration of EDTA and Sulfur Addition to
Control Sulfite Oxidation 5A-59
Optimizing die Operations in the Flue Gas Desulfurization Plants
of the Lignite Power Plant Neurath, Unit D and E and Improved
Control Concepts for Third Generation Advanced FGD Design 5A-81
Organic Acid Buffer Testing at Michigan South Central Power
Agency's Endicott Station 5A-101
Stack Gas Cleaning Optimization Via German Retrofit Wet
FGD Operating Experience 5A-127
Operation of a Compact FGD Plant Using CT-121 Process 5A-143
VOLUME 3
Session 5B - Combined SOx/NOx Technologies
Simultaneous SOx/NOx Removal Employing Absorbent Prepared
from Fly Ash 5B-1
Furnace Slurry Injection for Simultaneous SQ2/NOx Removal 5B-21
Combined SQ2/NOx Abatement by Sodium Bicarbonate
Dry Injection 5B-41
VT
I
-------
S02 and NOx Control by Combined Dry Injection of Hydrated
Lime and Sodium Bicarbonate 5B-67
Engineering Evaluation of Combined NOx/SC>2 Controls for
Utility Application 5B-79
Advanced Flue Gas Treatment Using Activated Char Process
Combined with FBC 5B-10.1
Combined SC>2/NOx Control using Ferrous • EDTA and a
Secondary Additive in a Lime-Based Aqueous Scrubber System 5B-125
Recent Developments in the Parsons FGC Process for Simultaneous
Removal of SOx and NOx 5B-141
Session 6A - Wet FGD Operating Issues
Pilot-Scale Evaluation of Sorbent Injection to Remove SO3 and HC1 6A-1
Control of Add Mist Emissions from FGD Systems 6A-27
Managing Air Toxics: Status of EPRI PISCES Project 6A-47
Results of Mist Eliminator System Testing in an Air-Water
Pilot Facility 6A-73
CEMS Vendor and Utility Survey Databases 6A-95
Determination of Continuous Emissions Monitoring
Requirements at Electric Energy, Inc. 6A-117
Improving Performance of Flushless Mechanical Seals in Wet FGD
Plants through Field and Laboratory Testing 6A-139
Sulcis FGD Demonstration Plant limestone-Gypsum Process:
Performance, Materials, Waste Water Treatment 6A-163
Session 6B - Clean Coal Demonstrations
Recovery Scrubber - Cement Application Operating Results 6B-1
The NOXSO Clean Coal Technology Demonstration Project 6B-17
vii
-------
Economic Comparison of Coolside Sorbent Injection and Wet
Limestone FGD Processes 6B-33
Ohio Edison Clean Coal Projects Circa: 1991 6B-55
Sanitech's 2.5-MWe Magnesia Dry-Scrubbing Demonstration
Project 6B-79
Application of DOW Chemical's Regenerable Rue Gas
Desulfurization Technology to Coal-Fired Power Plants 6B-93
Pilot Testing of the Cansolv® System FGD Process 6B-105
Dry Desulphurization Technologies Involving Humidification
for Enhanced SO2 Removal 6B-119
VOLUME 4
Session 7 - Poster Papers
Summary of Guidelines for the Use of FRP in Utility FGD
Systems 7-1
Development and Evaluation of High-Surface-Area Hydrated
lime for SO2 Control 7-13
Effect of Spray Nozzle Design and Measurement Techniques on
Reported Drop Size Data 7-29
High SO2 Removals with a New Duct Injection Process 7-51
Combined SOx/NOx Control Via Soxal™, A Regenerative Sodium
Based Scrubbing System 7-61
The Healy Clean Coal Project Air Quality Control System 7-77
Lime/lime Stone Scrubbing Producing Usable By-Products 7-93
Modeling of Furnace Sorbent Injection Processes 7-105
Dry FGD Process Using Calcium Sorbenis 7-127
Clean Coal Technology Optimization Model 7-145
SNRB Catalytic Baghouse Process Development and Demonstration 7-157
Reaction of Moist Calcium Silicate Reagents with Sulfur Dioxide
in Humidified Flue Gas 7-181
viii
-------
i
Commercial Application of Dry FGD using High Surface Area
Hydrated Lime 7-199
Initial Operating Experience of the SNOX Process 7-221
Progress Report of the NIPSCO - Pure Air - DOE Clean Coal II
Project 7-241
Development of a Post Combustion Dry SC>2 Control Reactor
for Small Scale Combustion Systems 7-253
Scrubber Reagent Additives for Oxidation Inhibited Scrubbing 7-269
Recovery of Sulfur from Calcium Sulfite and Sulfate
Scrubber Sludges 7-277
Magnesite and Dolomite FGD Technologies 7-285
SO2 and Particulate Emissions Reduction in a Pulverized Coal
Utility Boiler through Natural Gas Cofiring 7-293
Design, Installation, and Operation of the First Wet FGD for a
Lignite-Pored Boiler in Europe at 330 MW P/S Voitsberg 3 in Austria 7-321
VOLUME 5
Session 8A - Commercial FGD Designs
Mitsui-BF Dry Desulfurization and Denitrification Process
Using Activated Coke 8A-1
High Efficiency, Dry Flue Gas SOx, and Combined SOx/NOx
Removal Experience with Lurgi Circulating Fluid Bed
Dry Scrubber - A New, Economical Retrofit Option for U.S.
Utilities for Acid Rain Remediation 8A-21
incorporating Full-Scale Experience into Advanced Limestone
Wet FGD Designs 8A-43
Design and Operation of Single Train Spray Tower FGD Systems 8A-69
Selecting the FGD Process and Six Years of Operating Experience
in Unit 5 of the Altbach-Deizisau Neckarwerke Power Station 8A-93
Development and Operating Experience of FGD-Tedmique at the
Voelklingen Power Station 8A-121
Advantages of the CT-121 Process as a Throwaway FGD System 8A-135
ix
-------
Session 8B - By-Product Utilization
German Experience of FGD By-Product Disposal and Utilization 8B-1
The Elimination of Pollutants from FGD Wastewaters 8B-25
The Influence of FGD Variables jDn FGD Performance and
By-Product Gypsum Properties 8B-47
Quality of FGD Gypsum 8B-69
Chemical Analysis and Howability of ByProduct Gypsums 8B-91
Evaluation of Disposal Methods for Oxidized FGD Sludge 8B-113
Commercial Aggregate Production from FGD Waste 8B-127
x
I
-------
PREFACE
The 1991 SO2 Control Symposium was held December 3-6,1991, in Washington,
D.C. The symposium, jointly sponsored by the Electric Power Research Institute
(EPRI), the U.S. Environmental Protection Agency (EPA), and the U.S. Department
of Energy (DOE), focused attention on recent improvements in conventional sulfur
dioxide (SO2) control technologies, emerging processes, and strategies for complying
with the Clean Air Act Amendments of 1990.
The proceedings from this Symposium have been compiled in five volumes,
containing 111 presented papers covering i4 technical sessions:-
Session
Subject Area
I
Opening Remarks by EPRI^EPA and DOE Guest Speakers
1
Emission Allowance Panel Discussion
2
Clean Air Act Compliance Strategies
3A
Wet FGD Process Improvements
3B
Furnace Sorbent Injection
4A
Wet FGD Design Improvements
4B
Dry FGD Technologies
5A
Wet FGD Full Scale Operations
5B
Combined SOx/NOx Technologies
6A
Wet FGD Operating Issues
6B
Clean Coal Demonstratioins/Emerging Technologies
7
Poster Session - papers on all aspects of SO2 control
8A
Commercial FGD Designs
8B
FGD By-Product Utilization
These proceedings also contain opening remarks by the co-sponsors and comments
by the three guest speakers. The guest speakers were Shelley Fidler - Assistant,
Policy subcommittee on Energy and Power, U. S. Congress,
Jack . . S. Siegel - Deputy Assistant Secretary, Office of Coal Technology, U-S-
Department of Energy, and Michael Shapiro - Deputy Assistant Adminstrator,
Office of Air and Radiation* U. S. Environmental Protection Agency.
The assistance of Steve Hoffman, independent, in preparing the
manuscript is gratefully acknowledged.
The following persons organized this symposium:
• Barbara Toole 0*Neil - Co-Chair, Electric Power Research Institute
• Charles Drummond - Co-Chair, US. Department of Energy
• Brian K. Gullett - Co-Chair, US. Environmental Protection Agency
• Pam Turner and Ellen Lanum - Symposium Coordinators, Electric Power
Research Institute
xi
-------
AGENDA
1991SO2 CONTROL SYMPOSIUM
Opening Session
Session Chain M. Maxwell - EPA
1-1 EPRI Perspective - S.M. Dalton
1-2 EPA Perspective - M. Maxwell
DOE Perspective - P. Bailey (no written manuscript)
Guest Speakers
Shelley Fidler - Assistant, Policy subcommittee on energy and
Power, U. S. Congress
Jack S. Siegel - Deputy Assistant Secretary, Office of Coal _
Technology, U.S. Department of Energy
Michael Shapiro - Deputy Assistant Adminstrator, Office of Air
and Radiation, U. S. Environmental Protection Agency
Session 1 - Clean Air Act Compliance Issues/Panel
Session Moderator S. Jenkins, Tampa Electric Co.
Comments by:
Alice LeBlanc - Environmental Defense Fund
Karl Moor, Esq., Balch & Bingham
John Palmisano AER*X
Craig A. Glazer - Chair, Ohio Public Utilities Commission
xii
-------
I
Session 2 -Clean Air Act Compliance Strategies
Session Chain Paul T. Radcliffe - EPRI
2-1 Scrubbers: A Popular Phase 1 Compliance Strategy, P.E. Bissell,
Consolidation Coal Co.
2-2 Scrub Vs. Trade: Enemies or Allies? J. Piatt, EPRI
2-3 Evaluating Compliance Options, J.H. Wile, National Economic
Research Association, Inc.
2-4 Clean Air Technology Workstation, D. Sopocy, Sargent & Lundy
2-5 Economic Evaluations of 27 FGD Processes, R.J. Keeth, United
Engineers & Constructors
2-6 Strategies for Meeting Sulfur Abatement Targets in the UK Electricity
Supply Industry, WS. Kyte, PowerGen
2-7 Compliance Strategies for Future Capacity Additions: The Role of
Organic Acid Additives, C.V. Weilert, Burns & McDonnell Engineerir
Co.
2-8 IPL Petersburg 1 & 2 CAAA Retrofit FGDs, CP. Wedig, Stone &
Webster Engineering Corp.
2-9 Evaluation of SO2 Control Compliance Strategies at Virginia Power,
J.V. Presley, Virginia Power
Session 3A Wet FGD Process Improvements
Session Chain David R. Owens - EPRI
Overview on the Use of Additives in Wet FGD Systems, R.E. Moser,
EPRI
Results of High SO2 Removal Efficiency Tests at EPRTs HSTC, G.
Stevens, Radian
Results of Formate Additive Tests at EPRTs HSTC, M. Stohs, Radian
Corp.
FGDPRISM, EPRTS FGD Process Model-Recent Applications, J.Gl
Noblett, Radian Corp.
Additive Enhanced Desulfurization for FGD Scrubbers, G. Juip,
Northern States Power
Techniques for Evaluating Alternative Reagent Supplies, C.V. Weilert
Bums & McDonnell Engineering Co.
3A-7 Factors Involved in the Selection of Limestones for Use in Wet FGD
Systems, J.B. Jarvis, Radian Corp.
3A-8 Magnesium-Enhanced lime Reaction Tank Design Tests at EPRTs
HSTC, J. Wilhelm, Codan Associates
xiii
3A-1
3A-2
3A-3
3A-4
3A-5
3A-6
-------
Session 3B - Furnace Sorbent Injection
Session Chain Brian Gullett - EPA
3B-1 Computer Simulation of Reacting Particle-Laden Jet Mixing Applied to
SO2 Control by Dry Sorbent Injection, P.J. Smith, The University of
Utah
3B-2 Studies of the Initial Stage of the High Temperature Ca0-S02 Reaction,
L Bjerle, University of Lund
3B-3 Status of the Tangentially Fired LIMB Demonstration Program at
Yorktown Unit No. 2: An Update, J.P. Clark, ABB Combustion
Engineering Systems
3B-4 Results from LIMB Extension Testing at the Ohio Edison Edgewater
Station, T. Goots, Babcock & Wilcox
Session 4A - Wet FGD Design Improvements
Session Chain Richard E. Tischer - DOE
4A-1 Reliability Considerations in the Design of Gypsum Producing Flue Gas
Desulfurization Plants in UK, L Gower, John Brown Engineers &
Constructors Ltd.
4A-2 Sparing Analysis for FGD Systems, M. A. Twombly, ARINC Research
Corp.
4A-3 Increasing Draft Capability for Retrofit Flue Gas Desulfurization
Systems, R.D. Petersen, Burns & McDonnell Engineering Co.
4A-4 Development of Advanced Retrofit FGD Designs, C.E. Dene, EPRI
4A-5 Acid Rain FGD Systems Retrofits, A.J. doVale, Wheelabrator Air
Pollution Control
4A-6 Guidelines for FGD Materials Selection and Corrosion Protection, HS.
Rosenberg, Batelle
4A-7 Economic Comparison of Materials of Construction of Wet FGD
Absorbers & Internals, W. Nischt, Babcock & Wilcox
4A-8 Hie Intelligence & Economics of FJLP. in F.G.D. Systems, E.J. Boucher,
RPS/ABCO
xiv
-------
Session 4B - Dry FGD Technologies
Session Chain Michael Maxwell /Brian Gullett/Norman Kaplan - EPA
4B-1 Poplar River LEFAC Demonstration^. Enwald, Tamp>ella Power Ltd.
4B-2 1.7 MW Pilot Results for Duct Injection FGD Process Using Hydra ted
Time Upstream of an ESP, M. Maibodi, Radian Corp.
4B-3 Scaleup Tests and Supporting Research for the Development of Duct
Injection Technology, M.G. Klett, Gilbert/Commonwealth Inc.
4B-4 A Pilot Demonstration of the Moving Bed Limestone Emission
Control Process (LEC), M.E. Prudich, Ohio University
4B-5 Pilot Plant Support for MDI/ADVACATE Commercialization, C.'
Sedman, US. EPA
4B-6 Suitability of Available Fly Ashes in ADVACATE Sorbents, C. Singer,
US. EPA
4B-7 Mechanistic Study of Desulfurization by Absorbent Prepared from Coal
Fly Ash, H. Hattori, Hokkaido University
4B-8 Results of Spray Dryer/Pulse-Jet Fabric Filter Pilot Unit Tests at EPRI
HSTC, G. Blythe, Radian Corp.
4B-9 Results of Medium & High-Sulfur Coal Tests on the TV A 10-MW
Spray Dryer/ESP Pilot, T. Burnett, TVA
4B-10 Evolution of the B&W Durajet™ Atomizer, S. Feeney, Babcock &
Wilcox
4B-11 Characterization of the Linear VGA Nozzle for Flue Gas
Humidification, J.R. Butz, ADA Technologies, Inc.
4B-12 High SO2 Removal Dry FGD Systems, B. Brown, Joy Technologies, Inc.
Session 5A - Wet Full Scale FGD Operations
Session Chain Robert L. Glover - EPRI
FGD System Retrofit for Dalhousie Station Units 1 & 2, F.W. Campbell,
Burns & McDonnell Engineering Co.
Zinuner FGD System, W. Brockman, Cincinnati Gas & Electric
Results of on Investigation to Improve the Performance and Reliabiity
of HL&P*s Limestone Electric Generating Station FGD System, M.
Bailey, Houston Lighting & Power
Full-Scala Demonstration of EDTA and Sulfur Addition to Control
Sulfite Oxidation, G. Blythe, Radian
xv
5A-1
5A-2
5A-3
5A-4
-------
5A-5 Optimizing the Operations in the Flue Gas Desulfurization Plants of
the Lignite Power Plant Neurath Unit D and E and Improved Control
Concepts for Third Generation Advanced FGD Design, H. Scherer,
Noell, Inc.
5A-6 Organic Acid BufferTesting at Michigan South Central Power Agency's
Endicott Station, B. J. Jankura, Babcock & Wilcox
5A-7 Stack Gas Cleaning Optimization Via German Retrofit Wet FGD
Operating Experience, H. Weiler, Ellison Consultants.
5A-8 Operation of a Compact FGD Plant Using CT-121 Process, Y. Ogawa,
Chiyoda Corp.
Session 5B - Combined SOx/NOx Technologies
Session Chain Mildred E. Perry - DOE
5B-1 Simultaneous SOx/NOx Removal Employing Absorbent Prepared
from Fly Ash, H. Tsuchiai, The Hokkaido Electric Power Co.
5B-2 Furnace Slurry Injection for Simultaneous SO2/NOX Removal, BJC
Gullett, U.S. EPA
5B-3 Combined SO2/NOX Abatement by Sodium Bicarbonate Dry Injection,
J. Verlaeten, Solvay Technologies, Inc (124)
5B-4 SO2 and NOx Control by Combined Dry Injection of Hydrated Lime
and Sodium Bicarbonate, D. Helfritch, R-C Environmental Services &
Technologies
5B-5 Engineering Evaluation of Combined N0x/S02 Controls for Utility
Application, J.E. Qchanowicz, EPRI
5B-6 Advanced Flue Gas Treatment Using Activated Char Process
Combined with FBC, H. Murayama, Electric Power Development Co.
5B-7 S02/N0x Control using Ferrous EDTA and a Secondary Additive in a
Combined Lime-Based Aqueous Scrubber System, M.H. Mendelsohn,
Argonne National Laboratory
5B-8 Parsons FGC Process Simultaneous Removal of SOx and NOx, K.V.
Kwong, The Ralph M. Parsons Co.
xvi
-------
Session 6A - Wet FGD Operating Issues
Session Chain Gary M. Andes - EPRI
6A-1 Pilot-Scale Evaluation of Sorbent Injection to Remove SO3 and HQ, J.
Peterson, Radian Corp.
6A-2 Control of Acid Mist Emissions from FGD Systems, R-S. Dahlin,
Southern Research Institute
6A-3 Managing Air Toxics: Status of EPRI PISCES Project, W. Chow, EPRI
6A-4 Results of Mist Elimination System Testing in an Air-Water Pilot
Facility, A J. Jones, Radian Corp.
6A-5 CEM Vendor and Utility Survey Databases, J-L- Shoemaker,
Engineering Science, Inc
6A-6 Determination of Continuous Emissions Monitoring Requirements at
Electric Energy Inc., V. V. Bland, Stone & Webster Engineering Corp.
6A-7 Improving Performance of Flush!ess Mechanical Seals in Wet FGD
Plants through Held and Laboratory Testing, F.E. Manning, BW/IP
International Inc.
6A-8 Sulcis FGD Demonstration Plant Limestone-Gypsum Process:
Performance, Materials, Waste Water Treatment, E. Marchesi, Enel
Construction Department
Session 6B - Clean Coal Demonstrations
Session Chain Joseph P. Strakey - DOE
6B-1 Recovery Scrubber Cement Application Operating Results, G.L.
Morrison, Passamaquoddy Technology
6B-2 The NOXSO Clean Coal Technology Demonstration Project, L.G. Neal,
NOXSO Corp.
6B-3 Economic Comparison of Coolside Sorbent Injection and Wet
Limestone FGD Processes, D.C McCoy, Consolidation Coal Co.
6B-4 Ohio Edison's Clean Coal Projects: Circa 1991, R. Bolli, Ohio Edison
Emerging Technologies
6B-5 A Status Report on Sanitech's 2-MWe Magnesia Dry Scrubbing
Demonstration, S.G. Nelson, Sanitech Inc
6B-6 Application of DOW Chemical's Regenerable Flue Gas Desulfurization
Technology to Coal Fired Power Plants, LK Kirby, Dow Chemical
6B-7 Pilot Testing of the Cansolv System FGD Process, I.E. Hakka Union
Carbide Canada LTD.
6B-8 Dry Desulfurization Technology Involving Humidification for
Enhanced SO2 Removal, D.P. Singh, Procedair Industries Inc.
xvii
-------
Session 7 - Poster Papers
Session Chair Charles Sedman - EPA
7-1 Summary of Guidelines for the Use of FRP in Utility FGD Systems, W.
Renoud, Fiberglass Structural Engineering, Inc.
7-2 Development and Evaluation of High Surface Area Hydrated Lime for
SO2 Control, M. Rostam-Abadi, The Illinois State Geological Survey
7-3 Effect of Spray Nozzle Design and measurement Techniques on
Reported Drop Size Data, W. Bartell, Spraying Systems Co.
7-4 High SO2 Removals with a New Duct Injection Process, S.G. Nelson, Jr.
Sanitech, Inc.
7-5 Combined SOx/NOx Control Via Soxal™, A Regenerative Sodium
Based Scrubbing System , C.H. Byszewski, Aquatech Systems
7-6 The Healy Clean Coal Project Air Quality Control System, V.V. Bland,
Stone & Webster Engineering Corp.
7-7 Lime/Lime Stone Scrubbing Producing Useable By-Products, D. P.
Singh, Procedair Industries Inc.
7-8 Modeling of Furnace Sorbent Injection Processes, A.S. Damle, Research
Triangle Institute
7-9 Dry FGD Process Using Calcium Absorbents, N. Nosaka, Babcock-
Hitachi K.K.
7-10 Clean Coal Technology Optimization Model, B.A. Laseke, International
Technology Corp.
7-11 SNRB Catalytic Baghouse Process Development & Demonstration, K.E.
Redinger, Babcock & Wilcox
7-12 Reaction of Moist Calcium Silicate Reagents with Sulfur Dioxide in
Humidified Hue Gas, W. Jozewicz, Acurex
7-13 Commercial Application of Dry FGD using High Surface Area Hydrated
Lime, F. Schwarzkopf, Horian Schwarzkopf PE.
7-14 Initial Operatiing Experience of the SNOX Process, D.J. Collins, ABB
Environmental System
7-15 Progress Report of the NIPSCO - Pure Air - DOE Clean Coal II Project, S.
Satrom, Pure Air
7-16 Development of a Post Combustion Dry SO2 Control Reactor for Small
Scale Combustion Systems, J.C. Balsavich, Tecogen Inc.
xviii
-------
7-17 Scrubber Reagent Additives for Oxidation Inhibited Scrubbing, J.
Thompson, Process Calx, Inc.
7-18 Recovery of Sulfur from Calcium Sulfite and Sulfate Scrubber Sludges,
J. Thompson, Process Calx, Inc.
7-19 Magnesite & Dolomite FGD Technologies, D. Najmr, Ore Research
Institute
7-20 SOx and Particulate Emissions Reduction in a Pulverized Coal Utility
Boiler through natural Gas Cofiring, K.J. Clark Aptech Engineering
Services
7-21 Design, Installation, and Operation of the First Wet FGD for a ligiute
Fired Boiler in Europe at 330 MW P/S Voitsberg 3 in Austria, H.
Kropfitsch, Voitsberg
Session 8A - Commercial FGD Designs
Session Chair: Robert E. Moser - EPRI
8A-1 Mitsui-BF Dry Desulfurization and Utility Compliance Strategies, K.
Tsuji, Mitsui Mining Company Ltd.
8A-2 High Efficiency Dry Flue Gas SOx and Combined SOx/NOx Removal
Experience with Lurgi Circulating Fluid Bed Dry Scrubber; A New
Economical Retrofit Option for Utilities for Acid Rain Remediation, J.
G. Toher, Environmental Elements Corp.
8A-3 Incorporating Full-Scale Experience into Advanced Limestone Wet
FGD Designs, P.C. Rader, ABB Environmental Systems
8A-4 Design and Operation of Single Train Spray Tower FGD Systems, A.
Saleem, GE Environmental Systems
8 A-5 Selecting the FGD Process and Six Years of Operating Experience in
Unit 5 FGD of the Altbach-Deizisau Neckawerke Power Station, R.
Maule, Noell Inc.
8A-6 Development and Operating Experience of FGD Technique at the
Volkingen Power Station, H. Petzel, SHU-Technik
8A-7 Advantages of the CT-121 Process as a Throwaway FGD System, M.J.
Krasnopoler, Bechtel Corp.
xix
-------
Session 8B - By-Product Utilization
Session Chain Charles E. Schmidt - DOE
8B-1 German Experience of FGD By-Product Disposal and Utilization, J.
Demmich, Ncell Inc.
8B-2 The Elimination of Pollutants from FGD Wastewaters, M.K.
Mierzejewski, Infilco Degremont Inc
8B-3 The Influence of FGD Variables on FGD Performance and By-Product
Gypsum Properties,F. Theodore, Consolidation Coal Co.
8B-4 Quality of FGD Gypsum, F.W. van der Brugghen, N.V. Kema
8B-5 Chemical Analysis and Flowability of By-Product Gypsums, L-Kilpeck,
Centerior
8B-6 Evaluation of Disposal Methods Stabilized FGD & Oxidized FGD
Sludge & Fly Ash, W. Yu, Conversion Systems, Inc
8B-7 Commercial Aggregate Production from FGD Waste, C.L. Smith,
Conversion Systems, Inc
xx
-------
Session 8A
COMMERCIAL FGD DESIGNS
MITSUI-BF DRY DESULFURIZATION AMD DENITRIFICATION PROCESS
USING ACTIVATED COKE
K. Tsuj i
Mitsui Mining Company, Limited
2-1-1 Nihonbashi Muromachi
Chuoh Ward, Tokyo, Japan 103
I. Shiraishi
Mitsui Mining Company, Limited
1-3 Hibikimachi, Wakamatsu Ward
Kitakyushu City, Fukuoka, Japan 808
ABSTRACT
The problem of acid rain has been a growing concern in recent years.
In this paper, Mitsui-BF dry DeSOx/DeNOx process is introduced as an
effective process for reduction of SOx and NOx from flue gas. This
process can achieve 100% removal of SOx and over 80% removal of NOx
by contacting flue gas with activated coke and injecting NH3 for
DeNOx at temperature range of about 100-200° C (212-392 F) .
The dry desulfurization and denitrification process using activated
coke (AC) was originally researched and developed during the sixties
by Bergbau Forschung (BF)1), now called Deutsche Montan
Technologies. Mitsui Mining Company (MMC) concluded a licensing
agreement with BF to investigate, test and adapt the system to the
facilities in Japan where the regulations are stricter towards
SOx/NOx pollutants, as well as dust emissions from the utility
industry, oil refineries and other industries. There sure four
commercial plants of this process installed to coal fired boilers
and FCC units. These plants were constructed by MMC in Japan and
Uhde GmbH in Germany.
MMC also developed a technology to produce activated coke, used in
dry DeSOx/DeNOx process, based on our own metallurgical coke
manufacturing technology.
This paper provides the information on the details of MMC's
activated coke used in the dry DeSOx/DeNOX process in the former
section and of the dry DeSOx/DeNOx process using activated coke in
the latter section.
8A-1
-------
MMC'S ACTIVATED COKE FOR DRY DESOx/DENOx PROCESS
Activated coke (AC) is a formed carbonaceous material designed for
dry DESOx/DENOx process of flue gas cleaning. For this purpose, AC
is able to remove SOx and NOx, has high mechanical strength against
abrasion and crush during circulation and handling in the process.
There is a stable supply of AC and the price is reasonable.
Research and Development of Activated Coke Production in MMC
Laboratory tests of AC have been conducted since 1980 and the
following information has been obtained. '
• Selection of suitable raw materials from bituminous
coals and the technology of their pretreatment. ( It was
recently found that lignites and petroleum coke are
suitable starting materials.)
• Blending ratio of pretreated starting material and
binders produce high mechanical strength in activated
coke. (Roga index was introduced to represent proper
blending ratio.)
• Briquetting process and briquett's size.
• Heating program and atmosphere for carbonization and
activation process ( A high mechanical strength of AC is
obtained with a controlled carbonization condition and a
proper micro-porous structure of AC is grown with a
controlled chemical activation condition.)
• Chemical treatment technologies of AC micro-porous
surface have increased the ability to remove SOx and NOx.
Based on the results of laboratory tests, in 1982 MMC constructed
and operated a pilot plant, which had a capacity of 0.5 tons/day AC
production.
Table 1 summarizes MMC's activities for AC production and supply in
chronologic order since the pilot test. AC produced at this pilot
plant was tested on DeSOX/DENOx pilot plant at Tochigi and the
demonstration plant at Ohmuta. Each of these extended more than
4,000 hours and 16,000 hours, respectively. This confirms the
excellent performance of MMC's AC on the desulfurization and
denitrification efficiencies and on the mechanical strength against
abrasion during circulation in large scale plants.
8A-2
-------
In October 1986, a commercial plant with the capacity of 3,000
tons/year AC production begem operation at MMC's Kitakyushu Works.
Since 1987, AC has been supplied to the DeSOx/DeNox unit at Idemitsu
Aichi Oil Refinery. MMC's AC is also supplied to the DeSox unit of
an iron ore sintering plant at Nippon Steel Co., Nagoya Works, as
well as several other customers.
Characterization of MMC's Activated Coke
Table 2 shows the characteristics of MMC's activated coke compared
to activated carbon using criteria such as gas recovering process
and deodorizing process. BET surface area of activated coke is less
than one third of activated carbon. BET surface area of these carbon
materials represent their micro-porous structure, which becomes
larger as chemical activation condition is severe in manufacturing.
As the activation becomes more severe the yield of the product
decreases and the mechanical strength of the product falls. A
decrease in yield results in increased product cost and a decrease
in mechanical strength causes greater material loss during the
process. As a result, MMC's activated coke processed with a
temperate activation condition is one fourth to one third in price
and has a high mechanical strength compeared with activated carbon.
Activated coke also has advantages in the abilities to remove SOx
and NOx as compeared with activated carbon or a metallic catalyst.
The following are some characteristic features of activated coke in
desulfurization and denitrification found in laboratory studies.
Desulfurization with Activated Coke.
Figure 1 shows S02 adsorption capacities of activated coke and
activated carbon with fresh one and used one. With fresh one, S02
adsorption capacity becomes higher as BET surface area of adsorbent
increases. On the other hand, with the used one, which has
experienced several cycles of S02 adsorption and thermal desorption,
the BET surface area of materials generally tends to increase. S02
adsorption capacity of the used one of activated carbon decreases
drastically, while those of activated coke, having a smaller BET
surface area and still less micro-porous than activated carbon, are
less influenced. The decrease of S02 adsorption capacity on
activated carbon may be caused by a chemical modification of its
micro-porous surface through S02 adsorption and desorption.
As SOz removal efficiency in DeSOx process is approximately
estimated with the S02 adsorption capacity, it is supposed that
activated coke keeps a stable DeSOx performance during the process
and it has been confirmed through pilot and demonstration tests
mentioned later in this paper.
8A-3
-------
Denitrification with Activated Coke
Catalytic activities of carbon catalysts (activated coke and
activated carbon) for denitrification do not always depend on their
BET surface area. It is also true that their activities sure raised
by the chemical modification of their micro-porous surface through
S02 adsorption and desorption, (called SOx treatment). With both
fresh and used one, (after SOx treatment), the catalytic activity of
activated coke is superior to that of activated carbon. It should be
noted that the chemical surface structure of carbon catalysts is a
dominant factor for controlling the rate of denitrification over the
physical micro-pore structure.
In the DeSOx/DeNOx process, the catalytic activity of activated coke
for denitrification is enhanced3' by SOx treatment joining with NH3
treatment forming oxides groups and nitrogen-species on its micro-
porous surface, which also has been confirmed with pilot and
demonstration tests as mentioned later in this paper.
Figure 2 shows the catalytic activities of activated coke and
vanadium/titania catalyst for denitrification (called
selective catalytic reduction: SCR) in each working temperature
range. The SCR reaction occurs on these two catalysts with the same
mechanism :
4NO + 4NH3 + 02 -> 4N2 + 6H20
The catalytic active sites of these catalysts for SCR reaction are
mainly surface oxides, which are bonded to carbon in activated coke
and is doubly bonded to metal in vanadium catalyst. It is
undesirable that the surface oxide on vanadium is also active for
oxidation of S02 to S03 at the working temperature for SCR ( about
350° C) because of forming ammonium bisulphate deposition at down
stream of flue gas flow. On the other hand, it is an advantage that
the surface oxide on activated coke works in a relatively lower
temperature range and it catches S03 effectively.
8A-4
-------
MITSUI—BF DRY SIMULTANEOUS DESOX/DENOX PROCESS DESCRIPTION
A schematic of the Mitsui-BF DeSOx/DeNOx process is provided in
Figure 3. This process consists of three sections; adsorption, AC
regeneration and by-product recovery.
Adsorption Section
Adsorption section consists of two stages in simultaneous
DeSox/DeNOx process.
Activated coke (AC) moves from top to bottom through adsorbers
continuously. First, AC enters in the top of the 2nd stage, where
NOx reduction, with addition of NH3 occurs. The discharged AC from
bottom of 2nd stage enters in top of 1st stage, where SOx adsorption
occurs. AC adsorbed SOx up to its designed capacity is discharged
from bottom of 1st stage, and sent to regeneration section by bucket
conveyer.
If only DeSOx ->r only DeNOx (in case of no SOx in flue gas) is
required, a one stage process can be designed.
SOx Removal bv Adsorption at 1st stage
Flue gas, ranging from 100 to 200°C , passes through the 1st stage.
During this stage, SOx (SOz and SO^) is removed mainly by
adsorption. AC acts as an adsorbent for SOx at this stage, SOx is
adsorbed and held as sulfuric acid (partially as ammonium salts) in
micro-porous structure of AC. Chemistry of SOx removal is listed in
Table 3-1.
NOx Removal bv Selective Catalytic Reduction at 2nd Stage.
When NH-, is added while S02 concentration is high, the NH3 is
consumed by SOx forming ammonium salts and effective denitrification
is not achieved. Hence, the optimum performance is achieved by
injection of NH3 at a location where less concentrated S02 exists.
This approach would provide a greater denitrification rate at a
lesser NH3 consumption.
Then, the flue gas, which has had almost all of the SOx removed at
the 1st stage, is introduced in the 2nd stage.
During the 2nd stage, mainly NOx (NO and NO?) is decomposed into N2
and H,0 catalytically with the addition of NH3 as a
reductant. Surface oxides and nitrogen-species in micro-porous
structure of AC act as the catalytic active sites for N0x-NH3~02
(SCR) reaction. Chemistry of NOx removal is listed in Table 3-2.
8A-5
-------
Removal of Dust bv Filtration at Adsorber.
Moving bed of AC acts as granular filter for removal of dust in flue
gas. It is the third function of the AC dry process. MMC designed a
special louver for effective removal and discharge of dust.
Removal of Trace Elements
Halogen compounds and trace elements such as mercury vapor and
dioxines contained in flue gas are also removable by chemical or
physical adsorption into AC.
AC Regeneration Section
AC discharged from 1st stage is sent to regeneration section, where
sulfuric acid and its ammonium salts adsorbed into AC are thermally
decomposed and generate S02 concentrated gas. Xt is called S02 rich
gas (SRG), which is sent to by-product section. Chemistry of
regeneration section is listed in Table 3-3.
After cooling, the regenerated AC is sieved out fine dusts (AC fine
powder generated by abrasion plus fly-ash in flue gas caught by AC)
through vibration screen, then recycled back to the adsorber.
The lost AC by mechanical abrasion on circulation in the
process(called mechanical loss, depends on AC moving velocity and
distance) and by chemical consumption of S02 adsorption and
desorption, (called chemical loss, depends on SOx load of AC) is
made up of fresh AC after this section.
Bv-Product Section
SRG, generated in AC regeneration section, contains approximately
20-25% S02. It can be converted into either elemental sulfur,
sulfuric acid or liquid S02.
ADVANTAGES OF MITSUI-BF PROCESS
Two process flows for flue gas desulfurization and denitrification
of a power plant are compared as an example in Figure 4. Mitsui-BF
process using activated coke is the upper one in the diagram. The
lower one is a combination of two seperate processes before and
after air preheater. These processes are ; a catalytic
denitrification process using vanadium/titania catalyst at about
350° C before air preheater and a wet scrubbing desulfurization
process using lime milk after air preheater.
Mitsui-BF process has the following advantages compared with the
lower one in Figure 4.
8A-6
-------
Simultaneous SOx/NOx Removal Process
It is one of the advantages of Mitsui-BF process that both
desulfurization and denitrification are performed in a single
process. Dedusting and removal of trace elements from flue gas are
also achieved additionally.
Drv Process
Contact of flue gas with activated coke bed under dry condition has
been able to simplify the process flow and save the installation
space. No large scale waste water treatment , no S03 mist-seperator,
and no reheater of treated flue gas are necessary in Mitsui-BF
process.
Low Temperature Process
High removal efficiencies of SOx and NOx from flue gas aire achieved
at temperature ranging from 100 to 200° C in Mitsui-BF process.
Removal efficiency for SOx becomes higher with lowering flue gas
temperature in this range. High removal efficiency of SO, is also an
advantage of this dry process. On the other hand, removal efficiency
for NOx becomes higher with raising flue gas temperature in this
range. These two relationships are opposite. Therefore, suitable gas
temperatures need to be designed for each flue gas condition and
expected removal efficiencies. Usually, flue gas after recovering
its waste heat through heat exchanger(just before stack) is suitable
for Mitsui-BF process.
Optional Bv-Product
Either elemental sulfur, sulfuric acid or liquid S02 can be made
optionally from SRG generating in AC regeneration section.
DEVELOPMENT OF MITSUI-BF PROCESS
Table 1 also summarizes MMC's activities for dry DeSOx/DeNOx process
development and plant construction in chronologic order.
Process Development
MMC concluded a licensing agreement with Bergbau Forschung (now
called Deutsche Montan Technologies:DMT) for the dry DeSOx/DeNOx
process in 1980.
8A-7
-------
I
Pilot Test
During the years 1981-1983, MMC operated a pilot plant of
DeSOx/DeNOx process treating 1,000 Nm /Hr flue gas at Tochigi,
Japan. ' Main purposes of this plant were as follows.
• To obtain basic performance data of this process such as:
(1) correlation of SOx removal efficiencies versus flue
gas temperature, space velocity, AC retension time
in adsorber and other factors.
(2) correlation of NOx removal efficiencies versus flue
gas temperature, space velocity, NH3 injection
volume, and other factors.
• To obtain working knowledge of the engineering and
cperation of this process.
• To improve original process aiming at minimization of
process running cost and maximization of process
performance.
• To confirm the performance of MMC's activated coke (AC)
for this process.
Demonstration Test
In 1984, MMC constructed a dry DESOx/DeNOx plant treating 30,000
Nm /Hr flue gas from coal-fired boiler of Mitsui Coal Mining Company
at Ohmuta, Japan. Two-stage adsorbers, in which MMC's activated coke
is filled up and circulating, performed 100% of SOx removal and over
80% of NOx removal efficiencies. Figure 5 shows the correlation of
SOz removal efficiency with ranges of 700-1250 ppm of inlet S02
concentration, when NH3 was not injected. This is one of the
parameter test runs done during the demonstration operation. Perfect
removal and 87% removal of S02 was achieved till approximately 1,000
ppm and at l,250ppm of S02 concentration, respectively, with
designed AC retention time for this demonstration plant. The removal
efficiency of S02 can be kept at a high level by controlling AC
retention time to treat higher concentration of SO, containing flue
gas.
Although NHj was not injected in this test run, 16-22 % of NOx
removal efficiency was achieved. This may occur by a direct reaction
of NOx and carbon and/or a reaction with nitrogen-species on the
microporous surface of AC, which seemed to be formed through SOx and
NH3 treatments of AC in the DeSOx/DeNOx process as mentioned in
former section.
This demonstration test operation has extended over 16,000 hours.
Through this long term operation, stability and reliablity of the
system has been confirmed.
8A-8
-------
Elemental sulfur was recovered from SRG in the process combined
reduction-reactor with metallurgical coke as reductant (S02 -> H2S)
and Claus-unit (2H2S + S02 -> 3S + 2H20).
During this test, activated coke circulating in the plant was
sampled periodically and analysed. Figure 6 shows the
characteristics of sampled AC in accordance with plant operation
time.
• BET surface area of sampled AC increased with operation
in the early period. It suggests that AC micro-porous
structure expanded with SOx adsorption and desorpiton
in turn in the process. The reaction of carbon to S03,-
which formed through thermal decomposition of H2S04 in
the AC regeneration section, is the chemistry for
surface area expansion.
• S02 adsorption capacity of sampled AC was almost
constant, though AC microporous structure was
expanding. It means that SOx removal efficiency
depends on not only micro-porous structure, but
also the chemical structure of AC in practice.
• Strength index of sampled AC against abrasion was almost
constant. It means that mechanical strength of AC
doesn't go down in practice.
• AC chemical structure, which are represented by O/C and
N/C also increased with operation in the early period.
It suggests that oxide groups and nitrogen species on
AC microporous surface were formed by contacting with
oxygen containing gas(SOx, NOx, 02) and nitrogen
containing gas (NH3, NOx) in the process.
In accordance with an increasing amount of oxides and
nitrogen-species on AC chemical structure, NOx removal
efficiency of AC increased as shown in Figure 7.
PROCESS APPLICATION
In 1987, Mitsui-BF DeSOx/DENOx commercial plant started operation at
Idemitsu Kosan, Aichi Oil Refinery. This plant has been treating
flue gas (236,000 Nm3/Hr designed) from catalyst regeneration
section of Residue Fluid Catalytic Cracking Unit (RFCC). Performance
of this DeSOx/DeNOx plant has been very successful. 100% of SOx (S02
+ S03) removal efficiency and over 80% of NOx removal efficiency has
been achieved constantly at approximately 180° C. This plant is easy
to operate and experienced almost no trouble.
In 1990, MMC constructed AC DENOx pilot plant, which has been
commissioned to EPDC by the Japanese government. This pilot plant
was designed for treating 10,000 Nm /Hr flue gas at 140° C from &
fluidized bed combustion boiler.
8A-9
-------
In Germany, Chde GmbH, which is also a licensee of BF, constructed
two dry DeSOx/DeNOx commercial plants. One started operating in 1987
for treating flue gas (totally 1,110,000 Nm /Hr designed) from
lignite fired power plant at EVO's Arzberg power station- Another
began operation in 1989 for treating flue gas (323,000 Nm3/Hr
designed) from hard coal fired boiler of Hoechst AG at Fraiikfurt.
MMC cooperated with Uhde in the engineering of these two plants.
Mitsui-BF process is applicable to a broad range of flue gas
cleaning as follows:
- boilers
- furnaces for sintering and heating
- incinerators for trash and refuse
- regenerator of catalysts such as FCC
- chemical plants such as sulfuric acid plant and
others.
MITSUI-BF PROCESS ECONOMICS
Simultaneous DeSOx/DeNOx Process
Capital cost : US$ 220 - 240/KW
This is estimated based on the following conditions :
• 500 MW (250 MW x 2 units, 25% allowence for our system
included)
• S02 2000 ppm, NOx 326 ppm
• DeSOx efficiency : more than 90%
• DeNOx efficiency : more than 80%
• Civil, foundation works are excluded
• Including by-product equipment (elemental sulfur or sulfuric
acid)
DeSOx Only Process
Capital cost: US$ 140 - 160/KW
This is estimated based on the same conditions as above except DeNOx
efficiency.
8A-10
-------
DeNOx Only Process
Capital cost : US$ 70 -80/KW
This is estimated based on conditions as follows.
• 350 HW
• NOx 250 ppm, SOx less than 50 ppm
• DeNOx efficiency: more than 80%
• Civil, foundation works are excluded
Remarks : If the flue gas has no SOx, no dust like LNG boiler, we
can introduce fixed bed AC for DeNOx. Thus, our capital cost might
decrease further.
CONCLUSION
Mitsui-BF dry DeSOx/DeNOx technology, it's history of research,
development and commercialization has been provided in this paper
with both aspects of the activated coke used in this process and the
DeSOx/DeNOx process using activated coke. There are four commercial
installations of this process for cleaning flue gas from coal fired
boilers and oil refinery in Japan and Germany. Air, water and land
are essential for life, but they are no longer infinite resources.
We believe that the Mitsui-BF system is an important contribution to
emmission control technology and an effective method to keep our
atmosphere clean.
REFERENCES
1. K. Knoblauch, E. Richter and H. Juntgen. n Application of active
coke in processes of SOx and NOx removal from flue gases. n Fuel.
September, Vol. 60, 1981, p. 832.
2. Y. Komatsubara, M. Yano, I. Shiraishi and S. Ida. n Preparation
of active coke for the removal of SOx and NOx in the flue gases."
In Proceedings of 16th Biennial Conference on Carbon. 1983,
p. 325.
3. I. Mochida, M.Ogaki, H. Fujitsu, Y. Komatsubara and S. Ida.
" Catalytic Activity of coke activated with sulfuric acid for the
reduction of nitric oxide." Fuel. July 1983, Volume 62, p. 867.
4. S.M. Dalton. " Current status of dry NOx-SOx emission control
processes." In Proceedings of the 1982 Joint Symposium on
Stationary CoTnhngt-i on NOx Control , 1982, p. 32-1.
8A-11
-------
Fresh sample
Used Sasple
Activated
Coke
A O
~ •
Activated
Carbon
o
~
» 200
r*
O
in
cr
e
>s
4J
u
c
a
c
u
100
c
o
•H
iJ
a
u
o
n
•o
<
o
w
500
BET Surface Area (m?/g )
1000
Figure 1 The SO2 adsorption capacities of
Activated coke and Activated carbon
100
80
>s
U
c
o
•H
u
•^4
60
Metal—oxide Catalyst
c
>
o
e
9
oc
Activated Coke
20
2
50
100
150
350
400
250
200
300
SCR Reaction Temperature (°C)
Figure 2 The catalytic activities of Activated cch*
and Metal—oxide catalyst for SCR reaction
8A-12
-------
To
A.C.
Fan
Ak
I
* Metallurgical coke breeze was used
in the coke—reecter(Sulfer recovery)
Figure 3 Mitsui-BF dry DESOx/DENOx process with Sulfer recovery
Olitsui-BF Dry I Simultaneous DESOx/DEWOk Process]
' U~t33J [
V
Boiler
Hot ZP
Air
preheater
IOF
Nitsui-BF
PROCESS
' Simultaneous DESOx/D&Ox
¦ Dry Treatment
• After Air Heater
(Just before Stack)
Stack
[Combination of llith-tesperature DEKOx and Vet DESOx)
IDF BUF
—®-n—SH
SCR
V
Boiler
Hot EP at 3S0T:
Ketal Oxide Air
Catalyst preheater
Gas/Gas
Heater
Vet
FGD
Stack
Figure 4 Advantages of Mitsui— BF process for flue gas DESOx/DENOx
¦ Electrostatic precipitator(EP) also can be located after oir-preheater.
• Wet EP also can be additionally arranged after Wet FCD to remove aista.
8A-13
-------
[Test condition]
SO 2 : Parameter
NOx : 120-130ppm
NHj : not injected
Gas Temperature: 145—150*C
AC Retention time: Constant
o
a
ti
0c
X
o
z
*10
Figure 7 Correlation of NOx removal ability of AC
with elemental ratio of AC surface
8A-14
-------
Table 1
StfrMARY OP MMC's ACTIVITIES TOR AC PRODUCTION AND DRY DESOx/DENOx PROCESS
Year AC Production and Supply
1982 • Pilot plant(0.5Ton/Day)
• AC supply(lOTon) for Tochigi
1984 • AC supply for Ohrauta
Mitsui—Br Dry DESOx/DENOx Process
DESOx/DENOx pilot plant (1,OOONmJ/Hr)
test operation at Tochigi
[from Feb.1981 to Oct.1983J
DESOx/DENOx plant(30,OOONm'/Hr)starts to
operate at Mitsui Coal Mining,Ohmuta
1985 • AC supply for Matsushiraa
power station of EPDC
1986 • Commercial AC production
(3.OOOTon/Year)start to
operate at Kitakyusyu
1987 • Start to supply of AC for
Idemitsu-Aichi Refinery
• DESOx/DENOx plant(236.000Nn3/Hr)3torts
to operate at Idewitsu—Aichi Refinery
¦ DESOx/DENOx plant( 1,HO.OOONra^HrJatarta
to operate at Arzberg power station of
Eyo,Germany *1)
1988 • Start to supply of AC for
Nippon Steel—Nagoya Works
1989 • AC supply for Hoechst,Cermany • DESOx/DENOx plant(323,OOONra3/Hr)starts
to operate at Hoechst,Frankfurt,Germany
«1)
1990 • AC supply for Wakamotsu • Lovr-Teinperature DENOx pilot plant
Research Center of EPDC (10.000Nra3/Hr) at Wakaaatsu of EPDC «2)
#1) These plants were constructed by Uhde GmbH.Cermany.
MMC cooperated in enginieering.
#2) Commissioned to Electric Power Development Company(EPDC)
by Japanese government
8A-15
-------
Table 2
CHARACTERISTICS OF MMC'S ACTIVATED COKE COMPARED WITH ACTIVATED CARBON
Items MIC's Activated Coke
BET Surface Area (mJ/g) «D 150 - 250
Activated Carbon
850
Mechanical Strength (%) #2)
95
85
SO? Adsorption Capacity S3)
(mg~S0j/9)
60 - 120
220(with fresh one)
70(with used one)
NOx Removal Efficiency 94)
(*)
Price Patio (-)
80 - 85
1/4 - 1/3
60 - 70
01) Measured by CO? adsorption at 199K
•2) Defined as the yield of powder after 1,000 revolution in the drum tester
*3) SO? amount adsorbed under following conditions:
Adsorption : 100 "C * 3Hr
Contacting gas composition: S0j=2%, 0?=5%, H^O^lOX, N?=Balance
*4) Measured with fixed bed flow reacter under following conditions:
Samples amount:300cc, Temperature: 140"C, Space velocity: 400Hr~'
Contacting gas composition: NO=NH3=200PPM, 02=5%, H20=10X, N2~Balance
Table 3-1
CHEMISTRY OF SOx REMOVAL IN ADSORPTION SECTION
[Desulfurization into micro-porous structure of AC without NHj
: mainly proceeds in 1st—stage adsorber]
1/2 0» - H, SO. Cad.) Dissolution of SOj
<99.) * (Formation of sulfuric acid)
[Desulfurization into micro-porous structure of AC with NH3
: mainly proceeds in 2nd—stage adsorber]
SO, (ad. ) + H, O (ad.) - H, SO. Cad.) Formation of sulfuric acid
N H , Cs. ) + H , S 0 . (ad. ) — N H . H S O . Cad. ) Formation of anmonium bisulfate
N H , <(j. J + N H . H S O. (ad. ) — ( N H . ) i SO. (ad. ) Formation of ammonium sulfate
2NH, (g ) + H , SO, (ad.) — (Nil. ). SO. (ad. ) Formation of anunorium sulfate
»(,d.)radsorbcd atate •((. ):gas phase
8A-16
-------
Table 3-2
CHEMSTRY OF NOx REMOVAL IN ADSORPTION SECTION
(NOx Selective Catalytic Reduction with NHj on AC: proceeds in 2nd—stage]
1/4 0» (g.) — 1/2 0(ad. ) Dissociative chemisorption
NO(j. ) 0(ad. ) — N Ot (ad. ) Oxidative chemisorption of NO
NH> (k.) + OH (ad.) — N H . (ad. ) + O (ad. 3 Chemisorption of NH3
NO» (ad.) + NH. (ad.) — Nil. NOi (ad.) Surface reaction of
adsorbed species
NH. NO« (od. ) + 1/2 0(ad. ) Desorption of decoaposed produCj
— N, (*. ) + 3/2 Hi 0(s. ) + OH (ad.) from AC micro—porous surface
* NHaNO? doesn't identified.
NO + N H , *1/4 0« — N 1 * 3/2 H, 0 Total SCR reaction
(NOx reduction with nitrogen-species on micro-porous surface of AC]
HN-C (Surface) + NOi (ad. ) — N , (g.)+OH(id.) + 0 - C (Surface)
» Nitrogen—species doesn't identified.
(NOx direct reduction by carbon of AC]
C + NO. (ad.) - 1/2 N,
-------
-------
HIGH EFFICIENCY, DRY FLUE GAS SOx, AND COMBINED SOx/NOx REMOVAL
EXPERIENCE WITH THE LURGI CIRCULATING FLUID BED DRY SCRUBBER-
A NEW, ECONOMICAL RETROFIT OPTION FOR VS. UTILITIES
FOR ACID RAIN REMEDIATION
J.G. Toher
Environmental Elements Corporation
3700 Koppers Street
Baltimore, Maryland 21227
G.D. Lanois
Environmental Elements Corporation
3700 Koppers Street
Baltimore, Maryland 21227
Harald Sauer
Lurgi GmbH
Frankfurt am Main, Germany
Preceding page blank
8A-19
-------
8A-20
-------
ABSTRACT
The Lurgi dry flue gas desulfurization Circulating Fluid Bed (CFB) process has been in com-
mercial operation on five coal-fired utility boilers in Gennany since 1987. The process has
consistently demonstrated SOz removal efficiencies up to 97% since its introduction in 1984.
Several U.S. utilities are now seriously considering the CFB in their Phase I acid rain remedia-
tion planning.
Currently, the CFB is being further developed for combined SOx/NOx removal. Pilot study
results indicate removal efficiencies of 95%/85%, respectively.
The presentation covers:
• Operating data and experience from five plants
• Process design data
• Performance data (Le., 97% SOz removal on 6% sulfur coal)
• Economics showing favorable capital and O&M costs compared to wet flue gas desulfu-
rization processes and spray dryers
• Retrofit strategies with little or no footprint which provide options for plants with limited
space availability
• Description of advanced, combined SOx/NOx CFB process and performance results
Preceding page blank gA 21
-------
THE CFB DRY FGD TECHNOLOGY CONCEPT
Process Overview
The circulating fluid bed FGD process provides a viable, proven alternative to wet limestone
scrubbing (LSFO). The process uses a dry reagent and precludes the liquid phase absorption
mechanism used in LSFO and in lime spray dryer absorbers (LSD). The system is less compli-
cated to operate and easier to maintain because it does not require high maintenance mechanical
equipment such as grinding mills, abrasion resistant slurry pumps, agitators, rotary atomizers,
and sludge dewatering devices. The CFB's reagent flow is independent from water balance
around the reactor vessel.
In general, process operation begins with calcium oxide (pebble lime) hydrated on site and
injected dry into the flue gas on the cold side of the air preheater. A fluid bed of lime develops
in the reactor, providing the contact medium between gaseous sulfur oxides and the hydrated
lime. Dry recirculation of material from the downstream particulate collector is used to opti-
mize fresh lime consumption.
Reactor Operation
Gas enters the scrubber vessel at
the bottom and flows vertically
upward through a venturi
section. The venturi is designed
to achieve the proper flow
distribution throughout the
operating range of the vessel.
Inside the venturi, the gas is
first accelerated, then deceler-
ated before entering the upper
cylindrical vessel.
The upper vessel's height is
designed to accommodate the
mass of bed material required
for the desired Ca and S contact
time. The vessel is designed
with an internal gas velocity
range of 6 to 20 feet per second.
This range of gas velocities
supports boiler loads from 30%
to 100%.
All external inputs of recirculating material, fresh reagent and gas conditioning water are intro-
duced to the gas on the diverging wail of the venturi. The vessel has no internal mechanical or
structural components. When operating at a full load design pressure loss of 5 to 6 inches W.G.,
the gas residence time in the vessel is approximately 3 seconds.
Ash and
Absorption
Products to
Precipitator
Recirculated Ash
and Absorption
Products from
Precipitator
HumldHlcatton
Water Injection
Nozzles
Flue Gas from
Air Pieheater
from
Bed Removal from
Shutdown to Ash SOo
Figure 1. Details of CFB Reactor
8A-22
-------
CFB PROCESS DESIGN
Process
Gas from the boiler air preheater passes through the scrubber vessel, then through a particulate
control device before being exhausted via a fan to the suck. The gas path material is carbon
steel. .The existing stack can be re-used, a feature usually possible with any "dry" technology.
Flue gas entering the scrubber is evaporatively cooled to within 45*F of the adiabatic saturation
temperature with a water spray. This process is relatively insensitive to the chemistry of the
cooling water, therefore on-site waste waters may be consumed in the CFB.
The cooled gas passes up through the circulating bed of fresh reagent and recirculated material.
Abrasion between particles in the fluid bed continuously removes the outer layer of absorption
products and exposes the underlying surfaces of unused lime. The continuous contact between
panicles, combined with the evaporative cooling of the gases, optimize the overall consumption
of fresh lime.
The gases are cleaned of dry dust in the downstream particulate collector. As much as 98% of
the material collected is recirculated to the CFB to resupply the bed. This high recirculation
ratio keeps unused calcium in the process for up to 1/2 hour and boosts the performance capabil-
ity level up to 99% total sulfur capture. The vast particle surface area in the circulating bed
permits successful capture of virtually all of the S03 in the gas, eliminating the possibility of gas
path corrosion from condensed SO} aerosol mist. The disposal material from the process is
moved to an ash silo for discharge and eventually transported to a landfill or mine.
I—I
WATCH
TANK
NEW I OLD I
-© (>J •
I 1
» I
I I
FAN I STACK!
WATER
COOUNG,
WATER
-------
Control System
The simplicity of the process and limited number of control loops allow the system to be oper-
ated either by the existing boiler control room DCS system or by a separate PLC based architec-
ture. Only three continuous control parameters are used in the process:
1. The gas temperature exiting the CFB vessel is maintained by modulating the flow of
water returning from the high pressure nozzles. Opening the return flow control valve
reduces pressure at the nozzle tip, thus reducing the flow of water through the nozzle tip
and into the gas.
2. The stack SOz gas concentration is maintained by modulating the flow of fresh lime into
the CFB lime transport line. Rapid response to changes in stack S02 are possible due to
close coupling of the CFB lime transport system to the CFB reactor.
3. The mass of the fluid bed is maintained by modulating the disposal of material in accor-
dance with a pre-set differential pressure across the scrubber vesseL The pressure loss
associated with fluidizing and suspending the mass of the bed will remain constant
throughout the range of boiler load. This parameter is therefore used to regulate the
discharge of material and to insure proper flow of recirculated material to replenish the
bed-
All other control logic is discrete (yes/no) signals or typical particulate control logic.
Hytfrtfkx
HumidMcatlon
for Tomperoturs
Control
Gaslntet
otMO'F
C)«an
Gas
Outtetat
170'F
• Absorbw Praaui* Drop Controta Aiti Recirculation
• Gas Temperature Controls HurrMlflcatton Wat Of How Bat*
• Stack SO. Concentration Controls Fresh Lima Addition
ToDfeposal
Figure 3. CFB Process Control
During start-up of the boiler, the circulating bed of material is established at 30% boiler load.
Either the inventory of material in the surge bin or fresh lime can be used; the choice depends
upon how long the system has been shutdown. When a preset pressure differential across the
fluid bed is reached, a signal releases an interlock, allowing the flow of cooling water for
humidification. The process is now fully functional and automatically follows the boiler up to
full load.
8A-24
-------
PERFORMANCE
Operation and Maintenance
Figure 4 presents an overview of CFB scrubber operating experience in the power, process steam,
and cement industries.
Rated
Capacity
MW
Operational
Time
(YRS)
Capacity
Factor
(%)
Coal
Sulfur
(*)
Removal
Efficiency
(%)
Stolchiometry
as tested
Schwandorff
50
1 +
50
1.5
95
1.3
Borken
100
3+
<50
6.0
97
1.5
Siersdorf
2@ 85
3+
70
1.0
93
1.2
Utervaz
50*
3+
80
1.S*
90
1.2
Opel
50*
1
70
1.0
92
1.3
'Equivalent
Figure 4: CFB Installations
The experience base includes fuels with low to very high sulfur contents. Plant operations cover
both peaking stations and base loaded high capacity stations. Current installations are demonstrat-
ing system availability of 99% or higher. Operating and maintenance costs have been in line with
original design predictions.
A 1991 survey of power plant operations indicates that the CFB does not require a full-time
person to operate and maintain it. Two of the facilities surveyed relied upon their existing staff to
operate the newly installed CFB.
Waste Characterization
The dry waste product from this process has characteristics similar to LSD waste, therefore, it can
be transported and disposed of in the same manner. The product can be fixated with an optimum
amount of water and will achieve yield strengths in excess of 2000 psi with permeability equiva-
lent to clay (10"10 ft/sec). The material has compacted densities in the range of 80 lb/ft3.
Laboratory analysis of the material produced at Schwandorff show the following typical values
after mixing with 30% water:
• Density: 80 lb/ft3
• Compressive Strength: 2,600 psi @ 56 days
• Permeability: 12.2 x 10",0ft/sec @ 28 days
This product's ability to achieve lean concrete strength and clay like permeability make it a good
source of material to stabilize spent mines or reclaim open quarries and landfills.
8A-25
-------
FGD PROCESS ECONOMICS COMPARISON
This economic comparison evaluates current (1991) capital and O&M costs for dry Circulating
Fluid Bed (CFB), Wet Limestone Scrubbing with Forced Oxidation (LSFO), and Lime Spray
Drying (LSD) processes.
To make a fair and objective comparison of these three technologies, an established data base
was used. The format and methodologies of the recent EPRI FGD Economics Studies were
selected. Highlights of EPRI evaluation criteria are as follows:
Technical
• 300 MW PC Boiler
• 2.6% coal @ 13,100 b/# HHV
• 65% capacity factor
• 90% S02 removal
1.276 retrofit factor
• (3) 50% absorbers
• Not included
- Particulate removal equipment
- Stack modifications
- Boiler modifications
Economic
1/90 dollars
15 yr. plant life
• Labor $20/hr.
• Reagent
- Lime $55/ton
- Limestone $15/ton
• Disposal $8.15/ton
• 5% annual inflation
• 0.3% annual inflation (power & steam)
• Levelized constant/current dollar factors, and process/project contingencies all as per
1991 EPRI FGD report
8A-26
-------
Certain adjustments to the EPRI report criteria and cost assumptions have been made to develop
a comparison that reflects FGD process economics that are currently being considered for actual
Phase I Acid Rain remediation projects. These adjustments are tabulated along with explanatory
comments as follows:
CFB
LSFO
LSD
Number of Absorber
Modules
1
1
1
Total Process Capital
70 S/KW
124 S/KW
80 S/KW
Particulate Equipment
ESP added
(None)"
ESP added
Ca/S Ratio
(1.15T
(LI)*
125
Operating Labor
One Man Year
(4-5 man years)*
(3-4 man years)*
Maintenance Labor
(4.8% TPC)*
(4.4% TPC)*
Solids Disposal Cost
Ryasfa excluded
(Flyaii excluded)*
Flyash excluded
* Items in parentheses denote no change from the EPRI study
Figure 5. Adjustments to EPRI Study
Nntre to Figure 5
• Absorber Modules/Total Process Capital - Single 100% capacity modules are consid-
ered, thus the total process capital (IPC) was adjusted for CFB and LSD based on cur-
rent estimates. The LSFO TPC was adjusted based on the stated EPRI report cost reduc-
tion.
• Particulate Equipment - The EPRI study assumes there is existing particulate equip-
ment in place and no new particulate additions are needed. Current experience indicates
additional particulate equipment is usually needed for the CFB and LSD. A new full size
ESP is now included in the TPC at approximately 20-25 S/KW.
• Ca/S Ratio - The LSD Ca/S Ratio is set at 1.25 and thus the CFB and LSD are approxi-
mately in the same range of +/-10%.
- O&M Labor - Actual costs arc included based on 1991 studies of data obtained from
three operating CFB systems at utility plants.
¦ Solids - Assuming particulate equipment is in place and operational, the waste disposal
costs for the CFB and LSD are subtracted £nom the waste volume stated in the EPRI
study.
8A-27
-------
Figure 6 displays fixed and variable operating costs, fixed chaises, and total cost comparisons
for CFB, LSFO, and LSD. Levelized current dollars and levslized constant dollars are calcu-
lated based upon a IS year life cycle.
LovsKzsd Constant
Lavettzad Currant
First Y«ar Cost
Don arm
Dollan
cm LSFO L3Q
CFB LSFO LSO
CFB LSFO LSD
Fixed O & M Costs
$
(LSI S £05 S 1.59
$ 051 % 2J05 $ 1.59
i
0.70 $ 231 S 2.18
&llMriak 4Support)
Variable Opmtlng Costs
133 2.11 1.99
1.94 2.14 243
234 233 2.71
(FWsgent SaBd* DUp. Wafer 4 Pomm
D
FlxadChargos
4.96 7.81 £57
231 442 3.15
338 6.11 4.35
(DatoVEqulty flattam, Inccww Tom.
DipiiL. Ptvimtti Tax a Kwuianei)
Total Costs
S
7.40 $11.97 S 9.15
S 5J8 S 831 $ 6.77
s
7.22 $1135 S 924
Comparison (MHJLS/KWH)
Figure 6. FGD Process Economics
Economics Conclusions
• CFB fixed O&M costs are significantly less than LSFO and LSD due the system's sim-
plicity (minimum components) and very low O&M labor requirements.
• Variable operating costs for both dry systems (CFB and LSD) are comparable to LSFO
at 65% capacity factor and are sensitive as capacity factor changes.
• The fixed charge for the CFB (including particulate equipment) is the lowest of the
systems compared.
• While the LSFO capital cost could vary downward, additional costs not included but
typically required for LSFO (new stack, reheat, etc.) would tend to increase the fixed
charge component.
• While LSD is included in this study, it is generally not considered to be a practical option
for acid rain retrofits due to its significant plant space requirements.
• Total O&M and fixed charge costs indicate that the CFB is a viable economic choice for
FGD retrofit applications when compared to LSFO, even allowing for significant reduc-
tion in LSFO capital costs and/or capacity factor increases.
8A-28
-------
CFB OPERATING HISTORY
The CFB dry scrubbing technology emerged from fluosolids processes used in the 1950s and
'60s for calcining, roasting and coal burning. During the 1970s, CFB dry scrubbers were in-
stalled to control HF emissions from aluminum potlines, typically handling multi-million ACFM
at each installation. In the early 1980s the CFB was installed on incinerators for controlling HF,
HO, and SOr
The first CFB installation on a coal-fired boiler was in 1984 on a SO MW demonstration plant.
CFB systems were installed for commercial operation on utility power plants in 1987, 1988 and
1990. The most recent CFB system is planned for stan-up in 1993. Additional proposed projects
are expected to be finalized in the next several months. The number of developing CFB projects
suggests that the European and American power generating industry now regard the CFB as a
viable system for efficient FGD.
Current operating CFB installations are:
-ssaSfc Gas Volume (ACFM): 3x111,300
Figure 7
Company: VAW Bonn
Location: Schwandorff, Germany
Stan-Up: 1982
Application: Incinerator
Boiler Type: Traveling Grate
Fuel: Municipal Waste
Inlet S02 (PPM): 175
S02 Removal Efficiency: 60%
Particulate Oudet (GR/SCF): .031
8A-29
-------
Figure 8
Company: Baycrnwcrk
Location: SchwandorfF, Germany
Start-Up: 1984
Application: Utility Power
Capacity: 50 MW
Boiler Type: PC
% Sulfur: 2-3.5
% Ash: 8-10
Fuel: Lignite
Inlet SOz (PPM): 1,490
S02 Removal Efficiency: 95%
Particulate Outlet (GR/SCF): .062
Gas Volume (ACFM): 231,400
Figure 9
Company: Prussia Electric
Location: Borken, Germany
Start-Up: 1987
Application: Utility Power
^ Capacity: 100 MW
^ Boiler Type: PC
% Sulfur: 3-6.5
% Ash: 20-24
Fuel: Lignite
Met SO, (PPM): 4,450
, S02 Removal Efficiency: 97%
Particulate Outlet (GR/SCF): .033
? Gas Volume (ACFM): 585,600
8A-30
-------
Figure 10
Company: EBV
Location: Siersdorf, Germany
Stait-Up: 1988
Application: Utility Power
Capacity: 2 x 85 MW
Boiler Type: PC
% Sulfur. .8-1.6
% Ash: 23-30
Fuel: Bituminus Coal
Inlet S02 (PPM): 945
S02 Removal Efficiency: 93%
Paniculate Outlet (GR/SCF): .02
Gas Volume (ACFM): 2 x 315,000
Figure 11
Company: Buendner Cement
Location: Untervaz, Switzerland
Start-Up: 1988
Application: Raw Mill
Fuel: Biniminus Coal
Inlet S02 (PPM): 1,260
S02 Removal Efficiency: 90%
Paniculate Outlet (GR/SCF): .02
Gas Volume (ACFM): 236,000
Note: Sulfur source is from raw material
Figure 12
Company: Adam Opel AG
Location: Russelsheim, Germany
Start-Up: 1990
Application: Industrial Power
Capacity: 33 MW
Boiler Type: PC
% Sulfur .6-1.3
% Ash: 6-8
Fuel: Bionr I.ns Coal
Inlet S02(PFM): 945
S02 Removal Efficiency: 92%
Particulate Outlet (GR/SCF): .012
Gas Volume (ACFM): 2 x 82,300
RA-.TI
-------
CFB RETROFIT OPTIONS
The CFB offers several plant layout options and can be installed on plants with limited space.
Typically, these plants cannot be retrofit with LSFO or LSD systems.
In addition to its adaptable layout plans, the CFB conserves plant space because its reactor is
compact and close-coupled to the precipitator. The system frequently uses the existing suck,
and its support equipment can be conveniendy located in available plant space.
i-. nations where existing paniculate equipment requires upgrading, the CFB is more economi-
cs -"an LSFO systems because the CFB includes new particulate control equipment as a integral
component In such cases, th-, CFB requires less plant space.
Options for CFB arrangements are as follows:
II
New
Stack
Figure 13. Replacement of Existing ESP with CFB and New ESP
II
Additional
Electrostatic
Precipitator
RekU
Of required)
Boner
Figure 14. CFB Installed Upstream of Existing ESP
8A-32
I
-------
BoOer
Existing
Bectrostoflc
PredpMor
v w
New
Electrostatic
Predptoor
V V V
Figure 15. CFB Installed before Existing Stack and Downstream of Existing ESP
BoOer
Existing
Electrostatic
Precipitator
N/W
Stack
n n n n
New
Bectrostatic
Precipitator
VW
Figure 16. CFB Installed After Existing Stack Downstream of Existing ESP
Existing
Budding
ri n n_
New
Bectrostedic
Precipitator
vw
H n n_
BoBer
Existing
Dectroriattc
PredplWor
V w
SL
Stack
Figure 17. CFB Installed and New ESP Elevated above
Existing Boiler and ESP
8A-33
-------
COMBINED DeSOx/DeNOx PROCESS
A combined DeSOx/DeNOx process which uses one circulating fluid bed is currently being
developed. This process can also be used as a single process for either desulfuiization or NOx
reduction.
In the combined process, the CFB reactor is located upstream of the boiler and operates at
approximately 725*F. Hydrated lime is used as the sorbent for sulfur oxides and water is not
required for this process. The absorption products are mainly CaS04 (anhydrate) and approxi-
mately 10% CaSO,. This is a result of S02 oxidation during the NOx reduction. Usually, this
side reaction would be minimized in DeNOx systems because it creates corrosion problems.
However, in the combined DeSOx/DeNOx process, this side reaction is responsible for the
desulfurization process.
The DeNOx reaction is a selective catalytic reduction which uses ammonia as the reducing agent
and a catalyst. The catalyst is a fine powder of the active compound, FeS04 x 7 HjO, without a
supporting carrier. This catalyst was chosen based upon investigations conducted in the labora-
tory and at a pilot plant.
Selective Catalyst Reduction
4NH,+ 4N0+02£^ 4N2 + 6H20
4NH, + 2 NO, + O, 3N, + 6H,0
Catalyst
Support
Active Compounds
——lmm
7-U-l
SV-557mVm*
to2
V202
MoO,
wos
—10|im
• • •
S»-314.000»n»/m»
FeO/FeSO,
MnO/MnSQ4
Figure 18. Catalyst Comparison
8A-34
-------
initial laboratory tests of the combined process were done at the University of Karlsruhe and at
Chemtall, a subsidiary of Metallgesellschaft AG and Lurgi AG (2,3,4). Figure 19 shows the
results obtained from the laboratory CFB. This CFB used a catalyst with 03mm Si02 carrier
particles at a space velocity of 20,000 h*1.
F*«d Conditions
NO 510 vpm
0.9 "
0.8 "
510 vpm
550 vpm
7 vol T.
® 0.6 ~
1 1 1 1—
350 400
Temperature "C
Spoc* V«toctty. 20.000 h-'
Catalyst: Oimiti SlOjiphwas
FsSO^/MnSOA
Sotbcnh SinnCatOH),
partidw
Figure 19. Test Results from a Laboratory CFB
Performing Simultaneous S02andNOx Removal
The pilot CFB system was built at the Rheinisch Westfalische Elektrizitatsweike power station in
Dettingen, Germany. Initially, the catalyst used here had a silica or alumina support of 0.5mm
(0.0197 inch) and the active compound on the surface of the support. However, during the
plant's eighteen month operating period, the active compound without a supporting carrier
proved to be the most effective and economical catalyst.
The results from the pilot plant show that fine grained particles offer an advantage over honey-
comb catalysts because the particles provide a larger physical surface area. Very thin honey-
comb walls with a thickness of 1mm and a pitch of 6mm provide a catalyst surface area of
557 m2/m3. At an average particle diameter of 10 microns, the fine powdered catalyst obtains a
physical surface area of more than 300,000 m2/m3. If comparable course panicles of 50 microns
are used, the physical surface area is still 68,000 m2/m3.
8A-35
-------
Catalyst
Figure 20. Flow Diagram of DeSOx/DeNOx Pilot Plant
Figure 20 shows the simplified flow diagram of the pilot plant. The absorber is 600 mm (1.96
ft.) in diameter and 6 m (19.7 ft.) in height. Downstream of the absorber is a two-field electro-
static precipitator for dust collection.
SOj concentrations at the pilot plant varied between 450-630 ppm. NOx concentrations varied
between 170 - 320 ppm (calculated as NOz). At a Ca/S-mole ratio of 1.6 to 1.8,97% desulfuriza-
tion could be achieved. A NOx-reduction rate up to 88% was achieved at a mole ratio of
NHj:NO = 0.7 to 1. At higher NO raw gas concentrations up to 750 ppm, a DeNOx cleaning
efficiency of 94% was obtained.
NjO could not be detected in the clean gas from the combined DeSOx/DeNOx process. Ammo-
nia slip values were less than 5 ppm.
Tests conducted at this plant proved that the fly ash loading of the flue gas does not obstruct
NOx-reduction or desulfurizarion.
8A-36
-------
In summary, the features of the CFB process are:
for DcNOx;
• the NO^ removal is higher than 90% at 400°C
• a small particle sized caralyst with simple geometry,
• simple components,
• renewable during operation,
• environmentally acceptable (Fe/Mn),
• constant activity,
• resistant against -S02
• S02- oxidation is a positive additional effect,
• easy handling,
• and lower N20-formation
for desulfurization:
• the S02-removal is higher than 95% at 400°C,
• a small panicle sized CaOHj sorbent,
• no water injection,
• total SOj-removal,
• and no interference between DeNOx and DeSOx reactions.
8A-37
-------
' J V!
r /
V///X
Figure 21. Plant equipped with continuou
measuring and recording of raw and clear
gas concentrations of NO, S02, CO, 02, t)
NHj-slip and continuous feed control for
theNHj.
Figure 22. Details of all the additional measuring and control equipment.
8A-38
-------
REFERENCES
1. H_ Saucr, J JX Riley, GJHaug. "Operating Experience with a Dry FGD-Planr using a
Circulating Fluid Bed at the Lignite-Fired Power Station of PREAG in Borken, F.R.G."
Paper presented at the First Combined FGD and Dry SOz Control Symposium in Sl
Louis. MO/USA .October 25 -28,1988
2. W. Weisweiler. "Umweltfreundliche Entstickungs-Katalysatoren auf Basis Eisen/Mangan
Oxid/Sulfat DECHEMA-Jahiestagung 1989 Frankfurt/Main June 01 - 02, 1989
3. H_ Sauer, W.Weisweiler, JJD.Riley. "Simultaneous SOz- and NOx-Removal in the Circu-
lating Fluid Bed" Paper presented at The First International Power Technology Confer-
ence, Chicago, HI, October 31 November 2,1989
4. W.Weisweiler, E. Herrman, J. Zimmer, H.Sauer. "Der ziiiculicrende Wirbelschicht
Reaktor zur simultaneous trockenen Reinigung von NOi and SOs-haltigen
Feuerungabgasen" Proceedings GVC/VDI-Jahrestreffen
5. RJ. Keeth, PA. Ireland, P.T. Radcliffe."1990 Update of FGD Economic Evaluations."
6. EPRI Repon 1991, Project 1610-6, "Economic Evaluations of Flue Gas Desulfurization
Systems."
8A-39
-------
8 A-40
-------
Incorporating Full-Scale
Experience Into Advanced
Limestone Wet FGD Designs
P. C. Rader
E. Bakke
ABB Environmental Systems
Preceding page blank
8A-41
-------
8A-42
-------
I
Introduction
Utilities choosing flue gas desulfiirization as
a strategy for compliance with Phase 1 of the 1990
Clean Air Act Amendments will largely turn to
limestone wet scrubbing as the most cost-effective,
least-risk option. It has also become increasingly
apparent that employing high efficiency scrubbers on
larger, newer plants is the favored alternative for
implementing FGD.
The special demands of retrofitting scrubbers
into plants affected by acid rain legislation can be
effectively met through the use of advanced limestone
FGD technology. State-of-the-art single absorber wet
scrubbing systems can be designed to achieve:
* SO- removal efficiencies in excess of 95%,
* System availabilities in excess of 98%, and
* Byproducts which can be marketed or
landfill ed.
Further, these objectives can be accomplished at
levelized costs which compare favorably to other
clean air compliance options.
As a result of varying fuel characteristics,
site considerations, and owner preferences, FGD
plants for large central power stations ate typically
custom-designed. In order to avoid the risks
associated with new, first-of-a-kind technologies,
utilities have preferred to purchase FGD systems
from suppliers with proven utility experience and
reference plants as close as possible to the design
envisioned. This tendency favors suppliers with
broad, diverse experience bases.
Further, as the market for FGD systems is
regulatory driven, the has shifted
geographically in response to national environmental
policies. Although limestone wet scrubbing has
emerged as the overwhelming choice for SO:
emission control in coal-fired power stations, the
technology has evolved and been adapted to suit local
and regional '-°1 and economic situations.
Global suppliers are able to benefit from experience
and technological advances in the world market.
ABB is a leading supplier of wet FGD
systems worldwide having contracted for over 23,000
MW of coal-fired utility capacity (Table 1). With
market units in the U.S., Denmark, Italy, Sweden,
and Germany active in the design and supply of wet
FGD plants, ABB has a unique ability to incorporate
knowledge and experience gained throughout the
industrialized world to acid rain retrofit projects in
the U.S.
This paper describes the design of advanced
limestone wet scrubbing systems for application to
acid rain retrofits. Specifically, the evolution of
advanced design concepts from a global experience
base is disrussfd.
Preceding page blank
8A-43
-------
Designing for High Efficiency
Design Criteria
Unlike previous SO; emission regulations
which provided no incentive to scrub beyond the
legally permitted level, the 1990 Clean Air Act
Amendments have created a system of allowances and
credits which can be used to offset emissions at other
stations, banked, or sold. Hence, the economics of
FGD system design depend on the relationship
between the marginal cost of additional scrubbing
capacity, the value of emission allowances, and a
utility's future needs for these allowances.
While there is considerable speculation
concerning the various economic tradeoffs a clear
pattern of increasing SO: removal efficiencies is
emerging. Whereas plants purchased under the NSPS
provisions of the 1979 Clean Air Act Amendments
rarely required removal efficiencies in excess of
90%, acid rain retrofit plants are typically designed
for a minimum of 95%. This upward pressure on
SO; removal efficiencies has challenged designers to
develop approaches to maximize efficiency while
minimizing operating costs. At the same time,
reliability cannot be sacrificed.
High Efficiency Spray Tower Absorbers
The single loop, countercuirent spray tower
has been proven to be the most cost effective and
reliable device for removal of sulfur dioxide from
coal-fired utility flue gases. This is demonstrated by
the fact that spray tower absorbers have predominated
in FGD markets throughout the world. In fact, since
early 1990 about 80% of world-wide wet FGD orders
(totalling 22,500 MW) have been awarded to spray
towers. Advances in process design, control systems,
materials of construction, and operating philosophies
have enabled countercurrent spray towers to maintain
their competitive edge over competing technologies
such as packed or tray towers which tend to increase
operational risks in exchange for potential marginal
reductions in operating cost.
Although many aspects ate weighed in the
design of spray tower absorbers for high efficiency
acid rain retrofits, three very critical considerations
are L/G selection, gas/liquid contact, and power
consumption. The following paragraphs will discuss
these issues in more detail.
L/G Selection
The ratio of the quantity of slurry used to
treat a given quantity of flue gas is arguably the
single most significant design parameter in limestone
wet scrubbing. L/G impacts capital costs due to its
relationship to spray system design and operating cost
due to its effect on energy consumed by the recycle
pumps.
Table 2 outlines the effect of a number of
design parameters on the liquid-to-gas ratio selection
for both high and low sulfur applications. The
relationship between many of these parameters and
the L/G is straightforward and easily understood.
For example, there is little doubt that increasing the
design removal efficiency will result in a higher L/G
requirement, or that lowering the pH of the spray
slurry will require a higher L/G. The effects of
droplet size, bank spacing, and tower height are
commonly misunderstood and deserve further
discussion.
Droplet Size
The relationship of spray droplet size to
removal efficiency or L/G depends on the SO;
concentration in the flue gas. At high concentrations,
the overall absorption process is limited by the
availability of liquid phase alkalinity to neutralize the
absorbed SO:. Thus, Actors such as the slurry pH,
the presence of buffer additives, or the limestone
grind are relatively more important in determining the
overall absorption rate.
At lower gas phase SO; concentrations, more
than sufficient alkalinity is generally present to
neutralize the absorbed SO; and, consequently, the
limiting step in the absorption process becomes the
interfacial area available for SO, transfer. Hence,
droplet size has a relatively larger effect on removal
efficiency and L/G selection in low sulfur
applications.
In advanced limestone systems, the nozzle
design and operating pressure are selected to provide
the optimum droplet size for the design sulfur dioxide
regime. While most acid rain retrofit applications
call for nozzles with capacities of 250-400 gpm, ABB
has commercial experience with nozzles ranging from
125 to 850 gpm.
Improved hollow cone nozzle designs are
capable of providing the required atomization and
spray pattern characteristics at pressures as low as 7-
8 psi. Consideration may be given to utilizing two or
more nozzle configurations in high sulfur
8A-44
-------
applications. For example, low energy nozzles may
be employed in ihe lower stages where droplet size is
less critical due to the higher sulfur concentrations,
and higher pressure nodes with smaller droplets in
the upper stages where the SOw concentration is
lower.
Tower Height/Bank Spacing
There currently exists significant confusion
concerning the effect of tower height and spray bank
spacing on SO. removal efficiency and L/C selection.
Some investigators have conducted experiments which
have led them to conclude that a strong dependency
exists between overall tower height and removal
efficiency. Were this to be true, it would suggest
that greater spacing between spray banks and, hence,
taller towers are preferable.
In general, the conclusion that tower height
exhibits a strong effect on efficiency resulted from
experiments in which single spray banks at different
elevations were placed in service. The removal
efficiency was then recorded with different elevations
in operation. Significantly improved performance
was noted with higher elevations in service.
Neglected in this analysis is the fact that gas/liquid
contact conditions in the upper elevations are more
favorable due to development of more uniform gas
flow distribution.
Figure 1 depicts actually operating data from
full-scale ABB installations in which multiple
elevations were in operation. In this scenario, which
attenuates non-uniformities in gas flow distribution
atd more correctly ^mniatx actual operation, the
effect of tower height can be seen to be minimal
Hence, spray tower absorbers in advanced
limestone scrubbing systems can be quite compact in
size. In addition to the obvious capital cost savings
associated with compact towers, retrofitability is
enhanced. ABB spray towers are designed for 5-6
feet spacing between spray elevetions depending on
physical constraints of the header aad support
systems.
Gas/Liquid Contact
The demand for SO. removal efficiencies in
excess of 95 % requires very effective contact
between the flue gas and the spray slurry. Sneakage
or under-treatment of even a small percentage of the
gas can cause removal efficiencies to fall below the
anticipated level. Figure 2 demonstrates the effect of
con-ideal gas/liquid contact. In Case A, the curve
indicates that despite large increases in UC no
further improvement in removal efficiency occurs.
Simplistically. about 5 % of the flue gas is not being
treated properly in this example. In Case B where
gas/liquid contact was improved by utilizing different
nozzles it can be sees that extremely high efficiencies
can be obtained at reasonable liquid-to-gas ratios with
spray tower absorbers in limestone FGD systems
without the use of buffer additives.
Key factors related to efficient gas/liquid
contacting in spray tower absorbers include gas inlet
design and spray system design.
Cos Inlet Design
The two principal process considerations for
proper gas inlet design are:
• Uniformity of gas distribution at the
absorber inlet flange, and
• Sufficient flue gas momentum entering the
quench zone.
Uniform gas distribution at the absorber inlet
flange is clearly important to obtaining the desired
uniformity in the spray zone. Of particular
importance are side-to-side variations which would
result in higher gas velocities on one side of the
tower than the c- ier. ABB experience has
demonstrated that variations in excess of IS % RMS
are detrimental to tower performance. The required
uniformity is obtained by the placement of turning
vanes and perforated plates in the inlet duct and
confirmed by gas flow modelling.
ABB also employs a sweep bottom design to
improve performance aad minimi?* inlet duct
deposition. The sweep bottom inlet is designed to
divert a slightly higher fraction of the incoming flue
gas to the inlet duct bottom at a velocity which is
higher than the mean. By doing so. splash back of
slurry and gas recirculation which could le&i to solids
deposition in the wet/dry interface area are
minimi Further, this design assists in obtain
good front-to-back gas distribution in the tower.
In addition to uniform distribution entering
the tower, it is important that the flue gas have
sufficient momentum to insure turbulent contacting
and penetration of the incoming gas into the quench
zone. Both experience and wet flow
modelling studies have mdiratrri that high efficiency
spray towers should have inlet gas velocities of 2700-
3000 fpm. This feature along with the deflector
baffle (rain hood) over the inlet prevent build up of
8A-45
-------
deposits which could adversely affect performance or
service life of materials.
Spray System Design
Selection and placement of spray nozzles is
of critical importance in high efficiency spray towers.
To assist designers in laying out spray systems, ABB
has developed a computer model which analyzes and
provides a pictorial representation of spray flux and
pattern overlap. Given die spray nozzle characteristics
(e.g. spray angle, radial spray distribution, etc.) and
the header arrangement, the model determines the
degree and number of overlaps at a predetermined
distance from the nozzle tip.
Figure 3 is a typical example of the spray
coverage achieved by ABB designs. Generally,
overlaps on the order of 600-700% at a distance of 3
feet are achieved with 250-400 gpm hollow cone
nozzles.
Having achieved the desired spray coverage
from a single spray elevation, the overall gas/liquid
contact in the tower is further improved by staggering
the headers on adjacent elevations (Figure 4).
Staggering the headers insures full development of
the spray pattern from the nozzles and prevents the
formation of passages for flue gas sneakage.
The impact of spray system design is
graphically demonstrated by Figure 2. The principal
difference between the performance between Case A
and Case B is an improvement in spray coverage. In
Case A, poorly designed nozzles resulted in non-
uniform spray pattern development and,
consequently, substantial gas sneakage. When the
nozzles were replaced in one of die absorbers, side-
by-side comparisons showed substantial performance
improvement as demonstrated by die curve for Case
B.
Thus, ABB experience has shown that with
proper inlet and spray system design SO: removals in
excess of 98% are readily achievable without the use
of buffer additives with limestone spray towers. This
level of performance when coupled with the inherent
simplicity and reliability of countercurrent spray
towers are highly desirable in advanced wet
scrubbing systems.
Power Consumption
Because of the inherently low gas-side
pressure drop associated with spray tower absorbers,
a typical breakdown of power consumption by
component in an advanosd FGD system on a 2-3%
sulfur application would be as follows:
Component
Recycle (spray) pumps
Booster Cms
Oxidation blowers
Others
Percent of
Total Power
50%
20%
10%
20%
For this reason considerable attention has been given
to minimizing recycle pump power consumption in
advanced limestone FGD systems.
The principal determinants of recycle pump
power are L/G, tower height, and nozzle pressure.
Concepts used to optimize these variables are
riisnissnd in the following paragraphs:
Liquid-to-Cas Ratio
Under U.S. utility purchasing
practices, the cv--.it,d its consultants define the
basis for system desist -2y specifying variables such
as the fuel composition, required efficiency, and
reagent type, the supplier is provided with the basic
requirements of the process design.
Computer models based on full-scale,
commercial experience are the principal tools used by
ABB process engineers to select the proper L/G for
a given application. With the merger of Flakt,
Peabody Process Systems, and Combustion
Engineering into ABB Environmental Systems, a
single data base of performance and operating data
unparalleled in the industry has been created. This
data has been usod to calibrate performance models
which allow L/G to be selected and optimized as a
function of inlet SO^ concentration, removal
efficiency, buffer additives, rfiTaitwioirs,
tower design, and liquid phase chemistry. Table 3
summarizes the breadth and depth of ABB
commercial experience with variables affecting SO,
absorption.
Tower Height/Nozzle Pressure
<">w» inipnft.nl rrmipiw*!! in rlrtnmminf
total discharge head of the recycle pumps and, hence
power consumption, is the height from the surface of
the recycle tank slurry to the nozzle tip for a given
spray elevation (dimension h in Figure 5). In ABB
advanced limestone scrubbers this parameter is
by compact tower designs based on proven
commercial experience. AH aspects of the design
irwinHing the inlet duct geometry and aspect ratio.
8A-46
-------
height of the first spray elevation above die top of the
inlet, and the spacing between spray elevations have
been confirmed by full-scale operating experience.
Typical values for key dimensions are shown in
Figure 5.
Spray nozzle pressure drop is the other
major contributor to recycle pump energy
consumption. Process considerations which enter into
nozzle selection include:
• Droplet size distribution
• Uniformity of spray distribution
• Uniformity of gas distribution
• Pluggage potential
ABB has determined that hollow cone nozzles
operating at 7-8 psi provide the optimum balance
between power consumption and atomization for most
limestone scrubbing applications.
Commercial Experience Base
From the discussions above, it can be seen
that the design of high efficiency absorbers for coal
-fired utility applications is a complex task involving
numerous process and mrrhanical considerations.
While theoretical models developed from mass
transfer theory are very useful for predicting and
extrapolating tower performance, it is absolutely
critical that these models be rigorously calibrated
with full-scale, commercial operating data.
Since ABB's first spray tower absorbers
were placed in service in the U.S. in the late 1970's
to present day, an extensive program of testing and
performance characterization of commercial
installations has been used to create a detail
computerized database consisting of over 140
individual records. These test have been conducted
over a broad range of such key variables as sulfur
dioxide inlet concentration, removal efficiency, L/G,
spray tower configuration, nozzle type, gas velocity,
chloride concentration, and limestone stoichiometry.
The data generated by these tests has been
used to calibrate both internally developed design
models and EPRI's FGD PRISM. The predictive
capability afforded by these models allows ABB to
produce highly optimized designs for its commercial
offerings.
8A-47
-------
Designing for High Availability
Factors Affecting Availability
FGD availability is generally discussed in
terms of the percentage of hours that the system is
available for operation over a given period. Whereas
first generation FGD plants often exhibited periods of
very low availability, advances in design, materials,
and operational philosophies have enabled more
recently installed systems to achieve availability
records which often exceed 98-99% on a annual
basis. An ABB Flakt FGD system in Denmark at
Elkraft's Amager Power Station Unit 3 (Figure 6) has
recently completed a year's operation at 100%
availability. This is even more remarkable
considering that the 250 MW installation operates
8.000 hours per year at load factors as high as 80%
with a single absorber train (i.e. no spare).
ABB has determined that FGD system
reliability is primarily influenced by the following
factors:
* Design and sparing philosophy (e.g. process
design, oxidation, use of additives, materials
of construction, component selection, etc.)
* Service conditions (e.g. fuel sulfur and
chloride content, base vs. swing load, fly
ash levels, etc.), and
* Operation and maintenance (e.g. control of
key process parameters, preventive
maintenance, etc ).
While a detailed discussion of factors affecting FGD
system availability is beyond the scope of this paper,
it is useful to consider several characteristics of
advanced wet scrubbing systems and their relationship
to availability. Of particular interest are issues such
as the need for spare absorbers: operation in forced,
natural, and inhibited oxidation modes; the use of
chemical additives; and, process control and
operation.
Spare Absorber Vessels
The NSPS provisions of the 1979 Clean Air
Act Amendments essentially mandatrri that FGD
systems be equipped with spare absorber modules.
While the cost associated with spare absorber
modules is significant, considerable operational
flexibility is achieved. FGD system operators have
developed effective operation and maintenance
procedures based on rotating absorbers into and out
of service on a periodic basis. As a result of this and
improvements in design and materials, U.S. owners
of limestone wet scrubbers recorded availabilities at
or near 100% on an «""n«liTxl basis.
The 1990 Clean Air Act Amendments
however provide no specific incentives for the
purchase of spare absorbers. In both Europe and
Japan where this has always been the case, single
train FGD systems have been installed on steam
generators as large as 700 MW. Although there has
been some debate is to whether this technology is
directly translatable to U.S. high sulfur coal burning
power stations with high load factors. ABB believes
that single train systems will become the preferred
choice for most applicanons in the U.S. as well. In
the short term, some utilities are choosing an
intermediate course in which multiple modules with
no spare is specified. This permits partial load
operation in the event of an emergency scrubber
outage.
Economic studies including internal ABB
analyses have confirmed that 10-20% reductions in
Total Plant Cost (TPQ for a 500 MW acid rain
retrofit can be realized by replacing 3 x 50% capacity
trains with a single absorber. In addition to the
obvious savings in absorber and flue gas handling
system costs, substantia] reductions and simplification
in piping, controls, and electrical subsystems
contribute heavily to the overall savings. Provided
that availability of single absorber FGD systems is
equal to or better than multiple module systems with
spares, the economic benefits of the single train
approach are clear.
Design of high reliability single train FGD
systems requires superior design in several key areas.
ABB advanced limestone FGD systems incorporate
features such as the following to
availability:
• Absorber inlet ducts are constructed of
highly corrosion resistant materials such as
C-276 (solid plate or lining).
• Sweep bottom inlets with sufficient flue gas
momentum to prevent recirculation are
employed to eliminate deposition in the
wet/dry interface area.
8A-48
-------
* Hollow cone nozzles with large free passage
minimi7a the potential for phiggage.
* Proven multi-stage mist eliminator with top
and bottom wash system capability are
provided to minimi™ the need for manual
cleaning.
* Alloy or rubber-lined spray zones insure low
maintenance and excellent corrosion/erosion
resistance.
* Spare absorber recycle pumps insure
continued compliance with emissions
regulations.
Features such as those described along with the
owner's commitment to a comprehensive operation
and maintenance program will permit advanced
limestone FCD plants to achieve the high levels of
availability required by modem power stations.
Oxidation Mode
FGD systems are operated in one of three
modes with respect to oxidation.
* Natural Oxidation - Oxidation is
uncontrolled. Some oxidation (typically 10-
20 96 for mrriium to high sulfur coals) occurs
due to the presence of oxygen in the flue
gas. A mixture of calcium sulfite and
calcium sulfate is formed in the slurry
solids.
* Forced Oxidation - Air is injected into the
absorber recycle tank to achieve oxidation
percentages approaching 100%. High purity
calcium sulfate dihydrate (gypsum) can be
produced in such systems. Due to the
production of large crystals, high capacity
dewatering systems can be designed.
* Inhibited Oxidation - Oxidation inhibitors
such as thiosulfiue ion are used to minimis
oxidanon. Operation in this mode on result
in the process operating subsanirated with
respect to gypsum which has been shown to
be beneficial from the standpoint of scaling
and deposition. In addition, thiosulfate
addition has been shown to improve
dewatering characteristics of the slurry
solids in comparison to natural oxidation
processes.
ABB experience with wet FCD systems producing
gypsum is cnmnnriin Table 4.
Although operation in any of these three
modes is possible, ABB recommends that advanced
limestone FGD systems be designed for either forced
or inhibited oxidation. Experience has shown that
fewer problems with scaling and deposition
particularly in the mist eliminator area occur when
oxidation is either forced or inhibited.
From an overall cost standpoint, studies have
shown that forced oxidation is generally more cost-
effective than natural oxidation. However, the
margin is small enough that the choice between
forced and inhibited oxidation is primarily dictated by
waste product considerations. Gypsum production
from forced oxidation systems offers considerable
flexibility in that the product can be either sold or
landfilled. Although disposal options are ii™;t~i to
landfill with inhibited oxidation, simplicity of
operation, extremely clean operation, and the ability
to co-dispose flyash make inhibited oxidation with
byproduct fixation an alternative.
Use of Additives
Several types of chemical additives have
been employed in advanced limestone scrubbers.
Organic buffers are used to enhance SO, removal
performance and oxidation inhibitors are used to
minimi» gypsum scaling potential. In addition,
crystal growth modifiers and corrosion inhibitors
have been investigated.
ABB experience with performance additives
is summarized in Table 5. The roles of buffer
additives and oxidation inhibitors in advanced
limestone FGD systems are disnresnd in the following
paragraphs.
Buffer Additives
The addition of modrst amounts of organic
acids such as formic, adipic, and dibasic acid cause
substantial improvement in SO} removal efficiency.
Figure 7 presents operating data from a full-scale
installation at various DBA concentrations. The
improvement in removal efficiency is due to the
increased levels of soluble alkalinity caused by the
dissociation of the DBA. Although the removal
efficiency benefits are well documented both in terms
of process design and economics, the reliability
aspects of operation with buffer additives are often
overlooked.
8A-49
-------
First, the addition of buffer additives allows
the process to be operated at a lower pH than a
corresponding non-buffered system. This has been
shown to improve availability by promoting cleaner
operation in the mist eliminator zooe. Further, the
presence of the buffer attenuates pH swings as a
result of a mismatch between limestone feed rate and
SO; absorption rate (e.g. during load changes or coal
sulfur swings). Again, this tends to promote cleaner
operation.
Second, use of buffer additives in systems
designed for non-buffered operation may allow the
desired removal efficiency to be achieved with fewer
spray elevations in service. The corresponding
benefits are reduced duty factors on spray pumps,
headers, nozzles, etc. as well as increased operating
flexibility.
Third, in closed loop operation with high
concentrations of dissolved solids such as chlorides,
the use of organic acid additives are extremely cost
effective due to improved limestone solubility and
reduced scaling potential as a result of low pH
operation.
Oxidation Inhibitors
As described previously, thiosulfate ion
either added as the sodium salt or generated in situ by
elemental suliiir addition has been shown to
dramatically reduce the oxidation of sulfite to sulfate
in limestone scrubbers. If the level of oxidation and,
hence, the rate of sulfate formation can be depressed
sufficiently, the process can be operated subsaturated
with respect to gypsum.
The principal benefit of this mode of
operation is the elimination of gypsum scaling
potential. Inhibited oxidation systems are therefore
less subject to deposition and fouling of absorber
internals as a result of process upsets. As the
byproduct cannot be marketed, some loss in
flexibility with respect to solid disposal results.
Advanced limestone FGD systems can be
designed for operation with or without the use of
chemical additives. The positive economics of
organic buffer addition have been thoroughly
documented. Taken to the fullest extent, ".Kct.nri.i
reductions in both capital and operating costs are
possible. Currently, most systems are designed such
that the required removal efficiency over the range of
fuels is achievable without the use of organic buffers.
Tbe buffer additives are then employed primarily to
reduce operating costs and increase removal
efficiency.
The cost benefit of oxidation inhibitors is
more difficult to judge in tbe specification or design
stage. The increase in availability resulting from
subsaturated operation has not been sufficiently
documented to enable reliable predictions to be made
on a generic basis.
Process Control and Operation
Successful operation of single train limestone
wet FGD systems has been widely demonstrated. As
discussed previously, ABB's installation at the
Amager Power Station in Denmark has achieved
annual availabilities of up to 100%. A significant, if
not tbe major, contributor to high FGD system
availability relates to quality of tbe operation and
maintrn»n
-------
Openbility and maintainability must be
seriously considered in the design stage.
For example, sufficient access and lifting
equipment to service all key components
must be provided; key instruments which
require frequent calibration should be easily
accessible (at grade elevation where
practical); automated flushing and draining
of slurry piping systems should be provided,
etc.
Operator training programs with on-site
support from the process supplier prior to
commissioning and periodically thereafter
are extremely important.
In recognition of the importance of O&M
issues to reliable operation, ABB sponsors FGD
User's Conferences for owners of its systems.
Operations and maintenance personnel from all U.S.
and European FGD installations are invited to attend
a three-day conference held every 12-18 months.
Presentations and workshops related to reliability,
O&M procedures/training, new technical
developments, etc. are disniwd in a open forum.
These sessions have proved highly effective in
providing a better understanding of O&M problems
and issues to ABB engineers as well as a forum for
exchange of ideas among the owner/operators of ABB
systems. ABB is committed to continuing these
conferences as a mrans of improving the design and
reliability of its wet limestone scrubbing systems.
8A-51
-------
Advanced Limestone FGD System Costs
Advanced Technologies
The Tool PUai Cost (TPQ for retrofitting
FGD equipment on a coal-fired power station can
reach S200-250/KW according a recent update of the
Electric Power Research Institute's (EPRI) landmark
economic study of FGD technologies. These figures
underline the importance of cost in the selection of
high efficiency wet scrubbing as a compliance option.
Fortunately, however, advances in technology and a
combination of economic and market conditions have
caused FGD system costs to remain quite stable and.
in fact, decline in recent years.
The following paragraphs summarize the
results of an economic comparison among three
alternative limestone WFGD designs. The
alternatives are defined as follows:
Conventional refers to the limestone forced
oxidation system used by EPRI in comparing
costs of various FGD alternatives. The
economics generated by EPRI for this
technology serve as the benchmark for
comparing advanced technologies.
Advanced refers to present day ABB designs
featuring single absorber trains, high
technology spray tower designs, and state-
of-the-art materials. Costs are
representative of current commercial
offerings.
Third Generation refers to potential future
advances in wet limestone FGD
technologies. While still under
development, the technology described as
third generation represents possible future
directions for performance improvement and
cost reduction..
The plant design and economic evaluation
criteria, process design, total plant investment, and
operating costs for these alternatives are summarized
in Tables 6, 7, 8, and 9 respectively and disnissrri in
the following paragraphs.
Plant Design and Economic Criteria
The plant design and economic criteria and
methods (Table 6) used in this analysis are identical
to those used by EPRI for retrofit systems. These
criteria were selected because they form the basis for
a large body of work performed previously by EPRI
and others. Hence, the figures generated in this
analysis can be compared to other technologies in the
EPRI database.
The analysis is based on a mid-west
(Kenosha, WI) 300 MW power plant burning 2.6%
sulfur fuel. The plant ™""|<»I load capacity is
assumed to be 65% and. being a retrofit on an
existing power station, a plant life of IS years is used
in the analysis.
Capital and operating costs used by EPRI for
the limestone forced oxidation process were used as
the base (Conventional) case for this analysis. Cost
rstimatrs for the Advanced and 3rd Generation
processes were generated on the basis of internal
estimates. These costs were «n»™aii™n on the basis
of the economic criteria presented in Table 6.
All major operating costs are included on the
basis of estimates for consumables and labor. The
cost of consumables such as power, limestone, water,
etc. are based on actual estimates of the performance
and current costs. Q&M labor costs were developed
from estimates of the numbers of operations and
numt^aiw» personnel required on a per shift basis.
Other costs such as A/E fees, general
facilities work, and project contingencies were also
developed in accordance with the criteria and
methods developed in the EPRI study.
Design Comparison
The base (Conventional) case assumes a
limestone forced oxidation system designed to remove
90% of the incoming sulfur dioxide (Table 7) using
3 x 50% capacity absorbers (i.e. 2 operating/1
spare). The process uses thickeners and vacuum
filtration to produce gypsum for sale or disposal.
The technology described in the Conventional case is
typical of that offered in the late 1980's.
Hie principal differences between the
Conventional and Advanced design are as follows:
• The advanced design employs a single
absorber train.
• Improvements in spray tower technology
(i.e. lower power consumption, compact
8A-52
-------
design, etc.) have been included.
* Removal efficiency has been increased to
95%.
The Advanced design is considered typical of current
offerings in the industry.
The 3rd Generation design extrapolates
current designs to project potential cost savings due
to further improvements in absorber/mist eliminator
design. These improvements which are still under
development are expected to produce savings in the
design of Phase 2 acid rain retrofits and new plant
applications. In addition, the 3rd Generation system
is designed to take full advantage of the use of
chemical additives to enhance performance.
Economic Comparison
A comparison of the Total Plant Investment
(TPI) for the three processes is presented in Table 8.
As can be noted. Total Process Capital (TPC)
reductions in the range of 15-20% are achieved by
the Advanced technology when compared to the
Conventional. These savings are primarily due to the
impact of utilizing a single absorber train in the
Advanced process. Savings are realized in absorbers,
ductwork, piping, control, and electrical subsystems.
Further savings are realized in the 3rd Generation
system as a result of more compact absorber design
and reduced spray system costs associated with the
use of organic acid buffets to rnhanrr SO: removal.
The TPC cost savings are reflected in proportional
reductions in General Facilities, Contingencies, etc.
The cost reductions noted in the Advanced
and 3rd Generation technologies occur despite the
increase in design SOz removal efficiency (i.e. 95%
and 98% vs. 90% for the Conventional process). In
actual, practice ABB feels thai the observed savings
may not be quite as dramatic as indicated by this
analysis. The costs attributed to the Conventional
technology are probably somewhat higher than the
market price due to conservatism in the original
estimate and the presence of competitive market
pressures.
A similar comparison of Operating Costs
was performed and is presented in Table 9. Again,
substantial savings are seen in the Advanced and 3rd
Generation cases when compared to the
Conventional. Annual O&M Cost savings result
primarily from reductions in power consumption.
Improvements in spray tower technology and, in the
3rd Generation case, the use of organic acid buffers
contribute to substantially lower energy consumption
in the pumping systems. Total Operating Cost
savings in the Advanced and 3rd Generation result
from the Annual O&M Costs savings as well as
reduced Fixed Charges. Again, these savings result
despite the improved SO. removal efficiencies offered
by the Advanced and 3rd Generation technologies.
8A-53
-------
Summary
Although modern wet limestone scrubbing
has been in existence for nearly 30 years, evolution
of the technology continues. The earliest utility FGD
systems were based on designs adapt rri from wet dust
collector technology and the chemical process
industry. Often these systems attempted to remove
SO;, calcined limestone, and fly ash in a packed or
tray tower. A lack of understanding of the process
and mechanical design requirements resulted in poor
compooent performance; selection of unsuitable
materials of construction; scaling/plugging of
absorber internals and piping; and, numerous other
O&M problems. These problems understandably
lead to poor performance and availability of FGD
equipment.
During the early and mid 1970's, experience
with commercial installations coupled with R&D
programs conducted by suppliers and organizations
such as EPA and EPRI contributed significantly to
improved performance and reliability in limestone
wet scrubbers. Among other advances, the spray
tower was adopted as the preferred absorber design
due to its inherently low pluggage potential and low
power consumption. Improved materials of
construction and additional application experience
reduced the incidence of corrosion, erosion, and
general failures. The availability of more accurate
and reliable instrumentation permitted better process
control. Plant operations and maintenance personnel
developed O&M experience bases which allowed the
development of standard policies and procedures
at high availability operation.
As utility wet scrubber technology prepares
to enter its 25th year it continues to evolve in
response to industry needs, technological advances,
and competitive pressure. In the U.S., the concept of
high efficiency, single train, multi-bank spray towers
treating high sulfur flue gases while producing either
marketable or landfill gypsum is expected- to be the
preferred choice for many acid rain retrofit projects.
In Europe where many of these particular design
features are already commonplace, concentrated
efforts are underway to reduce capital and operating
costs to achieve competitiveness in local and world
markets.
Because of its global presence. ABB is active
in both of these areas. Through regularly scheduled
internal tx-frnirai meetings, standardization of process
and nw-i,«n;rai design, and an overall commitment to
the world FGD market, ABB is able to benefit from
knowledge and experience gained throughout the
industrialized world. This unparalleled capability is
applied to produce reliable, cost-effective wet
limestone FGD designs for U.S. acid rain projects.
8A-54
-------
Figure 1
Effect of Tower Height
Unit A
91.5% 91.3%
Unit B
92.5% 91.1%
Unit C
96.0% 94.3%
Spray Level in Service
Spray Level Not in Service
8A-55
-------
Figure 2
Gas Liquid Contact
00
>
c
O
E
-------
Figure 3
Computer -Simulation of
Absorber Spray Pattern
8A-57
-------
TYPICAL
* PASS MIST
eumnator
SECTION
typical
2 PASS UlST
ELIMINATOR
SECTION
RECYCLE
HEADERS
TYPICAL
SPRAY
NOZZLE
Figure 4
ABB Spray Tower Design
8A-58
-------
Figure 5
Key Spray Tower Dimensions
'<^<<<<<<<<<<<
-------
Figure 6
Elkraft Amager Power Station
(ABB FGD Building in Foreground)
SA-60
-------
Figure 7 - Influence of Buffer Additives
3500 ppm S02
160
100
Removal Efficiency at
a Constant L/G of 80
140
90
120
80
100
70
L/G at a Constant
Removal Efficency
of 90 Percent
80
2000
60
1500
1000
Buffer Additive, (ppm)
500
(
-------
TABLE 1
ABB WET FGD EXPERIENCE
WORLDWIDE
U.S. 20,250 MW
Italy 1,960 MW
Denmark 900 MW
Taiwan 700 MW
Total 23,810 MW
8A-62
-------
TABLE2
RELATIVE INFLUENCE OF KEY
PARAMETERS EFFECTING L/G
HIGH
LOW
INCREASE IN PARAMETER
SULFUR
SULFUR
S02 Removal Efficiency
Strong/Increase
Strong/Increase
Droplet Size
Weak/Increase
Minimal Effect
Spray Bank Spacing
Minimal Effect
Minimal Effect
Tower Height
Minimal Effect
Minimal Effect
pH
Strong/Decrease
Strong/Decrease
Buffer Additive Concentration
Strong/Decrease
Weak/Decrease
Chloride Concentration
Strong/Increase
Strong/Increase
Spray Flux
Minimal Effect
Minimal Effect
Absorber Gas Velocity
Weak/Decrease
Weak/Decrease
TABLE3
COMMERCIAL EXPERIENCE BASE
Parameter
Minimum
Maximum
Inlet S02 Concentration (ppm)
310
2.860
Removal Efficiency (%)
40.6
99.6
LVG (gal/kcf)
15
189
No. of Spray Elevations
1
5
Spray Flux (gpm/sq ft)
10
30
Spray Zone Height (ft)
13
38
Absorber Gas Velocity (fps)
6
15
Spray Droplet Size (micron)
1,320
2,950
pH
5.1
6.4
Buffer Additive Concentration (ppm)
0
3.030
Chloride Concentration (ppm)
200
20,000
8A-63
-------
Table 4
ABB Wet FGD Systems Producing Gypsum
Utility
Unit
Type
Start Up
Capacity fMW^I
Northern States Power
Sherburne 1
Disposal
1976
740
Northern States Power
Sherburne 2
Disposal
1977
740
Texas Utilities
Sandow4
Disposal
1980
545
Tennessee Valey Authority
Widow's Creek 7
Disposal
1981
575
Lower Colorado River Authority
: aye tie 3
Disposal
1988
435
Elkraft
imager 3
Commercial
1988
250
Louisville Gas & Electric
rrimble County 1
Disposal
1990
500
Taiwan Power
Jnkou t
Commercial
1992
350
Taiwan Power
Jnkou2
Commercial
1992
350
EN EL
SioiaTauro 1
Commercial
1994
660
EN EL
Sioia Tauro 2
Commercial
1994
660
Isefjordverket
AsnaesS
Commercial
1993
650
Total
6.455
Table 5
ABB Experience with Performance Additives
Utility
Unit
Capacity (MW)
DBA/Adipic Acid
Seminole Electric
Seminole 1
620
Seminole Electric
Seminole 2
620
Houston Lighting & Power
Limestone 1
750
Houston Lighting & Power
Limestone 2
750
Orlando Utilities
Stanton 1
450
Plains Electric
Escalante 1
235
Oxidation Inhibitors
Louisville Gas & Electric
Mill Creek 1
360
Louisville Gas & Electric
Mill Creek 2
360
Texas Municipal Power Authority
Sibbon's Creek 1
445
Seminole Electro
Seminole 1
620
Seminole Electric
Seminole 2
620
New York State E & G
Somerset 1
635
Houston Lighting & Power
J'mestone 1
750
Houston Lighting & Power
Jmestone 2
750
South Carolina Public Service
Cross 2
500
8A-64
-------
Table 6
Plant Design and Economic Criteria
Plant Size ( MW elec.)
300
Coal Heating Value (Btu/lb)
13,100
Coal Firing Rate (tph)
111
Max. Sulfur (%)
2.60
Max. Chlorine Content (%)
0.12
Ash Content (%)
9.10
Inlet Gas Flow (acfm nominal)
685,000
Inlet Temp. (deg. F)
300
Current Annual Load Factor (%)
65
Current Full load oper. (hrs/yr)
5,694
Plant Life (yr)
15
Discount Rate (%)
11.5%
Fixed Charge Rate, 15 yr (%)
19.2%
Power Cost ($/MW-hr)
50
Limestone Cost ($/Ton)
15
Dibasic Acid Cost ($/lb)
0.18
Table 7
Process Design Comparison
Advanced
3rd
Conventional
LS System
Generation
S02 Removal (%)
90
95
98
L/G (gpm/kacfm)
140
125
100
No. of Operating Modules
2
1
1
No. of Spare Modules
1
0
0
Absorber Velocity (ft/sec)
10
10
15
Absorber Diameter (ft)
29.8
422
34.4
Number of Operating Headers
5
4
3
Absorber Pressure Drop (in.W.G.)
5.00
5.00
6.50
Recycle Spray Nozzle Pres. (psi)
15.0
10.0
8.0
Avg. Net Header Elev. (ft)
50.0
47.0
43.0
Recycle Slurry Solids Cone. (%)
15.0
15.0
20.0
Solids Residence Tir-o (hrs)
25.0
18.0
15.0
S.R. (mole alk./mole S02 abs.)
1.10
1.05
1.03
Organic Acid Cone, (ppm)
0
0
500
Oxidation Air Stoichiometry (0/S02)
6.0
3.0
1.5
8A-65
-------
Table 8
Total Plant InvestmentComparison ($000)
Advanced
3rd
Conventional
LS System
Generation
Reagent Storage and Prep.
11.010
10,460
10,460
FGD Area
21,300
17,040
15,336
Flue Gas Handling
7.200
2,880
2,880
Waste Disposal
2.010
2,010
£010
Other
1.770
1,770
1,770
TOTAL PROCESS CAPITAL (TPrC)
43.290
34,160
32,456
General Facilities @ 10% TPrC
4.329
3.416
3,246
Engineering & Home Office Fees @ 10% TPrC
4.329
3.416
3,246
Project Contingency @ 12.5% TPC
7.507
5.924
5,628
Process Contingency @ 1.5% TPC
901
711
675
TOTAL PLANT COST
60.055
47,389
45.025
Interest during Construction
3,003
2,369
2,251
TOTAL PLANT INVESTMENT
63,058
49,758
47,276
Delta
Base
(13,300)
(15.782)
TOTAL PLANT INVESTMENT ($/KW)
210
166
15ft
Table 9
1
First Year Annualized Operating Cost Comparison ($000)
Advanced
3rd
Conventional
LS System
Generation
Power
1.519
1,144
949
Reagent
788
794
803
Organic Acid
0
0
24
Waste Disposal inc. flyash
114
116
118
Waste Water Treatment
3
3
3
Maintenance
649
512
487
Manpower (5 shifts)
2.500
1,500
1,500
Byproduct Sale
(156)
(160)
(163)
ANNUAL O&M COST
5.418
3.909
3.721
[Delta
Base
(1.509
(1.698)
Fixed Charges
12,107
9.554
9.07^i
TOTAL OPERATING COST
17.525
13,463
12.79^
TOTAL OPERATING COST ($/MW-hr)
10.26
7.88
7.49
8A-66
-------
Design and Operation of Single Train Spray Tower FGD
Systems
8A-67
-------
8A-68
-------
A. Saleem
General Electric Environmental Services, Inc.
200 North Seventh Street
Lebanon, PA 17042
ABSTRACT
This paper describes advances made over the past 20 years in the design of open
spray towers to achieve dramatic results in performance, reliability and cost
reduction. Today open spray towers are serving more than 25,000 MW of F6D
systems dealing with flue gases from a wide variety of boilers burning low to
high sulfur fuels. Desulfurization efficiencies of 90 to 99% are being
achieved with reliability close to 100%. Of particular interest are the most
advanced single train open spray tower FGD systems which have become the
industry standard in Japan, Europe and now in the United States. Important
design principles for achieving high reliability and performance will be
presented along with operating experience of the single train spray tower FGD
systems
Preceding page blank
8A-69
-------
Just as coal has cone to dominate electric power generation, limestone has come
to dominate flue gas desulfurization to meet the increasingly stringent
regulations for reducing sulfur dioxide emissions. This preeminence of
limestone reagent for sulfur dioxide control is dictated by cost, availability
and logistics of material handling. A typical 500MW power plant, burning 3%
sulfur bituminous coal, would generate about 15 tons per hour of sulfur
dioxide. The quantities of various possible reagents that will be needed, and
byproducts generated, per unit of sulfur dioxide removed from the flue gas are
shown in Figure 1. Limestone is the most abundant naturally occurring reagent
and has the lowest cost. Therefore, it is the most economic choice. This
position is further reinforced when byproduct disposition is taken into
consideration. Since the byproduct quantities are large, they need to be
either sold or stored as environmentally safe solid waste to avoid secondary
pollution problems. Byproducts of sulfur dioxide removal with limestone have
low solubility in water. Therefore, they can be disposed of as landfill after
proper treatment or converted into gypsum for which there is an outlet for
wall board and cement manufacture.
For reasons cited above, limestone (and to a lesser extent lime, which is
derived from limestone) have come to dominate the field of flue gas
desulfurization. However, the evolution of the limestone based processes has
been fraught with difficulties, primarily due to plugging, scaling and
inadequate performance. These difficulties can be traced back to a lack of
understanding of fundamental chemical reactions and their implications on
process and equipment designs.
LIMESTONE FGD PROCESS REACTIONS
The most efficient means of removing sulfur dioxide with limestone (or lime) is
the so called "wet" process in which an aqueous slurry of finely ground
8A-70
-------
limestone is contacted with the flue gas. Figure 2 shows a simplified block
diagram of the wet FGD process. Flue gas leaving the fly ash collecting system
is introduced into a suitable SO2 absorber4 in which SO2 is removed by
intimate contact with an aqueous suspension of limestone recycled from the
absorber slurry tank. Fresh limestone slurry is continuously charged into the
absorber tank for reaction with absorbed SO2. Reaction products are withdrawn
and sent for dewatering and further processing. Figure 3 shows the basic
chemical reactions of SO2 absorption and reaction with limestone. Reaction (1)
is common to all wet-scrubbing processes and shows the formation of sulfurous
acid, which must be neutralized rapidly to enhance the SO2 absorption.
Reaction (2) shows the neutralization of sulfurous acid with limestone. The
primary product of neutralization is calcium sulfite. Due to the presence of
oxygen in the flue gas, the secondary oxidation reaction (3) takes place which
converts a portion of the calcium sulfite to sulfate. Both calcium sulfate and
sulfite have low solubility in water and result in precipitation as shown in
reactions (4) and (5). Reaction (6) shows bisulfite formation, which is
favored by decreasing pH.
The enormous quantities of SO2 absorbed, reacted and precipitated from the
absorbing solution create an environment in which scaling and plugging can
readily take place. The development of proper safeguards against plugging and
scaling is of paramount importance for reliability of the wet FGD system.
PREVENTION OF PLUGGING AND SCALING
Process and absorber design go hand in hand in evolving an FGD system which is
free from the debilitating effects of plugging and scaling. From a process
point of view, the following criteria are essential:
LARGE LIQUID-TO-GAS RATIO. Both calcium sulfite and sulfate can form highly
supersaturated solutions by virtue of their low solubility in water.
Therefore, the liquid-to-gas ratio must be large enough to avoid any excessive
instantaneous supersaturation which can cause uncontrolled precipitation. The
minimum liquid-to-gas ratio can be estimated with a knowledge of the SO2
content of flue gas and the expected amount of sulfite oxidation.
SEED CRYSTALS. A well-known technique for controlling scale uses a suspension
of the crystals of the material being precipitated. These crystals not only
enhance the precipitation rate, but also provide host sites where preferential
precipitation takes place. Sufficient seed crystals of both calcium sulfite
and sulfate must be maintained at all times in the SO2 absorbing liquor1.
8A-71
-------
DELAY TIME. Because the process of precipitation proceeds at a finite rate,
the S02-absorbing liquor must be delayed in a tank after each pass through the
absorber to allow time for the precipitation reactions. Failure to provide
sufficient delay time increases the supersaturation, hence scaling. The exact
delay time is a function of the degree of supersaturation that is allowed to
take place during SO2 absorption, as well as the delay tank design. Plug flow
tanks are more efficient reactors than single backmixed tanks3.
ABSORBER DESIGN. From an absorber design point of view, Murphy's law must be
taken into account, ie, if anything CAN go wrong, it WILL go wrong! During
process upsets, such as loss of pH, loss of seed crystals, pump failures, etc.
the system will experience conditions that could cause plugging and scaling.
For this reason, an "open spray tower" is an ideal type of absorber for the wet
limestone process. Because SO2 absorption and reaction is carried out on
freely moving droplets, there are no gas-flow-restricting devices, thus making
the open spray tower virtually free from plugging and scaling problems even
under upset conditions.
HISTORY OF SPRAY TOWER DEVELOPMENT
The first formal development of a spray tower for SO2 removal with limestone
slurry was launched in 1970 at the Lakeview Generating Station of Ontario
Hydro, on a 4,000 cubic foot per minute slipstream of a 300MW coal fired
boiler. This pilot plant investigation clearly demonstrated the reliability
and performance of the open spray tower. The effects of such variables as
liquid-to-gas ratio, number of spray stages, nozzle size and pressure,
limestone stoichiometry, gas velocity, etc. on the mass transfer coefficient
(and hence performance) were investigated and correlated in an analytical
model. The results of this pilot plant investigation were published in the
Second International FGD Symposium in 1971.2
The first coiranercial application of the open spray tower took place in 1978 at
the 750 MW lignite-fired Monticello Station of TUELECTRIC, Texas. Enlarging
the spray tower from a mere 4,000 ACFM to 1,000,000 ACFM capacity was a
daunting challenge. Additional pilot tests at Monticello Station were
performed and coupled with 1/15 scale fluid dynamics model studies to point the
way toward a confident scal<_-up. The Monticello FGD Station went into
operation in January 1978 and its performance exceeded expectation with an SO2
removal efficiency of over 95% and a reliability factor of over 99.5%. The
operation and maintenance results of this landmark installation were published
8A-72
-------
in 1980 in the Sixth F6D Symposium in Houston, Texas.5 Since then, over 90
open spray towers have been put into successful operation by GE and its
licensees around the world. Many of these spray towers are single train
including gigantic sizes of 700 MW. The single train spray towers have
resulted in dramatic cost reductions. For instance, the 645MW Amercentrale #8
single train FGD system was contracted on a turnkey basis in July, 1985, at
S60/KW. After adding owner's costs tor foundations, auxiliaries and interest
during construction, the total cost of the FGD system amounted to S84/KW with
commercial operation commencing in March, 1988. The EPRI sponsored study
completed in 1990 compared costs of multiple trains and a single train FGD
system.8 A single train system was found to be 28.5% less costly compared to a
four 33% train system and 11.5% less costly compared to a two 50% train system.
IMPORTANT ASPECTS OF A SINGLE TRAIN FGD SYSTEM DESIGN
The single train system is defined as one in which the flue gas path has no
spare absorber and the entire gas is treated by a single absorber. Such a
system can be discussed in the context of two variations of the overall process
depending on the byproduct of flue gas desulfurization. Figure 4 shows a
system with no forced oxidation (NFO) in which the byproduct is a mixture of
calcium sulfite hemihydrate and calcium sulfate dihydrate. Since this
byproduct mixture is difficult to dewater, primary dewatering requires
thickeners. The secondary dewatering is done with drum filters or continuous
centrifuges. The final product, containing about 50 - 60% solids, is
thixotropic in nature, thus requiring further drying by mixing with flyash and
lime to make it suitable for disposal. It has little potential as a salable
byproduct.
Figure 5 shows a system with insitu forced oxidation (IF0) in which the
byproduct is calcium sulfate dihydrate, i.e. gypsum. Air is bubbled through
the absorber slurry in the integral recycle tank which oxidizes all sulfite
into sulfate. Since gypsum crystals are relatively large, the primary
dewatering is accomplished by simple hydrocyclones followed by secondary
dewatering in filters or centrifuges. The final product, containing about 90%
solids, is easy to handle and may either be disposed of as landfill or sold as
gypsum for wall board or cement manufacture. Salable gypsum, however, requires
gypsum washing during secondary dewatering to remove soluble salts such as
chlorides, thus resulting in a water blowdown to purge these salts from the
8A-73
-------
system.9 The water blowdown is treated in a waste water treatment system prior
to disposal.
The reliability of a single train FGD system is first and foremost dependent on
sound process design and design of the S02 absorber which is an integral part
cf the process. Avoiding plugging and scale is absolutely essential for
reliability as discussed earlier. In this section important aspects of spray
tower design, material of construction, process controls and equipment
redundancy requirements are discussed.
OPEN SPRAY TOWER DESIGN:
The spray tower can be a very efficient mass transfer device. However, to
achieve high performance requires a very intimate gas liquid contact so that no
channelling of gas is possible. Since positive control can be exercised on
liquid distribution through good hydraulic design, the key to achieving
intimate gas liquid contact is through uniform liquid distribution. Each spray
stage must have sufficient number of nozzles to cover the entire cross-section
of the spray tower in a highly overlapping manner. A good design should
provide a spray pattern overlap of over 150% within three feet from the tip of
the spray nozzles. Such an intense overlapping spray pattern has a very
beneficial effect on gas distribution as revealed by wet model tests.
GAS DISTRIBUTION:
An open spray tower does not need any gas distribution aids. The key to
getting good gas distribution, however, is uniform liquid distribution across
the cross-section of the tower as mentioned above. With uniform liquid
distribution each spray stage becomes an effective screen, and the resistance
created by dissipation of the enortrcc; energy of liquid against the rising gas
leads to good gas distribution. "Uii; beneficial effect of sprays on gas
distribution was first confirmed in a 1/15 scale model study in 1974 before
building the first commercial open spray tower.4 Confirmation of this
phenomenon has made it possible to confidently scale-up the open spray tower to
gigantic sizes while achieving over 98% SO2 removal efficiencies and extremely
low liquid mist carryover. Recently, mist eliminator carryover tests at the
645 MW single train open spray tower at EPZ's Amercentrale #8 unit in The
Netherlands were performed by the independent testing company, TuV. Using a
8A-74
-------
single stage mist eliminator, the mist values obtained ranged between 1.1 to
4mg/NMs with an average value of 2.5 mg/NM3 0.001 grains/SCFD). These
extremely low values were measured at gas velocities of 10.7 to 12.1
feet/second. The high SO2 removal efficiencies and low mist carryover values
are clear indications that spray-induced gas distribution in a well designed
open spray tower is sufficient for high performance.
Additional scale model tests were performed during 1990 and 1991 to study the
effect of tower outlet geometry and a perforated tray on gas distribution. The
results are summarized in Figure 6. With four spray banks in operation, the
standard deviation of 67 velocity points across the lower face of the mist
eliminator, (location "A"), ranged between 23.2% to 26.0% while gas outlet
geometry was changed from completely open to 45* and 60° conical discharge to
90° side discharge. These results clearly demonstrate that the effect of
sprays on gas distribution is very strong and practically unaffected by outlet
geometry. This controlling influence of sprays is further illustrated by the
results of installing a perforated tray with 46.6% open area above the first
spray stage. The standard deviation with a tray in place was 23.1%, indicating
no further improvement by a tray on spray-induced gas distribution of the open
spray tower.
Tests were also performed to determine the gas distribution at locations "B"
and "C", which are between the two stages of the mist eliminator system and on
top of the second stage mist eliminator, respectively. A standard deviation of
25.1 is obtained at location "B", which is comparable to that at location "A".
This is reasonable when considering that the first stage mist eliminator had
2.5 inch spaced continuous blades and a nominal resistance to gas flow of ^bout
0.1 inch WC. Therefore, the prevailing gas distribution at location "A" was
captured and displayed at location "B". As expected, some influence of the
outlet geometry on gas distribution begins to show at location "C" which is
above the second stage mist eliminator. The standard deviation at this
location varied from 11.3% for a completely open discharge to 16.7% for a 90°
side discharge. However, this variation is still small and within the norms of
good gas distribution.
The above data clearly show that a well designed, open spray tower has
excellent gas distribution and needs no gas distribution aids which can
compromise its reliability. It further points out that the gas outlet geometry
does not have a significant influence on gas distribution in the spray zone.
8A-75
-------
HASS TRANSFER:
The open spray tower becomes a very efficient SCte absorber once proper liquid
and gas distribution is achieved as discussed above. The mechanism of mass
transfer can best be explained by the classical film theory which envisages
very thin gas and liquid films on either side of the gas-liquid interface.
Turbulence in these films dies out so that molecular diffusion is the only
mechanism by which mass transfer occurs across the gas-liquid interface
boundary. The resistance to mass transfer is inversely proportional to the
film thicknesses. Since diffusion in gas is much faster than in liquid,
efficient gas absorption can best be achieved by minimizing diffusion through
liquid. In the case of SCte absorption in water, this means keeping the free
SO2 concentration at the liquid surface to a negligible level. Binding SO2
with a chemical reaction at the interface is one sure way to achieve this. In
a limestone slurry system, the reacting species are essentially calcium ions
generated through dissolution of limestone. Since limestone has low solubility
in water and also dissolves at a slow rate, it is necessary to have a large
liquid-to-gas ratio as well as a large amount of limestone surface area to meet
the demand for calcium ions at the gas-liquid interface for reaction with SO2.
Failure to meet this demand can result in inefficient SO2 absorption. The
higher the SO2 content, the higher the demand for calcium ions, hence, the
higher the requirement for liquid-to-gas ratio as well as limestone surface
area. For this reason, fine grinding of limestone becomes important for high
sulfur coal applications.
Theoretical modeling of mass transfer in commercial absorbers is possible, but
has little practical value, because of indeterminates such as gas and liquid
film thicknesses and gas-liquid interface area. The mass transfer film
thicknesses and gas-liquid interface area in a commercial absorber are
impossible to measure directly, thus requiring gross assumptions to fit the
experimental data and therefore making the theoretical models highly
subjective. For this reason, it is customary to describe industrial gas
absorption processes in analytical models based upon volumetric mass transfer
coefficients, which embody the indeterminates peculiar to a given absorber.
For an open spray tower, an analytical model was developed in 1970 which has
8A-76
-------
been refined and calibrated on numerous comnercial spray towers. The model is
briefly described here:
R = G (Yi - Y2> = Kga * V - (Yi-Y2)/Ln XI (1)
Y2
R - Rate of SO2 absorption (Moles/Hr)
G - Flue gas flow rate (Moles/Hr)
*1, *2 » Absorber inlet and outlet SO2 concentrations (Hole fraction)
XI « log mean driving force for mass transfer
Y2
v - effective volume of the absorber
Xga - overall volumetric mass transfer coefficient Moles/Hr. Ft3
Equation (1) above is analytical in nature and involves no assumptions, other
than that the concentration of free SO2 and therefore its back pressure at
gas-liquid interface is negligible. This condition, of course, is a
prerequisite for an efficient absorber design. Equation 1 can be rearranged to
reflect efficiency (E) as follows:
Xga = ^ . Ln (1/1-E) (2)
Equation 2 can be used to predict efficiency when the overall mass transfer
coefficient for a given absorber is known. The overall mass transfer
coefficient is a function of the absorber design and must be experimentally
obtained from pilot and full size operations. For 6E open spray towers, the
mass transfer coefficient has been correlated to three most sensitive and
important variables, namely...gas velocity, liquid density and inlet SO2
concentration as follows
Xga = C . U® . Ln/YlP (3)
Where
C - proportionality constant
U - gas velocity in feet/sec
L - liquid spray density gpm/ft3 of tower
a, n, p - correlation coefficients
I
Other variables such as pH and droplet size are not correlated. The pH cannot
be used as a variable if condition of negligible SO2 back pressure oust be met
for efficient absorption. Correlation with droplet size is not useful because
of significant collision and agglomeration that occurs in the spray zone.
8A-77
-------
MATERIALS OF CONSTRUCTION:
The WFGD process creates a very aggressive environment for corrosion and
abrasion. The flue gas path from inlet of the absorber to stack discharge must
be protected against acid attack due to adiabatic cooling and saturation of the
gas. Equipment in the gas path includes isolation dampers, absorber inlet
duct, absorber and its internals, outlet ducts, reheater system and stack. All
slurry handling parts are subject to both corrosive and abrasive attack. These
include absorber spray zones, tanks, agitators, pumps, pipes, valves and all
dewatering equipment. The gas flow path presents the greatest challenge
because of the large surfaces and weights involved. The choice is between
carbon steel with protective linings or alloy materials. Alloy selection is
further complicated by the presence of varying amounts of chlorides in the
scrubbing liquor. Cost of the WFGD system is greatly influenced by the choice
of materials.
Lined carbon steel results in the lowest cost. Industry experience with
linings, however, has been a mixed bag of successes and failures. Lining
carbon steel is both an art and a science. Careful attention must be paid to
such details as material formulations, surface preparation, application
technique, curing, temperature and humidity limitations, slurry impingement
intensities, inspection and timely repairs. Therefore, it is not surprising to
find widely varying industry experience on linings. Nonetheless, when linings
are properly selected and applied, lined carbon steel does provide excellent
service, as proven by more than 30,000 MW of GE design FGD systems. Figure 7
shows the history of various linings in GE FGD systems.
GE has also conducted extensive studies on corrosion of alloys in the simulated
FGD environment.6 Materials tested included 316 and 317 stainless steels,
high nickel alloys such as Alloy G, Inconel 625 and Hastalloy G. Material
samples were prepared to provide stressed conditions, welds and heat affected
zones as well as base metals. These samples were exposed to slurry
submergence, gas-liquid interface and saturated gas environment. The pH of the
slurry was controlled by adjusting SO2 injection and lime slurry additions.
The dissolved chlorides were controlled at three levels of 10,000, 20,000 and
30,000 PPM. Figure 8 shows the results of a three month study. Although
general corrosion rates of all materials were within tolerable limits, the
pitting corrosion varied significantly and was greatly accelerated by
increasing chloride levels. Only Inconel-625 shows resistance to pitting
8A-78
-------
corrosion. Therefore, when chlorides are present, Inconel or a similar alloy
such as Hastalloy C would appear to be the only alloys which can be relied upon
for long life. To reduce the cost of alloy construction, wallpapering of
carbon steel with thin alloy sheets has recently been applied successfully,
particularly in the outlet ducts where no abrasive environment exists.
Successful application of alloys also requires very stringent quality assurance
specifications and attention to such details as welding, prevention of surface
contamination and careful post-installation handling to avoid damage. Since
alloy wallpaper is not bonded to the carbon surface, any incursion of fluid
through the wallpaper due to any localized pitting or cracks can cause
widespread corrosion of carbon steel. Therefore, any wallpaper surface should
be frequently inspected for early detection of damage and repair.
PROCESS CONTROLS SIMPLIFIED:
The critical process controls revolve around pH, slurry density and water
balance. Optimum pH of the absorber slurry is essential to keeping SO2 removal
efficiency at peak performance. Absorber slurry density control keeps
suspended solids at optimum level to prevent scaling and to provide crystal
growth for maximum dewatering. Water balance control avoids any unnecessary
blowdown and maximizes mist eliminator washing for trouble-free operations.
These three process controls are fully automated, along with other operations,
and yet are kept simple to allow easy integration with boiler operation. The
Central Processing Unit (CPU) uses boiler load and absorber inlet SO2
concentration signals to anticipate the demand for reagent feed and the need
for product withdrawal. In both cases, the anticipatory feed forward and bleed
signals are subservient to pH and density respectively. The master pH signal
ultimately determines the amount of reagent fed to the absorber. Similarly,
the master density control signal determines the ultimate amount of bleed to be
withdrawn. The system can also be operated with pH and density controls only;
however, feed forward helps minimize wide fluctuations in pH and density.
Water balance is automatically controlled by total plant water inventory
management. This is simply accomplished by the CPU with level signals from all
storage tanks. Any change from the designed total water inventory determines
the permissible make-up water additions. Since most of the makeup water is
used for mist eliminator washing, the wash inventory signal is used to
determine the frequency of mist eliminator washing, hence makeup water
addition. Again the feed forward signal is used to determine washing frequency
of the mist eliminator which is further adjusted by water inventory signals.
8A-79
-------
Since the quantity of makeup water is relatively small, the mist eliminator is
sequentially washed in ten to twelve segments at a fixed intensity of about
1.5 gpm/ft2 of cross-sectional area. The frequency of segmental wash, hence
water input, is automatically adjusted by feed forward and water inventory
signals. The bottom face of the mist eliminator which captures the bulk of the
slurry droplets is washed in a cyclical fashion as frequently as permissible
within the water balance constraints. The upper faces of the mist eliminator
need less washing, therefore, only one segment is washed after each cycle of
bottom face wash.
The simplified, automatic control of the above three critical process variables
is readily integrated with boiler controls and ensures trouble-free peak
performance of the FGD system.
SPARING OF AUXILIARIES:
All critical process operations must be secured through installed spares. An
installed spare spray stage and recycle pump must be provided in the absorber
to allow on-load isolation and maintenance of the recycle pumps without loss of
SO2 removal efficiency. This philosophy must be followed in other operations
of the system. All pumping systems, such as bleed, reagent feed, reclaimed
water, mist eliminator wash, filter feed, etc., must have an installed spare to
allow on-load maintenance. Dewatering and reagent preparation operations can
be secured by installed spares and/or through storage capacities. Certain
critical process control instruments such as pH meters and density meters must
also have installed spares. A warehouse inventory of spare parts must be kept
up-to-date to facilitate timely maintenance.
RELIABILITY OF THE SINGLE TRAIN SYSTEMS:
With single absorber modules, the flue gas path is greatly simplified due to
elimination of gas manifolds, most dampers and associated controls. When
simplicity of the single train system is combined with design principles
outlined in this paper, high reliability is readily achieved. These
expectations are being clearly demonstrated by GE designed FGD systems. Figure
9 shows a list of single train FGD systems that GE and its licensees around the
world have supplied. Although the single train design is a common practice
overseas, the first commercial application of the single train system in the
United States is slated for the Harrison Station of Allegheny Power Systems.
8A-80
-------
Each of the three 640 MW boilers at the Harrison Station will be equipped with
a GE single train absorber to remove a minimum of 98% SO2 removal while firing
4.5% sulfur coal and using high magnesium lime reagent. One of these absorbers
will be built in the base of a 1,000 foot high chimney.
The performance record of some key single train FGD systems is shown in Figures
10 and 11. Sulfur dioxide removal efficiencies of 95-98% are being achieved
using limestone reagent with both low and high sulfur coals while producing
salable gypsum. Materials of construction are traditional carbon steel with
rubber or flakeglass linings for absorber and ducts. The reliability factor,
expressed as one minus ratio of forced outage hours over boiler operating
hours, is close to 100%. This remarkable performance of the single train
system is achieved while providing flue gas reheat. The reheat systems in the
U.S. have been a particular source of problems. Recently, the trend in the
U.S. is toward wet stacks which eliminate the troublesome reheat system. This
further simplifies the single train systems and ensures reliability.
CONCLUSION:
The state-of-the-art of flue gas desulfurization has been greatly enhanced by
the introduction of single train systems. The inherent simplicity of the open
spray tower has made it possible to confidently scale-up FGD systems to large
sizes. The dramatic cost savings coupled with excellent performance and high
reliability has made the single train open spray tower FGD system the ideal
choice for controlling sulfur dioxide emissions from power plants.
8A-81
-------
ACKNOWLEDGMENTS: Author is grateful to the following individuals and their
organizations for providing the performance data on single train open spray
tower FGD systems presented in this paper.
Mr. N.A. Doets, Director, EPZ, Geertruidenberg, The Netherlands
Mr. P. Kneissle, Director, OKA, Riedersbach, Austria
Mr. W. Schaller, Director, STEWEAG, Graz, Austria
Mr. Sunao Sato, Director, Thermal Power Department, EPDC, Tokyo, Japan
Mr. H. Weiler, Head of Dept. of Environmental Tech., STEAG, Essen, Germany
Mr. R. Hewitt, Supevisor-Engineering, TUELECTRIC, Dallas, Texas, USA
REFERENCES:
1. R. Lessing, "The development of a process of flue-gas washing without
effluent", Journal of the Society of Chemical Industry, Transaction and
Communications, November 1938
2. A. Saleem, D. Harrison, N. Sekhar, "S02 removal by limestone slurry in a
spray tower", Proceedings of the Second International Lime/Limestone
Wet-Scrubbing Symposium, November, 1971, New Orleans, Louisiana. Sponsored
by the U.S. EPA.
3. R.H. Borgwardt, "Increasing limestone utilization in FGD scrubbers", 68th
annual meeting of the American Institute of Chemical Engineers, Nov 16,
1975
4. A. Saleem, "Spray Tower: The Workhorse of Flue Gas Desulfurization",
POWER, October 1980
5. R. Hewitt, A. Saleem, "Operating and Maintenance Experience of the World's
Largest Spray Tower Scrubbers", 6th Symposium on Flue Gas Desulfurization,
October 1980, Houston, Texas. Sponsored by the U.S. EPA
6. N.L. Koshkin and M.C. Chen, "Wet Flue Gas Desulfurization System Alloy
Corrosion Test Program Test Results", presented at The 7th Symposium on
Flue Gas Desulfurization, May, 1982, Hollywood, Florida. Sponsored by the
U.S. EPA/EPRI.
7. W.H.P. Goossens and P.C. VanLoon, "First Year Operational Experience with
the Largest Single Absorber FGD in Europe," Proceedings of the 8th World
Clean Air Congress, 1989, The Hague, The Netherlands, 11-15 September 1989
8. S.M. Katzberger, C.E. Dene and R.J. Keeth, "FGD Retrofit Design
Improvements" presented at 1990 SO2 Control Symposium, May 8 - 11, 1990,
New Orleans, Louisiana. Sponsored by EPRI/EPA/DOE
9. A. Saleem, "GE's Worldwide Experience with IFO Based Gypsum Producing FGD
Systems", presented at The Second International Conference on FGD and
Chemical Gypsum, May 12 - 15, 1991, Toronto, Canada. Sponsored by ORTECH
of CANADA.
8A-82
-------
Figure l:L06isncs of various reagekts and byproducts
«oBg flM&STOgBiL
15 TOMS PER HOUR OF S02 REMOVED
REQUIRES:
15 TOMS OF LIME
OR
30 TONS OF LIMESTONE
OR
8 TONS OF AmOKIA
PRODUCES:
90 TONS OF 50% SOLID SLUOGE
OR
45 TONS OF GYPSUM
OR
23 TONS OF ACID
OR
7.5 TONS OF SULFUR
OR
31 TONS OF AWONIUM SULFATE
Figure 2: Simplified Block Diagram
of the Wet Limestone FGP Process
XliCK
tt.
ST
8A-83
-------
Figure 3: BASIC chemical reactions of
S02 ABSORPTION WTTH LIMESTONE
1.
SOz -f HzO
2.
CaCOj +
HiSOs
3.
CaSOs +
tfb
4.
CaSOs +
%HzO
S.
CaSOt +
2 HzO
6.
C»V>» 4.
feSOs
feSOj ABSORPTION
-> CiSOj + CQj + H2O NEUTRALIZATION
-> CaSO* — OXIDATION
->
-> CaSOs . tytiO
, CRYSTALLIZATION
-> CiSO* . 2HzO
=? ea(HSOs)* pH CONTROLLED
"CURE 4: GE NFO FLUE GAS DESIILFIIRIZATION PROPFSS
AMD
SULFUR DtOXDE
ABSORPTION
I--0
REAOENT PREPARATION
80UD8 DEWATBWQ
8A-84
-------
FIGURE 5: fit
IFO FLUE GAS DESULFURIZATTON PROCESS
OAS HANDLING
AND
SULFUR DIOXIDE
ABSORPTION
GYPSUM DEWATBBNG
R£A(£MT PREPARATION
FlGttRE 6: &&
A- \
A \
/
OUTLET COHORT
c
C
c
C
MHL
B
Mtt
B
MW.
B
MHL
B
1HMX.
A
M KL
A
W MX.
A
<
I
>
[aaaaaa]
A/\/\/\/\/\
r\/\/\/\/\/\
|AAAAAA|
LOCATION: A
(67 VEL. POINTS]
23.3
26.7
24.6
23.6
LOCATION: A WITH
*6.6% oea TRAY
(67 yel. points)
23.1
LOCATION: B
(68 TEL. POINTS)
2S.1
LOCATION: C
(60 TEL. POINTS)
11.3
11.2
12.6
16.7
8A-85
-------
FIGURE 7: GE EXPERIENCE WITH t TNTMfiS
AREA
MATERIAL
NUMBER
OF
MOOUIES
HW
SO. FEET
(X 1000)
YEARS
IN
SERVICE
1. SPRAY ZONE
• RUBBER LINING
39
9200
257
9
• FLAKESLASS PLASTIC
95
12000
32S
17
2. ABOVE SPRAY ZONE
• RUBBER LUUNB.
39
9200
141
9
• FLAKESLASS PLASTIC
95
12000
180
17
3. SUPPORT BEARS
• RUBBER LIMINE
149
23000
30
17
4. RECYCLE TAMC
• RUBBER LINING
30
9200
171
6
• RATE GLASS PLASTIC
9S
13000
195
17
S. INLET DUCT
¦ GLASS BLOCK
IB
3000
2S
S
• FLAKESLASS PLASTIC
22
4750
24
IS
6. OUTLET DUCT
• GLASS BLOCK
IS
2400
103
s
• FLAKESLASS PLASTIC
35
8000
240
IS
FIGURE 8: GE STUDY ON CORROSION OF ALLOYS IN FGD ENVIRONMENT
DATA ON PITTING OCCURENCE, DENSITY AND RATE* AT pH 5.2
CI -
10.000 P
m
CI ¦
20.000 P
PH
CI -
30.000 PPM
ALLOY
POSITION
BM
w
KAZ
W
HAZ
BR
H
HAZ
INCONEL 625
U
N
HASTALLOT G
U
*
028
020
0
N
B
020
048
0
632
ALLOY 904L
U
0
0188
•156
040
0
9
•
•24
0
H
B
•
0
•20
•
•48
e
•
SS TYPE 317 IN
U
0
•32
0
948
052
0
e
N
•
0
•20
820
•80
e
•
B
04
0
0
012
012
0100
9
0
SS TYPE 316 L
U
•200
0
0
0200
0
0
• 116
_
H
•S6
0
•
¦
•64
040
0
a
B
•
08
024
•
04
COOTOH POSITION
u ¦ ABOVE slurry
H - AT SLURRY/VAPOR INTERFACES
B - IWERSION IN SLURRY
commai
W . VELD
HAZ . HEAT AFFECTED ZCME
BR - BASE METAL
gTTDntSTTT
<10/SQ.IN.
ie-iooo/sq.M.
>looo/sq.lN.
• - LOCALLY >1000/SQ.IN.
NUMERICAL VALUES IIOICATE ESTMKTES PITTING RATE IN HllS/YR BASED ON LINEAR PROJECTIONS FROM 90 DAYS TESTS
8A-86
-------
Figure 9: installations utilizing single train spray towers of ge design
Size
UTILITY
Unit
MUE
* Sulfur Coal
%S02 Removal
Start Up
Austria
OKA
Riedersbach II
ISO
2.0 - 4.0
95 -
9B
1986
STEWEAG
Hellaeh
220
O.S
98 -
99
1986
VIENNA
Sinsnering
380
2.0 (oil)
97
1991
Japan
EPOC
Matsushima 11
500
1.3
97 -
99
1980
EPDC
Takehara #3
700
1.0
96 -
98
1983
KEC
Hinato
ISO
1.7
CEPC
Shin Onoda #1
SOO
1.0
96 -
9B
1985
CEPC
Shin Onoda #2
SOO
1.0
96 -
98
1985
EPOC
Ishikawa #1
1S6
1.0
96 -
98
1986
EPOC
Ishikawa #2
1S6
1.0
96 -
98
1987
HEC
Tsuruga
SOO
1.0
95
1991
Netherlands
EPZ
Amercentrale fS
645
l.S
90 -
94
1987
EPZ
Aaerccntrale #9
645
l.S
90
1992
AHSTERDAH
UNA 8
650
1.5
90
1992
United
States
APS
Harrison #1
640
4.5
98
1995
APS
Harrison 12
640
4.5
J9
1995
APS
Harrison #3
640
4.5
9B
1995
United
Kingdom
POWERSEN
Ratcliffe #1
SOO
3.0
90
1994
POWERGEN
Ratcliffe 12
500
3.0
90
1994
POUERGEN
Ratcliffe #3
SOO
3.0
90
1995
POUERGEN
Ratcliffe #4
SOO
3.0
90
1995
West Germany
RWE
Niederaussen
9x300
0.9
90
1987
STEAG
Hcrne IV
SOO
1.5
95.6
1989
STEAG
Charlottenburg
225
1.5
90
1987
BEUAG
Oberhavel
200
1.5
96
1988
HEW
Wedel
260
l.S
94
1987
Taiwan
TPC
Hsinta fl
500
1.2
90
1991
TPC
Hsinta 12
SOO
1.2
90
1992
All units except Harrison use limestone reagent and produce gypsum byproduct. Harrison uses high
MAGNESIUM LIME REAGENT.
8A-87
-------
Figure 10: Performance, Design and Operational Data of
Some Large Single Train Spray Tomer FGD Systems
00
*
80
Station
Hitiuihlaa II
ISOOHW)
IlK, Jipm
Tikihirt 1)
(700M)
CpDC, Jipin
RlldirsblCh tl
(l«MU)
OKA, Austria
Nil Itch 11
m ohw)
SUWAC, Austria
AMrcintril II
(49SHW)
CPJ, Hilhirlinds
Him* 14
(SOMMI
STCAS, itratny
Mont letllo 11 *
(7S0HU)
TUClCCTRIt, Tim
S02 Efficiency 1
17 • II
II - II
IS - 18
II - 99
90 - 14
95.1
10 - 95
Abierbir Inlit SOt PPH
)00 • S20
214 • (M
1749-JJJJ
<00 - 700
610 - 1049
1124 • 1S74
400 • (50
Abierbir Shall KitirUI
CS-F6L
CS-FGL
CS-RL
CS-Rl
CS-RL
CS-RL
CS - FCL
Abtorbor Intirmlt Pipit
SUS JUL
SUS Jill
CS-RL
CS-RL
CS-RL
CS-RL
CS-RL •> FRP
AbiorUr Rtcycl* Pvap
VST 130A
CS-RL
Alloy
Alloy
CS-RL
Alloy
CS-RL •> Alloy
lapollort
CSt Pr«-Krit«
--> c-m
Add Irlck
Abiorbor Outlst Duett
Stick Hittrlil
cs-roL
Add Irlck
cs-rci
Acid Irlck
CS-FGL
Add Irlck
CS-FCL
Acid Irlck
CS-RL
Add Irlck
CS-fCL
Add Irlck
Opiritort hr Shift
l.J
l.S
t
t
I
:
2,1
Routlni Hitntinmc* Cr«w
pir Shift
SJUI(l.S)
SAH(l.S)
SM
SWI (2)
SAN (O.S)
SAM
SAN (2)
CS-FGLi Cariom Stccl with Flaki Glasi Riihforcco Plastic Lihino
CS-Rlt Carioh Stccl with Ruiscr Lininq
SAH: Sahc as Boilim Haihtimancc Crcu
* HONTICCLLO 13 ts A THRCC TRAIN SYSTCH AND IHCLUOCO HERE FOR RIFCRtNCC
-------
Figure 11: Reliability Record of Some Large
Single Train Spray Tomer FGD Systems
Station
Htlsuihlai II (SOOHW)
EPOC, Japan
takchara ») (700HW)
CPOC, Jipin
Rlldtribich #2 (160HW)
OKA, AuttrU
BOH
FOH
RF
BOH
FOH
(IF
BOH
FOH
RF
>990
6851
0
IOOX
7))8
0
loot
SSS9
0
IOOX
1989
7)67
0
loot
870)
0
100X
5709
178
97X
1988
S902
0
IOOX
7216
0
IOOX
4557
0
IOOX
1987
7744
0
IOOX
760)
0
IOOX
4587
0
IOOX
1988
7705
0
IOOX
7ZS5
0
IOOX
....
....
....
Station
NtlllCh O (220HW)
STCWAG, Austria
Aatrctntral* 18 (645HU)
CPZ, the Ntthcrlandi
Hern* 14 (500HV)
STFAC, Gtrnany
Konltcollo »" (7S0HU)
miECIRIC, ttxai
BOH
FOH
RF
BOH
FOH RF
BOH FOH
RF
BOH
FOH
RF
1990
4400
0
IOOX
7508
122 )8X
8760 0
IOOX
7756
0
1001
1989
4400
0
IOOX
' 8422
606 )3X
7480
4
99.91
1988
4400
0
IOOX
....
6962
14
99.8*
1987
4400
0
IOOX
....
....
7026
)1
99.8*
1986
4400
0
IOOX
I
•
•
•
•
•
¦
•
767)
56
99.)*
BOH: Boiler Operative Hours
FOH: Forced Outaoe Hours Oue to FGO
RF: Reliability Factor; (1-F0H/B0H)* 100
• HOHTICELLO #3 IS a THREE TRAIN SYSTEM AND IHCLUDEO HERE FOR REFERENCE
-------
8 A-90
-------
SELECTING THE FGD PROCESS AND SIX YEARS OF
OPERATING EXPERIENCE IN UNIT 5 OF THE
ALTBACH-DEIZ3SAU NECKARWERKE POWER STATION
R. Maule
P. Necker
M. Straus
Neckarwerke, Elektrizitatsversorgungs-AG
Esslingen
Kuferstrasse 2
7300 Esslingen
Federal Republic of Germany
S. Negrea
Noell, Inc.
2411 Dulles Comer Park, Suite 410
Herndon, VA 22071
3780 Kilroy Airport Way, Suite 350
Long Beach, CA 90806
Preceding page blank
8A-91
-------
8A-92
-------
ABSTRACT
In 1981, this power station was evaluating various FGD processes suitable to meet stringent German
regulations. The paper discusses the various criteria which led to the selection of their SOj removal
system.
The paper covers some of the operational experiences gained after six years of using an FGD plant with
wet limestone, with gypsum by-product, using a two-stage advanced process.
Among other topics are experience with limestone as reagent, performance of alloys, contribution of
reheating systems, and utilization of the gypsum by-product in the construction industry.
The paper the operating results, as well as the costs of operating and maintaining this facility.
Recommendations are made for future users in areas such as material selection, spare module
scrubbers, based on the performance of this FGD plant.
Preceding page blank
8A-93
-------
INTRODUCTION
The Neckarwerke, AG, based in Esslingen is a Regional Utility Company in southern Germany near
Stuttgart. In 1990, the Neckarwerke had a network peak production of 1,456 MW. In 1990,
approximately 8.4 billion kWh were supplied, of which 50% were used by various industrial users.
Approximately 95% of this electric power was generated in our power plants. The annual turn-over
totaled l.S billion DM (approximately 857 million US dollars).
The distribution of various energy sources in the power generation of Neckarwerke and the comparison
with the similar total for the "old" Federal Republic of Germany for the year 1990 is shown in
Figure 1. This data gives a good understanding of the energy supply condition for the Neckarwerke
(NW) and the position NW had in the "old" Federal Republic of Germany in 1990.
At the end of the seventies, NW realized that the planned extension of the nuclear energy plant could
not be completed as fast as originally planned. Therefore, in addition to the planning of the nuclear
power plant (which originally was scheduled for completion in 1984/85 [actual date was 1989], it began
in 1979/80 plans for a coal power plant (Unit 5) with 420 MW net. This plant was scheduled to go
into operation in 1985. Existing regulations of 1981 required that a flue gas desulphurization plant
must be installed only for units with a firing capacity of more than > 4 TJ/h. The new plant was
designed for a firing heat capacity of 3.9 TJ/h. Nevertheless, it was decided, in order to ensure a
better environmental acceptance of this power plant, to voluntarily install the flue gas desulphurization
system for an initial partial gas flow of 40%. At that time, there were only 3 flue gas desulphurization
plants in operation in Germany. As can be seen in Figure 2, the regulations became continuously
stricter during the beginning of the 80*s. This was not a good situation, neither for the plant operators,
who at that particular time were constructing plants, nor for the FGD plant suppliers, who continuously
were confronted with new regulatory demands.
In 1983, the large-size power plant regulations stipulated that within five years old plants with a
remaining life of more than 30,000 hours have to be retrofitted with FGD-sy stems to comply with the
same emission limits required for new plants.
Today, the tendency in approving new plants is that, although the requirements, quantitatively
speaking, were not changed to this date by the Federal Environmental Protection Laws with the thought
. . according to the status of technology . . . ", yet they are much stricter with the emission limits.
8A-94
-------
DEVELOPMENT OF F GD-TECHNOLOG Y IN GERMANY
The basic development phases of FGD-technology in Germany are shown in Figure 3.
The first plants worked with lime as an absorption agent and calcium-sulfite/sulfate-slurry was the
end-product. Since only partial desulphurization was required, the reheating of the clean outlet gas
could be accomplished by mixing outlet gas with untreated inlet gas. The desulphurization was still
very inefficient.
Due to the difficulty with waste disposal of this type of end-product and due to Federal Government
Regulations for minimizing waste disposal, and the requirement to produce usable end-products, the
next development step was to oxidize the calcium-sulfite in order to obtain calcium-sulfate (i.e.
gypsum); this was first done in oxidation vessels, and was later integrated by in-situ oxidation into the
scrubber tower.
In Germany, similar to the United States, lime is in comparison to limestone relatively expensive.
Therefore, through research and experience, it was determined that lime as reagent could be replaced
by limestone, and today limestone is being used in the overwhelming majority of German FGD plants.
Because of the necessity at that time to increase the desulphurization performance to 85 %, the portion
of flue gas to be desulphurized had to be increased, and it became necessary to build reheating systems
for the flue gas. The reheating systems were mostly developed as regenerative heat exchangers because
of the familiarity of the power industry with these systems.
Due to the tendency of increasing the desulphurization efficiency towards 90% and higher and the
associated requirement to maximize the desulphurized flue gas stream, we will see in the future flue
gas desulphurization systems a minimization in leakage at regenerative heat exchangers, that is,
leakage-free systems for reheating.
THE DECISION FOR FLUE GAS DESULPHURIZATION OF
NECKARWERKE AG IN THE YEAR 1981
The decision regarding the plant concept for Unit 5 was made in accordance with the following criteria:
• to ensure high S(>2 removal efficiency
• to achieve a simplified operating procedure with clearly defined process parameters
• to ensure safe operation through simple construction
• to realize low operating costs by using simple reagents
• to satisfy a safe disposal by utilization of the end-product (gypsum)
• to use known power plant-type equipment and to standardize plant elements
• If repairs are necessary, the usual power plant repair techniques should be sufficient.
• to have a high degree of automatization according to power plant standards
With this requirement profile in mind, we decided on the following plant concept:
8A-95
-------
PROCESS
The two-loop NOELL-KRC process, with separate optimum regions for the pH-values was adopted.
In the quencher area for oxidation and limestone dissolution a pH around 4 and in the absorber loop
a pH of approximately 6 for SO2 removal was selected. The oxidation was integrated in the scrubber.
REAGENT PRODUCT LIMESTONE
Limestone in Germany is considerably cheaper than burnt lime (ratio 65 : 1SS). The advantage of
transporting calcined lime, due to the smaller volume (S6: 100), cannot overcome the price difference.
The cost of the limestone is independent of the primary energy price, which is an entirely different
situation for the calcined lime. TTie result was a low and stable operational cost.
TOTAL PLANT CONCEPT
• Since it was planned to build the system in two phases, 198S Phase I and 1987 Phase n,
Neckarwerke chose the erection of the plant with two lines.
• Stainless steel scrubbers with various alloy steel quality for various levels were selected.
• Stainless steel recycle pumps for the quencher and the absorber with the respectively
required material quality were chosen.
• The reheating takes place in the regenerative heat exchanger of the second line.
• The remaining droplet evaporation (untreated gas), after the scrubber, was planned with
untreated hot flue gas.
• Due to our climatic conditions and the regulations for noise protection, as well as to
provide better conditions for the plant maintenance, we decided for a total enclosure of
the FGD plant.
• The system is, to a great extent, fully automated.
• The FGD control room is integrated into the Main Control Center of the power plant.
The Unit S FGD plant schematic is shown on Figure 4 and the simplified process flow sheet can be
seen on Figure 5.
END-PRODUCT GYPSUM
Gypsum dewatering using the vacuum drum filter is
and further by production of gypsum briquettes.
Commercial, high quality gypsum is the end-product
disposal safety.
followed by drying with untreated hot flue gas,
which guarantees a high level of potential waste
8A-96
-------
END-PRODUCT GYPSUM
The gypsum preparation simplified schematic is shown in Figure 6.
The FGD-plant was supplied by NOELL-KRC with a construction period of 2-1/2 years and was
brought into operation in 198S with the first line and in 1986 with the second line, one year earlier than
originally planned. Further details are found in References 2, 3, 4, 5, and 6 listed at the end of this
paper. The technical data is contained in Figure 7.
OPERATIONAL EXPERIENCE
We have now gained an experience of six years of operation with a total of more than 30,000 operating
hours with this type of FGD technology, which can be judged in total as very good. Following are
some details of our experience:
EXPERIENCE WITH THE REAGENT PRODUCT
Limestone is suitable as absorbent when applying the correct process. In order to get a large number
of suppliers, the purity of the limestone was varied in tests. This, however, was not effective. A too
high inert portion in the limestone leads to problems during dewatering of gypsum. For a trouble-free
operation with high quality commercial gypsum as by-product, limestone with approximately 97%
purity is recommended. The particle size distribution should be as follows: 80% < 63 /im and
90% < 90 fjLm respectively.
EXPERIENCE WITH THE PROCESS
The two-loop-system with clear, controllable operating conditions for the absorber (high pH-value) and
for the quencher loop (low pH-value) has proven to be effective. If the pH-values are maintained and
the spray nozzles are correctly directed, deposits are prevented. The SO2 removal efficiency in the
scrubber is higher than 95 %. Therefore, it was possible to reduce the original packing depth of the
Wet Film Contact (WFC) from three layers (3 feet) to one layer (12 inches) without any significant
reduction in removal efficiency. The blown-in air, integrated in the scrubber sump for sulfite to sulfate
oxidation was built on the lance exhaust system located before the agitator. Thus the oxidation air can
be turned off when the unit is not in operation and the power consumption can be reduced.
EXPERIENCE WITH THE SELECTION OF MATERIALS
The materials used are shown in Figure 8 including the variation of material qualities for the alloy steel
scrubber. Due to material considerations, it is of great importance to separate the high-chloride-
contents (lower loop) quencher circulation (approx. 15,000 - 30,000 ppm CI) and the (upper loop)
lower-chloride-contents absorber circulation (approx. 3,000 - 5,000 ppm CI).
8A-97
I
-------
EXPERIENCE WITH THE SELECTION OF MATERIALS
Otherwise, corrosion problems in the upper circulation loop could arise. In addition, deposit problems
can be expected even with high-quality materials. Deposits are therefore to be prevented by optimizing
the spray levels and the flue gas exhaust, especially in the so-called bowl area separating the lower
from the upper loop. For that reason an additional wash system with spray was built into both scrubber
towers. The underside of the bowl was plated with a high quality alloy steel material; Figure 8 and
Reference 7 deals in detail with our experience in this domain. The use of alloy steels for the
scrubbers and pumps has proven to be basically effective. Corrosion can be monitored and eliminated
by suitable methods for power plants during inspection shut-downs.
Reference 8 deals in detail with the partially very problematic experience of rubber-lining use by some
FGD units in Germany.
Also successful was the use of all-metal pumps. We used similar-built pumps to improve the repair
and spare parts inventory. The same material was used (HA 28.5 similar to 1.4464 to simplify storage
needs). The quencher pumps reached a life of approximately 25,000 h, the absorber pumps are
expected to have a life of 40,000 h. The slide ring seals are made without seal water.
EXPERIENCE WITH REHEATING
By using untreated inlet gas for droplet evaporation before the regenerative reheating system (Regavo),
we experienced corrosion problems at the Regavo due to temperatures under the acid dew point. To
increase the SO2 removal efficiency of the total system (reducing leakage) and to avoid this corrosion,
the droplet evaporation system and the gas valve system at the Regavo was rebuilt. Today, this system
is run with treated (outlet) gas. In Figure 9 the principle used today for reheating is presented. For
bolts and other connecting elements corrosion-proof materials are used (partially Hastelloy and F.R.P.
elements).
EXPERIENCES WITH THE END-PRODUCT GYPSUM
The dewatering of gypsum with a vacuum drum filter leaves a moisture contents of approx. 10 - 12%.
With band filter systems, the same values can be achieved with a simpler operation.
The drying of gypsum with flue gas has not proven to be effective. CI was transferred from flue gas
to gypsum and resulted in high CI concentration in gypsum. We then converted the system to hot air
drying (partial stream from the economizer) and today the system works without any problems. Figure
6 shows today's gypsum preparation.
The gypsum quality is excellent. The gypsum can be supplied without limitation for use in the
construction industry.
8A-98
-------
EXPERIENCES WITH THE PLANT DESIGN
The relatively expensive decision for the complete enclosure of the FGD plant was proven to be correct
and met its original purposes.
The centralization and integration of the total instrumentation and control system into the main control
center of the power plant and the complete automatization of the plant has proven to be excellent.
In conclusion we can say that, with the modifications mentioned above, we would build this FGD plant
again. The FGD process selected has proven to be a success; we are satisfied with its operation and
the high SO2 removal efficiency.
Under our present regulations it would be unthinkable to put an FGD plant into operation stepwise and
it is of course essential to treat from the beginning the entire gas stream. Today, progress has been
made where for one boiler a single absorber tower with one ID fan can possibly be designed for plants
with a capacity of up to 700 MW. The conceptual scheme for such a plant is shown in Figure 9.
Our decision for a process with limestone as reagent and gypsum as end-product, and the good
experiences with the selected FGD process performance is totally confirmed by the market shares of
the various FGD plants built in Germany.
Reference 8 gives details about operational experiences with FGD-plants in the Federal Republic of
Germany.
From all selected alternative processes in Germany, 87% operate with limestone (or lime) and have
a gypsum by-product. Figure 10 shows the technology market distribution. The double-loop process
is represented in this market with a very significant share of approximately 20%. Recent developments
for FGD plants burning lignite with very high SO2 inlet loading to the FGD system, in the territory
previously known as "East Germany", have also selected this process.
COSTS FOR FLUE GAS DESULPHURIZATION
Investment Costs for Neckarwerke FGD Unit 5
The costs for FGD Unit 5 including interest on construction loans totaled 135 Million DM
(approximately 77 million US dollars). The breakdown of the respective contract work sections is
shown in Figure 11. With a net output of 420 MW, this results in specific investment costs of 320
DM/kW (approximately 183 $/kW). In Figure 12 the operating costs for full-load operation are
summarized. This table shows the specific items which make up the total operational costs. The main
costs are generated by the pressure drop produced by the FGD plant. This corresponds to a share of
1.4% of the gross power output of the unit. The use of lime instead of limestone would increase the
operational hourly costs by 16% due to the higher lime price (155 : 65), without the benefit of any
operational advantages.
For the operation of this FGD-plant a staff of 12 people is necessary (shift personnel, maintenance and
repair).
8A-99
-------
COSTS FOR FLUE GAS DESULPHURIZATION
For the calculation of capital investment costs, assumption of amortization is necessary. We believe
that the period of amortization for the FGD-plant must be considerably less than the period which is
usually applied for the other power plant components. Figure 13 shows the additional costs for the
kWh due to the FGD-system operation in relation to the annual operation hours and an amortization
period of 10 and 15 years. The additional costs by the operation of the FGD-system are between 1.5 -
2.5 pf/kWh (approximately 0.8 to 1.4c/kWh or 8 to 14 mills/kWh) for the average load of Unit 5.
It is to be considered that the FGD Unit 5 was designed as a new plant together with the new
construction of the total power plant. The labor share (German wage scale) is approximately 0.05
Pf/kWh (0.03C/kWh or .3 mills/kWh) or approx. 3% of total costs. Maintenance costs, as an average
of prior six years, are approximately 2% of investment costs; future average maintenance costs are
expected to be approximately 3-4%.
Based on the positive experience with our FGD-system for Unit 5, we estimate savings possibilities of
approximately 15 to 20% at new plants with the concept of using only one FGD line in lieu of our
present two-line system.
In addition to the FGD-system, the first large-scale DeNOx-system outside Japan was also installed in
Unit 5. This means that 33% of the total investment volume for Unit 5, which totaled approx. 900
million DM (approx. 515 million US dollars), was expended for environmental protection. Figure 14
shows the percentage distribution of investment costs for Unit 5.
SUCCESS OF ENVIRONMENTAL PROTECTION ACTIVITIES AT NECKARWERKE
The success gained with the installation of FGD plants in SO. emission reduction for our company is
most impressively shown by the emission values in Figure 15. Within five years, the construction of
the new plant Unit 5, the retrofit of two old plants with FGD-systems, and the commissioning of our
second nuclear power plant made it possible to lower the emission values by almost 90%. In addition
to cost mitigation, a considerable contribution is also provided by the nuclear energy in the total
emission reduction.
8A-100
-------
SUMMARY
The Neckarwerke AG decided in 1981 to build an FGD-system, Unit 5, at the power plant in Altbach-
Deizisau, with the following concept:
• Limestone as reagent
• . Gypsum as end-product
• Two-loop scrubber system
• Alloy steel scrubbers
• Fully-automated operation
After six years of operation of the FGD-system, we can confirm that the principle of the plant as built
by NOELL-KRC has been proven successful. Due to the favorable experience, Neckarwerke AG
would, today, decide on a one-line system with one scrubber tower and one I.D. fan (for the Unit and
the FGD system together) instead of a two-line system. Higher-quality stainless steel would be selected
for all areas of the absorbers but especially in the (lower) quencher area and in the area of the bowl.
During the six years of operation, we never experienced an interruption or non-availability of power
output due to the FGD-system. The FGD-system can, without any difficulty, be integrated into the
operation of the power plant. We are operating at an SO^ removal efficiency higher than 90%. The
S02 removal efficiency can be maintained in all operating situations.
Investment costs for the FGD system amounted to 127 million DM (approx. 72 million US dollars).
Depending on the annual hours of operation, FGD costs of operation amount to 1.5 - 2.0 Pf/kWh
(approx. 0.9 - 1.2C/kWh or 9 to 12 mills/kWh); the share of FGD operational costs is approx. 0.3
Pf/kWh (approx. 0.18C/kWh or 18 mills/kWh).
8A-101
-------
B. Forck. "The Emission Limitations for Power Plants in Europe." International ASME-VGB-
Joint Power-Generation Conference. Boston, MA. 1990.
R. Maier, P. Necker, J. H. Strauss, H. Hemming and E. Landgraf. "Erection and Function of
the Flue Gas System of the Power Plant Altback/Deizisau, Unit 5." VGB-Kraftwerstechnik 67.
Heft 4. April, 1987, pp 378-383.
R. Maule. "Selecting an Optimal SC>2 Reduction Program - A German Perspective." Forum on
T rpaft CW Options. May 1990. Research-Cottrell Companies USA, Sommerviiie, NJ.
P. Necker and J. H. Strauss. "Function and Costs of a Flue Gas Desulphurization System with
a Limestone-Gypsum Wet Process with Unit 5 of the Neckarwerke at Power Plant
Altbach/Deizisau as the Model." (in German) ECE-Workshop 1987, Esslingen.
B. Lehmann, P. Necker and J. H. Strauss. "Experience with the Flue Gas Cleaning Systems for
SO_ and NOx in the Power Plant of Neckarwerke Elektrizitaetsversorgungs-AG." (in German)
ECE-Se miliar 1991, Nuernberg
J. H. Strauss. "A Standard Solution Does Not Exist." (in German). Zeitschrift Enerpie 40. Nr.
2. July, 1988.
R. Maier, K. U. Buschmann und J. H. Strauss. "Experiences from Operation and Maintenance
after 26,000 Hours of Operation - FGD, Power Plant Altbach/Deizisau" (in German).
Vortragsband VGB-Conference "Kraftwerk und Umwelt 1991". Essen
P. Necker. "Development and Present Status of Flue Gas Desulphurization in Germany."
International ASME-VGB-Joint-Power-Generation Conference. Boston, MA. 1990
8A-102
-------
ELECTRICAL POWER GENERATION 1990
BY ENERGY RESOURCES
Neckarwerke WEST - GERMANY
Nuclear Energy
74,5 %
Nuclear Energy
38 %
Gas
Oil 2 %
Others
vV 1 %
A Hydro
4 %
Hard Coal
18,5 %
approx. 8,4 Billion KWh
peak output (1456 MW)
Souroa, BMWI. VDEW, DVQ
chased
Power
5,8 %
Lignite
20 %
Hard Coal
29 %
approx. 360 Billion KWh
peak output (62000 MW)
NW-492-EL:09.91:neo/zwl
Figure I.
-------
DEVELOPMENT OF THE LEGISLATION
REQUIREMENTS OF FLUE GAS
DESULPHURIZATION (FGD)
Year
Size of Unit
Emission Limits
1974
1977
1982
1983
1988
from 1990:
< 4 TJ/h (1111 MWth)
> 4 TJ/h
No FGD
1160 mg S02/m3
850 mg S02/m3
650 mg S02/m3
> 300 MWe,
New Plants
Old Plants
> 30 000 h Residual
Use
}
85 %-S02
Removal
Efficiency
max. 400 mg
S02/m3
Completion
Retrofit
Old Plants
Tendencies towards further
aggravation in obtaining approvals
(legal basis...
dynamization clause in the BlmSchG)
NW-494_2-EL:04.90:nec/zwi
Figure 2.
8A-104
-------
DEVELOPMENT STEPS IN
FLUE GAS DESULPHURIZATION
IN GERMANY
Start up
1977
1979
1982
85/86
Further
Development
Requirements
State Leg. (1974)
Federal Leg. (1983)
Reagent
CaO
CaC03
hi
Oxidation
without
with separate ox id.
integrated ox id.
Desulphurization
Flue gas volume %
Removal efficiency %
20
19
25
24
60
57
90
>85
98 - 100
> 90
Reheating
without
with
Leakage minimization
Leakage-free
End Produkt
Sulfite/Sulfate-slurry
Gypsum moist
Gypsum dried
Gypsum compacted
~
N W-498
*
~
*
:L:nec/zwl;9.01
Figure 3.
8A-105
-------
8
ION (UNTREATED GAS)
GYPSUM DRYER
vAISOMII 2
INLET GAS
5'/.
liakagi
POD I . ./
BOOSTER
FAN 2 BP
RIQINIRATIVC
REHEATER
STACK
ID PAN
PGD BOOSTER PAN 1
GYPSUM DRYER
Figure 4. Plant Schematic PGD Unit 5 -Neckarwerke 1991
-------
< limestone
limestone
silo
flue gas
absorber
absorber
A A A A A
A A A A A
flue gas
flue gas
AAA AA
f\ r\ r\ f\ f\ f\
v/yyy* sj \J
\y yy w w w w
absorber feed-tank
hydrocyclone
hydrocyclone
to gypsum suspension tank
Figure 5. Altbach Power Station, Unit 5 Flow Sheet, Absorber - Neckarwerke 1991
-------
water
hot air
from regenerative
air preheater
air to absorber
3
vacuum
drum rf OI7"
filters 1—
water gypsum-suspension
from hydrocyclones
&
wet gypsum
gypsum
silo
gypsum suspension
tank
waste water
waste water
treatment
double
roll
presses
) waste water to
sludge back
to FQD system
second step of treatment
Figure 6. Altbach Power Station, Unit 5 Gypsum Preparation - Neckarwerke 1991
-------
Major Design Data
FGD
Electric power (net)
MWel
420
Flue gas volume
m3/h
1.4 mio
Design of untr. S02-gas
mg/m3
2700
Limestone requirement
t/h
4
End product: gypsum
t/h
6
Energy requirement
MW
6,6
L/G - ratio
l/m 3(i.N.h.)
15
Water consumption
m3/h
68
Waste water
m3/h
18
Limestone silo
m3
1500
Gypsum silo
m3
2100
Vacuum drum filter
t/h
2 * 9
Gypsum drier
t/h
18
Compactors
t/h
2 * 9
NW-499
Figure 7.
8A-109
I
-------
1.4435 alloy 316
1.4539 alloy 904L
1.4539 alloy 904px
plated with
2.4819 alloy C-276
(retrofit measure) i
—o—
— o— •
—o—.
— o-
— O- -
— o- •
2.4856 alloy 625
Figure 8. Absorber - Neckarwerke 1991
8A-110
-------
00
*¦
GYPSUM DRYER
REOENERATIVE
REHEATING v
ABSORBER
LEAKAGE
17.
FAN
ESP
STACK
ID FAN
Figure 9. Standard Conceptual Schematic For FGD Plants Till Approx. 700 MW - Neckarwerke 1991
-------
FLUE GAS DESULPHURIZATION TECHNOLOGIES
IN GERMANY (AS OF 1990)
Remainder 6 %
Spray-/Dry-
Absorption
Lime/Limestone-
Gypsum-Process
87 % ^
TOTAL POWER PLANTS WITH FGD
100 % = 38000 MW
NW-l7l_2-EL:04.90:n»c/iwl
Figure 10.
-------
INVESTMENT COSTS
FGD - UNIT 5
CONSTRUCTION COSTS 11 Mio DM
PLANT DELIVERY 74 Mio DM
REGENERATIVE REHEATING 13 Mio DM
ERECTION 19 Mio DM
ELECTRICALS CONTROL 9 Mio DM
AND INSTRUMENTATION
WASTE WATER TREATMENT 1 Mio DM
127 Mio DM
INTEREST FOR CONSTRUCTION 8 Mio DM
TOTAL COSTS 135 Mio DM
SPECIFIC COSTS 320 DM/kW
NW-trO-EL:nec/zwi 1 $ " 1.7 DM; 1 DM " 0,6 $
Figure 11.
8A-113
-------
OPERATING COSTS
FGD - UNIT 5 (420 MW)
Costs per Hour
Full Load Operation
LIMESTONE 170 DM/h
POWER CONSUMPTION 6,6 MW 990 DM/h
WATER TREATMENT 70 DM/h
OTHERS 30 DM/h
1260 DM/h
OPERATING COSTS AT 420 MW
APPROX. 0,3 Pfg./KWh
1 $ - 1,7 DM ; 1 DM - 0,6 $
N W-160-EL;04.90:nec/Iwl
Figure 12.
8A-114
-------
FGD ALTBACH - DEIZISAU,
UNIT S: 420MWn
KWH KWH
AMORTIZATION SCHEDULE
10 YEARS
IS TEARS
3,0
2JJ
1,00
0,75
1,0
0.50-
03
0,25
i OPERATING COSTS
0 2000 3000 40Q0S000 4000
FULL LOAD OP.HRS/YEAR
1 $ » 1,7 DM; 1 DM - 0,6 $
Figure 13. Desulphurization Costs - FGD Unit 5 - Neckarwerke 1991
8A-115
-------
DISTRIBUTION OF INVESTMENT
COSTS UNIT 5
Plant Costs
603 Mio DM
900 Mio DM
33 % Investments for Environmental
Protection
FGD
Mio DM
DeNOx
62 Mio DM
Others
Noise Protection
Precipitator (ESP)
Hybrid Cooling
Tower
100 Mio DM
NW-168-EL:04.90:neo/zwl
Figure 14.
-------
S02 EMISSIONS OF NW 1985 - 1990
Total S02 a
Fossil
Total
Generation
(QWh el.)
1000'8 T/y
FQD Unit 5
1st stage -
Altbaoh/Delzlsau
FQD Unit 6/
2nd stage
Start up
QKN 2
QWh el. aOOO
7000
NW-493-ELineo/iwl:09.91
Figure 15.
-------
8 A-118
-------
DEVELOPMENT AND OPERATING EXPERIENCE
OF FGD-TECHNIQUE AT THE VOLKLINGEN POWER STATION
- Introduction
- Saarberg-H61ter FGD Process
- S-H-U Process with Formic Acid
- Gypsum
- Cooling Tower Discharge of Flue Gas
- Measurements of Cooling Tower Discharge
- Reheat System for Forced Draft Cooling Tower
- New Power Plant VSlklingen-Fenne
- Operating Experience
- Economics
- Prospects
Hans-Karl PETZEL
Saarberg Company
Trierer StraBe 1
6600 Saarbrucken. Germany
Preceding page blank
8A-119
-------
8 A-120
-------
Introduction
The growing awareness of ecological issues in Germany has led to the enactment
of lavs and regulations relating to emissions of large combustors. Since July
1st 1988 the Large Furnance Ordinance (13th Regulation on the Implementation
of the Federal Act on Protection against Emissions) specifies an emission
level of SO2 < 400 mg/n3 (140 ppm) and a removal efficiency of _> 85 %, the
most stringent of these tvo numbers applies. The legally binding standards
stipulate that none of the daily averages, calculated on the basis of half-
hourly averages, may exceed the concentration allowed. The law also requires
flue gases to have a minimum temperature at the stack outlet of 162 "F (72 eC)
after desulfurization. All wet flue gas desulfurization systems suffer from
the technical-physical drawback that the treated flue gas leaves the FGD-plant
almost watersaturated at a temperature of 122 °F to 131 «F (50 rC to 55 '•"C).
Therefore the desulfurized gas must be reheated, if it is discharged via a
power plant stack. Treated flue gases do not have to be reheated if discharged
via a cooling tower.
Saarbergverke AG is the second biggest mining company in the Federal Republic
of Germany and a power plant operator with an installed capacity of 2 400 MVel
at three sites.
At Vdlklingen - the site which I am responsible for - we operate the so-called
230 MW Prototype Power Plant. Prototype, because some very modern ideas for
power generation are realized for this plant, such as fluidized bed combustion
with a submersed heat exchanger, additional gas turbine for a combined cycle
and treating the full flue gas quantity in an FGD unit, installed into a
natural draft cooling tower.
The Prototype Power Plant VOlklingen is the first power plant with cooling
tower discharge of desulfurized flue gas.
It is in operation since August 1982 with meanwhile an operation time of
nearly 60 000 hours. The investment costs were subsidised by the Minister of
Research and Technology.
Preceding page blank
8A-121
-------
Adjacent to this unit, we operate Fenne III, a 170 MW unit, and since August
1982 a 210 MW power plant with 185 MW^ for district beating also with cooling
tower discharge of the desulfurized flue gas.
The largest units we are operating at Saarberg are a 707 MW unit and a 800 MW
unit.
8aarberg-H61ter FGD Process
The coal of the Saarberg Company has a sulfur content of 1 - 1.5 %. In the
seventies - without the request of a law - our company decided to develop a
wet desulfurization process in cooperation with the engineering company
H<er. The development of wet scrubbers for FGD systems on calcium basis has
started with lime as absorbing media as the reactive component. In a Saarberg
power plant the process was tested under normal operating conditions.
And so we started in August 1982 the Prototype Power Plant VGlklingen with
lime as absorbent. The FGD plant was not rubber lined. We operated at that
time our lime scrubber in the pH-range of about 8 - 10. The experience was
that scaling took place in several places in the FGD plant. Always after an
operation period of several weeks we had to shut dovn the plant to remove the
incrustations. To avoid scaling we tested the FGD-process with a lower
pH-range. Our demonstration plant was designed for a removal efficiency of
80 % and an emission level of < 540 mg SC^/m^.
To be in compliance with the emission law July 1st 1988 we reconstructed the
FGD-plant and commissioned again in 1986. Lime was replaced by limestone. As
the limestone scrubbers are operating in the pH-range of 4.5 - 5.5 the
FGD-plant had to be rubber lined. For some elements at the wet-dry interface
at high temperatures the construction material is Hasteloy.
8A-122
-------
The Bain disadvantage of a limestone unit is that a limestone unit needs
higher investment cost, because for a given absorption efficiency the scmbber
nodules have to be larger and the recycle pumps have to be larger as veil.
This disadvantage of higher investment cost is partly reduced due to the fact
that the oxidation to reach the final product gypsum may be done internally in
the scrubber so there is no external oxidizer necessary.
The lover pH-range of limestone has several advantages listed belov:
1. Oxidation may be done internally in the scrubber.
2. Due to the lover pH-value the primary product of the reaction
is not calcium sulphite CaSC^, but calcium bisulphite CafHSOj^.
Since calcium sulphite is hardly vatersoluble vhilst calcium
bisulphite has a much better solubility the risk for scaling in
a limestone scrubber is a lot less than in a lime scrubber. It
is our experience nov since 1986 that scaling takes no more
place.
3. To form gypsum from calciumsulphite an oxidation is necessary
and this reaction goes via calciumbisulphite. The more easy and
safer vay obviously is not to form calciumsulphite at all, but
to form calciumbisulphite directly.
4. At a lover pH-value the unit does not react sensitive against
deviations in volumetric flov rate or SOo inlet concentration,
but can rather easily handle those deviations.
5. Especially due to the longer retention time the gypsum cristals
formed in the limestone scrubber are larger and thus easier to
devater giving better gypsum qualities.
Additionally in the limestone scrubber the gypsum quality is
constant even at fluctuating operating conditions.
6. The main advantage of a limestone scrubber is the lover
operating cost since limestone is a lot cheaper than lime even
taking into consideration that the amount of limestone needed
is nearly tvice as much as the amount of lime vould be, even at
the same stoichiometric ratio.
8A-123
-------
8-H-U Process with Formic Acid
Compared to common limestone processes the S-H-U limestone process is
additionally buffered with formic acid.
This reduces the pH-range for the reaction to be in the range of appx. 4-5
and forms a lot more available calcium ions. All the advantages listed above
are even more valid for the process using formic acid compared to processes
not using formic acid.
The number of calcium ions available with formic acid is about 40 times as
high as the calcium ions available without formic acid addition.
The addition of formic acid enhances the transfer of sulfur dioxide from the
flue gas to the washing fluid. It buffers the washing fluid in the appropriate
pH range to ensure a high rate of utilization of the limestone reagent with
simultaneous high S02 absorption. The stoichiometric ratio of the absorbent in
the S-H-U-process is consequently only 1.00 to 1.02 even with limestone,
whereas for conventional limestone processes much higher stoichiometric ratios
are common-place.
Formic acid addition also has the advantage that SO2 abatement is far less
impaired when chlorides are present and requires some 20 % to 25 % lower
liquid-to-gas ratio with less energy needed for the washing fluid pumps.
The use of formic acid makes a compact combined scrubber feasible which is
economical in space and employs integrated cocurrent and countercurrent flow
stages.
Formic acid also has the advantage for the process technology that an
integrated oxidation of the SO2 bonded with the washing fluid can take place
to form gypsum (CaSO^) in the scrubber sump simply by injecting air. There is
therefore no need for a separate oxidizer.
8A-124
-------
Gypsum
The gypsum produced is of high quality and is utilized by the gypsum and the
cement industry. The composition and especially the reflectance of the FGD
gypsum is supervised by the industry.
The German market consumes about 2.5 Mio tons/a gypsum. The coal fired
stations produce about 2.5 Mio tons/a. That means the gypsum of the FGD-plants
is completely utilized. By centrifuges the gypsum is being devatered to a
residual moisture of approximately 6 to 8 %. Then it is used for example for
vail boards. For plaster the FGD gypsum is being additionally dried and
pelletized.
Cooling Tover Discharge of Flue Gas
As an alternative to reheating solutions Saarbergverke AG and Saarberg-H61ter
Umvelttechnik GnbH (S-H-U) have developed a different approach vhich is
internationally protected by patent. The desulfurized flue gas is mixed vith
the cooling air of the cooling tover and transported into the atmosphere. In
principle, no stack is retired for the normal operation of this type of pover
plant.
The application to avoid costly reheating of desulfurized flue gas made its
world debut 1982 in the 230 MW coal-fired Saarbergverke pover plant
VOlklingen. The S-H-TJ flue gas desulfurization unit is integrated into a
cross-flov natural-draft cooling tover. It is located in the center of the
cooling tover and measures approximately 31 m x 18 m. The four outlet tubes
vith a diameter of some 3 m end at a height of appr. 40 m. The height of the
cooling tover is 100 m.
The vails of the cooling tover are likely to be affected by harmful residues
since conglomerated vater drops are falling back into the tover as veil as
condensation of vapor occuring at the shell. Carbon dioxide is particularly
harmful. The causative factor is carbonization of the concrete vhich occurs
vhen carbonic acid is present simultaneously vith vater vapor.
8A-125
-------
Without adequate protection inside the shell, the concrete as veil as the
steel reinforcement vould be attacked.
As a result of tests we decided that concrete with Portland Cement P/45 F HS
and a water-cement factor of < 0.5 has to be used for the shell and the tower
pillars. Furthermore the inside of the shell oust be coated. As with certain
wind speeds the plume is touching the outside of the shell, also the outer
shell was protected.
In the Prototype Power Plant VSlklingen directly after the concrete work
continuously a primer was applied. Then the shell was coated twice with
epoxy-resin. After an operation time of 17 000 hours on the upper third inside
the shell we experienced peeling.
It is very likely, this was caused due to the fact that the temperature during
application was below + 5 °C, so that the primer could not be applied
correctly. In the time to the next summer the surface of the concrete was no
longer clean, so that the primer on the surface did not accomplish the actual
bond.
In a height between 80 and 100 m the inside of the shell was sand blasted, the
primer was applied again and twice coated with two layers of epoxy resin. We
inspected the shell. Up to now (60 300 h operation time) the coating of the
shell shows no fissuring and no cracks.
The cooling tower of the new plant was totally completed and then the shell
was cleaned by high pressure water. Then we followed the same procedure: a
primer was applied and then the shell twice coated with two layers of epoxy
resin.
Measurements of Cooling Tover Discharge
To substantiate the theory of cooling tower discharge comprehensive
measurements were taken in the winter 1984 and the stumer 1985 by several
university institutes. The results are published at several meetings.
It is essential that cooling tower plumes are able to pierce non-turbulent
layers of atmospheric inversion and to reach higher levels. This is a
particular advantage when there are non-turbulent weather situations (smog).
8A-126
-------
In our case the desulfurized flue gas is mixed with the cooling air and
piercing the inversion layer. Also during smog situation we are allowed to
continue the operation of the power plant.
The cooling tower and a stack have different effects on ground
level concentrations in the immediate area. Because of its
higher temperature and greater thrust the cooling tower plume
remains more conpact for a longer period.
This is due to the internal circulation mechanism. Since the
cooling tower plume disperses at a greater altitude, the
"emission" is lower with this type of discharge than in the
case of flue gases released from stacks.
Due to an additional separator installed in the flue gas the
concentration of dissolved pollutants and particles sticking
to droplets is further reduced by a factor of approx. 2.
|S A ARBERG|
Comparison of Flue Gas Discharge
via Stack or Cooling Tower
Stack
Owned Gas
Cooling Tower
cleaned a. C T. air
Velocity
5
1
14
3
6
10
1
Mass
Impulse
Amount of Heat
Draft
SO2 Concentration
2
V:
J
8A-127
-------
A further advantage of cooling tower discharge is that the concentration of
pollutants is reduced by a factor of 24 compared with stack discharge at the
emission outlet.
The theoretical predictions about the performance of the cooling tover, the
dispersal pattern of the plume in the atnosphere and the chemical and physical
transformations in the plume were confirmed by these measurements.
ti-J.s discharge of flue gas into the cooling tover and mixing into the plume is
tr.-v fully approved as alternative to the stack. The provisions of the clean
air standard are complemented by special appraisal.
Reheat System for Forced Draft Cooling Tover
The 170 MV unit Fenne III has two forced draft cooling tovers. Therefore ve
could not apply the cooling tover discharge. The desulfurized flue gas has to
be reheated. He decided for a heat exchanger system. The untreated flue gas
transfers the heat to the vater in a heat exchanger. Then in a second heat
exchanger the varm vater heats up the cooled desulfurized flue gas to a
temperature of above 162 F (72 C). The flue gas is then released via a stack.
The material of the heat exchanger tubes is a PTFE basis. This system also
operates vithout difficulties. Compared to the design of a regenerative heat
exchanger the advantages are
due to lover temperatures of the rav gas condensation does
occur giving lover concentrations of pollutants
due to the principle of the heat exchanger no leakage occurs;
thus the concentration of pollutants is lover
Nev Pover Plant Vfllklingen-Penne
The nev pover plant at the V&lklingen site has a counter flow natural draft
cooling tover. The FGD-plant is located besides the cooling tover.
The location of the FGD plant outside the cooling tover has the advantage that
the dimensions of the cooling tover only have to fit the thermodynamic
requirements.
8A-128
-------
The following comparison demonstates this:
Both units the Prototype Power Plant and the new power plant have
the same output
The basic diameter of the cooling tower for the
Prototype power plant is 90 m
new power plant is 60 m.
Operating Experience
Since 1986 we are operating all our FGD-units in the Saarberg company with
fornic acid, limestone and a low pH-value. All units are rubber lined.
material in the quenching zone is Hasteloy.
The straightforward technical construction of the basic flow diagrasn is
matched by an equally uncomplicated, fully automatic process. This ensures
that there is no need for any labor-intensive chemical operations in the power
plant. The FGD-unit is controlled on a side panel in the control room of the
plant. No additional operators are required in the control room to run the
FGD-unit. The fully automatic start-up and shut-down procedures via the
control units require no lengthy preparation or special measures, such as
separate start-up of the various pump circuits. The unit attains its full
desulfurization capacity within 5 to 7 minutes.
During operation since 1986 the power plant complex has suffered no
restrictions on availability whatever due to FGD. According to emission law
the power plants could be operated 240 hours/a without FGD. Due to the high
availability we installed no bypass and also no stack - also not for the new
power plant.
The cooling tower discharge of flue gas to avoid costly and energy intensive
reheating is gaining ground in Europe. The RWE (Rheinisch-Westfaiische
ElektrizitStswerke AG) was the first major German energy producer to decide to
discharge treated gases via cooling towers in all its power plants with
natural-draft-cooling-towers with a total capacity of 6.000 MV (6 x 600 MW,
8 x 300 MV). Meanwhile more than 20 units are in operation with cooling tower
discharge of desulfurized flue gas.
8A-129
-------
Oar Prototype Pover Plant Vdlklingen has the most experience vith an operation
of 60 000 hours. The advantages of cooling tower discharge are proved
no use of external energy for reheat
- no operating costs for reheat
availability of system 100 %
stackless pover plant
Bconoaics
Vhen ve compair the cooling tover discharge with a reheat system there are
additional investment costs not only for a gas heat-exchanger, but also for
steel construction foundation and additional flue gas ducts. Power is needed
for a rotating gas heat-exchanger and more significant for covering the
pressure drop of about 14 mbar.
Similar considerations apply to the costs of operating, maintaining and
cleaning a gas heat-exchanger. In addition you have to consider the costs for
a stack.
None of these considerable investment and operating costs were incurred vith
the Vdlklingen cooling discharge (stackless pover plant).
The solution adopted in Vdlklingen involves reduced investment costs of
approx. DM 10,2 MW ($ 6 MM) compared vith a conventional FGD unit vith
reheating and stack.
The V&lklingen Economics are
- lover energy consumption
no maintenance and cleaning
higher availability
5 - 7 % lover operating costs per year
8A-130
-------
Prospects
Vet-scrubbing desulfurization processes on lime or limestone will remain
dominant with over 90 % of the market.
There vill be a growing emphasis on those desulfurization systems which use
less energy (electricity) and, above all, those which obtain high levels of
abailability, for example by using organic additives.
In Europe, the main F6D end-product permitted in the future will be gypsum
alone. This can be used by the gypsum and cement industry. By contrast, there
will be little interest in Europe in end-products from so-called dry scrubbing
processes with unreacted lime components because of the need for expensive,
environmentally unacceptable special storage facilities.
The VBlklingen plant with an S-H-U flue gas desulfurization unit with cooling
tower discharge is in its tenth year of operation and has successfully
demonstrated that there is a more effective and far cheaper alternative to the
costly and energy intensive conventional reheating of desulfurized flue gases
and with a high availability.
8A-131
-------
8A-132
-------
ADVANTAGES OF THE CT-121 PROCESS
AS A THROWAWAY FGD SYSTEM
M. J. Krasnopoler
G. Shields
Bechtel Power Corporation
9801 Washingtonian Blvd.
Gaithersburg, MD 20878-5456
Y. Shoji
Chiyoda Corporation
2-12-1, Tsurumi-Chuo
Tsurumi-Ku
Yokohama 230, Japan
8A-133
Preceding page blank
-------
8 A-134
9
-------
ABSTRACT
Over 3,900 MW of flue gas desulfurization plants utilizing the CT-121 process are installed
and in design. The majority of these produce high-quality marketable gypsum. For the
North American market, however, process design of CT-121 has been optimized to
produce throwaway gypsum
Bechtel is currently marketing the Chiyoda CT-121 process for a number of United States
flue gas desulfurization (FGD) retrofit applications. These CT-121 designs typically
address high sulfur (4%) and high chloride (0.4%) coal and require high SOz removal
efficiencies (95% and above). The designs produce a throwaway gypsum byproduct and
often include zero liquid discharge from the FGD process.
Current designs use large non-redundant absorbers, called jet bubbling reactors (JBR),
that achieve consistent high SOz removals at relatively high liquor chloride concentrations.
The unique gas/fiquid contacting action of the JBR also removes most of the particulate
from the inlet flue gas. The large gypsum crystals produced in the CT-121 process allow
single-step dewatering of the gypsum slurry in a gypsum stack or vacuum filters.
This paper reviews the various design options available and emphasizes the advantages of
the CT-121 process in the United States retrofit FGD market These same advantages will
be important to the next generation of coal-fired power plants.
Preceding page blank
8A-135
-------
ADVANTAGES OF THE CT-121 PROCESS
AS A THROWAWAY FGD SYSTEM
INTRODUCTION
The Chiyoda Thoroughbred 121 (CT-121) process has been well proven in nine
commercial plants worldwide with an installed capacity totalling 1,600 MW. Current
designs underway total a further 2,450 MW. These installations have primarily been in
Japan where the high-quality gypsum is utilized either for wallboard or cement
manufacture. In the United States, the 40 MW Abbott plant at the University of Illinois
operates on high sulfur coal and produces byproduct gypsum which is landfilled due to
the lack of a commercial market The 100 MW Plant Yates, scheduled to startup in 1992,
will also produce throwaway gypsum.
This paper will demonstrate that the inherent advantages which the CT-121 flue gas
desulfurization (FGD) process has over other limestone forced oxidation processes can be
similarly applied to a throwaway byproduct FGD process. It will highlight the following:
Process and mechanical simplicity
High S02 removal efficiency
High particulate removal
Scaleup
Layout features
Single-step gypsum dewatering
Zero-liquid discharge capability
CT-121 PROCESS SIMPLICITY
The CT-121 process is an advanced, wet limestone FGD process that uses
gas/liquid contacting device, the jet bubbling reactor (JBR) as its absorber,
following general issues capture the simplicity of CT-121:
a unique
The three
8A-136
-------
1. All process reaction steps occur simultaneously in the JBR.
2. Complex slurry recycle systems are eliminated.
3. Byproduct gypsum dewatering is achieved in a single stage.
The simplicity of the CM 21 process is clearly depicted in Figure 1. This simplicity results
in significant benefits in terms of capital and operating costs.
The four mayor FGD process steps which occur simultaneously in the continuous liquid
phase of the JBR are:
SOz absorption
Oxidation of acid sulfites
Neutralization with limestone
Gypsum crystal growth
The significant innovation of the CT-121 process reverses the conventional concept by
making the JBR slurry the continuous phase. Conventional FGD using spray towers have
flue gas as the continuous phase. This change from other limestone FGD processes
enhances the mass transfer and chemical reaction mechanism and accounts for the
unique characteristics of the CT-121 process. The most important enhancement of
CT-121 is its low-pH operation and the resultant benefits yielded by different reaction
chemistry. CT-121's improved gas/liquid contact also increases the particulate removal
capability.
Process Description
Figure 1 shows the entire CT-121 process and Figure 2 depicts the heart of the process,
the JBR. Rue gas enters the system through a presaturator duct and then is finely
dispersed into the JBR slurry through gas spargers. SOz is absorbed as the flue gas
bubbles up through the slurry. The treated flue gas then flows up through the gas risers,
into the plenum above the upper deck, and out of the JBR to the mist eliminator and
chimney.
Oxidation air is introduced into the JBR to oxidize the acid sulfites (HS03") to sulfates
(SO/ + H*). Limestone slurry is added to neutralize the acid sulfates to form gypsum and
to maintain the pH. The JBR is designed for a slurry residence time of 15 to 25 hours to
achieve large crystal growth and efficient limestone conversion.
8A-137
-------
Makeup
Water
Presaturator-
nQ
Chimney
Mist
Eliminator
Jet Bubbling 1,
Reactor
Vacuum Filter
Gypsum
Slurry Pump
-Limestone
Flue &
Oxidation
Air Blower
Booster
Fan
Ball Mill
System
IN
Limestone
Slurry Pump
Gypsum
Figure 1. CT-121 Row Diagram. The CT-121 process Is simple because It eliminates mechanical equipment such as the
large recycle pumps and primary dewaterlng equipment of other FGD processes.
-------
Gas outlet
Inlet plenum
Figure 2. Gas Row in the JBR. The even spacing of the JBR's gas spargers and gas
risers assure uniform fiue gas distribution.
8A-139
-------
A bleed stream of JBR slurry is pumped to the dewatering system, either a vacuum filter 01
a gypsum stack. Return water from dewatering is pumped to the ball mill for limestone
slurry preparation and back to the JBR.
Process Features
The major FGD process steps occur in CT-121's JBR because the slurry is the continuous
phase. Conversely, in other FGD processes, SOz absorption and partial oxidation occur ir
the spray section while other reaction steps occur in the recycle tank. Large recycle
pumps circulate substantial flowrates of slurry from the recycle tank to the spray nozzles.
By contrast, the slurry circulation required in the JBR is supplied by large-diameter low-
speed turbine agitators and supplemental mixing from the flue gas and oxidation air
spargers. Thus, the CT-121 process eliminates the large recycle pumps and spray
headers, making it mechanically less complex than conventional spray tower FGD
systems.
Figure 3 summarizes CT-121's process chemistry and shows that H2S03 is rapidly
oxidized to H2S04 before it is neutralized by limestone. Thus, S02 absorption in the JBR i;
essentially independent of dissolved alkaline species. Conversely, conventional spray
tower designs operate at relatively high pH levels (5.5 to 6.0) and do depend on dissolved
alkaline species to provide the driving force for S02 removal.
Oxidation of dissolved sulfite to sulfate in the CT-121 process occurs virtually
instantaneously after SOz absorption. Thus, S02 backpressure from sulfite in solution is
eliminated and high removal efficiencies can be maintained. Complete oxidation and low
pH also eliminate the potential for scaling.
Operation at low pH is the key to scrubbing effectively with limestone, because of
enhanced limestone dissolution and sulfite oxidation. Limestone dissolution is 100 times
faster at CT-121's typical pH of 4.5 than at other FGD system's typical pH of 5.5. The
limestone dissolution rate is a function of hydrogen ion concentration or acidity:
r = K [CaC03] [H*]
where r is the limestone dissolution rate, K is the rate constant, [CaC03] and [H*] are the
concentrations of calcium carbonate and hydrogen ion. Since pH = -log10[H*], for
example, a decrease in pH of 2 corresponds to a 100-fold increase in hydrogen ion
8A-140
-------
*¦ h2so4
JBR Gas Sparger
o* \o
O O o
= gas phase
SO2 = "quid phase
\ \- solid phase
\CaS04»2H^cT\ rflrn3
C.
: \CaCO3V
Figure 3. Sequence of CT-121 process chemistry In the JBR. The enhanced mass transfer and Improved chemistry of
the CT-121 process allow low-pH operation.
-------
concentration and thus limestone dissolution rate. CT-121's fast limestone dissolution and
long J BR liquid residence time creates high limestone utilization, typically over 99 percent.
To summarize, CT-121's process features are:
Mechanical simplicity
Low pH operation
No potential for scaling
Complete oxidation
Extremely high limestone utilization
Large, easily dewatered gypsum crystals
HIGH SOz REMOVAL EFFICIENCY
The CT-121 process can consistently
achieve greater than 95 percent S02
removal over a full range of coal sulfur
and chloride levels and at significant
turndown ratios. As shown in Figure 4 ,
S02 removal may be increased by
increasing either the JBR pressure drop
or system pH. Pressure drop is
increased by increasing the
submergence of the sparger tubes and
pH is controlled by the limestone addition
rate.
Careful design and layout of the JBR gas
spargers and gas risers as well as
system backpressure assure excellent
gas distribution and high-level S02
removal even at turndown to 25 percent
of design gas flow. Figure 2 depicts the
gas flow through the JBR and shows the even spacing of the gas spargers and gas risers.
To achieve high S02 removal in a conventional spray tower, the liquid to gas (L/G) ratio
and consequently the recycle slurry flowrate must increase. One possible limitation of high
100
pH = 4,5
96
pH = 4.0
CO
88
JBR AP, inches w.g.
Figure 4. Typical CT-121 S02 Removal
Efficiency. The CT-121 process can
achieve high SOz removal efficiencies by
increasing the JBR pressure drop or pH.
8A-142
-------
L/G's is inadequate gas-liquid distribution, in particular where turndown is required. Thus,
the JBR uniform gas distribution is a significant advantage over spray towers.
The CT-121 systems on the Mitsubishi Petrochemical (85 MW) and Nippon Mining
(75 MW) CT-121 plants which burn high sulfur oil achieve 98 percent S02 removal. The
225 MW CT-121 process at Kashima treats gas from a high-sulfur oil-fired boiler and
removes 97 percent of the inlet S02. The CT-121 system at the 40 MW stoker-fired boiler
at the Abbott Plant achieved 93 to 96 percent SOz removal while operating on coals
ranging from 1.5 to 4.5 percent sulfur. The aggregate past performance of the CT-121
system establishes its capability to achieve consistent high S02 removal efficiencies while
burning high sulfur coal.
HIGH PARTICULATE REMOVAL
Due to the unique gas-liquid contactor,
the JBR, the CT-121 removes most of the
particulates from incoming flue gas. This
has been demonstrated in pilot and
commercial plants1. This high particulate
removal efficiency is due to the improved
gas-particulate/liquid contacting created
by:
Turbulent two-phase flow of flue
gas and JBR slurry
High gas-side pressure drop
Large interfacial contacting area
of small collapsing gas bubbles
As shown in Figure 5, the JBR achieves
high levels of particulate removal in the
sub-micron range2. This can
compensate for the performance of a
marginal electrostatic precipitator and
reduce particulate levels to meet New
Source Performance Standards (NSPS).
100
>*
O
C
©
90-
lj 80-
o
>
o
E
*
C£
©
o
3
O
O
CL
70
60
50
0.1
r
t
1
i
\
\
\
/
/
/
Particle Size, /j.
Figure 5. CT-121 Particulate Removal
Capability. The CT-121 process can
achieve high particulate removal
efficiencies. (Source: Reference 2.)
8A-143
-------
This benefit of lower particulate emissions is achieved without any increase in the capital or
operating costs of the CT-121 process.
The resultant particulates removed by the CT-121 scrubbing action collect in the gypsum
byproduct which can be safely landfilled. CT-121's particulate removal capability will be
increasingly more important as environmental regulations on air toxics are tightened and
the need to remove a larger percentage of heavy metals from stack gases increases.
SCALEUP
CT-121's unique gas/liquid contacting device, the JBR, is ideally suited for scaleup to
large sizes. Scaleup of the CT-121 process simply means increasing the number of JBR
spargers and increasing the JBR diameter to accommodate them. Each sparger in the
JBR is designed for the same set gas flowrate, so scaling up a JBR simply means
increasing the number of spargers. For example, to double the gas flow capacity, the
number of spargers are doubled. Uniform gas flowrate per sparger and SOz removal
efficiency are thereby maintained. Conversely, the scaleup of conventional spray tower
systems is limited as spray header designs become very complex and the risk of
maldistribution of flue gas or spray droplets increases.
The largest CT-121 unit in operation is a 350 MW oil-fired unit operating at over 90 percent
SOz removal on 2.9 percent sulfur oil. Currently in the design or construction phase is a
single absorber CT-121 system for a 500 MW coal-fired utility boiler, as well as a 1000 MW
single-absorber CT-121 system.
Chiyoda and its clients are confident in the scaleup of CT-121 to these large sizes because
of design simplicity and past operating history. Table 1 indicates an aggregate reliability of
major CT-121 units in excess of 99 percent. In all instances these units operate with a
single JBR. In fact, no CT-121 plan has ever used a spare absorber. The ability to
scaleup the CT-121 process to large capacities provides economies of scale and capital
savings. This should prove to be of major benefit to the future North American FGD
market.
8A-144
-------
Table 1
CT-121 PROCESS RELIABILITY PARAMETERS
Period:
From
To
MHtublihl
Petrochemical
Yokkalchl
May 11,1062
Dec. 31, 1990
Nippon
Mining Co.
Chita
Nov. 10, 1983
Dec. 31,1990
Toyama Kyodo
Electrlo Power
UnNI
July 9, 1904
Dec. 31,1990
Toyama
Kyodo
Electrlo Power
Unit 2
Aug. 23, 19S4
Dec. 31,1990
Kaihlma
Northern
Joint Power Co.
No* IB, 1985
Deo. 31,1990
Hokurlku
Electrlo Power
Kuujlma
July 24, 1987
Dec. 31, 1990
Unhrwtlty
of llllnoli
Abbot
Aug. 18,1988
Sept. 2, 1989
Hour* of op*fitton
71,043
57,071
46,408
47,572
39,859
17,908
8,805
Hour* called upon to operate
71,029
67,468
48,414
47,572
40,202
17,908
8,883
Reliability, %
98.8
99.3
100.0
100.0
99.1
100.0
99.1
Hour* available
72,208
69,475
53,387
51,248
42,905
29,142
8,638
Hour* In period
76,758
82,680
58,794
65,714
44,928
30,160
9,120
Availability, %
95.3
95.0
94.0
92.0
95.5
96.7
93.6
Reliability - Hours the CT-121 process was operated divided by hours the CT-121 process was called upon to operate
Availability = Hours the CT-121 process was available lor operation (whether operated or not), divided by the hours In the period
-------
SINGLE-STEP GYPSUM DEWATERING
The CT-121 process produces a completely oxidized, easily dewatered gypsum byproduct
that is easily dewatered and is suitable for landfill disposal without the need for fixating with
flyash. In Japan, most CT-121 installations produce wallboard or cement quality gypsum;
however, CT-121's process simplicity advantages apply equally to throwaway system
designs that predominate in the U.S.
Contrary to conventional forced oxidation FGD systems, CT-121 does not experience
crystal attrition due to the high speed recycle slurry pumps. Rather, CT-121 attains large
gypsum crystals due to the long residence time and high concentration of gypsum slurry
in the JBR.
The benefit of the large gypsum crystals from a CT-121 process is that they dewater
easily. At Abbott, absorber slurry with 15 to 25 weight percent solids is fed directly to the
vacuum filter which produces a 93 weight percent solids cake. Primary dewatering by
thickeners or hydrocyclones is eliminated. Table 2 shows a comparison of gypsum
dewatering from a CT-121 process and conventional in-situ forced oxidation process.
Elimination of the primary dewatering step creates significant advantages in terms of
capital and operating costs.
Table 2
GYPSUM DEWATERING
Other Forced Oxidation
Parameter CT-121 Process FGD Processes
Mean Crystal Size, m 80 - 100 50
Feed Slurry, % solids 15 — 25 40 (thickener/hydrocydone)
Vacuum Filter Cake, % solids 90 — 93 85 - 90
The gypsum byproduct produced by the CT-121 process is suitable for landfilling. Recent
tests show that CT-121 gypsum easily passed standard leaching tests used to define
hazardous wastes3.
i
8A-146
-------
An alternative to gypsum filtration is gypsum stacking. Gypsum stacking has been used
extensively in phosphoric acid production operations and simply involves filling a diked
area with gypsum slurry piped directly from the JBR. Return water from the gypsum stack
is pumped back to the FGD process. Gypsum is excavated from the center of the stack
and used to build up the walls of the dike. The solids content in the stack walls reaches
up to 80 percent because of the ease of dewatering. The walls can readily support the
vehicular equipment needed to build the stack. Gypsum stacking requires about the same
space as landfill disposal. It is environmentally safe and can eventually be grassed over.
The simplicity of dewatering the gypsum byproduct is a significant advantage for CT-121
over other throwaway FGD processes, whether filtration or gypsum stacking is utilized.
LAYOUT FEATURES
The smaller space requirements of the CT-121 process are an advantage to both retrofit
FGD plants and new generation plants. The CT-121 system is a highly compact because:
No spare absorber is needed.
Large non-redundant JBRs (up to 1000 MW) can be provided.
Large recycle pumps are not needed.
Thickeners and sludge fixation equipment are not needed.
A major cost impact in plant retrofits is the ductwork arrangement. Depending on site-
specific factors such as the location of the existing stack, routing the ductwork may be
difficult The CT-121 process offers the following advantages that can lead to economic
ductwork system design:
CT-121's JBR height is lower than other absorbers.
The outlet duct can be horizontal at almost any orientation.
The outlet duct can be vertical through the roof of the JBR.
CT-121's flexibility and the significantly lower height of the JBR compared to a spray tower,
reduce both the length and complexity of the ductwork.
8A-147
-------
ZERO-LIQUID DISCHARGE
Regulations governing surface discharge of wastewater in the U.S. are becoming
increasingly more stringent Thus, FGD processes which can function with minimum or
zero wastewater discharge will be of premium importance. In such processes, impurities
are concentrated in the scrubbing solution and are ultimately purged in the throwaway
gypsum.
The primary impurity affecting a closed water balance is dissolved chlorides which
originate from coal. These dissolved chlorides result in high dissolved calcium
concentrations, as CaCI2. The resulting "common ion effect" causes dissolved CaCI2 to
interfere with the dissolution of the CaC03 from limestone. Inhibition of CaC03
solubilization means less dissolved alkalinity and pH suppression. In extreme cases, in
conventional FGD processes, high chloride pH suppression has led to failure of the
system to maintain the design SOz removal efficiency.
The S02 removal efficiency of CT-121 is not impacted by such high chloride
concentrations. This is because of the unique process chemistry, explained earlier, which
allows CT-121 to operate at low pH and quickly solubilize fresh limestone feed. Operation
at 95 percent S02 removal with chloride levels up to 70,000 ppm has been demonstrated
in the laboratory. Further testing of this attribute in commercial facilities is planned.
A further complication of closed loop operation is the accumulation of dissolved fluoride
ions. (Fluoride and aluminum originate in the coal.) If the fluoride concentration is too
high, aluminum fluoride can precipitate from solution and coat undissolved limestone
particles, thereby causing limestone blinding which inhibits limestone dissolution and
utilization*. Aluminum fluoride blinding is not a problem in the CT-121 process because at
its lower pH, limestone is rapidly solubilized.
The ability of the CT-121 process to operate at high concentrations of ionic impurities and
achieve zero discharge is therefore much greater than conventional forced oxidation FGD
processes. This means CT-121 can be operated more efficiently over a wide range of coal
impurity levels. It also allows greater flexibility in the use of a variety of sources of plant
water for makeup to the CT-121 process.
8A-148
-------
Wastewater Treatment
An alternative to closed-loop operation is to purge a blowdown stream from the FGD
system. To keep the overall system zero-discharge, this blowdown must be treated,
typically in an evaporator/spray dryer wastewater treatment system. The blowdown is
concentrated to a brine in the evaporator with clean condensate returning to the FGD
system. The spray dryer evaporates the remaining water in the brine to produce a solid
powder for waste disposal. The blowdown flowrate, and so, the wastewater treatment
equipment will be smaller for a CT-121 system than other FGD systems because CT-121
can handle higher chloride concentrations.
The CT-121 FGD system can eliminate the spray dryer if a zero-discharge wastewater
treatment system is needed. Since CT-121 gypsum easily dewaters to over 90 percent
using a Bird-Young drum vacuum filter, the evaporator brine can be sprayed on the dry
gypsum byproduct The final gypsum and evaporator brine waste byproduct is still
relatively dry and is suitable for landfill disposal. By spraying the brine onto the gypsum, a
second separate solid waste from the spray dryer is eliminated. Figure 6 is a schematic of
the process steps.
The CT-121 process will tolerate higher chloride concentrations than other FGD systems.
Thus, in many cases the CT-121 FGD process can be operated with a closed water
balance. If wastewater treatment equipment is needed it will be smaller for a CT-121
system.
CONCLUSION
The advantages of the CT-121 process discussed here provide significant capital and
operating cost savings when applied to production of a throwaway byproduct gypsum.
The CT-121 process requires less space, process equipment, piping, and ducting than
conventional systems. The simple mechanical design results in high reliability,
straightforward operation, and low maintenance. Elimination of major mechanical
components such as a spare absorber, large recycle pumps, spray nozzles, and
thickeners reduces the capital costs and simplifies the operation and maintenance of the
CT-121 process.
8A-149
-------
VACUUM
filter
filtrate
to J3R
CONDENSATE
TEMPORARY
GYPSUM
STORAGE
EVAPORATOR
STEAM
BRINE
TO
CONDENSATE
TANK
Figure 6. Process Schematic of Wastewater Treatment and Gypsum Disposal. The
CT-121 process can eliminate the spray dryer in a zero-discharge wastewater treatment
system by spraying the brine on the 90+ % solids filter cake.
8A-150
-------
The CT-121 process can achieve over 95 percent SOz removal efficiency with limestone
and without chemical additives. Its high particulate removal efficiency translates into cost
savings when the existing electrostatic precipitator is inadequate and will provide added
benefits as air toxics regulations are implemented.
The improved chemistry and unique absorber of the CT-121 process simplify the overall
system design. It can eliminate or minimize zero-discharge wastewater treatment
equipment because of its ability to operate with high chloride concentrations. Primary
dewatering equipment is not needed because of CT-121 large gypsum crystal size. If
gypsum stacking is used, no dewatering equipment is needed. The unique advantages of
CT-121 make it a very attractive choice for future North American FGD applications.
REFERENCES
1. K. Wataya, A. Hon, N. Hashimoto, H. Koshizuka, and D. D. Clasen, "Operating Results
of Toyama Kyodo Electric Power's Chiyoda Thoroughbred 121 Hue Gas
Desulfurization System." Presented at the EPA/EPRI Ninth Symposium on Hue Gas
Desulfurization, Cincinnati, Ohio, June 1985.
2. Y. Hozumi and Y. Yoshizawa, "Numerical Analysis of Dust Particles Motion inside Gas
Bubble for Hue Gas Desulfurization in Jet Bubbling Reactor." Presented at the Forum
on Micro Ruid Mechanics, ASME, 1991.
3. D. D. Clasen, 'Commercial Status of the Chiyoda Thoroughbred 121 Hue Gas
Desulfurization Process." Presented at the Canadian Electric Association Seminar on
Hue Gas Desulfurization, September 1983, Ottawa, Canada.
4. J. B. Jarvis, R. W. Farmer, and D. A. Stewart, "Description and Mechanism of
Limestone FGD Operating Problems due to Aluminum/Huoride Chemistry." Presented
at EPA/EPRI Tenth Symposium on Hue Gas Desulfurization, Atlanta, Georgia,
November, 1986.
8A-151
-------
Session 8B
BY-PRODUCT UTILIZATION
GERMAN EXPERIENCE OF FGD BY-PRODUCT DISPOSAL AND UTILIZATION
J. Demmich
E. WeiBflog
G. Roeser
GFR
Aufbereitung und Verwertung
von Reststoffen mbH
8700 Wurzburg
Federal Republic of Germany
F. Ghoreishi
Noell, Inc.
3780 Kilroy Airport Way, Suite 350
Long Beach, CA 90806
ABSTRACT
Worldwide, increasingly restrictive requirements have been regulated for flue gas treatment at coal
fired power stations. Increased flue gas cleaning has led to an increase of hazardous materials,
originally contained in the fuel and combustion residues, finding their way in the fly ash and flue gas
cleaning systems process by-products.
The paper will review the types, quantities, and compositions of various residues from different flue
gas cleaning systems. Depending upon the physical and chemical properties of these materials, they
could be directly utilized or, they require treatment with suitable techniques in order to be safely
disposed.
Examples of FGD gypsum, ashes from fluidized bed combustors (FBC), and spray dry absorbers
(SDA) by-products will be discussed insofar as their potential for utilization in the gypsum and sulfuric
acid industries and/or the construction industry.
Process for treatment and disposal of non-utilizable residues carried out in Germany by the authors will
be covered. Among topics for discussion are disposal of mixtures of fly ash and gypsum, FBC ashes,
SDA products, and design of mono disposal site without additional binder, and comparative costs.
8B-1
-------
INTRODUCTION
During combustion of fossil fuels in power plants and industrial incinerators, various residues remain
due to incombustible particles of the coal, such as slag and fly ashes. Further flue gas cleaning, such
as removal of acid gaseous components, especially sulphur oxides, result in various by-products,
depending on the Flue Gas Desulphurization (FGD) process.
German legislation primarily requires to avoid and/or utilize those residues. Specific characteristics
or the composition of some residues can, however, interfere with their utilization.
Therefore, the question of utilization and disposal of residues can only be answered if detailed
information is available regarding quantities and compositions (Reference 1 to 5).
TYPE, QUANTITY AND COMPOSITION OF POWER PLANT RESIDUES
At this time, coal-fired power plants use various procedures for cleaning flue gas. Mainly, they are
dust collection and wet scrubbing processes. The various types of power plant residues, dependent on
die type of FGD cleaning process, are shown in Figure 1. Basically, residues can be divided into four
categories:
• Fly ashes, particulate, or dust collection without reaction products, which are not discussed
in this paper.
• Ashes from fluidized bed combustion with dry additive process as a mixture from fly ashes
and reaction products; i.e. desulphurization at high temperatures directly in the furnace.
• Residues from dry and semi-dry absorption FGD processes as a mixture from fly ashes and
reaction products with a fly ash content, which is dependent on the quality of the dust
collection system used; i.e. desulphurization at low temperatures after the boiler.
• Residues from the wet absorption process, basically FGD-gypsum and FGD waste water.
Most of these residue materials are dry and free flowing. Normally, only FGD-gypsum is produced
in moist, fine particles or, after drying/briquetting, in form of briquettes.
8B-2
-------
QUANTITIES
Depending on the type of the various FGD-processes, different quantities of residues are formed in
coal-fired power plants. Figure 2 shows the varying residue quantities obtained by combusting one ton
(1,000 kg) of hard coal by using the limestone scrubbing process, the semi-dry absorption process or
the fluidized bed combustion with dry additive process. The figures shown are relative, since they
depend on the type of coal and its contents.
The contribution of different types of power stations and power supply companies to flue gas
desulphurization and distribution of flue gas cleaning systems in respect to total electricity output in
Germany is shown in Table 1.
Residue output quantities from the combustion of coal and the related FGD are shown in Table 2.
COMPOSITION OF POWER PLANT RESIDUES
The FGD residues of coal-fired power plants differ at times considerably in their chemical and mineral
composition. Table 3 shows a comparison of the main and trace components of natural gypsum and
FGD-gypsum. The main components of residues from the dry flue gas cleaning process are shown in
Table 4.
UTILIZATION OF FGD-GYPSUM
Location, time of the year, and availability as well as quality are important for utilizing FGD-gypsum,
especially in the gypsum industry.
The requirements of the gypsum industry for the FGD-gypsum quality are stated in Table 5.
Normally, these requirements are achieved by using additional processes like hydroclone classification
and filtration/washing.
Following are important industries for gypsum utilization (Reference 9):
GYPSUM INDUSTRY
The gypsum industry is, of course, the main user of FGD-gypsum; FGD-gypsum is used in
construction (stucco and gypsum-plaster) and for gypsum wallboards.
CEMENT INDUSTRY
Following the gypsum industry, the cement industry is another large consumer of FGD-gypsum and
anhydrite.
In the cement production, approximately 5% milled CaS04 (a mixture of FGD-gypsum/anhydrite) is
added to the cement clinker as a retarding agent.
8B-3
-------
MINING INDUSTRY
In recent years, the mining industry also uses FGD-gypsum and the building material produced from
the gypsum. Part of the natural anhydrites can sometimes be substituted by FGD-gypsum after
processing it to anhydrate or a-hemihydrate.
OTHER UTILIZATION
FGD-gypsum from hard coal-fired power plants is sometimes used for earth and landscape projects
(Reference 4), particularly if there is no market demand or the product does not meet the required
specifications (Table S). For this use it is mixed with fly ash and, if necessary, with bonding agents.
FGD-gypsum of lignite-fired power plants, mixed with free calcium containing fly ashes, is almost
entirely processed to a fixated product ("stabilisate") and disposed in abandoned lignite mines above
ground.
This utilization will continue to be an important possibility of using either a seasonal or a regional
surplus.
The various utilizations of FGD-gypsum are listed in Figure 3.
PROCESSING AND CONDITIONING OF FGD-GYPSUM
Normally, FGD-gypsum is produced in fine particles with a moisture of appiox. 10% by weight. Only
where the existing utilization installations are retrofitted from natural gypsum (rocks) to FGD-gypsum, i
can it be used in this form. However, this requires a tremendous amount of handling. Therefore,
many power plants compact their FGD-gypsum with excess heat to form briquettes, which then can
be used like natural gypsum. See Figures 4-7.
In addition, energy and transportation costs can be saved, and the handling is significantly reduced.
These briquettes can be stored outside, occasionally covered, without costly storage installations or
silos, so that there is a buffer for the seasonal fluctuations in production.
"Converting'' into a- or j3-hemihydrate (HH) or anhydrite as further refining steps creates other
possibilities to introduce FGD-gypsum to the construction industry. Several new installations are
already being built or in operation for the burning of 0-HH through the "gypsum boiling process" and
for producing a-HH through autoclaving.
UTILIZATIONS FOR OTHER FGD-PRODUCTS
As already stated above, some more FGD-residues are produced, depending on the type of combustion
and the FGD-process (see Figure 1).
In the past, there were only limited utilization possibilities for these residues, since their physical
properties and their chemical/mineral composition have a vast variation in comparison to the washed
FGD-gypsum.
8B-4
-------
ASHES OF FLUIDIZED BED COMBUSTORS AND DRY ABSORPTION PROCESSES
The actual utilization potential of these FGD-products lies in their self-solidification properties due to
their contents of calcium sulfate (anhydrite) as well as parts of silicate and aluminate, and the free
calcium oxide (CaO).
Listed are examples for utilizations which have been realized:
• Drying and solidification agent for various sludges (e.g. sewage sludges).
• Additives for producing mining cements.
• Utilization in road and dam constructions.
• Application as a soil improvement material and for ground stabilization.
By mixing different power plant residues with ashes of fluidized bed combustors, water and, if
necessary, small amounts of hydraulic bonding agents, "late"-supporting mining cements can be
produced which conform with the environmental and hygienic requirements of mining. The
compressive strength requirements for "Iate'-supporting construction cements of > 2 MPa after 28
days can be achieved.
"Early "-supporting mining cements (>20 MPa after 2 days) can also be produced by using conditioned
FGD-gypsum and cement with up to 20% (by mass) ashes of fluidized bed combustors.
FGD-PRODUCTS OF DRY AND SEMI-DRY ABSORPTION PROCESSES
As discussed previously, there are different flue gas treatment systems on a dry or semi-dry basis.
There is either a separate dust collection system for the fly ashes or the fly ashes are collected together
with the FGD-product at the end of the flue gas stream.
Due to the low temperatures, the low residence time, and the low moisture in a semi-dry absorption
system, calcium sulfite hemihydrate (CaSQj - 1/2 H2O) is the main component in the residues of a dry
or semi-dry FGD-process. The low temperatures are also the reason that calcium sulfate forms as
CaS04 - 1/2 H20.
In addition, the semi-dry absorption process, in comparison to the dry absorption process, eliminates
the flue gas component HC1 and HF almost completely. They are found in the FGD-residue as CaClj
and CaFj.
The utilization of these residues is difficult in comparison with the fly ash of fluidized bed combustors;
that is because of the vast variations in the composition and the above-mentioned components.
If the contents of reactive fly ash and unreacted absorbents (calcium hydroxide) is sufficient, those
residues have a limited self-solidification property which can be used to produce filling materials by
mixing it with fly ashes of fluidized bed combustors as discussed above. A more elegant utilization
for the products of a dry or semi-dry absorption system is possible if the following main components
can be used:
• Sulphur carriers such as calcium sulphite and calcium sulphate
• Calcium oxide
• Mineral components of the fly ash
8B-5
-------
FGD-PRODUCTS OF DRY AND SEMI-DRY ABSORPTION PROCESSES
An actual possibility, is the gypsum-sulfuric acid process (also known as Muller-Kuhne-process) which
is run by the "Schwefelsaure and Zement GmbH" (DSZ) in Wolfen (recently founded by GFR and
located in what was previously referred to as East Germany).
For over four decades, according to the classical process, natural gypsum or anhydrite was reduced
with coke and was cracked thermally. Together with additives which contain silicon oxides, aluminum
oxides and iron oxides (sand, clay, residual metals), cement clinker was burnt in a rotary furnace. By
adding milled gypsum as a retarding agent, they produced approximately 100,000 tons per year of
Portland cement.
The cement clinker produced is low on alkali and chromate and therefore has a very interesting
marketing potential.
The highly concentrated sulphur dioxide gas, which is set free during the reducing, thermal cracking
of the anhydrite (calcium sulphate), was purified, cooled, and converted into sulfuric acid (94 - 96%),
or oleum (approx. 100,000 tons per year) by an additional contact process. Lignite was used as
primary energy source.
Meanwhile, this process was modified so that the above-mentioned raw materials can be replaced by
suitable residues (Figure 8).
Particularly the main component, the natural anhydrite, was replaced by the products of the dry and
semi-dry absorption process. The specific composition of those FGD-products has all the components
which are necessary for the cement clinker and sulfuric acid production and furthermore, they need less
cracking energy and reduction material. A comparison of the main reactions (old and new) is shown
in Figure 9.
By using other residues such as used casting sands, fly ashes, and residues with a high caloric value
(e.g. acid resins, acid tars, used oils, plastics) it is possible to convert this old and inefficient process
plant into an economic one.
At this time the plant is brought up to Western environmental standards while in operation.
PROCESSING AND DISPOSAL
There are seasonal fluctuations in production of the FGD-residues in Germany with less summertime
demand for electricity at the same time that construction industry business activity is strong. This
means that in the summer there may not be sufficient FGD-residues to meet demand. Conversely,
during winter, there is a slow down in the construction industry and larger an*, v.nts of FGD-products
are available.
Depending on the location of the power plant and the processing plant, or poor quality of the residue,
it is important to have waste disposal possibilities in addition to the above-noted utilizations. Because
the various FGD-products appear in dust or slurry form, it becomes necessary to condition them prior
to their disposal.
8B-6
-------
PROCESSING AND DISPOSAL
The specifications for disposing the FGD-products are:
• sufficient stability and material suitable for transportation with trucks (minimum pressure
strength)
• minimum water permeability
• environmentally acceptable leaching levels of heavy metals.
These specifications can be met by producing specific mixtures of different residues and occasionally
adding water to fixate the material. This stable material is normally handled and compressed in a
slightly moist condition with bulldozers and vibrating rollers. For several days up to weeks, after the
disposal, pozzolanic reactions take place which finally lead to a disposal product which is
environmentally safe. This mono-disposal needs an additional mineral seal at the base (several layers
of clay) with a water permeability of less than 1 x 10"^ m/s. After the disposal site is filled, it also
needs to be sealed on the top. In case water permeability is too high, an evacuation system for the
infiltration water and a cleaning system is necessary.
GFR developed and executed two different disposal methods, dependent on the FGD-products. They
are discussed below.
FLY ASHES AND FGD-GYPSUM SLURRY
While FGD-gypsum and fly ashes of hard coal power plants are utilized widely in the Federal Republic
of Germany, as presented above, the residues of lignite-fired power plants are mostly disposed after
processing.
For processing and disposing dry fly ashes and the residues of wet absorption systems they are mixed
together and occasionally FGD waste water and calcium hydroxide is added.
When the fuel is German lignite, fly ashes with a free calcium oxide content up to 30% (by mass) are
obtained. By adding FGD-gypsum slurry or FGD waste water to those fly ashes, an exothermic
slaking reaction takes place. In these cases, slaking is to occur prior to the actual mixing.
8B-7
-------
FLY ASHES AND FGD-GYPSUM SLURRY
The silica and alumina components of the fly ash react after mixing with calcium hydroxide, calcium
sulfate and respective parts of water to produce cementitious mineral phases. These very complex
reactions are also known as pozzolanic reactions.
Pozzolanic Reactions (Simplified):
X Ca(OH>2 + Y SiOj (amorph) + Z H20 >
X CaO • YSiC>2 H20 (C-S-H = Calcium - Silicate - Hydrate)
Ettiingite Formation:
3 CaO + A1203 + 3CaS04 + H20 >
3 CaO • AI2O3 - 3 CaS04 • 32 H20 (Ettiingite)
These solidification reactions have the capability to, at least partially, immobilize heavy metals by
causing them to adhere to the crystal structure.
Products of Dry and Semi Dry Absorption Systems, Fly Ashes of Dry Additive Systems and Fluidized
Bed Combustors, and Other not Utilizable Fly Ashes.
These dry residues, which normally consist of varying fractions of fly ashes and FGD-products, as well
as excess absorbents (see above), are not disposable "as is", because they produce dust and have
uncontrolled leaching properties of hazardous components. Therefore, a special conditioning (mixing
the different FGD by-products together with water) is necessary to produce a disposable fixated
product.
CONDITIONING PROCESSES
Special mixtures of residues are made according to results of laboratory tests. An exactly defined
amount of water is added to achieve a good solidification. The product is disposed in a slightly moist
condition. A flow sheet of a processing plant is shown in Figure 10. Figure 11 shows an aerial
photograph of a processing plant which is operated by GFR.
The fixated product, processed as described above, is disposed by bulldozer in layers (Figure 12) and
is compacted several times by a vibration roller (Figure 13). This disposal technique and the occurring
solidification reactions meet the specifications (Table 6) of the permitting authorities. The necessary
quality assurance of the disposal end-product is continuously monitored by regulatory agencies and our
own test personnel. This supervision is done using drilling cores. Examination of those cores is done
for their density, single-axial compression strength, permeability of water, and the leachability.
These conditioning processes have been developed by GFR for various FGD-products and fly ashes,
which cannot be usefully utilized, or directly disposed otherwise in a guaranteed environmentally safe
manner.
8B-8
-------
SUMMARY
It was shown that the FGD-processes used in Germany result in various FGD-product compositions,
depending on the system used, which have various utilization potentials based on their specific
composition. While FGD-gypsum can be used almost entirely in the gypsum and construction
industries, other FGD-products have a limited utilization, mostly in the mining industry.
For certain dry and semi-dry absorption by-products, a recently developed high-quality utilization
potential exists, using the components contained in the residues for the production of cement and
sulfuric acid.
In conclusion, it should be emphasized that even with all the utilization possibilities, it is advisable to
plan for safe disposal of some residues. To achieve this, process technologies are necessary which take
into account the properties of the various FGD-products and fly ashes.
8B-9
-------
REFERENCES
1. Residues from Flue Gas Cleaning (in German) Insert 29, "Mull und Abfall" (publ. by
D. O. Reimann/J. Dcmmich), Erich Schmidt Verlag Berlin, 1990
2. Ministry of Environment Baden Wurttemberg: Disposal of Residues from Flue Gas Cleaning (in
German) - Part H: GroBfeuerungsanlagen - Issue 1 "Luft, Boden, Abfall", 1988
3. Winkler, Gruber, Hammerschmid, Rentz: Disposal of Residues from Flue Gas Cleaning at Large
Combustion Plants Cement - Lime - Gypsum (in German), 41st Year, No. 11, pp. 576-582, 1988
4. VDEW/VGB-GemeinschaftsausschuB "Residues and Waste Materials" (in German); Utilization
Concepts for Residues from Coal-burning Power Plants - REA-Gips, March 1986
5. VDEW/VGB-GemeinschaftsausschuB "Residues and Waste Materials" fm German); Utilization
Concepts for Residues from Coal-burning Power Plants - Aschen, September 1988
6. M. Hildebrand: Status of Flue Gas Cleaning in the Western States of the Federal Republic of
Germany VGB-Seminar "Flue Gas Cleaning and Disposal of Residues in Power Plants, Industry
and Heating and Power Plants" (in German) Cottbus, November 29 - 30, 1990
7. W. vom Berg: Utilizing Combustion Residues (in German) Publication, see 6.
8. J. Beckett: Comparison of Natural Gypsum and FGD-Gypsum (in German) VGB-Conference
"Kraftwerk und Umwelt 1989", Essen
9. F. Wirsching: Utilization of FGD-Gypsum Publication, see 6.
10. J. Demmich, E. WeiBflog: Utilization and/or Disposal of the Semi-Dry Absorption Process
Product (in German) Publication, see 6.
11. F. Risse, et al. By-products from Coal-Buming Power Plants and Residues from Garbage
Incinerating Plants in the Federal Republic of Germany (in German) VGB Kxaftwerkstechnik 71
(1991) Issue 5, pp. 504 - 508
12. E. WeiBflog, J. Demmich: The Disposal of Ashes from Fluidized Bed Combustion and/or Flue
Gas Desulphurization (in German) VGB-Conference "Kraftwerk and Umwelt 1989", Essen
13. F. Wirsching and E. WeiBflog: Environmentally Safe Disposal of Residues from Hard Coal
Burning Power Plants (in German) VGB-Kraftwerkstechnik 68, Issue 12 (Dec. 1988)
14. F. Wirsching in Winnacker-Kuchler, Vol. 3 Anorganical Technology II (in German) (Carl-
Hauser-Verlag Munchen - Wien 1983), pp 262 - 275
8B-10
-------
TYPES OF FLUE GAS CLEANING RESIDUES
IN REFERENCE TO THE CLEANING PROCESS
EMISSION CONTROL MEASURE RESIDUES
OUST FILTER
FLY ASHES
\ FA* FROM FBC
FA FROM DAP
DSP RESIDUE
DRY ABSORPTION PROCESS
} SEMI DRY ABSORPTION PROCESS
SDAP RESIDUE
FGO GYPSUM
WET ABSORPTION PROCESS
FGD WASTE HgO
DRY ADDITIVE PROCESS
FBC' WITH ADDITIVE
' FLIWISCD BCD COWUSIIOH
7 rLY ASCS
Figure 1.
8B-11
-------
METHODS OF FLUE GAS PURIFICATION (EXAMPLES) AND
RESPECTIVE RESIDUES OF COAL FIRED POWER STATIONS
COAL
1000 kg
DRY
COMBU8TION
8LAQ
jso - 70 kg
FLUIDIZED BED
COMBU8TION WITH
DRY AB80RPTI0N
BOTTOM A8HI 30 - SO kg
D.8. • DRY 80LI08
R.P. • REAOTION PR00UCT8
F.A. ¦ FLY ASH
AB. • ABSORBENT
DEDU8TING
F.A. 40 - 100 kg
4 (D.8.)
SEMI-DRY
ABSORPTION
WET ABSORPTION
FQD GYPSUM 146 kg (D.8.)
DEDUSTING
F.A./R.P./AB
176 - 130 kg
(D.8.)
DEDUSTING
F.A./R.P./AB. 110 - 140 kg
I (D.8.)
Figure 2.
-------
WAYS FOR THE USE OF FGD-GYPSUM
GYPSUM WALLBOARD
CONSTRUCTION PLASTER
CONSOLIDATION REGULANT
FOR CEMENT PRODUCTION
FGD-GYPSUM
CaS04*2H20
LANDFILL, RECULTIVATION,
ROAD CONSTRUCTION
MORTAR FOR CONSTRUCTION
IN UNDERGROUND MINES
FILLING BY
PNEUMATIC TRANSPORT
Figure 3.
8B-13
-------
-------
Figure 6. Rotating Briquetting Press
Figure 7. Use of Processed FGD-Gypsum E.G. for Mining-Construction
8B-15
-------
FLOW CHART : GYPSUM-SULFURIC ACID-PROCESS
WITH CHEMICAL/ENERGETICAL USE OF RESIDUES
STAKE SULFURIC ACID
RESIDUES RICH IN
HEAT OF COMBUSTION
(ACID TARS)
AND
LIGNITE COAL
FGD-GYPSUM
SLURRY
CONTACT I
PROCESS|
S02
FGC-PRODUCT3 iDOSAGEh
PROCESS GAS | so2/
PURIFICATION i plUE GAS
CASTING SAND -»-JDOSAGEl.
CEMENT
CLINKER
FLY ASH FROM
LIGNITE COAL
BURN-UP
(Fe304)
C-CONTAINING
FLY ASH
CEMENTS
PORTLAND. BLASTFURNACE.
-~^DOSAGEH
FLUE GAS
PURIFICATION
MIXING
(RAW MATERIAL
PRODUCTION)
ADDITION BY MILLING
- GYPSUM
- BLASTFURNACE SLAG
- AND/OR FLY ASHES
REDUCTION AND/OR
THERMAL DECOMPOSITION
CLINKER PRODUCTION
f (ROTARY KILN)
FLY ASH CEMENT
Figure 8
REDUCTION/DECOMPOSITION OF CaSQ4
2CaS04 * C * 2Si02—~- 2CaSi03 * C02 * 2S02 - ENERGY
COMPARABLE REACTIONS OCCUR WITH AI203 OR Fe203.
DECOMPOSITION REACTIONS OF CaSQ3«1/2SQ2
1) » 300 C INERT GAS
CaS03* 1/2H20 ~ CaS03 - 1/2H20 (DEHYDRATATION)
2) > 600 C
CaS03 ~ CaO - S02 (DECOMPOSITION)
3) > 680 C
4CaS03 » 3CaS04 * CaS (DISPROPORTIONING)
4) > 780 C
3CaS04 + CaS ~ 4CaO - 4S02 (COMPROPORTIONING)
Figure 9
8B-16
-------
PROCESSING PLANT FOR DISPOSAL
INSTALLATIONS
FOR
DECHARGING AND
DOSAGE
PUG MILL
STABILISATE
PRODUCTS
FGD-
FLY ASHES
Figure 10
8B-17
-------
Figure 11: Processing Plan: for FGD-Stabilisate
83-18
-------
Figure II: SDrcacir;^ oi" Siabiissaie bv Bu'.lcozer
SB-19
-------
HARD COAL
FIRED P.S.
LIGNITE COAL
FIRED P.S.
OIL
FIRED P.S.
TOTAL
TOWER/MEGA-
WATTS (ELEC.)
26.090
10.950
354
38.194
P.S. = POWER STATIONS
WET
PROCESSES
SEMI-DRY
ABSORPTION
DRY ADDITIVE
FLUIDIZED BED
COMBUSTION
QUANTITY OF
TOTAL POWER
SUPPLY (%)
1
86
7,7
3,7
Table 1: Contribution of Different Types of Power Stations of Power Supply Companies to Flue
Gas Desulfurization and Distribution of Flue Gas Purification Systems in Respect of
Total Electricity Output
GRANULATE
COARSE AS^ES
(M10 T/A)
FLY-ASHES
(HIO T/A)*
RECYCLING
RATE
m
HARD COAL
FIRED POWER
STATIONS
3,17
3,11
90, 6
LIGNITE COAL
FIRED P.S.
1,21
4,87
2,1
FGD—
GYPSUM
(HIO T/A)*
RECYCLING
RATE
%
FBC
SDA+
MIO T/A
JAP
HARD COAL
FIRED P.S.
1.0
> 90
APPR.
0,39
APPR.
0,35
APPR.
0,08
LIGNITE COAL
FIRED P.S..
1,2
-
-
-
-
Table 2: Residue Output From Combustion of Coal and Flue Gas Purification
8B-20
-------
NATURAL GYPSUM
FGD-GTPSUN
MOISTURE
WEIGHT-*
2
7
10
P„-VAL'JE
WEIGHT-*
6
7
5
8
CnS04*7H70
HEIGHT-*
7B
95
98
99
CI
HEIGHT-*
<
0,01
-
0,01
NOjO
WEIGHT-*
0,02
0,01
INERT
COMPONENTS
WEIGI1T-*
5
20
-
As
PPM
0,22
3,79
0,21
2,70
Pb
PPM
0,46
7.1,1
0,27
22,0
Cd
PPM
0,03
0, 30
0,003
0,29
Cr
PPM
0,65
24,9
1,02
9.72
Ni
PPM
0,3
13,4
0.3
12,9
llg
TPM
0,006 -
0,05
0,03
1,32
Table 3: Main and Trace Components of Natural Gypsum in Comparison to FGD-Gypsum
PARAMETER
DAP AS1I
FRC ASH
DSP PRODUCT
SDA PRODUCT
FLY ASH
RESIDUAL C
»JaS04
25 - 80
3 -25
3 - 25
50 - 80
3 - 25
15 - 30
1 - 10
0.1- 2
3 - 85
0.5 - 10
<.:aS04«HII20
«:aso,«wuo
CaCO,
3 - 10
3 - 10
5 - 25
5 - 30
10 - 30
5 - 35
5 - 70
5 - 30
CaO
Ca (Oil) 2
CaC 12
3 - 30
0.1 - 0.5
7 - 25
0.1 - 1.0
10 - 30
2-1
1 - 15
2-8
Table 4: Main Components of Residues from Dry Rue Gas Cleaning Process
(Typical Extent of Fluctuation)
8B-21
I
-------
REQUIREMENTS OF GYPSUM INDUSTRY
FOR FGD-GYPSUM QUALITY
FREE MOISTURE
CaS04*2H20
MgO WATER-SOLUBLE
CHLORIDE
Na20
SULFUR DIOXIDE (S02)
pH-VALUE
COLOR
ODOR
TOXIC COMPONENTS
« 10* (WEIGHT PERCENT)
» 96%
« 0.1*
< 0.01*
< 0.06*
< 0.26*
6-9
WHITE (DEGREE OF
WHITENESS;80*)
NEUTRAL
NONE
Table 5
Disposal-Site-Requirements for Processed FGD-Residues
1. _>_ 30 Weight-% Fly Ash of Sufficient Pozzolanic Reactivity
2. >_ 3 Weight-% Ca^H^CaO
3. Optimal Field Density >1.0 Gramms/CM^ (Dry Solids)
4. Solidity, Useability for Vehicles Right After Compression, Setting Time
5. Unconfined Compressive Strength _>_ 0,5 MPA (28 D)
6. Permeability: K < 5.10"® m/s
7. Leachability: Testing Procedure Dev-S4; Class II Disposal Site Category (For
Mineral Residues) Due to the Proposal of a Guide-Line For North-Rhine Westphalia,
Pan 2, 1987
Table 6
8B-22
-------
The Elimination of Pollutants from FGD Wastewaters
8B-23
-------
8B-24
-------
M. K. Mierzejewski
Infilco Degremont Inc.
2924 Emerywood Parkway
Richmond, Virginia 23294
ABSTRACT
Limestone forced oxidation systems producing saleable gypsum and throwaway wet FGD
systems operated in an open loop configuration both yield wastewater streams. In
order that pollutants, such as heavy metals, are not merely shifted from the gas
phase to the liquid phase, special treatment techniques must be applied to these
wastewaters.
Over the last 6 years a specific treatment process has been developed to eliminate
pollutants from these wastewaters, and 22 such treatment plants are currently
operating in Europe, with 2 more under construction. The first such plant in the
U.S. is due to be started up in 1992.
This paper reviews the factors influencing the composition of these wastewaters and
describes the technology of their treatment.
Preceding page blank
8B-25
-------
INTRODUCTION
The latest amendments to the 1970 Clean Air Act will undoubtedly lead to an increase
in the number of flue gas desulfurization (FGD) installations at fossil-fuel fired
power plants. In most cases it appears likely that the sludge produced in either
dry or wet FGD systems, will be disposed of to landfill, i.e., the FGD system will
be of the "throwaway" sludge type. At some power plants, certain factors may allow
selection of an FGD system which produces gypsum as a marketable product, precluding
the need to landfill any sludge. Under such circumstances, a wastewater stream will
be produced, requiring appropriate treatment to eliminate pollutants so that it may
be discharged into the environment or, if this is restricted, to be treated to a
higher standard and reused on the power plant.
The wet limestone FGD system which produces a marketable product incorporates forced
oxidation of calcium sulfite, the intermediate by-product, to calcium sulfate
dihydrate, CaS04.2H20, known as gypsum. Such a system is termed LSFO (limestone
forced oxidation), sometimes with the designation WB for "wallboard", the
predominant use of such gypsum.
The gearing-up of the utility industry to address tighter S02 and NO. discharge
limits in the U.S. is similar to that seen earlier in West Germany, with the Federal
TA Luft and GFAVO programs started in 1983. By July 1, 1988, West German stations
larger than 110 MW were required to meet new S02, NO, and flyash limits. This
represented a relatively short period, similar to that now faced by U.S. utilities,
with the Phase I deadline of the Clean Air Act Amendments being January 1, 1995. In
total, Phase I will affect 110 power plants in the U.S., with a combined capacity of
almost 90,000 MW. Phase II will impose stricter emission limits still for the start
of the next century.
In West Germany, 86% of FGD installations are wet and the majority of these are of
the LSFO-WB type"'. The shortage of available land area for disposal essentially
pushed all the utilities toward this technology and, since there were (and still
are) strict controls on aqueous discharges, there was a parallel growth and
development of appropriate FGD wastewater treatment technology.
8B-26
-------
The demand for gypsum on the one hand and the cost of landfilling on the other will
be strong determinants of how many LSFO-WB installations will be seen in the U.S.
Of the utilities that have started seriously evaluating different FGD technologies,
about six to date have elected to use this technology on selected power plants. In
each case, the problem of wastewater treatment will need to be addressed.
ORIGINS OF WASTEWATER
Except on small units, cost analyses have shown that dry scrubbing is not an
economic means of achieving compliance<2). Although capital costs are generally
lower than wet limestone systems, overall operating costs, in terms of $/ton S02
removed, are high, due to high reagent consumption. In addition, dry scrubbing may
not a feasible economic option where the existing electrostatic precipitators (ESP)
are unable to remove the additional solids load, necessitating costly modification
or replacement.
It is likeiy that most of the stations on the Phase I list which will be installing
FGD equipment will adopt wet scrubbing alternatives. There are many different wet
scrubbing technologies available but, from the point of view of wastewater
(discounting processes like the Wellman-Lord), the choice comes down to whether the
sludge produced is disposed of to landfill, i.e., a throwaway process, or is washed
and sold as a useful by-product.
A system producing a throwaway sludge can be designed either to operate at a
relatively low chloride concentration in a so-called "open loop", in which case a
wastewater stream is produced, or at a higher chloride concentration, in a "closed
loop", without the inconvenience of the wastewater (see Figure 1). In both cases
water is entrained in the sludge formed in the scrubber. The sludge may be
dewatered to, say, 80% dry solids (d.s.), with the 20% residual moisture having the
same composition as the liquid phase in the scrubber. Hydrogen chloride is present
in the flue gas, and is taken out under the wet alkaline conditions in the scrubber.
Without a chloride blowdown, this species would build up, adversely affecting the
efficiency of S02 removal<3). In a closed loop, the chloride concentration in the
scrubber is elevated so that the moisture entrained in the sludge represents the
chloride sink, rendering a separate wastewater stream unnecessary. Open loop
operation, conversely, does not allow sufficient chlorides to be "lost" from the
system without an additional blowdown from the scrubber, and it is that blowdown
which represents the wastewater stream.
8B-27
-------
The apparent advantage of not requiring wastewater treatment in closed loop
operation may be out-weighed by the cost of more expensive materials of construction
of the scrubber and all associated wetted parts, in order to withstand the more
corrosive environment due to the higher chloride concentration. Which option is
selected at a particular installation depends on case-specific factors.
Limestone reacts with S02 in the scrubber to form calcium sulfite which, on further
oxidation (if selected) forms gypsum. The reaction product is continuously removed
from the scrubber and fresh limestone is returned to the reservoir. In both, open
and closed loops the dewatering of the sludge gives rise to a liquid stream: in an
open loop only a portion of this stream, depending on chloride content, cake
dryness, etc. can be returned to the scrubber, in the form of slurry make-up,
without jeopardizing the chloride balance. In closed loop operation the complete
stream can be returned.
This "open loop versus closed loop" choice does not exist in LSFO-WB systems, since
current requirements in the U.S. call for the saleable product to have no more than
120 mg chloride per kg of dry solid. This requirement, together with the fact that
a much higher dryness (90% d.s. minimum) is demanded by the gypsum user, means that
the sludge cannot be used as a chloride sink. Saleable gypsum must be washed
thoroughly to achieve the purity standards required: the washings, being of low
salinity, can be re-used in the scrubber. Figure 2 shows sources of wastewater on
an LSFO-WB scrubber.
In summary, a throwaway system may or may not produce wastewater, whereas a system
producing a saleable cake always will.
FGD wastewater flows are typically low, and for a given unit size are related to the
chloride concentration in the scrubber and to the calorific value and chloride
content of the coal being burned. Typical values lie in the range 0.05-0.2 gpm/MW
scrubbed. Other than sulfur, the majority of which is taken out in the scrubber,
the wastewater contains most of the pollutants present in the coal, limestone and
makeup water. In order that it can be discharged into the environment, certain
chemical species, particularly heavy metals, must be removed.
FACTORS DETERMINING FGD WASTEWATER CHARACTERISTICS
FGD wastewaters are typically acid, highly saline solutions with varying quantities
of suspended solids, heavy metals, chlorides, fluorides and COD. Table 1 shows
8B-28
-------
typical compositions before and after treatment, taken from numerous operating
Degremont installations. Since FGD operates on a continuous basis, the wastewater
is produced and treated likewise, such that only limited cooling occurs as it passes
through the gypsum dewatering stages and collects in an equalization tank.
Treatment is normally carried out at about 40-50°C.
The pH of the wastewater usually li=s in the range 5.0-6.5 S.U. Buffered systems
using carboxylic acids are designed to maintain a pH of 4.8-5.5 S.U., the optimum
for the reaction of sulfur dioxide with limestone.
Coming as it does from a warm gypsum-rich environment, the wastewater is normally
supersaturated with gypsum when it reaches the wastewater treatment plant (WWTP),
although the degree of this supersaturation depends to a large extent on the nature
of and retention time in the concentrating and dewatering stages upstream of the
WWTP proper. Magnesium, aluminum and iron are all elements of major importance in
the design of a FGD wastewater treatment facility. Since pH elevation is a key
process in treatment, these metals will come out as hydroxides, and represent a
solids load to the clarifier/thickener.
The high concentration of chloride (10-40 g/1) at which the scrubber operates is the
same as that in the wastewater, balanced principally by calcium and magnesium
cations. The resulting corrosive property of the wastewater has considerable impact
on the WWTP materials of construction. Another halogen, fluorine, is present in
coal and also volatilizes as the hydrogen compound; however, its lower concentration
in coal and the low solubility of its calcium salt render it only a minor
constituent of FGD wastewaters.
The Chemical Oxygen Demand (COD) of FGD wastewaters, typically about 100-150 mg/1,
is due largely to unreacted calcium sulfite (though this contribution is low) and
organics present in the makeup water.
The quantity and quality of the wastewater is determined by the following factors:
• rated capacity of the unit boiler
• scrubber chloride concentration
• efficiency of flyash removal
• type and efficiency of dewatering installation
• type of FGD process
• chemical composition of coal, limestone and makeup water
8B-29
-------
Rated Capacity of The Unit Boiler
All other factors being equal, the wastewater flow will be directly proportional to
the unit capacity.
Scrubber Chloride Concentration
Similarly, the flow is directly proportional to the chloride concentration at which
the scrubber operates. The higher design chloride concentrations seen in recent
years has led to the need for smaller wastewater plants.
Efficiency of Flvash Removal
A poorly optimized ESP will allow large quantities of flyash to pass into the
scrubber, with the attendant risk of contaminating the gypsum to the extent that it
is not marketable. However, even a wel1-operated ESP will have difficulty in
intercepting small flyash particles, in the 5-10 um range, which will then be taken
out in the wet scrubber, from where they may pass into the wastewater purge.
There is an additional effect caused by the passage of fine flyash particles into
the scrubber. Many heavy metals appear in the aqueous phase of the wastewater by
volatilizing during the combustion of the coal and then, in the boiler outlet duct,
condensing on flyash particles. The metals are then leached from the flyash at the
low pH conditions encountered in the scrubber. This effect is enhanced by the fact
that the small particles, unintercepted by the ESP, have high specific surface areas
for the condensation of these elements1*'.
Type and Efficiency of Dewaterinq Installation
Dewatering of the gypsum slurry is achieved by belt presses, vacuum filtration or
centrifugation, usually preceded by a pre-concentration stage such as hydrocyclones.
Since the overflow from these units represents the flow to the WWTP, their operating
efficiency determines the total suspended solids (TSS) in the influent. In
particular, flyash particles, with a density similar to water, frequently escape
capture in these units and appear in the wastewater.
8B-30
-------
Type of FGD Process
The use of chemical additives to modify the conditions inside the scrubber, for
example to inhibit sulfite oxidation or to enhance S02 absorption, has been used on
several FGD systems. An example of the latter is the use of dibasic acid (DBA), a
mixture of adipic, glutaric and succinic acids which, although more frequently used
in natural oxidation or inhibited systems, has been proposed for LSFO. In Germany,
one FGD supplier uses formic acid, at concentrations of about 1 g/1, in a similar
way to DBA, i.e., as a pH buffer to obtain optimum S02 absorption. Being in.
solution, these carboxylic acids appear in the wastewater, considerably elevating
its COD.
If, in addition to desulfurization, the flue gas is treated to remove N0X through
the use of, for example, Selective Catalytic Reduction, ammonia is added to the flue
gas stream. Although the process is optimized so as to minimize ammonia wastage,
some ammonia will pass through the DeNO, system and be removed in the air preheater
and the ESP. The small amount remaining after the ESP will be taken out in the
scrubber, giving rise to 50 mg/1 ammonium ion or less in the wastewater.
In installations where there is a prescrubber ahead of the limestone scrubber, the
pH of the blowdown will be very low (0-2 S.U.) with a high chloride, fluoride and
heavy metal content, in addition to high TSS levels due to flyash (up to 6 wt.%).
Chemical Composition of Coal. Limestone and Makeup Mater
As would be expected, the compositions of the coal, limestone and makeup water
determine the chemical analysis of the FGD wastewater influent. The coal is, of
course, the primary source of sulfur, chloride and many metals which appear in the
wastewater. The limestone similarly affects wastewater composition, particularly as
regards its aluminum, iron and magnesium content. The limestone purity will also
affect the TSS in the wastewater since impurities such as fine clay particles can
form suspensions which are difficult to remove during the gypsum dewatering phase.
In addition to being used in slurry preparation and mist elimination, makeup water
is added to the scrubber to compensate for evaporation losses. The contaminants in
this water are concentrated by evaporation in the power plant cooling system — if
this is the source of the water — and in the scrubber itself.
8B-31
-------
Availability of Analyses
The numerous factors above would suggest that FGD wastewaters can be quite variable
in composition and this is, in fact, the case. It is also in the nature of the FGD
projects that the WWTP has to be designed at about the same time as the FGD system.
Consequently, the design of the WWTP has to be based on the known analyses of coal,
limestone and make-up water. Only a water treatment company with a wide experience
of treating FGD wastewaters can determine the appropriate treatment process and
provide meaningful discharge guarantees. Extensive tests will be performed at the
new FGD installation at Northern Indiana Public Service Company's Bailly Station in
1992-1995 under the Department of Energy's Clean Coal Technology Program(s>. These
tests will evaluate different coals and should generate a wealth of information on
how this parameter affects the wastewater.
TREATMENT OPTIONS
The degree of treatment of the wastewater is determined by the National Pollutant
Discharge Elimination System (NPDES) permit discharge limits imposed t/ the regional
EPA. Since these limits are getting progressively tighter, increasingly higher
levels of treatment are being required.
The minimum level of treatment encountered has been the adjustment of pH to 7±2 S.U.
and the removal of suspended solids to less than 30 mg/1. Most NPDES permits do
however limit heavy metal concentrations and, in addition, there may be limits on
fluoride, sulfide, ammonia, COD, etc. Where the power plant is located close to the
sea, the chloride content of the discharge, being of the same order of magnitude as
that of seawater, about 20 g/1, has not been a problem. To date, this same
reasoning has been applied to inland stations located on large lakes or rivers.
However, some installations, because of state requirements or their specific
locations, are having to address the removal of chlorides, and hence total dissolved
salts (TDS), from the blowdown.
The options of wastewater treatment are as follows:
i) treat the FGD WW in a dedicated physico-chemical treatment plant, using
coagulation, sedimentation, etc.
ii) treat the FGD WW by evaporation/crystallization.
8B-32
-------
The choice between these options is straightforward, with evaporation/
crystallization only being selected where there is an absolute zero water discharge
requirement. The advantages of evaporation/crystallization, namely its small
footprint and the fact that it produces high quality water reusable elsewhere on the
power plant, should be balanced against the disadvantages of very high power
consumption and high sludge production. More importantly, there is little
experience in the evaporation of FGD wastewaters at present; indications are that a
high level of pretreatment may be required.
The possibility of co-treating the FGD wastewater with other power plant wastewaters
should only be considered if the required treatments are complimentary and guarantee
effluent values can be met. In particular, there is a risk of heavy metal
complexation in co-treatment, making them difficult to remove.
Host of the emphasis to date has been on the elimination of pollutants physico-
chemically in dedicated treatment plant and numerous such installations exist in
Germany and Japan — although these differ greatly in concept (compare (6) and
(7)) — with the first due to start up in the United States in 1992.
ELIMINATION OF POLLUTANTS BY PHYSICO-CHEMICAL MEANS
The "complete" WWTP described below will render the water non-scale forming,
substantially remove cadmium, mercury, nickel, copper, zinc, lead and other metals,
and clarify it prior to discharge. The treatment scheme was developed by the
Philipp Mueller Company of Degremont Group in Europe through a combined use of
related technologies. Experience in the treatment of brines, the removal of heavy
metals, fluorides, etc., together with numerous designs of clarifiers, thickeners
and filters, was used in developing a design for the treatment of these unique
wastewaters. Over the last 6 years some 22 installations have been started-up and
the process scheme has been refined to that shown in Figure 3.
The first stages of treatment, namely oxidation, pH elevation/desaturation, heavy
metal precipitation and coagulation are carried out in individual reaction tanks,
vigorously mixed by axial flow agitators. This is important in view of the
different retention times and mixing energies required.
The reaction tanks provide a series of complimentary chemical environments, where
reagents are dosed either proportional to flow or to adjust the pH, precipitating
various species from solution. The wastewater leaving these tanks contains all
8B-33
-------
influent solids, precipitated products and recycled gypsum solids (see later).
Typical chemical dosing is shown in Table 2. The next stage of treatment provides
for the settling of these suspended solids and hence the clarification of the water.
Other stages of treatment include the conditioning and dewatering of the sludge
settled in the clarifier, and final pH adjustment. In some European installations,
COD is removed by adsorption on granular activated carbon (GAC).
It has been found that the optimum design for the system is to elevate the reaction
tanks such that the water flows by gravity through the complete plant; in this way
floe particles are not destroyed by re-pumping and are able to settle well in the
clarifier downstream. With only one fixed recycle and this type of WWTP is easy to
operate.
The treatment stages are described more fully below.
Treatment Stages
Oxidation. This stage will only be required if the influent sulfite concentration
exceeds 100 mg/1 SO,2* and there is a discharge limit for COD. In cases where
oxidation of calcium sulfite to sulfate in the absorber is incomplete, this reaction
can be brought near completion by blowing air through the liquid in the first
reaction tank. The approximate sizing of this, and other, reaction tanks is given
in Table 3.
Elevation of pH/Desaturation. Typically, FGD wastewaters have pH values in the 5.0-
6.5 S.U. range. In this stage, an alkali, either calcium hydroxide or sodium
hydroxide, is added to obtain pH 8.5-9.2. If the wastewater originates from a
prescrubber, and is highly acid, pH adjustment is achieved in two stages (two
tanks): the pH in the first tank is adjusted to about pH 4.3, and in the second tank
to pH 8.5-9.0, the optimum pH for treatment.
Since pH elevation is an important part of the wastewater treatment process, the
concentrations of the abundant metals which form insoluble hydroxides and
oxyhydroxides (i.e., aluminum, magnesium and iron) are important in the sizing of
the treatment facility, as they represent a major solids load to the
clarifier/thickener. Host of the heavy metals also present in the waste stream are
removed at these higher pH values, but their much lower concentrations represent a
negligible load.
8B-34
-------
The other function of this tank is the "desaturation" of gypsum from the wastewater,
since this compound has a tendency to supersaturation. If not brought to
equilibrium, this supersaturation can exist for several hours; when equilibrium is
achieved, calcium sulfate comes out of solution, causing severe scaling in the
clarifier, pipes, etc. In order to accelerate the rate of desaturation, and to have
it take place in a controlled manner, "young", thin sludge is returned from the main
clarifier/thickener to the reaction tank. This sludge flow rate is equal to about
50% the influent flow, so as to maintain a minimum concentration of 2.5-3.0 wt.%
solids in the tank. This sludge contains already-formed gypsum crystals which act
as crystallization nuclei for the calcium sulfate in solution. In this way, gypsum
"scales" sludge particles rather than the equipment.
Heavy Metal Removal. The removal of heavy metals as their hydroxides alone would
not allow the strict regulatory requirements to be met. Since the solubility
products of all heavy metal sulfides are lower than their corresponding hydroxides,
it follows that the addition of sodium sulfide, for example, should allow stricter
standards to be met. In fact, rather than this reagent, an organosulfide such as
sodium trimercapto-s-triazine is used. These organosulfide-heavy metal complexes
have very low solubility products, of the same orders of magnitude as the inorganic
sulfides. One advantage of using the organosulfide is that it is non-toxic. Unlike
sodium sulfide, whose use must be carefully controlled to ensure that the toxic
sulfide ion does not escape in the discharge, there is no risk with this reagent if
dosed in excess.
As well as allowing better removal, this two-stage precipitation of heavy metals,
both as hydroxides and sulfides, has an added advantage. Except in the case of
metals such as zinc and lead, a higher pH will tend to remove more of the heavy
metal from solution — an argument for using as high a pH as practicable. However,
at about pH 10, all the magnesium in solution would precipitate, and even so, at the
selected pH of 8.5-9.0, some magnesium hydroxide, Mg(0H)2, does precipitate,
representing additional solids to be dewatered. Since Mg is not usually subject to
discharge controls and, further, Mg(0H)2 is difficult to dewater, keeping the pH
lower rather than higher is beneficial. Selection of a limestone with a low MgCO,
content is also important since each gramme of magnesium introduced by the limestone
yields 2.42 g of Mg(0H)2 sludge.
Coagulation. Iron (III) chloride is used as the mineral coagulant of choice in FGD
wastewater treatment in preference to aluminum sulfate because of its wider pH range
of application and the denser floe it produces. As this reagent hydrolyses to form
8B-35
-------
iron (III) hydroxide, other metals are co-precipitated and some soluble organic
matter is destabilized.
Flocculation. Clarification and Thickening. The latest WWTP designs use a combined
reactor-clarifier-thickencr for the separation of influent solids and precipitation
products from the wastewater'*'. In the reaction zone, there is a high energy input
where sludge is recycled and densification occurs. The fast-settling nature of this
dense sludge enables high rise rates (and hence smaller equipment) to be used in
clarification. The solids settle in the raked thickener portion, while the
clarified water passes through honeycomb-type lamellar modules, which trap any rogue
particles that may not have settled. Effective gypsum desaturation is essential if
these lightweight modules are to function well, and not act as a matrix for scale
deposition.
The sludge composition determines the surface loading and rise rate used for the
thickener design. Typically, sludges with a high metal hydroxide content (greater
than ca. 30%) will require low solids loadings and rise rates of approximately 0.1
gpm/sq.ft., while those with a high gypsum content can be applied at rise rates of
up to 0.4 gpm/sq.ft. and higher loadings. Sludge is withdrawn from the thickener
unit in two zones using progressing cavity pumps: older, thicker sludge is removed
from the cone bottom of the unit for dewatering, while younger sludge is removed
higher up the cylinder to be returned to the desaturation stage.
COD Removal. If required, COD removal is effected after clarification. While the
sulfite can be removed by oxidation, the organics need to be removed by adsorption
through GAC. Where carboxylic acids are dosed, and are responsible for high COD
values, they can be removed biologically'".
dH Adjustment. The pH value of the wastewater is re-adjusted prior to discharge
using hydrochloric or sulfuric acid.
Sludge Dewatering. The 10-20 wt. % sludge extracted from the thickener is
transferred to an agitated sludge holding tank where lime, if available, is dosed to
10-15% by weight to aid dewaterability and, as has been found with other power plant
wastewaters00', to fix the heavy metals in the sludge once it has been landfilled.
Dewatering is achieved by filter presses operating at 225 psig (15 bar g) so as to
produce a cake of 50-55 wt.% dry solids content for easy truck disposal.
8B-36
-------
Materials of Construction
1 The precautions taken against corrosion in the scrubber, due to the high chloride
environment of FGD, have similarly to be adopted in the design of the WWTP.
All wetted parts (tanks, pipework, mixers, pumps) should be rubber-lined carbon
steel, or FRP. In the case of the former, precautions have to be made against oils
and greases in the wastewater (which in any case are unlikely to exceed 5 mg/1) as
these can perish the rubber lining. In the case of FRP, this should be resistant to
abrasion by gypsum and flyash particles. Where rubber lining of wetted parts is not
possible, as for example with progressing cavity pump rotors or knife gate valves,
then special alloys (e.g., Hastelloy C) must be used.
The high suspended solids concentration of the wastewaters, as well as its corrosive
nature, necessitates the use of appropriate instrumentation such as magnetic flow
meters, diaphragm type pressure indicators and suitable self-cleaning pH probes.
OTHER ISSUES
I Zero Discharge
Pilot tests are scheduled in Europe later this year on the evaporation of FGD
wastewaters using a vertical tube falling film unit. To date, very little
information is available on the treatment of FGD scrubber blowdown by this means,
although it is known that the requirements for a very low influent TSS and a non-
scaling characteristic will necessitate some pretreatment. Further, if the salt
reject is evaporated to give a dry cake, which may perhaps be used as road salt,
prior heavy metal removal will be required. In effect, then, the zero discharge
option may mean a complete physico-chemical plant, as described above, preceding the
evaporator/crystal 1i zer.
Biological Treatment
If organic acids start being widely used for LSFO systems, then WWTP will have to be
adapted to allow for treatment to remove the resultant COD. So far, two
installations in Germany, Bexbach and Fenne, have biological trickling filters after
physico-chemical treatment, to remove formic acid dosed in the FGD process. Pilot
tests conducted prior to installation of these treatment stages indicate removals in
8B-37
-------
excess of 90% from influent levels up to 1,000 mg/1 COD, with chloride
concentrations up to 9,000 mg/1.
With the expected higher chloride concentrations in modem FGD systems, further work
will have to be performed on COD removal in FGD wastewaters. Experience in the
treatment of other wastewaters indicates that inhibition of biological processes
starts at about 10,000 mg/1 chloride.
Sludge Disposal
Unlike the gypsum produced in the scrubber which, if 9556 pure CaS04.2H20 and
complying with other purity standards, can be sold, the sludge from the WWTP needs
to be disposed of to waste. Transportation and lining costs, as well as licensing
issues need to be addressed. The costs associated with such disposal partially
offset the savings made through not having to landfill the gypsum from the scrubber.
Some tests have been performed in Europe where the WWTP sludge was dumped on the
coal pile, combusted and then taken out with the bottom ash.
SUMMARY
Although some of the stations required to achieve Phase I compliance by January 1,
1995, will use an FGD process producing saleable gypsum, it is not clear how
widespread this technology will become. This will be determined to a large extent
by the market for gypsum in the construction, agricultural and other industries.
It is only where this process is selected that a wastewater stream will definitely
be produced. Economics have determined this route to be favored in Germany;
conversely, in the U.S. throwaway systems have been favored. Throwaway systems can
be operated on an open or closed loop basis, and only in the former case will
wastewater treatment be required.
The wastewaters are saline, corrosive waters containing heavy metals at levels
unacceptable for environmental disposal. A full knowledge of these waters and
experience of current treatment practices ensures the removal of these *nd other
pollutants.
ACKNOWLEDGEMENTS
My thanks to Johannes Weinig of Philipp Mueller Company, Stuttgart, Germany for his
review of this paper.
8B-38
-------
REFERENCES
1. Weiler, H. & Ellison, W., "Progress in European FGD and SCR Applications",
EPA/EPRI S0Z Control Symposium, New Orleans (1990)
2. Metzler, A.R., Gallardy, P.B., McLaughlin, B.R., Ireland, P.A. and Spengel,
W.J., "Evaluation of Acid Rain Alternatives", Conference Papers, III & IV. 3rd
International Power Generation Industries Conference, Orlando, Florida, 1990,
pp. 225-245.
3. Downs, W., Johnson, D. W., Aldred, R. W., Tonty, L. V., Robards, R. F. and
Runyan, R. A., "Influence of Chlorides on the Performance of Flue Gas
Desulfurization", EPA/EPRI Symposium on Flue Gas Desulfurization, New Orleans,
(1983).
4. Taylor, H.R.G., Heaton, R. and Baty, R., "The Impact of Flue-Gas
Desulphurization on the Water Environment", J.IWEM 3, pp. 227-234 (1989)
5. Wrobel, B., and Seelaus, T.J., "Acid Rain Compliance - Advanced Co-current Wet
FGD Design for the Bailly Station", Conference Papers, III & IV. 3rd
International Power Generation Industries Conference, Orlando, Florida, 1990,
pp. 343-357.
6. Bursik, A. & Dieterle, E., "FGD Wastewater Treatment - State of the Art in the
Federal Republic of Germany" Proceedings, 49th International Water Conference,
Pittsburgh IWC-88-38, pp. 373-379 (1988)
7. Etoh, Y., Takadoi, T. & Itoh, I., "Sludge Reduction in Coal-Fired Power Plant
Flue Gas Desulfurization Wastewater Treatment", Proceedings 41st Purdue
Industrial Waste Conference, pp. 545-553 (1986).
8. Mierzejewski, M.K., Rovel, J.M., and VandeVenter, L.W., "The Use of a High-Rate
Combined Reactor/Clarifier/Thickener for the Treatment of Industrial
Wastewaters", Proceedings, 44th Purdue Industrial Waste Conference, pp. 519-526
(1989).
9. Frick, B.R., "Flue Gas Purification - Shifting Pollution from Air to Water?",
Vom Wasser 65.Band, pp. 145-156 (1985) (Germany).
10. Manzione, M.A., Merrill, D.T., McLearn, M.E., Chow, W., Stine, J.F., Kobayashi,
S., and Martin, W.J., "Removal of Trace Elements from Power Plant Waste Streams
by Iron Adsorption/Coprecipitation", Proceedings 49th International Water
Conference, Pittsburgh IWC-88-36, pp. 361-372 (1988).
8B-39
-------
FLUE
GAS
G-0SB3 LOOP
-o*
RECTCLE
RETURN
CAKE
NO WASTEWATER
Figure 1. Open and closed loop operation of an
FGD scrubber. By operating at a higher chloride
concentration (e.g. 40,000 mg Cl~/1) the closed
loop system avoids a separate wastewater stream
as chloride blowdown.
8B-40
-------
out
CLEAN FUJE CAS
CAS
DEMISTER WATER
AJR
FCO
WASTEWATER
Figure 2. Schematic of water use on a LSFO-WB scrubber
(after Weiler and Ellison [1]).
8B-41
-------
SULFIDE
LIME —i
— FERRIC CH.0R1DE
POLYMER
FGO
WASTE
WATER
i r
uj
©
a-UDaBKYOE^j
©
FILTRATE
CAKE
TREATED
WATER
Figure 3. Flow scheme for physico-chemical
treatment of water
1. Oxidation 2. pH Elevation/Description 3. Heavy Metal Removal
4. Coagulation 5. Clarification/Thickening 6. Sludge Conditioning
7. Sludge Dewatcring
-------
TABLE 1
TYPICAL RANGES OF CONSTITUENTS
IN FGD WASTEWATERS BEFORE AND AFTER TREATMENT
PARAMETER
BEFORE *
AFTER *
CHLORIDE (CI)
10,000 - 40,000
10,000 - 40,000
SULFATE (SO.)
1,500 - 8,000
800 - 2,500
NITRATE (NOj)
300 - 1,400
300 - 1,400
FLUORIDE (F)
30 - 200
<15
CALCIUM (Ca)
4,000 - 20,000
5,000 - 25,000
MAGNESIUM (Mg)
200 - 5,600
200 - 2,000
SODIUM (Na)
75 - 1,200
75 - 1,200
IRON (Fe)
30 - 400
<0.5
ALUMINUM (A1)
50 - 800
<2.0
ARSENIC (As)
0.05 - 3.0
<0.05
BORON (B)
20 - 40
20 - 40
CADMIUM (Cd)
0.04 - 0.5
<0.1
COBALT (Co)
0.05 - 0.4
<0.1
CHROMIUM (Cr) -TOTAL
0.3 - 5.0
<0.1
COPPER (Cu)
0.1 - 0.85
<0.1
MERCURY (Hg)
0.05 - 0.8
<0.05
NICKEL (Ni)
0.2 -6.0
<0.1
LEAD (Pb)
0.1 - 3.0
<0.1
SELENIUM (Se)
0.2 - 1.0
<0.2
ZINC (Zn)
0.4 - 8.0
<0.1
AMMONIUM (NHJ
50
50
COD
100 - 150
100 - 150
PH
4-7
7-9
TSS
300 - 10,000
<15
~-All units are mg/1; pH in Standard Units.
8B-43
-------
TABLE 2
CHEMICAL DOSING IN FGD WASTEWATER TREATMENT
CHEMICAL
DOSE
UNITS
CALCIUM HYDROXIDE (95% PURE)
ORGANOSULFIDE (eg. TMT 15)
FERRIC CHLORIDE (as FeCl,)
POLYMER
SULFURIC ACID
1,500-4,000*
50
30-50
1-1.5
•irk
g/m3
ml product/m3
g/m3
g/m3
* assuming influent pH 5-6, and dependent on water composition.
** dependent on effluent pH required.
TABLE 3
REACTION TANK SIZING
TANK. FUNCTION
HYDRAULIC
RETENTION
TIME (MINS)
OXIDATION
60
pH ELEVATION/
DESATURATION
60
HEAVY METAL REMOVAL
15
COAGULATION
5
POLYMER FLOCCULATION
12
8B-44
-------
The Influence of FGD Variables on FGD Performance
and By-Product Gypsum Properties
8E-45
-------
-------
F. W. Theodore, M. R. Stouffer, D. C. McCoy, and ti. Yoon
Consolidation Coal Company
Research and Development
Library, Pennsylvania 15129
J. E. Smigelski, P. J. Szalach, and J. A. Weist
New York State Electric and Gas Corporation
Binghamton, New York 13902
D. R. Owens
Electric Power Research Institute
Palo Alto, California 94303
ABSTRACT
The paper describes a project designed to better understand the influence of wet FGD
process variables on FGD performance and the production of salable gypsum. Results
obtained from FGD pilot plant testing at the EPRI High Sulfur Test Center are
discussed. The variables studied were dissolved chloride content (15,000-75,000
ppm), limestone grind size, and process configuration (single-tank and double-tank).
Short-term (24-hour) runs were made to determine S02 removal and limestone
utilization as functions of the process pH. These results were used to set the
conditions for the long-term (5-day) gypsum production runs. The FGD process
performance and the preliminary characteristics of the product gypsum are reported.
The gypsum produced during the long-term runs will be tested as a feed material for
wallboard production. New York State Electric and Gas Corporation, Consolidation
Coal Company, EPRI, Empire State Electric Energy Research Corporation, and U.S.
Gypsum are project participants.
Preceding page blank
8B-47
-------
INTRODUCTION
In anticipation of increasing disposal costs for wet flue gas desulfurization (FGD)
waste, New York State Electric and Gas (NYSEG) initiated a project directed at
converting the Kintigh (formerly Somerset) Station to a forced oxidation FGD to
produce a salable gypsum by-product. The focus of the work described in this paper
is the relationship between FGD performance and the resulting gypsum quality. The
overall project objectives are:
1. To expand the existing data base for forced oxidation FGD, with
emphasis on the effects of FGD operating variables on gypsum proper-
ties.
2. To provide data to support the conversion of the inhibited oxidation
wet limestone FGD system at NYSEG's Kintigh Station to a forced
oxidation FGD system to produce a salable gypsum by-product.
3. To evaluate the influence of FGD gypsum properties on wallboard
manufacture.
In order to achieve these objectives, NYSEG assembled a project team with expertise
in all areas of gypsum utilization from FGD operation to wallboard manufacture. In
addition to NYSEG, the co-funding project participants are the Electric Power
Research Institute (EPRI), the Empire State Electric Energy Research Corporation
(ESEERCO), the United States Gypsum Company (USG), and Consolidation Coal Company
(Consol).
The initial phase of this project was a literature study of forced oxidation FGD and
the impact of FGD operating variables on gypsum properties. The literature study
indicated that there are research areas with potential to advance forced oxidation
FGD technology and the production of by-product wallboard-quality gypsum. In
8B-48
-------
addition, the study identified those design variables which could be set based on
literature information. A staged experimental program in conjunction with process
economic studies was recommended.
Based on this preliminary work, a pilot-scale test program was developed to provide
data on forced oxidation FGD performance for the conversion of the Kintigh Station
FGD system to forced oxidation. Such data would allow confirmation of the
feasibility of the conversion and would be used in economic studies to optimize the
process design. This pilot test program focused on the effects of key process
chemistry variables and was conceived as a first-test phase, to be followed by
additional test phase(s) for process optimization, if necessary.
This paper discusses the results obtained from this initial phase of the test
program. The paper will focus on the influence of the key FGD variables on FGD
performance and the resultant gypsum properties, namely, purity and crystal size.
A detailed evaluation of the by-product gypsum quality and its influence on
wall board products is in progress and will be reported at a later date.
TEST PROGRAM
The pilot-scale program was conducted at the EPRI High Sulfur Test Center (HSTC)1
wet FGD pilot plant (4 MWe). The specific test objectives were:
1. To determine the effect of process configuration on FGD performance
and gypsum properties.
2. To determine the effects of process variables, including pH, dissolved
solids (CI" concentration), limestone grind, and L/G ratio on FGD
performance and gypsum properties.
3. To generate data for process optimization for the Kintigh forced
oxidation conversion.
4. To generate gypsum samples for by-product quality evaluation.
Test Variables
The pilot test program was conducted in two test blocks: (1) process configuration
tests, and (2) process variable tests. The process configurations evaluated were
8B-49
-------
a combined oxidizer-reaction tank system (single-tank mode) and a separated oxidizer
and reaction tank system (double-tank mode). The process configuration tests were
conducted with the other process conditions held constant: 35000 ppra CV (average),
89% -325 mesh limestone, and 130 L/G. The process conditions used for the
configuration tests were within a range which represents the projected conditions
for the Kintigh Station conversion. The L/G ratio used in the pilot tests was
higher than projected for Kintigh to reflect differences in absorber mass transfer
capacities between the pilot and Kintigh equipment. The process variable test block
was designed to investigate the effects of key process operating variables (i.e.,
dissolved solids concentration, limestone grind, and L/G ratio) with a fixed process
configuration (double-tank).
The FGO design and operating parameters which were not varied for the tests were
fixed to match those expected in the converted Kintigh FGO system. The inlet S02
concentration was held constant for the entire program at 2000 ppm S02. The
limestone now used at Kintigh was used in the pilot tests. Concentrations of
dissolved chemical species (CV, Mg2*, Na*) were maintained in the same ratio as
currently found at Kintigh. The liquid retention time for both the oxidizer and
reactor tanks were set to approximate their counterparts at Kintigh, namely, 4.4 min
and 3.8 min, respectively. The solids content of the circulating slurry was
controlled at 12%. Finally, a hydroclone and vacuum belt filter were used for
dewatering of the product gypsum.
The oxidation air system was not optimized in this program; rather, the air sparge
rate was set to achieve 99%* oxidation efficiency. Because the pilot plant oxidizer
height was limited, substantially more air was required than expected for commercial
size systems.
Test Procedure
There are several factors that determine gypsum purity. Two components that are
related to both gypsum purity and FGD performance are the calcium sulfite
concentration (percent oxidation) and calcium carbonate concentration (limestone
utilization). As noted above, the degree of oxidation was maintained constant at
a high level, typically greater than 99%, during the program. Limestone utilization
was used as a performance parameter because it is a governing variable in the by-
product gypsum purity and because it controls solid-phase alkalinity which controls
S02 removal. Therefore, limestone utilization and S02 removal together define the
performance of a forced oxidation FGD system. Testing proceeded on this basis to
8B-50
-------
evaluate the effects of key process variables on utilization and S02 removal. A
series of short-term, pilot plant parametric tests were conducted. Short-term tests
were 24 hr long, which was based on the time required for the system limestone
loading to reach equilibrium. These tests consisted of measuring the limestone
utilization and S02 removal at different absorber feed slurry pH levels with
limestone grind, chloride level, and L/G held constant. Figure 1 presents typical
short-term test data, the utilization and S02 removal as functions of feed slurry
pH. The data from the test series were analyzed to determine optimum operating
conditions for long-term tests. The purpose of the long-term pilot plant tests was
to generate gypsum samples, representative of solid-phase equilibrium and stable
steady-state operation, for use in by-product quality evaluation studies. Long-term
runs were greater than five days in length. This allowed for about three solid-
phase residence time changeouts.
Measured S02 removals in the HSTC pilot plant absorber were expected to be less than
for the Kintigh absorbers at the same L/G ratio and process conditions, primarily
because of increased wall effects in the smaller unit. In addition to wall effects,
other factors affecting mass transfer could not be duplicated in the pilot plant
(for instance, absorber gas distribution, absorber slurry distribution, slurry
droplet size, etc.). Hence, the pilot tests were conducted at higher L/G ratios
(most at 130 gal/1000 acf) than used at Kintigh (currently, 79 gal/1000 acf). In
order to better define the correlation between the HSTC pilot plant performance and
Kintigh Station absorber performance, some pilot runs were conducted under inhibited
oxidation conditions to simulate Kintigh operation.
RESULTS AND DISCUSSION
As stated earlier, the objectives of the test program were multipurpose and a
complete analysis of the test results is beyond the scope of this paper. In
addition to summarizing results, the following discussion will focus on how the
results influence the conversion of Kintigh to a forced oxidation unit producing
salable gypsum. Preliminary data are presented which characterize the gypsum
produced during the five long-term runs. The gypsum from these five runs will be
analyzed further for acceptability as a wallboard feed and the results will be
reported at a later date.
In dealing with the results of this study, we will focus on the limestone
utilization, since this variable is a key in establishing the by-product gypsum
8B-51
-------
purity. Low utilization results in gypsum containing high levels of unreacted
CaCOj. CaCOj is an inert material that adversely affects the properties and
economics of wallboard production. The chlorides and other soluble impurities can
be washed out, but carbonates will remain in the gypsum. If the goal is to sell
gypsum, then a high (90-95+%) limestone utilization must be achieved, and the
desired S02 removal must be achieved by adjusting other control variables, (L/G,
etc.). The F6D run data are summarized in Table I.
FGD Performance
Process Configuration. The two process configuration modes discussed in this paper
are the single-tank and double-tank. The two major differences for the double-tank
configuration tests, as compared with single-tank tests were: (1) longer solids
residence time (about double) due to additional tank volume, and (2) separation of
the oxidation tank upstream from the limestone addition point. The process
configuration test block was not designed to distinguish between the effects of
these two factors on the performance.
Figure 2 shows the FGD performance (S02 removal versus limestone utilization) for
the two process configurations. As can be seen, there is little if any difference
in performance between the two configurations at limestone utilizations below 90%.
As the limestone utilization increases above 90%, the double-tank configuration
achieves greater S02 removal. At a very high utilization (98%), the S02 removal is
5-8 % absolute greater for the double-tank case. The graph suggests that if one is
attempting to obtain high limestone utilization (high purity gypsum) greater than
90%, then the double-tank configuration case becomes attractive.
The probable reason for better performance of the double-tank system is that at a
constant utilization, the double-tank system was operable at a higher pH. The
increase in operating pH gave a higher liquid-phase alkalinity, which accounted for
the higher S02 removals. The increased residence time and separation of the
oxidation and limestone addition steps from each other are likely responsible for
the increase in operating pH level for the double-tank configuration. As noted
above, system equipment limitations during the test program prevented isolation of
the relative importance of these two factors.
In the case of the Kintigh Station, the existing FGD plant was designed for a future
retrofit and has the tankage in place for a double-tank system. Because the Kintigh
Station already had equipment for a double-tank configuration and because the pilot
8B-52
-------
double-tank results were equal to if not better than the results for the single-tank
system, the remainder of the program was conducted in a double-tank mode.
Chloride Concentration. The base case chloride concentration used in this study was
35,000 ppm. This level of chloride was chosen since it is in the middle of the
chloride range being considered for the Kintigh Station. As stated earlier and as
shown in Table I, the other dissolved soids species (Mg2* and Na") were increased
and decreased in proportion to the CI* concentration in order to maintain the same
ratio as currently found at Kintigh. The chloride concentration is determined by
the plant water balance. The design chloride level for Kintigh will depend on the
cost of chloride control versus the costs for improving FGD performance at higher
chloride levels.
Figure 3 shows the FGD performance achieved during the pilot tests at four different
chloride levels (15,000, 35,000, 55,000 and 75,000 ppm). The data for the lower
three chloride levels include data from long-term runs. The 75,000 ppm chloride
data were taken during a series of short-term runs. The individual data points are
plotted over regression curves generated from the entire data set. As shown in
Table I, the chloride concentration in the three high chloride level runs deviated
from the 75,000 ppm CI" set point. This fact explains the deviation of the
individual points from the regression curve (Figure 3).
The detrimental effect of higher chloride levels on FGD performance is obvious from
the results shown in Figure 3. For any given limestone utilization, the absorber
S02 removal decreases with dissolved chloride.
Furthermore, at higher limestone utilization, the S02 removal is sensitive to small
changes in utilization. This is not an unexpected result because as the limestone
utilization levels are increased, the solid-phase alkalinity decreases, which
reduces the availability of reagents in the absorber if slurry circulation (L/G)
remains constant. The result emphasizes that at the high utilization required to
produce a high purity by-product gypsum, the resulting S02 removal is more sensitive
to system chemistry than at lower utilization.
Limestone Grind Size. The limestone grinds used in this study are related to that
used at the Kintigh Station in the following ways. The coarsest grind, 67% -325
mesh, is equivalent to the current Kintigh grind. The middle grind, 89% -325 mesh,
corresponds to the limit achievable at Kintigh without an additional ball mill. The
8B-53
-------
finest grind (94% -325 mesh}, would require the addition of another ball mill at
Kintigh. For the three limestone grind sizes tested in the pilot plant, use of
finer grinds moderately improved desulfurization performance. The exact improvement
was influenced by the levels of other variables. As an example, the influence of
grind size appeared to be stronger at a low chloride level (15,000 ppm) than at a
higher chloride level (55,000 ppm).
Figure 4 shows S02 removal versus utilization curves for the three different grind
sizes at a 55,000 ppm CI' concentration. At constant limestone utilization, SOj
removals increased by 2 to 4% absolute from the coarsest to the finest grind. For
example, at 95% utilization, S02 removals were evenly spaced (77%, 79%, and 81%,
respectively). At lower utilizations, however, the middle grind size (89% -325
mesh) performance approached that of the coarse grind; whereas the finer grind gave
3% absolute higher S02 removal than the other two. It was concluded that for the
projected Kintigh Station operating conditions (85% S02 removal and 95% limestone
utilization), the base case grind (89% -325 mesh) gave satisfactory performance and
would eliminate the need for a new mill to meet Kintigh's proposed conditions.
Liouid-to-Gas Ratio fL/G). At the completion of the last long-term run (51), a
series of short-term runs was made to investigate the effect of L/G on S02 removal.
The conditions for these runs, coarse limestone grind, 15,000 ppm CI", and a pH of
approximately 5.9, were comparable to those of run 51. The limestone utilization
was not strongly affected by L/G over the narrow range tested. The utilization
varied randomly by less than 4% absolute for an L/G range of 130-162 gal/1000 acf.
Below is a summary of the data.
L/G, gal/1000 acf
130
142
162
205
65
100 1
PH
5.91
5.91
5.91
5.90
6.11
5.92 |
SO, Removal, %
80.7
83.3
85.3
92.2
57.1
67.6 |
NTU = In (SO, in/SO, out)
1.64
1.79
1.92
2.56
0.85
1-13 1
Limestone Utilization, %
95.8
94.4
98.3
--
--
I
The S02 removal data are plotted as number of transfer units (NTU) versus L/G in
Figure 5. As shown, the S02 removal expressed in NTUs is directly proportional to
L/G, which corresponds to previous natural oxidation FGD data.2 Based on this
series of runs, it is assumed that for other conditions used in the pilot plant, the
S02 removal in NTUs is a linear function of L/G and that utilization is independent
8B-54
-------
of L/G. This is a very useful relationship since it allows calculation of required
L/G for any S02 removal, assuming performance is known at one L/G level.
Gypsum Crystals
Samples of gypsum from both long- and short-term tests were analyzed for crystal
size using the Malvern Laser Diffraction Analysis method. The results shown in
Table II are the averages of duplicate analyses of each sample. The absolute
measured crystal size is dependent upon the specific method and procedure used for
the size determination. However, by using only one method for size determination,
the relative influence of process variables on crystal size can be evaluated.
Table II shows the crystal size (mass mean diameter) and the size distribution
%<\Qfm and %>4Q«ni. The single largest variable to influence the crystal size is the
process configuration (double-tank versus single-tank). By comparing the crystal
size for runs 6 and 17, it can be seen that the mass mean diameter increased from
24/an with the single tank to 31.fyon with the double tank. In addition, the percent
-lQum decreased from 12.9% to 5.1%. The larger crystal size obtained with the
double tank configuration may solely be due to increased system residence time. As
mentioned earlier, the ability to separate the influence of retention time from the
two-tank configuration was not possible in this program.
The process variables had little if any influence on the resultant crystal size
distribution, especially when one considers only the results from the long-term
runs. When the crystal size data from the short-term run at high chloride content
(run 31) are included in the analysis, it appears that the crystal size increases
slightly with increasing chloride content. This would confirm earlier work3 by
EPRI/Radian which showed that the gypsum crystal growth rate increases with TDS
levels from 30,000-240,000 mg/L. At lower TDS levels, their results indicated
little dependence of crystal growth rate on TDS level.
The conclusion is that with the exception of the possible minor influence of
chloride concentration on the crystal size, the process chemistry variables studied
had no effect on crystal growth. This is an important observation since crystal
size and shape can be expected to be stable during normal variations in FGD
operation.
The typical shape of the gypsum crystal produced in the program is illustrated by
Figure 6, a micrograph from run 37 material. The crystal is block-shaped as opposed
8B-55
-------
to needle-like or plate-like. A complete analysis of the by-product gypsum will be
reported later. Material from all the long-term runs (6, 17, 29, 37 and 51) is
available for evaluation.
A preliminary inspection of the product crystals produced during this test program
showed that the crystal size and shape were typical of other by-product gypsums.
In addition, a product sample was subjected to a series of batch tests to determine
its suitability for producing wallboard. The results of these tests showed that a
high quality board was produced from the test material. The test details will be
reported at a later date. Based on these facts, the resultant crystal size and
shape generated during this program should be suitable for wallboard production if
the other product specifications such as moisture, chloride, purity, etc., are met.
Scaleuo
In order for the data obtained during the testing to be more useful in projecting
the operation of the Kintigh F6D converted to forced oxidation, the scaleup factor
for the pilot FGD was determined. This was accomplished by comparing the
performance of the pilot and Kintigh FGD systems under conditions where process
chemistry was held constant.
The final testing done under this program was a series of short-term runs in the
inhibited oxidation mode. Absorber slurry circulation (L/G) was the only variable
changed during this test series. The other conditions for these runs are shown in
Table III. Figure 7 graphically shows S02 removal as a function of L/G. As in the
forced oxidation case, S02 removal plotted as NTUs is a linear function of L/G.
The S02 removal and L/G data from Kintigh are shown in Figure 7. The other
conditions for the Kintigh tests are shown in Table III. Although there was some
difference in the conditions between the pilot plant and Kintigh Station, the
simulation was close enough to allow an estimation of the HSTC pilot plant scaleup
factor. Based on the results of these tests, the L/G required at Kintigh is
expected to be about 55% of the L/G required at the HSTC.
Summary of Results and Conclusions
The forced oxidation test program conducted at the HSTC was successfully completed.
The data collected show the influence of key process variables on system perfor-
mance. The results of the test work will be used to guide design efforts in
8B-56
-------
converting the inhibited oxidation FGD system at Kintigh to a forced oxidation
system. Major program conclusions are presented below.
1. The test program results were positive, indicating that conversion of
Kintigh to a forced oxidation system to produce salable gypsum is
technically feasible. Based on pilot data and on an approximate
scaleup correlation between the pilot and Kintigh absorbers, 85% S02
removal at 95% limestone utilization should be achievable at 55,000
ppm CI", 67% -325 mesh limestone grind, and an L/G ratio of about 95
gal/1000 acf, an achievable value at Kintigh.
2. The pilot tests generated gypsum samples at varying process conditions
for characterization and by-product quality evaluation.
--Gypsum purity has been determined for all samples. High purity was
achieved by maintaining high limestone utilization.
--Laser diffraction particle size analyses were conducted by Consol to
determine the effect of process configuration and process variables on
gypsum crystal size.
3. FGD performance with the double-tank configuration was somewhat better
than with the single-tank configuration in the pilot tests.
--S02 removals were slightly higher with the double-tank configura-
tion, especially at higher limestone utilizations. For example, S02
removal was about 5% (absolute) higher at 98% utilization, but
removals were essentially the same below about 92% utilization.
--The gypsum crystals produced in double-tank testing were larger than
those from single-tank testing (31/an versus 24/an mass mean diameter).
4. FGD performance was significantly impaired by increasing levels of
chloride in the process liquor. Chloride was varied over a range of
15,000 ppm to 75,000 ppm, with the other species (Mg2*, Na*) held in
the same ratio to CI* as for current operating conditions at Kintigh.
8B-57
-------
--With increasing dissolved solids, S02 removals were lower at the
same limestone utilization. For example, with the 89% -325 mesh
limestone grind, S02 removals at 95% utilization were about 86%, 82%,
and 79% at chloride contents of 15,000, 35,000, and 55,000 ppm,
respectively. The decline in S02 removal was roughly linear with
increasing chloride over the test data range.
5. FGD performance improved moderately with finer limestone size for
three grinds tested (67%, 89%, and 94% -325 mesh).
--S02 removal increased moderately at constant limestone utilization
as limestone grind size was reduced. For instance, at 55,000 ppm CI"
and 95% utilization, the S02 removal increased by about 4% (absolute)
from the coarsest to the finest grind tested.
--The effect of limestone grind on performance appeared to be somewhat
stronger at lower chloride concentration (15,000 ppm) than at higher
CI concentration (55,000 ppm).
6. The effect of L/G ratio on FGD performance was determined under forced
oxidation conditions.
--S02 removal increased significantly with increased L/G ratio.
Expressed as number of transfer units (NTU), the S02 removal was
directly proportional to L/G ratio over a range of 65 to 205 gal/1000
acf. The correlation of NTU with L/G was strong, despite the use of
different spray header levels and nozzle flow rates to achieve the L/G
variation in the pilot absorber.
--Limestone utilization was not affected significantly by varying L/G
over the range in which it was measured, 130 to 160 gal/1000 acf.
7. Pilot testing under inhibited oxidation conditions confirmed that the
mass transfer capacity of the HSTC pilot plant absorber is consider-
ably less than that of the Kintigh absorbers operated at the same L/G
ratio. The results suggest that an L/G ratio of roughly 140 gal/1000
acf in the pilot plant provides about the same mass transfer capacity
8B-58
-------
as the Kintigh absorbers at current operating conditions (L/G - 79
gal/1000 acf).
ACKNOWLEDGEMENTS
The authors would like to recognize the operating and maintenance staffs at the EPRI
HSTC for their contribution to the success of the test program. In addition, the
contribution made by Greg Stevens, Radian, and Michael Delallo, Gilbert/Common-
wealth, was key to the successful completion of the program.
The authors wish to acknowledge the financial support of the HSTC cosponsors: EPRI,
NYSEG, ESEERCO, U.S. Department of Energy, and Consol.
REFERENCES
1. R. E. Moser, J. M. Burke, and S. M. Gray. "Results of Wet FGD Testing at EPRI's
High Sulfur Test Center." In Proceedings of the EPA/EPRI First Combined FGD and
Dry S0: Control Symposium. EPRI CS-6307, RP 982-41, St. Louis, Mo., October 1988.
2. S. M. Gray and J. M. Burke. EPRI High Sulfur Test Center Report: Wet FGD
Baseline Tests. Electric Power Research Institute (RP 1031-9), Palo Alto, Calif.,
1990.
3. F. B. Meserole, T. W. Trofe, and D. A. Stewart. "Influence of High Dissolved
Solids on Precipitation Kinetics and Solid Particle Size," Presented at the
EPA/EPRI Symposium on Flue Gas Desulfurization. New Orleans, La., November 1983.
The work reported in this paper is the result of research carried out (in part) at
EPRI's High Sulfur Test Center (HSTC) located near Barker, NY. We wish to
acknowledge the support of the HSTC cosponsors: New York State Electric &
Gas, Empire State Electric Energy Research Corp., Electric Power Development
Co., Ltd., and the US Department of Energy. The cosponsors provide valuable
technical review of the work in progress as well as funding test center
operations.
8B-59
-------
i --1
COMMON CONDITIONS
CHlOfBDE CONCENTRATION : 38.000 ppm
LIMESTONE GAIMO : ®% US MESM
L/C RATIO : 130 Q^/1000 ad
O $02 REMOVAL
o UTKiZATIOM
(ABSORBER INLET ANALYSES)
~I—
48
~I—
SO
—I-
5.2
—r~
5.4
—r~
s.e
—r~
S.B
ABSORBER INLET pH
rigor* 1. SO, KOOfAL AID LOBS1M mUZATJOa AS A V0BCTI0B OF ABMUO IBIS B 1
8B-60
-------
90
•6
•0
o
75
COMMON CONOmONS
CONFIGURATION
UMCSTONC GMMO
L/C
O 15.000
O 36.000
O 55.000
A 75.000
70
99% -325MCSM
130
7S
M
74
90
at
UTHiZATtOM. % (ASSOMSt* M. AMAL.I
ri|«r« j. yTF1"1 nxvoiBABCz coKvn lot pob
-------
COMMON CQfomows
CONFIGURATION : OOU«L£ TAA
O&OMDC CONCENTRATION : 15.000 ppm
UMESTONE CRMO : (7% -32S Ml
pM ; S3
8B-62
-------
£9-a8
uru uniA <¦? muiu bzuis mi oun m cl azMn •«*¦¦¦* man toi 40 aomraoo 'i
p* ooouv** ouvw on
00Z 091 021 08 0*
I ¦ 1 ' I I I I I 1__
101U31SH O
H9UM» O
m nsvi 335
-------
Table I
DATA SUMMARY1
Dissolved Solids
Run2
Eif
Grind
cr
Mg2*
Na*
Ca2*
S02 Removal
UT*
* -325 Mesn
ppm
ppm
ppm
ppm
percent
percent
1
4.80
88.8
43806
9062
2450
214.2
67.3
98.8
2
5.20
88.8
41015
8667
2026
211.7
69.8
98.0
3
5.40
88.8
37577
7911
1763
190.9
73.8
96.6
4
5.60
89.0
—
—
80.0
96.1
4A
5.60
89.0
36488
8044
1841
181.3
82£
96.1
5
5.90
89.0
34140
7019
1978
170.6
90.4
73.5
6
5.70
89.0
—
7344
1983
171.2
84.0
94.1
6A
5.70
89.0
38326
7521
2168
187.6
84.2
92.0
6B
5.70
89.0
35671
7566
2096
162-8
84.6
93.4
17F
5.81
86.8
34625
7027
2137
208.8
794
97.0
17G
5.80
88.4
34424
7199
1895
2060
80.8
95.4
17H
590
88.4
33962
7175
2018
197.1
84.9
92.3
171
580
91.0
34566
6889
2188
184.3
825
96.4
100
600
90.7
35347
7097
2019
179.9
87.4
86.1
8RR
5.42
90.7
34192
6792
1960
171.5
75.1
99.3
24A
5.59
89.7
3740
958
92.3
77.5
__
24B
5.60
89.4
14809
3432
803
86.6
76.4
99.5
25
5.91
89.4
14337
3171
736
84.0
83.6
97.9
26
6.11
89.4
14624
3211
751
93.0
86.0
96.0
27A
6.16
89.4
14752
3343
737
90.6
88.5
81.9
278
6.16
89.4
14840
3514
763
90.5
896
78.6
28
6.06
89.7
15362
3245
898
98.6
86.4
93.7
29A
6.11
89.7
14541
3221
886
93.9
86.8
91.2
29C
6.11
893
15348
3038
86.0
93.1
29D
6.11
893
15009
3054
817
90.0
85.8
92.0
30
5.41
90.4
79412
14644
4323
424.8
69.8
96.4
31
5.60
90.4
71208
13156
4023
382.8
77.8
91.7
32
5.80
90.4
64231
12671
3730
370.2
8Z2
81.8
33
5.41
89.9
53908
9332
2837
296.7
69.1
99.1
34
5.61
89.9
55007
9904
3123
279.8
74.0
97.4
35
5.81
89.4
53362
10241
3016
263.6
81.0
90.7
36
5.90
89.4
53096
10100
3014
262.4
84.3
80.4
37A
5.71
884
10449
_
2560
78.9
95.1
37B
5.71
882
52217
10423
3074
7772
78.9
95.8
37C
5.71
882
_
280.5
79.0
94.6
37D
5.70
88.2
54385
10899
3138
283.8
79.7
95.6
38
5.41
92.2
51585
8985
2979
246.4
72.0
99.3
39
5.70
93.7
53345
11421
3122
268.2
80.2
95.8
40
5.89
93.7
46126
9978
2724
274.4
85.3
88.2
41
5.95
93.7
48219
9962
2916
260.5
86.5
73.8
42
5.80
94.4
45951
—
—
—
85.2
84.6
43A
5.41
66.3
_
10487
73.7
438
5.41
66.3
—
10173
2868
74.7
43C
5.41
66.3
56557
11033
3024
253.8
74.5
95.4
44
5.61
66.3
45878
11062
2939
276.8
78.5
94.4
45
5.80
67.0
54970
10674
2809
280.9
85.3
74.8
46
5.70
67.0
57914
11340
3046
288.7
82.0
86.9
47
5.42
66.5
11315
2733
775
75.6
75.0
99.1
48
5.72
66.5
15467
2911
817
103.7
75.6
98.1
49
5.91
66.2
14531
3047
887
109.4
80.5
94.4
50
6.00
66.2
13829
2782
752
107.4
88.6
75.7
51A
5.91
66.2
2540
751
—
82.4
—
51B
5.91
66.2
—
3165
—
91.5
80.8
95.1
51C
5.91
67.3
14609
3181
711
94.2
79.6
—
510
531
67.3
13646
2759
746
84.4
80S
94.4
51E
5.91
67.3
14318
3079
711
91.6
80.7
95.8
1) Slurry solids, 12%. UG = 130, 2000 ppm inlet S02
2) Runs 1-6B were in the single-tank configuration
3) Absorber inlet
4) Limestone Utilization
8B-64
-------
Table II
GYPSUM PARTICLE SIZE DISTRIBUTION1
Distribution
Run2
Liquor a
Grind
Etf.
UT*
Mass Mean
Run Tvoe
ppm
% -325 Mesh
percent
Diameter, am
% <10sm
6
long term
35,000
89
5.7
93
24.0
12.9%
29
long term
15,000
89
6.1
92
32.6
6.1%
51
long term
15,000
67
5.9
95
34.2
5.0%
17
long term
35.000
89
5.8
96
31.4
5.1%
37
long term
55,000
89
5.7
96
34.0
7.2%
42
short term
55,000
94
5.8
85
34.4
6.4%
31
shon term
75,000
89
5.6
92
35.6
7.9%
% >40«m
11.2%
25.1%
28.2%
22.5%
30.8%
32.0%
37.8%
1) l/G ratio: 130.
2) Run 6 was in trie single-tank configuration.
3) Absorber Inlet.
4) Limestone utilization.
Table III
COMMON OPERATING CONDmONS FOR INHIBITED OXIDATION RUNS
Operating Conditions
Configuration
Dissolved Solids
Concentrations, ppm
cr
Mg2*
Na*
Ca2*
Limestone Grind
% <325 mesh
Inlet S02 Content, ppm
Reaction Tank. pH
Slurry Solids, wt %
Thiosulfate (SjOJ
Concentration, ppm
Oxidation, %
Solids Residence Time, Hr
Limestone Utilization, %
CaC03 Loading, g/L
HSTC Pilot Tests
2 Tanks
11300
2750
805
3000
67%
1750
58
18.2%
498
17.8%
ca 55
93%
10.3
Kirnigh FGD
2 Tanks
12000-15000
ca 2900
ca 870
70%
ca 1500
5.6-5.65
12%
80
6-10%
ca 21
87%
11.2
8B-65
-------
8B-66
-------
Quality of FGD Gypsum
Preceding page blank 8B"67
-------
8B-68
-------
F.H. van der Brugghen
N.V. KEMA
P.O. Box 9035
6800 ET ARNHEM
The Netherlands
ABSTRACT
In the Netherlands 8 large coal-fired power stations are or will be equipped with
FGD installations based on the lime(stone) gypsum technology. The gypsum produced by
those installations has to be marketed. To achieve this, the quality requirements
specified by the gypsum industry have to be met. The gypsum quality attained by the
different power stations will be assessed in relation to those standards i.e.
chemical composition, crystal size and colour. The influence of the design and
operation of the different FGD plants will be described. Furthermore, some remarks
will be made about the trace element content, especially radioactive elements and
mercury.
Preceding page blank
8B-69
-------
INTRODUCTION
Power generation in the Netherlands is currently mainly based on natural gas and
coal and it is most probable that this will remain so during the next decade. The
coal-based generating capacity will even increase.
Abatement of air pollution caused by sulphur dioxide and nitrogen oxides is an
important issue of the environmental policy in the Netherlands. A voluntary
agreement (covenant) has been signed between the authorities and the Electricity
Generating Board to limit the sulphur dioxide emissions resulting from power
generation to 18,000 t/a by the year 2000. This means that by that time the full
coal-fired generating capacity has to be equipped with FGD systems.
These FGD systems will be based on the lime(stone)-gypsum technology, which will
mean that the output of gypsum will increase compared to the current situation.
An important factor in this respect is the quality of the gypsum. The gypsum
industry has formulated requirements to ensure the quality of their products and to
enable proper operation of their existing facilities.
After a short introduction to power generation in the Netherlands and its air
pollution aspects, an overview will be given of the experience gained in Dutch FGD
¦yRtems concerning the influence on gypsum quality of plant lay-out, operating
parameters and the raw materials used. Some remarks will be made about trace
elements, especially about radioactive elements and mercury.
ELECTRICITY GENERATION IN THE NETHERLANDS
Large-scale electricity generation in the Netherlands is in the hands of four
regional, publicly owned power companies. In 1990 their total installed generating
capacity amounted to 15,142 MW, the net output was 57.6 TWh. At present coal and
natural gas are by far the most important primary energy sources. The share of
nuclear power is only modest, while the use of fuel-oil is almost negligible. This
is shown in Table 1. This table also show9 the division of the generating capacity,
among the different fuelling possibilities.
8B-70
V
-------
To maintain a reliable, •conomic and socially acceptable power ganaration ayatam,
old units have to be decommissioned and replaced by new ones based on state-of-the-
art technology. Furthermore, the still growing demand has to be met. In this
respect, plans for the following 10 year period are published biannually in tha
so-called "Electricity Plan". The Electricity Generating Board "Sep", the
coordinating body for large-scale electricity generation and transportation, ia
responsible for this plan, which has to be approved by the Minister of Economic
Affairs.
According to the "Electricity Plan 1991-2000", the following new units will be built
before the year 2000:
# 3 x 600 MW pulverized coal/gas firing (1993, 1994, 1997)
# 1 x 250 MW demonstration IGCC (integrated coal gasification combined
cycle) (1993)
# 1 x 600 MW IGCC (1999)
# 1,700 MW natural gas fired combined cycle system (1996)
# 5 x 250 MW natural gas fired combined heat and power (1996-1999).
A number of older units will be decommissioned.
Table 2 shows a comparison between the maximum possible generating capacities for
each fuel under the current situation, and as it will be in the year 2000.
The coal generating capacity will increase at the cost of fuel-oil, whereas the gas
capacity remains stable. Coal and gas will remain the dominant fuels.
AIR POLLUTION ASPECTS OF POWER GENERATION IN THE NETHERLANDS
Abatement of air pollution caused by sulphur dioxide and nitrogen oxideB is an
important issue of the environmental policy in the Netherlands. It will be obvious
that the electric power industry has to take its share of the often costly measures:
# since 1981 the use of natural gas sharply increased at the cost of
fuel-oil. This was made possible by a change in government policy con-
cerning the use of the domestic natural gas resources
# the existing large coal-fired power plants with a lifetime extending
beyond the year 2000 were retrofitted with FGD
8B-71
-------
• the smaller units with a restricted lifetime have to fire coal with a
sulphur content below 0.8%
• the new coal-fired units will be equipped with FCD.
The results achieved by those measures are shown in Figure 1. The switch from
fuel-oil to natural gas had far reaching positive consequences during the early
eighties. After 1985 the influence of retrofitting FGD became clear. Despite the
start-up of 3 units (1,500 MW) that were converted from oil to coal firing, no
increase in total annual SO; emissions occurred.
The current level of SO. emissions is 40,000 Mg/a, and has to be fully attributed to
the coal-fired units. To achieve the reduction of acidification formulated in the
National Environmental Plans of 1989 and 1990, a further decrease of this emission
level is necessary.
One possible tool is the tightening of the emission limits stipulated in the
ordinance of May 1987 in the framework of the Air Pollution Act. A draft revision
has already been published with a planned date of enforcement of 1 January 1992. The
current emission limit for new coal-fired power plants with a thermal capacity over
300 MW is 400 mg/m,3, as well as 85% desulphurization. The value in the draft
ordinance is 200 mg/rn,3.
However, in the meantime, a voluntary agreement, between the Federal Government, the
Provinces (as licensing authorities) and the Electricity Generating Board ("Sep")
has been signed in June 1990. As a result of this covenant, by the year 2000
emission levels will be achieved that are considerably lower than might be possible
with revisions of the legal emission limits. On the other hand "Sep" can select the
most cost-effective way of realizing the goals of this covenant.
The total yearly SO. emissions resulting from large-scale electricity generation
will be limited to 18,000 Mg/a by the year 2000, with an intermediate target of
30,000 Mg/a by the year 1994. An additional 4,000 Mg/a will be allowed for possible
upsets of FGD installations, in case the flue gas has to be bypassed directly to the
stack. By law this bypassing is limited to 240 hours per year and may not surpass
72 hours per event.
The emission ceilings for NO, are 55,000 Mg/a by the yeai T.994 and 30,000 Mg/a by
the year 2000, compared to the current level of 75,000 Mg/a.
FLOE GAS DESULPHURIZATION IN THE NETHERLANDS
Four factors will influence the SO> emissions of the electric power industry in
Netherlands during the next decade:
8B-72
-------
• the share of coal in electricity production
9 the presence of flue gas desulphurization installations and their re-
liability
9 the sulphur content of the coal
t.he reliability and sulphur removal efficiency of IGCC systems.
The ^hcire of coal in electricity production is difficult to predict and will be
dependent on developments in the world energy market.
In 1975 the last colliery in the Netherlands was closed. At present all coal has to
be imported, mainly from the USA, Columbia, Australia and Poland. As a rule those
coals are relatively low in sulphur. The mean sulphur content of the coals fired in
1989 was 0.7%.
IGCC is a new technology and to test its reliability a demonstration unit is under
construction. The expectations concerning sulphur removal are very high. The
expected value for the 250 MW demonstration plant is 98.5*.
Flue gas desulphurization has been introduced in the Netherlands in 1985. Table 3
gives a survey of the current situation. At present 2,686 MW or 71* of the total
coal firing capacity is provided with FGD. By the year 2000 all coal-fired units,
4,486 MW in total, will be desulphurized.
All systems are based on lime(stone) scrubbing with gypsum as by-product. This
technology was selected because of its high reliability, relative simplicity,
acceptable costs and the applicability of the by-product.
Two basic FGD technologies are applied: the German Bischoff system, with Tebodin and
Cotnprimo as the Dutch contractors; and the American GEESI system, with Hoogovens
Technical Services (ESTS) as the contractor. The Bischoff installations currently
use hydrated lime as absorbent and seawater for suspension make-up; the ESTS
installations use limestone and riverwater. Major components of the FGD instal-
lations are the spray towers. The two technologies use different arrangements for
the suspension spray nozzles, and different arrangements for the injection of
oxidation air. It i3 noteworthy that Amer-8 has the largest 3ingle spray tower in
the world.
A continuous purge flow from the main scrubbing loop enables the removal of gypsum.
The gypsum removal process comprises two steps: concentrating and dewatering.
The concentration of gypsum in the suspension is increased either by the use of
hydrocyclones or a settling tank. The concentrated suspension is dewatered.
8B-73
-------
Three methods for this final dewatering are used: batch centrifuges, continuously
operating decanters and vacuum belt filters.
Finally the product has to be stored under cover. To that end, use is made of silos
with loading and unloading facilities, warehouses with fixed unloading facilities
and warehouses with Bhovels for unloading.
The marketing of FGD gypsum is in the hands of "Vliegasunie", a f Istj. . ounded by the
four large power companies. To date, it has been possible to sell the gypsum, mainly
to the gypsum industry. With increasing coal-based generating capacity, a growth of
gypsum production is expected, as is shown in Figure 2. To maintain the complete
sale of gypsum it is necessary to guarantee a constant and good quality.
REQUIRED GYPSUM QUALITY
The major applications for FGD gypsum are: as a raw material for the production of
building materials such as blocks, wallboard and plaster; and as a retarder in
cement. The quality of the FGD gypsum has to meet the requirements of the industry.
A major consideration of the cement industry is the handling properties. The
specification of the gypsum industry has to guarantee the manufacture of products of
good and constant quality as well as proper operation of the fabrication facilities,
and comprises chemical composition, crystal size and shape and colour. The required
gypsum quality is specified in the contracts between power plant and consumer.
Table 4 summarizes several specifications used by the gypsum industry. '
A and B axe specifications of the gypsum industry. The VDEW/VGB specification has
been drawn up in cooperation between the electric power industry and the gypsum
industry in the FRG. There are only minor differences between those specifications.
The moisture content of the gypsum is codetermining for the energy use during
calcining. Furthermore, a high moisture content is often accompanied by a high
concentration of soluble compounds such as chloride, sodium and potassium, which is
due to insufficient removal of the liquid of the scrubbing suspension.
The required CaS04.2H:0 content is high compared to natural gypsum (normally about
80%) and is influenced by the inert content of the absorbent, the fly ash content of
the flue gas and the amount of unreacted limestone. Low concentrations of those
inert materials are desirable to decrease the chance of erosion in the manufacturing
equipment.
The content of CaSO^.O.SH^O is limited because of the fear of smell problems when
acid reacting additives are used during manufacture of gypsum product. Furthermore,
sulphite seems to have a negative influence on the hardening process. The CaSO,
content is directly coupled to the effectiveness of the oxidation in the hold tank, j
8B-74
-------
The concantration of aoluble halides and alkalis la closely related to the effec-
tiveness of Che dewatering. The limitations are prescribed because of a reduced
fixing of the gypsum to the paper during wallboard manufacture and because of the
fe&r of efflorescence during use of the ready product.
A much discussed subject is colour or "whiteness", which is mostly referred to BaS04
aa standard. The colour is negatively influenced by the moisture content of the
gypsum, so it has to be stipulated that a sample be dried at 40 °C. The colour is
influenced by contaminants like iron and manganese compounds.
The presence of ammonium is of special interest for units with an SCR installation.
The extremely low value is based on fear of smell problems and the possibility of
the growth of micro-organisms in ready products.
The term "toxic compounds" in the VGB/VDEW specification has not been furthar
defined, but can refer to heavy metals or certain organic compounds like "dioxinea"
and PAH's.
INFLUENCE OF OPERATING PARAMETERS OF THE FGD PLA'JT ON GYPSUM QUALITY
The crystal shape and especially crystal size are very important factors during the
dewatering stage. The influence which the operating parameters of the scrubber loop
have on the particle size and the shape of the gypsum crystals is still rather
erratic. Parameters involved are concentration of solids and dissolved compounds in
the scrubber liquid and the mechanical damage caused by the circulating pumps.
There is a general tendency that a reduction of the solids content in the scrubber
liquid results in an increase in particle size. The influence of soluble compounds
is still not clear. High sodium chloride concentrations in the seawater systems
initially seemed to promote better crystal growth, but doubts remain. Influence of
chloride concentrations between 4 and 30 g/1 could not be established. The use of
low RPM pumps largely prevents mechanical damage to the crystals.
The forced oxidation mode of operation is very satisfactory: complete oxidation ia
achieved without problems. In some installations it is even questionable whether
injection of oxidation air is necessary at all. In those installations the oxygen
concentration in the flue gas (4-4.Si) seems to be sufficient for proper oxidation.
The proper operation of the hydrocyclones is essential when vacuum belt filters are
used for dewatering. Bad separation of brown mud and other fine material results in
bad filtration and washing properties of the cake, which in turn can result in
gypsum with a high moisture and chloride content. Regular control of the apex of the
hydrocyclones is necessary. Dewatering with batch centrifuges is less sensitive to
hydrocyclone separation characteristics. In this case the solids concentration in
the underflow of the hydrocyclones has to be reduced with scrubber suspension to
allow proper filling of the centrifuges.
8B-75
-------
On* FGD plant is equipped with an oversized ¦ act ling tank for thickening of the
gypsum suspension. The overflow of this thickener contains less than 50 mg/1 solids
which means that the underflow contains all the gypsum and most of the contaminants.
The underflow is fed to batch centrifuges.
For a proper functioning of the vacuum belt filters it is essential that the
consistency of the feed allows rapid and even distribution across the surface of the
filter cloth. Solids concentrations between 600 and 1,000 g/1 are used. The cake
thickness is kept between 30 and 40 mm by adjusting the belt speed. Hashing of the
filter cake is done with tap-water, either cold or warm. Moisture contents of 8% or
better can be achieved as is shown in Table 5.
A feed with gypsum crystals of a small mean diameter is more difficult to wash and
dewater. This has to be corrected in the scrubbing loop, but as already has been
mentioned, the influence of operating parameters on crystal size is still rather
erratic. Furthermore, the dewatering characteristics are also negatively influenced
by increased amounts of brown mud and other fines. This can be improved by adjusting
the setting of the hydrocyclones.
Problems have been encountered with tearing of the cloth, which was often caused by
the knife fitted at the end of the belt for removal of the filter cake.
The batch centrifuges guarantee with almost 100% certainty a gypsum moisture content
below 8%. Values of 6% have been reached (see Table 5). The influence of the
particle size of the crystals on dewatering properties is less pronounced.
For a proper operation of the batch centrifuges attention has to be paid to the
washing cycle of the centrifuge itself. This will prevent inbalance. Furthermore, a
concrete foundation is preferable to a steel structure.
Because of their good performance in several German FGD installations, the rela-
tively low price, low energy consumption and small size, one of the power companies
decided to select continuous centrifuges for gypsum dewatering. The operation of
these decanters turned out to be extremely sensitive to the consistency of the
gypsum slurry feed. A small crystal size or a high content of fine material resulted
in bad dewatering. Furthermore, the apparatus were very prone to abrasion damage,
caused by contaminants of the gypsum such as fly ash and the insoluble constituents
of the limestone. A large-scale effort has been made to improve the performance by
testing filter cages of different designs and by increasing the wear resistance of
the feed worm. The lifetime of the filter cages remained about 400 hours, whereas
the lifetime of the feed worm could be extended to 3,000 hours by a plasma sprayed
layer of Cr;Oj. Changes in operation, such as using hot water for cake washing,
decreasing the number of revolutions, and using acid water to flush the filter cage
did not result in sufficient improvement. Because the insufficient reliability of
those decanters could give rise to a decreased availability of the complete FGD
installation, the power company decided to install two batch centrifuges and use the
decanters as a stand-by option.
8B-76
-------
A critical it«n in the gypsum specification is "colour" or "whiteness". Most gypsum
is faintly coloured: yellow, brown or grey. This colour can be attributed to
components such as iron, manganese, fly ash and unburnt carbon, that are among tbe
constituents of the brown mud. One utility company decreased the amount of these
contaminants by changing the absorbent from limestone to hydrated lime and indeed
gained some "whiteness".
Another possibility is the preferential removal of brown mud from the scrubber
system. This has been done by treating part of the topflow of the gypsum hydro-
cyclones with specially designed purge hydrocyclones. The topflow from these purge
hydrocyclones containing the brown mud is discharged to the waste water treatment
plant, and the underflow is returned to the scrubber system as "seed" crystals.
Marked improvements have been realized.
Brown mud removal from the FCD system with the gypsum thickener, by a twofold
increase of the upward flow, did not result in an improvement of the "whiteness".
Finally the gypsum has to be stored under cover prior to transportation to the
industrial consumers.
A silo has been installed in a situation where limited space was available. This
Eurosilo is filled and emptied from above. Separation between good and bad quality
gypsum is not possible which can lead to the deterioration of a larger amount of
good quality gypsum.
An advantage of a warehouse is that a part of the floor area can be reserved for
poor quality gypsum, so that mixing with good product can be avoided. A prerequisite
is that a close eye is kept on the gypsum entering the warehouse. The moisture
content is especially important in this resptrt. A method for the continuous
measurement of the moisture content of gypsum on a conveyor belt is under develop-
ment.
During transportation and the initial phase of the storage the gypsum is still, warm
(40-50 °C) and some evaporation of moisture takes place. This effect has been
quantified in the Eurosilo system, where the moisture content dropped about 1%.
TRACE ELEMENTS
The concentration of the compounds mentioned in the gypsum specification is of the
order of at least 100 pg/g, but gypsum also contains trace elements with a concen-
tration of about 1 (jg/g.
Those trace elements, such as heavy metals, are present in the flue gas (volatile as
well as a constituent of the fly ash), in the lime or limestone and in the make-up
water. Some of them leave the FCD plant with the gypsum. Table 6 gives a summary of
8B-77
-------
the results of the analyses of gypsum from two power plants in the Netherlands (A
and B).
Table 6 also shows the results of the analyses of a limestone used in power 1
station A, and a sample of natural gypsum from Spain, as well as the outcome of a
large-scale research effort by the German power and gypsum industry [2] which
comprised 15 samples from different sources. All values in the table are within the
band width of the German study. The final conclusion from this Btudy was that FGD
and natural gypsum can safely be used as a building material.
Special attention has been given to radioactivity because of the unfavourable
reputation of phosphogypsum in this respect. Table 7 shows a comparison of 7GD
gypsum with phosphogypsum, natural gypsum and several commonly used building
materials in the Netherlands [1]. The specific activity of FGD gypsum is comparable
to natural gypsum and is significantly lower than those of commonly used building
materials. From this point of view no objections to the use of FGD gypsum exist.
The Cs-137 is contamination caused by the Chernobyl accident. The high value was
measured in samples taken in 1987, the low value was measured in 1989. In samples
that were taken before the accident no significant amounts of Cs-137 could be
detected.
Finally, attention has been given to one of the volatile trace elements, mercury. In
laboratory experiments it was found that during calcining at 140 °C for 1 hour, .
between 5 and 30% of the mercury present in FGD gypsum samples evaporated. The mean
value was 10%. This mercury evaporation can also occur in industrial calciners and
may cause emission of mercury to the atmosphere. The total emission and the emission
concentration will be dependent on the method of calcining, the calcining tempera-
ture and the mercury content of the gypsum. The mercury content is dependent on coal
quality, efficiency of mercury removal by the scrubber and operation of the FGD
plant. In the Netherlands no legal limit for mercury emissions from this type of
installation exists. In Germany the TI-Air limits the emission concentration to
200 pg/m„5.
At the Gelderland-13 power station a tentative programme has been carried out to
study the behaviour of mercury in an FGD installation (4] with the intention of
finding ways of reducing the mercury content of FGD gypsum by changing plant
operation. This FGD plant is particularly suitable for this purpose because it has
been built in two stages with two different designs for the gypsum concentration
(3]. Stage 1 has a settling tank, stage 2 has hydrocyclones. The principal findings
of this programme were:
* mercury removal from the flue gas in the scrubber was between 70 and 80%
• the mercury content of the coarse, rapidly settling fraction of the solid
material in the scrubbing suspension was 0.3-0.7 pg/g
8B-78
-------
9 the Mercury eontMt of th« brown mud was 20-40 ftg/g
9 the amount of mercury in solution was 5-10 pg/l.
The most obvious way of reducing the mercury content of gypsum is a sharp separation
between brown mud and gypsum followed by proper washing of the filter caJce.
Furthermore, it is necessary to reduce the mercury content of the brown mud. This
can be done by reducing the residence time of the brown mud in the installation, by
taking specific measures to remove it.
It has already been mentioned that the settling tank of stage 1 is oversized, and no
separation between brown mud and gypsum occurs (Figure 3). Separation occurs in the
centrifuges. The centrate contains 68 g/1 solids with a mercury content of about
20 fjg/g. Because the centrate is returned to the thickener, a build-up of brown mud
in the loop thickener centrifuges takes place. The only outlet for the mercury-
containing brown mud is the gypsum. By treating part of the centrate with purge
hydrocyclones a general reduction of the brown mud, and thus mercury level, is
possible in this loop and a second way out can be created in this way.
Zn stage 2 of the FGD plant, efficient separation between brown mud and gypsum is
accomplished in the hydrocyclones (Figure 4). Part of this effect is nullified
because the underflow has to be diluted with scrubbing suspension to enable proper
filling of the centrifuges. Build-up of mercury-containing brown mud takes place in
the whole system. One way out is again the gypsum, the second way out has been
created by the installing of the purge cyclones for treatment of part of the topflow
from the hydrocyclones. Indeed the presence of these purge hydrocyclones results in
a lower mercury content in the gypsum produced by stage 2: 0.60 pg/g versus
0.97 fjg/g for stage 1. Minor adaptations in the system directed at the removal of
brown mud can indeed lead to a decrease in the mercury content of the gypsum, is has
to be taken into account that an increased discharge of brown mud can create
overloading problems in the waste water treatment plant.
FINAL REMARKS
Flue gas desulphurization is a common practice in the electric power industry in the
Netherlands. Properly designed installations and great dedication of the power plant
crews makes the availability of the FCO systems close to 100%.
Zn addition to SO; removal, attention has to be paid to maintaining the gypsum
quality in line with the requirements of the gypsum industry. Zn particular, the
operation of the gypsum concentrating and dewatering steps has to be closely
watched.
8B-79
-------
Batch centrifuges -turned out to be least sensitive to variation* in gypsum prop-
erties. Proper operation of vacuum belt filters is more sensitive to these prop-
erties, and the plant crew's close attention is required. ^
Colour improvements could be realized by changing from limestone to hydrated line
and by optimizing the operation of purge cyclones.
Currently it can be said that the vast majority of the gypsum produced is in line
with the specifications. Trace elements such as heavy metals and radioactive
elements do not interfere with the applicability.
Summing up, it can be said that the environmental measures taken in the coal-fired
power stations in the Netherlands resulted in a drastic decrease in SO, emissions
without creating secondary environmental problems: the gypsum produced finds its way
into the building industry.
ACKNOWLEDGEMENT
The substance of this paper is the reflection of several years of experience in PCD
operation and a number of special development efforts. He thank Mr. N.A. Doets,
managing director of Amer power station, Mr. J.A. Mout, head of the chemistry
department of Borssele power station, Mr. H. van der Berge, head of the chemistry
department of Maasvlakte power station, Mr. H.P. Klink, head of the chemistryj
department of Gelderland power station, and Mr. H.W. Hoeksema, environmental
engineer of the electricity production company for East and North Netherlands, for
discussions about their efforts to maintain a high FGD gypsum quality standard.
REFERENCES
1. C. van der Lugt. "Radiation Aspects of the Firing of Coal and the Use of Ply
Ash." {In Dutch.) Eneroiesceetrum. December, 1984, pp. 270-283.
2. J. Beckert. "Vergleich von Naturgips und REA-Gips." (Comparison of Natural
Gypsum and FGD Gypsum.) (In German.) VGB Conference Kraftwerk und Omwelt.
Essen, April 1989.
3. F.W. van der Brugghen, J.M. Koppius-Odink, B.C. Kemper and H.P. Klink.
"Operational Experience with the FGD Plant of Gelderland-13 Power Station in
Nijmegen (the Netherlands)." In Desulohurisation 2. Technologies and Strat-
egies for Reducing Sulphur Emissions. Institution of Chemical Engineers.
Symposium Series No. 123. Sheffield, March 20-21, 1991, pp. 67-82.
4. H.W. Hoeksema and F.W. van der Brugghen. "Minimizing Trace Element Concen-
trations in FGD Gypsum." Second International Conference on FGD and Chemical
Gypsum. Toronto, May 12-15, 1991.
8B-80
-------
902
- -A- NOx
~e~ partioulatM
900
-a
a
200
100
-------
400
320
240
160
1990 1988 2000
Figure 2. Expected FGD gypsum production
8B-82
-------
thiofc«Mr
oantrifuges
from scrubber
0.3
waste wator
25
—<—
(mercury flow in g/h)
gypsum
from scrubber
to waste water
from thickener
from centrifuge
gypsum
Solid Content
Suspension
72 g/1
0.2 g/1
270 g/1
68 g/1
Mercury Content
Solids
1.4 pg/g
50 pg/g
7.5 pg/g
20 pg/g
0.98 pg/g
Figure 3. Mercury flow in thickener-centrifuges loop of
stage 1 FGD Gelderland-13
8B-83
-------
13
to
top
itar
18
purge
cyclones
hydrocyclone
0.2
centrifuges
feed tank
collecting vessel
(mercury flow in g/h)
Solid Concent Mercury Content
of Suspension of Solids
from scrubber 110 g/1 2.6 pg/g
centrifuge feed 490 g/1 1.5 pg/g
underflow 1,500 g/1 0.7 pg/g
topflow 7 g/1 39 pg/g
gypsum - 0.6 pg/g
to purge cyclone 22 g/1 7 pg/g
to waste water 9 g/1 27 pg/g
Figure 4. Mercury flow in hydrocyclone-centrifuge area of stage 2
FGD Gelderland-13
8B-84
-------
Table 1
FINAL CONSUMPTION AND GENERATING CAPACITY
ACCORDING TO FUELS IN 1990
Final Consumption
coal
fuel-oil
gas
nuclear
233,640 TJ
1,639 TJ
247,600 TJ
36,876 TJ
45 %
0.3%
47.6%
7.1%
Generating capacity according to fuelling possibilities
fuel-oil
gas
oil/gas
oil/coal
gas/coal
nuclear
37 MW
3,075 MW
7,729 MW
1,705 MW
2,062 MW
508 MW
Table 2
MAXIMUM POSSIBLE GENERATING CAPACITY
FOR EACH FUEL IN THE YEARS
1990 AND 2000
coal
fuel-oil
gas
nuclear
3,767 MW
9,471 MW
12,866 MW
508 MW
2000
5,336 MW
4,628 MW
12,872 MW
508 MW
Table 3
CURRENT STATUS
OF FGD
INSTALLATIONS
IN THE
NETHERLANDS
Capacity
Comnanv
Power Station
Location
MW
Technology
Statu*
EPON
Gelderfand-13-I
Nijmcgcn
300
Chcmico/Miuui
operational
ETON
Gelderiand-13-II
Nijmegcn
302
GEESI
operational
EPZ
Amer-8
Geeruuidenberg
645
GEES]
operational
EPZ
Boruele-12
Bonacle
402
BiacbofT
operational
EPZ
Amcr-9
Gecnnudcnbcfg
600
GEESI
cooatniction
EZH
Maaivlakte-l
Rotterdam
518
BiacbofT
operational
EZH
Mauvltktc-2
Rotterdam
518
BiacbofT
operational
EZH
Maa«vlakie-3
RoaenUm
600
tobcaekcted
licencing
UNA
Hemweg-8
Amacrdam
600
GEESI
conabuction
8B-85
-------
Table 4
FGD GYPSUM
SPECIFICATIONS OF THE GYPSUM INDUSTRY
VPEW/VGB
Moisture weight %
CaS04.2HjO weight %
sulphite (as SOj) weight %
CI weight %
CaCOj weight %
HgO (soluble) weight %
Ra^O weight %
KjO weight %
FejOj weight %
SiOj weight %
PH
colour* "whiteness"
NH4
toxic components
<10
<95
< 0.25
0.02
2.5
0.05
0.05
0.06
0.5
2.5
5-8
>75**
8-10
<95
< 0.25
< 0.01
0.05
0.04
0.1
1.5
1.0
.5-7.5
>70
<10
<95
< 0.25
< 0.01
< 0.1
< 0.06
5-9
>80
0
none
reference BaSO, is 100%
after calcining
Table 5
COMPOSITION OF GYPSUM PRODUCED BY
POWER STATIONS IN THE NETHERLANDS
Centrifuge/Limestone Belt filter/Lime
moisture
weight
%
6.1 - 7.0
7.8
CaS04.2Hj0
weight
%
96 -98
97
pH
7.5 - 7.7
7.7
SOj
weight
%
0.01 - 0.02
-
CaCO,
weight
%
1.2 - 2.5
0.03
A 1,0,
weight
%
0.5 - 0.7
0.07
Fep,
weight
%
0.12 - 0.17
0.02
SiO,
weight
%
0.3 - 0.5
0.3
MgO
weight
%
0.05 - 0.07
0.06
KjO
weight
%
0.05
0.01
Ha.fi
weight
%
0.005
0.04
CI
weight
%
0.002- 0.005
0.002
colour*
46 -50
86
BaS04 standard as a reference
8B-86
-------
Tabl* 6
TRACE ELEMENT CONCENTRATIONS
IN FGC GYPSUM
(P9)
AS
Cd
Cu
Cr
Hg
Ni
Pb
Zn
1.0-1.4
0.05
0.7
4 -8
1.2
1.0-2.5
5 -7
<7.5
B
< 3
0.05
7
<10
1.3
5
11
<10
Natural
Gypsum
1.7
0.33
7
6.8
0.02
3
3
4.0
Limestone
2
0.08
1.2
1.4
<0.2
0.7
6.5
<5
Beckert
0.21- 2.70
0.03- 0.29
1.10- 8.5
1.0 - 9.7
0.03- 1.32
0.55-12.9
3 -22
2 -53
Table 7
RADIOACTIVITY OF GYPSUM AND
BUILDING MATERIALS
Ra-226
FGD gypsum 1- 5
Natural gypsum 5
Phosphogypsum 20-110
Sand 8
Gravel 9
Portland cement 47
Sand/lime brick 8
Brick 45
Th-232 K—40 C8—137
0.8-3 8-22 0.2-1.3
6 50 -
- 5 -
8 2S0
9 150
19 240
9 270
40 600
8B-87
-------
8B-88
-------
Chemical Analysis and FlowabQity of
ByProduct Gypsums
Preceding page blank
-------
8B-90
-------
L. A. Kilpeck
Senior Design Engineer
Centerior Energy
P. 0. Box 94661
Cleveland, Ohio 44101-4661
B. E. Dose, Jr.
Associate Chemical Engineer
B. E. Basel,
Projecc Engineer
Burns & McDonnell Engineering Co.
4800 E. 63rd Street
Kansas City, Missouri 64130-4696
ABSTRACT
I Several forced-oxidized FGD processes produce synthetic gypsum in the U.S. Knowledge
of the chemical and physical characteristics of synthetic gypsum is requisite to assess
handling and disposal alternatives. Results of chemical, physical properties, and
flowability testing of several representative FGD system synthetic gypsums are
presented.
Physical properties of synthetic and natural gypsum samples are discussed, including
freezing point, abrasion resistance of several liner materials, and particle size.
Physical property data also cover a range of moisture levels and ambient and
near-freezing temperature conditions. Gypsum flowability is assessed based on direct
shear tests, coefficient of friction, and critical moisture content. Derived
flowability results include hopper and chute slopes, arching dimensions, outlet sizes,
stable rathole diameter, and rail car hopper configurations. Chemical analyses of
solids and water leachate are presented for the synthetic and natural gypsum samples
for selected parameters.
Test program results are compared to published literature values and are summarized in
connection with material handling requirements.
Preceding page blank
8B-91
-------
INTRODUCTION
Many utilities are considering the addition of wee limestone forced oxidized (LSFO)
flue gas desulfurization (FGD) systems which will produce gypsum byproduct in Che form
of vacuum filter cake. This paper is based on a study which was conducted to
accomplish the following primary objectives:
1. Obtain comparative chemical analyses of several natural and byproduct
gypsum samples.
2. Assess flowablllty and related physical characteristics of
byproduct gypsum for use in design of material handling equipment.
METHODOLOGY
The study consisted of a literature search to compile information on properties of
gypsum and testing of FGD byproduct and natural gypsum samples from several sources.
Commercial laboratories performed chemical analysis, flowablllty and physical
properties testing, particle size analysis, and wear testing.
Chemical analyses were performed on three natural gypsum samples and on three byproduct
gypsum filter cake samples obtained from operating LSFO FGD systems as follows:
Natural Byproduct
Iowa Plant A
Nova Scotia Plant B
Oklahoma Plant C
Byproduct and natural gypsum samples were prepared and analyzed in accordance with
standard analytical methods summarized in Table 1.
CHEMICAL ANALYSIS OF SOLIDS
Chemical analyses of "as received" natural and byproduct gypsum samples are presented
in Table 2.
The trace elements found in the natural and byproduct gypsum samples are compared to
8B-92
-------
literature values in Table 4. The levels of trace elements found in the gypsum samples
were compared to literature values and found to be similar to those previously
reported.
CHEMICAL ANALYSIS OF LEACHATE
Chemical analyses of the aqueous leachate from natural and byproduct gypsum samples are
summarized in Table 3. Analysis results vere compared to contaminant levels
(MCL) specified by the EPA (40CFR141 and 40CFR143) which are summarized in Table S.
Ohio standards (Table 6) require constituents in the leachates to be less than 15 times
the drinking water standards. Overall, the measured concentrations in the leachates
shown in Table 3 are well below the required Ohio levels.
Literature values for aqueous leachate and runoff from byproduct gypsum are compared
in Table 7 to averages for the natural and byproduct gypsum samples analyzed. The
values found in the gypsum samples analyzed are very similar to literature values
reported in Table 7 for gypsum disposal site underdrain leachate and pond runoff.
These literature data were pertinent because gypsum is similar to the samples tested.
Total water soluble salts, excluding calcium sulfate, in the leachates tested were
estimated by subtracting calcium plus sulfate from TDS. Dsing this technique, water
soluble salts for the natural gypsum samples ranged 68 to 217 mg/1 compared to the
byproduct gypsum samples which ranged 160 to 322 mg/1. Gypsum is soluble in water with
reported solubility ranging from 2440 to 2640 mg/1 at 10*C to 30*C, respectively (1).
Tables 3 and 7 indicate that natural and byproduct gypsum samples contained water
soluble salts and constituents other than calcium sulfate. Reported values for water
soluble salts in several natural and synthetic gypsums are listed as ranges in Table
8 (2.). In natural gjrpsum the salts are found to be deposited evenly over the surface
of crystals, resulting from the slow, near equilibrium deposition (2.). Contrastingly,
synthetic gypsum, which is precipitated much more rapidly from solution than the
natural gypsum, could have co-precipitated salts distributed throughout the particle
as well as on the surface (2).
Constituent values for leachate were used to calculate corresponding concentrations
present in gypsum solids. Water soluble salts present in the natural gypsum samples
analyzed ranged 1360 to 4340 mg/kg (or ppm) compared to byproduct gypsum samples which
ranged 3200 to 6440 ppm. The byproduct gypsum values are within the range of
literature values displayed in Table 8. Similar projections for chloride show a range
of 40 to 60 ppm for the natural gypsum samples analyzed and a range of 180 to 520 ppm
8B-93
-------
for Che byproducc gypsum samples tested. These ranges may be compared Co the
commercial-grade gypsum specifications for chloride in Table 9 which range from 100 Co
400 ppm.
PARTICLE SIZE ANALYSIS
Particle size analysis for Che byproducc gypsum samples showed the following arichmecic
mean size values on a volume or mass basis compared Co mean size on a frequency or
councs basis.
Byproducc Gypsum Mean Farcicle Sizes
Mass Frequency
Basis Basis
Size. inn jim
Plane A 41.28 9.804
Plane B 29.25 9.411
Plane C 43.65 26.97
To examine any pocencial differences in the byproducc gypsum parcicle morphology
(scruccure) not revealed by che parcicle size analysis, opcical color phocomicrographs
under polarized lighc ac 100X magnification were obcained. The phocomicrographs for
Plane A, Plane B, and Plane C are presenced in Figures 1, 2, and 3, respectively.
ii^t
The phocomicrographs for Plane A and Plane C show many parcicles of similar size whil
Plane B shows more parcicles of a smaller size range. The phocomicrographs disclc
significant differences in parcicle morphology. Plane C parcicles appear
characceriscic of gypsum and are blocky eo rounded in shape and relaclvely uniform in
size, approximately 25 Co 50 fm (3. 4. 5). The Plane A parcicles appear Co include
large gypsum needles which range in size from 25 fan pieces Co needles near 250 /an in
lengch (.6) . Numerous gypsum parcicles in Che 25 Co 50 pm range are rounded in form.
Calcium sulfiee parcicles in che form of small placelecs also appear co be present (3.
5). The presence of fly ash particles in Che form of spheres is noced in all three
samples, more so for Plane A and Plane B. Small calcium sulfite placelecs appear co
be presenc in che Plane B sample. The mosc outstanding characceriscic of che Plane
B sample is Che visual indicacion of a large fraceion of gypsum needles which appear
Co range in lengch from about 25 faa Co 100 fta with proporcionacely varying narrow
widths (6) .
The chemical composicion of Che gypsum solids as shown in Table 2 indicates obvious
differences in levels of calcium sulfate, calcium sulfite, calcium carbonate or
unreacced limestone, and fly ash and ineres. Significant differences were also
observed in percenc oxidation levels for each byproducc; percent oxidation was 85.
78.8, and 99.7 percenc, respeccively, for Plane A , Plane B , and Plane C. The
8B-94
-------
correlation of larger mass mean particle size with increasing oxidation level is
depicted graphically in Figure 4.
Many design and process related factors that can influence the shape, size, and size
distribution of byproduct gypsum solids and similar materials, have been described by
others (3. 4. 5. 6. 7. 8). In naturally oxidized waste materials containing a large
fraction of sulfite, the presence of sulfate has not been found to affect the physical
properties of the wastes (.5). However, for gypsum waste, in which the major component
is sulfate, laboratory tests have shown that small amounts of sulfite have an effect
on the physical properties (3. 5. 7). For example, samples of completely oxidized
material exhibited substantially higher unconfined compressive strength than a material
containing 5 percent sulfite (5) . The unconfined compressive strength of the latter
material was approximately equal to that of FGD waste that is predominantly calcium
sulfite (5.) .
Thus, oxidation level and particle size and morphology can have measurable influence
on various material physical properties.
BYPRODUCT PHYSICAL PROPERTIES TESTING
A number of material and physical properties were determined on byproduct samples from
Plant A and Plant B as summarized below.
Byproduct Gypsum Physical Properties
Plant A
Plant B
Moisture
Free (w/o hydrated water)
Percent
17.2
20.3
Total (w/hydrated water)
Percent
29.3
28.7
Angle of Repose
Degrees
35-39
33-36
Specific Gravity
2.68
3.08
Modified Proctor
Optimum Moisture
Percent;
23.2
28.7
Max. Dry Height
Lb/Ft3
91.4
88.4
Freezing Point
° C
-2
-2
Permeability
Cm/Sec
4.6E-06
1.2E-05
Values for specific gravity were significantly higher than the value for pure gypsum
at 2.31 (9). A contributing cause for the discrepancy may be the use of a drying
temperature of 110° C which would cause loss of waters of hydration.
Abrasion Testing
Abrasion tests were conducted on the "as received" Plant A sample on liners of Type
8B-95
-------
304 stainless steel with a 2B finish (SS304-2B), aged carbon steel (A-36), and high
density polyethylene (HDPE)(Tivar-88) with a new, smooth surface.
The results of the abrasion tests revealed that a thin coating of byproduct tended to
form on the surfaces of the SS304-2B and A-36 carbon steel liners during the course of
the test, and, consequently, no liner loss was measured. Vhen compared with tests
conducted on similar HDPE liners vith an Eastern U.S. coal and a Western U.S. coal, the
Plant A byproduct may be considered less abrasive than coal.
BYPRODUCT FLOWABILITY TESTING
Flowability direct shear tests and byproduct/liner friction tests were performed on the
Plant A and Plant B byproduct gypsum samples. A total of 114 direct shear flowability
tests and 32 wall friction tests were performed. The testing program covered several
moisture levels, temperature conditions, and liner materials.
The critical strength and frictional properties measured during the flowability testing
program were used to compute the minimum hopper slopes, outlet dimensions, funnel flow
dimensions, and stable rathole diameters that are required for reliable flow of the
gypsum byproducts.
Flowability Direct Shear Tests
Instantaneous and time consolidation direct shear flow tests vere conducted on each
gypsum byproduct under continuous instantaneous flow and three-day extended (time
consolidated) storage conditions. Time consolidation testing vas only performed near
the critical moisture content, defined as the moisture level which results in the
steepest flow function, or the least flowable condition. Instantaneous flowability
tests vere conducted on each "as-received" sample at a "near-freezing" temperature,
defined as one degree centigrade above the freezing point (-2* C).
The Plant A sample displayed the highest overall strength at the "as-received" state
vith 17.2 percent free moisture. The Plant A byproduct exhibited a moderate gain in
strength and cohesion after a three-day storage period under consolidating pressures.
The Plant B byproduct displayed the greatest overall strength during instantaneous
testing and showed substantial gain in strength after time consolidation.
Freeze Testing
The results of the freezing tests at minus 1° C indicate that the "as-receive
byproducts gained little or no strength under the low temperature conditions
compared to the strengths at ambient room temperature conditions.
8B-96
L
-------
Liner Friction Tests
Shear tests on composite byproduct/liner samples, using Plant A and Plant B byproducts,
were used to determine the wall friction coefficient for SS304-2B, mild steel with a
corroded surface, and HDPE with both a smooch (new) surface and abraded (in-service)
surface.
The results of the tests indicated that the high density polyethylene liner,
particularly the smooth polyethylene, offered the least resistance to sliding with
either byproduct. For both samples, frictional resistance to sliding on the smooth
HDPE liner decreased slightly vich lower byproduct moisture, while resistance to
sliding on the abraded HDPE liner increased slightly vich lower byproduct moisture.
Vail friccion tests at the low temperature conditions indicated that only Che SS304-2B
liner exhibiced any increased resistance to byproduct sliding when compared with the
sliding resistance at ambient room Cemperature conditions.
RAIL CAR DESIGN
Hopper
To promote reliable gravity flow of gypsum byproduct, the hoppers of the rail cars used
to transport the byproduct should incorporate a mass flow design, described as follows:
Mass flow is a bulk material flow paccern which develops when che encire
storage vessel contents are set in motion as material is withdrawn from
the outlets. This flow pattern results in a first-in, first-out flow of
material. The entire volume of material in the storage vessel is moving
with no stagnant or dead zones. To achieve reliable withdrawal, che
ouclet dimension must exceed Che mass flow arching dimension of Che bulk
material.
A mass flow design is required because the anticipated raChole diameters for both
materials are significantly larger Chan che ancicipaced arching dimensions. Inhibiting
the formation of stable racholes decreases the likelihood chat a steep-walled, scable
material channel will form in the rail car. Formation of such stable channels would
necessitate car shaking for removal of all byproduct. Hopper recommendations for
gravity mass flow, based on no-arching condition, in a plane-flow hopper for Plant B
byproduct in continuous flow are shown as follows.
8B-97
-------
Hopper Recommendations for Gravity Mass Flow
Byproduct
Min.
Min.
Stable
Free
Hopper
Hopper
Outlet
Rathole
Moisture.%
Surface
Slope.Dee
Dimension.Ft
Dia..Ft
20.3
SS304-2B
79
9.0 x 27.0
24.0
HDPE-Smooth
68
9.2 x 27.6
13.7
SS304-2B
83
4.8 x 14.4
22.7
HDPE-Smooth
66
5.15 x 15.45
8.5
SS304-2B
88
4.95 x 14.85
18.7
HDPE-Smooth
64
5.35 x 16.05
Liner
The entire sloping surfaces of the rail car hoppers should be lined with HDPE because
flow of gypsum byproduct can be initiated on significantly shallower slopes than that
possible for either of the steel liners. The HDPE has a lower moisture absorption than
the steel liners, thus helping fine, wet, bulk materials slide at low temperature
conditions. Based on the results of the abrasion tests, reasonable liner life is
expected with hopper cars lined with HDPE.
Car liner selection should also consider the anticipated temperature environment to
which the loaded rail car will be subjected. If loaded rail cars might be passed
through a "thaw shed" in the winter time, then the resultant effects on any rail car
liner, particularly polymeric lining materials, should be evaluated. Such concernsi
not exist with either Type 304 stainless steel or mild steel liners.
Other Liner Applications
Liners made of SS304-2B and HDPE can also be applied to the hoppers and chutes in a
variety of material handling equipment and in the beds of dump trucks.
Rotary Dump Rail Cars
Rotary dump cars could be an alternative byproduct transportation mode using rail cars.
Flow of bulk material from a rail car hopper rotated 140 to 160 degrees from vertical
would be considered an asymmetric plane-flow regime. Rail cars lined with HDPE or
ultra-high molecular weight (OHMW) polymer will likely promote flow of the gypsum
byproduct if the rail car is rotated 145 to 155 degrees from the vertical. Cars lined
with SS304-2B will likely require rotation of at least 160 degrees to evacuate the
contents. In either application, additional car cleanout will likely be required.
CORRELATION OF RESULTS
For damp FGD gypsum (95 percent gypsum, average particle size SO /xm) with 9 perc^
free moisture, instantaneous values of unconfined yield strength plotted against t^P
major consolidation stress show that the characteristics lie in the worst flowability
8B-98
-------
zone (very cohesive and non-flowing) as defined by Jenike. This is shown in Figures
5 and 6 compared Co data for Plant A and Plane B byproduct gypsums, respectively (10.
11) . The flowabilicy characteristics of damp FGD gympsum are worse than for 20 percent
moisture raw coal or 10 percent moisture wet sand (.10). After a period of seven days
time consolidation, the FGD gypsum flowability characteristics degrade significantly
(10. 11). The Plant A and Plant B byproducts significantly exceed the poor flowability
characteristics of damp FGD gypsum described in the literature.
Several of the properties investigated in the study of Plant A and Plant B byproduct
gypsums appear to be largely responsible for their flowability behavior in comparison
to the literature. The properties having the greatest effect on flowability are free
moisture, particle size, particle morphology, and percent oxidation.
CONCLUSIONS
The overall purity of the FGD byproduct gypsum samples tested was well below that of
most high-grade or commercial-grade gypsum for which flowability characteristics are
described in the literature. The presence of impurities, such as fly ash, unreacted
limestone, calcium sulfite, and inerts, contributed to the poor flowability of the
byproduct gypsum tested compared to "high grade" or commercial grade gypsum in the
literature. The characteristics of the particles associated with these impurities in
the byproduct gypsum samples tested appeared to cause significant departure in their
flowability behavior from that of high-grade gypsum. The presence of moisture in any
byproduct gypsum creates poor flowability characteristics.
Chemical analysis data for the byproduct and natural gypsum samples tested showed
levels of trace elements which were similar to those reported in the literature. The
concentrations of constituents analyzed in the aqueous leachates from the byproduct and
natural gypsum samples tested were well below the required Ohio levels.
REFERENCES
1. GARLANGER, J.E. AND INGRA, T.S., ARDAMAN & ASSOCIATES, INC., "EVALUATION OF
CHIY0DA THOROUGHBRED 121 FGD PROCESS AND GYPSUM STACKING, VOLUME 3: TESTING AND
FEASIBILITY OF STACKING FGD GYPSUM", ELECTRIC POWER RESEARCH INSTITUTE, CS-1579,
RESEARCH PROJECT 536-3, NOVEMBER 1980.
2. BARBER, J.W., ET AL, "CHARACTERIZATION OF SYNTHETIC GYPSUM", PROCEEDINGS OF A
CONFERENCE PRESENTED BY ORTECH INTERNATIONAL. TORONTO 1991; MAY
12-15:pp.26.1-26.15.
3. CROWE, J.L. AND SEALE, S.K., "FULL-SCALE SCRUBBER SLUDGE CHARACTERIZATION
STUDIES", ELECTRIC POWER RESEARCH INSTITUTE. FP-942, RESEARCH PROJECT 537-1, FINAL
REPORT, JANUARY 1979.
8B-99
-------
u
5
6
7
8
9
10
11
12
13
FEENEY, STEVE ET AL. "IN-SITU FORCED OXIDATION RETROFIT AT MICHIGAN SOUTH CENTRAL
POWER AGENCY'S ENDICOTT STATION", PROCEEDINGS OF CONFERENCE PRESENTED BY ORTECH
INTERNATIONAL. TORONTO 1991; MAY 12-15:pp.l0.1-10.18.
HURT. P.R. ET AL, "DISPOSAL OF FLUE GAS DESULFURIZATION WASTES: EPA SHAWNEE FI^
EVALUATION FINAL REPORT", EPA 600/7-81-103 (NTIS PB 81 212 482), U.S.
ENVIRONMENTAL PROTECTION AGENCY, JUNE 1981.
MAKKINEJAD, DR. DAVID ET AL, "VARIABLES AFFECTING THE PRODUCTION OF COMMERCIAL
GRADE GYPSUM FROM THE FGD PROCESS" , PROCEEDINGS OF CONFERENCE PRESSENTED BY ORTECH
INTERNATIONAL. TORONTO 1991; MAY 12-15: pp.11.1-11.12.
ROSSOFF, J. ET AL, "DISPOSAL OF BY-PRODUCTS FROM NONREGENERALBLE FLUE GAS
DESULFURIZATION SYSTEMS: FINAL REPORT", EPA 600/7-79-046, U.S. ENVIRONMENTAL
PROTECTION AGENCY. FEBRUARY 1979.
WILTERDINK, D.W. ET AL, "GYPSUM-AN FGD BYPRODUCT", COAL TECHNOLOGY '85.
INTERNATIONAL COAL UTILIZATION EXHIBITION AND CONFERENCE, pp.187-199.
WIRSCHING, DR. FRANZ, "FGD GYPSUM IN RESEARCH, DEVELOPMENT, PRODUCTION AND
APPLICATION". PROCEEDINGS OF CONFERENCE PRESENTED BY ORTECH INTERNATIONAL. TORONTO
1991; MAY 12-15: pp.18.1-18.18.
RADEHACHER, F.J.C. ET AL, "ENVIRONMENTALLY-CONSCIOUS STORAGE AND HANDLING OF HUGE
QUANTITIES OF COAL AND DAMP FLDE GAS GYPSUM AT POWER STATIONS BY MEANS OF
EUROSILOS", BULK SOLIDS HANDLING. FEBRUARY 1988, pp.13-17.
WRIGHT, H. . "LARGE SCALE HANDLING OF BULK. MATERIALS", BULK SOLIDS HANDLING.
FEBRUARY 1990, pp.175-179.
KOCMAN, VLADIMIR, "MODERN METHODS FOR THE ANALYSIS OF FLDE GAS AND BYPR0d(J
GYPSUMS", PROCEEDINGS OF A SEMINAR PRESENTED BY ORTECH INTERNATIONAL. TORONTO
1988; NOVEMBER: pp.155-164.
STEFFAN, PATRICIA AND GOLDEN, DEAN, "FGD GYPSUM UTILIZATION: SURVEY OF CURRENT
PRACTICES AND ASSESSMENT OF MARKET POTENTIAL" , PROCEEDINGS OF CONFERENCE PRESENTED
BY ORTECH INTERNATIONAL. TORONTO 1991; MAY 12-15: PP.4.1-4.18.
8B-100
-------
55TO
«
Figure 1. Byproduct gypsum - Plant A; 100X magnification
Figure 2. Byproduct gypsum - Plant B; 100X magnification
«
Figure 3. Byproduct gypsum - Plant C; 100X magnification
*Please note that the illustration^) on this page have been reduced 10% during printing
8B-101
-------
£
>->
o
ro
E 10:
<
DL
T 1—I—I I II I I
"i i i i m 11 r
10
1
% GREATER THAN
1111
Plant A (85% ox.) H— Plant B (79 % ox.) Plant C (99% ox)
-a- Ref. (0) (98 %ox) -X- Ref. (8) (34 % ox.)
Figure 4. Oxidation state vs. particle size
-------
o
-------
Table 1
SUMMARY OF ANALYTICAL METHODS
Designation
C471-87
C471-87 (6)
C471-87 (8)
C471-87 (9)
C471-87 (13)
03987-85
#41
EPA 160.1
EPA 300.0
EPA 310.1
EPA 340.2
EPA 3050
EPA 6010
EPA 7470
EPA 9040
SM303E
Source
ASTH*
astm'
astm'
astm'
ASTH1
ASTM'
EPRI2
SV-8453
EPA*
SW-8463
SW-8463
SW-8463
SW-8463
SW-8463
SW-846,
EPA1
Title
Hethod for Chemical Analysis of Gypsun and Gypsun Products
Free Water
Carbon Dioxide
Silicon 0ioxide and Insoluble Hatter
Sulfur Trioxide
Method for Shake Extraction of Solid Wastes with Water
Analysis of Sulfite in Scrubber Liquor and Slurry Solids by
lodine-Arsenite Titration
Residue. Filterable. Gravimetric, Dried at 180"C
The Determination of Inorganic Ions in Water by Ion
Chromatography
Alkalinity. Titrimetric (pH 4.5)
Fluoride. Potenticmetric. Ion Selective Electrode
Acid Oigestion of Sediment Sludges and Soils
Inductively Coupled Plasma Atomic Emission Spectroscopy
Mercury in Liquid Waste (Hanual Cold-Vapor Technique)
pH Electrometric Measurement
Arsenic. AA Hydride (Atonic Absorption - 6aseous Hydride)
1. American Society for Testing and Haterials.
2. FGD Chemistry and Analytical Methods Handbook. Volune 2: Chemical and Physical Test Methods. Revision 1. Electric
Power Research Institute. EPRI CS-3612.
3. Test Methods for Evaluating Solid Waste. SW-846. EPA-600/4-79-020. EPA.
4. EPA-600/4-84-017. EPA.
5. Standard Methods for Water and Wastewater. EPA.
8B-104
-------
Table 2
Gypsum Solids Analysis
Natural Gypsum
Detection
Analysis
Units
Limit
lowa
Nova Scotia
Oklahoma
Sulfur Trloxlde(S03)
%
31.25
39.75
35.50
Free Water
%
<0.10
<0.10
<0.10
Silicon Dioxide and
%
10.78
9.40
5.00
Insoluble Matter
Carbon Dioxide (C02)
%
1.07
0.57
0.33
Loss on Ignition (550C)
%
16.79
19.65
20.14
Sulfite (S03)
%
0.066
0.054
0.154
Total Aluminum
mg/kg
6
824
113
176
Total Arsenic
mg/kg
10
<10
<10
<10
Total Boron
mg/kg
5
9.99
14.4
12.4
Total Cadmium
mg/kg
0.5
<0.5
<0.5
<0.6
Total Barium
mg/kg
0.5
5.84
2.08
2.13
Total Calcium
mg/kg
1
166214
199588
172421
Total Chromium
mg/kg
1
<1.0
<1.0
<1.0
Total Cobalt
mg/kg
1
<1.0
<1.0
<1.0
Total Copper
mg/kg
1
1.22
<1.0
1.72
Total Iron
mg/kg
1
1109
160
90
Total Lead
mg/kg
<5
<5
<5
Total Magnesium
mg/kg
1
4340
714
611
Total Manganese
mg/kg
1
20
14.6
8.03
Total Molybdenum
mg/kg
1
<1.0
2.66
<1.0
Total Nickel
mg/kg
1
4.19
<1
<1
Total Phosphorus
mg/kg
10
42
32.2
43
Total Potassium
mg/kg
10
433
92.4
53
Total Selenium
mg/kg
10
<10
<10
<10
Total Silver
mg/kg
1
<1.0
<1.0
<1.0
Total Sodium
mg/kg
1
228
111
450
Total Sulfur
mg/kg
25
124758
158779
141916
Total Zinc
mg/kg
1
5.86
1.67
1.81
Byproduct Gypsum
Avg. Avg.
Plant A
Plant B
Plant C
Natural
Byproduct
40.75
36.50
35.75
35.50
37.67
15.94
18.62
9.68
<0.10
14.75
4.98
2.99
2.20
8.39
3.39
3.83
6.56
0.46
0.66
3.62
30.22
31.86
28.62
18.86
30.23
7.20
9.80
0104
0.09
5.70
165
282
268
371
238
<10
<10
<10
<10
<10
18
52
19
12.26
29.67
<0.5
<0.5
<0.5
<0.5
<0.5
0.97
3.19
2.06
3.35
2.07
198724
197162
176564
179408
190817
<1.0
1.03
5.85
<1.0
<2.63
<1.0
<1.0
<1.0
<1.0
<1.0
<1.0
2.25
1.52
<1.31
<1.59
335
838
269
453
481
<5
<5
<5
<5
<5
245
1521
182
1888
649
17
101
2.13
14.21
40.04
<1.0
<1.0
<1.0
<1.55
<1
<1
1.22
<1
<2.06
<1.07
128
42
154
39.07
108
67
151
39
192.8
85.7
<10
<10
<10
<10
<10
<1.0
<1.0
<1.0
<1.0
<1.0
107
105
246
263
153
159828
145666
142690
141818
149395
2.01
5.86
4.81
3.11
4.23
-------
Table 3
Gypsum Water Leachale Analysis
Natural Qypsum
Detection
Analysis
Units
Limit
Iowa
Nova Scotia
Oklahoma
Total Dissolved Solids
mg/L
10
2310
2253
2182
Fluoride
mg/L
0.1
0.2
0.3
0.1
(without distillation)
Chloride
mg/L
1
3
3
2
Sulfate
mg/L
10
1451
1429
1489
Nitrate Nitrogen
mg/L
0.2
5.1
8.5
1
PH
S.U.
7.28
3.32
5.75
Alkalinity
mg/L as
C03
CaC03
mg/L as
0.01
0.02
<0.01
<0.01
HC03
CaC03
10
12
<10
<10
Aluminum
mg/L
0.05
0.24
0.31
0.25
Boron
mg/L
0.05
0.21
0.38
0.38
Cadmium
mg/L
0.005
<0.005
<0.005
<0.005
Barium
mg/L
0.005
0.02
0.01
0.007
Calcium
mg/L
0.01
642
639
625
Chromium
mg/L
0.01
<0.01
<0.01
<0.01
Cobalt
mg/L
0.01
<0.01
<0.01
<0.01
Copper
mg/L
0.01
<0.01
<0.01
<0.01
Iron
mg/L
0.05
<0.05
0.05
<0.05
Lead
mg/L
0.05
<0.05
0.07
<0.05
Magnesium
mg/L
0.01
1.61
0.30
0.20
Manganese
mg/L
0.01
0.01
<0.01
<0.01
Molybdenum
mg/L
0.05
<0.05
0.12
<0.01
Nickel
mg/L
0.01
<0.01
<0.01
<0.01
Phosphorus
mg/L
0.10
<0.10
<0.10
<0.10
Potassium
mg/L
0.10
1.54
0.73
1.29
Selenium
mg/L
0.10
<0.10
<0.10
<0.10
Silver
mg/L
0.01
<0.01
<0.01
<0.01
Sodium
mg/L
0.01
9.32
2.18
22
Sullur
mg/L
0.10
535
527
531
Zinc
mg/L
0.01
<0.01
<0.01
<0.01
Arsenic
mg/L
0.005
<0.005
<0.005
<0.005
Mercury
mg/L
0.005
<0.005
<0.005
<0.005
Byproduct Qypsum
Avg. Avg.
3lant A
Plant B
Plant C
Natural
Byproduct
2479
2476
2285
2248
2413
3.6
2.5
4.0
0.2
3.4
26
9
18
2.7
17.7
1492
1503
1491
1456
1495
1.2
05
2.3
4.9
1.3
8.28
7.94
6.01
5.45
7.41
0.35
0.17
<0.01
<0.01
<0.18
19
0.21
<10
<10.7
<9.7
0.33
0.21
0.41
0.27
0.32
0.64
2.60
0.66
0.32
1.27
<0.005
<0.005
<0.005
<0.005
<0.005
0.007
0.01
0.02
0.012
0.012
708
651
634
635.33
664.33
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.01
<0.01
<0.01
<0.01
<0.05
<0.05
0.26
<0.05
<0.12
<0.05
<0.05
<0.05
<0.06
<0.05
2.92
18.0
0.28
0.70
7.07
0.16
0.71
0.05
<0.01
0.31
<0.01
<0.05
<0.05
<0.06
<0.04
<0.01
<0.01
<0.01
<0.01
<0.01
<0.10
0.18
0.16
<0.10
<0.15
<0.10
<0.10
0.36
1.19
<0.19
<0.10
<0.10
<0.10
<0.10
<0.10
0.01
<0.01
<0.01
<0.01
<0.01
1.66
1.82
7.81
11.17
3.76
570
689
514
531
558
<0.01
<0.01
<0.01
<0.01
<0.01
<0.005
<0.005
<0.005
<0.005
<0.005
<0.005
<0.005
<0.005
<0.005
<0.005
-------
Table 4
LITERATURE VALUES FOR TRACE ELEMENTS
IN NATURAL AND BYPRODUCT GYPSUH SOLIDS
(tag/kg or ppm)
Avg.1
Avg.1
Range®
Range4
Range*
Natural
Bvoroduct
Bvoroduct
Bvoroduct
Natural
Zinc
3.11
4.23
2.9 - <7
2.0
4.0 - 6.0
Cacfariua
<0.5
<0.5
-
<1.0
<0.2
Chrooiun
<1.0
<2.63
2.2 - 11
3.0 - 10.0
<5.0 - 5.0
Nickel
<2.06
<1.07
-
<5.0
<5.0
Cobalt
<1.0
<1.0
0.05 - 0.46
<1.0
<2.0
Copper
<1.31
<1.59
D.6 - 3.2
3.0
3.0 - 6.0
Lead
<5
<5
-
<1.0
2.0
Tin
-
-
-
<5.0
<50.0
Hoi ybdenua
<1.55
<1
0.35 - 1.74
5.0 - 6.0
<5.0
Fluorine
-
-
95 - 729
321 - 475
20.0 - 105
Arsenic
<10
<10
0.14 - 1.74
<5.0
1.0 - 1.4
Antimony
-
-
0.028 - 0.06S
<5.0
0.8 - 1.0
Mercury
-
-
0.04 - 0.36
<1.0
<0.2
Seleniun
<10
<10
<1 - 5.2
<5.0
<0.2
Notes:
1. Analyzed for this study. Iowa. Nova Scotia. Oklahoma.
2. Analyzed for this study. Plant A. Plant 8. Plant C.
3. From five FGD gypsuas in Reference (12).
4. From two synthetic gypsuas in Reference (2).
5. From two natural gypsuas in Reference (2).
Table S
EPA DRINKING WATER REGULATIONS
Arsenic (As)
Bariua (Ba)
Cadsiua (Cd)
Chrcraiun (Cr)
Lead (Pb)
Hercury (Hg)
Nitrate (NO, as N)
Seleniin (Se)
Fluoride (F)
A1 minus (Al)
Chloride (CI)
Color
Copper (Cu)
Corrosi vi ty
Fluoride (F)
Foaming Agents
Iron (Fe)
Manganese (Hn)
Odor
pH
Silver (Ag)
Sulfate (SO.)
TDS
Zinc (Zn)
Primary Drinking Water Regulations
WaxiwuB Contaminant Level (mo/11*
0.05
1
0.010
o.os
o.os
0.002
10
0.01
4.0
Secondary Drinking Water Regulations
Haiti mun Contaminant level (mg/1)**
O.OS to 0.2
250
15 color units
1.0
Non-Corrosive
2.0
0.5
0.3
O.OS
3 threshold odor ninber
6.5 to 8.5
0.1
250
500
5
Maxima permissible level
These regulations control contaminants in drinking water that primarily affect the aesthetic
qualities relating to public acceptance of drinking water. The regulations are not Federally
enforced but are intended as guidelines for the states.
8B-107
-------
Table 6
OHIO LEACHATE REGULATIONS
Contanrinant Kaximun Contaminant level fpom)
Arsenic 0.75
Bariira 15.0
Cadniin 0.15
Chraaiua 0.75
Lead 0.75
Mercury 0.03
Seleniun 0.15
Stiver 0.75
Cyanide 3.0
Phenolics 4.5
Table 7
LITERATURE VALUES FOR ANALYSIS OF A0UE0US
LEACHATE FROM BYPRODUCT GYPSUMS
(mg/1 except as noted)
Avg.'
Avg.a
Underdrain3
n
rono
Pond*
Analysis
Natural
Byproduct
Leachate
Runoff
Runoff
Total Dissolved Solids
2248
2413
4260
2400
2287
Fluoride
0.2
3.4
-
-
-
Chloride
2.7
17.7
1450
1.0
47.0
Sulfate
1456
1495
1150
1500
1450
Nitrate Nitrogen
4.9
1.3
1050
-
-
pH
S.4S
7.41
6.6
7.0
7.1
Alkalinity
<1.07
<9.7
84
470
25
A1 mi nun
0.27
0.32
4.5
-
-
Boron
0.32
1.27
57
1.4
5.0
Cadmiun
<0.005
<0.005
0.025
-
-
Bartun
0.012
0.012
3.4
-
-
Caldua
635.33
664.33
580
1300
410
Chromi L*n
<0.01
<0.01
<0.03
-
-
Cobalt
<0.01
<0.01
<0.1
-
-
Copper
<0.01
<0.01
<0.1
-
-
Iron
<0.05
<0.12
<0.1
-
-
Lead
<0.06
<0.05
0.15
0.0020
0.4400
Hagnesius
0.70
7.07
372
16.0
27.0
Manganese
<0.01
0.31
1.20
-
-
Nolybdenia
<0.06
<0.04
<1.0
-
-
Nickel
<0.01
<0.01
<0.1
-
-
Phosphorus
<0.10
<0.15
-
-
-
Potassiua
1.19
<0.19
-
-
-
Seleniun
<0.10
<0.10
0.020
0.0340
0.0070
Silver
<0.01
<0.01
<0.05
-
-
Sodiua
11.17
3.76
38
2.5
2.0
Silicon (Si02)
-
-
142
-
-
Zinc
<0.01
<0.01
0.07
•
-
Arsenic
<0.005
<0.005
<0.01
0.1560
0.0800
Mercury
<0.005
<0.005
0.006
0.0021
0.0070
Notes:
1. Analyzed for this study. Iowa, Nova Scotia. Oklahoma
2. Analyzed for this study. Plant A. Plant B. Plant C.
3. Shawnee FGD Waste Disposal Site H in Reference (5).
4. Shawnee FGD Waste Disposal Site J in Reference (5).
5. Shawnee FGD Waste Disposal Site K in Reference (5).
8B-108
-------
Table 6
LITERTATURE VALUES FOR SOLUBLE SALTS IN SEVERAL
NATURAL AND SYNTHETIC GYPSUMS (2)
(ppn except as noted)
Range
Range
Natural
Synthetic
K
21 - 47
22 - 1532
Na
7 - 195
18 - 2134
"9
12 - 26
12 - 212
CI
15 - 111
21 - 412
KC1
31.5 - 89.6
42.0 - 509.1
NaCl
0 - 112.7
0 - 233.9
MgCij
0
0 - 37.7
Cad,
0
0
0 - 9.9
0 - 3082.0
tajSO,.
21.6 - 465.4
0 - 6586.3
MgSO.
59.4 - 128.7
59.4 - 1002.0
Ca
0
0
SO,
62.0 - 417.5
197.7 - 4754.5
Sub
796.4
354.7 - 7009.5
Equiv f/Ton
1.59
0.7 - 14.0
Table 9
COtWERCIAL-GRADE GYPSUH SPECIFICATIONS (13)
National
Georgia-
U.S.
Gypsus
Pad f i c
Gypsus
West
Parameter
Co.
Corp.
Co.
Japan
Germany
Gypsu* content. nin X
94
90 .
95
95
80-95
Calcium sulfite content, max X
0.5
-
2.0
0.25
0.25
Total soluble salts, max ppo
-
-
600
1000
-
Sodiua content, max ppo
250
200
75
-
GOO
Chloride content, max ppn
400
200
120
-
100
Hagnesiua content, max ppn
250
-
50
-
1000
Free water, max X
1
10
10
10
10
pH
6-8
3-9
6.5 - 8
6.5 - 8
5-9
Inerts. max X
3.0
-
1
-
-
Source: Ellison. William, and Edoard Hasner. "FED Gypsus Use Penetrates U.S.
Wall board Industry.' Power. February 1988.
8B-109
-------
8B-110
-------
EVALUATION OF DISPOSAL METHODS
FOR OXIDIZED FGD SLUDGE
For Presentation at the
EPRI-EPA-DOE 1991 SO2 Control Symposiums
December 3-6.1991
Washington. D.C.
by
W. C. Yu
Manager, Process Engineering
Conversion Systems. Inc.
Horsham, PA
Preceding page blank
8B-111
-------
8B-1I2
-------
Introduction
The implementation of wet flue gas desulfurization — in
response to the Clean Air Act of 1990 — will cause many power
generators and state regulatory personnel to face important
decisions on the disposal of large volumes of resultant solid
waste. Even with the selection of forced oxidation technology,
it is widely recognized that the vast majority of flue gas desul-
furization by-products will be disposed.
This paper analyzes the water quality issues associated
with gypsum stacking, macroencapsulation of gypsum, and
the stabilization/fixation of gypsum. Water quality issues include
leachate quality, leachate generation, runoff management, and
groundwater impact. The following analysis uses both field and
literature data to measure the environmental impact of the
three most discussed disposal options.
Leachate Management
Comparison of Leachate Quality
Leachate quality for each disposal option was determined by
reviewing pertinent literature for reported chemical compositions
of FGD waste and tneir corresponding leachates. The analy-
sis focused on chemical analyses of forced oxidized gypsum
sludge produced by wet limestone scrubbing of medium to
high sulfur, bituminous coal. The results reported in this anal-
ysis are based on documented forced-oxidized FGD wastes
from six generating stations.
The chemical composition of a leachate produced from a
waste is governed by the chemical composition of the waste
itself. FGD forced oxidized gypsum wastes produced by wet-
limestone scrubbing consist primarily of flue gas reaction prod-
ucts. excess unreacted limestone, fly ash. and scrubber liquor.
Limestone contains many inert materials that do not react
during the scrubbing process but become incorporated into
the solid phase of the FGD waste. Magnesium (Mg2-) is a
common constituent of limestone, often combining in the scrub-
ber with sulfate (S042 ) to form magnesium sulfate (MgS04).
MgSOa is soluble in water and may leach from the FGD waste
when landfilled. Excess, unreacted limestone also becomes
incorporated in the FGD waste, thereby increasing CaC03
concentration of the sludge. Dissolution of CaC03 from the
landfilled FGD waste will result in a leachate high in Ca2- with
a moderately to high pH.
The coal type adds to the chemical character of the FGD
waste as a source of both volatile compounds and fly ash.
Volatile elements such as chlorine, lead, bromine, fluoride, and
selenium are scrubbed from the flue gas. along wit.: the sulfur,
and incorporated in the FGD waste. High sulfur coals require
greater quantities of limestone in the scrubber, and therefore
produce wastes and tend to produce leachates with greater
concentrations of sulfur and limestone compounds. In addi-
tion. coal containing unusually high amounts of chloride pro-
duces leachates with high chloride concentrations.
Fly ash is the major source of trace elements in FGD
waste. These elements often become incorporated in the FGD
leachate. Although particle control devices are commonly
installed prior to the FGD scrubbers, small amounts of fly ash
are captured by the scrubber, and trace metals added to the
FGD waste.
The composition of scrubber sludge is extremely variable,
changing from station to station and hour to hour within each
station (Baker 1978). In general, the unoxidized solid sludge of
the FGD slurry is characterized primarily of calcium sulfate
dihydrate (CaS042H20; gypsum), calcium sulfite hemihydrate
(CaS03l/2H20). calcium carbonate (CaC03>, and inert con-
stituents (Summers et al. 1983). With oxidation, approximate-
ly all of the calcium sulfite hemihydrate is ox^;zed to calcium
sulfate dihydrate.
Like the scrubber sludge, the chemical composition of the
scrubber liquor is extremely variable. In general, the scrubber
liquor is high in total dissolved solids, sulfate, chloride, calcium,
magnesium, and sodium. Numerous trace elements, such as
arsenic, boron, chromium, copper, iron. lead, mercury, nickel,
selenium, and zinc occur at varying concentrations in scrubber
liquors (Baker 1983). Typical pH values of forced-oxidized
scrubber liquors produced from wet-limestone scrubbing of
eastern coals range from 5 to 7.
Leachate is the result of either rain water or groundwater
coming in contact with a solid waste material, and incorporat-
ing the soluble components of the waste material into the water.
In the case of forced-oxidized FGD waste, the solid waste
material is mainly calcium sulfate dihydrate (CaSO^HjO).
gypsum. Rain water in contact with gypsum will dissolve the
gypsum until saturation is reached.
At 25C. the concentrations of calcium and sulfate in water
saturated with rsspect to gypsum would be approximately 400
mg/L and 960 mg/L respectively. Total dissolved solids (TDS)
is the sum of the dissolved constituents in water. Assuming
that calcium and sulfate were the only two constituents dis-
solved in the water as it infiltrates through the gypsum, the
resulting TDS would be approximately 1.360 mg/L However,
forced-oxidized gypsum waste also contains elements asso-
ciated with limestone and the fly ash: therefore. TDS of forced-
oxidized gypsum waste leachates are typically higher than
1.360 mg/L.
In this analysis, the three disposal or sludge management
practices evaluated have water budgets that are unique to that
particular practice. This is important because leachate and
runoff constituents would be expected to vary as a function of
the amount of waler contacting the sludge, the duration of the
contact, and the initial quality of the water. Each disposal prac-
tice is evaluated below and the quality and quantity of leachate
and runoff volumes are projected.
Preceding page blank
8B-113
-------
Disposal Option 1:
Wet-Impoundment/Gypsum Stacking
In this cits00sal option, the impounded scrubber slurry water
provides a continual source of recharge for leachate genera-
tion. as opposed to the other disposal options in which leachate
generation is dependent upon rainfall. The pond water is a
combination of supernatant liquor from slurry and rain water.
Usually, the supernatant liquor is decanted from the pond and
recirculated as process water to the FGD scrubber system. As
a result of the recirculating process, the supernatant liquor
builds up high concentrations of ions, particularly chloride ard
sulfate. In turn, the pond water acquires the high ion concen-
trations of the supernatant liquor.
A summary of literature-reported chemical analysis of
forced oxidized gypsum sludge pond water is listed m Table 1.
The water is high in total dissolved solids (TDS). ranging from
6.000 mg-L to over 10.000 mg/L. The TDS consists mainly of
sulfate (1.400 - 3.050 mg/L). chloride (890 - 6.600 mg/L). cal-
cium (550 - 1.300 mg/L). and magnesium (540 - 1.100 mg/L).
Numerous metals such as aluminum, arsenic, beryllium, boron,
cadmium, chromium, cobalt, copper, iron. lead, mercury, nick-
el. selenium, silver, and zinc often occur at trace concentra-
tions. Typical pH values of the pond water range between 5
and 7.4.
Published leaching studies of gypsum indicate that ion
concentrations decrease in the leachate with increased pore vol-
ume displacement. The initial high leachate quality is similar to
the pond water quality, and represents the release of intersti-
tial pond water stored within the pores of the gypsum With
increased pore volume displacement, constituent concentrations
in the leachate decrease to levels expected for rain water at sat-
uration with respect to gypsum. Table 2 presents leachate con-
centrations which may be expected after 50 pore volume dis-
placements from gypsum. Concentrations of TDS (2200 mg/L).
sulfate (1200 mg/L) and chloride (100 mg/L) presented in Table
2 are less than the expected pond water concentrations pre-
sented in Table i. The time required for 50 pore volumes to
leach from i cubic meter of gypsum is approximately 50 to
100 years. Therefore, leachate concentrations similar to the
pond water should be expected from this disposal option tor a
long time following its closure.
Disposal Option 2:
Macroencapsulation of Gypsum RItercake
In this option, dewatered gypsum filtercake is placed in clay
lined cells and capped with clay. The results of ASTM and EP
Toxicity leaching tests of dewatered gypsum filtercake are list-
ed in Table 3. Leachate compositions produced from the gyp-
sum filtercake are similar to the leachates produced from the
wet-impoundment/gypsum stacking option. The leachates are
high in TDS (2.000 mg/L to 10.500 mg/L). with the TDS con-
sisting mainly of calcium (1.420 - 2.220 mg/L). chloride (3330
- 3930 mg L). sulfate (788 - 1800 mg/L) and magnesium (465
- 740 rng/L). Reported pH values of gypsum filtercake leaerfiate
vary from 5.0 To 6.6. Although the resute in Table 3 ar report-
ed for gypsum containing no fly ash, the metals aluminum,
arsenic, banum. beryllium, cadmium, chromium, copper. non.,
lead, manganese, mercury, nickel, selenium, ana silicon can be'
expected to leach from the gypsum filtercake at trace con-
centrations. The occurrence of tnese metals may indicate that
some fly asn escapes the particulate control devices, that the
metals are inherent at trace concentrations in the limestone, or
that the metals are a combination of both fly ash and limestone
sources.
The analysis presented in Table 4 are resutis of leacfiaxe coi
lected from a macroencapsulated test cell. The TDS concen-
trations presented in Table 4- are significantly higher than the
results reported from standard leaching tests (Table 3) and
range from 2. OOO mg/L to 46.000 mg/L The dominant ion con-
tributing to the high TDS is chloride with concentrations ranging
from 272 to 38.145 mg/L [Table 4). As mentioned earlier, the
facility shows the impact of a tight water loop and a coal high in
chlonde content, thereby producing a leachate high in chloride.
Disposal Option 3:
Disposal of Stabilized/Fixated Gypsum
Table 5 presents a compilation of literature-reported values ol
leachates produced from standard leaching tests, as wel as
from leachate collecteo from a field disposal site of stabi-
lized/fixaied gypsum. In all cases, the gypsum was treated with
the addition of fly ash and small quantities ol lime to produced
a stabilized/fixated material using CSI'sPoz-O-Tec^ technol"
ogy. Leachate concentrations produced from the stabilized/fix-
ated material are significantly lower than the leachates pro-
duced from dewatered gypsum. TDS values of leachates Irom
stabilized/fixated gypsum range from 177 to 2372 mg/L. an
order of magnitude lower than the leachates produced in the
previous three options (2.000 to 10.000 mg/L). Expected chlo-
ride concentrations (40 to 80 mg/L) are one to three orders ol
magnitude lower in the leachate from the stabilized/fixated
gypsum, as compared to the dewatered gypsum. Sulfate ana
calcium concentrations are somewtiat lower in leachates from
the stabilized/fixated material, as well. In addition, leacnate
produced from stabilized/fixated gypsum has a higher pH than
leachate from dewatered gypsum, ranging from 6.6 to 10.5.
Similar to leachate from dewatered gypsum, the metals arsenic,
boron, cadmium, chromium, lead, mercury, nickel, selenium,
and silver are still expected to leach from the stabilized/fixat-
ed gypsum at trace concentrations.
Predicting Leachate Quality
Although laboratory leaching tests are more commonly used to
determine leachate compositions, the test results often do not
adequately characterize the leachate produced from coal com-
bustion by-products. The reason for the discrepancy is lhati
8B-114
-------
the standardized leaching tests (EP. TCLP. and ASTM) are
designed to mimic the biological decomposition conditions typ-
ical of disposed hazardous wastes or municipal wastes in a
sanitary landfill. FGD waste, however, is composed of inor-
ganic constituents which, for the most part, do not undergo
active biological decomposition.
In an attempt to predict the leachate quality from oxdized-
gypsum. the Electric Power Research Institute (EPRI)-devel-
oped computer model Fossil Fuel Combustion Waste Leaching
Code (FOWL) was utilized (Hostoetler et al. 1988). The FOWL
model is the result of numerous EPRI studies conducted for
the purpose of predicting leachate compositions from fossil
fuel comtx-stion by-products.
The model is designed to calculate quantities, aqueous
concentrations, and release durations of selected inorganic
constituents released from fossil fuel combustion by-products.
Given an initial waste composition, rainwater infiltration rate, and
physical charactenstics of the waste form, the FOWL model
estimates leachate compositions as a function of time by com-
bining a geochemical model with a water balance model.
Constituents considered by the model indude Al. As. B. Ba.
Ca. Cd. Cr. Mo. S. Si. Sr. Cu. Fe. Mg. Na. Ni. Se. and Zn. The
model assumes that the concentrations of each of these con-
stituents are controlled by the solubilities of particular solids.
Constituent concentrations are determined for deionized water
at equilibrium with respect to the solids phases at various val-
ues of pH.
A major limitation of the model is that the pH of the waste
is assumed to be constant. Although this may be the case for
the solid waste material, the pH of the leachate will most prob-
ably vary: this is not considered in the model. The model also
assumes that the rainwater infiltrating through the waste is
similar in composition to deionized water. In many regions of
the U.S.. particularly in the eastern U.S.. rainwater is more
acidic than deionized water. Some of the leachate constituents
are modelled based on empirical relationships as opposed to
mechanistic thermodynamic data. The model also does not
consider redox reactions which govern the teachability of redox-
sensnrve constituents such as iron and sulfur. Finally, the model
does not consider some elements, such as chloride, which are
significant in FGD wastes.
The input parameters required for the model indude phys-
ical characteristics of the waste material including its landfill
geometry, bulk density, and initial and saturated moisture con-
tents. Chemical characteristics of the waste material input to the
model include its solid composition, leach able fractions of indi-
vidual constituents, and its pH. Net infiltration through the wast?
matenal is also input to the model.
For this exercise, two disposal scenarios were cent-
ered: First, the disposal of gypsum filtercake alone: and sec-
ond. the disposal of stabilized/fixated gypsum filtercake. For
the gypsum filtercake alone the solid material was assumed
to contain 90 percent CaS0i2*H2O and 10 percent other con-
stituents. The pH of the gypsum filtercake was assumed to be
5.5. the mean literature- reported value (Table 3). The solid
matenal of the stabilized/fixated gypsum filtercake was assumed
to have a 0.5 to 1 ratio of fly ash to filtercake on a dry weight
basis.
The teachable fractions of each constituent listed in Table
"I correspond to the fraction of the total amount of that particular
constituent expected to leach from the solid waste material. The
model assumes leachable fractions for each of the elements
based upon laboratory leaching tests performed by Ainsworth
and Rai (1987). Use of Ainsworth and Rai's (1987) data in this
modeling exerdse is limited for two reasons. First, the leachable
fractions calculated by Ainsworth and Rai (1987) were obtained
in nitric add solutions and are considered in the model even
though the model assumes that dionized water is the leaching
solution. Second. Ainsworth and Rai (1987) did not conduct leach-
ing tests of forced oxidized FGD wastes, the subject of this anal-
yses. The leachable fractions considered in the model are for
unoxidized FGD wastes only. Conveniently, the user of FOWL
has the option to use alternative value for the leachable fractions.
In this analysis, only the leachable fraction for calcium was
changed to a value of 1.0. since calaum is reported as a major
ion in leachates of forced oxidized FGD wastes.
The FOWL estimated leachate concentrations of almost all
of the constituents for the gypsum filtercake are significantly
lower than literature-reported values (Table 3). In particular,
caldum (393 mg/L). sulfur (315 mg/L). and TDS (756 mg/L)
produced by FOWL are an order of magnitude lower than those
reported by the literature (Table 3). Even if a chloride con-
centration of 3.600 mg/L were added to the leachate compo-
sition expected by FOWL, the resultant TDS would still be half
of the mean value reported by the literature (Table 3). Leachate
compositions expected for the stabilized/fixated gypsum fil-
tercake are doser. but still lower than the literature-reported
value (Table 5).
The consistent underestimation of leachate constituents by
the FOWL model, as compared to literature sources, is a result
of the model's limitations. By assuming that only deionized
water is in contact with the solid waste material, the model
ignores that most disposed FGD waste contains residual pro-
cess water in its pore spaces. This process water is typically
high in major ions and TDS which dominates the leachate char-
acteristics when released. By ignoring pH changes and redox
processes, the model is not also considering geochemical pro-
cesses which highly influence the concentrations of many con-
stituents. Finally, by ignoring major ions such as cntoride and
potassium, the model will always underestimate the TDS of
the resilient leachate.
Comparison of Leachate Generation Rates
The volume of leachate potentially produced from each dis-
posal option was estimated using the Hydrologic Evaluation
of Landfill Performance (HELP) computer model. The HELP
model, developed by the United States Environmental
Protection Agency (US EPA) and The Army Corp of Engineers
(Schroeder et al.. 1984). uses dimatologic. soil, and design
data to produce volume estimates of runoff, drainage, and
8B-115
-------
leachate expected from various landfill designs. Inherent to
the model is a detailed data base of dimatotogcal data lor 102
cities in 45 states, enabling the estimation of leachate and
runoff volumes from landfill designs located in each of the eight
states of interest.
As is the case for all models. HELP is based on many
simplified assumptions which can produce results quite differ-
ent from actual values. In particular, the climatological data
inherent to the model consists of daily rainfall data for only five
years, from 1974 to 1978. This data may not be representative
of unusually wet or dry conditions. In addition, the model can
not be used to estimate the large volumes of runoff expected
from high intensity rain storms, typical of five-year, ten-year or
greater storm events.
The HELP model is additionally restricted in that it can not
estimate the volume of leachate and runoff generated from a
pond. For disposal Option 1 - wet-impoundment/gypsum stack-
ing method, leachate generation from beneath the pond was
estimated using the following modified Darcy equation:
Q = KAT *H/L. (1)
In Equation (1). Q is the volume of leachate generated. K
is the hydraulic conductivity of the soil layer underlying the
pond. A is the area of the pond. T is time. H is the head of
standing water in the pond, and L is the thickness of the soil
layer beneath the pond. The pond was assumed to have a
constant head of 15 feet of water, and was underlain by a one
foot layer of low permeability (1 x 10—7 cm/s) day. For a pond
area of 100.000 square feet, a leachate generation rate of
165.610 ft3/yr was estimated. Similarty, a pond area of 50 acres
(2.180.000 ft2) would produce an estimated 3.610.298 frVyr
of leachate.
Input parameters required for the model included clima-
tological data, soil characteristics, and the landfill design.
Climatological data were utilized from the model's internal data
base. This data included daily rainfall amounts, mean month-
ly temperatures, mean monthly solar radiation, leaf area indices,
and winter snow cover. Physical characteristics (permeability,
porosity, field capacity, and wilting point) typical of clay liners
and soil covers are also taker: from the model's internal data
base. The physical parameters for each disposal option includ-
ing permeability, initial water content, porosity, and thickness
of the waste material were input manually.
Initially, all of the model simulations were conducted under
the assumption that each landfill design option was located in
one city: Cleveland. Ohio. This assumption allows for direct
comparison of leachate volumes produced between disposal
options. Two simulations were conducted for each disposal
option, the first assuming the landfill was actively being filled and
contained no cover (Active/Open), and the second assuming
that the landfill was closed with some type of cover
(Closed/Inactive). For the Active/Open simulation, the landfill
thickness was assumed to be half of the total thickness, while
for the Closed/Inactive simulation, the total landfill thickness
was used.
Disposal Option 1:
Wet-Impoundment/Gypsum Stacking
Although the HELP model could not be used to simuls^
leachate generation from the pond, it was used to estima™
leachate and runoff generation rates from the gypsum stacks.
For this disposal option, it was assumed that the gypsum stacks
were underlain by a one foot sand drain and a one foot layer
of compacted clay.
Table 6 presents expected leachate generation rates
assuming the gypsum stacks are underlain by both a sand
drain and compacted clay while the pond is underlain by com-
pacted day only. From 2 to 30 years, leachate expected from
this disposal scenario increases from 3.600.000 ft3/yr to
4,600.000 ft3/yr. Comparing this result to a case with a liner
and no sand drain, the inclusion of the sand drain cuts the
expected leachate volume in half during the active years of
the disposal facility. Upon closure, leachate generation is
expected at 970.000 ft3/yr. This estimate is one order of mag-
nitude less than the leachate volume estimated lor the dosed
facility lined only with compacted day.
Disposal Option 2:
Macroencapsulation of Gypsum Filtercake
The expected leachate generation rates from the macroencap-
suiation of gypsum filtercake without fly ash are irKluded in Table
7. This option includes a one foot clay liner, a sand drain, and ulti-
mately. a day cap. Leachate generation from this option inae^M
es from 72.000 ft3/yr for the first five years of operation to 420.(^|
ft3/yr by the end of 30 years. Upon closure. 420.000 ft3/yr of
leachate is expected to be generated from the macroen capsula-
tion of gypsum filtercake. Approximately half the amount of leachate
is expeded from macroencapsulation of the gypsum filtercake as
compared to the wet-impoundment/gypsum stacking method.
Disposal Option 3:
Disposal of Stabilized/Fixated Gypsum
This disposal option is expected to produce the lowest vol-
umes of leachate as compared to the other disrosal options.
Leachate generation rates expected to be produced from dis-
posal of stabilized/fixated gypsum ranged from 0 ft3/yr in the first
five years of operation to 61 ft3/yr by the end of thirty years
(Table 8). This disposal option, due to the impermeability of
the fixated material, indudes neither a liner nor a cap. Upon do-
sure. 73 ft3/yr of leachate is expected from the disposal of sta-
bilized/fixated gypsum. The leachate generation rates expect-
ed from this disposal option are three orders of magnitude
lower, less than 0.1%. of the leachate generation expected
from disposal Option 2 (the second lowest leachate produc-
er). The low leachate volumes expected from this option are
directly attributable to the low permeability (1x10-? cm/sec) of.
the stabilized/fixated gypsum.
8B-116
-------
Leachate and Runoff Generation
by Geographic Area
The leachate generation rates discussed above were deter-
mined assuming that a landfill representative of each dispos-
al option was located in one central city: Cleveland. Ohio. To
determine the effect of geographic region on leachate gener-
ation rates, the eight states were grouped into four of the US
EPA designated rainfall zones. New York. Pennsylvania, north-
ern Ohio, northern Indiana, and northern Illinois are located in
rainfall zone i. Southeastern Illinois, southern Indiana, south-
em Ohio. Kentucky, and Tennessee are located in rainfall zone
2. Florida and southwestern Illinois are located in rainfall zone
3 and 4. respectively. HELP model simulations were conduct-
ed for one city considered to be centrally located within each
rainfall zone. The cities chosen for zones 1 through A includ-
ed Cleveland. Ohio: Lexington. Kentucky: Orlando. Florida:
and East St. Louis. Illinois, respectively.
Simulations were conducted assuming that landfill Option
2 - disposal of gypsum fittercake (alone) in an unlmed landfill
- was constructed in each central city. For best comparison of
expected leachates generated in each city, all landfills were
assumed to occupy a 100.000 square foot area. In addition,
model simulations were conducted for both an active and closed
landfill.
The expected leachate volumes did not vary between
cities, but were all approximately 62.000 ft3. These results sug-
gest that leachate generation rates should be similar for land-
fills located anywhere within the US EPA designated rainfall
zones i through 4. given similar conditions as assumed m this
study. Although rainfall amounts are expected to vary from
region to region, so do other climatic factors such as evapo-
ration. temperature, solar radiation, and snow cover — all fac-
tors which affect infiltration. The net effect of all these factors
results in similar leachate generation rates for each rainfall
zone.
Assessment of Relative Potential for Groundwater
Impact
The potential for groundwater impact is determined by both the
quality and quantity of the leachate expected from each dis-
posal option. Typically, the potential for leachate to impact
groundwater is determined by comparing the chemical charac-
ter of the leachate to drinking water standards. Therefore, the US
EPA regulated drinking water standards are included on the
tables listing the leachate chemical analysis. However, even if
the leacnate quality exceeds the drinking water standards, it
may have no measurable impact on the groundwater if ihe
amount of leachate generated is low. Work is currently under-
way to integrate the above findings with the US EPA's verti-
cal/horizontal spread model (VHS) to measure groundwater
impact.
Disposal Option 1:
Wet-Impoundment/Gypsum Stacking
As illustrated in Table 1. the leachate expected from disposal
of gypsum by wet-impoundmem/gypsum stacking may exceed
the maximum contaminant levels (MCLs) for fluoride, arsenic,
cadmium, chromium, lead, and selenium. In addition, the sec-
ondary maximum contaminant levels (SMCLs) for chloride,
sulfate. TDS and pH may be exceeded. Assuming that long-
term leaching of the gypsum waste produces leachates simi-
lar in chemical character to that illustrated in Table 2. SMCLs
for sulfate. TDS. and pH may still be exceeded. Combining the
poor quality of leachate with the large volumes expected sug-
gests that the potential for groundwater to be impacted from this
disposal option is quite high.
Disposal Option 2:
Macroencapsulation of Gypsum Filtercake
As mentioned previously, leachate quality for disposal of gyp-
sum filtercake with and without fly ash is expected to be sim-
ilar (Table 3). These leachates may exceed the MCLs for
arsenic, cadmium, chromium, lead, and selenium. In addition,
the SMCLs for chloride, sulfate, manganese. TDS. and pH
may be exceeded. Long-term leaching of the gypsum filter-
cake can produce leachates that no longer exceed the MCLs.
but do exceed the SMCLs for sulfate and TDS. The volume of
leachate expected to be generated from this option is less than
for Option i: however, it still has the potential to impact ground-
water. The potential for groundwater to be impacted from this
disposal option is considered to be moderate.
Disposal Option 3:
Disposal of Stabilized/Fixated Gypsum
The leachate quality for the disposal of stabilized/fixated gyp-
sum may exceed the MCLs for cadmium, chromium, lead, and
selenium, and silver (Table 7). In addition, the SMCLs for sul-
fate. TDS. and pH may be exceeded. However, the leachate vol-
umes estimated from this disposal option are extremely low. With
such low volumes of leachate released, the leachate constituents
have the potential to become diluted to background levels upon
mixing with the groundwater. The potential for groundwater impact
from this disposal option is considered to be low.
Runoff Management
Comparison of Runoff Quality
Runoff from each disposal option is a combination of rain water
and the soluble components of the gypsum. The concentra-
tion of constituents in the runoff is dependent upon the con-
tact time of the runoff water with the gypsum. Typically, the
8B-117
I
-------
contact time of the runoff with the gypsum is quite low; there-
fore. ion concentrations in the runoff water are expected to be
less than ion concentrations reported in leachates. Table 4
lists leachate and runoff concentrations from disposed gyp-
sum filtercake. Runoff concentrations are significantly less than
the leachate concentrations. TDS of the runoff was 10.000
mg/L as compared to 37,000 mg/L in the leachate.
Although there is a lack of runoff water quality reported in
the literature, the following generalizations can be made. Since
leachate compositions for disposal Options 1 and 2, without
co-disposal of fly ash. are similar, then runoff produced from
these options are expected to be similar. The runoff is expect-
ed to have less metals, less total dissolved solids, and only
comparable levels of the more soluble constituents such as
sulfate and chloride as compared to their leachate results.
Given the low ion concentrations expected in leachates from
the stabilized/fixated gypsum, the runoff concentrations are
also expected to be even lower and the lowest of all disposal
options.
Comparison of Controlled Discharges
In this report, controlled discharges are considered to be the
combined discharges of runoff and discharges from a sand
drain if present beneath the landfilled gypsum. Addition of both
runoff and collection from the sand drains provides a total esti-
mate of the volume of water requiring management and treat-
ment for each disposal option. Discharge estimates from runoff
and from the sand drain for each disposal option were esti-
mated using the HELP model. In general, disposal practices pro-
ducing the largest volumes of uncontrolled discharges (leachate)
produce the lowest volumes of controlled discharges. Upon
closure of a facility, runoff is expected to contact the cover
material and not the gypsum: therefore, treatment of runoff
from a closed/inactive facility is not expected to be needed.
As a result, controlled discharges estimated for each dispos-
al option upon closure considers discharges estimated from
the sand drain only and not runoff.
Disposal Option 1:
Wei-impoundment/Gypsum Stacking
The wet-impoundment/gypsum stacking option considering
both the pond and the gypsum stacks that are lined with com-
pacted day but without a sand drain produce the lowest con-
trolled discharges. No runoff is anticipated to leave the pond.
Therefore, during this facility's first year of operation, zero con-
trolled discharges are estimated. Runoff, however, is expect-
ed from the gypsum stacks and is estimated at 83,000 ftVyr for
the second year of operation and at 420,000 ft3/yr at the end
of thirty years. Beyond thirty years, runoff generated from the
closed gypsum stacks will not require treatment, and there-
fore. controlled discharges requiring treatment at that time are
considered to be zero.
Upon addition of the sand drain, the estimated controlled
discharges triples for this disposal option and ranges from
260.000 ffVyr after the first two years of operation to 4.300.000
ft3/yr after 30 years. Beyond thirty years, controlled discharges
from the sand drain are estimated at 3.700.000 fFVyr Without
the sand drain, the 3.700,000 fPVyr would have been released
as uncontrolled leachate to the groundwater.
Disposal Option 2:
Macroencapsulation of Gypsum Filtercake
Macroencapsulation of gypsum filtercake produces moderate
amounts of controlled runoff as compared to the other disposal
options. Controlled discharges are estimated at 1.000.000 ft3/yr
during the thirty active years of the facility. Beyond thirty years,
approximately 76,000 ft3/yr is anticipated as discharge from
the sand drajn.
Disposal Option 3:
Disposal of Stabilized/Fixated Gypsum
Controlled discharges estimated from the disposal of stabi-
lized/fixated gypsum are similar to the controlled discharges
estimated for macroencapsulation of gypsum filtercake
(1.100.000 ft3/yr); however, field data to date shows that this
controlled discharge can be reused or discharged without treat-
ment. Upon closure of a stabilized/fixatated FOS landfill, zero
controlled discharges requiring treatment are anticipated. These
results combined with the low anticipated leachate discharges
with low constituent concentrations shows that this disposal
option is dearly superior in terms of lowest potential for ground-
water and surface water impacts.
Conclusions
The above data dearly set forth the potential impact to ground-
water quality by each of the most widely considered gypsum dis-
posal options. Physical properties of materials and disposal
economics for each option are important considerations and wil
be presented in subsequent papers.
Given the above modeling results, only the stabilization/fix-
ation option can provide assurance of little or no groundwater
degradation, even on sites with little available existing ground-
water dilution for leachate. Similarfy, stabilization/fixation offers
wider operating choices to the management of surface runoff
— an increasingly complex problem in today's power plants.
Finally, it is believed that the superior results in all the
areas of water quality management studied for stabilization/fix-
ation demonstrate that this technology can meet the tightening
groundwater protection requirements that are integral to gen-
erator compliance.
8B-118
-------
Table 1
Literature Reported Values of
Forced Oxidized FGD Pond Water (a)
Drinking
Constituent/
Range
Mean
Water
Parameter
(tng/L)
(mgrt-)
Standards
Maior Cations
Calcium
550-1300
946
Magnesium
540-1100
775
Sodium
62-116
89
Potassium
5.9-43
27
Maior Anions
Chloride
890-6600
2500
250
Sulfate
1438 - 3050
2500
250
Fluoride
2.4 - 6.5
4
2-4
Trace Constituents
Aluminum
02 - 0.6
0.4
Antimony
1.4
1.4
Arsenic
<0.003 - 0.09
0.04
0.05
Barium
1
Beryllium
0.004 - 0.05
0.025
Boron
95-140
115
Cadmium
0.003 - 0.009
0.004
0.01
Chromium (total)
0.09 - 0.51
0.23
0.05
Cobalt
0.1
0.1
Copper
0.01 - 0.4
0.15
1
Iron
0.02
0.02
0.3
Lead
<0.01 - 0.67
0.27
0.05
Manganese
0.05
Mercury
<0.0002 - 0.002
0.001
0.002
Molybdenum
Nickel
0.33 - 0.5
0.42
Selenium
0.035 - 0.14
0.09
0.01
Silicon
Silver
0.005
0.005
0.05
Tin
Vanadium
Zinc
0.02 - 27
6.4
Parameters
TDS
6694-10756
9012
500
PH
5.8 - 7.4
6.57
6.5 - 8.5
(a) Sources: Aerospace Corporation. 1977. Disposal of By-Products from Nonregenerable Rue Gas Desulfurization
Systems Second Progress Report. Trepared for Industrial Environmental Research Lab.. Research Triangle Park.
N. C. PB-271 728.
USEPA. 1980. Disposal of Flue Gas Desulfurization Wastes Shawnee Field Evaluation — EPA 625/2-80-028.
Morasky. et al. 1981. Evaluation of Gypsum Waste Disposal By Stacking. In Proceedings: Symposium on
Flue Gas Desulfurization - Houston. October 1980; Volume 2.
8B-119
-------
Table 2
Literature Reported Values of Leachale from Forced
Oxidized FQD Waste (a) alter 50 Pore Volumes
Drinking
Constituent/ Concentration Water
Parameter (mg/L) Standards
Malor Cations
Calcium 100
Magnesium
Sodium
Potassium
Major Anions
Chloride 100 250
Sulfate 1200 250
Fluoride 0 08 2 • 4
Trace Constituents
Aluminum
Antimony
Arsenic 0.005 0.05
Barium 1
Beryllium <0.005
Boron
Cadmium 0.002 0.01
Chromium (total) <0.003 0.05
Cobalt
Copper 1
Iron 0.3
Lead 0.01 0.05
Manganese 0.05
Mercury <0.0005 0.002
Molybdenum
Nickel
Selenium 0.006 0.01
Silicon
Silver 005
Tin
Vanadium
Zinc 0.04
Parameters
TDS 2200 500
pH 5 5 6.5 ¦ 8 5
(a) Source: USEPA (1979) Disposal ol By-Products Irom Non-
Regenarable Flue Gas Desulfurization Systems: Final Report, EPA
600/7-79046.
Table 3
Literature Reported Values of Leachale From
Forced Oxidized FQD Flltercake (a)
Drinking
Constituent/ Range Mean Water
Parameter (mg/L) (mg/L) Standards
Major Cations
Calcium 1420 • 2220 1788
Magnesium 440 - 740 528
Sodium
Potassium
Malor Anions
Chloride 3330 • 3930 3630 250
Sulfate 788-1800 1136 250
Fluoride 2 - 4
Aluminum
1
1
Antimony
Arsenic
<0.01 • 0.02
002
0.05
Barium
0.05 - <0 5
0 15
1
Beryllium
0.05
0.05
Boron
Cadmium
<0.01 -0 02
001
0.01
Chromium
<0 05 - 0.06
0.05
0.05
Cobalt
Copper
007
007
1
Iron
0.07 - 0 2
001
0.3
Lead
0.12 - 1.43
0.45
0.05
Manganese
0.12
0.12
005
Mercury
0.0008 - 0 001
0.0009
0.002
Molybdenum
Nickel
006
0.06
Selenium
0.02 - 0.065
004
001
Silicon
<0 05 • 0 05
005
Silver
005
Tin
Vanadium
Zinc
Parameters
TDS
2138- 10490
6000
500
PH
5-66
54
6.5 8.5
(a) Source: USEPA (1979) Disposal ol By-Products from Non-Regenarable Flue Qas Desulfurization
Systems: Final Report, EPA 600/7-79 046.
-------
Table 4
Chemical Composition of Leachate and Runoff
from Forced Oxidized FGD Filtercake
Drinking
Constituent/
Leachate
Runoff
Water
Parameter
Range
Mean
Range
Mean
Standard:
(mg/L)
(mg/L)
(mg/L)
(mg/L)
Maior Anions
Chloride
272 - 38145
21770
560-19203
6138
250
Sulfate
390-1018
678
544 -1590
1111
250
Fluoride
0.98
0.98
0.59 -1-2
1
2-4
Trace Constituent
Aluminum
Antimony
Arsenic
<0.01 - <0.05
<0.05
<0.01 - 0.53
0.26
0.05
Barium
1
Beryllium
Boron
Cadmium
<0.01 - 0.07
0.01
<0.1 - 0.3
0.175
0.01
Chromium
<0.05-0.10
0.05
<0.05 - 0.05
0.03
0.05
Cobalt
Copper
<0.05 - 0.5
0.07
<0.05 - <0.5
0.14
1
Iron
<0.1 - 5.2
0.88
<0.1 - 0.3
0.1
0.3
Lead
<0.05 - 1.2
0.09
<0.05 - 0.40
0.21
0.05
Manganese
<0.05 - 0.87
0.43
<0.05 - 0.37
0.2
0.05
Mercury
<0.001 - <0.002
<0.002
<0.001 - <0.002
<0.001
0.002
Molybdenum
Nickel
<0.05 - 0.6
0.09
<0.05 - 0.26
0.14
Selenium
<0.01 - <0.05
<0.05
<0.01 - <0.02
<0.01
0.01
Silicon
Silver
0.05
Tin
Vanadium
Zinc
<0.05-0.12
0.05
<0.05 - <0.1
<0.1
Parameters
TDS
2096-46588
36811
2716-18700
10708
500
PH
6.0 - 7.4
6.4 (d)
6.4 - 8.1
6.7
6.5 - 8.5
8B-121
-------
Table 5
Literature Reported Values of Leachate From
Stabilized/Fixated Forced Oxidized FGD Waste (a)
Constriuenv
Parameter
Maior Cations
Calbum
Magnesium
Sodium
Potassium
Range
(mg/L)
67-647
Mean
(mg/L)
520
Drinking
Water
Standards
Maior Anions
Chloride
Sulfate
Fluonde
40-80
85-1356
60
886
250
250
2-4
Trace Constituents
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Molybdenum
Nickel
Selenium
Silicon
Silver
Tin
Vanadium
Zinc
0.005 - 0.032
<0.5 - <1
0.25-1.6
<0.01 -0.10
<0.05 - 0.20
<0.1
<0.1
<0.05 - 0.56
0.0008 - 0.0052
0.05
0.002 - 0.043
<0.05
<0.01 - 0.05
<0.1
0.014
<1
0.925
0.04
0.09
<0.1
<0.1
0.15
0.0014
0.05
0.018
<0.05
0.03
<0.1
0.05
1
0.01
0.05
1
0.3
0.05
0.05
0.002
0.01
0.05
Parameters
TDS
pH
177 - 2372
6.6-10.5
1865
7.6
500
6.5 - 8.5
(a) Source: Conversion Systems. Inc.
Source: Golden. D.M. 1981. EPRI FGD Sludge Disposal Demonstration and Site Monitoring Projects: In
Proceedings: Symposium on Flue Gas Desulfurization Houston. October 1980: Volume 2. EPA- 0600/9-81-09b.
8B-122
-------
TABLE 6
Option 1 - W«t-poodlng of FGD Sludge and Stacking el Settled Material
Average Annual Discharge Rare
(cube feet per year per
100.000 square feet drposaJ area)
Peak Datfy Discharge Rate
(Cubic feet per day per
i00.000 square feet disposal area)
Controlled
Uncontrolled
Controlled
Uncontrolled
Pona area bned witn 1 foot of compacted day
Leacnate
Lined FGO Stack Area • OpervActive
Runoff
Drainage from drain Layer
Leacnate
Lined FGO Stack Area - CiosedMnactive
Runoff (Does not need to be treated)
Drainage from drain layer
Leacftate
33.647
69.556
49357
165.610
12.999
12.999
5.640
797
222
372
Average Annual Discharge Rate
(Cu&c feet per year)
Peak Daily Discharge Rate
(Cub* feet per day)
Controlled
Uncontrolled
Controlled
Uncontrolled
Year
Pond Area
Open Stack Area
Closed Stack Area
(Square feet)
(Square feet)
(Square feet)
1
2.180.000
0
0
0
3.610298
0
9.884
2
2.180.000
248.520
0
256.480
3.642.603
15.998
10.008
S
2.180.MC
994.080
0
1,025.920
3.739.518
63.992
10.380
10
2.180.000
1.242.600
1.242.600
1.895.711
3.933.349
82.748
10366
15
2.180.000
1242.600
2.485*200
2309.021
4.094.875
85.507
11.429
20
2.180.000
1.242.600
3.727.800
3.122.331
4256.400
88266
11.891
25
2/CO.OOO
1.242.600
4.970.400
3.735.641
4.417.926
91.024
12.353
30
2.180.000
1.242.600
6213.000
4.348.951
4.S79.451
93.783
12.815
>30
0
0
7.455.600
3679.860
969.153
16.551
12.658
TABLE 7
Option 2 - Marrn unnpiiiliWnn of dewatered FGD sludge
Average Annual Distiiarge Rate
(cubic feet per year per
100.000 square feet dprtsa' area)
Peak Daily Discharge Rate
(Cube feet per day per
100.000 square feet disposal area)
Case r
Lined FGD drsposaJ area • Qpen/Aarve
Runoff
Drainage from drainage layer
Leaoiaie
Lined FGO disposal area • Closed/Inactive
Runoff (Does not need to be treated)
Drainage from drainage layer
Leacnate
Controlled
143307
12346
Uncontrolled
10864
10367
Average Annual Discharge Rate
(Cubic tern per year)
Controlled
7882
91
Uncontrolled
32
29
Peak Daily Discharge Rate
(Cubic feet per day)
Controlled
Uncontrolled
Controlled
Uncontrolled
Open Area
(Square feet)
Ck»ed Area
(Square feet)
5
667.080
0
1.038343
72.472
53.186
213
10
667.080
667.080
1.050.604
141.628
53253
407
15
667.080
1.334.160
1.062.865
210.784
53320
600
20
667.080
2.001240
1.075.126
279.940
53386
794
2S
667.080
2.668.320
1.087387
349.096
53.453
987
30
667.080
3335.400
1.099.648
418252
53.520
1.181
>30
0
4.002.480
73^66
414.937
400
1.161
8B-123
-------
TABLE 8
Option 3 - Unltoed dtipoiil of stabtltoed/ftxsted FGD sludge
Average Annual Drsctiarge Rate
(cubic feet per year per
100.000 squat® feel dposai area)
Peak Daily Discnarge Rate
(Cubic feet per cay per
100.000 ware feet otsposai area)
Case i
Controlled
Uncontrolled
ConxroUed
Uncontrolled
Unbned FGD disposal area • Open/Active
Runotl
Leacftaie
Uniirted FGD ooposai area • OoseG/inacuve
Runoti {Does not neeo to oe treaied)
leacnate
Open Area
(Square feet)
Ctoseo Area
(Souare feet)
185.123
Average Annual Dtscnarge Rate
(Cubic feet per year)
Controlled
Uncontrolled
9.950
0.1
Peak Daily Discharge Rate
(Cubic teet per day)
Controlled
Uncontroltoo
5
10
15
20
25
30
>30
610.400
610.400
610.400
610.400
610.400
61C.400
0
0
610.400
1.220.800
1.831.200
2.441.600
3.052.000
3.662.400
1.129.991
0
60.735
0
1.129.991
12
60.735
1
1.129.991
24
60.735
1
1.129.991
37
60.735
2
1.129.991
49
60.735
2
1.129.991
61
60.735
3
0
73
0
4
8B-124
-------
COMMERCIAL AGGREGATE PRODUCTIOH
PROM POD WASTE
Charles L. Smith
Conversion Sysfceu, Inc.
200 Welsh Road
Horsham, Pennsylvania 19044
8B-125
-------
8B-126
-------
ABSTRACT
A commercial quality aggregate can be made from flue gas desulfurization
wastes. This aggregate is capable of meeting American Society for
Testing and Materials specifications for construction aggregate — even
the stringent Los Angeles Abrasion Test for toughness. The result of a
12 year development program, this aggregate can be lightweight or normal
weight.
The rationale behind this development is two fold, the first reason is
that FGD is expensive, an added cost to power generation- Any possible
byproduct that can reduce overall costs, including minimizing landfill
space consumed, is a benefit.
A second fact that lent emphasis to the program is that although the
United States consumes over one billion tons of aggregate per year,
aggregate is in scarce supply in many parts of the country. It is
becoming increasingly more expensive as regulatory constraints cause the
opening of new sources to replace exhausted old ones, to be extremely
difficult, if not impossible.
The paper presents test data on aggregate and aggregate end uses
focusing on concrete masonry units. A strong advantage is the ability
to use either forced oxidation or unoxidized FGD waste in aggregate
production.
Preceding page blank
8B-127
-------
INTRODUCTION
Our reason for being here today is elementary; the Clean Air Act
Amendments have initiated another surge of pulverized coal electric
generating facility scrubbing. The logical sequence of events,
comparable to the early wave of scrubbing in the 1970's and early
1980's, is that we clean up the air and produce substantial quantities
of solid waste; this, in turn, must be dealt with in an environmentally
safe manner. Similar to that earlier period, there are applications
where coal cleaning or coal switching are the answers to CAAA
compliance, but for the overwhelming majority of power plants, scrubbing
is the most cost effective answer. Scrubbers come in several general
classes, as we all know. Regenerative processes have been built in
limited numbers, dry scrubbers have been well proven for limited S02
reduction levels, but the bulk of past and future scrubbers appear to be
wet throw-away processes producing either calcium sulfite hemihydrate or
calcium sulfate dihydrate (gypsum). Historically, disposal of throw-
away scrubber sludges was a problem. The solution most commonly used is
fixation by dewatering followed by fly ash and lime interblending.
REUSE
There have been many efforts to make constructive use of the
cementitious character of fixated scrubber sludge, beyond its initial
purpose as an environmentally safe landfill disposal method.
Significant success has been achieved in utilizing this material as a
roadbase medium, with somewhere in excess of a quarter million tons sold
for that purpose. This tends to be an erratic market; road construction
is not a continuous demand.
Focusing on potential usages that are more continuous and uniform in
quantity demanded, the search quickly narrows in on aggregate as a
promising market.
8B-128
-------
AGGREGATE IN THE UNITED STATES
This nation consumes over one billion tons of aggregate annually. In
many areas of the country, aggregate is in scarce supply and must be
transported from significant distances. The difficulty is enhanced as
regulatory constraints tighten on existing sources. Permitting for new
aggregate sources, to replace exhausted quarries, has become extremely
difficult.
Following this rationale, production of an aggregate from the
pozzolanically cementitious scrubber sludge - fly ash - lime composition
was selected as a desirable target. It has taken twelve years work to
bring this concept to the commercial level achieved as of this
symposium.
Simple economics will point out that the cost of extracting a natural
aggregate, as opposed to the cost of manufacturing one, is low.
However, aggregates are transportation sensitive; that is, hauling
costs are large in relationship to the actual value of the aggregate.
Note specifically that aggregate uses are predicated upon volume, not
weight. In concrete, concrete block, and in highway construction, the
aggregate fills a volume; as long as it has sufficient strength and
durability, the weight of that volume is not of importance. If
available at a comparable price, a 60 pound per cubic foot aggregate
(with adequate physical properties) can substitute for a 120 pound per
cubic foot aggregate. The cost of hauling is significantly less.
CHEMISTRY OF THE PROCESS
The Poz-O-Tec process or comparable processes, are in operation at over
35 locations in the U.S., production of lightweight aggregate can be
carried out on wastes from wet scrubbing by lime, limestone, or dual
alkali scrubbers. The scrubbing can be with or without forced
oxidation, as long as the general waste composition falls largely within
the ranges given in Table I.
8B-129
-------
It should be obvious that the demand for a commercial aggregate product
is greater than those for a hardening landfill disposal process; quality
control constraints for the process are more demanding.
LIGHTWEIGHT AGGREGATES
Until the 1970's, when energy costs jumped drastically there were many
lightweight aggregates produced in this country by heat processes.
Hundreds of heat expanded clay, shale, slate, and perlite sources were
spread throughout the county. Because of energy costs, few are left.
There are a very few plants producing no-heat or low-heat,
cementitiously set aggregate.
The majority of lightweight aggregate is utilized in the production of
concrete masonry units, where there are specific advantages to
lightweight aggregate:
Lower transportation costs,
Lower block laying costs,
Superior insulating characteristics,
Superior acoustical characteristics,
Superior fire resistance.
To achieve acceptability as a lightweight aggregate, the potential
product must comply with American Society for Testing and Materials
requirements. The largest market for a lightweight aggregate is in
concrete masonry units, therefore compliance with ASTM C-331 STANDARD
SPECIFICATION FOR LIGHTWEIGHT AGGREGATES FOR CONCRETE MASONRY UNITS is
critical.
Table II provides these requirements along with typical values for the
new FGD waste-derived aggregate. In examination of this Table, it is
8B-130
-------
noted that there are five particle size gradations listed in the
specifications; these are a function of the physical crushing and
screening, and are not relevant to the material itself. Table III
indicates that the vital characteristics, with given specifications, are
met by the new aggregate. Figure I shows the material.
Greater Yield
Typical aggregates for block making are siliceous or calcareous in
nature. When sized for concrete masonry unit production, the new
product has from about 50 to 75% of the bulk density of conventional
types. It is obvious that an increase in blocks produced per ton of
aggregate of up to 75% can be obtained. This advantage has been readily
understood by users. Figures II and III show blocks being produced, and
ready for shipment, respectively.
Labor
If the concrete masonry units produced are substantially lighter, then
they are less taxing for the mason to handle. The result is more blocks
laid per day per man. Or, in the areas in the United States that have
very specific union rules, the difference can reduce the number of men
required for placement.
Improved Insulating Characteristics
Insulation in building materials is a function of density, or entrapped
air. Historically, lightweight concrete masonry units have
substantially superior insulating characteristics, as delineated in
Table VI. Air conditioning, or home heating, efficiency is increased,
resulting in a distinctive economic advantage.
Improved Fire Resistance
Along with improved insulating characteristics, the thermal flow in
these concrete masonry units provides resistance to degradation by fire.
A University of Ohio fire wall test gave a superior, two hour
8B-131
-------
rating to concrete blocks produced from the new lightweight aggregate
(Figure 4).
Acoustical Insulation
Another clear advantage of lower density block is acknowledged to be
lower sound transmission characteristics.
THE NEW AGGREGATE
The new FGD waste-derived lightweight aggregate (Poz-O-Lite) required
many years to fully develop. Final block producer plant trials were
completed in mid 1991. Test data given in Tables IV and V are on
concrete masonry units produced at two commercial facilities, using Poz-
O-Lite. The tests were by independent testing laboratories. Compare
the data with Table III which gives ASTM C-90 SPECIFICATION FOR HOLLOW
LOADBEARING CONCRETE MASONRY ONITS.
Concrete masonry units produced from the new Poz-O-Lite aggregate meet
ASTM C-90.
COMMERCIAL PRODUCTION FACTORS
Those pulverized coal burning power stations whose wet scrubbers produce
an oxidized or unoxidized waste have the raw material for commercial
production. Although all testing has not been completed, the
indications are that nearly every FGD lime or limestone scrubber waste
can be used in the Poz-O-Lite process. Oxidized wastes seem to give a
superior product.
CONCRETE AGGREGATE TESTS
The previous discussion has focused upon the largest market for
lightweight aggregate, in concrete masonry units. Developmental scale
tests on Poz-O-Lite for concrete usage have been promising. There seems
to be full capability for meeting ASTM C332 LIGHTWEIGHT AGGREGATE FOR
STRUCTURAL CONCRETE. Similarly, laboratory tests have indicated the
8B-132
-------
ability to meet even ASTM C33 STANDARD SPECIFICATION FOR CONCRETE
AGGREGATE, including the rigorous Los Angeles Abrasion Resistance
requirement.
STATUS
The commitment in shifting from a landfill disposal to a commercial
product requires caution. The developmental program that generated the
new concept was begun over 12 years ago. Since long term durability is
an essential in building materials, the first concrete block test walls
and test buildings were constructed nearly 10 years ago. Figure 5
shows the first building, constructed in 1982, still standing today with
no degradation of the concrete masonry units utilized in its
construction. A decade of performance testing would seem promising. As
of this writing, full scale commercial production of Poz-O-Lite has
begun.
8B-133
-------
TABLE I
JUUKE OF COHSTITUXXTS: FOD WASTE SOLID PHASE
Calcium Sulfite Heraihydrate
1
to
98%
Calcium Sulfate Dihydrate
1
to
99%
Calcium Carbonate
0
to
30%
Fly Ash
0
to
65%
Magnesium Sulfate Hexahydrate
0
to
4%
Calcium Hydroxide
0
to
3%
Sodium Chloride
0
to
3%
8B-134
-------
TABLE II
POD WASTE - DERIVED AGGREGATE FOR KASORRY
ASXM C-331
SYNTHETIC AGGREGATE 1
• FIVE GRADATIONS
CONTROLLED BY SCREENING j
OPERATION 1
• UNIT WEIGHT (LB.FT3)
Coarse Sizes 55 Maximum
Fine 70 Maximum
Combined Sizes 65 Maximum
54 I
62 [
60 1
• ORGANIC IMPURITIES
"Little or None"
NONE
• STAINING
None
NONE
POPOUTS
None
NONE
CLAY LUMPS AND FRIABLES
2% Maximum
<1%
Drying Shrinkage
0.1% Maximum
0.074%
8B-135
-------
TABLE III
CONCRETE MASONRY UKITS FROM rOD WASTE - DERIVED AGGREGATE
A8TM C-90
CONCRETE MASONRY OMITS TEST DATA
COMPRESSIVE STRENGTH (PSI)
1000 MIN (AVERAGE)
GROSS AREA
800 MIN (INDIVIDUAL)
| UNIT WEIGHT (LBS/FT3)
| LIGHT WEIGHT
< 105
| MEDIUM WEIGHT
105 TO 125
| NORMAL WEIGHT
> 125
ABSORPTION (LB/FT3)
LIGHT WEIGHT
18 MAX
MEDIUM WEIGHT
15 MAX
| NORMAL WEIGHT
13 MAX
I
|| DRYING SHRINKAGE (%)
0.065 MAX
8B-136
-------
TABLE IV
CONCRETE MASONRY UVITS: PRODUCER B
Run #1 Run #2
STRENGTH (psi) 1466 1155
BLOCK WEIGHT (LBS) 34.2 32.5
DENSITY (LB/FT3) 118.5 114.8
ABSORPTION (LB/FT3) 12.65 12.77
DRYING SHRINKAGE (%)
MAX:0.065% 0.0440 0.0597
8B-137
-------
TABLE V
CONCRETE MASOKRT UMITS: PRODUCER C
Run #1 Run #2 Run #3 Run #4 Run #5 Run #6
COMPRESSIVE 1259 1397 1294 1388 1154 1242
STRENGTH
(PSI)
BLOCK 32.2 32.0 32.0 32.1 31.8 31.3
WEIGHT
(LBS)
UNIT
WEIGHT
(LB/FT3) 115.6 114.4 113.4 114.8 111.5 111.4
ABSORPTION
(LB/FT3) 15.7 16.1 16.5 15.7 17.2 17.2
(%) 13.6 14.1 14.6 13.6 15.5 15.5
DRYING
SHRINKAGE
(%) 0.057 0.025 0.033 0.045 0.045 0.039
8B-138
-------
TABLE VI
IVSULATIKG CHARACTERISTICS OF CONCRETE KASOHRT UHITS
RESISTANCE
("F/BTU/HR/FT2)
CONDUCTIVITY
(3TU/HR/FT2/°F)
8 x 12 x 16"
Conventional Aggregate
1.28
0 .78
Lightweight Aggregate
2.63
0.38
8x8x16"
Conventional Aggregate
1.11
0.90
Lightweight Aggregate
2.27
0.44
8B-139
-------
-------
Figure 3. FGD Lightweight: Aggregate Block
Ready For Shipment
Figure 4. Block Fire Wall Testing
At University Of Ohio
SB-141
-------
Figure 5- Initial Building Using FGD
Byproduct Aggregate Stands Without Flaw
After Nearly Ten Years.
8B-142
------- |