]SPA-600/R-93-064d
Api'il 1993
Proceedings: 1991S02 Control Symposium
Volume 4. Session 7
crc Sponsors:
Electric Power Research Institute
B_ Toole O'Neil
3412 Hill view Avenue
Palo Alto. CA 94304
ILS. Department of Energy
Charles J. Drummond
Pittsburgh Energy
Technology Center
P.O. Box 10340
Pittsburgh, PA 15236
ILS. Environmental Protection Agency
Brian K. Gallett
Air and Energy £ki£ineeriiig
Research Laboratory
Research Triangle Part. NC 277H
-------
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
US. 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 US. 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 (SQ2/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
-------
TECHNICAL REPORT DATA
(PIrew read tntirurtiatis ittt llir rct'rrur Itcforc CotnpUt
^ hi poht NO. ?.
i-:PA-G00/H-93-0G4d
3.
a. title ano subtitle
Proceedings: 1991 SO2 Control Symposium. Volume 4.
Session 7
0 ncronT OAT£
April 1993
6. PERFORMING ORGANIZATION COOE
7. AUTHORISl
Miscellaneous
8. PERFORMING OROANIZATION REPORT NO.
TR-101054 (1)
9. performing oroanization name ano aooress
See Block 12
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
IS'A (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 PERlOO 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. Cosponsorea by EPRI and DOE. Vol. 1 is Ooenine Session and Sessions 1-3B,
Vol. 2 is Sessions 4 and 5A, Vol. 3 is. Sessions 5B and 6. and Vol. 5 is Session 8.
16. abstract proceedings document the 1991 SO2 Control Symposium, held December
3-fir 1991, in Wochi'ngtnn, TIT, anH joinf-ly spnndnrpH hy t-Vio Powpr Ppceorrh
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 HI technical papers on development, operation, and commercialization of
wet and dry FGD, clean coal technologies, and combined sulfur oxide/nitrogen oxide
(SOx/NOx) processes. y '
17. KEY WORDS ANO DOCUMENT ANALYSIS
I. DESCRIPTORS
b.tOENTI FIE AS/OPEN ENDED TERMS
c. COSATI Fidd/Groop
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
373
20. SECURITY CLASS (This page J
Unclassified
22. PRICE
EPA Form 2220-1 <9-73) 7« 353
¦
-------
CONTENTS
PREFACE xi
AGENDA xii
VOLUME 1
Opening Session
EPRI Perspective OS-1
EPA Perspective OS-5
DOE Perspective OS-9
Guest Speakers
Shelley Fidler - Assistant, Policy Subcommittee on
Energy and Power, U.S. Congress OS-11
Jack S. Siegel - Deputy Assistant Secretary, Office of Coal
Technology, US. Department of Energy OS-19
Michael Shapiro - Deputy Assistant Administrator, Office
of Air and Radiation, US. Environmental Protection Agency OS-29
Session 1 - Clean Air Act Compliance Issues/Panel 1-1
Session 2 - Clean Air Act Compliance Strategies
Scrubbers: A Popular Phase 1 Compliance Strategy 2-1
Scrub Vs. Trade: Enemies or Allies? 2-21
Evaluating Compliance Options 2-39
Clean Air Technology (CAT) Workstation 2-49
Economic Evaluations of 28 FGD Processes 2-73
Strategies for Meeting Sulfur Abatement Targets in die
UK Electricity Supply Industry 2-93
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 EPRTs 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 Mbdng
Appued to SO2 Control by Dry Sorbent Injection
Studies of the Initial Stage of the High Temperature
CaO-SQ2 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
Rue 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
Acid Rain FGD System Retrofits
Guidelines for rGD Materials Selection and Corrosion Protection
Economic Comparison of Materials of Construction of Wet FGD
Absorbers and Internals
The Intelligence & Economics of FRP in F.G.D. Systems
4A-1
4A-25
4A-41
4A-61
4A-79
4A-99
4A-125
4A-141
Session 46 - Dry FGD Technologies
LTFAC Demonstration at Poplar River
1.7 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 CLEQ Process
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 Hign 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
Evolution of the B&W Durajet™ Atomizer
Characterization of the Linear VGA Nozzle for Flue
Gas Humidification
High SO2 Removal Dry FGD Systems
4B-151
4B-173
4B-189
4B-205
Session 5A - Wet Full Scale FGD Operations
FGD System Retrofit for Dalhousie Station Units 1 & 2
Zimmer FGD System: Design, Construction, Start-Up
and Operation
Results of an Investigation to Improve the Performance and
Reliability of HL&Fs Limestone Electric Generating Station
FGD System
Full-Scale Demonstration of EDTA and Sulfur Addition to
Control Sulfite Oxidation
Optimizing the Operations in the Flue Gas Desulfuiization Plants
or the Lignite Power Plant Neurath, Unit D and E and Improved
Control Conoepts for Third Generation Advanced FGD Design
Organic Acid Buffer Testing at Michigan South Central Power
Agency's Endicott Station
Stack Gas Cleaning Optimization Via German Retrofit Wet
FGD Operating Experience
Operation of a Compact FGD Plant Using CT-121 Process
5A-1
5A-17
5A-37
5A-59
5A-81
5A-101
5A-127
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 SO2/NOX Removal 5B-21
Combined SQ2/NOx Abatement by Sodium Bicarbonate
Dry Injection 5B-41
VI
-------
SO2 and NOx Control by Combined Dry Injection ol" Hydrated
lime and Sodium Bicarbonate
Engineering Evaluation of Combined N0x/S02 Controls for
Utility Application
Advanced Flue Gas Treatment Using Activated Char Process
Combined with FBC
Combined SO2/NOX Control using Ferrous *EDTA and a
Secondary Additive in a lime-Based Aqueous Scrubber System
Recent Developments in the Parsons FGC Process for Simultaneous
Removal of SOx and NOx
Session 6A - Wet FGD Operating Issues
Pilot-Scale Evaluation of Sorbent Injection to Remove SO3 and HC1
Control of Acid Mist Emissions from FGD Systems
Managing Air Toxics: Status of EPRI PISCES Project
Results of Mist Eliminator System Testing in an Air-Water
Pilot Facility
CEMS Vendor and Utility Survey Databases
Determination of Continuous Emissions Monitoring
Requirements at Electric Energy, Inc.
Improving Performance of Flushless Mechanical Seals in Wet FGD
Plants through Field and Laboratory Testing
Sulas FGD Demonstration Plant Limestone-Gypsum Process:
Performance, Materials, Waste Water Treatment
Session 66 - Clean Goal Demonstrations
Recovery Scrubber - Cement Application Operating Results
The NOXSO Clean Coal Technology Demonstration Project
vii
-------
Economic Comparison of Coolside Sorbent Injection and Wet
Limestone FGD Processes
Ohio Edison Clean Coal Projects Circa; 1991
Sard tech's 2-5-MWe Magnesia Dry-Scrubbing Demonstration
Project
Application of DOW Chemical's Regenerable Flue Gas
Desulfurization Technology to Coal-Fired Power Plants
Pilot Testing of the Cansolv® System FGD Process
Dry Desulphurization Technologies Involving Humidification
for Enhanced SO2 Removal
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 Tediniques 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 Sorbents 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
6B-33
6B-55
6B-79
6B-93
6B-105
6B-119
viii
-------
Commercial Application of Dry FGD using High Surface Area
Hydrated Lime
Initial Operating Experience of the SNOX Process
Progress Report of the NIPSCO - Pure Air - DOE Clean Coal n
Project
Development of a Post Combustion Dry SO2 Control Reactor
for Small Scale Combustion Systems
Scrubber Reagent Additives for Oxidation Inhibited Scrubbing
Recovery of Sulfur from Calcium Sulfite and Sulfate
Scrubber Sludges
Magnesite and Dolomite FGD Technologies
SO2 and Particulate Emissions Reduction in a Pulverized Coal
Utility Boiler through Natural Gas Cofiring
Design, Installation, and Operation of the First Wet FGD for a
Lignite-Fired Boiler in Europe at 330 MW P/S Voitsberg 3 in Austria
VOLUME 5
Session 8A - Commercial FGD Designs
Mitsui-BF Dry Desulfurization and Denitrification Process
Using Activated Coke
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 US.
Utilities for Add Rain Remediation
Incorporating Full-Scale Experience into Advanced limestone
Wet FGD Designs
Design and Operation of Single Train Spray Tower FGD Systems
Selecting the FGD Process and Six Years of Operating Experience
in Unit 5 of the Altb ach-Deizisau Neckarwerke Power Station
Development and Operating Experience of FGD-Technique at the
Vcelklingen Power Station
Advantages of the CT-121 Process as a Throwaway FGD System
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.on FGD Performance and
By-Product Gypsum Properties 8B-47
Quality of FGD Gypsum 8B-69
Chemical Analysis and Flowability of ByProduct Gypsums 8B-91
Evaluation of Disposal Methods for Oxidized FGD Sludge 8B-113
Commercial Aggregate Production from FGD Waste 8B-127
-------
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 US. 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 14 technical sessions:-
Session
Subject Area
I
Opening Remarks by EPRIJEPA and DOE Guest Speakers
1
Emission Allowance Panel Discussion
2
dean 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
dean Coal Demonstratioins/Emerging Technologies
7
Poster Session - papers on all aspects of SQ2 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 Shdley Hdler - Assistant
Policy subcommittee on Energy and Power, U. S. Congress,
Jack . . S. Siegel - Deputy Assistant Secretary, Office of Coal Technology, US.
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, U-S- 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
-------
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, W5. 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 EPL Petersburg 1 & 2 CAAA Retrofit FGDs, CP. Wedig, Stone &
Webster Engineering Corp.
2-9 Evaluation af SO2 Control Compliance Strategies at Virginia Power,
J.V. Presley, Virginia Power
Session 3A Wet FGD Process Improvements
Session Chain David R. Owens - EPRI
3A-1 Overview on the Use of Additives in Wet FGD Systems, R.E. Moser,
EPRI
3A-2 Results of High SO2 Removal Efficiency Tests at EPRI's HSTC, G.
Stevens, Radian
3A-3 Results of Formate Additive Tests at EPRTs HSTC, M. Stohs, Radian
Corp.
3A-4 FGDPRISM, EPRTS FGD Process Model-Recent Applications, J.Gl
Noblett, Radian Corp.
3A-5 Additive Enhanced Desulfurization for FGD Scrubbers, G. Juip,
Northern States Power
3A-6 Techniques for Evaluating Alternative Reagent Supplies, C.V. Weilert
Burns & 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
•viii
-------
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 4c McDonnell Engineering Co.
4A-4 Development of Advanced Retrofit FGD Designs, CLE. 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, US.
Rosenberg, Batelle
4A-7 Economic Comparison of Materials of Construction of Wet FGD
Absorbers & Internals, W. Nischt, Babcock & Wilcox
4A-8 The 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 LLFAC Demonstration^. Enwald, Tampella Power Ltd.
4B-2 1.7 MW Pilot Results for Duct Injection FGD Process Using Hydrated
Lime Upstream of 2m 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, TV A
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, JJL 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.
Zimmer FGD System, W. Brockman, Cincinnati Gas & Electric
Results of on Investigation to Improve the Performance and Reliabiity
of HL&Fs Limestone Electric Generating Station FGD System, M.
Bailey, Houston Lighting & Power
Full-Scale 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 Germain Retrofit Wet FGD
Operating Experience, H. Weiler, Ellison Consultants
5A-8 Operation of a Compact FGD Plaint 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 Removed Employing Absorbent Prepared
from Fly Ash, H. Tsuchiai, The Hokkaido Electric Power Co.
5B-2 Furnace Slurry Injection for Simultaneous SO2/NOX Removal, B.K.
Gullett, US. 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. Cichanowicz, EPRI
5B-6 Advanced Flue Gas Treatment Using Activated Char Process
Combined with FBC, H. Murayama, Electric Power Development Co.
5B-7 SO2/NOX 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 Flushless Mechanical Seals in Wet FGD
Plants through Field 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 Kiiby, Dow Chemical
6B-7 Pilot Testing of the Cansolv System FGD Process, L.E. Hakka Union
Carbide Canada LTD.
6B-8 Dry Desulfurization Technology Involving Humidification for
Enhanced SO2 Removal, D J. 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
SC>2 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 SC>2 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, AS. 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 Flue Gas, W. Jozewicz, Acurex
7-13 Commercial Application of Dry FGD using High Surface Area Hydrated
lime, F. Schwarzkopf, Florian Schwarzkopf PE.
7-14 Initial Operating Experience of the SNOX Process, D.J. Collins, ABB
Environmental System
7-15 Progress Report of the NIPSCO - Pure Air - DOE Clean Coal n Project, S.
Satrom, Pure Air
7-16 Development of a Post Combustion Dry SO2 Control Reactor for Small
Scale Combustion Systems, Ja_. 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 lignite
Fired Boiler in Europe at 330 MW P/S Voitsberg 3 in .Austria, H.
Kropfitsch, Voitsberg
Session 8A - Commercial FGD Designs
Session Chain 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
8A-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, Noell 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^. 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 7
POSTER PAPERS
SUMMARY OF GUIDELINES
FOR THE USE OF FRP IN UTILITY FGD SYSTEMS
Winston J. Renoud
Fiberglass Structural Engineering, Inc.
316 E. McLeod Road, Suite 1
Bellingham, Washington 98226
ABSTRACT
In recent years, Fiberglass Reinforced Plastic (FRP) has demonstrated
successful performance as the material of construction for FGD
systems. Excellent corrosion resistance, high reliability, low
maintenance and attractive initial cost give FRP the potential for
extensive use in future FGD systems.
Despite its potential, successful use of FRP has been hampered by a
lack of clear standards for design and fabrication. Many utility
companies are inexperienced in purchasing FRP equipment, and there has
not been adequate information available to project engineers and
managers in order to control quality.
EPRI is presently developing a guideline of value to power industry
purchasers of FRP equipment used in FGD systems. The guideline
provides a basic understanding of FRP composite material and design
issues, as well as an overall strategy for quality control. This
paper will briefly discuss this guideline.
7-1
-------
Fiberglass Reinforced Plastic (FRP) has great potential for service in
the process environments found in Flue Gas Desulfurization (FGD)
systems. It has very good resistance to corrosion and seismic damage,
is light weight, and is easily maintained. In many cases, FRP is
lower in its initial life cycle costs than alternative materials.
These attributes have made FRP a desirable material in Flue Gas
Desulfurization.
FRP's distinctive characteristics maJce it unique from the standpoint
of materials and structural engineering. The lack of practical
engineering knowledge at the level of the end-user is probably the
greatest drawback to this material. Proper design and quality control
are critical in custom, high performance FGD equipment made of
advanced materials like FRP, and it is difficult to assure adequate
performance without knowledge of the related engineering.
Unfortunately, there are few good sources of information relating to
application of FRP in FGD service.
EPRI is in the final stages of preparing and offering a document which
covers most of the information that an owner needs in order to under-
take and manage an FRP project. Background and specific information
are presented in such a way that end-users will understand not only
the options available to them but also the basis for sound decisions.
This paper provides a summary of the topics addressed in the upcoming
EPRI guidelines. The information presented in the guidelines will be
specifically limited to the design and fabrication of equipment built
entirely from fiberglass and intended specifically for utility FGD
systems. Furthermore, this guideline will consider only the wet
process slurry systems for SO, removal. The three main processes of
this type utilize limestone, lime, or alkaline fly ash as reagents.
POTENTIAL APPLICATIONS OF FRP IN FGD SYSTEMS
The corrosion resistance of FRP makes it a good candidate for use in
the construction of FGD equipment where the inlet quenching systems
limit continuous operating temperatures to 220°F or less. Since no
two systems are exactly alike, the physical and chemical conditions of
each system must be analyzed on a case by case basis to optimize the
use of FRP materials. FRP laminates can be tailored to meet the
conditions present in each zone of the FGD process. This is
accomplished by controlling resin selection, laminate construction and
flexibility, type of glass reinforcements, addition of fillers and
methods of cure.
7-2
-------
The main factors that affect the suitability of fiberglass as a
construction material for FGD systems include: The continuous
temperature of the process, the type of process (alkaline fly ash
systems are more abrasive than others), the chemical composition of
the gas stream (coal, sulfur, fluoride, and chemical content), the
source of make up water, the reheat methods employed and the operation
and maintenance policies. The use of FRP may also be influenced by
the addition of chemicals to the process stream to enhance S02 removal
or to reduce the build-up of hard deposits.
BASICS OF FRP
Laminated composites consist of several individual layers of
reinforcement bonded together by saturating with a resin matrix.
Lamination is used to combine the special characteristics of each
layer and optimizes the overall, or "composite" properties and
performance of the laminate. An example of a familiar laminate is
plywood, which gains its strength and stability from the alternating
grain orientation and grade of its various plies.
FRP laminates consist of glass reinforcing filaments embedded in a
matrix of chemically resistant thermoset resin. Thermoset resins are
unsaturated liquid polymers that react in the presence of a catalyst
by cross-linking their molecular chains to form a solid mass. The FRP
composites obtain their high tensile strength from the glass
reinforcement and chemical resistance from the surrounding thermoset
resin matrix.
One of the advantages of FRP composite materials is that the physical
and chemical resisting characteristics can be custom designed for
various applications through careful selection of the type of resin,
the orientation and placement of the glass reinforcements and through
the inclusion of organic and inorganic fillers and other additives.
By modifying these various attributes, the FRP designer has great
versatility in determining strength and stiffness, resistance to
corrosion and abrasion, durability, and high temperature performance.
FRP structures are largely immune to many of the electromechanical
corrosion mechanisms present in FGD systems that eventually destroy
corrosion resistant metal alloys through stress corrosion cracking,
intergranular corrosion, pitting and crevice corrosion. In addition,
FRP composites are not chemically affected by the majority of chloride
and sulfite trace elements found in FGD systems that have been shown
to increase the susceptibility of stainless steel to pitting and
stress-corrosion cracking.
On the other hand, FRP composites have lower operating temperatures
than metallics. The maximum recommended operating temperature for FRP
in gas stream applications is generally in the range of 300° to 400°
F. Repeated temperature cycling over a period of time may lead to the
formation of defects in the laminate such as cracks in the resin
matrix, subsurface blistering, breakdown of glass to resin bonds, or a
reduction of physical properties.
7-3
-------
FRP composites generally have higher coefficients of expansion than
metallics, which must be taken into consideration in the structural
design of the equipment. Abrasion may also require special
consideration in some areas of the FGD system. The properties of most
FRP laminates vary with direction due to the oriented nature of the
glass reinforcement, so it is practical to design a cost effective FRP
composite to meet the special conditions found in FGD systems.
There are three main fabrication processes employed to produce
fiberglass corrosion-resistant equipment for industrial service.
These processes are hand lay-up, filament winding and field winding.
Where a choice exists, the method of fabrication should be specified
by the engineer to best satisfy the laminate requirements.
FRP EQUIPMENT DESIGN
The proper design of FRP equipment entails complexities beyond the
typical issues confronting designers. The special nature of the FRP
composite is easily overlooked, even by competent engineers. In
addition to familiarity with the general principles of equipment
design, there are two major areas of specialized knowledge that are
critical.
Possibly the most important and most frequently misunderstood design
issue is that of physical properties, including the various failure
modes under different types of loads. The physical properties vary
widely with the materials and fabrication methods employed, and
analysis is complicated by the bi-modular behavior of many laminate
types. In addition, the long-term success of FRP equipment will
require developing a failure criteria based on parameters other than
ultimate strength; FRP equipment may be damaged by severe operating
environments at a stress level that is well below ultimate stress.
Another requirement for FRP design is an understanding of appropriate
design and analysis procedures. Too often, analytical techniques
developed for traditional materials are extended to use with
composites.
There are many design guides and standards which have been developed
to aid the engineer in designing FRP equipment. These guides and
standards have been developed over the many years of growth in the FRP
industry, and, with the increase in research, available technology,
and the experience of design engineers, they naturally vary in
completeness and approach.
It should be recognized that specifying any particular document does
not ensure that the equipment will be designed properly. In other
words, compliance with a design standard does not necessarily mean
that the design will be adequate for the intended service. It is
essential that the design engineer be experienced with and
knowledgeable of the attributes of FRP, understanding the special
nature of laminates and their fabrication techniques. Sound
engineering judgment and the ability to accurately analyze the
structural properties of FRP laminates and their application to FGD
service are necessary to obtain equipment adequate for long term
service.
7-4
-------
The design specification for FRP equipment should provide adequate
engineering details to allow the design methodology and completeness
to be thoroughly reviewed. The "design package" should include a
complete set of engineering calculations and detailed construction
drawings and specifications, which will ensure that assumptions made
during design are valid.
PROCUREMENT STRATEGY
There are several main players which can be involved in procuring FRP
equipment. The ultimate user of the equipment (Owner) and the builder
of the equipment (Fabricator) are always involved, and, on smaller
projects, may handle all required functions without additional
assistance.
In many cases, however, one or more additional parties are included in
the procurement chain. An engineering firm is often responsible for
global design of an expansion or modification. This firm may have
total responsibility, or the Owner may retain responsibility for some
portions, such as major pieces of equipment.
Whether the engineering firm is involved or not, certain kinds of
specialty equipment may be procured as "packages". Many firms exist
to provide the process, engineering, and equipment for these
subsystems, retaining responsibility for the proper performance of the
process as well as the equipment. In most cases, these process
companies have the FRP equipment fabricated to their own
specifications, then resell the equipment as a part of the system
package.
Any one of the parties mentioned can optionally employ an independent
engineering firm for the detailed design of the equipment.
The design of the FRP equipment can be divided into two segments,
typically performed by different parties. The configuration of the
equipment is dependent on the overall design of the system and
process. The Owner, engineering firm or process company normally
provides this level of design, determining design criteria such as
temperatures and pressure.
The second element of the design is the detailed design, which
includes the structural engineering, material selection, ana
development of specific construction details and procedures. The
fabricator of the equipment historically has provided the detail
design for FRP equipment, based on the configuration or process design
which has been provided to them. Fabricators often work to
specifications provided by the Owner, engineering firm or process
company, but these specifications vary widely as to quality,
completeness and approach.
Relying on the fabricator for detailed design is an approach used
widely in the purchase of all types of industrial equipment, and is
not unique in the procurement of FRP equipment. Unlike equipment
fabricated from traditional materials such as steel, however, FRP
fabrication practices and equipment can vary widely among suppliers.
7-5
-------
In a few cases, owners have developed in-house expertise with regard
to FRP equipment detail design and specification. This practice is
most common in the chemical industry, where FRP equipment usage is
high and a basic understanding of plastics may already exist.
Development of in-house FRP design expertise or standards has been
limited to a handful of major companies, and is not common among major
engineering firms and process companies.
There are two distinct approaches normally taken by buyers of FRP
equipment, differing in the allocation of design responsibility
between the buyer and the supplier. To a large extent, equipment is
purchased on a "performance" basis, meaning that the supplier provides
the detailed design as a part of the purchase and is responsible for
assuring that it performs as intended. Since true long-term
performance cannot be measured initially, it is often difficult for an
Owner to know whether the equipment will prove adequate in long-term
service.
The alternative approach is for the buyer to provide a detailed design
of the equipment prior to requesting bids, removing the design
responsibility from the supplier. The Owner's "descriptive"
specification reflects decisions as to how conservative the design
should be, and further allows competitive bids on a single
predetermined design.
In reality, most specifications fall somewhere between the two extreme
approaches, with the owner making some of the decisions and specifying
as such, while leaving much of the design responsibility with the
supplier. When additional parties are involved in the procurement
chain, each party normally assumes some portion of the design
responsibility. Owners can exercise greater control over the long-
term success of FRP equipment by active involvement in the design
decisions which affect quality.
QUALITY MANAGEMENT
In order to have the longest possible service life from FRP equipment,
attention must be paid to the quality issues in several areas. In
general, these include:
• Process design and identification of design criteria
• Appropriate material selection and equipment design
• Fabrication quality
• Verification of design and fabrication (testing)
• System maintenance
The Owner can establish procedures that allow for control of the key
decisions affecting quality. In order to accomplish this, it is
important to become acquainted with the special nature of FRP and the
complexities of analysis and fabrication. The Owner should adopt an
appropriate procurement strategy that allows subcontractors to provide
equipment or services in a manner which promotes quality.
7-6
-------
One of the first important steps an Owner can take to assure overall
quality of the FGD equipment is to insure that appropriate design
criteria are established. The party responsible for the process
design of the equipment would normally develop these parameters to be
used for equipment design. In order for the equipment to be designed
efficiently and procured for the lowest possible cost, the criteria
must be realistic and accurate. A qualified designer using methods
accurate for composite analysis will include adequate design (safety)
factors to assure long term adequacy based on realistic criteria.
While it is important to state conrervative values (criteria) such
that potential service extremes are taken into account, this should be
identified in terms of frequency. This allows the designer to
incorporate these criteria into the design process in an appropriate
manner. It is generally best to incorporate the necessary safety
factor at the time of detail design.
A natural tendency exists at all levels of design to be somewhat
conservative with each design decision. When this happens in the
determination of the design criteria, the impact can be magnified
dramatically as each progressive level of design is accomplished.
Overstating design criteria is wasteful, in that final designs can
become excessively conservative and expensive. In the extreme case,
overly conservative design criteria could cause a decision to change
materials completely when it would not otherwise be necessary. The
party responsible for process design should be made aware of the
special nature of FRP composites. It may be possible to modify the
process somewhat in order to bring service conditions into a more
suitable range for FRP.
Even the highest level of effort applied to the design and engineering
of FRP equipment will have little bearing on the success of a project
unless all phases of fabrication and installation are carried out by
competent, experienced people in accordance with an established
quality management program. An effective quality management program
will evaluate the vendor's expertise, define the fabrication
activities to be performed and establish levels of inspection required
to assure vendor compliance with the project specifications,
engineering drawings, and related contract documents. The intent of
the program is to prevent deficiencies in fabrication and to assure
that the level of reliability designed into the equipment is actually
achieved in the fabrication and installation process.
Inspection levels are established by the power plant after a careful
evaluation of the need for, and the extent of control desired. The
level of need for controlling quality is based on the level of
consequence if the equipment were to fail. Equipment critical to
plant operation, operator safety or significant environmental concerns
receive the highest inspection priority.
7-7
-------
Quality management: programs, as they relate to fabrication and
installation of FRP FGD equipment, encompass four main themes:
• Qualifying Vendors
• Vendor's Quality Control Program
• Owner's Quality Assurance Program
• Qualifying Inspectors
SYSTEMS MAINTENANCE
Fiber reinforced plastic equipment used in FGD applications is subject
to demanding thermal and corrosive conditions. These conditions, when
properly considered and accounted for in the equipment design and
materials selection, will be pushed to its limits of performance due
to elevated temperature upsets or unexpected chemical concentrations
in portions of the FGD system.
FRP is often specified where maximum corrosion resistance in
aggressive environments is required. When used in FGD systems, it is
normal and expected that FRP equipment will degrade over a period of
time and that periodic inspection and regular maintenance will be
required to assure system reliability.
These factors point to the need for a comprehensive inspection and
maintenance program designed to ensure the long term use and
uninterrupted availability of the power plant. A recommended approach
to maintaining FRP is through a predictive maintenance (PM) inspection
program combining routine, scheduled inspections with a detailed
reporting system for tracking, documenting and recording defects.
With good, current knowledge of equipment condition, maintenance needs
can be projected and developing problems can be dealt with on a
planned basis.
Predictive maintenance and repair decisions involve subjective risk.
The object is to evaluate the probability of equipment failure and
consequence of failure if repair or replacement is not done versus the
cost of the repair if it is done. If unnecessary repairs are
specified, costs will be high for little or no gain. If repairs are
delayed and the equipment becomes unserviceable, costs for operational
losses and reduced service life will be high also. Finding the
balance between risk and cost is the challenge facing the inspector
and engineering specialist.
The least risky choice is to repair all defects when they are found.
For defects repaired with a small effort or cost, this is reasonable.
If the cost of the repair is great, however, the inspector is in a
position to evaluate the probability of damage if the repair is
delayed. If the probability of further damage is low, the cost of
repair might well be delayed or avoided for the intended life of the
equipment.
By making a detailed comparison of current and previously documented
defects, a trend can be established that will allow the plant to track
degradation and help predict the point at which the equipment must be
7-8
-------
replaced or repaired. In addition, previous inspection documentation
is invaluable for predicting if a defect will penetrate through the
corrosion liner and into the structural wall before the next scheduled
inspection.
The primary elements of a successful PM program are: Planned
inspection intervals, quantitative documentation of defects,
knowledgeable evaluation, and carefully administered repairs. The
intent of PM is to detect problems in their early stages, so
maintenance costs and the threat to safety and reliability caused by
unexpected failure of equipment are reduced.
An effective PM program for FRP will include the following elements:
# A regularly scheduled program for interior and exterior
inspection, with frequency of inspection based on the
age and condition of the equipment, severity of
service, and history of previous defects.
# The use of inspection tools and methods suited to FRP
materials for determining surface hardness, extent of
subsurface defects, corrosion liner thicknesses, and
overall condition of FRP laminates.
# A system for documentation and reporting that allows
inspection findings to be chronologically updated so
that changes in the condition of the equipment can be
tracked and, based on long term trend evaluations,
future maintenance needs projected.
# Inspection personnel who have the ability to interpret
the significance of specific defects, who are
knowledgeable and experienced in the inspection of FRP
equipment, and who are capable of recommending
appropriate repairs.
# A defined plan for repairs, if repairs should become
necessary, including the specification of materials and
methods of repair and verification of compliance
through inspection.
The need for ongoing inspection of FRP FGD equipment cannot be over
emphasized. Inspection is as important for assuring long term,
uninterrupted use of the equipment as the original design and
engineering used to build it.
COMPLETE GUIDELINES
The final Guidelines for the Use of FRP in Utility FGD Systems will
address each of the above subjects in far greater detail, providing
power plant owners with useful information in purchasing and
controlling quality of FRP equipment. Appendices will serve as a
further valuable resource, identifying relevant terms, current
industry standards and information obtained from the manufacturers of
resins.
7-9
-------
Intentionally Blank Page
7-10
-------
DEVELOPMENT AND EVALUATION OF HIGH-SURFACE-AREA
HYDRATED LIME FOR SOz CONTROL
M. Rostam-Abadi and D. L Moran
Illinois State Geological Survey
Champaign, IL 61820
7-11
-------
Intentionally Blank Page
-------
ABSTRACT
A process for producing high-surface-area (HSA) hydrated lime has been developed at the
Illinois State Geological Survey (ISGS). The process has been optimized in a batch, bench-
scale reactor and has been evaluated as being technically feasible and economically
attractive. The "know how" is available to tailor properties of hydrated limes for specific S02
removal applications. A patent application covering the process was filed with the U. S.
Patent Office in 1990. An integrated, continuous lime hydration process optimization unit
(POU) capable of producing 20-100 pounds of HSA hydrated jime per hour has been
designed, constructed, assembled and tested. The POU will provide engineering and scale-
up data for rapid commercialization and transfer of the technology to Gme industries.
S02 removal efficiencies of five ISGS HSA hydrated limes with surface areas ranging from
40 to 60 rr^/g were evaluated in pilot-scale dry sorbent injection test facilities in Consolida-
tion Coal Company's Coolside unit, in Research-Cottrell ESTs boiler economizer system,
and in the U. S. EPA/Acurex innovative furnace reactor under furnace sorbent injection
conditions. Calcium utilization data were also obtained at economizer temperatures in a
differential flow reactor at U. S. EPA/Acurex. The ISGS HSA hydrates showed up to 84%
SQz removal (Ca/S=2) under furnace sorbent injection conditions (a 22 mVg commercial
hydrates removed 60%) and up to 80% calcium conversions under boiler economizer
conditions (59% for the commercial hydrate). Tests at Research-Cottrell EST under boiler
economizer conditions showed 64% SOz removal (Ca/S=2) for a 48 rr^/g HSA hydrate
compared to 45% for its commercial counterpart Under Coolside conditions (Ca/S=2,25 * F
approach to adiabatic saturation) the system S02 removals ranged between 67 and 81 % for
the HSA hydrates. The commercial hydrate removed about 50% SOj.
7-13
-------
INTRODUCTION
Dry sorbent injection (DSI) processes which inject hydrated lime powders to adsorb S02
from stack gases can be preferred for small and older power plants because these
processes are easier to install, require less space and have lower projected capital cost than
wet scrubbers using lime- or limestone-water slurry. The Electric Power Research Institute
(EPRI) estimates that 450 power plants east of Colorado, 150 MW or less in size and
accounting for 35.000 MW total capacity, could benefit by installing DSI systems (1).
The operating costs of DSI processes depend primarily on the expense and performance
of the sorbents used. The S02-capturing performance of commercial hydrated lime in these
processes is generally less than 30% of its theoretical capacity, representing less than 60%
S02 removal at Ca/S ratio of 2 (2, 3. 4). A more active sorbent that achieves substantially
higher S02 removals than commercially available hydrated limes would make these
processes suitable for more power plants, especially those burning high-sulfur coal. The
reduction in sorbent needed to achieve a given emissions level lowers not only the cost of
sorbent, but also expenditures for storage, handling and disposal depending upon
comparable delivered sorbent costs.
A major objective of past research has been to identify sorbent properties that enhance S02
capture in various DSI processes. Because the chemistry and physics of S02 capture vary
from one process to another, the properties of hydrated lime which are desirable for a
specific S02 removal application may not be desirable for other applications. For example,
in furnace sorbent injection (FSI) process, hydrated limes with high surface area, small
particle size and large pores are desired (5,6). In duct injection/humidification (Coolside),
hydrated limes with high surface area and porosity are preferred (7,8). In boiler economizer
process, a balance between surface area and particle size is needed to achieve maximum
S02 removal (9). In a boiler economizer process, the product not only should have a high
surface area (above 40 m2/g) but also particle diameters of 2 microns or less. If the particle
diameter is large (above 2-5 microns and depending on S02 concentration) Ca(OH)2, reacts
mainly with COz rather than S02 since the latter reaction is controlled by film diffusion (in a
typical flue gas generated by burning high sulfur coal, the concentration of C02 is 30 to 50
times greater than that of SOJ. Thus, it is desirable to provide calcium hydroxide having
high surface area, high porosity, and small particle size for use in DSI processes.
Commercial hydrated limes typically have surface areas in the range of 10-25 rr^/g, mean
particle diameters of 1.7-10 micrometers, and pore volumes in the range of 0.1-0.25 cc/g.
Several investigators have prepared small quantities of lime hydrated having surface areas
in the range of 35-50 m2/g by using an aqueous alcohol solution (10, 11). Rheinische
Kalksteinwerke GmbH, Wuelfrath, has developed and patented a process for the
manufacture of alcohol-water hydrated lime (WUELFRAsorp) having surface areas in the
range of about 35-55 m2/g (12). A production plant with a capacity of 6 tons of WUELFRA-
sorp per hour is currently in operation. Reportedly, small quantities of hydrates having
surface areas as high as 80 m2/g have been obtained by calcining (dehydrating) commercial
hydrated lime and rehydrating it using an alcohol-water hydration method (13).
7-14
-------
ISGS HYDRATION PROCESS
Since 1986, the Illinois State Geological Survey (ISGS), a division of the Illinois Department
of Energy and Natural Resources (ENR), has been developing a process to produce high-
surface-area (HSA) hydrated lime with far more activity for adsorbing S02than commercially
available hydrated lime. HSA hydrate prepared by the ISGS method has considerably
higher surface area and porosity, and smaller mean particle diameter and crystallite size
than commercial hydrated lime. A patent application covering the process was filed with the
U. S. Patent Office in 1990. Research Corporation Technologies (Tucson, AZ), a marketing
and licensing organization, has accepted the responsibility for patent prosecution, marketing,
and licensing of the ISGS process.
In the ISGS process, lime is hydrated at atmospheric pressure with an aqueous solution of
a reagent (one-step process) followed by an optional post-hydration wash step (two-step
process). The one-step process is capable of producing a hydrate with surface areas
ranging between 35 and 50 m2/g using commercial quicklimes as feed materials. Higher
surface area hydrates, above 50 m2/g, can be made using the two-step process. With either
the one-step or the two-step hydration methods, hydrates with even higher surface areas
(up to 85 rr^/g) can be made using specially-prepared quick limes.
During the past five years more than 500 experiments have been conducted to optimize the
ISGS hydration process and identify key parameters influencing hydrate properties (14,15,
16,17). These tests were conducted both at gram quantities in a laboratory-scale unit and
in 5-7 pound quantities in a bench-scale batch hydrator. The dependence of the hydrate
properties (surface area, porosity, particle size and crystallite size) important for S02
capture on operating conditions including hydration temperature, contact time, type and
quality of lime, quantity of water and reagent, and the amount of reagent used in the post-
hydration wash step have been investigated. Based on the results of these tests, optimum
conditions for producing HSA hydrates from different limes have been determined. Typical
properties of HSA hydrates prepared by the ISGS method and hydrates available
commercially are compared in Table 1. In addition to their considerably higher surface area
and porosity, the ISGS hydrates have smaller mean particle diameter and crystallite size than
commercial hydrated lime.
ECONOMIC EVALUATION OF THE ISGS HYDRATION PROCESS
In 1989, a technical and economic evaluation of the ISGS hydration process was conducted
by Arthur L Conn and Associates, Ltd. to determine the probable cost and the suitability
of the ISGS hydrate for manufacture on a commercial scale (18). This preliminary study was
performed on the basis of a lime hydration plant large enough to handle the demand of a
1000 MW coal-fired power plant site using a 12,000 Btu/Ib Illinois coal containing 3.5% sulfur.
The study concluded that the process to manufacture HSA hydrates by the ISGS method
is technically feasible and that conventional commercial equipment should be applicable.
The projected cost of production of the ISGS HSA hydrated lime was evaluated to be only
marginally greater than the cost of currently available, commercial hydrated lime. The cost
benefit gained by the large increase in S02 adsorption capacity for HSA hydrates far
outweighs this small increased sorbent cost.
7-15
-------
S02 REMOVAL PERFORMANCE OF ISGS HYDRATES IN DSI PROCESSES
Sorbents
The samples tested included five HSA hydrates prepared according to the one-step ISGS
process. WUELFRAsorp, and a commercial hydrated lime. For the pilot-scale DSI tests
reported in this paper, more than 1000 pounds of five HSA hydrates were prepared in the
batch, bench-scale hydrator. The chemical analyses and the physical properties of the test
sorbents are presented in Table 2 and 3. The surface areas of the HSA hydrates ranged
from 40 to 60 rrrVg compared to 22 m2/g for the commercial hydrate. The surface area of
the WUELFRAsorp sample tested was about 40 m2/g. Mean particle diameters of the HSA
hydrates were between 2.3 and 4.5 micrometers compared to 15 micrometers for the
WUELFRAsorp and 20 to 25 micrometers for the commercial hydrates. The storage history
of the WUELFRAsorp prior to shipment to the ISGS was not known. The sample could have
lost some reactivity while in storage.
Test Facilities
Pilot-scale DSI tests were performed with several ISGS hydrates to evaluate their SOz-
removal efficiencies under typical conditions of burning high-sulfur coals. Pilot-scale tests
were performed in Consolidation Coal Company's 100-kW Coolside pilot unit (Library. PA);
in Research-Cottrell ESTs 146-kW. boiler economizer furnace (Irvine, CA); and in U. S.
EPA/Acurex 14-kW, Innovative Furnace Reactor under FSI conditions (Research Triangle
Park, NC). Boiler economizer tests were also performed in a bench-scale flow reactor under
differential conditions with respect to S02 concentration at U. S. EPA/Acurex Corporation.
Furnace Sorbent Injection Tests
The S02 removal performances of the hydrates under furnace sorbent injection conditions
are shown in Figure 1 and Figure 2. The fuel used in these tests was natural gas doped
with SOj. In Figure 1. S02 removal is plotted as a function of Ca/S ratio for the ISGS HSA31
hydrate, and the commercial hydrate at 1042 and 1127 "C. The performance of the
commercial sorbent was more temperature dependent than for the HSA31 hydrate in the
temperature range studied. The ISGS HSA31 hydrate removed 53% S02 at Ca/S=1 and
about 85% S02 at a Ca/S=2, independent of furnace temperature. The commercial hydrate,
at Ca/S=2, captured about 58% at 1042 "C and 68% at 1127 "C. In Figure 2, S02 removal
for the HSA1, HSA2, HSA3. WUELFRAsorp and the commercial hydrates at 1200 "C are
shown. The ISGS HSA hydrates, regardless of surface area, captured about 84% S02 at
a Ca/S=2 compared to about 75% for the WUELFRAsorp. and about 60% for the
commercial hydrate. At Ca/S= 1. only the ISGS HSA hydrates had S02 removal greater than
50% (ranged from 53 to 60%).
7-16
-------
Boiler Economizer Tests
The S02 removal performance of HSA30 hydrate and the commercial hydrate is shown in
Figure 3. These data were obtained by Research-Cottrell EST. The fuel used was natural
gas doped with S02 (3000 ppm). The SOz removals obtained at 540 'C are plotted as a
function of Ca/S ratio in Figure 3. At a Ca/S ratio of 1, the HSA30 captured 39% S02
compared to 28% for the commercial sorbent At a Ca/S=2, HSA30 captured 64% S02
compared to only 45% for the commercial hydrate.
The calcination utilization data obtained at economizer temperatures in the differential flow
reactor at the U. S. EPA/Acurex is shown in Figure 4. The conversion of calcium to calcium
sulfite (or sulfate) is shown for the test sorbents. Calcium conversions achieved were 80,
75 and 51% for HSA1, HSA2 and HSA3, compared to 61% for the WUELFRAsorp and 59%
for the commercial hydrate. The highest S02 removal was observed with the HSA1 (50
rrrVg) suggesting that there maybe an optimum hydrate surface area for maximizing S02
capture under boiler economizer conditions.
Coolside Tests
The results of Coolside tests conducted at Consolidation Coal Company are shown in
Figure 5. The common conditions were 300 "F inlet flue gas temperature, 1500 ppm inlet
S02 content (dry basis), and 125 " F adiabatic saturation temperature (corresponding to 25 " F
approach to adiabatic saturation temperature). The flue gas flow rate was set at 175 scfm,
which provided a 2.0 second humidifier residence time. The S02 removals reported include
capture both in the humidifier and the baghouse.
Figure 5 shows the effect of the Ca/S molar ratio on system (Humidifier + baghouse) S02
removal in the pilot unit tests. The error bars represent one standard deviation. Points
without error bars have standard deviations smaller than the plotting symbol. The highest
S02 removal was observed using HSA2. The average system removals were 33, 53 and
81% at 0.5,1.0 and 2.0 Ca/S, respectively, using HSA2. This represents 154, 71 and 62%
(relative) higher SOz removal than the commercial hydrated lime. Sorbents HSA1 and HSA3
showed lower S02 removal than HSA2. but higher removal than the commercial hydrated
lime at similar Ca/S ratios. At 0.5, 1.0 and 2.0 Ca/S ratios SOz removals were 28, 47 and
67%, respectively (representing 115, 52 and 34% relative improvement over commercial
hydrate), using HSA1, and 30, 44 and 65%. respectively, using HSA3 (representing 131,42
and 30% relative improvement over commercial hydrate). Using the best performing
commercial hydrated lime, the S02 removals were 13, 31 and 50% at 0.5,1.0 and 2.0 Ca/S,
respectively.
Optimum Hydrate Surface Area for SO, Capture in DSI Processes
Figure 6 show the effect of hydrate surface area on S02 capture (at Ca/S = 2.0) under
different DSI conditions. These data were obtained from Figures 1-5. It is seen that there
is an optimum surface area of hydrate for maximum sulfur capture in each of the three
systems. Under FSI conditions, HSA hydrates with surface areas ranging between 40 to 60
rr^/g showed the highest S02 capture (about 85%). The S02 removal for the WUELFRAsorp
7-17
-------
and the commercial hydrate were 78% and 60%, respectively. Enhanced performance of
the ISGS HSA hydrates could be related to their smaller particle size, higher surface area,
and greater pore volume than for the other hydrates. Pore volume analyses of raw sorbents
indicate the volume of pores between 0.01 and 0.1 micrometers was substantially higher for
the HSA1 and HSA2 hydrates than for commercial hydrated lime. Pore volumes of hydrated
limes are expected to correlate with the pore volumes of the corresponding calcines formed
during calcination of hydrated lime. Due to the increase in molar volumes when converting
from CaO to CaSO< (16.9 vs. 46.0 cm3/mole), pore plugging is known to limit the sulfation
reaction. Sorbents with a high volume of larger pores are expected to capture more SO^
Under boiler economizer conditions, the calcium utilizations for HSA1, HSA2, HSA3,
WUELFRAsorp and the commercial hydrate were 80%, 75%, 51%, 61% and 55%,
respectively. The maximum S02 removal was observed for the 50 rr^/g HSA1 hydrate.
Under Coolside conditions the S02 removal increased with increasing surface area and was
67% for HSA1, 85% for HSA2, 65% for HSA3,and 50% for the commercial hydrate.
Based on the data shown in Figure 6 and Table 3, the optimum hydrates for maximizing S02
removal under FSI or boiler economizer conditions appear to have surface areas in the
range of 40 to 50 m2/g, mean particle diameters below about 1.5 micrometers (as measured
by sedigraph), and pore volumes above about 0.25 cc/g. Hydrates with surface areas
above 50 m2/g would likely be more effective for capturing S02 if their particle size could be
significantly reduced (i.e.. to less than 1 micrometer).
Hydrated limes having optimum properties for S02 removal in the Coolside process have
surface areas of at least 55 m2/g (e.g. about 60 m2/g), mean particle diameters of below
about 5 micrometers, and pore volumes above about 0.3 cc/g. For the Coolside process,
particle size of hydrates appears to be of secondary importance.
PROCESS OPTIMIZATION UNIT
In November 1990, the ISGS received a grant from the Illinois Department of Commerce and
Community Affairs under the Governor's Challenge Grant Program to build and operate an
integrated, continuous lime hydration process optimization unit (POU) capable of producing
20-100 pounds of HSA hydrated lime per hour. The goals of the program are to generate
design, construction, and operation data necessary for the private sector to scale-up the
process to a commercial level, produce samples for further testing and to evaluate the
economics of the process.
The POU has been designed, constructed and assembled. The mechanical operability tests
of the individual units have been performed and baseline data have been obtained. Seven
shakedown tests, a total of 40 hours of operation, have successfully been performed with
the hydration reactor. The mechanical operability and performance of the reactor has been
demonstrated in these tests. The results of the shakedown tests with the hydration reactor
have indicated that there are no serious problems with continuous operation and scale-up
of the process and that the quality of the hydrated lime is not affected. Process conditions
for optimizing product based on the type of dry sorbent injection processes are being
investigated. This "know-how" will help to tailor properties of hydrated lime products for
specific S02 removal applications.
7-18
-------
ACKNOWLEDGEMENTS
This work was funded in part by grants from the Illinois Coal Development Board and the
Illinois Coal Industry Committee through the Center for Research on Sulfur in Coal and the
Illinois Department of Commerce and Community Affairs. The authors thank Dr. Gullet at
U. S. EPA. Mr. Bruce at Acurex Corporation, Dr. Yoon and Dr. Withum at Consolidation Coal
Company, and Mr. Bortz at Research-Cottrell EST for obtaining DSI test data reported in this
paper.
REFERENCES
1. Toole-O'Neil, B., 1990. Private communication.
2. Beitiel, R. J.; Gooch. J.; Dismukes. E.; Muzio, L Proc. Svmp. on Dry SO-, and Simul.
SOJNO. Control Technol. 1986.1. 6-1. EPA-600/-85-020a (NTIS PB85-232353).
3. Snow. G.C.; Lorrain, J.M.; Rakes, S.L Proc. Joint Svmp. on Dry SO, and Simul.
SOJNO- Control Technol. 1986,1, 6-1, EPA-600/9/86/029a (NTIS PB87-120465).
4. Bortz, S. J.; Roman, V.P.; Yang, C.J.; Offen, G. R. Proc. Joint Svmp. on Dry SO* and
Simul. SQJNO. Control Technol. 1986, 2, 31.
5. Gullett, B. K.; Bruce, K. R. First Combined FGD and Dry SO, Control Symposium. St.
Louis. MO. Oct. 25-28 1983. 6-7, EPA-600/9-89-0369(NTIS PB89-172159).
6. Newton. G. H.; Chen. S.L; Kramtich. J.C. AlChE J. 1989, 35f6V 988-994.
7. Yoon, H., Stouffer, M. FL, Rosenhoover, W. A., and R. M. Statnick. 1985a.
"Laboratory and Applied Development of Coolside SOz Abatement Technology." In
Proceedings of the Second Pittsburgh Coal Conference. U. S. DOE. Pittsburgh, PA.
8. Yoon, H., Ring. P. A. and F. P. Burke. 1985b. "Coolside S02 Abatement Technology:
1 MW Field Tests." Paper presented at the 1985 Coal Technology Conference,
November 12-14, Pittsburgh, PA.
9. Bortz. S. J.; Roman, V., Offen, G. R. First Combined FGD and Dry SO, Control
Symposium. St. Louis. Mo. Oct. 25-28. 1988, Paper No. 10-7.
10. Muzio, L J; Boni, A. A.; Beittel, R.; Offen. G. R. 1986 Joint Symposium on Dry SO,
and Simultaneous SOJNO. Control Technologies. Volume 1, 1986.
11. Blythe, G.; Smith, R.; McElroy, M; Rhudy, R.; Bland. V.; Martin, C. 1986 Joint
Symposium on Dry SO, and Simultaneous SOJNO. Control Technologies. Volume
2. 1986.
12. Bestek, H., et al.. U. S. patent 4,636,379, June 1987.
7-19
-------
13. Gooch, J. P., Be'rttel. FL and Dismukes, E B.. In Proceedings: 1986 Joint Symposium
and Dry SO-, and Simultaneous SO-JNO. Control Technologies, volume 1, 1986.
14. Rostam-Abadi, M.; Moran, D. L; Harvey, FL D.; Frost FL R., "Enhancement of S02
Sorption Capacity of Lime/Limestone". Final Technical Report to Center for Research
on Sulfur in Coal, Carterville, Illinois, 1986.
15. Rostam-Abadi, M.; Moran, D. L; Harvey, R. D. "Development of High Surface Area
Hydrated Lime for Sulfur Dioxide Control", Final Technical Report to the Center for
Research on Sulfur in Coal, Carterville, Illinois, 1988.
16. Rostam-Abadi, M.; Moran, D. L; Harvey, R. D.; "Development and Evaluation of High
Surface Area Hydrated Lime for Sulfur Dioxide Control," Final Technical Report to the
Center for Research on Sulfur in Coal, Carterville, Illinois, 1989.
17. Rcstam-Abadi, M.; Moran. D. L "Evaluation of High Surface Area Hydrated Lime for
Sulfur Dioxide Control," Final Technical Report to the Center for Research on Sulfur
in Coal, Carterville, Illinois, 1990.
18. Conn. A. L; Bowman, W. H.; Kotharie, M. M., "Economic Evaluation of a Process for
Preparing High Surface Area Hydrated Lime," Final Report to the Center for
Research on Sulfur in Coal, Carterville, Illinois, 1989.
7-20
-------
Table 1
Typical properties of commercial and ISGS hydrated Gme
ISGS Commercial
Surface area (m2/g) 35-85 10-20
Mean particle diameter 0.5-2.5 1.5-5
(micrometer)
Crystallite size 140-160 220
(angstroms)
Pore volume (cc/g) 0.25-0.6 0.1-0.25
Table 2
Chemical analysis of test sorbents
Ca(OH)2 Wt%
Sorbent Ash Carbon Hydrogen CO, CaO (by TGA) moisture
HSA-1
74.35
1.03
2.80
NA*
73.85
892
1.0
HSA-2
74.49
122.
2.82
2.93
71.28
862.
0.4
HSA-3
74.75
0.90
2.82
1.60
73.44
91.6
0.8
Commercial
75.42
0.41
2.70
1.51
72.62
93.0
1.0
* not available
7-21
-------
Table 3
Physical properties of hydrates tested for S02 removal
Samole ID
BET
Surface Area
(rrP/a)
Pore Volume®
(crrrVal
Parti cle Size
frrecranieters)
HSA1
50
0.24
0.7"(3.5)b
HSA2
60
0.36
0.8*(3.5)b
HSA3
40
0.18
1.1"
HSA30
48
NAd
(4-5)b
HSA31
47
NAd
(2.3)b
WUELFRAsorp
40
NAd
(15)b
commercial hydrate 22
0.16
1725(2^
•measured by Sedigraph Particle Size Analyzer
bmeasured by Microtrac Particle Size Analyzer
cpores smaller than 0.25 micrometers
"not available
7-22
-------
CO
>
3
I
3
100
80
60
40
20
3000 ppm SO2
(104Z 1127 C)
.P (1127 C)
-
y' ^(1042 C)
1
m/ /
/ /t*
- ///
¦ HSA31 (47 n?/g)
V 1 1
• Commercial (22 nftg)
1 1 1 1
0.5
1.5 2 2.5
Ca/S Ratio, mol
3.5
Figure 1. Comparison of FSI sulfur cSoxide removals for commercial
and HSA31 hydtate.
100
90
80
2
70
3
"5.
(0
60
0
«
50
X
0
•o
40
30
CO
20
10
0
- 1000-3000 ppm SOo
- 1200 C
^ 2QQ0
SO2 REDUCTION
REQUIREMENTS
(typical Irois coal)
1995~ ~
HSA3 (40 m^g)
HSA2 (60 m^g)
HSA1 (50 m^g)
WUELFRAsorp (40 rr^/g)
Commercial (22 nr^/g)
2
Ca/S. mol
Figire 2. Comparison of FSI sulfur dfoxide removals for ISQS HSA. commercial
and WUELFRAsorp hydrates.
7-23
-------
100
80
tSGS KSA30 (48 n#g)
Commercial (22 m?g)
3
1 1.5
Ca/S ratio, mol
2.5
Figure 3. Comparison of pSot-scale boier economizer siifur cioxkle
removals for commercial and HSA30 hydrates.
o
o
<
o
100
80
Flow reactor
. 3000 ppm SO-
2 sec
_ 540 C
eo -
40
20
CO^M HSA-2 BESTBC
HSA-1 HSA-3
SURFACE AREA nf/g
20 50 60 40 40
HYDRATED LMES
Figire4. ConTpariscn of boler economizer calcium conversions lor
tSGS USA, commercial and WUELFRAsorp hycfrates
7-24
-------
100
80
60
40
20
25 F approach to adabatic saturation
1500 ppm SO2
2000
¦ SCb REDUCTION
REQUIREMENTS
- (typical IBnois coaO
1995
-
/Mr Jr
¦
HSA2 (60 m^g)
fly*
•
~
HSA1 (50 m2/g)
HSA3 (40 m^g)
r 1
~
<
Commercial (22 m2/g)
1
1 2
Ca/S ratio, molar
Figure 5. Coolside sulfur dioxide removals fhumkffier + baghouse) for
commercial and ISGS HSA hycfraies.
7-25
-------
100
80
60
40
20
0
80
60
40
20
0
80
60
40
20
0
FSI
HSA31 (47 mfrg)
HSA1 (50 rrfrg)
HSA2 (60 m?/g)
HSA3 (40 m2/g)
WUELFRAsorp (40 m^g)
Commercial (22 n#g)
BOILER ECONOMIZER
j I i I ¦
COOLSIDE
I ¦ I
20 40 60
Surface area, m^g
80
Figure 6. Sulfur capture for hydrates as a function of strface area.
7-26
-------
EFFECTS OF SPRAY NOZZLE DESIGN
AND MEASUREMENT TECHNIQUES ON REPORTED DROP SIZE DATA
Wes Bartell
Jerry Ferrazza
Rudi Schick
Spraying Systems Co.
North Ave. at Schmale Rd.
Wheaton, IL 60189-7900
7-2?
-------
Intentionally Blank Page
-------
ABSTRACT
Accurate spray nozzle drop size data is a critical component in the field of flue gas
desulfurization (FGD). Nozzle manufacturers have worked independently
responding to drop size requirements as specified by scrubber designers. Typically, a
Volume Median Diameter drop size of 2500 microns is specified regardless of flow
rate, operating pressure, or nozzle design.
In order to accurately assess drop size performance for a given nozzle, the method
for testing should be fully understood. The type of instrument, method of collection,
manner of interpretation of raw data, and reporting techniques all have a strong
influence on the results.
This paper provides an overview of common drop size measurement techniques,
emphasizing the importance that nozzle design has on resulting drop size as well as
other factors such as capacity and pressure. Also, some of the more commonly used
methods and instruments used for drop size data collection and the importance of
repeatability of test results when comparing data from different sources is discussed.
Finally, a critical look at reporting techniques and their potential affect on drop size
data is presented.
7-29
-------
hH-fcClB OF SPRAY NOZZLE DESIGN
DISCUSSION
Lefs first examine measurement techniques. The American Society for Testing and
Materials (ASTM) recognizes two different types of drop size sampling techniques
known as spatial and temporal (also known as flux-sensitive).
The spatial technique (Le. spatial distribution) is implied when a collection of droplets
occupying a given volume are sampled instantaneously (see Fig. 1). Generally,
spatial measurements are collected with the aid of holographic means or high speed
photography. This type of measurement is sensitive to the number density in each
class size and the number of particles per unit volume.
The flux technique (i.e. flux distribution) is implied when individual droplets that
pass through the cross section of a sampling region are examined during a interval of
time. Flux measurements are generally collected by optical measurements that are
capable of sensing individual drops.
The flux distribution may be transformed to a spatial distribution by dividing the
number of samples in each class size by the average velocity of the drop in that size
class. If all drops in a spray are moving at the same velocity the flux and spatial
distribution are identical
However, the spray will generally exhibit differences in drop velocities that vary
from class size to class size. In addition, these differences depend on the type of
nozzle, capacity and spraying pressure.
7-30
-------
SPATIAL vs. FLUX SAMPUNG
SPATIAL
• AVERAGED OVER FINITE
VOLUME
• INSTANTANEOUS SAMPLE
• SENSITIVE TO RELATIVE NUMBER DENSITY N(d). PARTICLES/VOLUME
MEASUREMENT
VOLUME
• -U-Y
••
o
•
¦
FLUX (TEMPORAL)
• TIME AVERAGED
• SENSITIVE TO PARTICLE FUUX
MEASUREMENT"
CROSS-SECTION
fkmaylug 9j*mm Ofc 1991
(Reproduced with permission.)
Fig. 1
7-31
-------
Flux/Spatial Ratio vs. Pressure
tr 1.15-
1.1-
U_
1.05-¦
14
20
5
7
3
15
9
11
13
17
19
21
Pressure (pa)
4H-160 2HHSJ-901780 2BSJ-901780 -H- 4CF-150/90
Fig. 2
Fig. 2 clearly shows the relationships between the spatial and flux samples for the
nozzle listed at the indicated conditions. This data does not suggest that a
relationship between the two methods is fixed, or that a certain correction factor
could be developed.
The flux/spatial comparative drop size ratio relationship will vary with pressure and
approaches unity at a condition where all droplets would be traveling at the same
velocity. Therefore, it is important to combine measurement techniques and
equipment for measuring drop diameters, and for the spatial measurement technique,
to reconcile the drop size distribution by applying velocity correction values to each
class size in the distribution.
7-32
-------
In Example 1, while the same nozzle is vised, the spatial data without velocity
correction consistently indicates smaller drop sizes.
Example 1
4CF-150-90 Whirljet® Hollow Cone Nozzle,
150 GPM @ 7 psig, 90° Spray Angle
Pressure (PSIG) Flux Spatial
3 3910 3490
5 3470 3050
7 3150 2750
10 2820 2450
15 2520 2170
20 2410 2070
The selected range of the instrument will also effect the end results. In Example 2
the same nozzle was tested using ranges of 100 to 6200 microns and 200 to 12,400
microns. For the first range, droplets larger than 6200 microns were discarded.
While droplets larger than 6200 microns are few in number they will be a sizeable
percent of the volume of liquid sprayed. Comparing the spatial samples we see a
difference of 400 microns or 14.6 percent
Example 2
4CF-150-90 Whirljet®
VMD in Microns
Pressure (psig) Range 6200
Flux Spatial
7 2690 2340
Range 12,400
Flux Spatial
7 3150 2740
In this example, the difference between the 6200 upper bound spatial data (2340) and
the 12, 400 upper bound flux data (3150), is 810 mir -^ns or a 25.8 percent difference
for the same nozzle and pressure. In comparing nozzles of similar design, from
different manufacturers, unless the sampling technique and range are known,
evaluating the drop size performance of the two designs becomes subjective if not
invalid.
7-33
-------
With measurement techniques addressed, let's look at four common nozzle designs
used in FGD applications and review the importance of flow rate and pressure when
considering drop size.
Nozzle Selection
Nozzle
Description
Flow Rate
4H-160 Fulljet®
2HHSJ-901780 Spiraljet®
2BSJ-901780 Spiraljet*
4CF-150-90 Whirl Jet®
90° full cone spray pattern 160 GPM at 7 psi
90° full cone spray pattern 178 GPM at 10 psi
90° hollow cone spray pattern 178 GPM at 10 psi
90° hollow cone spray pattern 150 GPM at 7 psi
The above nozzles have the same spray angle but the flow rates are given at the
rated pressure established by the manufacturer for each nozzle design. Drop size
tests are often conducted at the rated pressure. If a 150 GPM flow rate is required
then three of the four nozzles must be operated at lower than rated pressures which
will result in larger drop sizes. A proper comparison can only be made if drop size
data is collected for all four nozzles at the same flow rate and pressure.
Figure 3 indicates that the four nozzle designs also produce four distinctly different
volumetric spray distribution patterns which influence the composite drop size
populations.
The Fulljet nozzle, in this case the 4H160, has an internal vane which causes a
swirling of the fluid, creating a cone shaped pattern- Typically this type of nozzle
design produces a full cone pattern with larger droplets, yet will show a fairly wide
overall drop size range.
The Spiraljet® nozzle (2HHSJ-901780) creates a pattern by shearing off portions of a
column of water as it passes through the inside of the spiraL This type of nozzle is
typically made with two deflective angles on the spiral which creates a cone within a
cone spray pattern with the two edges of the cones having the largest droplets.
The hollow cone Whirljet® nozzle (4CF-150-90) forms a spray by having the fluid
enter tangetially from an inlet into a circular chamber exiting through a smaller
concentric orifice. This method of spray formation can create a wider overall spray
pattern and potentially a smaller droplet size for the same flow rate and pressure.
The hollow cone Spiraljet® nozzle (2BSJ-901780) uses shearing foroes to turn the solid
column of water into a hollow cone spray pattern. Spray formation is similar to the
full cone Spiraljet except that only a single deflection angle is used.
7-34
-------
4H-160 FullJet*
2HHSJ—901780 Full
Cone SpIrolJet®
2BSJ—901780 Hollow
Cone SpIralJet*
4CF-150-90
WhirlJei*
-------
VMD Data for 12400
1
Micron Range
Pressure
Flow Rate
(PSIG)
GPM
FLUX
SPATIAL
4H-160 Fulljet®
3
106
4750
4090
5
137
4180
3630
7
160
3880
3350
10
189
3600
3080
15
225
3330
2790
1 20
260
3160
2610
| 2HHSJ-901780 Full Cone Spiraljet®
3
98
4330
3900
5
127
3880
3420
7
150
3600
3150
10
178
3340
2900
15
218
3060
2640
20
252
2870
2480
2BSJ-901780 Hollow Cone Spiraljet*
3
98
4100
3700
5
127
3590
3250
7
150
3300
2980
10
178
3020
2720
15
218
2730
2460
20
252
2550
2280
4CF-150-90 Whirljet®
3
98
3910
3490
5
127
3470
3050
7
150
3150
2750
10
179
2820
2450
15
220
2520
2170
20
254
2410
2070
Fig. 4
Figure 4 further illustrates how design can effect drop size. Note that each spray
pattern provides a range of drop sizes. This range varies with spray nozzle type,
capacity, spraying pressure, and to a lesser extent, spray angle.
7-36
-------
Now, lets look at some of the more commonly available methods and instruments
available for drop size data collection. Since repeatable test results are essential in
comparing nozzle performance data, it is essential to use testing procedures that take
into account all die potential variables in die sampling technique for both methods
and instrumentation.
By spraying water into a pan of oil and shuttering the spray, it is possible to count
and size individual droplets using a microscope. This technique is still used by some
researchers. Problems involve drop coalescence, inadequate sample size and the fact
that very small droplets will be deflected away from the oil by air currents at the
surface due to the spray velocity. Also, larger droplets can and do break up from
impacting the surface.
The same type of method is used when spraying a dye onto a stationary card or
water on to liquid sensitive paper. Again, the small drops might be deflected away
from the target and the large drops can break-up from the impact Data collected by
these "intrusive" methods depends on a number of uncontrolled variables making
such test results generally non-repeatable.
While drop size data was being collected in the early 1950's using methods such as
flash photography, probably the first real breakthrough in droplet sizing technology
was die development in 1961 of an automated imaging analyzer (Fig. 5).1
Basically, the Electronic Imaging Analyzer incorporates the spatial measurement
technique using a strobe light to illuminate the spray and record the image with a
vidicon tube. The image is scanned and the droplets are sized and separated into
different classes. Resulting data can be mathematically corrected using velocity data
to give a flux distribution. Sources of error early in the development of this device
included blurring, depth of field variations, and vidicon tube saturation. These
sources were recognized and corrected.
The imaging type analyzer is still actively promoted by some nozzle manufacturers.
The limited availability of this type of instrument, however, prevents independent
researchers and other interested members of the drop size analyzer community from
verifying data arrived at from a particular test or comparing performance from
similar designs.
More recently the development of commercially available drop size analyzers makes
it feasible to verify drop size results by independent sources. This new breed of
analyzers incorporates lasers, special optics and digital circuitry to minimize imaging
error. Some of the most commonly recognized manufacturers of laser measurement
instruments include Malvern, Particle Measuring Systems (PMS), and Aerometrics.
1The Electronic Imaging Analyzer was developed at Spraying Systems Co. by Dr.
Verne Dietrich and built by the Dage Division of TRW, Michigan City, Indiana. The
design was awarded U.S. Patent 3275733 in September of 1966 and is currently in its
second generation.
7-37
-------
SLOT
OPTICS ,
TELEVISION CAMERA—. /
, L lOItT SOURCE
j (STROBE)
MONITOR
CONSOLE
SPRAY ANALYSER
Spraying Systems Co.
Spray Nozzlas and Accessor!**
North Amu* at tdimli ltd. • ML Sax 7SOO
WkMton. U. 00180-7000
Ret:
Revision No.
Data ShMt No.
11833-1
»»¦!>< HH» i 6*. 1
(Reproduced -with permission.)
Fig. 5
7-38
-------
COMPONENTS OF THE MALVERN
PARTICLE ANALYZER
TRANSMITTER
SPRAY
RECOVER
n
AMPLIFIER
A/D CONVERTER
PRWTER COMPUTER
Reprinted by permission of
Malvern Instruments, Inc.
Southborough, Massachusetts
Fig. 6
The Malvern Analyzer which is considered a spatial sampling device, utilizes the
principle that a spray droplet will cause laser light to scatter (Fig. 6) through an angle
dependent on the diameter of the droplet The scattered light intensity is measured
using a series of semicircular photo diodes. Theoretically, the distance of the
individual photo diodes from the centerline of the laser and the light intensity
functions are all that are needed to calculate the droplet size distribution. A curve-
fitting program is used to convert the light intensity distribution into any of several
empirical drop size distribution functions. Since the Malvern has some self
diagnostics, potential sources of error are easier to identify. The instrument must be
aligned and calibrated periodically using reticle slides with known etched drop
distributions.
7-39
-------
Perhaps the biggest source for error with this type of instrument is multiple light
scattering. If the spray is too dense, there is a possibility that the scattered light from
one droplet might be scattered again by other drops further down the beam axis.
The Malvem is equipped with a "obscuration level" indicator which can be used to
determine if the spray is too dense, but such a determination is often difficult To
circumvent this in the lab the technician typically moves the nozzle farther away or
uses special shielding to permit only a portion of the spray to enter the sample area.
Particle Measuring Systems, also known as PMS, produces both diffraction type
instruments and imaging type instruments known as Optical Array Probes. The PMS
Optical Array Probe (Fig. 7) is a flux sampling instrument. As the droplets pass
through the sampling plane, the droplets are sized and counted providing
information which can be used to determine velocity. The two dimensional grey
scale OAP can provide droplet measurement in two ranges, 100 to 6200 microns and
200 to 12,400 microns, and is currently the most sophisticated offered by PMS.
The PMS OAP Grey Scale probes applies advanced technology utilizing extensive self
diagnostics. These probes will reject drop images which are out of focus or which do
not meet a series of other acceptability tests automatically. Problems with PMS units
usually center on improper calibration or maintenance. The optics tend to get wet
easily, and cleaning and alignment require some skilL Also, dense sprays tend to
overload the circuitry and sample area reductions are often necessary. Sample area
correction factors and drop distribution curve fitting equations are needed and are
left up to the operator to include in the analysis.
The Aerometrics Phase Doppler Particle Analyzer or PDPA (Fig. 8) is a point
sampling device and is a flux-sensitive instrument Point sampling refers to an
instrument that focuses on a portion of the total spray pattern and requires targeting
several test points within the spray in order to obtain a composite sample of the
spray flux distribution.
The PDPA uses a low power laser that is split into two beams by utilizing a beam
splitter and a frequency module. The two laser beams intersect again into a single
beam at the sample volume location. When a drop passes through the intersection
region of the two laser beams, an interference fringe pattern is formed by the
scattered light Since the drop is moving, the scattered interference pattern sweeps
past the receiver aperture at the Doppler difference frequency which is proportional
to the drop velocity. The spatial frequency of the fringe pattern is inversely
proportional to the drop diameter.
Aerometrics offers an optional fibre optic probe which isolates the instrument from
the spray and eliminates the potential for error due to vibration caused by direct
contact with larger capacity sprays.
Other drop sizing instruments which are commercially available, generally use lasers
and operate on principles which we've previously discussed.
7-40
-------
OPTICAL SYSTEM DIAGRAM FOR AN OPTICAL
ARRAY SPECTROMETER PMS MODEL
0AP-2D GA2
SPRAY NOZZLE
OBJECTIVE
60 mm
45° MIRROR
43" MIRROR
PHOTOUODES
SECONDARY
ZOOM LENS
PRINTER COMPUTER
PHOTOOIOOE ARRAY
Reprinted by permission of
Particle Measuring Systems, Inc.
Boulder, Colorado
Fig. 7
7-41
-------
COMPONENTS OF THE AEROMETRICS PDPA
MEASUREMENT VOLUME
SCATTERED LIGHT INTERFERENCE PATTERN
OCT
nLTERED DOPMER BURST SIGNALS
Fig-8
Reprinted by permission of
Aerometrics Inc.,
Sunnyvale, California
7-42
-------
In the application of these instruments if small regions of the spray are examined, as
with the PMS or Aerometrics, it is important to test several points within the spray in
order to obtain a combined result which is representative. Ideally, the combined
drop size distribution will also be weighted relative to the volumetric flow at the
various sampling points. This step is critical and is often ignored. In some
situations, the effect is small, as with air atomizing or Whirljet® type nozzles. In
other nozzle designs, improper weighting can affect the results by as much as 50%.
The distance that the nozzle is located from the sample region and the orientation
also have an effect on the data collected. Smaller droplets will be entrained into the
air flow pattern created by the nozzle and will show up in increasing concentrations
as the distance from the nozzle increases in the regions of higher entrained air flow.
Also, very small drops may evaporate completely before they get to the sampling
region. Therefore, in order to provide accurate comparisons between nozzles of
different design each nozzle test should be conducted at severed locations reflecting
the spray flux distribution and normalized to reflect the whole drop size population.
However, a standardized procedure for obtaining composite data has not been
established at the present time.
All of the issues mentioned so far can lead to differences in the data obtained, even if
the instrumentation is in perfect working order. However, proper calibration and
maintenance of the measuring equipment can't be overlooked. A great potential for
error exists when everything seems to be working fine but something is just slightly
out of adjustment Properly scheduled calibration tests are important, particularly in
labs where many people use the equipment
While instrumentation issues are of great importance, the reporting method is the
greatest potential source of bias for drop size data. This directly effects those
responsible for evaluating nozzle performance and making recommendations. There
are a number of formats used for reporting drop size data. When evaluating data,
particularly from different sources, it is extremely important to know the type of
instrument and range used, the sampling technique, and the percent volume for each
size class in order to make for valid data comparisons.
Following the procedure as outlined by the ASTM Standard E799 (Eg. 9) provides an
exact method used for collection and reporting. In this example the drop size data is
tabulated using the ASTM standard, and an empirical (Rosm-Rammler) curve fitting
formula is then applied to further classify each drop size as a percent of the volume.
To verify if the correct instrument range was used the ASTM guidelines state that the
largest drop size reported should make up less than one percent of the volume of the
spray flux. Looking again at Fig. 9, under percent volume undersize, we see that the
largest drop size reported was 5333 microns or 0.2 percent of the volume. This
verifies that a large enough collection range was vised and assures the accuracy of the
test
7-43
-------
REPRESENTATIVE DROP SIZES 4c DISTRIBUTION
4CF-150/90 Whirljet
20PSI
07-25-1990
-Dropsize Analyzer: PMS-OAP-2D-GA2 (12400 pm max)
-Sampling Method: Flux (TEMPORAL)
-All values computed utilizing the procedures for determining spray
characteristics as outlined by ASTM (standard E799).
UPPER BOUND = 5332.65 (pm)
LOWER BOUND = 317.46 (jim)
(BOUNDED CURVE)
DROP
PERCENT VOLUME
PERCENT COUNT
DIAMETER
UNDERSIZE
UNDERSIZE
(pm)
317
0.20
000
368
031
838
427
0.47
16.95
496
0.71
25.13
575
1.08
33.10
667
1.65
40JB6
774
230
4839
898
3.80
55.68
1041
5.74
62.69
1208
8.63
6937
1401
1Z87
75.66
1626
18.98
81.46
1886
27.48
8&64
2188
38.77
9L06
2538
52.72
9456
2944
68.13
97.06
3416
8255
98 57
3963
93.05
99.28
4597
98.29
9951
5333
99.80
9955
TOTAL CURVE BOUNDED CURVE
(|im)
(um)
(ARITHMETIC MEAN)
Djo1
0
1039
(SURFACE MEAN)
Djo-"
0
1263
(VOLUME MEAN)
D30:
0
1484
(SURFACE/LINEAR MEAN)
Pa:
1294
1536
(-EVAPORATIVE- MEAN)
1619
1774
(SAUTERMEAN)
2026
2050
(DeBROUGERE [HERDAN] MEAN)
Dc:
2503
2514
(VOLUME MEDIAN DIAMETER)
Dyoj:
2470
(NUMBER MEDIAN DIAMETER)
^N05:
804
[BOUNDED]
(DIAMETER AT Max dVOLUME/dDIAMETER)
®VOL MODE"
2414
(RELATIVE SPAN [(Do^Do^/DvqJ)
RSF:
14)075
(COEFF. OF VARIANCE l(Dv05/DNIo.s]>
CV:
34)737
Fig. 9
7-44
-------
BETE OROPLET ANALYSIS STSTEM
COMPOStIE REPORT NOZZLE: 3 FvT 4460120
TEST OATE 4/l9/8«
TESTS INCLUOEO: 1247 1248 1249
PRESSURE 10.0 PS1
CEMTERLINE OIST 60.00 INCHES
AZ. ANGLE SO.O OSG
CTL. ANGLE 60 - 300 DEG
OIAHETER
(MICRONS)
12S.9
-
IS8.S
158.S
-
199.S
199.5
-
251.2
251.2
-
316.2
316.2
-
398.1
398.1
-
501.2
501.2
-
631.0
631.0
-
794.3
794.3
-
1000.0
1000.0
-
1258.9
1258.9
-
1584.9
1584.9
-
1995.3
1995.3
-
2511.9
2511.9
-
3162.3
3162.3
-
3981.1
3981.1
-
5011.9
5011.9
-
6309.6
6309.6
-
7943.3
X
OROPS
OCCURRANCE
529
4.33
1013
6.30
1639
13.42
1457
11.93
IS88
13.00
1233
10.10
134S
11.01
1206
9.88
858
7.03
539
4.41
363
2.97
208
1.70
137
1.12
SO
0.41
27
0.22
13
0.1.1
5
0.04
2
0.02
12212. 100-00
SURFACE
X
AREA
VOLUME
0.20
0.02
0.58
0.07
1.38
C.19
1.90
0.32
3.27
0.70
3.93
1.05
6.59
2.19
9.S3
4.02
10.69
S.66
10.67
7.12
11.23
9.37
10.36
10.95
10.64
14.05
6.09
10.09
S.43
11.54
3.76
9.61
2.31
7.41
1.42
S.66
100.00 190.00
CUK X
CLASS
VOLUME
CHECt
0.02
0.000
0.08
0.000
0.27
0.000
0.60
0.000
1.30
0.001
2.35
0.001
4.S4
0.003
6.56
0.005
14.22
0.006
21.33
0.008
30.70
0.011
41.64
0.013
55.69
0.016
6S.76
0.012
77.32
0.013
86.93
0.011
94.34
0.008
100.00
0.006
100.00
AVERAGE DIAMETERS (MICRONS} :
ARITHMETIC MEAN * 547.61
SURFACE MEAN - 722.28
VOLUME MEAN x 9S9.87
SAUltR MEAN . 1695.21
WIGHT MEAN . 2734.
VOILHL MEDIAN « 2306.22
SAMPLE SIZE CHtCK = 0.03
MAXIMUM DIAMETER - 6910.21
MINIMUM OIAMETER « 155.73
TOTAL OROPS IK SAMPLE « 12212.
TOTAL OUT Of FOCUS « . 1824
TOTAL FRANCS IN SttCLf « 600
AVL. UWOPS rtR I KAMI j iXJ.'Jt,
OCVIAIION * 0.76
RUAIIVi SPAN « ( S52G.00 - b*j2.4tf)/ ^300.22 « 2.03
Fig. 10
with P—
riS^^.S,OCiati°° °f C—E»gi=«rsmHousTo"U^ ^"1^'
-------
In (Fig. 10)2 while a similar format was used, the type of analyzer and range are not
reported. Examining the percent volume, if the largest drop size class makes up a
larger percentage of the total volume the upper limit of measurement may be too
low. Therefore, a VMD droplet size which is substantially smaller than is actually
occurring in the spray flux may be reported.
When using a curve to report drop size data such as volume median diameter (Dto5)
vs. pressure, enough data points should be reported when the curve is plotted to aid
meaningful interpolation of data between data points. Since performance variations
between different pressures are not necessarily linear, comparing Figure 11A and 11B,
a small difference in pressure may produce a significant change in trending the
median drop size. In any case such a graph should be used to supplement a more
complete reporting format, not as a substitute.
Volume Median Diameter vs. Pressure
5000
® 4600"
® 4400"
O 4200-
£ 4000-+
£ 2400-
O 2200-
Pressure (psi)
4H-160
2HHSJ-901780-
2BSJ-901780
4CF-150/90
Fig. 11A
7-46
-------
Volume Median Diameter vs. Pressure
5000
4000" •
El
1800-1—T
Pressure (psi)
4H-160
2HHSJ-901780
2BSJ-901780
4CF-150/90
Kg. 11B
7-47
-------
CONCLUSION
Providing accurate drop size data is the responsibility of the nozzle manufacturer. It
is also apparent with the wide range of nozzle designs and capacities that no single
instrument or technique is universal. Accurate data collection requires a dedicated
facility which uses commercially available instruments and offers the potential for
verifying test results. Most of all, a trained staff familiar with the latest calibration
and collecting techniques is essential for reporting the results of any test
REFERENCES
ASTM, "Standard Practice for Determining Data Criteria ar.d Processing for Liquid
Drop Size Analysis" Designation E799-81 (1981)
Bete, David L. Soule, Lincoln S. and Walker William F.; Solving Corrosion Problems
in Air Pollution Equipment Paper No. 19. The Evolution of Durable, Efficient, Non-
Plugging Slurry Spray Nozzles for Utility FGD Systems" National Association of
Corrosion Engineers (1984)
Dodge, Lee G. "Comparison of Drop Size Measurement for Similar Atomizer" Report
No. SWRI-885812, Southwest Research Institute (1986)
Dietrich, V.E., "Drop Size Distributions For Various Types of Nozzles", The 1st
International Conference on Liquid Atomization, ICLASS Proceedings, Tokyo (1978)
Hebden, W. E., Shah A.M., "Spray Droplets Analysis By High Speed Strobe
Photography", The 1st International Conference on Liquid Atomization, ICLASS
Proceedings, Tokyo (1978)
Saleem, Abdus "Spray Tower: The workhorse of flue-gas desulfurization" Power
(1980)
"A Survey of Spray Technology for Research and Development Engineers" Bulletin
No. 65 Penn State U. (1956)
Fulljet®, Spiraljet® and Whirljet® are registered Trademarks of the Spraying Systems
Co., Wheator., Illinois.
7-48
-------
High SO2 Removals with a New Duct Injection Process
7-49
-------
Intentionally Blank Page
-------
s. Nelson Jr.
Sanitech Inc.
1935 E. Aurora Road
Twinsburg, Ohio 44087
ABSTRACT
Sanitech has developed a new BOrbent material specifically formulated for
retrofit duct-injection applications. These line—based sorbents are tiny enough
to be entrained in the flue gas stream, yet large enough to be easily removed
once saturated with sulfur dioxide. The sorbents distribute their reactive lime
species over the large internal surface area of an inexpensive, sponge-like
mineral support, which carries with it the optimum amount of moisture needed for
acid-gas removal. Laboratory runs suggest the potential for greater than 90% SO2
removal and 60% net lime hydrate utilizations for duct-injection retrofits with a
small, high-ratio baghouse, enabling sorbent recycle. Plans are now being made
for scaled-up demonstrations.
These "sponge-sorbentb" hold particular promise for economical application at
many older, moderately-utilized, power plants and those plants with existing
particulate-control problems or potential toxic-emission problems.
7-51
-------
INTRODUCTION
The initial promise of sorbent injection technologies, both furnace-injection and
duct—injection, was that they would cost very little to install on coal-burning
power plants. In practice, however, weak sorbent performance has plagued these
technologies, with unassisted S02 removals of typically 40% to 60%; and low
sorbent utilizations, of from 20% to 40%. Moreover, particulate-control problems
have resulted. A' removal efficiency of only 50% will not allow a plant to burn
medium- or high-sulfur coal and still meet the 1.2 lb S02/MM Btu requirement of
the 1990 Clean Air Act. Utilizations of only 30% mean that net sorbent costs
will be high and that huge amounts of unused reactants, mixed-in with ESP fly
ash, will require waste disposal.
Sanitech has developed a new duct-injection sorbent specifically formulated to
solve these problems. For over four years, the company has been scaling-up a
dry, regenerable magnesia technology which utilizes expanded industrial
minerals — granular vermiculite or perlite — as reactant supports. The company
has recently discovered how to endow less expensive calcium-based supported-
sorbents with the same high reactivity and performance as magnesia and do so with
even cheaper grades of the support materials. While not regenerable, the calcium
sorbenta are recyclable.
These proprietary sorbents are made from very fine, inexpensive, almost waste-
grades of vermiculite or perlite. The tiny sorbents, called "sponge-borbents,"
can be injected into a humidified flue gas stream after the electrostatic
precipitator and then be captured on a fabric filter, where additional SO, removal
takes place, much like the Electric Power Research Institute's HYP AS concept.1
The sponge—norbents are hundreds times smaller than Sanitech's regenerable,
magnesia-based, panel-bed "Kag*Sorbents,"J yet they are still over a thousand
times larger than typical fly ash or lime inject;ant particles. Consequently, a
very small baghouse can be used for collection — a "microbaghouse" — which
operates at a high air-to-cloth ratio (A:C). With this extra, but lower-cost,
particulate collection device in line, the plant can avoid an otherwise-needed
ESP-upgrade, or could incorporate a furnace-injection or second duct-injection
front-end process as well.
7-52
-------
SPONGE—SORBENT CHARACTERISTICS
The Interrelated keys to the performance of sponge—sorbenta are their size,
moisture capacity, and reeyclabi.li.ty.
Sorbent Size and Shape
Fine sponge—sorbents are simple to prepare. Very fine vermiculite or perlite,
such as #5 vermiculite ore, is thermally expanded in the traditional, automated
manner, where it increases in volume by a factor of about ten. Vermiculite
expands like an accordion, exposing hundreds of thin mica plates; perlite pops
like an exploding onion, similarly exposing a high internal macro-surface area.
These expansions allow for a very high loading of reactive species, in this case
hydrated lime, with a resulting sorbent composition of approximately 60 wt%
Ca(OH)j. The material is then thermally activated and moistened before use. See
the process flow diagram in Figure 1.
The key to the sorbents' application and low capital cost requirements is their
size. The sorbents are about 0.20 millimeters on a side: small and light enough
so that they can be entrained in the gas stream, yet large enough so that they
present no problems in being removed from the gas once saturated. Fly ash and
traditional lime injectants are thousands of times smaller than the sorbents.
Fly ash frequently has a mass mean diameter of from 10 to 20 microns; hydrated
lime particles, injected dry or as a slurry, generally average 5 microns in
diameter. Particulate removal efficiency and cost are usually inversely
proportional to particulate size. It is much easier and cheaper to separate a
single, saturated 0.20 mm diameter sponge-sorbent from the gas stream than 40,000
equivalent, but tiny, 5-micron lime particles.
Because sponge-sorbents are so large, it may be possible to use fabric filters
that are less than one-fourth the size of those required to filter out the usual
5-micron particles. Such high-ratio baghouses have been used historically to
capture sawdust or perlite particles, which, like sponge-sorbents, are large and
porous and result in low pressure drops. The Electric Power Research Institute
is looking at high-ratio pulse-jet baghouses as retrofit back-ups to aging,
marginal ESPs. They may also see use with furnace sorbent injection retrofits.
Traditional baghouses, or increases in ESP specific collector area (SCA), are
major capital cost items. Consequently, if practicable, a high-ratio baghouse
could cut the capital cost of a sorbent-injection retrofit by 70% or more.
Because many of the fines that pass through the existing ESP will impinge upon
and be captured by the sponge-sorbents, this device may also help aging, low-SCA
ESPs meet tighter PSD, toxics, or PM-10 particulate emission regulations. In
preliminary microbaghouse trials, Sanitech has seen significant fine-particulate
captured by the sponge-sorbents, although this ability has yet to be quantified.
7-53
-------
Moisture Capacity
The key to sorbent performance is water. It has long been known that the higher
the amount of evaporated water in the flue gas, that is, the lower its approach
to adiabatic saturation, the better the lime sorbent utilization and SO, removal.
A high interaction between sorbent and injected humidification water has also
been shown to result in better performance. Scavenging, that is injecting dry
hydrated lime before humidification sprays to maximize water droplet-lime
collisions, results in greater performance than lime injection after
humidification, although utilizations are still low.3 Injecting lime and
humidification water together in a liquid slurry, of roughly 5 to 20% solids,
produces even better results: about 60% SO. removal with 35% utilization,4
although in this case the excess water causes the calcium hydrate crystals to
agglomerate, decreasing the ultimate performance.
Because of their sponge-like mineral supports, the new sorbents can be heavily
loaded with water, yet remain free-flowing and only moist to the touch. When
exposed to the gas stream, they carry their water tightly with them, right along
with the widely-exposed lime. The optimum quantity of water is exactly where it
is needed. As the water evaporates into the flue gas from the sponge supports,
the sorbents are preferentially cooled, dropping the wet-bulb temperature in the
micro-vicinity of the sorbents and accelerating S02 sorption. Because much of the
necessary water for gas cooling is evaporated from the moist sorbents themselves,
the humidification requirement of the flue gas stream is significantly reduced.
Reeve1ab11itv
The key to sponge-sorbent economics is their ability to be recycled. Fine lime
particles injected into flue gas streams typically form a sulfated shell around
an unutilized core that renders them largely useless for further sulfur capture.
Sponge-sorbents, on the other hand, can be recycled by simply remoistening and
reinjecting them into the gas Btream until their ultimate utilization is reached.
Because sorbent injection is after the ESP, flyash is not recycled along with the
sorbents.
In practice, a significant amount of the baghouse-collected sorbent would be
combined with a small amount of fresh, make-up sorbent, and re-entrained. The
spent portion of the collected sorbent can be abraded back into its two
constituent components. The dry, well-saturated lime sulfate fraction would go
to disposal, as is customary. Because it is twice as utilized as typical lime
injectants, however, only half the waste material is produced. The low-grade
mineral supports can be either recycled back to be made into new sponge-sorbents
or used in agricultural soil-conditioning or reclamation markets. If necessary,
they can be easily compressed for more compact disposal. Thus, this new SO,
removal process requires no water processing and has no other waste streams.
7-54
-------
RECENT IN—DUCT LABORATORY TESTING
Sanitech recently initiated in-duct testing of the sponge-Borbents. These trials
were carried out during three days of testing at the U.S. Environmental
Protection Agency's 40-acfm duct-injection facility at the Aix and Energy
Engineering Laboratory in Research Triangle Park, N.C. This facility, which has
been used in the development of the ADVACATE process, has been described
elsewhere.1 Acurex Corporation, the site contractor, ran the trials.
The baghouse at this facility, unfortunately, was sized at a typical air-to-cloth
ratio of approximately 3 feet-per-minute. This low face velocity was only
sufficient to keep about one-half of the sponge-sorbent material on the bags.
Half of the sponge-sorbents' reactive species immediately dropped to the baghouse
hopper, becoming unavailable for SOa removal. Nonetheless, the results achieved
in these runs were very promising.
Most important were the recycle runs, which simulated expected utility practice.
Partially-saturated sorbent from previous runs was backmixed with a small amount
of fresh sorbent, at a 4:1 weight ratio, remoistened, and reinjected into the gas
stream. The recycle ratio used, on a total reinjected-solids basis, was less
than that of the ADVACATE process. The saturation of the backmixed sorbent,
however, was a little less than a true simulation of steady-state. The inlet
concentration of S02 was 1500 ppro and the baghouse inlet was maintained at a 20°F
approach to adiabatic saturation.
The recycle results are shown in Figure 2. The stoichiometric injection ratio
represents the fresh lime that must be purchased per cycle — the important
determinant for cost calculations. Particularly noteworthy was the 88 percent
SO] removal observed at a fresh-lime-consumption stoichiometry of only 1.5:1.
This translates to a sorbent utilization of 59 percent.
These results were promising because:
• Due to the limited testing time, the tests were run without a prior
parametric determination of the optimum operating conditions for the new
sorbents in this system;
• Half of the sorbents' reactivity was not available for S02 removal in this
baghouse, as it would with a custom-designed, high-ratio baghouse; and
• Higher removal rates for the ADVACATE process have been observed when it
has been tested at larger scales than the lisited-mass-transfer, 2-inch-
diameter-duct apparatus used in these runs.4 Better performance might be
similarly extrapolated when sponge-sorbent injection is scaled up.
7-55
-------
ECONOMICS AND MARKET APPLICATIONS
Initial results suggest a potentially low-cost process. By boosting lime
utilization, the fine, inexpensive vermiculite supports more than pay for
themselves, while providing a dry, retrofitable method of 90-percent SO] removal.
Early cost estimates of sponge-sorbent retrofit costs range from about $300 to
$650 per ton of S02 removed, depending on the assumptions. If the fabric filter
costs are not attributed to S02 control, the cost is significantly less. Costs
are most sensitive to the plant operating factor, whether the use of a high-ratio
microbaghouse is possible, and the possibility of support reuse and resale. As
this technology is further developed and scaled-up, more thorough cost analyses
will be made.
An obvious market niche for this technology are those plants that are already
experiencing particulate control problems. To the extent that the added fabric
filter used for sponge-sorbents recycle avoids the cost of an electrostatic
precipitator replacement or upgrade, the sulfur-removal costs are largely only
for the sorbent variable costs. This technology may also make sense as a
polishing filter with another SOj-removal process or for units where toxic
emission control are needed.
FUTURE EFFORTS
Development efforts have continued since the small-scale tests. Plans are being
made for megawatt-scale pilot plant runs and possible host sites are being sought
for an eventual full-scale demonstration.
REFERENCES
1. R. Carr, et al. "Pilot-Scale Evaluation of the HYPAS SOj and Particulate
Removal Process," Joint EPRI/EPA 1990 SO, Control Symposium, New Orleans,
Louisiana, May, 1990.
2. S. Nelson. "Field Testing of a New SOs and NOx Sorbent," Joint EPRI/EPA
1990 SO2 Control Symposium, New Orleans, Louisiana, May, 1990.
3. W. O'Dowd. "Duct Injection Experiments at DOE-PETC," Technical Opdate No.
18, Duct Injection Technology Development Program, Jan/Febr 1991.
4. M. Klett & L. Felix. "Beverly DITF Test Results," presented at Fourth Duct
Injection Technology Working Croup Meeting, May 7, 1991.
5. W. Jozewicz, et al. "Development and Pilot Plant Evaluation of Silica-
Enhanced Lime Sorbents for Dry Flue Gas Desulfurization," Journal of the
Air Pollution Control Association, p.796, June, 1988.
6. B. Hall, et al. "Current Status of ADVACATE Process for Flue Gas
Desulfurization," Air & Haste Management Association Annual Meeting,
Vancouver, British Columbia, June, 1991.
7-56
-------
~OMDtr-
ICATI9*
1.0 sm
Figure 1. Sponge-Sorbent Flowsheet
100
80
± 60
c
W 40
20
RECYCLE RUNS
baghouse outlet x'
1500 ppm SO,
4:1 recyde
20F approach
0.5 i :.5
Injection Ratio of Fresh Lime
Figure 2. Sponge-Sorbent Performance
7-57
-------
Intentionally Blank Page
7-58
-------
Combined SOx/NOx Control Via Soxal™,
A Regenerative Sodium Based Scrubbing System
7-59
-------
Intentionally Blank Page
7-60
-------
Carolyn Byszewski
David Hurwitz
AQGMECH Systems
7 Powder Horn Drive
Warren, New Jersey 07059-5191
ABSTRACT
AQUATECH™ Systems, a business unit of Allied-Signal Inc.,
proposes to demonstrate the technical viability and cost
effectiveness of the SGXAL ccntoined SGx/NQx: ocntrol
technology en a 2-3 1W equivalent flue gas slipstream frtxi
Niagara Mchawk Power Corporation's Dunkirk No. 4 steam
Station. The No. 4 Boiler is a 100 MW coal-fired utility
boiler currently burning coal. Tt>is demonstration will be
funded primarily by the US Department of Energy-PEIC and
Allied-Signal Inc. with in-kind ccaitribufcicros both frcm
Niagara Mcha&Sc Power Corporation and cofunded by NYSERDA. and
ESEER00. Hie SCKAL process rarhines high efficiency
sulfur dioocide removal and improved NQx removal frcm flue
gas using a ryriiiTrn based scrubber solution and regeneration
of the spent scrubber liquor using AQUAIECH Systems'
proprietary bipolar membrane technology. The SCKAL process
is applicable to both utility and industrial scale boilers
using either high, or low sulfur coal.
-------
INIRDDOCTIGN
Oconnercially available flue gas 'treating options fear sulfur
dicoade removal inclnrie both throwaway and regenerative
processes. The limestone scrubbing process is the most
ccBsnonly used throwauay process; however, it is cumbersome
to operate because of the tendency of the scrubbing solution
to scale in prvmess equipment. In addition, large amounts
of limestone reagent are cxre=aimpd and the calcium containing
sludge formed in the sulfur removal reactions requires
extensive lanflfin area for disposal.
Double aivaii processes have been developed to address the
operational concerns of limestone scrubbing systems. In
these processes the flue gas is scrubbed with a soluble
sodium based salt solution. Lime or limestone is then used
to convert the spent sodium bisulfite solution to calcium
sulfite and sulfate for disposal while regenerating the
sodium scrubbing solution far recycle to the flue gas
scrubber. The use of these sodium based systems has greatly
improved scrubber reliability; however, large amounts of
lime or limestone reagent are ccnsumed, sodium reagent
losses are higher than anticipated, and substantial amounts
of calcium containing sludge are generated that require
landfill disposal.
Hie Wellman-Iord process combines the operational advantages
of sodium based flue gas scrubbing processes with a thermal
deocHnpositian step to regenerate the t^aeiit sodium bisulfite
solution for recycle to the flue gas scrubber *ftiile
producing a salable sulfur byproduct. The sodium sulfate
formed in the system, hewever, cannot be easily regenerated
and results in a net waste bLuaam even in the best operated
units. The formation of sulfate also restricts the use of
the WeUmanr-Iord process to those boilers using high sulfur
ooal vhere oxidation in the scrubber is less extensive.
Perhaps the greatest disadvantages of the Wellman-Lard
process are its high capital and operating costs relative to
other flue gas treating systems.
7-62
-------
Hie SOXAL process is an extremely premising flue gas
treating system that effectively avoids the operating and
eooncmic disadvantages of the competitive flue gas treating
systems. It uses highly effective sodium sulfite scrubbing
solution to remove greater than 90% of the sulfur dicoa.de
contaminant in the flue gas with a unrrinnun of operating
problems. Hie scrubbing solution is regenerated for recycle
in a relatively sinple electrochemical process using a
patented bipolar membrane system. Hie first stage membrane
stack regenerates the sodium bisulfite solution, while the
second stage membrane stack regenerates the sodium sulfate
as caustic and dilute sulfuric acid. Hie first stage
regeneration step recovers a oonoentrated stream of sulfur
dioxide suitable for subsequent processing to salable sulfur
or sulfuric acid.
Coupled with conventional urea/methanol injection techniques
for combustion zone NOx control, the sodium scrubber can
further reduce NOx emissions up to 90%. Hie urea injection
step reduces NO to N, gas and NO_ gases. Methanol injection
will act to oxidize remaining NO or NR. slip gases to NO_.
Hiese NO_ gases will either be reduced to N_ gas in tne
sodium ailfite scrubber or be scrubbed fanning sodium
nitrate in the scrubber solution. Hie nitrates will
ultimately be purged frcm the system along with the sodium
sulfate stripper bottoms or the dilute sulfuric acid ^
from the second stage membrane stack: Hie optional second
stage membrane stack is not part of this demonstration.
Hie SOXAL process is nearly closed-loop? wmrmai net wastes
are generated. Hie operation is simple and reliable with
relatively lew operating and capital costs. Hie process is
compact and can be readily retrofit to any sodium scrubber
system.
It is applicable to both large utility and smaller
industrial boilers using either high or low sulfur coal and
offers combined SOx/NOx control in a single system .
7-63
-------
rKDCESS EESCKEPnCN
The SOXAL Process (Figure 1) carbines sodium based flue gas
scrubbing techniques with regeneration of the spent
scrubbing solution using AQUMCECH Systems bipolar membrane
technology. Hie regenerated solution is recycled to the
flue gas scrubber. Hie concentrated sulfur dioxide stream
recovered from the spent scrubbing soluticn is converted to
salable sulfur via Allied-Signal's Catalytic Sulfur
Reduction Technology or sulfuric acid losing conventional
canmercial processes.
The flue gas scrubber uses a soluble sodium sulfite to
remove greater than 90% of the sulfur dioxide in a staged
spray absorber tower. Sodium carbonate or sodium hydroxide
replace the small amount of sodium lost in the process. The
high solubility of the sodium sulfite allows for the
efficient absorption of sulfur dioxide at relatively low
liquid to gas ratios without excessive scale formation
problems.
The makeup sodium carbonate and recycled scrubbing solution
are added directly to the flue gas scrubber where they react
with dissolved sulfur dioxide to form sodium sulfite as
shown belcw:
Na2C03 + SOz > Na2s°3 + (X>2 (la)
2 NaOH + S02 > + 1^0 (lb)
The sodium sulfite formed in these reactions, and in the
regeneration of spent sodium bisulfite scrubbing liquor,
then reacts with additional dissolved sulfur dioxide to form
sodium bisulfite:
Na2S03 + + S02 > 2NaHS03 (2a)
Seme of the sodium sulfite is oxidized in the flue gas
scrubber to form sodium sulfate. This sulfate species does
not react with sulfur dioxide. It is converted in the
regeneration system to sodium hydroxide, for recycle to the
scrubbing tcwer, arri dilute sulfuric acid, for sale, plant
usage or disposal.
7-64
-------
This recovery of sodium sulfate via the second stage
membrane stack minimizes the loss of sodium values in the
process.
NOx
Removal of NOx contaminants by urea/methanol injection into
the flue gas is arampl i shwi in a two stage process. First,
urea addition reduces approximately 50-70% of the ND
contaminants in the flue gas to N_. Ttiis reaction occurs
at 1600-1900 r according to the inactions shown in 3a and
3b.
NH^CO'Nf^ > 2 NHj + 00 (3a)
NHj + NO > N2 + HjO (3b)
The remaining NO contaminant in the flue gas is oxidized to
NO. by reaction with methanol. Addition of methanol also
reauces ammonia slip, thereby avoiding deposition of salts
in downstream ducting and plume visibility problems. Uiis
reaction proceeds at 1000-1500 F as shown in 4a and 4b.
OljCH + 3 02 —> 2 HD2 + 002 + 1^0 (4a)
H02 + NO —> N02 + CH" (4b)
NQx vAiicii has been previously converted to NO_ via
(.»mimerciany available urea/methanol injection in the
boiler/eccaianizer sections respectively, will be reduced by
the sodium sulfite according to the reaction as shewn in 5a:
N02 + 2 Na2S03 —> 1/2 N2 + 21^2^ (5a)
or scrubbed in the absorber forming sodium nitrate salts as
follows:
2N02 + 2 ^0 —> 2HN03 + (5b)
7-65
-------
Note that Na_SO is oxidized to Na-SO. in reaction (5a).
Ihe Na_SO generated in this fashion Is eventually recovered
as a solid crystalline product, or processed in the optional
secondary AQQA3ECH Unit.
For this demonstration, the boiler will not be retrofit with
the NQx injection facilities. Instead, the ND- level will
be simulated by injection of NO. gas directly into the flue
gas slip stream ducting upstream of the absorber. Hie
levels of NO injection will be equivalent to the levels
measured conmercially at utilities viiicb have been retrofit
with the NQx reduction technology as well as a predicted
level for the Niagara Mchawk Dunkirk Steam Station No. 4
Boiler.
Regeneration of the spent scrubbing liquor is achieved
electrodialytically using the AQUAIECH System's bipolar
membrane system. Hie bipolar membrane separates water
molecules into hydroxen and hydroxyl ions. Hie sodium
bisulfite solution is directly regenerated in a
two-compartment AQUAIECH stack (Figure 2) to form the
original sodium sulfite scrubbing solution and gaseous
sulfur dioxide.
During regeneration the sodium bisulfite molecules
dissociate and the sodium cations migrate across the cation
selective membrane towards the cathode. It then rambines
with hydroxide anions produced in the bipolar membrane to
form sodium hydroxide. Most of this sodium hydroxide reacts
with the sodium bisulfite to form sodium sulfite. Hiis
solution could then be routed to em optional
three-compartment AQUAIECH stack to pick-up additional
sodium values firm the sodium sulfate steam after sulfur
dioxide recovery:
NaH503 + Na+ + OH- > Na^SC^ + 1^0 (6)
Hie bisulfite ion that is not transferred across the cation
selective membrane combines with the hydrogen ions from the
bipolar membrane to form sulfurous acid. Nearly saturated
sulfurous acid solution is continuously recycled around the
two-acnpartanerrt cell stack and adiabatically flashed at
7-66
-------
reduced pressure to recover sulfur dioxide gas and water.
This concentrated sulfur dioxide sLrvam is a suitable
feedstock far producing either sulfur or sulfuric acid using
ccnmercially available technologies.
NaHSCXj + H+ - Na+ > (7a)
I^S03 > 1^0 + S02 (7b)
The three-acnpartinent cell stack (Figure 3) can be employed
to recover sodium values frcn sodium chloride, sorl'hTm
sulfate, farmed by sulfite oxidation in the flue gas
scrubbing tcwer and sodium nitrate formed by the removal of
N0_ gases. Thespi sodium values are ultimately collected in
the sodium sulfite stripper bottoms stream. This stream
can optionally be treated in the second stage membrane stack
to form caustic and a dilute miypri acid. The dilute
sulfuric acid, nitric acid and hydrochloric acid stream
produced in the three-ccopartment stack can be neutralized
by reaction with limestone to form gypsum far disposal or
used in-house for icn-excbange regeneration or cleaning of
process units.
FREV1CCS DEVELOPMENT WSK
The proposed 2-3 MW Proof-of-Ccncept program is considered
the next logical step in the development of the SOXAL
process. Although an components of the SOXAL process have
not been demonstrated in an integrated operation, each has
been demonstrated in > » mmiH-ncial cperation or in pilot scale
tests.
The tuD-cccpartment AQCRIECH cell stack technology has been
demonstrated using pilot size cell stacks over a one year
test period. This 0.5 19? equivalent test pmgram proved the
ease of cperation, system reliability and membrane
durability. Test results showed that the efficiency of the
system remained stable with the bisulfite conversion at or
near 100% throughout the test.
7-67
-------
Hie three-ocrpartment cell stack technology has been
demonstrated in both laboratory and ccxmnercial operations
for sulfate and fluoride salt conversions. The first
AQUAIECH three-ccnpartment application has been
demonstrated with a irixpfi potassium fluoride/potassium
nitrate salt in a aamnercial stainless steel pickle liquor
regeneration unit at Washington Steel, Washington,
Pennsylvania.
Allied-Signal has developed the bipolar membrane over
approximately the past fifteen years. Die membrane used in
any proof-of-concept program will be produced on an existing
o.iiiniprcial production line. Continuous membrane
refinements are anticipated to provide even more efficient
membrane.
EBOCESS ECONOMICS
Die cost effectiveness of the SOXAL process had been
estimated by Steams-Catalytic under an Electric Power
Research Institute (EERI)contract (SD-3342 Vol.5) in 1986.
This initial stucty found the SOXAL process to be the lowest
cost regenerative process. It also found the process to be
cCTnpetitive with the lowest cost throwaway processes.
More recently this study was reevaluated by United
Engineering and Constructors. Ihese eocmmics are based on
a nominal 300 MW unit burning 2.6 wt% sulfur coal at 65%
operating capacity. Operating costs include capital, fixed
and van'able costs. Ihese ecancmics were not available at
the time of this writing, however based on economics
generated by Allied-Signal Inc. in connection with the
DOE-EETC program it is believed that the SOXAL process
continues to remain competitive.
DEMONSTRATION PROGRAM
Die SOXAL Demonstration Program began September 10, 1991 and
is anticipated to take approximately 22 months to complete
as shewn by the schedule (Figure 4).
7-68
-------
During the 6 months of scheduled operations period, data
will be collected Iron the SOXAL system to define:
1) SO. and NCoc removal efficiencies
2) Current efficiency for the regeneration unit.
3) Sulfate oxidation in the absorber.
4) Make-
-------
References:
1. K. N. Mani. «yi-TrvjiaIvsis Water Splitting Technology.
Amsterdam: Elsevier Science Publishers B.V., 1991, pp.
117-138.
2. K. N. Mani, F. P. Chlanda and C. H. Byszewski,
AOOAIESCH Membrane Technology for Reoovrv of Acid/Base
Values frcm Salt Streams.
Amsterdam, Desalination, 68 (1988), pp. 149-166.
3. K. J. Liu, K. Nagasubramanian and F. P. Chlanda.
Application of Bipolar Membrane Technology : A Novel
Procss for Control of Sulfur Dioxide frum Flue Gases.
Amsterdam, Journal of Membrane Science, 3(1), 1978, pp.
57-70.
4. K. N. Mani and F. P. Chlanda. SOXAL™ Process Results
of Tahoratorv and Pilot Studies.
Pittsfcur^i, Pennsylvania, Paper presented at 2nd Annual
Pittsburgh Goal Conference, September 1985.
7-70
-------
Figure 1
SOXAL" Process
Genera! Flow Configuration
Treated Hue Gas
Liquid SO;
SO, Rich
Hue Gas
Spent
Scrubbing
Solution
Sulfur
Dioxide
Recovery
Unit
Primary
AQUATECH
Cell Stack
Steam
Scrubber
Soda Ash
Na2S04 Solution
Regenerated
Scrubbing
Solution
Secondary
AQUATECH
Cell Stack
7-71
-------
Figure 2
Primary Membrane Stack
0
©j
Anode
©
Na*
oh;
Base
H*
©
Na*
Acid
©
OH'
Base
©
H*
Acid
©
NaHS03 Solution
OH
Na*
Base
H*
Acid
Na*
1-0
Cathode
Base
_L
•h2so3
Solution
Na2S03 + NaOH
Solution
7-72
-------
Figure 3
Secondary Membrane Stack
Na^SO,
drag stream after*
S02 recovery
©H —r
Anode Base
Acid (HjSOJ-
OH=-
Na'
— H*
so;—
Acid
© © JO © 0 0 0
Salt
OH=~
—Na*
Base
— H*
so;-
Acid
Salt
Base Solution
-from primary stack
-Water
-Na*
Base
h
0
Cathode
¦ Recovered
base solution
to absorber
7-73
-------
£
Flyuro 4
SOXM. Demonstration Program DOE PETC Contract I DBAC2291PC91347
100 1 I 1092 T
CCJl_.T A MANPOWER
CI AM
DESIGN OASIS it
Il 'if PI.AN
I'NCl.. DESIGN A
COMIKACriNC
I'IMIC.. EAI1 . A
iii iaii 1:0 urs.
Ill: WORK
IMSIAI.1.AIION
illAKIOOWN
parametric a
REI. TESTING
IJISSASSEMGLY
CONCEPTUAL
111 .'SIGN 500MW
rc analysis
I 003
S O N D J
M A M J
asonuIj r MAM.I .1
-------
THE HEALY CLEAN COAL PROJECT
AIR QUALITY CONTROL SYSTEM
V. V. Bland
S. M. Rosendahl
D. M. Shattuck
Stone & Webster Engineering Corporation
7677 East Berry Avenue
Englewood, Colorado 80111-2137
7-75
-------
Intentionally Blank Page
-------
ABSTRACT
The Healy Clean Coal Project is a 50 MW nominal new coal fired power
plant demonstrating two 350 MMBtu/hr TRW coal combustion systems
designed to burn high ash coals while yielding low NOz emissions.
Limestone is introduced into the combustor discharge to create flash
calcined material (FCM). The FCM is collected in a fabric filter
located downstream of a spray dryer. From the fabric filter, the
collected material is slurried and fed into the spray dryer through the
atomizer as the active alkali for capture of sulfur dioxide. The
integration of the combustion system and the spray dryer/fabric filter
with reactivated FCM is expected to result in low NOx and SOz emissions.
This paper presents the design of the unique air quality control system *
and the status of the project.
7-77
-------
INTRODUCTION
The Healy Clean Coal Project: (HCCP) is jointly sponsored by the Alaska
Industrial Development and Export Authority (AIDEA) and the U.S.
Department of Energy (DOE). The HCCP was selected by DOE in Round III
of its Clean Coal Technology Program.
The primary objective of the HCCP is to conduct a cost-shared project
that will demonstrate a new power plant design which features innovative
integration of an advanced combustor and heat recovery system coupled
with both high and low temperature emission control processes. The
parties anticipate that, if the demonstration project is successful, the
technology could become commercialized during the 1990's and will be
capable of (1) achieving significant reductions in the emissions of
sulfur dioxide and the oxides of nitrogen from existing facilities to
minimize environmental impacts such as transboundary and interstate
pollution and/or (2) providing for future energy needs in an
environmentally acceptable manner.
The demonstration project is proposed to be built adjacent to the Golden
Valley Electric Association (GVEA) existing Healy Unit No. 1 pulverized
coal power plant. The site is located near Healy, Alaska. Alaskan
bituminous and subituminous coals will be tested. GVEA will operate and
maintain the new power plant facility.
Coal from the adjacent Usibelli Coal Mine (UCM) will be pulverized and
burned at the proposed facility to generate high-pressure steam that
will be used by the steam turbine generator to produce electricity.
Emissions of SOz and NOx from the plant will be controlled using TRWs
combustion systems with limestone injection, in conjunction with a i
boiler supplied by Foster Wheeler. Further S02 and particulate removal
will be accomplished using Joy Environmental Equipment Company's (Joy)
Activated Recycle Spray Absorber System.
The total project activities include design, permitting, procurement,
fabrication, construction, start-up, testing, and reporting of results.
Construction of the demonstration facility is expected to start in the
spring of 1993 and continue for 2.5 years. Following completion of the
demonstration test program, the plant is expected to continue to operate
and be maintained as a commercial utility electric generation station.
The proposed HCCP is to be a nominal 50 MWe facility consisting of two
pulverized coal-fired combustion systems, a boiler, a spray dryer
absorber with activation and recycle equipment, a fabric filter, a
turbine generator, coal and limestone pulverizing and handling
equipment, and associated auxiliary equipment.
The specific objectives of the HCCP demonstration are to: (1)
demonstrate the use of Alaskan, low-sulfur bituminous and subutuminous
coals of medium to high ash and moisture content; (2) demonstrate the
feasibility of large utility boiler repowering capability of the TRW
Combustion System; (3) demonstrate large utility boiler retrofit
capability of the TRW Combustion System on oil designed boilers with no
derating and on pulverized coal and cyclone furnace design boilers with
improved performance, and lower NOx, SOz, and particulate emissions; (4)
demonstrate the enhanced capability of the TRW Combustion System for
simultaneous NOx and SOz removal when combined with back-end S02
7-78
-------
absorption -techniques and furnace air staging to maintain emissions at
the nominal values of 0.2 lbs of NOx and 0.01 lbs of particulate matter
per MMBtu, and greater them 90% removal of S02; (5) demonstrate the
energy efficiency of the integrated technology as compared to pulverized
coal (PC) and fluidized bed combustor processes; and (6) determine the
cost effectiveness of the technology especially in terms of reduced
operating costs due to the system's capability to bum low grade/waste
coals.
The air pollution control system that will be demonstrated by the
project (Figure 1) incorporates the following major components:
• TRW Coal Combustion System
• Foster Wheeler boiler
• Joy Spray Dryer Absorber Fabric Filter System
The integrated air pollution control process that results from the HCCP
configuration of these components has been designed to minimize
emissions of S02> NOx, and particulates from the facility while firing
a broad range of coals.
N0X emissions are reduced in the coal combustion process by use of the
fuel and air-staged combustor system and a boiler that controls fuel and
thermal-related conditions which inhibit nitrogen oxide formation. The
slagging combustor/boiler system also functions as a limestone calciner
and first stage S02 removal device in addition to its heat recovery
function. Secondary and tertiary S02 capture are accomplished by a
single spray dryer absorber vessel and a fabric filter respectively.
Ash collection in the process is first achieved by the removal of molten
slag in the coal combustors followed by flyash particulate removal in
the fabric filter system downstream of the spray dry absorber vessel.
This paper describes the unique integrated air pollution control process
of the HCCP application. Subsequent sections discuss additional details
of the NOx, S02 and particulate control systems. A final section of the
paper addresses the current status of the project.
AIR QUALITY CONTROL SYSTEM
The TRW Combustion System will be designed to be installed on the boiler
furnace to provide efficient combustion, maintain effective limestone
calcination and minimize the formation of NO* emissions. As shown in
Figure 2, the main system components include a precombustor, main
combustor, slag recovery section, tertiary air windbox, pulverized coal
and limestone feed system, and combustion air system. Figure 3 shows
a schematic of the general boiler arrangement and the combustion system
installation for the HCCP. In this unique arrangement, the slagging
combustors are bottom mounted on the boiler hopper to yield optimum
operation and cost benefits.
The coal fired precombustor is used to increase the air inlet
temperature to the main combustor for optimum slagging performance. It
burns approximately 25-40 percent of the total coal input to the
combustor. Combustion is staged to minimize NO* formation.
7-79
-------
The main slagging combustor consists of a water-cooled cylinder which
is sloped toward a slag opening. The remaining coal is injected axially
into the combustor, rapidly entrained by the swirling precombustor gases
and additional air flow, and burned under substoichiometric (fuel-rich)|
conditions for NOx control. The ash contained in the burning coal forms
drops of molten slag and accumulates on the water-cooled walls as a
result of the centrifugal force resulting from the swirling gas flow.
The molten slag is driven by aerodynamic and gravity forces through a
slot into the bottom of the slag recovery section where it falls into
a water-filled tank and is removed by the slag removal system.
Approximately 80 percent of the ash in the coal is removed as molten
slag.
The hot gas, containing carbon monoxide and hydrogen, is then ducted to
the furnace from the slag recovery section through the hot gas exhaust
duct. To ensure complete combustion in the furnace, additional air is
supplied from the tertiary air windbox to NO, control ports and to final
overfire air ports located in the furnace.
Pulverized limestone, for S02 control, is fed into the combustor as
shown in Figure 2. While passing into the boiler most of the limestone
is decomposed to flash calcined lime by the following reaction:
CaC03 + heat - CaO + C02T (1)
The mixture of this lime and the ash not removed by the combustors is
called Flash Calcined Material (FCM). Some sulfur capture by the
entrained CaO also occurs at this time, but the primary S02 removal
mechanism is through a multiple step process of spray drying the
slurried and activated FCM solids (Figure 4).
Once FCM is produced in the furnace via equation (1), it is removed in
the fabric filter system. A portion of the material is transported to
disposal. Most of the material however, is conveyed to a mixing tank,
where it is mixed with water to form a 45% FCM solids slurry. The lime
rich FCM material is slaked by agitation of the suspension. A portion
of the slurry from the mixing tank passes directly through a screen to
the feed tank, where the slurry is continuously agitated. The remainder
of the slurry leaving the mixing tank is pumped to a grinding mill,
where the suspension is further mechanically activated by abrasive
grinding.
By grinding the slurry in a mill, the FCM is activated by a mechanical
process whereby the overall surface area of available lime is increased,
and coarse lime particle formation is avoided. Thus, the mill enhances
the slaking conditions of the FCM, and increases the surface area for
optimal S02 absorption. FCM slurry leaving the tower mill is
transported through the screen to the feed tank.
Feed slurry is pumped from the feed tank to the SDA, where it is
atomized via rotary atomization using JOY/Niro dry scrubbing technology.
Sulfur dioxide in the flue gas reacts with the FCM slurry as water is
simultaneously evaporated. The dry reaction product is removed via the
SDA hopper or baghouse catch. Sulfur dioxide is further removed from
the flue gas by reacting with the dry FCM on the baghouse filter bags.
7-80
-------
The HCCP is an integrated system for the combustion of coal and control
of all emissions. The slagging combustor, furnace, and enhanced recycle
SDA system all play a part in reducing emissions from the plant- The
slagging combustor inhibits NOx production, generates the FCM for
capture of sulfur dioxide, and reduces the potential amount of fly ash
by up to eighty percent. The furnace further contributes to the NO,
reduction process and begins the sulfur dioxide removal process. The
recycle/reactivation SDA system, which includes the pulse-jet baghouse,
completes the collection of particulate and sulfur dioxide.
Removal of any single component in the integrated system results in
ramifications on other components. For example, removal of the slagging
combustor and replacement with low NOz burners increases the ash loading
out of the furnace by nearly four hundred percent, eliminates the
production of FCM which requires the conversion of the
recycle/reactivation SDA system to a conventional lime spray dryer
system, and possibly increases NOz emissions. Replacement of the spray
dryer with a wet scrubber eliminates the need to generate FCM since all
of the particulate would be collected upstream of the wet scrubber in
a fabric filter or electrostatic precipitator where there is no way of
separating fly ash from FCM.
NO. Control
Emissions of NOz are expected to be demonstrated to levels significantly
below EPA New Source Performance Standards (NSPS) in the boiler by using
slagging combustor technology and known combustion techniques.
The HCCP combustors achieve NOx control as a combination of the
following (two) factors:
1. The combustor functions as a well-stirred reactor under
substoichiometric conditions for solid fuel combustion;
converting the solid fuel components to a hot, partially
oxidized fuel gas in an environment conducive to destroying
the complex organic fuel bound nitrogen compounds which could
easily be oxidized to N0X in the presence of excess oxygen.
2. The combustor water cooled enclosure additionally absorbs
approximately 10 to 25 percent of the total available heat
input to the combustor.
These two conditions together reduce the potential for encountering
combustion temperatures in the furnace sufficient for decomposition of
molecular nitrogen compounds in the combustion air into forms which can
produce thermal NOx emissions as excess oxygen is made available.
When the exhaust gases leave the combustor, the coal has already been
mixed with approximately 80 to 90 percent of the air theoretically
necessary to complete combustion. A portion of the remaining 10 to 20
percent is then allowed to mix slowly with the hot fuel gases exiting
the combustor and entering the furnace. The hot gases radiate their
heat to the furnace walls at rates faster than combustion is allowed to
occur so that gas temperatures slowly decay from those at the furnace
entrance. After the furnace gases have cooled sufficiently, a second
and possibly third stage of furnace combustion air injection is
performed as necessary to complete the coal combustion process in an
7-81
-------
oxidizing, controlled manner so that, combustion gas temperatures are
maintained below the thermal N0Z floor where significant NOz formation
begins. This is in contrast with a traditional coal-fired furnace where
the pulverized coal is burned in suspension at high excess air rates.
Resulting gas temperatures from PC furnaces typically rise significantly
above the 2800°F temperature maintained in the slagging combustor and
downstream furnace. In the traditional furnace, the pulverized coal is
relatively poorly mixed with conventional low N0X wall burner/suspension
firing techniques, and local areas of combustion in the presence of
stoichiometric oxygen create hot zones within the flame. These hot,
turbulent stoichiometric zones can produce significant NOx levels in the
area of burner throats. This tendency for high, localized NOx formation
is minimized with the slagging combustor through slow, controlled mixing
of furnace combustion air with the partially cooled, well-mixed fuel
gases discharging from the combustor into the lower furnace NOx control
zone.
The general relationship between low NOz emissions and combustor
stoichiometry resulting from tests at TRW's 50 MMBtu/hr Cleveland
demonstration facility are shown in Figure 5. This facility operates
with low excess air and no overfire air or NO, ports in the furnace.
The curve shows that NOx emissions (while firing bituminous coal) are
minimized for the Cleveland combustion process at approximately 30
percent of the current NSPS when the slagging combustor system is
operated at a stoichiometry of 0.70 and when all of the final combustion
air is added at the combustor exit nozzle at the entrance to the boiler
furnace. The HCCP will demonstrate additional NOx reduction techniques
including furnace NO, ports and furnace over-fire air injection.
NOx results obtained while firing the HCCP performance coal (Table 1) in
the Cleveland slagging combustor system differ from results obtained
with other, less volatile fuels. Figure 6 presents NOx emissions from
the facility as a function of main combustor stoichiometry. As can be
observed from the figure, very low N0X emissions are achievable by
operating the combustor near or slightly below stoichiometry while
firing the HCCP performance coal. These results indicate that N0X
emissions are only mildly dependent on combustor stoichiometry thus
providing operating flexibility.
SO- Control
Emission levels of S02 are controlled to and below NSPS levels using the
recycle/reactivation SDA system of Joy Environmental Equipment Company.
The coals, fired in the HCCP combustion system (shown in Table 1), are
low sulfur, high moisture, low heating value fuels from a nearby mine.
While the project will demonstrate higher SOz removal efficiencies, the
sulfur content is so low, only a 70% sulfur dioxide removal efficiency
is required to satisfy NSPS requirements. Coal 1 is a run-of-mine coal,
where care was taken in the mining operation to minimize the amount of
overburden and lenses included with the coal. Coal 2 is the performance
coal and consists of 50% run-of-mine and 50% waste coal. The waste coal
being a lower heating value fuel with significantly more ash. An
advantage of the slagging combustor is that it can burn low quality
7-82
-------
coals and still meet emissions and performance requirements.
The ash Etna lysis shown in Table 1 indicates that the ash contains a
relatively high amount of calcium oxide. As previously discussed, only
20* of the coal ash leaves the boiler as fly ash. This means that only
about 20% of the calcium oxide shown in the ash analysis is carried over
to the spray dryer.
Tests were performed at the TRW facility in CI eve lam d and Joy-Niro's
facilities in Copenhagen to confirm design conditions for the HCCP. The
primary purpose of the tests in Cleveland was to generate FCM that could
be used in the Niro test facility. Coal and limestone that are to be
used by the HCCP were used for the tests. Preliminary results from the
Niro tests show that 70 percent sulfur dioxide removal is attainable at
a Ca/s ratio of 1.7, with 90* removal attainable at slightly higher
stoichiometrics. These tests were accomplished by heating the FCM
slurry. Testing is to be performed in the near future to determine the
effect of mechanical activation (grinding) of the FCM.
An extensive test program has been proposed to prove various aspects of
the sulfur dioxide removal process. It is anticipated that the test
program will continue for 2.5 years following startup. The test program
will evaluate primary S02 removal (removal in the furnace), secondary S02
removal (removal in the SDA), and tertiary SOz removal (removal in the
fabric filter) at various loads. Additional tests will evaluate the
following:
• adiabatic humidification of the flue gas in the SDA, with and
without limestone injection in the combustor;
• adiabatic humidif ication of the flue gas in the SDA, combined
with dry injection of FCM or flyash downstream of the air
heater outlet;
• and dry injection of recycled FCM between the SDA outlet and
the fabric filter inlet, with and without adiabatic
humidification in the spray dryer.
The test program will attempt to confirm that pilot plant FCM spray
drying tests which have been performed at Niro Atomizer's test facility
in Copenhagen can be economically scaled up to a commercial sized
facility.
Particulate Control
Particulate emissions control on the HCCP is obtained via the slagging
combustors and by a pulse-jet baghouse. Each of ten fabric filter
compartments will contain 225 six inch diameter fiberglass bags. The
effective length of each bag is 20'-0" and the gross air-to-cloth ratio
is 2.8:1. The HCCP will demonstrate the effectiveness of a pulse jet
baghouse in removing the FCM particulate emissions.
It should be noted that a significant portion of the coal ash never
leaves the furnace with the flue gases, since it is estimated that
approximately eighty percent of the ash in the coal will leave the
slagging combustors as slag.
7-83
-------
PROJECT STATUS
The HCCP is currently in the permitting and design phase. A one year
meteorological and air quality data collection program has been
conducted. Field studies in support of the Environmental Impact
Statement and other necessary permits have also been completed. Permit
applications to Federal and State agencies are in various stages of
development.
In addition to the coal and FCM tests reported, significant engineering
studies were prepared including the heat rejection system, final site
selection, coal quality, and combustor arrangement. Preliminary station
arrangements, piping & instrument diagrams, site plans and other design
documents were issued. Procurement of equipment and systems is
continuing.
A definitive project cost estimate will be prepared in the first quarter
of 1992 for AIDEA and the DOE which will provide the basis for
decisions regarding completion of the project.
CONCLUSION
The HCCP features a unique coal combustion and air pollution control
process that has the potential to significantly:
• reduce emissions associated with coal fired steam generation
systems;
• increase efficiency of the coal combustion process;
• and reduce operational costs and impacts.
7-84
-------
REFERENCES
1. Healy Cogeneration Project Demonstrating TRW's Entrained Combustion
System with Limestone Injection and Joy Technologies' Activated
Recycle SDA System, Clean Coal Technology III Proposal, Volume II -
Demonstration Pro~iecrt Proposal. August, 1989.
2. Healv Coal Firing at TRW Cleveland Test Facility. Final Report to
Alaska Industrial Development and Export Authority, August, 1991.
7-85
-------
H—\. J_i t J
T
9VtMm
TOitfn
AOVAMCCD TECHNOLOGIES
IPUtOMOQuMCAlOi
— — — — BOTTOM ASM TO VAST!
Figure 1. HCCP Process Flow Diagram.
Figure 2. TRW Coal Combustion System.
7-86
-------
SECTIONAL
SIDE ELEVATION
FRONT
ELEVATION
Figure 3. HCCP Combustor Installation.
FE£D T
Figure 4. Joy Recycle/Reactivation SDA System.
7-87
-------
500
400
r?
o
*
® 300
>
2
0.
g 200
100
0
0.50 0.60 0.70 0.80 0.90 1.00
COMBUSTOR STOCHOMETRY
Figure 5. NOx Emissions for Bituminous Coal Versus
Combustor Stoichiometry.
HEALV TESTS (OX 6-e.t» MW8TUHR)
200 "
ISO
100 H
~ o
nBO0"0
~
0 ~
orfgf
8?
~
1
~
0.5
0.7 0.9
COMBUSTOR STOCHIOMETRY
1.1
Figure 6. NOx Emissions for HCCP Coal Versus
Combustor Stoichiometry.
7-88
-------
TABLE 1
COAL AMD ASH ANALYSIS
Coal 1
Proximate Analysis
Moisture, % 26.35
Ash, % 8.20
Volatile, % 34.56
Fixed Carbon, % 30.89
Total, % 100.00
HHV, Btu/lb 7,815
Ultimate Analysis
Moisture, % 26.35
Ash, % 8.20
Carbon, % 45.55
Hydrogen, * 3.45
Nitrogen, % 0.59
Sulfur, % 0.17
Oxygen, * 15.66
Chlorine, * 0.03
Total 100.00
Elemental Ash Analysis
Silicon Dioxide, % 38.61
Aluminum Oxide, % 16.97
Titanium Dioxide, % 0.81
Ferric Oxide, % 7.12
Calcium Oxide, % 23.75
Magnesium Oxide, % 3.54
Potassium Oxide, % 1.02
Sodium Oxide, % 0.66
Sulfur Trioxide, % 5.07
Phosphorus Pentoxide, % 0.48
Strontium Oxide, % 0.23
Barium Oxide, % 0.44
Manganese Oxide, % 0.06
Undetermined, % 1.24
Total, % 100.00
Coal 2
25.11
16.60
30.73
27.51
100.00
6,960
25.11
16.60
40.57
3.07
0.53
0.15
13.94
0.03
100.00
65.69
11.09
0.52
4.90
10.62
1.87
1.16
0.65
2.28
0.30
0.11
0.22
0.04
0-55
100.00
7-89
-------
Intentionally Blank Page
7-SO
-------
LIME / LIME STONE SCRUBBING PRODUCING USABLE BY-PRODUCTS
D. PAUL SINGH
PROCEDAIR INDUSTRIES INC.
625, PRESIDENT KENNEDY, 14th FLOOR
MONTREAL, QUEBEC, H3A 1X2
KNOT PAPAJEWSK1
THYSSEN ENGINEERING Gnbh
AM THYSSENHAUS 1
D - 4300 ESSEN 1
7-91
-------
Intentionally Blank Page
-------
ABSTRACT
This paper is in two (2) parts. The first part deals with plants
manufacturing gypsum, form the sulphur captured in desulphurization
plants. It is intended to describe, by using some of these plants as
examples, developments made in the past few years and the results that
have become available. It outlines the steps in the development of plans,
which were followed in retrofitting the FGD plants to the Generating
Stations, and our experiences gained.
Between 1983 and 1988, in the Federal Republic of Germany, Generating
Station capacity of more than 30,000 MWe were retrofitted with flue gas
desulphurization plants. The overall objective was to produce industrial-
usable products, from the sulphur captured in these desulphurization
processes. The requirements of these desulphurization plants were, in
some cases, in excess of what could be achieved by the, then, state-of-
the-art technologies. The conditions required over 90% of the
desulphurization capacity to be met, using lime/lime stone scrubbing
technology. Thyssen, as one of the leaders of lime/lime stone scrubbing,
delivered plants for some 8500 MWe of capacity; and desulphurization
efficiencies up to, and above, 97% were attained. Also, over a twelve
year operating period, boiler availability was 98% to 99%.
The second part of the paper discusses a newly developed process, called
the "Bertin Process", a method of producing elemental sulphur from either
wet or dry SOj lime scrubbing technologies. This process involves two
stagor. In the first stage, biotransformation of the sulfites or sulfates
to .l\t» occurs, with recycling of the calcium carbonate to the
desux;Surization plant. The second stage involves the recovery of
elemer. al sulphur from HjS, by means of updated REDOX conversion. This
paper a Oentifies major components of the system, together with results
obtainet on a pilot plant in France.
1. i»r VODOCTTOM
Worldwide, various processes have been developed for cleaning the flue
gases emitted from fossil fuel power plants. For SOj removal, wet
processes, generally using limestone as a reagent and producing commercial
quality gypsum as a by-product have found general acceptance from a
technical and also from an economic stand point. In West Germany during
the 80's, wet process flue gas desulfurization systems were installed on
approximately 37,000 MWe. The fact that West Germany was behind the USA
and Japan in installing this equipment benefitted the installations by
virtue of the fact that the designers of these plants gained insight and
technical knowledge on the plants installed in the 70's, particularly
those in the USA. The experience of these early plants together with the
very strong Federal and State Legislation requirements imposed by West
German Authorities lead to the emergency of a whole new advanced
generation of FGD Designs.
The Thyssen SOj flue gas desulfurization (FGD) systems, built recently, in
Germany and the USA involve 8500 MWe. To obtain the reliability,
efficiency and the gypsum quality required two (2) distinct generations of
design. Developments of these designs over more than ten (10) years have
formed a basis for a third generation design, which has been installed in
West Germany. The enhanced third generation design has been supplied on
1,000 MWe of plants in West Germany. (See Figure 1).
7-93
-------
The process is characterized by a combined spray and falling film zone.
The absorption sump is equipped with a forced oxidation system for
elevated limestone utilization and high gypsum quality. This improved
process enables the equipment to meet stricter qualification criteria
compared with the requirements of the first and second generation FGD
systems as follows:
• Liam Stone Reagent Utilization of 97%+
• Reliability of 99%, no spare absorbers or multiple boilers on a
single absorber
• Achieves 95 to 98% SO, removal
• Single absorbers of 750 MWe size
• Scaling and Plugging are eliminated
• Includes oxidation and high lime stone utilization - The selection
of the pB level in the process insures high level efficiency and
high by-product gypsum quality at very economical design criteria.
• Reduce air requirements of forced oxidation result in high quality
gypsum as required by German standards with """« gypsum
rejection.
• Low parasitic power of approximately 1.5% of the generator output
• A single central power plant control system providing tight process
control and operating stability.
2. IMPORTANT FACTORS INFLUENCING THE FGD PROCESS
The optimum performance of the FGD system is controlled by the following
parameters from a chemical view point:
a) pH of the suspension slurry
b) The reactivity of the lime stone
c) Crystallization of the CaSO* X 2H2O gypsum - supersaturation
d) The composition and solids contents of the suspension
e) The suspension volume, volumetric size of the reaction vessels, L/G
ratio, droplet size/weight distribution, height of the spraying
zone.
3. OPERATING EXPERIENCE OF THE THYSSEN FGD LIMESTONE SYSTEM
Wet flue gas deBulferization (FGD) processes are in operation at many
power plants around the world. The Thyssen FGD system with lime stone was
developed in the early 1980 's; since then it has been installed on ten
coal fired boilers in the USA and West Germany with a total capacity of
approximately 2,500 MW. The earlier generations of design have been
successfully modified over the last ten (10) years as a result of on going
project improvement programs. These have enhanced the overall process
operation and system reliability.
The first application of the Thyssen system was on a 360 MW boiler burning
low sulfur (.5%) coal. The end product was unoxidized sludge i.e. calcium
sulfite (CaSOj). Simple dewatering equipment was installed. In
comparison, the latest-generation FGD system has shown successful
operation on boilers burning up to 2.5% sulfur coal, SO2 removal
efficiencies up to 97%. These criteria have been achieved with lime stone
reagent simultaneously producing salable gypsum containing predominantly
calcium sulfate (CaS04).
The capabilities of the system allows the operator to maximize the overall
plant operation with regards to fuel selection and end product quality.
If the operator likes to produce FGD gypsum for sale, either for use as a
cement additive, soil conditioner or for wallbo&rd manufacturing
facilities, the waste disposal cost is drastically reduced. Even at
installations where FGD gypsum cannot be sold due to marketing
restrictions the oxidized end product (calcium sulfate) is frequently
preferred by utilities because of its far better dewatering characteristic
compared to the unoxidized sludge (calcium sulfite).
7-94
-------
4. RECENT DEVELOPMENTS
Evolution of the Thy seen FGD system has resulted in the development of
Vary Large Absorber VLA modules (up to 62' diameter) . These absorbers are
capable of handling the flue gas from approximately a 750 MW boiler (flue
gas flow rates up to 2,000,000 ACFM per minute. Thyssen has two (2) such
VLA's in operation in West Germany since 1988. Expediency of design as a
result of dealing with many retrofit applications, saw the evolution
towards large towers to handle the severe space restrictions. Development
of the design of the VLA'o occured after extensive field measurements in
full scale plants and numerous gas flow studies. For these large
absorbers, study of several major design requirements such as, gas liquid
contact, nozzle orientation, slurry agitators, were required. These,
together with consideration of structural requirements, were incorporated
into final designs and have now been demonstrated on operating units.
The state-of-the-art technology not only covers improved materials of •
construction, such as, alloys and linings, but also mechanical refinements
such as the use of hydroclones in combination with belt filters or
centrifuges for dewatering, mechanical seals for slurry pumps and
different types of reheat systems, etc.. Of course the modern plant also
included state-of-art process controls and instrumentation. Kany of theBe
design features are included in the plants described in this paper.
Details and operating data from the two (2) installations at Rheinich
Westfaelische Elektrizitaets Werke Frimmersdorf Units P & Q are presented
later.
5. PROCESS DESCRIPTION
A) Absorber Loop
The untreated flue gas is contacted with recirculating limestone slurry at
a pH of 5 to 5.4 in the absorber. A number of spray levels and a Falling
Film Contactor (FFC) are used to provide the necessary mass transfer
capacity for absorption of the SO;. The slurry in the absorber loop flows
down to the Absorber Recirculation Tank (ART). Fresh limestone slurry is
added to the ART based on process demand measured by the incoming SO, level
and the required outlet emission level. The ART is sized to provide
dissolution of limestone as well as desupersaturation of calcium sulfite.
The absorber loop operates on a closed loop concept and the bleed to the
dewatering system is controlled by density to approximately 15 to 17% by
weight.
B) Dewatering
The absorber slurry which contains 15 to 17% by weight of solids, is
passed through the dewatering system. Both primary and secondary
dewatering stages are required for most plants. Unoxidized sludge
(calcium sulphite) has typically been dewatered in a thickener as a first
step, although hydroclones may also be used to dewater the unoxidized
sludge. For oxidized sludge (gypsum), hydroclones have generally been
selected as the primary dewatering device over thickeners because of their
lower coat, compact size and superior particle size separation. Vacuum
filters have been used for secondary dewatering which produces an end
product with a moisture content between 60 and 88% solid for calcium
sulfite. If oxidized gypsum sludge is dewatered with a vacuum belt filter
an average moisture content of less than 10% is achieved. The dewatered
sulfite cake is typically blended with fly ash (and may be lime) to
produce a low permeability waste suitable for landfill. The dewater
gypsum product can either be used for landfill or sold for various process
applications.
7-95
-------
6. OPERATING DATA
RHEINICH WESTFAELISCHE ELEKTRIZTAES WERKE, FRXMKERSDORF UNITS P & Q
A) General
Frimmersdorf power atation is located in North Rhein Westfaelen near
Cologne Heat Germany. Units P fi Q are rated at 305 MWe steam generator,
the fuel used is ignite. The air pollution control system includes
electrostatic precipitators followed by two (2) 50% I.D. fans. From there
the raw flue gas at Unit P is ducted to a regenative reheater (Lungstrom
type) before being ducted to the Thyssen FGD scrubber. At unit Q the raw
gas is ducted directly to the Thyssen FGD scrubber.
B) FGD Pesion Basis
FGD system design parameters are summarized as follows:
Flue gas design flow rate 2,350,00 NaP/h wet (3,814,000 m3/h wet) or
1,460,000 SCFM (2,250,000 ACFM) at 170°C (338°F). The design SOj inlet
concentration is 5500 mg/Nm3 dry at 6% Cu or 1925 pptndv. The maximum SO,
emission in the stack, is limited by regulation, to less than 400 mg/Nm*
dry at 6% O, or 136 ppmdv. This corresponds to an overall removal
efficiency of 92.8%. Under these conditions, the individual absorber
efficiency was guaranteed at 95% with 100% gas treatment.
C) Reaoent Subsystem
The limestone subsystem is rated at 25 tons/hr feed rate of limestone
powder to the slurry tank. The limestone slurry is stored in a slurry
feed tanlc prior to pumping to the Absorber Recirculation Tanks (ART'S).
The sizing of the limestone preparation equipment is based on the maximum
inlet SO. rate, that is to say the highest design sulfur containing coal
at MCR boiler load. The addition of fresh limestone slurry to each ART is
regulated to maintain the desire stoichiometric ratio and pH level using
the inlet SOj mass flow rate signal. pH override signals are incorporated
to make sure that the absorber module chemistry operates under the most
favorable conditions.
D) Absorber Modules
Each FGD system includes one (1) single VIA absorber module 62 feet in
diameter. Each module has its integrated Absorber Recirculation Tank
(ART). No spare modules are installed. Uniform gas distribution to the
absorber is accomplished with proper ductwork and system layout which
eliminates the need for any bypass control dampers. The absorber towers
are carbon steel lined with corrosion resistant rubber lining, while
stainless steel is used for all slurry recirculation pumps. A wetted
Falling Film Contactor (FFC) is installed in the tower to achieve the
required 95% SO. removal. Silicon carbide slurry spray nozzles are used.
The vertical demister consists of two stageB, each provided with their own
water wash header and nozzles. The lower demister is washed once every
hour, while the upper stage demister is washed only once every four (4)
hours.
E) Reheater
In the case of Unit P, the raw gas coming from the I.D. fan is cooled from
170°C (338°F) to approximately llOt (230°F) in the regenerative heater
before entering the absorber. The absorber exit gas at approximately 65°C
(149*F) is ducted to the regenerative heater and leaves the heater to the
stack at approximately 120°C (248*7).
For Unit Q no reheat system was installed. The saturated absorber outlet
gas at 65°C (149°F) is directly ducted to the existing cooling tower.
7-96
-------
F) Gvdgum Dewatering
The produced gypsum slurry in 'the ART is completely oxidized by the
integrated insitue forced oxidation system before leaving the ART. The
bleed to the dewatering system is controlled by density. The dewatering
system consists of a hydroclone system as pre dewatering device and a
vacuum belt filter. Each FGD unit has one dedicated 100% dewatering
system and one common is installed as standby.
7. PERFORMANCE DATA TOR FRIMMERSDORF
A) General
This section deals with the test results obtained during the Owner's
acceptance test. The following parameters were used in determining the
absorber performance:
e sOj concentration at absorber inlet, outlet and stack
e Pressure drop across absorber and system
e Absorber gas flow rate
• Absorber droplet carry over
e Gypsum dewatering rate and gypsum quality
Test were conducted at 100% and 80% MCR load. Limestone usage and waste
product characteristics and production were evaluated by obtaining and
analyzing various samples at different locations in the FGD system.
German DIN standards were used for the sampling at SO?, CO-, O? and
particulates.
B) SO-. Removal Efficiency
SO] concentration in the flue gas was measured at the inlet and outlet of
the absorber module and in the stack for two cases of gas flow rate (100%
and 80%). During the test power production was around 300 MW, the
measured average total flow gas at the FGD inlet was 2,150,000 Nm'/hr wet
(1,460,000 SCFM). The test therefore were ran at conditions very close to
the design basis.
The average SO; inlet concentration was around 1650 ppmdv. The individual
absorber tower efficiency for Units P & Q were measured at 97% and 98%.
Test results at the 80% flow rate and 100% flow rate showed no significant
difference in SO] removal efficiency between the two rates. During the
test the limestone stoichiometric ratio was measured at 1.05 to 1.08. The
ART pB ranged between 5.4 and 5.6. CaCO, content in the reagent averaged
95% and MgCO, content about .2%. The limestone grind during the test were
found to be an average of 90% finer than 90 micron.
The excellent SOj removal efficiency was demonstrated during every test.
The results are predictably consistent with prior data from other
installations using the Thyssen FGD systems. Very efficient contact of
limestone slurry and the flue gas in the absorber and the enhanced mass
transfer contact area provided by the Falling Film Contactor zone (FFC)
are the main reasons for the high SO. removal rate. Reagent utilization
overall was found to be 98% or higher. The high SOj removal efficiency was
obtained without the use of any additives.
D) Other Performance Parameters
Power consumption, maJce-up water consumption and pressure drop were also
measured. During the test, the power consumption was found to be
approximately 5% lower than the guaranteed value. Hater consumption was
197.5 gpm. Pressure drop measured across the absorber was 95 mm WC (3.74
in WG) and 120 mm wc (4.72 in. WG) across the regenerative reheater.
7-97
-------
E) ^YT**""1 Dwaterina
The under flow from the hydrocyclone system was measured at 55 to 65%
aolide. The moisture content of the vacuum filter cake ranged between 9
and 10% during the test. The purity of the gypsum filter cake was
determined at 96-98% gypsum.
1. INTRODPCTION
The Bertin process provides an environmentally safe method for recovery of
solid sulfur from flue gas from coal fired boilers or SO2 producing
chemical processes. The process takes place in three (3) distinct stages.
In the first stage, the SO* is removed from the flue gas, by either wet
scrubbing, or by All-Dry scrubbing. The second stage consist of
biotransformation of the sulfites or sulfates to BjS through bacterial
action. The transformation occurs in an anaerobic digester. The third
stage consist of catalytic transformation of the BjS into solid sulfur
utilizing a modified Redox process.
The environmental benefits of the Bertin System are that the reagent
material is recycled from the second stage back to the first stage and
also that the sulfur in the SO2 is recovered as solid sulfur.
2. ABSORPTION OF THE GASEOTJS SO,
The first stage of the process consist of absorption of the SO2 as SO,"2.
This is achieved in a flue gas desulfurization plant using a reagent. The
reagent can be CaCO, (calcium carbonate) or sodium bi-carbonate or sodium
carbonate. The flue gas desulfurization can be performed with the
equipment described in the early part of the paper, in order to meet the
SO2 removal efficiency. For smaller size utility and industrial
installations, the wet scrubbing process can be replaced with Procedair's
enhanced All-Dry scrubbing, which is described in a separate paper. This
All-Dry system provides the same efficiency as wet scrubbing systems. The
SO; can also be removed in fluid bed boilers by the addition of lime stone
to the feed. In the second two cases the calcium sulfate or calcium
sulfite CaSO, or CaSO, are recovered in the fly ash.
3. BIOTRANSFORMATION OP SULFITES AND SULFATES
The transformation of SO/2 into B~S takes place in an aerobic digester.
This digester is mechanically agitated. The digester is first fed with an
anaerobic sludge from a municipal waste water treatment plant. This
sludge contains the sulfur reducing bacteria. The bacteria also need
carbon so that the S04- can be transformed into BJS. The carbon is
normally provided in the anaerobic sludge from waste water treatment
plants molasses or other sources. The fermentation temperature is about
35*C (95*F) and the pH can vary between 6 and 9. The sulfate concentration
in the mixture is maintained at about 50 gr/L. When the scrubber liquor
in the first stage is calcium carbonate (CaOO,) or ammonia carbonate (NH«)2
COj, the reagent is regenerated in the digester and recycled to the first
stage.
The HjS that is produced by the action of the bacteria on the sulfates and
sulfites is evolved as both gaseous BjS and also carried dissolved in the
water formed in the system. The water, containing the E~S, is sent to a
conventional stripper tower where the BjS is removed from the water by
means of air. The clean water is circulated back to the first stage of
the system when a wet scrubbing system is used.
7-98
-------
4. OXIDATION OF HYDROGEN SULFITE TO ELEMENTAL SULFA
Oxidation of -the hydrogen sulfide to elemental sulfur is achieved by means
of oxidizing catalyst dissolved or suspended in liquid media. The
modified Redox process used, requires Fe2*/Fe3* as catalyst. The BjS
evolved in the digester and the air and HjS released in the stripping
column are passed through a gas liquid tower where the H2S is absorbed in
a solution containing the catalyst. The HjS is oxidized to water and solid
sulfa is precipitated according to the following reaction:
2Fe3* + H^S -» Fe2* + 2H* + S
Fe:* iron is simultaneously generated into Fe3* by the excess oxygen
present in the gas:
2Fe:* + 1/2 O, -• 2Fe3* + Oj
This means that the overall reaction is as follows:
H?S + 1/2 O, - Hp + S
All these reactions occur at ambient pressures and temperature and there
is no reagent usage. The absorber towers use inclined trays to avoid
plugging problems due to precipitation of the solid sulfur. The sulfur is
separated from the liquid by conventional dewatering techniques. The
clean catalyst solution is recycled for reuse.
The recovered solid sulfur cam further be purified in a sulfur melter to
obtain the required purity for acid generation.
7-99
-------
I
ii £/
7-100
-------
CflpUtfon rie SO2
(procAM icc 011 Inimldle)
| IHolranfifornintton \
| ric£ srtilflies nn «utfnlcn en lljS |
| pni'bfidtfvlcs fliilfnto-ririaelrlces |
Transfortwtflft* cafaljrlfquc
lie HjS en mnifra sollile
J
I
I
S02 + Cn003 ~COj i C11SO3'
J
'1/2 Cfrj I- C«C03 * HjS-
f?
n2o + .1/2 c*
Recydngo
1/2 02
S + H20
i
PROCEDE BTOSOTlffRE (Memcni dn SOtf
SCITKMAJS CINIJTIQIJBS POUR IA DKSULPURATION
1 pjiorl do ortxmt orRanlqne
-------
Intentionally Blank Page
-------
Modeling of Furnace Sorbent Injection Processes
7-103
-------
Intentionally Blank Page
7-104
-------
A. S. Damle
Research Triangle Institute
P. O. Box 12194
Research Triangle Park, NC 27709
ABSTRACT
One possible approach to increase the cost effectiveness of furnace sorbent injection processes is to
enhance S02 removal by injecting the sorbent in a slurry form. Various steps involved in S02
removal by furnace sorbent slurry injection include evaporation of slurry droplets, sorbent particle
heat-up, activation of sorbent by calcination, reduction in active sorbent surface area by sintering,
and sulfation of calcined sorbent An existing model for sulfation and sintering of calcined sorbent,
CaO, based upon a simple grain structure was adapted in a user-interactive computer program for
simulation of the overall S02 removal process. The slurry droplet evaporation, particle heat-up, and
sorbent calcination steps were assumed to be instantaneous in the overall model compared to the
sulfation step. The effect of sorbent injection on furnace time/temperature profile and particle size
reduction was taken into account
Available dry and slurry sorbent injection data collected at the Ontario Hydro Research Division's
Combustion Research Facility were compared with model predictions with good agreement Injection
of sorbent in a slurry form was found to enhance S02 capture in two ways:
• reduced sorbent sintering due to cooling of gases, and
• reduction in sorbent particle size during slurrying process.
Both of these factors improve sulfur capture. The model conectly predicted the observed effects of
injection temperature and particle size on S02 capture for a given sorbent
7-105
-------
MODELING OF FURNACE SORBENT INJECTION PROCESSES
INTRODUCTION
A need for economically feasible S02 abatement in older coal-fired utility plants has promoted
significant interest in low-cost retrofit flue gas desulfurization processes capable of moderate S02
removal. In-fumace injection of limestone in coal-fired boilers is suitable for retrofit applications due
to lower capital cost and shorter lead time for installation. Typically, such a process results in up to
45 percent S02 removal, with only about 20% calcium utilization. Recent pilot-scale studies at the
Ontario Hydro Research Division's Combustion Research Facility (CRF) have indicated that up to 70
percent S02 removal could be achieved by injecting the limestone sorbent as an aqueous slurry.1
The effectiveness of the in-fumace slurry injection process depends on slurry droplet size, injection
temperature, and sorbent properties. For successful application of such a concept, it is necessary to
understand the mechanisms responsible for the increase in S02 removal.
In-fumace injection of limestone and hydrated lime, either in dry or slurry form, involves several steps
leading to S02 capture which include: evaporation of water in slurry droplets, heat-up of sorbent
particles, calcination of sorbent to form highly active CaO, reduction in surface area of active sorbent
by sintering, and sulfation of CaO in the presence of excess oxygen to form CaS04. A number of
experimental and theoretical studies have been conducted to describe the calcination, sintering, and
sulfation steps. A brief analysis of available information on each step is presented describing their
contributions to the overall process and relative time scales. A user interactive computer program
for the overall S02 removal process is described next Existing diy and slurry sorbent injection data
collected at Ontario Hydro CRF are discussed and compared with model predictions.
SLURRY DROPLET EVAPORATION AND PARTICLE HEAT-UP
After injection of a sorbent slurry droplet in a furnace, it is entrained in the flue gas and is exposed to
a certain temperature profile depending upon the location of the injection nozzle in furnace,
dispersion of droplets, and an existing gas time/temperature profile in the furnace. The time required
for evaporation of a droplet can be estimated by determining the rate of heat transfer from the gas to
the droplet and the rate of mass transfer from the droplet to the gas phase. The same heat transfer
relations also apply in determining the time required by the dried droplets to heat up to the gas
7-106
-------
temperature. The cSspersion/mbeing of droplets in the gas phase can influence the localized
temperature and heat transfer rates. However, such information is not usually readily available.
Uniform mixing of sorbent droplets/particles in the gas phase may thus be assumed for simplicity.
The drying behavior of droplets at moderate temperatures in a spray chamber has been stutied
2 3
extensively. For fine slurry droplets of less than 20 jim diameter, the droplets are entrained
rapidly in the gas phase, and the heat and mass transfer coefficients may be approximated by the
respective Nusselt and Sherwood numbers of 2.4
The heat transferred from gas phase to the droplet is used to evaporate the moisture. Thus, the
interrelated simultaneous heat and mass transfer relationships were used to estimate drying times of
a 20 pm droplet with an initial moisture content of 80 percent (20 percent solids slurry) and for
various furnace injection temperatures. The approximate drying times were found to be 1.8,1.4, and
1.1 ms for furnace injection temperatures of 1,000, 1,200, and 1,400 °C, respectively. The drying
time of a droplet at a given temperature is proportional to the square of the droplet diameter.
During slurry droplet evaporation stage, the droplet temperature is held close to its boiling point (i.e. ~
100 °C for atmospheric pressure operation). Upon drying, the dried particle is heated from 100 °C to
the gas temperature by heat transfer from the gas phase to the sorbent In case of dry injection, the
sorbent particle is heated from its initial feed temperature. The decomposition of the sorbent to CaO
may possibly begin during particle heat-up. However, for the purpose of estimating the partide heat-
up time, the sorbent decomposition may be assumed to occur after partide heat-up. For a 10 jim
limestone partide. with an initial temperature of 100 °C injected in furnace at 1,200 °C, the time
required to heat the partide to 1,000,1,100,1,150, and 1,190 °C temperature may be estimated to
be 0.6, 0.9,1.1, and 1.8 ms. respectively. (Properties of air and sorbent at a mean gas film
temperature were used.) The partide heat-up time is again proportional to the square of the partide
diameter.
Furnace Time/Temperature Profile
The heat consumed during water evaporation, sorbent particle heat-up and endothermic caldnation of
sorbent reduces gas temperature significantly. Since the furnace gas temperature profile dictates the
rates of various subprocesses, the furnace temperature drop due to soi'oent injection must be taken
into account The amount of sorbent added into the furnace depends upon the Ca:S ratio used and
the gas-phase S02 concentration. The amount of water used depends upon the weight fraction of
solids in slurry. The resulting expression for furnace temperature drop, AT, due to limestone sorbent
injection is given as:
7-107
-------
A r = 1.042 x 10"5 (SR)
396.31 + 0.28 T^j + (— - 1)(561.0 + 0.545
T/n/)] •
(D
A similar expression may also be written for hydrated lime sort>ent injection:
A7" = 7.712 x 10"® (SR)
317.89 + 0-28 TlnJ + (JL - 1)(561.0 + 0.545 rto/)
(2)
where
AT is drop in furnace temperatures, °C
CSo2 is S02 concentration in furnace in ppm;
SR is CaS stoichiometric ratio;
Tjnj is furnace temperature at the point of sorbent injection °C; and
w is weight fraction of solids in sorbent feed.
CALCINATION OF SORBENT
Following heat-up of the injected sorbent particles/slurry droplets, the limestone or hydrated lime
sorbent is calcined at furnace conditions to produce highly porous and reactive CaO. In-situ calcined
CaO has shown to be more reactive than injection of pre-calcined CaO in the furnace.5 During
calcination, calcium carbonate or calcium hydrate decomposes producing CaO according to the
following reactions:
CaCQ (s) - CaO (s) + CQ (£f)
w)
Ca(OH)2 (s) - CaO (s) + «jO (0 .
The thermal decomposition of sorbent involves three potential rate controlling steps: (1) heat transfer
to the surface and then through the CaO product layer to the reaction interface, (2) mass transfer of
C02 away from the interface through the product layer, and (3) the chemical reaction. A number of
studies have been conducted to determine the controlling step during calcination. Although earlier
studies conducted with large particles (1 to 2 cm diameter cylinders) indicated dominance of C02
mass transfer and heat transfer,6,7 recent data by Borgwardt8 and analysis of earlier data by Beruto
and Searcy9 clearly indicate the chemical kinetics as the controlling mechanism for calcination of
small dispersed particles. The studies by Borgwardt8 with small particles ranging from 1 to 90 jm*
show that the rate of decomposition of CaC03 at any time is directly proportional to the BET surface
area of the remaining unreacted limestone:
7-108
-------
d(CaCOJ iA
^ ~Ke ^COO,- v '
The reaction rate constant Kc, exhibited an Arrhenius type correlation with temperature. Calcination
of hydrate can occur at a much lower temperature than that of carbonate. Thus, at furnace
conditions, hydrate may be expected to decompose much more rapidly than carbonates. Data from
Bortz and Flament10 indicate that a hydrate decomposes about 50 times faster than a carbonate.
Flow reactor data obtained with 10 jim limestone particles by Borgwardt8 indicate that at a
temperature of 1,075 °C, 85 to 89 percent of limestone was converted to CaO in a residence time of
0.13 to 0.23 s. At furnace temperatures greater than 1,075 °C, the carbonate decomposition would
be even more rapid. Data collected by Mozes et al.11 at the Ontario Hydro CRF indicate essentially
complete calcination within the first section of furnace with a residence time of - 0.2 s. At the same
time, very little sulfation of calcined CaO was observed (<5%) during the same period.
From the overall modeling standpoint the time scale of the calcination step with respect to sulfation is
very important. Because the calcination is much more rapid than the sulfation reaction, the two
processes may be considered sequential with the calcination step being independent of sulfation.
One important parameter evolved in calcination is the active sorbent surface area The solid phase
decomposition process is not understood well enough to predict the surface area generated. For a
given application, the surface area of freshly calcined sorbent must be assigned based upon
experimental data. The literature data (e.g., ref. 8) and the modeling studies by Silcox et al.5 indicate
the carbonate calcine areas of the order of 50 to 60 m2/g, whereas those of hydrate calcines in the
range of 100 m2/g.
Another important parameter evolved in the calcination stage is the sorbent porous structure and
porosity. A nonporous CaC03 or Ca(OH)2 may be expected to produce porous CaO of 54 and 49
percent porosities, respectively. Any porosity inherent in limestone may be expected to increase the
porosity of the calcine. Studies by Mozes et al.11 with Beachville limestone with an initial porosity of
17 percent indicated calcine porosities in the range of 25 to 48 percent, significantly less than 54
percent instead of being greater than 54 percent This could be due to the inherent difficulties
associated with calcine sampling without altering its porosity and surface area as noted by Milne et
al.12 The porous structure of the calcine is important in selecting an appropriate model for sulfation
reaction. Recent scanning electron microscopy (SEM) analysis of calcined sorbent samples by Milne
et al.12 indicate a sorbent structure composed of small interconnected grains. These studies also did
not indicate any significant change in overall sorbent particle dimensions during calcination.
7-109
-------
Sinterina of Calcine
In a number of experimental studies, the surface area of active CaO has been shown to decrease
with time at high enough temperatures.5,12'16 The surface area loss is attributed to sintering process.
Sintering is the mechanism by which small solid particles coalesce when heated to high enough
temperatures below their melting point Sintering affects the surface area and porosity of the sorbent
and reduces its effectiveness for S02 capture. Sintering commences immediately after formation of
high surface areas during calcination and continues throughout the sort>enfs residence time through
the sulfation temperature window in the furnace. Thus, surface area reduction by sintering must be
taken into account simultaneously with the sulfation process.
A basic two-sphere model was used by German and Munir17 to develop a correlation for surface area
reduction:
(vr-*'-
where S is calcine surface area, at time t. SD is initial surface area, Ks is the sintering rate coefficient
and y is a constant dependent upon the mechanism of grain migration.
Although this equation can fit sintering data well, it indicates that two calcines of same surface area
can have different sintering rates based upon the initial starting surface areas in each case. An
empirical approach suggested by Silcox et al.18 has also been shown to fit data well:
= -KJiS - SJ2 . (6)
at
where S^ is an asymptotic surface area that a calcine can reach after a long period of sintering. The
sintering process has been shown to depend upon gas phase C02 and H20 concentrations by
Borgwardt.14 The C02 and HzO concentrations in a given furnace are likely to be in a narrow range,
thus the abovft simple expression may still be adequate in a given case. As shown by Borgwardt8 an
Arrhenius typ<- • -jrrolation can be used to describe variation in Ks with temperature, Ka = A exp (B/T).
Along with surface area, the sintering process has also been shown to reduce porosity of sorbent
material. As explained by Borgwardt,15 the sintering process may involve an induction period for
porosity decline, which has been shown to be about 3 min at 950 °C. Because this order of
magnitude of time is much greater than the furnace residence time, the effect of sintering on porosity
may be ignored in the furnace injection modeling.
7-110
-------
SULFATION OF CALCINED SORBENT
Sulfation of the calcined CaO is a n on catalytic gas-solid reaction and involves several potentially
controlling steps: (1) transfer of S02 from the bulk gas phase to particle surface, (2) diffusion of SOz
to the reactive surface within a particle, and (3) reaction between SOz and solid CaO. These steps
are further complicated by the fact that the reaction product CaS04 has a greater molar volume than
the CaO sorbenL The product is formed on the reaction surface, and, due to expansion, the product
formation will constrict or even completely fill the porous structure of the calcine, thus limiting the
sorbent utilization.
To accurately predict the S02 removal efficiency and sorbent utilization, the continuous effect of
product formation on the porous structure must be adequately taken into account The product
buildup on the reactive surface provides a mass transfer resistance for S02 transport to the reactive
surface, which is taken into account by defining an effective diffusion coefficient Various models
reported in literature, try to express the sorbent porous structure mathematically, e.g., a single pore
model,19 a random pore model,20 a pore tree model,21 or an overlapping grain model.22 A number of
modeling and experimental studies have been reported specifically for the CaO and S02
reaction.5,23"33 The lime deactivation mechanism in most of these models is considered to be the
product buildup and slow diffusion of S02 through the product layer.
In the SEM analysis of the calcined sorbent its structure appears to be composed of small spherical
interconnected grains.12 The grain structure model may be based upon a simple individual grain
structure, as in Silcox et al.,5 or an overlapping interconnected grain structure, as discussed by
Lindner and Simmonsson.22 The complex overlapping grain model involves more model parameters,
but has the ability of describing porosity reduction with sintering. The simple grain model cannot
realistically simulate porosities lower than 26 percent However, the initial porosities of calcined
sorbent are expected to be greater than 50 percent, and the sintering induced porosity reduction may
not be significant during the furnace time scale of 1 to 2 s. The porosity of sulfated sorbent may
decrease with a high degree of conversion; however, conversions during furnace injection processes
have typically been less than 30 percent. Thus, a simple grain model may be adequate for
simulation of in-fumace sorbent injection processes. The reduction of surface area due to sintering
can easily be taken into account by increasing the grain size appropriately with time, and the porosity
of the sorbent may be assumed to decrease solely because of product formation as dictated by
conversion.
7-111
-------
Grain Model Formulation
The grain model formulation as described by Hartman and Coughlin28 for the S02-lime reaction
considers a spherical porous lime particle as composed of solid spherical nonporous CaO grains of
uniform size, separated by pores. The reacting S02 is first transported from bulk gas phase to the
particle surface, which then diffuses through the porous network to a grain reaction surface. As the
reaction proceeds, a shell of the CaS04 product is formed on the surface of the grain. This product
layer provides an additional mass transfer resistance to the diffusing S02. The sulfation of individual
CaO grains is assumed to follow a shrinking core model. The sulfation reaction is assumed
irreversible. The lime particles are assumed to be spherical, with no temperature gradients within a
particle. The reacting surface is assumed to be CaO, i.e., the calcination is assumed to be complete
in a grain before sulfation commences. The final product is assumed to be CaS04 due to an excess
of oxygen in the gas phase.
The external mass transfer of SOz to particle is expected to be much faster as compared to internal
pore diffusion for small particles below 50 (im and thus is not expected to be a rate controlling step.34
The transport of sulfur dioxide within a spherical particle of lime can be described by a diffusion
equation:28
£?C 2 dC N n t7i
-D,'0 • m
where C is the gas-phase S02 concentration at a radial position, R, R is radial distance from center,
N is S02 consumption rate, and De( is an effective pore diffus'rvity of S02. The effective diffusivity is
given by:
D -
Dm + Dk
where Dm is gas phase molecular diffus'rvity and Dk is Knudsen pore diffusivity.
The S02 reaction rate, N, depends upon both the intrinsic kinetic rate and diffusion of S02 in product
layer. For a first order reaction kinetics with respect to S02
concentration:
N-
' (9)
D.
1 _ _L
7-112
-------
where
Kr = intrinsic reaction rate coefficient,
A = surface area per unit volume,
C = gas phase S02 concentration at a radial location R within a lime particle,
n = reaction order, = 1 for first order
D, = diffusivity of S02 in the product layer,
r = radius of reaction interface within spherical grain,
r^ = outer radius of partially sulfated grain.
The reaction rate is expressed in terms of a reaction order n. The S02-lime reaction has been
studied in a differential reactor by Borgwardt35 with lime particles in the size range of 0.01 to 0.1 cm.
These data indicated the reaction order with respect to SOz of 1 for particles smaller than 0.05 cm.
However, recent experiments by Borgwardt and Bruce25 indicated that for very small particles of 1
ixm, the rate of CaS04 formation increased with the 0.62 ± 0.07 power of the S02 partial pressure.
At these small particle sizes, the effect pore diffusion was eliminated. The reactivity of these particles
also increased with the square of the BET surface area of the calcine, which may be interpreted as
due to dominant diffusion resistance in the product layer. In flow reactor experiments by Coutant et
al.36 with 90 nm particles, a reaction order of 0.5 was indicated. Thus, a reaction order of 0.5 to 0.6
with respect to S02 concentration seems applicable for particle sizes relevant to furnace injection.
To utilize the grain model formulation as described above, correlations are needed for the intrinsic
reaction rate coefficient, Kr and product layer diffusivity, D(. The fundamental data obtained with 1
jim particles25 have been used by several investigators to obtain correlations for K,. and Ds.
Borgwardt and Bruce found that an Arrhenius type correlation best described the variation of Ds with
respect to T 25 The reaction rate coefficient may also be expressed with an Arrhenius type
correlation.5
The reaction surface area. A, is obtained from the sorbent grain radius, r^. This surface area is
reduced with time due to sintering process. In the grain model, this may be accounted for by
increasing the grain radius, r^ continuously with time. The sintering rate correlations are used to
change r^ appropriately. The sintering and sulfation processes may thus be considered
simultaneously in a grain model.
OVERALL MODEL STRUCTURE
The sorbent calcination, sintering, and sulfation processes are complex and details of these
processes are not very well understood. Each of these processes depends upon several important
7-113
-------
parameters, e.g., calcine surface area, porosity, and various kinetic correlations. Determination of
these parameters experimentally is difficult, and only limited data are available which often provides a
combined effect of these processes. Thus, in developing an overall mode) of this process, emphasis
needs to be placed on simplicity and keeping the required model parameters to a necessary
minimum.
The overall time scale of the in-furnace processes is about 1 to 2 s. The heat transfer steps of slurry
evaporation and particle heat-up are of the order of a few milliseconds. Thus, these steps may be
assumed instantaneous with respect to furnace time scale. The dispersion/mixing of sorbent in the
gas phase is assumed to be uniform. The change in the furnace time/temperature profile resulting
from sorbent injection is assumed to be instantaneous. The sorbent calcination step for carbonates
has been shown to be of the order of 0.1 to 0.2 s. For hydrates, this step is much faster. The
calcination step provides the surface area and porosity as the input parameters needed for the
sulfation model. Although the carbonate decomposition time is significant with respect to overall time
scale, this step was considered instantaneous with respect to sulfation in this effort for the following
reasons: (1) calcination kinetics is an onder of magnitude faster than sulfation kinetics; (2) the
present understanding does not provide any estimate of fresh calcine surface area and porosity, the
key parameters for sulfation reaction, which still need to be assumed; and (3) calcination and
sulfation may be considered sequential processes, with calcination being independent of sulfation.
The fresh calcine surface area and porosity may simply be used as user-supplied input parameters
for the overall model, which now primarily would be simultaneous sintering and sulfation model. A
sorbent structure composed of individual uniform sorbent grains represents a simple porous structure
defined by a single parameter, grain radius. The grain radius is directly related to the calcine surface
area. The calcine sorbent porosity is assumed independently as another user-supplied parameter.
This simple formulation is also able to incorporate sintering and surface area change by simply
increasing the grain radius with time.
Sulfation and Sintering Model
An existing computer code "S02EPA" developed by Dr. G. Silcox of the University of Utah, based
upon the simple grain structure, was adapted in the overall model in this effort This program
simulates simultaneous sintering and sulfation of calcined CaO under furnace conditions. This code
uses finite difference formulation to solve the equation describing S02 diffusion and reaction. The
reaction order with respect to S02 concentration is assumed to be 0.55. Due to the nonlinearity
introduced by this reaction order, a Newton-Raphson iteration technique is used to solve the
difference equations. Details of this code and solution procedures are given by Silcox et al.5
7-114
-------
The input parameters to this model include furnace S02 concentration, Ca:S ratio for sorbent
injection, furnace time/temperature profile as "seen* by injected particles, sorbent particle diameter,
initial and asymptotic BET surface area of calcine, initial porosity of calcine, weight fraction of CaC03
in raw stone, porosity of product layer, partial pressure of Oz in furnace gases, and residence time.
Additional input parameters related to numerical procedures also need to be specified. The
adaptation of the "S02EPA" code in the present work involved:
• changing the user specified furnace time/temperature profile to take into account
heat effects;
• allowing the user to specify an injection temperature in an existing furnace
temperature profile; the program recalculates the temperature profile as "seen" by
the sorbent;
• modifying the Newton-Raphson iteration procedures to allow faster convergence;
• allowing for particle size change by fragmentation as specified by user;
¦ allowing all input parameters to be entered interactively; and
• allowing changing various kinetic rate coefficients.
User-Interactive Data Input
A user-interactive computer program 'FSI' was developed in this effort to facilitate simulation of
furnace sorbent injection processes.37 This program allows: (1) entering all the data required to run
a simulation interactively, (2) saving all the data in a user designated file, (3) reading data from an
existing data file, and (4) running the sulfation model using the current data values. The model
output can either be saved in a user-specified file, or printed directly on a printer. At startup, all the
data values are initialized with default values. All the data input screens are self-explanatory, and
additional instructions are available in help files.
The user supplied inputs are divided into five categories: (1) sorbent parameters, (2) furnace
time/temperature profile, (3) sorbent injection operating conditions, (4) model simulation parameters,
and (5) reaction and diffusion rate coefficients.
MODEL SIMULATIONS
The concept of in-fumace injection of a sorbent in a slurry form for enhanced S02 removal has
recently been demonstrated in pilot-scale studies at Ontario Hydro CRF.1 In earlier studies, the
effectiveness of dry injection of sorbents was investigated.11,38 Thus, these data provide direct
7-115
-------
comparison of results of sorbent injection in both dry and slurry form. The data collected in these
studies were used for model simulation in present studies.
Experimental Database
The dry sorbent injection data collected at Ontario Hydro CRF include tests conducted with two
different limestone sorbents: Beachville limestone and PL Anne limestone. The chemical and
physical properties of these sorbents relevant to model simulation are given in Table 1. The details
of the combustion research facility (experimental conditions and operating procedures) are given in
references 1,11, and 38.
Bulk of the experimental data simulated a furnace quenching rate of 500 °C/s. The soibents were
injected at different locations in the furnace corresponding to injection temperatures in the range of
1,000 to 1,350 °C. The sorbent to sulfur ratios during these tests ranged from 1.5 to 3.0.
The dry sorbent injection S02 capture data indicate substantial influence of sorbent injection
temperature on SOz capture for all sorbents. The different sorbents showed differing reactivities
toward S02. The sorbents tested differ significantly in the porosities and sorbent particle sizes. The
slurry injection data indicated 5 to 30 percent enhancement in S02 capture compared to dry sorbent
injection under similar conditions. The effect of injection temperature in slurry injection tests was less
dramatic than that during dry injection. In both cases, the optimum temperature for sorbent injection
was found to be about 1,200 °C.
Model Parameters
Model simulations require a number of sorbent properties and rate coefficients. Sorbent properties
include particle diameter, surface area and porosity of fresh calcine, and particle fragmentation
estimate if any. The rate coefficients include the two coefficients in the two-parameter Arrhenius
correlations for the intrinsic reactivity, product layer diffusivity, and sintering rate (K^, Ds1, D^,
KsV and K^, respectively). During these simulations the surface area of fresh calcine from
carbonates was assumed to be 60 m2/g. The inherent sorbent porosities were taken into account in
estimating the calcine porosities; thus the porosity of calcine from Beachville limestone was taken as
0.62 and that from PL Anne limestone was taken to be 0.7. A number of simulations were conducted
with dry injection data to identify a common set of rate coefficient values for the data involving
calcium carbonates. The values found to give reasonable fit with data were:
Krt = 50.0, Dal = 1.85 x 10"3, and K,, = 25.0.
The activation energies were = -12540, D^ = -21200, and = -19566, all in K.
7-116
-------
The activation energies were kept constant in all simulations. For slurry injection simulations it was
necessary to change one or more rate coefficients to improve fit with observed data. The values
used in these simulations are given in Table 2.
Dry Injection Simulation
The simulation results are shown graphically in Figures 1 and 2 to highlight the effect of sorbent
injection temperature and Ca:S ratio on S02 capture. As seen from these figures, the model
predicted the observed trends in the effect of temperature and Ca:S ratio with good agreement in the
data and model predictions. For Beachville limestone, the model predictions indicated 1,200 °C as
an optimum injection temperature for S02 removal as observed experimentally.
During these simulations with Beachville limestone and PL Anne limestone the same reaction and
diffusion coefficients were used. Thus the differences in the observed S02 capture were explained
by smaller particle size and high porosity of Pt. Anne limestone sorbent Additional simulations
indicated that decreasing particle size below 3.0 jun did not further increase SOz capture significantly;
thus the pore diffusion resistances appear to be important only for particles greater than 3.0 jim in
diameter.
The model was able to explain the low S02 capture for sorbent injections at higher them optimum
temperatures for both sorbents. Such sorbent injection leads to rapid sintering and loss in surface
area while the temperature is still high enough to prevent sorbent sulfation due to reaction equilibrium
considerations. Thus the surface area is reduced without any S02 capture resulting in lower overall
sorbent reactivity.
Slurry Injection Simulation
These simulation results are shown graphically in Figures 3, 4, and 5 to highlight the effect of
injection temperature and Ca:S ratio on S02 capture. During the slurry injection experiments, two
injection modes were identified depending upon the orientation of the injection nozzles with respect to
furnace gas flow.1 The injection mode presumably influenced the actual furnace temperatures as
'seen' by the injected droplets, and significant differences in S02 capture were observed in the two
injection modes at a given furnace location. From the modeling standpoint, the difference in the
injection modes can only be accounted if the difference in the injection location temperature and the
temperature actually experienced by the droplets is known. Since this information could not
accurately be obtained, the furnace temperatures reported at the injection location were used in
model simulation regardless of the injection mode.
7-117
-------
As seen from Figure 3 for Beachville limestone slurry injection, the model appears to simulate the
cocurrent injection mode results better than the countercunrent mode results. The model is seen to
correctly predict the trend in the observed effect of injection temperature. The location of optimum
temperature peak as predicted by model matches closely with cocurrent mode data.
As seen from Figures 3 and 4, the model predicted the SOz capture levels at the optimum
temperature very well; however, the model underpredicted the S02 capture observed at lower
temperatures. The observed reactivities at lower temperatures did not drop with temperature as
rapidly as the model predicted. The data plotted in Figure 5 show the effect of Ca:S ratio on S02
capture at the optimum injection temperature. Good agreement is seen between the model
predictions and observed values.
At temperatures higher than the optimum level, the model correctly predicted increased SOz captures
when compared with the dry injection results. This results from reduced sintering effect at high
temperature injections because of substantial cooling of the gas phase. At a Ca:S ratio of 3 the
injection of sorbent in the slurry form decreases the gas temperature by an additional 60 to 70 °C,
thus preventing rapid sintering without sulfation at high temperatures. As seen from Table 2,
simulation of slurry injections data required greater intrinsic reactivity and greater product layer
diffusivity values for improving the agreement between model and data. This certainly indicates
enhanced reactivity of sorbent by injecting the sorbent in a slurry form beyond that explained by the
obvious sintering effect
The sorbent particle size in slurry injection mode is determined from the slurry droplet size and the
solids fraction in slurry. This assumption explained the observed lower S02 captures obtained with
17 jim slurry droplets as opposed to 6.0 jim slurry droplets of PL Anne limestone sorbent The effect
of particle size results from pore diffusion resistances which were found to be important for dry
particle sizes greater than about 3.0 jim. Any further decrease in particle size did not appear to
improve S02 capture significantly in model simulations. The 3.0 iim dry particle size corresponds to
about 5.0 to 6.0 (im slurry droplet size for 40 percent by weight slurry. Thus the simulations concur
with the experimental finding of optimum slurry droplet diameter of 6.0 jim.
CONCLUSIONS
The following conclusions may be drawn from the furnace sorbent injection modeling efforts and
simulation of experimental data
• The model simulations of dry sorbent injection pilot plant data indicated very good
agreement The model correcth- predicted the effect of injection temperature and
7-118
-------
particle size. The lower S02 capture observed by injecting sorbent at a higher
than optimum temperature was attributed to rapid sintering and surface area loss
without any significant S02 capture at high temperatures.
The model successfully simulated observed S02 captures during slurry injection at
optimum temperatures. The model underpredicted observed S02 captures at
lower than optimum temperature injections.
The model indicated reduced sintering effects in slurry sorbent injection as
compared to dry sorbent injections. This fact improved performance substantially
at higher than optimum temperature injections.
Slurrying process reduced effective sorbent particle size especially for Beachville
limestone which was found to improve S02 capture.
The model predictions were in better agreement with cocurrent slurry injection
data than countercurrent slurry injection data. The model, however, correctly
predicted the existence of optimum injection temperature as observed.
ACKNOWLEDGEMENT
Financial support for this work provided by the Canadian Electrical Association (Contract No. 824 G
714) is gratefully acknowledged. Helpful discussions and pilot plant data provided by Dr. G. Thomas
and Mr. R. Man gal of Ontario Hydro Research Division and Dr. D. Smith of Sask Power are sincerely
appreciated. Special thanks are due to Dr. Geoff Silcox of the University of Utah for providing a copy
of the 'S02EPA' computer program and for useful discussions.
REFERENCES
1. M.S. Mozes, R. Mangal, R. Thampi, and M. McDonald. In-Fumace Sorbent Slurry Injection for
S02 Control. Final Report for CEA, Project Number. 712G619, 1989.
2. W.R. Marshall, Jr. "Heat and Mass Transfer in Spray-Drying." Trans. ASME, 1955, vol. 77, pp.
1377-1385.
3. W.H. Gauvin and S. Katta "Basic Concepts of Spray Dryer Design." ASChEJ, 1976, vol. 22,
no. 4, pp. 713-724.
4. A.S. Damle. Modeling of S02 Removal in Spray Dryer Flue Gas Desulfurization System. U.S.
EPA Report 600/7-85-038. NTIS No. PB1361651AS. December 1985.
5. G.D. Silcox, S.L Chen, W.D. Clark, J.C. Kramlish, J.F. LaFond, J.M. McCarthy, D.W. Pershing,
and W.R. Seeker. Status and Evaluation of Caldtic S02 Capture: Analysis of Facilities
Performance. EPA-600/7-87-014 (NTIS PB87-194783), EPA, Washington, DC. May 1987.
6. C.N. Satterfield and F. Feakes. "Kinetics of the Thermal Decomposition of Calcium Carbonate."
AlChEJ, 1959, vol. 5. p. 115.
7. A.W.D. Hills. "The Mechanism of the Thermal Decomposition of Calcium Carbonate.' Chem.
Eng. Sd., 1968, vol. 23, p. 297.
7-119
-------
8. R.H. Borgwardt. 'Calcination Kinetics and Surface Area of Dispersed Limestone Particles."
AlChEJ, 1985. vol. 31, p. 103.
9. D. Beruto and A. W. Searcy. "Use of Langmuir Method for Kinetic Studies of Decomposition
Reactions: Calcite (CaCOg)." J. Chem. Soc. Faraday Trans., 1974, vol. 7, p. 2145.
10. S.J. Bortz and P. Flament "Recent IFRF Fundamental and Pilot Scale Studies on the Direct
Sort>ent Injection Process." In Proc. 1st Joint Symp. Dry SO^NOx Control Techno!., vol. 1, July
1985, p. 17-1, EPA-600/9-85-020a (NT1S PB85-232353).
11. M.S. Mozes, R. Mangal, R. Thampi, and D.L Michaciw. Pilot Studies of Limestone Injection
Process Phase I: Simulating Lakeview TGS Quenching Rate. Ontario Hydro Research Division
Report No. 86-62-K, May 1986.
12. C.R. Milne, G.D. Silcox, D.W. Pershing, and D.A. Kirchgessner. "Calcination and Sintering
Models for Application to High Temperature, Short Time Sulfation of Calcium Based Sorbents."
Ind. Eng. Chem. Res., 1990, vol. 29, p. 139.
13. R.H. Borgwardt, N.F. Roache, and K.R. Bruce. "Method for Variations of Grain Size in Studies
of Gas-Solid Reactions Involving CaO." Ind. Eng. Chem. Fundam., 1986, vol. 25, pp. 165-169.
14. R.H. Borgwardt. "CaO Sintering in Atmospheres Containing Water and Carbon Dioxide." Ind.
Eng. Chem. Res., 1989, vol. 28, pp. 493-500.
15. R.H. Borgwardt. "Sintering of Nascent Calcium Oxide." Chem. Eng. Sci., 1989, vol. 44, no. 1,
pp. 53-60.
16. B.K. Gullet and K.R. Bruce. "Pore Distribution Changes of Calcium-Based Sorbents Reacting
with Sulfur Dioxide." AlChEJ, 1987, vol. 33, pp. 1719-1726.
17. R.M. German and Z.A. Munir. "Surface Area Reduction During Isothermal Sintering." J. Am.
Ceram. Soc., 1976, vol. 59, p. 379.
18. G.D. Silcox, J.C. Kramlich, and D.W. Pershing. "A Mathematical Model for the Rash Calcination
of Dispersed CaC03 and Ca(OH)2 Particles." Ind. Eng. Chem. Res., 1989, vol. 28, pp. 155-160.
19. P.A. Ramachandran and J.M. Smith. "A Single Pore Model for Gas-Solid Noncatalytic
Reactions." AlChEJ, 1977, vol. 23, p. 353.
20. S.K. Bhatia and D.D. Perlemutter. "A Random Pore Model for Fluid-Solid Reactions.
II: Diffusion and Transport Effects." AlChEJ, 1981, vol. 27, p. 247.
21. G.A. Simons. "The Pore Tree Structure of Porous Char." 19th Symp. Int. Combustion,
Combustion Inst, 1982.
22. B. Lindner and D. Simmonsson. "Comparison of Structural Models for Gas-Solid Reactions in
Porous Solids Undergoing Structural Changes." Chem. Eng. Sci., 1981, vol. 36, pp. 1519-1527.
23. T. Bardaki. "Diffusional Study of the Reaction of Sulfur Dioxide with Reactive Porous Matrices."
Thermochimica Acta, 1984, p. 287.
24. S.K. Bhatia and D.D. Perlemutter. "The Effect of Pore Structure on Fluid-Solid Reactions:
Application to the S02-Lime Reaction." AlChEJ. vol. 27, p. 247.
7-120
-------
25. R.H. Borgwardt and K. R. Bruce. "Effect of Specific Surface Area on the Reactivity of CaO with
S02." AlChEJ, 1986, vol. 32, p. 239.
26. P.G. Christman and T.F. Edgar. "Distributed-Pore-Size Model for Sulfation of Limestone."
AlChEJ, vol. 29, pp. 388.
27. C. Georgakis, C.W. Chang, and J. Szekety. "A Changing-Grain-Size Model for Gas-Solid
Reactions." Chem. Eng. Sci., 1979, vol. 34, p. 1072.
28. M. Hartman and R.W. Coughlin. "Reaction of Sulfur Dioxide with Limestone and the Grain
Model." AlChEJ, 1976, vol. 22, p. 490.
29. G.A. Simons and A.R. Garman. "Small Pore Closure and the Deactivation of the Limestone
Sulfation Reaction. AlChEJ, 1986, vol. 32, p. 1491.
30. G.A. Simons, A.R. Garman, and A.A. Boni. The Kinetic Rate of S02 Sorption by CaO."
AlChEJ, 1987, vol. 33, p. 211.
31. J.A. Cole, J.C. Kramlich, W.R. Seeker, and G.D. Silcox. "Fundamental Studies of Sorbent
Reactivity in Isothermal Reactors." In Proceedings: Second Joint Symposium on Dry S02 and
Simultaneous SO/NOx Control Technologies, 1986.
32. B.K. Gullet and K.R. Bruce. "Effect of CaO Sorbent Physical Parameters Upon Sulfation."
Presented at EPA/EPRI First Combined FGD and Dry SOz Control Symposium, SL Louis, MO,
October 1988.
33. C.R. Milne, G.D. Silcox, D.W. Pershing, and D.A. Kirchgessner. "High Temperature Short Time
Sulfation of Calcium Based Sorbents: II Experimental Data." Ind Eng. Chem. Res., 1990, vol.
29, p. 2201.
34. G.D. Silcox. "Analysis of S02-Lime Reaction System: Mathematical Modeling and Experimental
Studies with Emphasis on Stoker Application." Ph.D. Dissertation, the University of Utah, Salt
Lake City, 1985.
35. R.H. Borgwardt. "Kinetics of the Reaction of S02 with Calcined Limestone." Environ. Sci.
Technol., 1970, vol. 4, p. 59.
36. R.W. Coutant, R. Simon, B. Campbell, and R.E. Barrett Investigation of the Reactivity of
Limestone and Dolomite for Capturing S02 from Flue Gas. EPA Report APTD 0802 (NTIS PB
204385). October 1971.
37. A.S. Damle. Sulfation Model for In-Fumace Calcuim-Based Sorbent Slurry Injection. Draft Final
Report for CEA, Project No. 824 G714, September 1991.
38. M.S. Mozes, R. Man gal, and R. Thampi. Sorbent Injection forSOs Control: (A) Sulphur Capture
by Various Sorbents and (B) Humidification. Ontario Hydro Research Division Report No. 88-
63-K, July 1988.
7-121
-------
Table 1
CHEMICAL AND PHYSICAL PROPERTIES OF SORBENTS
Beachville Limestone PL Anne Limestone
MgO g/kg 8.0 4.8
CaO g/kg 524.0 535.4
LOI 434.0
BET area, m2/g 1.3 2.9
MMD, ]im 8.6 3.9
p, g/cm3 2.6 2.3
Porosity % 17.0 55.0
Table 2
RATE COEFFICIENTS USED IN SIMULATIONS
Sorbent
Beachville limestone
Pt. Anne limestone
Beachville limestone
PL Anne limestone
Mode
_Kn_
Dry
50
Dry
50
Slurry
100
Slurry
100
Pi
1.85 X10"3 25
1.85 x 10"3 25
1.85 X10"3 20
3.7 xlO"3 25
7-122
-------
1 / \
/ ¦ y
~ PL Anna
Lmartaoa
Ca:S-3.6
•
¦
Baachviila *
Limaslana
CcS-30
Unas-Model Prediction
Symbotj-Exparimamal Data
—1_ .1
•
«
TO
60 _
Hydated Ums.
3
A
m
O
*4
8
1000 1100 1200 1300
I reaction Tamparature
Figure 1. SO2 Capture a« a Function of Injaction
Tamparature-Dry ln|actk>a
1400
Una» Modal Predcfen
Symbote-Exparfmantal Data
CaS Ratio
Flgura 2. SO2 Captura as a Function of Ca:S
Ratio-Dry Injactkon.
70
Experimental Data
90
SO
30
20
10
1000
1200
1300
1100
Infrrton Tnn—lir*. *C
Flgura 3. SO2 Captura as a Function of
Injection Tamparatura Baachvllla
Umastona Slurry Injactlon.
7-123
-------
80
to
8
I ~
Model Pr*4cbon. 1.334 ppm SO^
MoM Pwdflion. MAppmSOj
Experiment Dflta.1,334 ppmSOa
E^arfewnal Data. M^ppmSOj
30
20
1000
1100
1200
1900
Injection Temperature, *C
Flgura 4. SO2 Captura as a Function of Injactlon
Tamparatura-Pl Ann* Umaaton* Slurry Injaetlon.
TO
R. Aim* UmMlon*
-l.320ppm.SOj
Tl. Ann. Uwlent
BMppmSOi
ao
I
5 90
O
<«
s
40
20
to
CkSRU)
Flgura 5. SO2 Captura aa a Function of Ca:S
Ratio-Sturry Injaetlon.
7-124
-------
DRY FGD PROCESS USING CALCIUM SORBENTS
Hiroyuki Nosaka
Kouji Kuramashi
Kure Works, Babcock-Hitachi K.K.
6-9 Takaramachi Kure Hiroshima, Japan 737
Hirofumi Kikkawa
Kure Research Laboratory, Babcock-Hitachi K.K.
3-36 Takaramachi Kure Hiroshima, Japan 737
Takeo Komuro
Hitachi Research Laboratory, Hitachi Ltd.
832-2 Horiguchi Katsuta Ibaragi, Japan 312
7-125
-------
Intentionally Blank Page
7-126
-------
ABSTRACT
This paper describes the Hitachi Simplified FGD (Flue Gas Desulfurization) System.
As this is a dry FGD system which does not handle a sorbent as a slurry and is
constituted with a few number of auxiliary equipment, compared with semi-dry or wet
systems, the system can be operated easily. Furthermore, it has also features such
as no waste water and small installation space. There are two basic processes in
this system — DLI (Duct Lime Injection) Type and FLI (Furnace Limestone Injection)
Type.
DLI Type is the system in which hydrate lime (Ca(0H)2) as a sorbent is injected
into the duct pneumatically, then flue gas including sorbent is cooled and
humidified by water injection in the reactor. Reaction between Ca(0H)2 and sulfur
dioxide (SO2) is enhanced by this humidification. As sorbent injection in the
boiler furnace is not carried out, this system can be applied easily not only coal-
fired boiler but also oil-fired boiler whether it is new or retrofitting. Uhile in
the FLI Type, limestone (CaC03) as a sorbent is injected into the boiler furnace
where CaC03 is decomposed thermally to produce calcium oxide (CaO) which reacts
with SOj partially in the furnace. In addition, the reactor is also provided to
increase SO2 removal efficiency by cooling and humidification of flue gas similar
to DLI system. FLI system is most applicable to the coal-fired boiler in case the
CaC03 is more suitable for obtaining than Ca(0H)2«
In both DLI and FLI processes, SO2 removal efficiency was achieved above 70 % and
80 % at Ca/S molar ratio of 2 and 3 respectively.
Hitachi Ltd. and Babcock-Hitachi K.K. have performed various tests for this FGD
system, and are proceeding with application to commercial plant.
In this paper we also describe sorbent that had larger specific surface area than
Ca(0H)2 by heat treatment of Ca(0H)2 with certain additives. It was confirmed that
it's reactivity was higher than that of Ca(0H)2»
7-127
-------
DRY FGD PROCESS USING CALCIUM SORBENTS
INTRODUCTION
Recently, the problem of environmental pollution due to acid rain has been focused
worldwide. One of the major factors of acid rain is its releasing of sulfur oxide,
which is generated by combustion of fossil fuels, from thermal power plants.
Hitachi Ltd. and Babcock-Hitachi K.K. have been engaged in the development of FGD
systems for more than thirty years, and in addition to wet limestone-gypsum systems
(1). Ue have established various FGD systems such as dry activated carbon process,
dry absorbents made from fly ash process (2), etc.
As the problem of acid rain has become a global concern, various types of FGD sys-
tems are in demand. Recently the FGD system which is a simple system, easy to
operate and is low in cost, although SO2 removal efficiency is not as high as that
obtained in wet FGD systems, has been particularly demanded in certain countries
due to regional circumstances. In response to these needs, Hitachi Ltd. and
Babcock-Hitachi K.K. have developed and established Hitachi Simplified FGD System
based on research and development as well as extensive experience so far.
PROCESS DESCRIPTION
There are two basic processes — DLI (Duct Lime Injection) and FLI (Furnace Lime-
stone Injection) Types. These two processes vary according to the point at which
the sorbent is injected into the system. In the DLI type, the sorbent is injected
into the flue duct downstream of the Air Preheater, whereas in the FLI Type, the
sorbent is injected into the boiler furnace. Selection between the two available
processes depends upon the kind of boiler fuel (Coal or Oil) and desulfurization
sorbent used (Lime or Limestone). Both processes are provided with a reactor where
humidification and SO2 absorption occur.
Flow diagrams of the two processes are outlined in Figure 1.
Key parameters for SO2 removal are as follows.
• Ca/S molar ratio
• gas temperature (reactor outlet)
• residence time in furnace
• temperature at sorbent injection zone in the furance
TEST PROCEDURE AND RESULTS FOR DLI AND FLI PROCESS
2,000 Nm^/h Pilot Plant
Flow sheet and photograph of this plant are shown in Figure 2 and 3 respectively.
7-128
-------
Flue gas for testing was prepared by injecting SO? and fly ash into the flue gas
from the boiler. Under the standard condition, SO2 concentration in the inlet gas
was kept at 1800 ppm. After the sorbent was injected pneumatically into the flue
gas, it was cooled by humidification in the reactor. The injection nozzle of sor-
bent was located on the inlet duct of the reactor. Humidification was carried out
by injection of fine mists from dual fluid spray nozzles and these mists evaporated
completely in the reactor.
Having reacted with SO2 in the reactor, the sorbent was collected in ESP downstream
of the reactor. SO? concentration in the inlet and outlet gases were measured by
an infrared SO2 analyzer in order to determine SO2 removal. The dimensions of the
reactor were 0.85 m in diameter and 10 m in height. Inlet gas temperature of the
reactor was 150 °C.
Test Furnace
A photograph and outline of the test furnace are shown in Figures 4- and 5 respec-
tively. The furnace was a vertical type having four sorbent injection nozzles,
paired on the front and rear wall. Sorbents injection points were located on the
zone where the temperature were as high as 1050 °C and 1150 °C, and the sorbent was
injected pneumatically into the test furnace.
Reacted sorbent was collected in the bag filter downstream of the furnace.
Collected sorbents were injected into the pilot plant reactor for FLI process
evaluation. The flow sheet is shown in Figure 6.
Specification of the test furnace were as follows.
* Coal consumption : 4 t/h
• Combustion gas flow rate : 30,000 Nm3/h
Evaluation of PL I Process
In this process, Ca(0H)2 was injected into the reactor, while water was sprayed
separately. The average particle size of fine mist from the dual fluid spray
nozzle was approximately 50 um derived by a laser diffraction technique.
Figure 7 shows the dependence of SO2 removal on the outlet gas temperature, the
relative humidity and injected Ca/S molar ratio. As shown in this figure, SO2
removal increased as the outlet gas temperature decreased (relative humidity
increased) or Ca/S molar ratio increased (3). It was found, however, that the out-
let gas temperature should be kept at least 10 °C ove. '.he saturated temperature to
avoid deposition of the reacted sorbent in the downstream duct or ESP.
Test results indicated that SO2 removal as high as 70 and 80 % or more was achieved
at Ca/S molar ratio of 2 and 3 respectively at an approach-to-satuation temperature
of 10 °C.
The effect of SO2 concentration in the flue gas on the SO2 removal was also inves-
tigated. The result is shown in Figure 8. In the range of 500 - 3000 ppm of SO2
7-129
-------
concentration, the SO2 removal is considered to be independent of SO2 concentra-
tion.
Evaluation of FLI Process
Evaluation of FLI process was carried out by the result of the furnace and 2000
Nm^/h pilot plant.
^ -his process, limestone (95 % pass through 325 mesh) was used as an sorbent.
-Is*' fired for this study was a bituminous coal containing 0.57 % of sulfur.
The residence time from the injector to the first gas cooler was approximately 1
second and injection velocity of the sorbent was more than 30 meters per second.
SO2 removal in the test furnace was determined by measuring SO2 concentration in
the flue gas at the outlet with and with-out sorbent injection.
From the previous study (4) on the effect of the temperature of sorbent injection
zone in a furnace on SO2 removal, the optimum injection temperature was found to be
1000 - 1100 °C. From this result, sorbent injection points were determined. The
gas condition in the pilot plant was adjusted equal to the outlet gas condition
from the test furnace.
The overall SO2 removal is shown in Figure 9. SO2 removal in the test furnace was
above 30 % at Ca/S ratio of 3, while overall SO2 removal by injecting collected
sorbent in the test furnace into the reactor of the pilot plant was achieved more
than 70 % and 80 % at Ca/S molar ratio of 2 and 3 respectively.
During the above tests, no adverse influences due to fouling or slagging in the
furnace were observed.
TEST PROCEDURE AND RESULTS FOR HIGHLY REACTIVE SORBENT
Sorbent Preparation
As a next step of our development, we have researched about improvement of SO2
removal for sorbent. We found that it was possible by additive and heating.
Preparation of highly reactive sorbent is as follows.
Sorbent were prepared by slurrying Ca(0H)2 with fly ash and heating them in an
autoclave. The heating temperature was varied between 60 - 130 °C. Under the
standard condition, 60 g of Ca(0H)2 was slurried with 40 g of fly ash in 400 ml of
distilled water. After being heated for a given period, the slurry was filtered
and dried at 110 °C in an oven. The dry sorbents were later used for tests.
When sodium silicate (water glass, which contains approximately 18 % of Na20, 37 %
Si02 and the balance of water) was added to enhance the reactivity of the sorbents,
it was first dissolved in distilled water and mixed with Ca(0H)2 and coal ash.
Unless otherwise stated, the procedure for preparation was the same as mentioned
above. In this paper, "the amount of sodium silicate added to Ca(0H)2" was defined
as the weight percent of sodium silicate solution including water in it.
7-130
-------
Fixed Bed Reactor
The reactivity of sorbents with different surface areas was determined in, a fixed
bed reactor. The flow sheet is shown in Figure 10. The sorbents were dispersed in
alumina sand to avoid channeling and supported on the glass filter in the reactor.
The cylindrical pyrex reactor (5 cm in diameter and 15 cm in height) was placed in
an oven at temperature controlled between 60 - 90 °C. Simulated flue gas was
prepared by mixing N2, CO2, O2, SO2 and NO gases from gas cylindeT-.. The gas flow
rate was 3.0 liters per minute and consisted of 12 % C02,3 % O2 ppm SO2, 300
ppm NO and the balance N2. Steam was added to the gas by evaporating distilled
water in an evaporation chamber. Downstream of the reactor, the gas was cooled by
ice and the SC7 concentration in it was measured by an infrared SO2 analyzer to
determine the reactivity of the sorbents.
The amount of sorbent placed in the reactor was adjusted so that the Ca(0H)2 load-
ing was always 30 milli-grams. The reactivity of the sorbents is express'.! as "Ca
conversion" which is defined as the molar percent of Ca reacted with SO2 wr:tri the
total mole of SO2 supplied in the reactor is equal to that of Ca in the sorbent.
Evaluation of Highly Reactive Sorbents
The technique to prepare highly-reactive sorbent from Ca(0H)2 and fly ash is well
known (5, 6). In the present study, the effects of slurry heating conditions and
additives on the reactivity of sorbents were evaluated.
Before measuring the reactivity of sorbents towards SO2, the effects of some
factors on the specific surface areas of sorbents were investigated. The results
are shown in Figures 11 to 13.
Figures 11 and 12 show specific surface area of sorbents as a function of heating
time at a given temperature and as a function of heating temperature at a given
time respectively. The specific surface area increased according to the slurry
heating temperature or heating time.
In order to increase specific surface area of sorbents without increasing slurry
heating temperature or time, sodium silicate was added to Ca(0H)2 and fly ash.
Figure 13 indicates the effect of the amount of sodium silicate on the specific
surface area of the sorbents. The axis of abscissas in Figure 13 represents the
weight percentage of sodium silicate to Ca(0H)2- The specific surface area
increased by adding sodium silicate.
The reactivity of sorbents were studied in a fixed bed reactor. Figure 14 sum-
marizes the relationship between Ca conversion and the specific surface area of
sorbents which were prepared at various conditions (slurry heating temperature,
time and the amount of sodium silicate added were varied). In these experiments,
the temperature and relative humidity in the reactor were kept at 70 °C and 50 %
respectively. The data show that the reactivity of sorbents at a given condition
is proportaional to their surface area.
The above results show that the reactivity of sorbents can be multiplied by
addition of fly ash and sodium silicate. In the ...xt step, reactive sorbent with
specific surface area of 35.5 m^/g was injected into the reactor of the pilot plant
7-131
-------
and its SO2 removal was compared with that of Ca(0H)2 (specific surface area : 13
m2/g). This reactive sorbent was prepared by slurrying 57 % Ca(0H)2, 38 % fly ash
and 5 % sodium silicate at 125 °C for 3 hours. Figure 15 shows SO2 removal of the
reactive sorbent and Ca(0H)2 as a function of outlet gas temperature. At the same
Ca/S molar ratio, SO2 removal of the reactive sorbent was much higher than that of
Ca(0H)2- Even at lower Ca/S molar ratio (Ca/S = 1.4), the reactive sorbent
performed equal SO2 removal to that of Ca(0H)2 at Ca/S molar ratio of 2, resulting
in about 30 % decreased sorbent consumption at the same SO2 removal.
CONCLUSION
The following conclusions can be drawn from the results of the present study :
1. In both DLI and FLI processes, SOj removal was achieved above 70 %
and 80 % at Ca/S molar ratios of 2 and 3 respectively.
2. In the range of 500 - 3000 ppm, the SO2 removal is independent of
SO2 concentration in both processes.
3. SO2 removal increased in accordance with the lowering of the gas
temperature at reactor outlet. However, it should be kept at
least 10 °C over the saturated temperature to avoid deposition of
the reacted sorbent in the downstream duct or ESP.
4. In view of SO2 removal and thermal decomposition of limestone, op-
timum injection temperature in the furnace should be 1000-1100 °C.
5. In the reactor, downstream duct and ESP, scaling was not caused by
humidification.
6. The reactivity of sorbents at a given condition is proportional to
their surface area.
7. The specific surface area of the sorbents increased by adding
sodium silicate as well as by heating slurry longer or at higher
temperature.
REFERENCES
1. H. Kuroda et al. "AN ADVANCED FGD SYSTEM FOR LARGE COAL FIRED BOILER" 1990 S02
Control Symposium, New Orleans, 1990.
2. S. Kudo et al. "DRY FLUE GAS DESULFURIZATION RPOCESS USING ABSORBENT MADE FROM
FLYASH". The First Combined FGD and Dry SO2 Control Symposium, St. Louis, 1988
3. P.S. Nolan et al. "Results of the EPA LIME Demonstration at Edgewater" 1990 SO2
Control Symposium, New Orleans, 1990.
4. H. Kikkawa et al. Proceeding of the 24th Autumn Meeting of Japan Chemical Engi-
neering Society, 1991.
5. J.R. Peterson, "Lime/Fly Ash mateiral for Flue Gas Desulfurization : Effects of
Alminum and Recycle Materials" 1990 SO2 Control Symposium, New Orleans, 1990.
6. J.C.S. Chang et al. "Reactivation of Edgewater LIMB Solids by the Advocate
Process for In-Duct SO2 removal" 1990 SO2 Control Symposium, New Orleans, 1990
7-132
-------
DLI - TYPE
BOILER
<1
V
REACTOR
FLUE GAS
APH
| LIME |
¦AIR
WATER
ASH + WASTE MATERIAL
FLI - TYPE
REACTOR
FLUE GAS
ESP
BOILER
WATER
STACK
AIR
APH
LIME-
STONE
ASH + WASTE MATERIAL
Figure 1. Flow Diagram of the Hitachi Simplified FGD System
7-133
-------
SPECIFICATION OF THE PILOT PLANT:
GAS FLOWRATE:
REACTOR
Avft Abtorp^nt
Diameter of Reactor:
fi650 x 15.C00 mm H
Air
50;
Figure 2. Flow Sheet of 2,000 m'/h Pilot Plant
Figure 3. Photograph of Pilot Pl2nt
7-134
-------
Figure 4.
3notcgrapn oi
Test Furnace
Primary Gas Cccier
Secondary
Gas Cooler
J\ 1_
LClj^|l050°C
1 so2
Detectinc—1
r^° cl:
Point
1150 °C
Swich Valves
Air Compressor
A t/h Coal Burner
Figure 5. Outline of Test Furnace
7-135
-------
Test Furnace
PILOT PLANT
CaCC>3 Injection
Coal
Bag Filter
Sorbent_L
Silo
Ash + CaO Oi
1
_ £_1 t
Furnace SO2
Bag Filter
Reactor
A
I
Air
-Water
Figure 6. Test Flow Sheet for FLI Process
100
_ 80
>
o
E
£ 40
CM
o
1/1 20
0
Figure 7. Effect of Reactor Outlet Gas Temperature
and Relative Humidity on SO= Removal
Ca/S
SO2 = 1800ppm
60 70 80
Gas Outlet Temperature (°C)
60 50 40 30
Gas Outlet Relative Humidity (%)
7-136
-------
100
_ 80
£
1 60
o
E
(LI
cr: 40
CM
O
uo
20
0
0 1000 2000 3000
Inlet SO2 ( ppm )
Figure 8. Effect of Reactor Inlet S02 Concentration
on S02 Removal
3 —A A A A
2 -o O O O
1 —~ —~ ~ ~
Ca/S
Reactor Outlet
Gas Temp. = 65°C
1 1 1 1 1 i_
100
80
60
40
20
0
0
Over all
(Furnace + Reactor)
Reactor Outlet
Gas Temp = 65°C
Furnace
.. -cr
-
Temp, at Furnace
Injection Zone=1050°C
Figure 9.
12 3 4
Ca/S Molar Ratio (-)
Overall S02 Removal in FLI Process and
S03 Removal in Furnace
7-137
-------
h2o
N2, S02, NOx
Sorbent
Gas Mixer
Evaporator Filter
Oven
to SO2 Analyzer
Trap
Figure 10. Flow Sheet of Fixed Bed Reactor
E
40
4—
Temp: 100 °C
Lime : Ash : 6:4
Heating Time (h)
Figure 11. Specific Surface Area of Sorbent
as a Function of Heating Time
7-138
-------
pi
Heating Time: 3h
Lime : Ash : 6:4
CM
40
< 30
20
60
80
100 120 140
40
Heating Temperature (°C)
Figure 12. Specific Surface Area of Sorbent
as a Function of Heating Temperature
60
-S1 50
E
40
o
0>
L_
<
30
20
0>
Heating Condition
: 100°Cx3h
0
10
20
30
40
50
Sodium Silicate (wt%)
Figure 13. Relationship between Amount of
Sodium Silicate and Specific
Surface Area
7-139
-------
70
60
50
c
o
j2 40
CD
>
C
o
u 30
E
-------
Adiabatic Saturation Tern p. = 55°C
(Ca/s = 1.4)
Case B
(Ca/s = 2.0)
Case A : Ca(OH)2 +Sodium Silicate
+ Fly Ash
Case B : Only Ca(OH)2
60 65 70 75
Temperature (°C)
80
85
Figure 15. Relationship between Reactor Outlet Gas
Temperature and S0a Removal
7-141
-------
Intentionally Blank Page
7-142
-------
CLEAN COAL TECHNOLOGY
OPTIMIZATION MODEL
B. A. Laseke and S. B. Hance
International Technology Corporation
Air Quality Services
11499 Chester Road
Cincinnati, Ohio 45246
7-143
-------
Intentionally Blank Page
7-144
-------
ABSTRACT
Title IV of the Clean Air Act Amendments (CAAA) of 1990 contains provisions for the miti-
gation of acid rain precipitation through reductions in the annual emission of the acid rain
precursors of sulfur dioxide (SOj) and nitrogen oxide (NOJ. These provisions will affect
primarily existing coal-fired power-generating plants by requiring nominal reductions of 5 mil-
lion and 10 million tons of SOz by the years 1995 and 2000, respectively, and 2 million tons
of NOx by the year 2000 relative to the 1980 and 1985-87 reference period.
The 1990 CAAA Title IV provisions are extremely complex in that they establish phased regu-
latory milestones, unit-level emission allowances and caps, a mechanism for inter-utility trad-
ing of emission allowances, and a system of emission allowance credits based on selection
of control option and timing of its implementation. The net result of Title IV of the 1990 CAAA
is that approximately 147 gigawatts (GW) of generating capacity is eligible to retrofit S02 con-
trols by the year 2000.
A number of options are available to bring affected boilers into compliance with Title IV. Mar-
ket share will be influenced by technology performance and costs. These characteristics can
be modeled through a bottom-up technology cost and performance optimization exercise to
show their impact on the technology's potential market share. Such a model exists in the
form of an integrated data base-model software system. This microcomputer (PC)-based
software system consists of a unit (boiler)-level data base (ACIDBASE), a cost and perform-
ance engineering model (IAPCS), and a market forecast mocjel (ICEMAN).
7-145
-------
INTRODUCTION
Title IV of the Clean Air Act Amendments (CAAA) of 1990 contains provisions for the miti-
gation of acid rain precipitation through reductions in the annual emission of the acid rain
precursors of sulfur dioxide (SO^ and nitrogen oxide (NOJ. These provisions will affect
primarily existing coal-fired power-generating plants by requiring nominal reductions of 5 mil-
lion and 10 million tons of S02 by the years 1995 and 2000, respectively, and 2 million tons
of NOx by the year 2000. These reductions are relative to the 1980 and 1985-87 baseline
regulatory period. The resulting emission cap will be held constant, thus requiring additional
removals in response to load growth.
The 1990 CAAA Title IV provisions are extremely complex in that they establish phased-in
regulatory milestones, unit-level emission allowances and caps, a mechanism for inter-utility
trading of emission allowances, and a system of emission allowance credits based on selec-
tion of control option and timing of its implementation. The net result of Title IV of the 1990
CAAA is that approximately 147 gigawatts (GW) of generating capacity is eligible to retrofit
SOz controls by the year 2000.
A number of options are available to bring affected boilers into compliance with Title IV. Mar-
ket share will be influenced by technology performance and costs. These characteristics can
be modeled through a bottom-up technology cost and performance optimization exercise to
show their impact on the technology's potential market share. Such a model exists in the
form of an integrated data base-model software system designated as the Control Technolo-
gy Optimization Model. This integrated data base-model set is a microcomputer (PC)-based
software system developed specifically for the application of flue gas cleanup (FGC) and
clean retrofit and repowering technologies to coal-fired utility boilers.
7-146
-------
DESCRIPTION
The integrated PC-based software system (Control Technology Optimization Model) consists
of a unit (boiler)-level data base (ACIDBASE), a cost and performance engineering model
(1APCS), and a market forecast model (ICEMAN). Each component in the software system is
described in the following subsections in order of progression in the software system.
ACIDBASE
ACIDBASE is a boiler-level data system that contains information for domestic fossil-fuel-
generating units for the record years of 1985 to 1987. ACIDBASE provides a separate data
file for each individual record year of 1985, 1986, and 1987. These separate record year files
are united through internal programming to form a 1985 to 1987 composite derivative data
base known as 'BASELINE.* The BASELINE data base was created to provide information
for the 1985 to 1987 reference ("baseline") regulatory period specified in the 1990 CAAA.
The BASELINE data system is the 'run version" of ACIDBASE that is mated to the rest of the
components in the integrated software system.
ACIDBASE originated largely from a number of public domain data bases that exist primarily
(but not exclusively) in computerized form. Specifically, information was extracted from De-
partment of Energy (DOE) Environmental Impact Assessment (EIA) Form 767 (Annual Steam
Electric Utility Design and Operation Report), the DOE Federal Energy Regulatory Commis-
sion (FERC) Form 1 (Annual Report of Public Electric Utilities, Licencees, and Others), DOE
EIA Form 412 (Annual Report of Public Electric Utilities), DOE FERC Form 423 (Cost and
Quality of Fuels), DOE EIA Form 860 (Annual Electric Generator Report), and DOE Rue Gas
Desulfurization Information System (FGDIS). Supplemental information was also obtained
from a number of noncomputerized sources: McGraw-Hill's Electric Utility World and the
National Coal Association's Steam Electric Plant Factors.
7-147
-------
Using the computerized public domain data bases as the primary sources of information,
computer tapes of the 1985 through 1987 annual filings for the various data systems were
obtained, downloaded, ported into the PC environment, quality checked for data complete-
ness and accuracy, and stored in a PC-based data base management system.
A schematic showing the origination and organization of ACIDBASE is provided "m Figure 1.
A listing of the data elements contained in ACIDBASE is shown in Figure 2.
INTEGRATED AIR POLLUTION CONTROL SYSTEM
The Integrated Air Pollution Control System (IAPCS) cost engineering and performance
model was developed to estimate the costs and predict the performance of SOj, NOx, and
particulate matter (PM) emission control systems for coal-fired utility boilers. This computer-
ized model includes conventional and emerging technologies that effect pre-, in situ, and
post-combustion emission control. The model can accept any combination of the technology
"modules" built into the system. Interactions are reflected in a material balance tabulation at
the exit of each module. Alterations in material balance are used to account for integrated
performance and cost effects. The emission control technologies contained in IAPCS can be
selected in either "isolated" (single technology) or "integrated" (multiple technology) config-
urations. The power of IAPCS lies in its ability to reflect integrated effects of various control
configurations. This allows the analyst to identify synergistic interactions and thus optimize
performance and cost in terms of integrated cost-effectiveness.
The IAPCS performance and cost-estimating computer model includes both conventional and
emerging flue gas cleanup and clean coal technologies (Table 1). The model accepts any
combination of these technologies. Output from the model reports reductions in SOz, NOx,
and PM emissions and associated capital, annual, and cost-effectiveness values (dollars per
ton of pollutant removed across the entire emission control system).
A unique and important feature of the model is the parameter files. As each module was
developed, the important design, cost, and performance parameters (i.e., cost and perform-
ance drivers) were included in a parameter file. These parameters are assigned default val-
ues that may be changed by the user for a given application. The parameter file is designed
to permit the user to modify the important values to reflect those of choice or need.
7-148
-------
Supplemental
Data Sources
Primary
Data Sources
•c
Form 767
ACIDBASE
Form 1
Form 412
1985
Record
Year
1986
Record
Year
1987
Record
Year
Integration
\QA/QCy
Form 423
Form 860
FGDIS
Combined Year
Generator
Program
Interface with IAPCS
Cost Engineering and ¦<¦
Performance Model
BASELINE
(1985- 1987
Composite Data Base)
Figure 1. Origination and organization of Acldbase.
-------
Oris code
Utility
Plant
State
Unit No.
Boiler No.
Service date
Status
Geographic regions
Bolter
Nameplate rating
Net dependable capability
Summer capability
Winter capability
Boiler supplier
Draft type
Firing type
Firing configuration
Burner type
Bottom type
rumace type
Furnace depth
Furnace width
Furnace height
Boiler roof-nose
Boiler bottom-nose
Furnace plan area
Boiler special conditions
Multiple fuel
Gas availability
Coal rate
Coal consumption
Heat rate
Capacity factor
Load profile
Heat input
Steam rate
Steam temperature
Steam pressure
Steam cycle
Coal rank
Heat content
Sulfur, content
Pyritic sulfur
Organic sulfur
Sulfate sulfur
Ash
Moisture
Volatile matter
Fixed carbon
Chlorine
Ash alkalinity
Pyritic sulfur
Ferrous oxide percent
Sodium oxide
Emission Controls
Gas flow rate
Gas temperature
SO2 standard
NOx standard
PM standard
PM collector service date
PM collection application
PM type
PM design
PM removal
ESP collection area
Fabric filter AC ratio
SO2 control service date
SO2 control application
SO2 control type
SO2 control design
SO2 emission rate
NOx control application
NOx control type
NOx control design
NOx emission rate
Acid emissions
Acid emission rate
Figure 2. Acidbase data elements.
7-150
-------
Table 1
lAPCS CONTROL TECHNOLOGY MODULES
Category Technology
Precombustion Coal supply options (CSO)
In situ Low-NO, combustion (LNC)
Limestone injection multistage burner (LIMB)
Rebuming (NGR)
Natural gas substitution (NGS)
Posicombustion Selective catalytic reduction (SCR)
NOXSO
Gas conditioning (GC)
Dry sorbent injection (DSI)
Lime spray drying (LSD)
Bectrostatic precipitator (ESP)
Fabric filter (FF)
Rue gas desuifurization (FGD)
Plant Life extension (LE)
Demolition and decommissioning (D/D)
Repowering
Process
Coal substitution/blending (CS/B)
Physical coal cleaning (PCC)
Low excess air
Overfire air (OFA)
Low-NO,, burner (LNB)
Low-NO, concentric firing (LNCFS)
Furnace sorbent injection
Natural gas
Natural gas
Hot-side
Coal
Natural gas
S03
Spray humid ification
NahcoUte
Trona
Spray dryer absorber (SDA)
Duct spray drying (DSD)
Cold-side
Reverse-air
Lime/limestone
Reboiering
Refurbishment
Atmospheric FBC (AFBC)
Pressurized FBC (PFBC)
Integrated gasification combined cycle
(IGCC)
7-151
-------
A brief synopsis of lAPCS's capabilities, describing input features, processing, and output
reports is summarized below:
Inputs are provided in either interactive (single-record run) or batch (multiple-
record run) modes. The batch mode accesses boiler records through the BASE-
LINE data base.
The model incorporates the basis and format of cost estimation used by the
Electric Power Research Institute (EPRI).* The capital cost is built "bottom-up*
from calculated process conditions based on user inputs, estimating direct capi-
tal costs by process area, which include both direct and indirect costs. The
annual cost is also built "bottom-up," estimating fixed and variable operating and
maintenance costs and annual carrying charges.
Several of the files and routines are "system-wide" (i.e., not limited to one particu-
lar control technology module). These system-wide features are an important
aspect of the model's integrated operation. They include emission calculations,
boiler performance, fans, waste disposal, and economic/financial files.
The model provides the user with the following reports: user input summary,
module-specific output, boiler performance, material balance, emission reduction,
and costs.
At the initiation of a run, the user can 'optimize' a selected control system. A
target emission rate On lb/106 Btu) may be entered, and the system performance
and costs will be run automatically. This optimization routine allows the user to
alter the effective efficiency of a chosen control device through bypassing a frac-
tion of the gas stream or by changing the capture efficiency by altering its design
and operating parameters.
The model is designed on a modular basis. This permits a control module in the
run stream to pass on its altered conditions to any succeeding module(s) in the
downstream flue gas path. These data are then used to generate the design
and performance characteristics and cost estimates. The architecture of a mod-
ular program is such that it offers the user the greatest flexibility for revising any
existing control technology and for adding technologies.
ICEMAN
The IAPCS Cost Effectiveness Maximization Analysis (ICEMAN) model is a market forecast
model that selects least-cost technology solutions over time per conditions imposed by the
user. A least-cost solution is based on cost-effectiveness of dollars per ton ($/ton) of pollu-
tant (SOj/NOJ removed. Conditions imposed by the user include technology slate, target
* TAG-TechnicaJ Assessment Guide, Volume 1: Electricity Supply-1989. EPRI P-6587-L,
Volume 1: Rev. 6, Special Report, September 1989.
7-152
-------
global (i.e., tons per year) and unit-level (i.e., lb per 10s Btu) emission reduction rates, and
time horizon (i.e., year or years of analysis extending from 1995 to 2030).
ICEMAN is a postprocessing computer program that is slaved to IAPCS for its input data.
Although ICEMAN exists as a separate and independent software system, it cannot function
without IAPCS located upstream to provide input data requirements. Moreover, since ICE-
MAN processes a vast array of records (multiple boiler-multiple technology combinations),
IAPCS must be run in a "batch" mode, which requires that the BASELINE data base be situ-
ated upstream of IAPCS, providing input data requirements.
ICEMAN accepts output from IAPCS for the boiler population-technology slate case under
consideration. The output from IAPCS contains information needed by ICEMAN for process-
ing (cost-capital, annual, and cost-effectiveness and performarice-S02 and NOx removed).
ICEMAN processes the file to satisfy the global emission target selected by the user. ICE-
MAN'S first pass through the data set selects the most cost-effective technologies across the
entire boiler population set and sums the corresponding S02 (or NOJ tonnage. The selec-
tion process proceeds from the most cost-effective technology-boiler value in the entire data
set to the least cost-effective value (i.e., lowest numerical $/ton value to highest numerical
$/ton value). The summed S02 (or NOJ tonnage is dynamically compared with the pre-
selected global emission reduction target. When convergence is achieved (i.e., the summed
tonnage satisfies the preselected target), the process stops and results are reported. If con-
vergence is not achieved on the first pass, ICEMAN makes another pass through the data
set until convergence is achieved. In simplistic terms, the most cost-effective value not se-
lected in the first pass (in the entire data set) is selected and "tested." The test determines if
the second pass selection provides more SOa (or NOJ removal than the first pass selection.
If yes, the second pass value is retained and the first value is discarded and the new sum
value is compared with the preselected target value for convergence. If no, the first pass
value is retained and the process continues with the next most cost-effective value in the
entire data set This process continues in systematic fashion until convergence is achieved.
The process then stops and results are reported.
ICEMAN assigns one technology "solution* to each boiler that is selected in the convergence
process. The final result is that each selected boiler is assigned one technology.
7-153
-------
ICEMAN also features the ability to forecast over time. This option allows the user to select a
time horizon over which emissions are kept constant as a function of a user-provided emis-
sion cap. In effect, ICEMAN "holds" selections and makes new selections as the time hori-
zon is changed. For example, this feature allows ICEMAN to prepare market forecasts under
conditions of a fixed S02 emission cap beyond the year 2000 Phase II regulatory milestone.
7-154
-------
SNKB CATALYTIC BAGHOUSE PROCESS DEVELOPMENT AMD DEMONSTRATION
K. E. Redinger, P. Chu, and G. A. Farthing
The Babcock & Wilcox Company — Research and Development Division
Alliance, Ohio 44601
J. M. Wilkinson
The Babcock & Wilcox Company — Environmental Equipment Division
Barberton, Ohio 44203
R. W. Corbett
U.S. Department of Energy — Pittsburgh Energy Technology Center
Pittsburgh, Pennsylvania 15236
H. Johnson
Ohio Coal Development Of£ice
Columbus, Ohio 43266
•7-155
-------
Intentionally Blank Page
7-156
-------
ABSTRACT
The SOx-NOx-Rox Box™ (SNRB) process is a Babcock & Wilcox (B&H) patented
advanced air pollution control system that provides for significantly
reduced SOx, NOx, and particulate emissions from coal-fired boilers. The
process utilizes a high-temperature catalytic baghouse for the integration
of SOx reduction through the injection of an alkali sorbent such as hy-
drated lime or sodium bicarbonate, N0X removal through ammonia injection
and selective catalytic reduction, and particulate collection. The advan-
tages of the process include: compact integration of the emission control
technologies into a single component, dry sorbent and by-product handling,
and improved SCR catalyst life due to lowered SOx and particulate levels.
The basic SNRB concept has been developed over a period of several years at
B&W through a series of pilot-scale research programs initiated in 1979.
This successful pilot evaluation of emission control performance of the
concept led to selection of the technology for demonstration in the second
round of the U.S. Department of Energy (DOE) Innovative Clean Coal Technol-
ogy Demonstration Program. The current project focuses on the design,
installation, and operation of a 5-MWe equivalent SNRB facility at the Ohio
Edison R. E. Burger Plant. Installation of the demonstration facility has
been completed and startup activity is underway. The demonstration will be
used to generate information for commercialization of the SNRB technology.
7-157
-------
INTRODUCTION
B&W's patented SOx-NOx-Rox-Box process — also known as SNRB — has been
developed to provide boiler operators with a cost-effective way to simulta-
neously control their emissions of the oxides of sulfur (S0X) and nitrogen
(N0X), and particulate matter (Rox). Briefly, the process (shown in Figure
1) comprises the injection of both ammonia and dry sorbent upstream of a
fabric filter (baghouse). A catalyst for the selective catalytic reduction
(SCR) of N0X is installed inside the filter bags, providing for the de-
struction of N0X as the flue gas/ammonia mixture passes over the catalyst.
S0X is absorbed by the sorbent both in the flue gas duct, and as the sor-
bent resides on the filter bags in the baghouse. Since the S0X and N0X
removal processes require operation at elevated gas temperatures (550* -
900"F), specially woven high-temperature fabric filter bags are used.
Through the integration of the SOx, NOx, and particulate removal processes
into a single unit, lower capital cost and space requirements are achieved,
and operating procedures are simplified, when compared to a combination of
conventional emissions control system comprising separate wet scrubber,
SCR, and particulate removal systems.
B&W is currently conducting a multi-phase project funded by the U.S.
Department of Energy (DOE) - PETC, the Ohio Coal Development Office (OCDO),
and the Electric Power Research Institute (EPRI) under the DOE's Innovative
Clean Coal Technology program. The objective of the B&W project is to
continue the commercial development of the SNRB concept in a 5-MWe field
demonstration unit at Ohio Edison's R. E. Burger plant located near
Shadyside, Ohio. Other members of the project team include Ohio Edison
(host utility), Norton Company (catalyst supplier), and Minnesota Mining
and Manufacturing - 3M (bag supplier). The minimum emission control
targets for the project include the cost-effective reduction of SOx, NOx,
and particulate emissions in the following manner:
• 70% SOx removal
• 90% NOx removal
• Particulate emissions in compliance with the New Source
Performance Standards (NSPS) — 0.03 lb/million Btu
Phase I — Design and Permitting — of the program involved the development
7-158
-------
of a detailed engineering design for the 5-MWe SNRB facility, and the
preparation of the required environmental and permitting documents. The
design activities were supported by the operation of a 1500 ft3/minute SNRB
laboratory pilot unit located at BtW's research center in Alliance, Ohio.
Installation of the 5-MWe facility was completed under Phase II — Procure-
ment, Installation, and Startup. Following startup and shakedown activi-
ties in late 1991, full operation of the facility will begin in early 1992.
A 7-month testing program will follow, wherein detailed operating and
performance data will be generated over a wide range of operating condi-
tions. The test work will be conducted as part of Phase III — Demonstra-
tion Operation and Restoration — of the program. Phase III will include
a detailed engineering and economic analysis of the integration of the
technology into an existing boiler plant. This paper is intended to
provide a summary of the status of the development of the SNRB technology.
SNRB PROCESS DESCRIPTION
The SNRB process comprises a single process unit — a pulse-jet fabric
filter (baghouse) — located upstream of the boiler's combustion air
preheater and operating at a temperature 550* - 900*F. One possible
arrangement of the SNRB process applied to a utility boiler system is
illustrated in Figure 2. A specially woven ceramic fabric is used for the
fabrication of the filter bags to permit reliable baghouse operation at
these elevated temperatures. Absorption of SOx is accomplished through the
injection of a finely divided dry alkali sorbent into the flue gas stream
upstream of the baghouse. The SOx in the flue gas reacts with the sorbent
while the latter is dispersed in the flue gas stream as it flows through
the ductwork and baghouse, and while it resides on the filter bags in the
form of filter cake. A catalyst for the selective catalytic reduction
(SCR) of NOx is installed inside the filter bags — the clean side of the
bags in a pulse-jet baghouse. By injecting ammonia into the flue gas
upstream of the baghouse, NOx is converted to harmless N2 and H20 as the
gases pass over the catalyst. Particulate matter — fly ash ;.:;d spent
sorbent — is removed as the flue gas passes through the filter bags.
Calcium-based sorbents such as calcium hydroxide (hydrated lime) will
generally provide the most cost-effective approach for SNRB applications in
the eastern Dnited States. These sorbents require baghouse operating
temperatures near 850*F for optimal SNRB SCX removal performance. Typi-
cally, flue gas temperatures at the outlet of the economizer are lower than
7-159
-------
850"F. Therefore, it may be necessary to remove some economizer heat
transfer surface from service to obtain the desired baghouse operating
temperature. Detailed analysis of the integration of SNRB with the exist-
ing boiler design will be required to optimize overall steam cycle effi-
ciency .
Sodium-based sorbents such as sodium bicarbonate (NaHC03) may the preferred
approach for the application of SNRB in the western U.S. where natural
deposits of these materials occur. Baghouse operating temperatures in the
range of 500* - 800*F will be used for these systems. If maximum N0X
removal performance (90+%) is needed, the lower operating temperature of
the SNRB baghouse for sodium applications may result in the need to use a
promoted SCR catalyst for N0X removal in these systems. The higher sorbent
costs may be offset by the use of less-expensive bag fabrics at the lower
temperatures.
The S0x-N0x-Rox Box process has several potential benefits:
• On* Major Component. Capital cost and space requirements
are reduced by performing the S0X, N0X, and particulate
removal operations in a single piece of equipment.
• X.onger Catalyst Life. SOx and particulates are removed from
the flue gas stream upstream of the SCR catalyst, minimizing
catalyst poisoning, pluggage, and erosion concerns.
Dzy Materials Handling. Reagent preparation, handling, and
disposal costs are minimized because both the fresh sorbent
and waste streams are dry.
• "Unpronoted Catalyst." The SCR catalyst used for calcium-
based SNRB systems can be an unpromoted zeolite material.
This avoids the potential hazardous waste disposal concerns
associated with promoted catalysts containing metals such as
vanadium.
• Ko rlue Gas Reheat Required. N0X removal upstream of the
boiler air heater eliminates the need for flue gas reheat
for optimal N0X removal performance, as is required in many
conventional SCR systems.
• Increased Boiler Thermal Efficiency. The SNRB process is
one of the few S0X removal processes offering the potential
for a decrease in plant net heat rate due to the removal of
S03 upstream of the air heater — virtually eliminating acid
dew point concerns in the combustion air preheater.
7-160
-------
PROCESS DEVELOPMENT WORK AT B£W
Development of the SNRB process began at BSW in 1979 when a series of
laboratory screening tests was performed to evaluate the applicability of a
variety of materials to the catalytic reduction of N0X. Materials such as
fly ash, transition metals, and a Norton Company zeolite catalyst were
evaluated. Development work then proceeded through a series of pilot-scale
test programs conducted in baghouses ranging in size from 350 ft3/minute to
3000 ft3/minute. It was at this point that 3M's Nextel™ ceramic fiber was
identified as a potential material for fabrication of the high temperature
bags. Nextel can be used on a continuous basis at temperatures up to
1400"F, with brief temperature excursions up to 2200"F. These in-house
development programs eventually led to two OCDO-sponsored testing programs
wherein the process concept was further refined, and preliminary perfor-
mance data was obtained CI.) -
The successful early pilot test results were used to pursue funding under
the DOE Innovative Clean Coal Technology Development Program. In 1989, a
proposal was accepted by DOE and OCDO for demonstration of the technology
on a larger, intermediate-scale pilot facility. Additional laboratory
pilot testing under this program has been completed in support of the
design of a 5-MWe demonstration facility (2,2)- Erection of the demonstra-
tion facility was completed in September 1990. A preliminary economic
analysis based on the laboratory pilot results has been completed to help
focus the testing activity in the demonstration facility. The facility
will be operated to assess the commercial readiness of the SNRB technology.
LABORATORY PILOT TEST RESULTS
In the early test programs, it was not possible to evaluate a full-size,
integrated bag/catalyst arrangement. The bags used in the 3000 ft3/minute
baghouse were 4 inches in diameter and 10 feet long, whereas a commercial
bag is more likely to be 6-1/4 inches in diameter and 20 feet long. The
catalyst was also located in the exhaust plenum of the baghouse, as opposed
to being integrated into each filter bag assembly. Since both of these
design features could have a significant impact on bag cleanability, the
inability to assess these factors was a serious limitation of the early
pilot test programs.
7-161
-------
Further, the effect on S0X removal performance of sorbent injection into or
upstream of the boiler economizer could not be fully evaluated. In a
typical commercial, calcium-based SNRB application, the sorbent will be
injected ahead of, or into, the boiler economizer. While passing through
the economizer, the flue gases, and hence the sorbent, experience a sharp
drop in temperature as heat is extracted. It was anticipated that the
time-temperature history of the sorbent could have a significant impact on
the SOx removal process, but the equipment used for the earlier tests did
not lend itself to a comprehensive evaluation of these effects. Therefore,
while these earlier tests provided encouraging results, it was concluded
that additional SNRB laboratory pilot tests were necessary to support the
design of the 5-MWc field demonstration facility.
The primary objective of the most recent series of SNRB laboratory pilot
tests was to develop design, operating, and performance specifications for
the 5-MWc field demonstration facility. The major issues addressed in-
cluded the:
• Performance and operability of commercial-size, integrated
bag/catalyst arrangements
• Effect of economizer injection on SOx removal performance
• S0X and N0X removal performance
• Compatibility of the SOx and N0X removal processes
• Particulate and NH3 emissions downstream of the baghouse.
The results of these tests also provided information needed to support a
preliminary economic evaluation of the SNRB process.
A schematic of the SNRB laboratory pilot-scale test facility is illustrated
in Figure 3. The facility is comprised of a pulverized coal-fired test
furnace (5-million Btu/hr thermal input), insulated ductwork, sorbent and
ammonia injection systems, a heat exchanger to simulate a utility boiler
economizer, a high-temperature SNRB baghouse, and an ID fan. The baghouse
contained 12 commercial-size fabric filter bags — 6-1/4 inches in diameter
and 20 feet long — yielding an overall collection area of approximately
375 ft2. The 1500 ft3 (actual) /minute baghouse vas equipped with an on-
line pulse-jet cleaning system. The cleaning cycle was initiated by a set-
point limiting the differential pressure across the baghouse.
7-162
-------
High-temperature, woven, 3M Nextel™ bags were used along with an
unpromoted zeolite SCR catalyst from the Norton Company. The catalyst was
incorporated into each bag filter assembly.
The heat exchanger used to simulate a utility boiler economizer was a
water-cooled, natural convection unit equipped with three independent banks
of water-cooled tubes. By varying the number of banks in service, various
time-temperature profiles and baghouse inlet temperatures were investi-
gated. The sorbent injection system consisted of a solids feeder, air
eductor, and injection nozzle. Sorbent injection was performed at various
locations to investigate the effect of residence time at the injection
temperature. Two sorbent injection points were located upstream of the
simulated economizer. The third was downstream of the economizer, but
upstream of the baghouse. Ammonia vapor from a cylinder was diluted with
air before being injected into the flue gas upstream of the baghouse.
An Ohio *8 coal containing 2.5 - 4.0% sulfur was fired to produce a flue
gas containing 1500 - 2500 ppm S02, 600 - 800 ppm NOx, and 3 - 4% 02- The
coal was selected to be representative of the coal burned at Ohio Edison's
R. E. Burger Plant, thereby simulating expected flue gas conditions in the
5-MWe facility. The desired flue gas operating temperature was obtained by
mixing flue gas extracted from two different locations in the water-cooled
convection pass of the test furnace. S02, NOx, and 02 concentrations in
the flue gas were continuously monitored at the furnace outlet (before
sorbent and NH3 injection), baghouse inlet, and baghouse outlet. Tests
were typically conducted over a 3 to 4-hour period to ensure steady-state
test conditions.
Testing was conducted over a range of operating conditions. Major operat-
ing parameters varied during the test program included sorbent injection
and baghouse operating temperatures, Ca/SOx and NH3/N0X stoichiometrics,
and sorbent residence times. The laboratory pilot results are briefly
summarized in the following three sections.
Particulate Emissions
Woven, seamless bags fabricated of 3M Nextel™ ceramic fibers were used for
particulate collection. One of the key objectives of the laboratory pilot
tests was to demonstrate, on a continuous basis, particulate emissions in
7-163
-------
compliance with the NSPS requirement of 0-03 lb/million Btu. Of particular
concern was the cleanability of the fully-integrated bag/catalyst assembly
Figure 4 illustrates the particulate removal results achieved throughout
the SNRB laboratory pilot test program. Particulate emissions were deter-
mined using both the EPA Method 5 and 17 sampling techniques. As shown in
the figure, particulate emissions less than 0.03 lb/million Btu were
achieved for the majority of tests. Of equal importance was the fact that
the bags could be repeatedly and reliably cleaned using conventional pulse-
jet baghouse cleaning technology. The data collected after mid-December
were obtained after the fabric filter bags had been removed from the
baghouse. It is believed that some of the bags were damaged during their
re-installation, leading to the higher particulate emission measured during
the subsequent tests. The data points shown in February were obtained
during tests with a low-sulfur coal, where inlet dust loadings were much
lower than for the Ohio #8 coal tests.
NOz Rawuval
The primary NOx-related objective of the laboratory pilot tests was to
determine the design and operating specifications required to achieve at
least 90% KOx removal. Of particular concern were the design of the cata-
lyst and bag/catalyst supports, the required catalyst temperature and
ammonia stoichiometry, and the amount of ammonia "slip". The NH3/NOx
stoichiometry is defined as the molar ratio of injected ammonia to the NOx
in the flue gas, and ammonia slip refers to the amount of unreacted ammonia
passing through the SNRB system and exiting the stack. To achieve this
objective, the NH3/NOx ratio and catalyst temperature were varied over a
range of operating conditions to maximize NOx reduction and minimize NH3
slip. The catalyst design — shape, formulation, and space velocity (flue
gas flow rate divided by catalyst volume) — was also investigated during
the tests.
The effect of NH3/NOx ratio on NOx removal is illustrated in Figure 5. At
NH3/NOx stoichiometrics less than 1, the NOx removal increases from 70% to
90% as KH3/NOx stoichiometry increases from 0.8 to 1.0. At NH3/NOx stoi-
chiometrics greater than 1, NOx removals level off. In addition, operation
at these higher stoichiometry levels also results in higher emissions of
unreacted NH3. These results are also shown in Figure 5. This unreacted
ammonia is primarily of concern due to the potential formation of trouble-
7-164
-------
some ammonium bisulfate deposits as the flue gases pass through the down-
stream heat exchanger (air heater). At ammonia stoichiometrics greater
than 1.1, a sharp increase in NH3 slip was observed. At stoichiometries
lower than 1.1, NH3 slip was typically less than 20 ppm. The catalyst
design and arrangement have been modified for the field demonstration to
reduce NH3 slip.
Several tests were conducted to determine the effect of baghouse operating
temperature on NOx removal performance. The baghouse temperature was
varied from 675" - 850*F to determine baghouse operating conditions for
optimal NOx removal performance. As indicated in Figure 6, NOx removal
improved with increasing baghouse temperature up to about 850*F. Operation
at temperatures significantly above 850*F is not desirable due to certain
irreversible changes in catalyst effectiveness. On the basis of these
results, it was concluded that a baghouse temperature of 800* - 850"F was
consistent with both the project goal of 90% NOx removal, and the
temperature limits set by the nature of the catalyst.
SOz Ramoral
The primary S02-related objective of the SNRB laboratory pilot tests was to
determine the design and operating specifications required to achieve at
least 70% S02 removal with a hydrated lime sorbent. Overall SO2 removal
performance is affected by sorbent characteristics and operating variables
such as injection temperature, baghouse temperature, residence time, and
calcium/sulfur (Ca/S) stoichiometry. It was of particular interest to
evaluate the effect of heat extraction (via the simulated economizer)
between the point of sorbent injection and the inlet of the baghouse on S02
removal performance. A major portion of the testing was therefore devoted
to the relative amounts of S02 removal occurring in the ductwork and
baghouse. All of the laboratory pilot tests were conducted with a single,
commercially-available hydrated lime sorbent. Hence, no information was
developed on the potential benefits of using other sorbent materials.
Evaluation of alternative sorbents is planned for the field demonstration.
The cost of the S02 sorbent is a major contributor to the overall cost of
operating a SNRB system. The Ca/S stoichiometry, which is defined as the
molar ratio of injected Ca(OH)2 to S02 in the flue gas required to achieve
the S02 removal goal, is thus an important operating consideration. Tests
7-165
-------
were conducted over a wide range of stoichiometric conditions — Ca/S ranged
from 1.0 to 4.0 — to determine the impact on S02 removal performance. As
illustrated in Figure 7, it was found that above a stoichiometry of about
2.5, only slight improvements in S02 removal were obtained with increasing
stoichiometry. The approach of recycling the spent sorbent to improve
sorbent utilization was also investigated, but was found to be ineffective.
The reaction of the Ca(OH)2 and/or its derivatives with C02 in the flue
gases — which competes with the S02 reaction — renders the sorbent ineffec-
tive for re-use without intermediate processing.
In order to evaluate S02 removal performance over a range of temperatures
typical of utility boiler economizers, the sorbent injection temperature
was varied from 800* - 1100"F, resulting in baghouse temperatures of 600" -
850"F. Over the range of temperatures tested, S02 removal was essentially
independent of injection temperature. This suggests that the S02 removal
performance of commercial SNRB systems will not be adversely affected by
variations in economizer temperature within the tested range. However,
while the majority of S02 removal occurs in the duct, additional removal in
the baghouse — as the sorbent resides on the bags in the form of a filter
cake — is required in order to meet the minimum goal of 70% S02 removal.
The effect of baghouse operating temperature on this incremental baghouse
S02 removal is also illustrated in Figure 6. As the baghouse operating
temperature increases from 730* to 840"F, the total S02 removal increases
from 60% to 80%. The 70% sulfur removal goal is exceeded at baghouse
operating temperatures above about 800*F. It is hypothesized that the
incremental baghouse S02 removal is due to the fact that at the elevated
baghouse temperatures the sorbent continues to dehydrate, thereby generat-
ing additional surface area. In the operating range of 600* - 1000*F, the
rate of reaction of Ca(OH)2 with S02 is essentially constant. However, S02
removal is also a strong function of the available surface area of the
sorbent. If significant dehydration continues as the sorbent resides on
the filter bags, fresh surface is generated as the water vapor finds its
way out of the sorbent particles. This newly-created surface area then
becomes available for subsequent reaction with S02.
NQz/SOx Removal Compatibility
Another important factor addressed by the laboratory pilot tests is the
7-166
-------
compatibility of the S0X and N0X removal processes. After all, the SNRB
process requires that both objectives be accomplished in the same piece of
equipment. Figure 6 also illustrates the compatibility of these processes.
As previously mentioned, baghouse temperatures in the range of 800* - 850*F
are required to meet the NOx removal performance goal of 90% NOx removal.
The figure illustrates that this same range of baghouse operating tempera-
tures is required to meet the S02 removal goal of 70%. This important
result means that the two processes can be combined into a single SNRB
baghouse without sacrificing the performance of either system.
POTT.TMTKAKY ECONOMIC EVALUATION
The laboratory pilot tests provided a mechanism for assessing the impacts
of various design features and operating conditions on the overall perfor-
mance of the SNRB system. This information was subsequently used to
develop design specifications and testing plans for the 5-MWc SNRB demon-
stration facility. A preliminary economic assessment of SNRB was also
conducted in conjunction with the pilot tests to provide complementary
information on the cost impacts of the various potential design, perfor-
mance, and operating specifications for a commercial SNRB system. The
results of this study were subsequently used to prioritize the factors to
be investigated in the 5-MWc facility and to determine what additional
testing may be required.
Figure 8 provides a pictorial summary of the major factors contributing to
the overall levelized cost (in $ / ton of SOx + NOx removed) of operating a
500-MWc SNRB system. The case depicted is for a calcium-based SNRB system
in a new boiler application. The various costs indicated comprise all of
the incremental costs associated with the presence of the SNRB system
relative to those of the "base case" boiler system without the SNRB system.
For example, the capital cost component identified for the baghouse is
actually the incremental cost of the SNRB baghouse relative to the base
case particulate collector. Likewise, the cost indicated for waste dis-
posal is the incremental cost of waste disposal for the SNRB system rela-
tive to the fly ash disposal costs of the base case. In this regard it
should be pointed out that the base case boiler system did not include SOx
or NOx removal systems.
A detailed discussion of the economic analysis is beyond the scope of this
7-167
-------
paper. The purpose of presenting these results is to indicate the type of
information used to assess the sensitivity of the overall cost of a SNRB
system to the various cost factors. In particular, the figure clearly
illustrates the major influence exerted by: 1), the cost of the ceramic
filter bags and 2), the cost of the S02 sorbent.
The design of the 5-MWe facility and the associated test plan were there-
fore tailored to address the sorbent cost issue. In addition, another
testing program, utilizing the SNRB laboratory pilot baghouse, has been
developed to assess alternative filter bag fabrics.
TECHNOLOGY DEMONSTRATION
The commercial readiness of the SNRB technology is being evaluated through
the design, construction, and operation of a 5-MWe field demonstration
facility at Ohio Edison's R. E. Burger plant. A public dedication ceremony
was held at the R. E. Burger plant on May 9, 1991, and B&W Construction
Company's work on the 5-MWe facility was completed in September 1991.
Operation of the facility will begin in February, 1992, and continue
through September. The overall goal of the 5-MWe tests is to develop the
design, operating, and performance information needed by B&W to commercial-
ize the SNRB technology.
The specific objective of the field demonstration tests is to optimize the
SO2, N0X and particulate removal efficiencies during long-term operation on
fully integrated, commercial-size components. A test plan has been devel-
oped to facilitate an evaluation of the key operating parameters on SO2,
N0X and particulate removal performance. Several alternative hydxated lime
sorbents, in addition to commercial hydrated lime, will be evaluated for
S02 removal.
The cost effectiveness of the SNRB technology and the impacts of the
process on auxiliary plant equipment and operation will be evaluated. The
SNRB demonstration facility will allow evaluation of issues which could not
be adequately addressed at the laboratory pilot scale, including:
• Performance of the fully-integrated system on a long-term
basis
• Predictive performance curves for commercial applications
7-168
-------
• Control philosophy foe response to boiler load changes and
upsets
• Pressure drop across the SNRB baghouse modules and bag
cleaning procedures required to provide reliable, consistent
long-term operation
• Catalyst deactivation (catalyst life)
• Operating costs for the system
The findings of the SNRB 5-MWe field demonstration will be detailed in a
final report. The field demonstration will also support a detailed eco-
nomic and technical feasibility analysis of the technology, development of
the design specifications required for a commercial-scale system, and a
SNRB process control philosophy.
FuturM of the 5-Wa Facility
The SNRB field demonstration facility draws a 5-MWe (equivalent) flue gas
slip stream from Boiler No. 8. The SNRB baghouse consists of six indi-
vidual modules each containing 42 bag/catalyst assemblies, compared to the
single laboratory pilot module which contained 12 similar bag/catalyst
assemblies. It is designed to handle about 30,000 ft3/minute of flue gas
at a air-to-cloth ratio of 4:1. Other major features include:
• A Bailey Network 90™ system for integrated process control
• Automated ammonia injection system
• Five sorbent injection locations
A propane-fired heater for control of the sorbent injection
temperature
• Baghouse inlet and outlet flue gas heat exchangers to permit
flue gas temperatures to be accurately controlled
• Automated, pneumatic sorbent feed and ash disposal systems
A schematic of the 5-MWe SNRB field demonstration facility is shown in
Figure 9. A more detailed description of the facility is provided else-
where (1) .
7-169
-------
Filtar Fabric Dmlopautt
Initial economic analysis and pilot investigations identified the critical
role the filter bags will play in the technical and economic viability of
the SNRB process on a commercial scale. The filter fabric assessment test
program is designed to obtain extended term durability data for several
alternative high temperature bag fabrics. The test will also include an
evaluation of an alternative catalyst design.
The SNRB laboratory pilot baghouse was moved from BiH's Alliance Research
Center to the City of Colorado Spring's Martin Drake Plant located in
Colorado Springs, Colorado. Erection of the new pilot facility has been
completed. The pilot baghouse will be operated continuously over at least
a one-year period. Continuous operation will be interrupted periodically
to permit removal and examination of selected filter bags for analysis of
changes in the bag fabric characteristics over time. Intermediate results
from the pilot-scale test will be used to support consideration of a
possible extension of the base demonstration facility test plan in which an
alternative bag fabric would be evaluated under actual SNRB operating
conditions.
The pilot baghouse will draw flue gas from the economizer outlet of the
145-MWe coal-fired boiler. Inlet flue gas temperatures are expected to
range from 650" to 750"F. The baghouse will contain 12 full-scale bags
(6-1/4 inches in diameter, 20 feet long). Three alternative bag fabrics,
as well as alternative weave patterns, will be evaluated. A standard
pulse-jet bag cleaning cycle based on tubesheet pressure drop will be
maintained. The test will not involve full simulation of the SNRB process
since SO2 sorbent addition and ammonia injection will not be included.
However, significant data on bag fabric durability at elevated temperatures
under normal cleaning pulse flexure conditions will be obtained.
Enginaarlnj Study
The final phase of the program will include a detailed engineering study by
B&W s commercial Environmental Equipment and Energy Services Divisions for
the application of SNRB technology to one or more candidate commercial
boilers. This study will focus on the boiler plant modifications required
for integration of the SNRB technology with conventional boiler design.
7-170
-------
Economizer modifications and combustion air preheater design and operating
limitations will be specifically addressed.
smaaior
At baghouse operating temperatures of 800* - 850"F, the laboratory pilot
tests demonstrated that:
• Particulate emissions less than 0.03 lb/million Btu are
attainable with the Nextel™ filter bags.
• 90% NOx removal can be achieved at a NH3/NOx stoichiometry of
0.95 - 1.05. NH3/NOx stoichiometry is the primary parameter
affecting NOx removal performance. Baghouse operating
temperature appears to be of secondary importance.
• 70 to 80% S02 removal can be achieved over a Ca(0H)2/S02
stoichiometry range of 2.0 - 2.5. It appears that baghouse
operating temperatures of 800* to 850*F are necessary to
achieve 70% S02 removal.
Future testing will focus on assessment of alternative bag filter fabrics
and S02 removal performance with alternative sorbents. These two areas
appear to present the greatest opportunity for reducing operating and
capital costs associated with the SNRB technology.
Construction of the 5-MWe SNRB demonstration facility has been completed.
Shakedown of the process equipment and control systems is underway. In-
stallation of the filter bags and catalyst will be completed in January
1992. Operation of this larger scale facility in 1992 will support contin-
ued efforts to commercialize the technology. Potential sites for a commer-
cial demonstration of the technology are being pursued.
ACXKOHLEDGMENTS
The authors express their thanks to the D.S. Department of Energy (PETC),
the Ohio Coal Development Office, and the Electric Power Research Institute
for supporting the SNRB flue gas cleanup project. Additional thanks are
due to Ohio Edison for their support of the 5-MWe field demonstration, the
City of Colorado Springs Utilities for their support of the filter fabric
tests, 3M for supplying the Nextel™ bags, and the Norton Company for
supplying the SCR catalyst. The technical advice and encouragement offered
7-171
-------
by all of these organizations is also gratefully acknowledged.
RIRRDXZS
1. Chu, P., Downs, W., and Holmes, A.R., "Sorbent and Ammonia Injection
Opstream of a High-Temperature Baghouse," Environmental Progress, Vol.
9, No. 3, August 1990, pp. 149-155.
2. Chu, P., Kudlac, G.A., Wilkinson, J.M., and Corbett, R.W., ""Simulta-
neous SOx/NOx/Particulate Removal in a High Temperature Baghouse —
Clean Coal 2 Program Opdate," AFRC 1991 Spring Meeting on NOx Control,
Developments, and Commercial Applications, March 1991, Hartford, CT.
3. Kudlac, G.A., Farthing, G.A., Szymanski, T., and Corbett, R.W., "SNRB
Catalytic Baghouse Laboratory Pilot Testing," AIChE 1991 Summer Na-
tional Meeting, August 1991, Pittsburgh, PA.
4. Wilkinson, J.M., et al., "5-MWe SNRB Demonstration Project," EPRI/EPA
Ninth Symposium on the Transfer and Utilization of Particulate Control
Technology, October 1991, Williamsburg, VA.
7-172
-------
HIGH -TEMPERATURE
CERAMIC BAG
ALKALI
INJECTION
AMMONIA
ALKALI
RICH ASH
INLET FLUE GAS
• NO,
• SO,
• PARTICULATES
^ ASH
Figure 1. SOx-NOx-Rox Box Procass
M4, SUPPLYH
AMMONIA
HANDLING A
PUECnON
BOILER
COMBUSTION
ZONE
\ ECONOMIZER
OUTLET
- FUEL
- COMBUSTION AIR
BOILER
BOTTOM
ASH
'
'
DISPOSAL
(EXISTMG)
7
/
CATALYST
AIR
HEATER
HOT
BAGHOUSE
h€n
STACK
HKjH*TEMP.
BAGS
Figure 2. Commercial SKRB Application
7-173
-------
PULSE-JET
BAGHOUSE W/
CATALYST
COMBUSTOR
WATER STEAM
STACK
WASTE
ID FAN
ECONOMIZER
NH, INJECTION
SORBENT INJECTION
0-
GAS SAMPUNG LOCATIONS
Figur* 3. SHRB laboratory Pilot Facility
0.08
0.06
Particulate
Emissions 0.04
pbs/million BtuJ
0.02
0.00
s With Catalyst
¦ Without Catalyst
1
Method 5
Method 17
0.03 lbs/million Btu (NSPS)
J3L
Bags Removed
Ll
SI
IS]
I
M
1
1
PL
f
P
(II
SEP OCT NOV
DEC JAN FEB
rigor* 4. Particulate Kaissions Results
7-174
-------
NOx Removal
[%] Ammonia
Slip [ppm]
Ammonia/NOx Stoichiometry
Figure 5. Effect of HH3/HOx on MOx Removal Performance
1 1 /
~ a °
%8 °D
~° o2
C
780-840 F Catalyst Tetiipcuture ^ c
°
cP NOx
~
Removal
¦
. Stoichiometric
NOx Removal
1 l
¦ ¦
¦
Ammonia Slip
0.40 0.60 0.80 1.00 1.20 1.40
100.0
80.0
SOx or NOx
Removal [%]
60.0
40.0
20.0
0.0
600
NOx Removal
Removal
700 800
Baghouse Temperature [F]
900
Figure 6. Sff«ct of Baghouse Operating Temperature on SOx and HOz
Removal Performance
7-175
-------
Total SOx
Removal [%]
100
80
60
40
20
0
&
&
A
&
Aflfc a
a A
*
A
&
A
950-1050 F Injection Temperature |
>780 F Bag Temperature 1
1.5
2.5 3
Ca/S Stoichiometry
3.5
Figure 7. Effect of Ca/S StoicfcLiametry on S02 Removal
Baghouse
Auxiliary
Waste Disposal
Catalyst Replacement
Filter Bag Replacement
O&M Cost Factors
Capital Cost Factors
(Shaded)
Filter Bags
Catalyst
Other Equipment
Indirect Costs
Fixed O&M
SOiSorbent
Figure 8. Breakdown of SNRB Levelized Costs
7-176
-------
H
BOILER
AIR
HTR.
N/
PROPANE-
AIR-
EXt9T1NQ FLUEWORK
)-
COMBU9TOA
5 MW 8UP8TREAM
•50*F
10,250 8CFM
EXHAUST A
NH|
AIR
3
FLUE OAS
MONITOR
AIR
FLUE OA8
COOLER
AIR
wj
EXISTING
ELECTROSTATIC
PRECIPITATOR
(ESP)
WW
EXISTING
STACK
RETURN SLIPSTREAM
ITS • 300'P
FLUE
OAS
MONTOR
_a
FLUE
OAS
MONITOR
BAOHOUSE
TfTTTt
BOOSTER
FAN
FLUE OAS HEAT L/C"?
RECOVERY UNIT
8ILO
OONVEYINO
SYSTEM
y
CONVEYING
SYSTEM
8ELF UNLOADING
SORBENT TRUCK
5=3Q
SILO
BAOHOUSE
PRODUCT
SOUD WASTE
TO DISPOSAL
J
Figure 9. Schematic of the 5-MW, SNRB Field Demonstration Facility
-------
Intentionally Blank Page
7-178
-------
Reaction of Moist Calcium Silicate Reagents with Sulfur
Dioxide in Humidified Flue Gas
7-179
-------
Intentionally Blank Page
7-180
-------
Wojciech Jozewicz
Acurex Corporation
Environmental Systems Division
P.O. Box 13109
Research Triangle Park, NC 27709
Gary T. Rochelle and David E. Stroud
Department of Chemical Engineering
University of Texas
Austin, TX 78712
ABSTRACT
Experimental results are presented for enhancing SO2 reaction by the addition of
moisture to calcium silicate solids. The presence of up to 30-percent initial free
moisture increased short-time (<5 s) conversion of calcium silicate solids and of
physically mixed Ca(OH)2/fly ash solids. The conversion decreased when the initial
free moisture increased beyond 30-50 percent. The tendency of solids to agglomerate,
as represented by the critical moisture content, was a function of surface area. The
critical moisture content was directly equivalent to the pore volume of the solids.
7-181
-------
REACTION OF MOIST CALCIUM SILICATE REAGENTS
WITH SULFUR DIOXIDE IN HUMIDIFIED FLUE GAS
INTRODUCTION
Dry sorbent duct injection is a viable alternative as a tail-end process for the removal
of SO2 from the gases emitted from coal-fired power plants. Moist solid sorbents are
injected into ductwork containing humidified gases from coal fired power plants. The
sorbent reacts with the SO2 in the flue gas and forms calcium sulfite and sulfates. The
resulting product is removed from the system by particulate control equipment as a dry
solid.
Within the process, gas-solid reaction occurs between the SO2 in the flue gas and the
reagent particles injected into the duct leaving the power plant. Various properties of
the sorbent can be examined in order to better understand the reaction according to
first principles. One area of focus is to analyze the effect of the initial moisture of the
sorbents injected into the duct. Experiments to document the effects of various
parameters on the reactivity of the sorbents with the SO2 were conducted in an
isothermal, packed bed reactor to simulate in-duct conditions and bag filter conditions
within the ADVACATE process (dry sorbent duct injection). Significant improvements
in the operational potential of this process will result from effective understanding of
these parameters, and these advances, combined with Federal and State legislation
to abate acid rain, could lead to widespread industrial application of this technology.
Details of this work can be found in references [1], [2], and [2].
EXPERIMENTAL TECHNIQUES
Sorbent Preparation
The two types of reagents used most often in this study were commercial hydrated lime
supplied by Mississippi Lime Company and calcium silicate sorbent prepared by
heating a slurry of Clinch River or Meredosia fly ash with Ca(OH)2. Fly ash and
Ca(OH)2 were slurried for 3-24 h at 90°C, most often at a fly ash loading of 3 g/g
7-182
-------
Ca(OH)2- In soma cases, fly ash was ground on a bench-scale attritor prior to slurrying
with Ca(OH)2- Also tested were solids produced by physically mixing fly ash and
Ca(OH)2- See Table 1 for information on the fly ashes used.
Table 1
REAGENTS TESTED FOR S02 REACTIVITY
Slurrying Sorbenl
Fly Ash Source
Fly ash
Type
Time
P>]
Temperature
[•CI
Fly Ash Loading
[g/g Ca(OH)£
Surface
Area
lm2/ql
Pore
Volume
fcn^/ql
Median Pore
Size
TA1
Meredosia
U
8
90
3
6.9
0.039
225
Meredosia
G
3
90
3
27.4
0.122
178
Meredosia
G
3
90
2
17.5
0.100
229
Meredosia
G
3
90
1
16.6
0.132
239
Meredosia
G
8
90
3
33.2
0.167
201
Meredosia
G
24
90
3
4-7.0
0.213
180
Clinch River
G
3
90
3
30.2
0.132
174
U=Unground. G=Ground.
Meredosia U surface area =1.2 rr^/g
Meredosia G surface area = 6.8 rr^/g
Clinch River G surface area = 5.6 rr^/g
Agglomeration
To study the effects of moisture content on agglomeration of the calcium-based solids,
a 5-g sample of the calcium based solids was placed on an 80-mesh sieve (177 jim
between wires) and shaken for 10 min. Tests were run on samples with varying
amounts of moisture to determine the amount of moisture that caused half the sample
to be retained by the 80-mesh sieve. This amount is called the critical moisture (grams
of water/grams of wet solids expressed as a percentage).
A standard, 8-in diameter, full-height (2-in depth to cloth), 80-mesh sieve with pan and
cover from Brainard-Kilman was used. Water was added dropwise to 5 g of a dry
powdered sample in a glass beaker. The water was mixed in manually with a metal
chemical spatula. The mixture was spread against the beaker wall and repeatedly
chopped in an attempt to distribute the water uniformly and break any clumps. The
sample was weighed before wetter was added and after the water was mixed in to
determine the initial moisture content. After the moist sample was weighed,
7-183
-------
subsequent steps were performed without delay to minimize evaporation. Each
complete test took about 15 min.
The moist sample was spread uniformly over the sieve, the lid placed on, and the
assembly placed in a mechanical sieve and the material collected in the pan was
weighed. This procedure usually was repeated on 3 or more samples at different
moisture content until the critical moisture was bracketed and could be determined by
linear interpolation. A fresh sample was used for each test, but all samples came from
the same experiment.
Reactor
Experiments for reaction times of 0.5-400 s (most often 0.5-5 s) were performed in the
Short-Time Reactor (STR). The system used a fixed-bed reactor centered around a
pneumatically actuated sample slider assembly. Moist solids were dispersed on
quartz wool and placed in the STR. No conditioning of the solids prior to the contact
with SO2 has been implemented. Exposure to the SO2 containing process gas was
preceded by a thorough dispersion of solids on quartz wool. The amount of excess
moisture (free moisture) was determined as weight loss detected by a
thermogravimetric analyzer upon heating from room temperature to 100°C.
After reaction with the SO2 containing gases, the solids were slurried in an aqueous
solution of HCI (to dissolve calcium) and H2O2 (to oxidize sulfite to sulfate). The flask
containing the slurry was sealed and agitated for at least an hour. The resulting
solution was filtered and then diluted. Atomic absorption was used to analyze the
solution for Ca and ion chromatography was used to determine the sulfate content.
Calcium conversion was calculated as the ratio of sulfate to calcium in the solids.
THEORY OF REACTIVITY WITH MOIST SOLIDS
The key variables that affect the gas-solid reaction between SO2 and Ca(OH)2-based
sorbents during dry-sorbent injection are the relative humidity and SO2 concentration
of the flue gas and water content and surface structure of the sorbent. Previous spray
dryer studies involved the reactivity of sorbent particles suspended in liquid droplets,
which as they dried, passed through the liquid, free moisture, and equilibrium stages
of reactivity. In the equilibrium stage, the particle is assumed to contain surface water
in equilibrium with the relative humidity of the flue gas. While the total reactivity of
such a solid has been studied, there have been no efforts to decouple these effects.
Mass transfer theory may be applied to the gas-solid reaction system in the presence
7-184
-------
of moisture in excess of equilibrium on the surface of a reactive sorbent. The
theoretical model below describes the reaction occurring when a wet solid is drying
and relates the evaporative drying and the SO2 capture mechanism.
The model is based on these assumptions:
1. Both water evaporation and SO2 absorption are controlled by gas
phase diffusion.
2. Diffusion occurs in the pores by the Knudsen mechanism.
3. Liquid water is present on the solid predominantly in the pores.
The fluxes of SO2 and H2O are given by:
d(SOg/area) _ NsQ2 = -(Ds0l) (1)
d(HpO/area) = Nh;0 , ,(Dh,o) (2)
where:
Nj = molar flux:[mole i/(area-time)]
x = diffusion distance; [length],
Ci = concentration i; [mole i/volume], and
Dj = Knudsen mass diffusion coefficient [length2/time]
The Knudsen diffusion coefficient is inversely proportional to the square root of the
diffusing species. The ratio of the fluxes gives a relationship between the water
evaporated from the solid and the SO2 captured:
aS°2 _ Dsog(Pso?) _ -/ 18 Pso? f3.
AHzO - DH2o(PH20*-PH20) " V 64 P*H2o - PhzO ™
where:
ASO2 = SO2 captured; [moles/g sorbent]
AH2O = H2O evaporated; [moles/g sorbent]
PSO2 = SO2 partial pressure in bulk gas
7-185
-------
PHaO = H2O partial pressure at gas-liquid interface
P*H20 = equilibrium H2O partial pressure at the gas-liquid
interface.
This result assumes that the equilibrium SO2 partial pressure is zero. The gas/liquid
interface is assumed to be at the dry bulb temperature of the gas. The equilibrium H2O
partial pressure is taken as the H2O vapor pressure at that temperature. The model
allows one to predict the conversion of the porous sorbent as a function of the quantity
of water evaporated from the sorbent while in contact with SO2.
In each experiment, the initial water on the reactive sorbent, the relative humidity of the
flue gas, and the SO2 content of the gas are known. Assuming that all of the initial
water evaporates during an experiment, the expression is modified to predict the
additional Ca utilization of the sorbent based on these parameters:
This representation does not account for reaction that occurs when the sorbent
contains water in equilibrium with the flue gas relative humidity. To account for this
effect, data must be taken to obtain the instantaneous conversion for the reaction of
the sorbent in equilibrium with the SO2 containing gas under the same conditions as
the sorbent with initial free moisture. We have assumed that the total conversion
should be the sum of the expected conversion with equilibrium moisture and the
conversion predicted by Eq. (4).
RESULTS
Agglomeration of Moist Solids
The tendency of solids to agglomerate was quantified by measuring the critical
moisture by the method described above. Calcium silicate solids were prepared from
Mississippi Lime Hydrate and fly ash or silica fume and tested for critical moisture
content Solids with a greater value of critical moisture have a greater capacity for
retaining moisture while still being free-flowing.
When critical moisture was expressed as moisture ratio (g water per g of solid), it
correlated absolutely with solids pore volume (cm3/g) as shown in Figure 1. Each cm3
A/l8 pso? fAH2Q>|
y 64 (PWPhjO) I Ca J
(4)
where Ca = Ca content of the sorbent; [moles/gram sorbent]
7-186
-------
of absorbed water at the critical moisture corresponds to one cm3 of particle pore
volume. When expressed in terms of critical moisture the relationship is
Critical Moisture (%) = vSTml] <5)
Agglomeration occurs when the volume of free moisture exceeds the pore volume.
Pore volume and surface area can be related by the cylindrical pore model. Figure 2
shows the relation between the values for calcium silicate sorbents which is given by
the equation
Pore Volume = 0.0055*[Surface Area] - 0.0055 (6)
The slope of this correlation implies a pore radius of 110 A. Critical moisture increased
with increasing surface area of solids. The combination of equations (5) and (6) from
Figures 1 and 2 gives the correlation
Critical Moisture (%) ° ,.0 ^OOSS-'tSurtace Area] W
A linear best fit of the data in Figure 3 gives
Critical Moisture (%) = 0.425*[Surface Area] +1.3 (8)
To test the effect of fly ash loading on critical moisture, solids were prepared from
slurries of ground Clinch River fly ash (7.5 m2/g) with calcium sulfite hemihydrate,
gypsum, and varying fractions of Ca(OH)2 present. The fly ash loading was varied
from 0.25 to 4 g/g Ca(OH)2. Extremes of no Ca(OH>2 and no fly ash were represented
by dry mixtures of ash, hemihydrate and gypsum (1:0.75:0.25) and by Ca(OH)2 alone.
Figure 4 shows the effect of Ca(OH)2 fraction on both the critical moisture and surface
area of solids prepared by slurrying at 90°C for 8 h. Both properties peak in the solids
with a fly ash loading of 1.0. With greater amounts of Ca(OH)2. excess unreacted
Ca(OH)2 may dilute the properties of the solid product. Lesser amounts of Ca(OH)2
may starve the slurry solution for alkalinity. Also shown in Figure 4 are data for solids
prepared without hemihydrate or gypsum present. For a given fly ash loading, the
presence of hemihydrate and gypsum in the slurry results in solids with greater critical
moisture and surface area.
7-187
-------
Reactivity Results
Reactivity of solid sorbents with SO2 was measured as a function of the initial moisture
content of the solids. The experiments in the Short Time Reactor were performed with
three types of solids:
1. Unreacted mixtures of Mississippi lime hydrate and Meredosia fly
ash
2. Calcium silicate prepared from unground Clinch River ash
3. Calcium silicate prepared from Meredosia ash with various levels
of grinding.
Physical mixtures of 6 mg Ca(OH)2 and 20 mg unground Meredosia fly ash were
exposed to 2000 ppm SO2 in the STR operated at an 11 °C (20°F) approach to
saturation (approximately 60 percent relative humidity [RH]). Baseline experiments
were performed with dry Ca(OH)2 only and then with the dry mixture. Following
baseline tests, solid mixtures were used with 3.9, 12.4, and 18.3 percent initial free
moisture. Figure 5 gives the conversion of calcium as a function of contact time for dry
Ca(OH)2 and for the mixtures. Higher conversions were measured for solids with
increased level of initial free moisture than for dry Ca(OH)2. For a given contact time,
the conversion generally increased with increased initial free moisture. The effect of '
initial free moisture was more pronounced at longer contact times (60 and 400 s). For
example, the conversion at 5 s contact time was 2.8 and 4.7 percent and at 400 s
contact time was 7.3 and 36 percent for dry Mississippi Ca(OH)2 and a mixture of dry
Mississippi Ca(OH)2 and moist Meredosia fly ash with 18.3 percent initial free
moisture, respectively.
In a different series of experiments, calcium silicate solids (30.2 m2/g) were prepared
by slurrying 3 g ground Clinch River Fly Ash/g Ca(OH)2 at 90°C for 3 hours. Baseline
reactivity experiments were conducted with equilibrium moisture. Considerably
higher conversion was measured for solids produced by slurrying Ca(OH)2 and
ground fly ash than previously measured for dry Ca(OH)2 alone. For example, with 5 s
contact time, the conversion of dry fly ash/Ca(OH)2 sorbent was 7.4 percent, compared
to 2.8 percent for dry Mississippi Ca(OH)2. Solids were then prepared by re-wetting
calcium silicate solids (30.2 m2/g surface area) to yield 3 levels of initial free moisture:
8.5, 12 and 40 percent. The results are shown in Figure 6, giving the conversion of
calcium as a function of contact time with 2000 ppm SO2 at approximately 60 percent
7-188
-------
RH. For each initial moisture tested, the conversion increased with contact time.
Except for 0.3 s contact time, the conversion increased with initial free moisture.
The comparison of conversion measured for all three types of solids is shown in
Figure 7. With the 5 s contact time, calcium silicate solids consistently yielded higher
conversion than physically mixed fly ash/Ca(OH)2 solids. The physically mixed fly/ash
Ca(OH)2 was tested with up to approximately 20 percent initial free moisture. Calcium
silicate solids were tested with up to approximately 60 percent initial free moisture.
Attempts to produce a physically mixed Ca(OH)2 and fly ash with initial free moisture
above 20 percent were unsuccessful. Figure 7 gives calcium conversion with 5 s
contact time in the STR at 2000 ppm SO2 and approximately 60 percent RH. The 5 s
conversion of calcium silicate solids leveled off at 30-40 percent initial free moisture.
Increasing the amount of an initial free moisture above 40 percent resulted in
decreased conversion with SO2. With short contact time (0.5-1 s) of calcium silicate
solids with SO2, the effect of an increased amount of initial free moisture was weak,
and the maximum conversion of calcium was only about 5 percent. Significantly
increased conversion as a result of increased levels of initial free moisture took place
for 3 and 5 s contact time, as evident from Figure 8.
Reactivity of four calcium silicate solids with surface area varying from 6.9 to 47.0 m2/g
and with an initial free moisture content of up to 60 percent is presented in Figure 7.
Figure 9 shows the general effect of contact time and initial moisture on the reactivity
of the calcium silicate solids with the SO2. As contact time increases, Ca conversion
increases. As the initial moisture increases for a given contact time, Ca conversion
generally increases until a maximum initial moisture is reached beyond which
conversion dramatically decreases. Maximum conversion with S02was measured for
each solid. The maximum conversion increased with increasing solids surface area.
For 6.9, 27.4, 33.2, and 47.0 m2/g solids, the corresponding maximum conversion with
SO2 was 12.3, 15.2, 17.5, and 22.2 percent, respectively. The initial free moisture
content for which the maximum conversion for given solids was measured (optimum
moisture content) varied with specific surface area as shown in Figure 10. Increasing
the amount of initial free moisture above the optimum caused conversion to decrease.
A possible explanation for the maximum conversion is the agglomeration of the solids
at greater moisture content. Figure 10 also shows the critical moisture content of the
solids estimated by Eq. (8). The critical moisture content has the same trend as the
optimum moisture content giving maximum conversion. However, the critical moisture
content is less then half of the optimum moisture content.
7-189
-------
The theory models the trend of the Ca conversion well up to the critical moisture
content of each sorbent Above the critical moisture content the solids may form large
agglomerates. These agglomerates may not dry completely in the short-time
experiment and the assumptions of the theory would not apply.
Analysis of data presented in Figure 7 indicates that SO2 conversion of solids
containing initial free moisture is a weak function of solids surface area below optimum
moisture content of 6.9 m2/g solids (30 percent initial free moisture). However,
increasing the surface area of solids provides for higher optimum moisture content
with resulting higher conversion of solids with SO2.
CONCLUSIONS
Initial free moisture increases short-time (<5 s) conversion of calcium silicate solids
and of physically mixed Ca(OH)2/fly ash solids.
Agglomeration of calcium silicate solids due to moisture is a function of their surface
area. The critical moisture content at which agglomeration occurs is directly
proportional to solids pore volume. The presence of calcium sulfite hemihydrate and
gypsum in a slurry of ground low calcium fly ash results in solids with increased critical
moisture and surface area.
ACKNOWLEDG ME NTS
This work was funded by the Department of Energy (Contract DE-AC22-88PC88874)
and by the Texas Advanced Technology Program (Grant No. 003658-334). Stephen
P. Beaudoin [l] initially developed the theory. Harold L Johnson assisted in the final
preparation of this paper.
REFERENCES
1. Beaudoin. S.P.. The Effect of the Reactivity of CafOH^-Based Sorbents For
Flue Gas Desulfurization. M.S. Thesis, The University of Texas, Austin, Texas,
August 1990.
2. Stroud, D.E., Agglomeration of Damp Calcium Silicate Sorbents For Flue Gas
Desulfurization. M.S. Thesis, The University of Texas, Austin, Texas, August,
1991.
3. Jozewicz, W. and G.T. Rochelle, "Theoretical Approach for Enhanced Mass
Transfer Effects in Duct Flue Gas Desulfurization Processes," DOE Topical
Report for Task 4 - Novel Techniques, Contract No. DE-AC22-88PC88874,
September 17, 1991.
7-190
-------
o
IA
-S®
s-
V
«
w
o
Q£
vs
©
s
0.5
0.4
0.3
0.2
0.1
0.0
-i—i—.—i—.—|—i—i—i—i—r
i r ¦ r-
0.0 0.1 0.2 0.3 0.4
Pore Volume (cc/g)
0.5
FIGURE 1: Correlation of critical moisture with pore volume for calcium silicate
solids.
M
O
o
E
0.5
0.4
0.3 -
£ 0.2
U
o
^ o.i
o.o
1 1 ' 1 ' 1 ' 1 ' 1 1
- Pore Volume =0.0055(SA)-0.0055
i
.D"
i '
¦
1 1 | 1 1
'£
~ ¦
¦
-
m ¦
83-'"
¦ -
- . i . i i . i . i
"
i
10 20 30 40 50 60 70 80
Surface Area (m2/g)
FIGURE 2: Correlation of particle pore volume with specific surface area (solid
symbols represent data for sorbents in Table 1; open symbols are for additional
samples represented in Figure 1; dotted lines and equation are a least-squares fit of
the data}.
7-191
-------
¦ rKv>r\* Rjver
~ Ground
A Harrington
O Silica Fume
CJ
Urn
3
W'O
V5
o
S
* ~
L.
u
5
10
50
100
Surface Area (m2/g)
FIGURE 3: Correlation of critical moisture with surface area. Slurry times range from
15 min to 24 h. Dotted line represents the correlation predicted by Eq. (7); the solid
line is a least squares fit of the data given by Eq. (8).
70
i 50
Surface Area
es 3
oi
k. M
< ©
U
CO —
«E CQ
I- V
S ^
Critical Moisture
o.o
0.2
0.6
0.8
0.4
1.0
Fly Ash Fraction (g Ash/[g Ca(OH)2 + Ash])
FIGURE 4: Effect of Ca(OH)2 fraction on the surface area and critical moisture
content of solids prepared by slurrying for 8 h ground Clinch River fly ash, calcium
sulfite hemihydrate and gypsum, weight ratio 4:3:1 (respectively) held constant each
time.
7-192
-------
100
o
CO
>
c
o
O
CS
o
10
;
1 T H
4
•
¦
O
Dry. Hydrate
¦
•
¦
~
3.9 % Moisture, Mix
¦
O
12.4% Moisture. Mix
X
B
X
18.3% Moisture, Mix
X
0
f
0
~
|
0
j
¦
~
•
x ^
0
¦
i a a
.
* ^ 0 0
~ 0
9
01 , 1
¦
0.1 1 10 100 1000
Time (s)
FIGURE 5: Effect of initial free moisture on the reactivity of Ca(OH)2 physically mixed
with moist Meredosia fly ash (STR, 2000 ppm SO2, 60% relative humidity).
100
<0 10
o
"52
C
o
O
CO
O
1 1 1 1
1 O 8.5% Moisture
1 O
~ 12% Moisture
~
0 40% Moisture o
X 12.4% Moisture. Muc
! 0 i X
: t
i <-> i
e :
X
~
: ~ 0 j
: k ° x !
ft X
~ T
O X
•
x—i i
0.1 1 10, . 100 1000
Time (s)
FIGURE 6: Effect of initial moisture on the reactivity of 30.2 m2/g surface area
calcium silicate solids prepared from Clinch River fly ash. Reactivity comparison for
physically mixed and chemically reacted fly ash/Ca(OH)2. (STR, 2000 ppm SO2. 60%
relative humidity, 11 °C approach to saturation) Mix is a mixture obtained by physically
mixing Meredosia fly ash with Mississippi hydrated lime.
7-193
-------
o
(O
a>
>
c
o
O
CO
O
36
30
24
18
12
O 6.9 m2/g
~ 27.4 m2/g
~ O 33.2 m2/g
X 47.0 m2/g
S— Theory
1
1
1
/'
A Mix
• 30.2 m2/g
X
¦
i •
9 9
0
¦
;x O
¦ ^ * c
a O jO i
• ^
3/
* 0
•
~
•
c
0
*0
3
" ^
1
•
10
60
70
20 30 AO 50
Initial Free Moisture, (wt%)
FIGURE 7: Effect of solids surface area, initial free moisture, and sorbent type on the
reactivity of moist solids and comparison with Additive theory. (STR, 2000 ppm SO2,
60% relative humidity, 5 sec contact time)
20
16
>0
12
c
o
¦«
® 8
c
o
O
w A
O 4
1
1
;
* - - - -*
/
/
/
/
0
X
X
V
X'"""
/ «
* /!
/
.
\
\
Kx%
\
V
1 ' ' 1
* ,*
« *
—
*
-e
N
N
L
\
* » __
¦
!
>0 |0.5§\J \! _
i
•
i ° r€>
i i i
10
20
30
40
50
60
70
Initial Moisture [%]
FIGURE 8: Effect of contact time on the reactivity of Clinch River calcium silicate
solids, surface area 30.2 m2/g. (STR, 2000 ppm SO2, 60% relative humidity)
7-194
-------
30
25
20
o
"to
w
©
>
C
o
O
CO
O
15
10
~
o
x
Dry, No Moisture Added
10% Initial Moisture
30% Initial Moisture
40% Initial Moisture
60% Initial Moisture
Contact Time (sec)
FIGURE 9: Effect of initial free moisture on the reactivity with SO2 of 47.0 m2/g
sorbent (ground Meredosia fly ash/Ca(OH)2; loading 3 g/g; 60% relative humidity,
2000 ppm SO2).
60
o
"co
h—
c
o
O
50 ~
40
30 ~
20
Optimum Moisture
Maximum Converse
Critical Moisture
60
50
40
o
to"
30 2
20 —
1 0
10
20
30
40
50
Surface Area (m /g)
FIGURE 10: Effect of surface area on optimum moisture and maximum conversion.
(STR, 2000 ppm SO2, 60% relative humidity, 5 sec contact time)
7-195
-------
Intentionally Blank Page
7-196
-------
Commercial Application of Dry FGD using High Surface
Area Hydrated Lime
7-197
-------
Intentionally Blank Page
7-198
-------
F. Schwarzkopf
Consultent
1 Half Penny Circle
Savannah, Georgia 31411
H.P. Hennecke
A. Roeder
Rheinische Ka 1 ksteinwerke QnbH
Wilhemstrasse 77
5603 Wuelfrath, FRG
ABSTRACT
Details are provided of the performance and economics of first-of-its-kind
advanced technology for production and use of high surface area hydrated lime
(HSH) presently employed by 40 west European industrial emission sources mainly in
Germany, inclining coal fired boilers, for removal of 902, HC1 and other flue gas
pollutants. A description is also given of practical engineering experience in
retrofitting and commissioning of dry sorption by HSH in bituminous coal service.
Recent technological advancements discussed include ccnmercial operations using
HSH bolstered with carbonaceous additive to achieve simultaneous removal of toxic
organics and trace metals along with acid gas components. An assessment is made
of the applicability of the technology for retrofit installations in medium/high
sulfur coal service in the U.S.A. A cost benefit evaluation compares dry
scrubbing with HSH against semi-dry (spray dryer) scrubbing using quicklime.
INTRODUCTION
As an outgrowth of the massive national S02 emission reduction program that began
in Germany in 1983, dry sorption technology utilizing unique, high surface area
hydrated limp (HSH) has gained extensive, large-scale utilization in an array of
Industrial and municipal services. A ccnmercial HSH production facility in that
country supplies an array of boilers, incinerators and other sources, providing
tailored reagent for cost effective control of acid gas and toxic emissions,
typically without need for an excessive, impractical increase in flue gas
humidity.
7-199
-------
PROGRESS IN LIME SORBENT DEVELOPMENT FOR DR5f REMOVAL
Dry sorption is a simple process for flue gas cleaning characterized by low
investment and is therefore especially attractive for many retrofit applications.
The disadvantage of generally higher reagent consumption in dry operation is
further compensated by use of novel improved reagents or mixps of such reagents
designed for advantageous use in the specific application. To sorb diverse acid
waste gas components such as hydrogen fluoride, hydrogen chloride and sulfur
dioxide, lime hydrates with high surface area (HSH) have been developed (1J and
these materials are now in commercial use (2J. Alternatively, regular hydrated
lime is mixed during hydration with components such as sugar to enhance reagent
effectiveness for increased efficiency of S02 removal (3) . An additional
important development is the use of limp hydrate, particularly HSH, intermixed
with pulverized lignite coke (a carbonaceous additive) to gain simultaneous
removal of acid gas components, trace quantities of mercury, cadmium, etc. (4j, as
well as chlorinated organics: PCDD (diaxin) and PCDF (furan).
TmtirnuMi lime reagents such as HSH have the following advantages over normal
hydrated lime:
• A smaller quantity of reagent is required.
• Higher pollutant removal rates are achieved.
• The quantity of wastes to be disposed is less.
Obviously such improvements must not be negated by excessive reagent cost. High
reagent cost for such improved reagent supply may be tolerable as long as dry
sorption is competitive with other processes as in S02 removal applications for
utility boilers. A notable example of improved lime sorbent is the present,
first, commercial-scale production of HSH (tradename-WUELFRAsorp) and its use by
more than forty (40) industrial and municipal emission sources in western Europe,
primarily Germany.
COMMERCIAL HSH PRODUCTION IN EUROPE
A commercial scale unit, previously described in detail in the U.S. (2), began
operation two years ago at a rate of six metric tons per hour (40,000 tons per
year). This sorbent is produced in an alcohol hydration process protected by US
and other patents and primarily characterized by manufacture of a product with
warranted BET surface area of greater than 35 m2/g. This value is twice that of
regular lime hydrate, slaked with water, which has a BET valixe of approximately 17
n^/g. Compared with conventional hydrate, HSH is superior because of very small
7-200
-------
grain size ranging between 2.0 and 2.5 micrometers. Furthermore, HSH has very
high flowability and, injected with carrier air, is readily distributed throughout
the waste gas to be cleaned, even at carrier air rates well below l m3 per
kilogram HSH. The disadvantage of HSH is low bulk density, approximately 300
grams per liter (18 lb/ft3), which is 25* less than the 400 grams per liter (24
lb/ft3) density of normal lime hydrate, and greater storage volume is needed per
weight unit of lime stored.
The cannercial HSH production in Germany has a CaO content of more than 73.0% with
a fixed alcohol content of approximately IX. In addition to HSH as is, mixtures
of it with up to 20* pulverized lignite coke, (designated WELFRAsorp C-20), have
been available. There is presently a marked increase in demand for such tailored
lime/coke mixture calling for supply with a coke content of only 5-8*.
The present FOB-factory selling price of HSH manufactured in Europe is
approximately 40* above the price of normal hydrate. However, diverse HSH
consumers readily compensate for this premium price by benefit of savings and
improvements such as decreased sorbent consumption, smaller waste disposal
quantity and, above all, assured compliance with mandated emission limits.
Ongoing improvements in economics of the manufacturing process are expected to
lower the FOB-factory price to a level of only 20-25* above the selling price of
normal hydrate. It is likely that this reduced level of price premium will also
be applicable in HSH supply from future U.S. facilities for manufacture of HSH.
The anticipated substantial decrease in manufacturing costs can be achieved in new
production plants by process modifications to the effluent return to the
continuous lime hydration reactors 1 and 2, shown as iton A in Figure 1. This is
significant in lime hydration because in the presence of alcohol the reactivity of
water with calcium oxide is depressed to a degree that varies with the amount of
alcohol present. The contemplated design changes allow control of process water
reactivity learning to higher production rates as well as more complete hydration
reaction. This in turn will allow elimination of the batch vacuum operation
illustrated by item B of Figure 1. These significant improvements are the direct
result of the learning curve that has been made available through sustained
operation of commercial facilities for HSH production.
OVERVIEW OF COMMERCIAL SUPPLY OF HSH PRODUCT
HSH is consumed in a wide range of European industries and municipal services, the
nine most important of which are shown in Table 1. They are listed in the order
7-201
-------
of decreasing consumption, i.e. special waste incinerators (hazardous waste
incinerators) consume the largest portion and the glass industry the smallest
amount of the total present production of 10,000 annual metric tons. The supply
to utilities for flue gas S02 removal currently ranks sixth. As is to be
expected, the production facility is only operating at 25* of capacity since
WUELFRAsorp and WUELFRAsorp C are new products for flue gas cleaning which
continue to be systematically introduced to candidate consumers via testing
programs through which the demand is being expanded so as to gain a more
substantial market.
COMMERCIAL APPLICATION OF HSH PRINCIPALLY FOR HCL REMOVAL FROM WASTE GAS STREAMS
The degree of removal of HC1 by dry sorption is proportional to the magnitude of
the surface area of the hydrate solids used. Application of HSH for this service
is therefore especially effective. Moreover, the rate of removal is independent
of the reaction temperature over a wide range of temperatures.
Sinale-Staqe Operation
The flow sheet of a relevant single-stage flue gas cleaning facility serving a
hazardous waste incinerator is shown in Figure 2 and the performance results are
given in Table 2. The HSH is injected at a rate of 200 kilograms per hour at
180°C (356°F). The fabric filter collecting the gasborne solid waste is operating
at the same temperature. It is of critical importance to this reagent user that,
despite high and widely varying HC1 and S02 concentration in the raw gas, the
legally required emission limits be maintained in a manner reflected by Table 2.
The system operates with very high reliability removing more than 99* of HC1.
Two-Stage Gas Cleaning
In a recent new vise of HSH at a municipal waste incinerator served by two-step gas
cleaning, a snail portion of the HSH is injected at 250°C (482°F), upstream of
flue gas cooling using a heat exchanger. Downstream of gas cooling, HSH
intermixed with 8* pulverized lignite coke is injected in combination at 150°C
(302°F). The performance results are as shown in Table 3. HC1 removal is in
excess of 99% and PCDD (dioxin) and PCDF (furan) are lowered by 98%. Especially
significant in this example is the very law consumption of HSH, 15 kilograms per
ton of municipal waste, as contrasted with estimated consumption of 35-40
kilograms per ton of waste for alternative normal hydrate. Moreover, the
specified emission limit of less than 10 mg HC1 per normal cubic meter, newly
applicable on March 1, 1994, cannot be achieved with normal limp hydrate.
7-202
-------
COMERCIAL APPLICATION OF HSH PRINCIPALLY FOR S02 REMOVAL
A survey of the recent U.S. literature an the subject (8_ through 16)
that removal of S02 from utility boiler flue gas is influenced by the temperature
and the partial pressure of water vapor. If additional humidification is used,
the location at which the water is injected in relation to the point of hydrate
injection is important. Also, additives such as sodium hydroxide, if used in the
humidification water, improve S02 removal.
European coranercial performance experience for three S02 removal applications
using HSH is as follows:
S02 Removal Without Humidification
In 1987 the first dry sorption application for HSH began operation using a supply
of reagent from early pilot plant manufacturing facilities and providing dry S02
removed for two 7.5 nW(t) anthracite coal fired boilers at 170-185°C (338-36a°F)
flue gas temperature. At that time the required flue gas retention time for S02
capture was not known and therefore, a reactor vessel (5) was installed after the
HSH injection point. A simple vertical cylindrical design was provided as per
schematic Figure 3, with rotating wall scrapers to eliminate build-up on the inner
walls. Numerical performance data are shown in Table 4 under Case 1. Although
this type industrial fuel, unique to Germany, has very low fuel hydrogen content,
the supply in this case was high in surface moisture. Therefore, flue gas
hi modification was not necessary to gain targeted S02 removal. The system was
tested over a wide range of Ca/S solar ratios, comparing HSH with norma] hydrate.
In this plant S02 removal greater than 50* must be achieved, viiich, per Figure 4,
is possible with HSH use at a Ca/S ratio of approximately 1.7. Using normal
hydrate the Ca/S ratio to achieve this removal rate would, as per Figure 4, be
greater than 3.7.
Reduced Temperature Service When Firing Anthracite
In coranercial systems serving two 35 nW(t) anthracite coal fired boilers, as well
as a similar system for a 70 nW(t) unit in Poland presently under construction,
HSH along with humidification water is added as per Figure 3 via a venturi jet
feeder upstream of a reactor vessel (5). The German systems operate at an S02
removal efficiency of more than 50% to meet the mandated emission limit. This is
possible in this ccomercial operation by using HSH at a Ca/S solar ratio of 1.8
with an operating temperature of approximately 95°C (203°F). See the performance
description in Table 4 for Case 2 in this anthracite coal application with very
low raw gas water content. Note that even after humidification accompanied by a
7-203
-------
reduction in the gas temperature to 95°C (203°F) the water vapor content is very
much less than it Mould be for a bituminous coal application. The operator of
this system has found that lowering the temperature by further water addition to
less than 90°C (194°F) to achieve further humidification cooling allows removal
efficiency of more than 65* at the same Ca/S ratio of 1.8. This level of
humidification corresponds to an absolute humidity of less than 0.03 lb water
vapor per lb of dry gas, substantially less than unhumidified flue gas from
bituminous coal fired boilers, indicating that HSH will typically provide
comparable or better performance than this in US applications without need for use
of flue gas humidification means. Additionally, the plant operator has found that
the rotating scrapers in the reactor vessel need not be operated continuously.
High Sulfur Service .
As illustrated in Figure 5, HSH is used for high (90+*) efficiency desulfurizatian
of 250°C (482°F) flue gas of 6,000 mg/Nm3 (2,160 ppn) S02 strength issuing from a
tunnel kiln used to cure tar bonded magnesite bricks. This application is thus
directly comparable in flue gas S02 level to U.S. high-sulfur coal fired boiler
service. In the first of two HSH injection steps a small quantity of the reagent
is added to the raw flue gas prior to humidification to adsorb the small amount of
SOS present. After humidification cooling of the gas to a temperature of 68 to
80°C (154 to 176°F) the balance of HSH is injected to achieve S02 removal in a 2m
diameter by 8m high reactor vessel of special design. Provision is also made for
partial recycle to the lower part of the reactor of reaction product solids from
the downstream flue gas dedusting step. As indicated by a family of curves
entered in Figure 4, S02 removal efficiency is extremely sensitive to flue gas
temperature in the reactor. Note that at 68°C (154°F) and a Ca/S molar ratio of
3.0, 90* S02 removal is achieved, whereas above 75°C (167°F) the S02 removed is
less than 70*. The flue gas wet bulb temperature is approximately 54°C (129°F) and
thus, 90* removed operation via HSH at 68°C (154°F) requires 14°C (25°F) approach
to the wet bulb. The guaranteed stack S02 emission level of 500 mg/Nm3 (180 ppm)
is readily maintained at a stoichiometric ratio of 2.8. Severe peaks in raw gas
S02 strength are generated during periodic charging of the tunnel kiln, but the
gas cleaning system has been able to meet these conditions without difficulty.
Further, while very substantial gas residence time is provided by the reactor
vessel, it has been determined that only 2 to 3 seconds gas residence time is
needed to gain performance achieved. Waste residue collected in the fabric filter
(Figure 5) is amorphous calcium sulfite, which converts to calcium sulfate after
extended storage or under conditions of increased temperature.
7-204
-------
Optimized Point of Entry of Water When Required In S02 Service
Pilot test work has verified that the location of the flue gas humldification step
is of substantial importance in optimizing gas cleaning operation. Using a pilot
plant system as per Figure 6 with a flue gas flow rate of 350 normal cubic meters
per hour the location of water injection in relation to feed of normal lime or HSH
was tested using a synthetic gas containing 3,600 mg S02 per normal cubic meter
(1,300 ppm). Testing was focused only on gas cleaning performance of the reactor
proper with a flue gas retention time of 3 seconds and without taking into account
removal in the downstream fabric filter. The effect of changing this water
injection location is shown in Figure 7 data, wherein 35 kg/h water addition
corresponds to a water vapor content of 10.5%. Best results cure obtained by
adding one half of the H20 at each of locations 1 and 2 in Figure 6. If only half
of the water is added at either location 1 or 2, significantly higher removals are
achieved when the water addition is made downstream of the hydrate feed. This
leads to the conclusion, also given in diverse publications (6_, 12, and 13), that
for optimum performance the lime hydrate must be finely dispersed in the gas
before humidifying water is added.
CONTEMPLATED FUTURE USE OF HSH IN INTEGRATED GAS CLEANING SERVICE (7J
A tailored, enhanced, dry alkaline reagent such as HSH intermixed with lignite
coke lends itself advantageously to simultaneous removal of acid gases along with
chlorinated organics and/or trace heavy metals as may be appropriate for air
toxics emission reduction. One such application is cleaning of flue gas from
municipal solid waste incineration, which, effective March 1, 1994, must be
upgraded in Germany to meet emission limits as follows:
• HC1: 10 mg/Nm3 (6 ppm)
• S02 : 50 mg/Nm3 (18 ppm)
• Dicocin and furan: 0.1 ng/Nm3, tcocic equivalent (TE)
Since these new emissions limits are not typically achievable in the ccirnuonly used
spray dryer type gas cleaning systems, large-scale tests have been conducted over
extended periods to show the benefit and adequacy of system performance
augmentation by HSH. Results summarized in Table 5 indicate that a two stage
process as per Figure 8, with a spray dryer followed by dry injection of HSH,
meets the new emission limits while at the same time actually providing a
reduction in the overall amount of lime used. As indicated in Table 5, the
customary spray dryer system alone, without meeting the emission limits, consumed
lime reagent in the amount of 250 to 305 kilograms per hour of lime hydrate. With
7-205
-------
HSH injection incorporated as a second contact stage upstream of the dry
collector, only an overall total of 220 kilograms per hour of hydrate (milk of
lime hydrate plus HSH) were needed. HSH with 20* pulverized lignite coke was used
in these tests. The IE (tcocic equivalent) reduction of chlorinated organics from
6.3 nanograms per normal cubic meter in the feed gas to the extremely low value of
0.015 nanogram per normal cubic meter in the clean gas indicates an unnecessarily
large amount of coke use. Therefore, in further testing, the lignite coke
addition will be reduced to the level of 8-10*.
HIGH TEMPERATURE SORPTION
In a manner that parallels current major field developmental activity in the US
(10) aimed at economical, dry primary S02 capture via sorbent injection at the
boiler economizer, extensive pilot plant tests with good results have been
performed in Germany at 450°C (842°F). It has been verified that S02 removal is
directly proportional to hydrate surface area. Moreover, flue gas humidification
is not necessary because S02 is converted to the form, calcium sulfate, in this
mode. Apparently the crystal (chemically bonded) water of the lime hydrate plays
a contributing role in major enhancement of S02 sorption at such elevated
temperature. Commercial installations of this type have not yet been carried out.
Sub-scale tests have been performed in the U.S. at an S02 strength of 2,600 ppm
(7,200 mg/Nm3) and Ca/S of 2 leading to late 1991, proof of concept system
demonstration under US Department of Energy and Electric Power Research Institute
funding (10) using German-manufactured HSH. With injection at 1000°F (538°C) of
HSH having a surface area greater than 35m^/g, S02 removal of 60 to 65* can be
achieved without flue gas humidification. In conjunction with a modest degree of
flue gas humidification downstream of the air preheater for the purpose of
in?)roving ESP particulate removal (11J a further improvement in S02 removal is
accomplished.
USE OF HSH FOR RETROFITTING IN U.S. MEDIUM/HIGH SULFUR COAL SERVICE
Based on German connercial experience described herein, including high-sulfur
service, it can be seen that the high surface area of HSH permits its use in
highly diverse ways:
• In conjunction with duct injection at between 135 and 185°C (275 and
365°F) it siffords means to achieve 65* S02 removal, i.e. more than the
array of other dry injection technologies, with flue gas
humidification limited to that required to cool to 90°C (194°F) with a
7-206
-------
very substantial 50°C (70+°F) approach to wet bulb.
• It is expected to provide 65* S02 removal solely by econonizer
injection without flue gas humidification. This removal will be
augmented through a limited amount of flue gas humidification
downstream of the air preheater for the purpose of improving ESP
particulate removal efficiency.
• Further, as an alternative to lime spray dryer use, it may be
employed downstream of the air preheater in conjunction with flue gas
humidification to achieve 85-90* S02 removal.
These single-stage design alternatives for U.S. bituminous coal service with 2.6*
sulfur lelvel (apprax. 1,700 ppm: 4,700 mg/Nm3), the reference coal for EPRI PGD
cost program data (17), are displayed in Table 6.
Optimum use of HSH for high removal efficiency in such high-sulfur coeil service
may call for multi-staging, e.g. economizer injection at low Ca/S ratio augmented
by moderate humidification and residence time downstream of the air preheater.
Table 7 offers a trial comparison at the 300 trWe capacity level of such HSH system
design for 85-90* S02 removal (Case 5) against cost program data for wet PGD, lime
spray drying and an array of dry injection technologies previously evaluated by
EPRI (17).
Favorable economics of such optimal HSH use, projected from cocmercial operation
and testing in Germany, is the result of the unique match of its physical and
chemical properties with the diverse sorption regimes at the boiler exit.
Moreover, without great investment in gas cleaning equipment, HSH effectively and
advantageously utilizes available residence time and water vapor concentrations.
KEb'EKENCKS
1. M. Rostam-Abadi and D.L. Moran. Prospectus "High-Surface-Area Hydrated
Lime" Research at ISGS. Program Review. Illinois State Geological Survey.
Champaign, Illinois. August 1990.
2. F. Schwarzkopf; H.P. Hennecke and A. Roeder. "High Surface Area Hydrate -
WUELFRAsorp",- Cannercial Production and Application". ASIM-Symposium,
San Francisco, California, June 19th, 1990.
3. J.W. College and J. Vlnaty, Dravo Corporation, Pittsburgh. "Removal of
S02 from S02-containing gases" - US-Patent 4.626,418, December 2, 1986.
7-207
-------
4. K.D. Henning, H. Ruppert, K. Keldenich and K. Knoblauch. "Tmpraegnierte
Alctivkohlen zur Quecksilber-Qitfernung". WLB "Wasser. Luft und Betrieb.
p. 38—42.
5. H.F. Bach. T. Stumpf and A. Roeder. "Teilchen-Touch, Rauchgasreinigung durch
Trockensorption mit einejn Kalkhydrat hoher spezifischer Oberflaeche In einem
Spezialreaktor". Special print, Energie Spektrum, Heft 7, July 1987,
p. 1-4.
6. T. Huenlich, R. Jeschar and R. Scholz. "Sorptianskinstik von S02 aus
Verbrennungsabgasen bei niedrigen Temperaturen''. Zement-Kalk-Gips.
(44. Jrg) No. 5/1991, p. 228-237.
7. B. Morun, T. Sturcpf, P.-U. Schmidt, P. Strodt. "Die Nachruestung von •
Rauchgasreinigungsanlagen zur Einhaltung der 17. BTmSchV mit dem
modi fizierten, konditionierten Trockensorptionsverfahren (MKT-Verfahren)
am Beispiel der Spruehsorption einer Muellverbrennungsanlage in NFW".
Muell und Abfall, August 1991, p. 490-498.
8. R.E. Tischer. "Furnace And In-Duct Sorbent Injection". Acid Rain Retrofit
Seminar, National Lime Association, Philadelphia, Pennsylvania,
January 9-10, 1991.
9. D.J. Helfritch and S.J. Bortz. "Dry Sorbent Injection for Combined S02
and NQX Control". ASME Power Generation Conference, Boston, Massachusetts,
October 21-25, 1990.
10. D.J. Helfritch, S.J. Bortz, and R. Beittel. "Dry Sorbent Injection For
Combined S02 and NQX Control". Proceedings Seventh Annual Coal Preparation,
Utilisation, and Environmental Control Contractors Conference.
U.S. Department of Energy, Pittsburgh, Pennsylvania, July 15-18, 1991.
11. R. Beittel and J. Farr. "Cottrell/DOE Advanced Flue Gas Cleanup".
Report Riley Stoker Program No. 641-89801, November 14, 1990.
12. M. Babu, R.C. Forsythe. "Results of 1.0 NW BTU/Hour Testing And Plans For
A 5 W Pilot Halt Program For S02 Control". Report sponsored by the U.S.
Department of Energy through the Pittsburgh Energy Technology Center under
contract number DE-AC22-85PC81012.
13. H. Yoon et al. "Advanced Duct Sorbent Injection for S02 Control",
DOE Conference, Pittsburgh, Pennsylvania, July 15-18, 1991.
14. J. Butz et al. "Development of an Improved Nozzle For Flue Gas
Humdificatian", DOE Conference, Pittsburgh, Pennsylvania, July 15-18, 1991.
15. W.A. Walsh, Jr. "The Linear VGA Dry Sorbent Injection Mixer", July 1991.
16. M.G. Klett et cil. "Scaleup Tests and Supporting Research For The
Development Of Duct Injection Technology", DOE Conference, Pittsburgh,
Pennsylvania, July 15-18, 1991.
17. P. Radcliffe. "FGD Costs", ECS Update (EPRI), Fall 1990, No. 20. p. 10-12.
7-208
-------
Ground
quick'
Time
4J5 t/b
:-_j-
Water/alcohol
solution
metering
Process 0
water
1J t/h
V
Scale jc
3
Mixer
rrrfl
/
ytt ri
Reactor 1
Reactor 2
Scnibb«r
Alcohol tank
60 kg/h
Cooler
1AAA.
Hydrate
6 t/h
I
Rgure 1. Flowsheet of HSH (WUELfRAsorp®) production
Liquid
waste
1
t-, 700°C
nU
T>isxrt
HT—burner
<30 mg HCI/Nm3 Fobrjc
filter
c^Heot i
M
180°C>
P
HSH / Air
A
M
p=©
Residue
o
Figure 2. SWI with dry sorption only
Yenturi Water
Clean
gos
hO
Fabric filter
Figure 3. Coal—fired power plant with
dry sorption system
7-209
-------
100
90
80
7?
70
60
<
>
o
50
•>
LlJ
40
an
cT
30
to
20
10
0
I
Pin R im
6&*C
..••• ..-I
i"rSL
I
-- t *
-jo-
i
ucu
* *
6c
>
\
Tnal hydrate
1 2 3 *
STOICHIOMETRY [Co(OH)2/ S02]
Figure 4. SO^Ory—Sorption in Power Plant (Case 1; Tab. 4)
and Behind a Refractories—Tunnel—Kiln (Fig. 5)
Tunnel—Win, •*
gas c=l^_^250oC
6000 Nm3/h
6000 mg
S02/Nm3
HSH
HjO C$^-
~L
A
~nrv
Fabric filter
02m
68-80 °C
:z
3C>
200—500 mg
SOj/Nm3
HSH
"O Residue
app. 120 kg/h
-2*
Optional rec. residue
app. 100 kg/h
Solids discharge
Figure 5. Reactor for low temperature SO? removal
with HSH
Synthetic
gas
3600 mg SO,/Nm3
(1300 ppm)
r¥
SO. H.0 HYDRATE
2-3 kg/h
80 *C
Length of test duct: 22m
Dtamstar : 15cm
—I
T
350 NmJ/h
9
'Volume flow
S02measurement
by electrochemical cell
Figure 6. Schematic of pilot system
7-210
-------
HSH
-with 38 m2/g
Hydrate with 18 mz/g
STOICHIOMETRY [Ca(OH)2 / S02 J
Figure 7b. Test Conditions as In Fig. 7a
Comparison of Hydrates
I i
25 kg/h
10.53 moisture
50% position 1
50% position 2
12 kg/h position 2
6.5% moisture
I ^ I I
12 kg/h position 1
6.5% moisture
STOICHIOMETRY [Ca(0H)2/ S02]
Figure 7a. SO^-Removal. 80 C
Position of"H20 Injection
_ Milk of lime with
X 270 kg Ca(0H)yh
Raw gas with
1500 mg HCI/Nm3
500 mg SOj/Nm3
45000 NmVh
¦ Milk of lime with
1 165 kg Ca(0H),/h
Spray
dryer
ESP
» HSH
¦ 19—2+ kg/h
System 1
€)
System 2
Figure 8. Scheme of a MWI working w'rth sproy dryer only (System 1)
and with the combination spray dryer/dry sorption (System 2)
7-211
-------
Table 1
DftMrrc m? rvMynTcrT&r rpui Tf&'i'ifw m? ulhj
AnkTMCi Ut VAaIHCJw/^Au Ab* LJLwAJI xUN v6 jlaJw i>
231 CLEANING OF WASTE GASES
Rc&iilc
1
5
6
7
8
Service
SWI (special waste
incinerator)
WI (municipal waste
incinerator)
Aluminum industry
(Recycling)
Ceramic industry/
refractories
Industrial furnace
Power plait
Burning of sewage sludge
Thermal treatment
of contaminated soil
Glass industry
Pollutants
HCl/S02/PCDD/PCDF/Eg
HCl/S02/PCDD/PCDF/Hg
HCl/PCDD/PCDF
S02
S03/S02/HC1
S02
HC1
S02/Hg
S02
With/Without
Coke Additive
Without/
With
Without/
With
Without/
With
Without
Without
Without
Without
Without/
With
Without
Note; Total amount of reagent product sold is app. 10,000 metric t/year.
7-212
-------
Table 2
SPECIAL WASTE INCINERATION WITH DRY SORPTION
FUEL FIRED: 1.6 m3/h HIGHLY CHLORINATED. ORGANIC-LIQUID COMPOUNDS
VOLUME OF GAS: App. 12,000 Nm3/h
DOSING RATE, NORMAL HSH: App. 200 kg/h
HATER VAPOR (MOISTURE): App. 25%
HC1 Concentration S02 Concentration
(mq/Nm3) (mg/Tta3)
Raw gas 4,000 to 10,000 50 to 3,000
Clean gas <30 <60
Removal Efficiency
HC1 S02
99.2 - 99.7 86 - 97
Stoichiometric Ratio
(to Acid Compounds)
App. 3.0
7-213
-------
Table 3
MUNICIPAL. WASTE INCINERATION WITH DRY SORPTION
Capacity
Volume of gas
Temperature before air cooler
HSH dosing rate before cooler
Temperature after cooler
Ligni te-coke-containing
HSH dosing rate sifter cooler
Raw gas pollutants
(Preliminary information)
Clean gas pollutants
Removal efficiency
Stoichiometric ratio, molar, (Ca/HCl)
or
3 x 5 t waste/h
3 x 26,000 Nm3
App. 250°C (482°F)
15 kg/h
App. 150°C (302°F)
60 kg/h
App. 1,000 mg HCl/Nm3
n.d. S02*
App. 56-82 ng TE/Nm3 PCDD/-DF
App. 10 mg HCl/Nm3
n.d. S02
App. 1 ng TE/Nm3
99% HC1
n.d. S02
App. 98.2-98.8* PCDD/-DF
2.6
App. 15 kg HSH/t waste
* n.d. = not determined
7-214
-------
T&ble 4
COAL-FIRED POWER PLANTS WITH DRY SORPTION SYSTEMS
Flue gas volume Ita3/h
Temperature after °C
heat exchanger
Flue gas water vapor
content without
water Injection %
Water for humidi-
ficatian m3/h
Case 1:
2 x 7.5 nW(t) Units
Each 10,400
170-185
(338—365°F)
App. 6*
Case 2:
2 x 35 aW(t) Units
Each 22,000
135-140
(275-284°F)
App. 2*
None
1.2
Water vapor content
after water injection *
Temperature after °C
water injection
Pollutant concen-
tration in raw gas
Pollutant concen-
tration in clean gas mg/Nm
S02 removal
efficiency %
Stoichiometric ratio
Unchanged
Unchanged
mg/Nm S02:1,350 (486 ppn)
See Figure 4
See Figure 4
See Figure 4
App. 6.5
App. 95
(203°F)
S02:1,800 (648 ppn)
HC1: 400 (252 ppn)
S02: 800 (288 ppn)
56**
App 1.8
* Unique fueling with anthracite coal (low coal hydrogen content)
** As per systen operation to achieve 50% S02 removal. With humidification
cooling of gas below 90°C (194°F), > 65% S02 removal is obtained.
7-215
-------
Table 5
COMPARISON OF POLLUTANT CXWCEOTRATIONS WITH SYSTOl 1 (SPRAY DRYER)
AND SYSTai 2 (COMBINATION SPRAY DRYER/DRY SORPTION)
HC1 (ma/Nm3)
Raw Clean
SQ2 (ma/Nm3)
Raw Clean
TE (ng/Nm)
Raw Clean
Total Consump-
tion of Milk
nf T.itnt» + HSH
(ko/h)
System 1
(Spray Dryer)
Entire test
period
Single day
1,160
1,363
29
29
387
353
100
96
n.d.
n.d.
n.d.
n.d.
250
305
System 2
(Spray Dryer w/
Downstream HSH/
20% Coke Inj.)
Qitire test
period 1,093 6 250 42 n.d. n.d. 220
Single day 1,416 8 195 43 6.3 0.015 220
Note: Limit values to be met by existing system according to 17. BImSchV
by March 1, 1994 are:
HC1 < 10 mg/lta3
S02 < 50 mg/Ttor
PCDD and PCDF (as TE, toxic equivalent, factored) < 0.1 ng/Nm3
n.d. = not determined
7-216
-------
Table 6
SINGLE STAGE APPLICATION MEANS FOR HSH IN
2.6* SULFUR BntMENOUS COAL SERVICE
Injection
temperature, °C
Operating
tanperature, °C
Ca/S ratio
Humidification
Approach to
gas wet bulb, °C
S02 removal
efficiency, *
Major capital
equipment
Duct Injection,
Medium Efficiency
135-185
(275-365°F)
App. 90 (194°F)
1.8
Cool to
to 90°C (194°F)"
50 (70+°F)
60
None
Economizer Injection,
Medium Efficiency
538 (1000°F)
Below injection
temperature
2.0
None. Osed to
augment ESP per-
formance.
No humidificatian
65
None
Duct Injection,
High Efficiency
135-185
(275-365°F)
68-72
{154—162°F)
App. 3.0
Yes
14-17
(25-31°F)
85-90
< 3 seconds gas
residence time*
* Estimated flue gas residence time after HSH injection and gas humidificatian
" Possibly a higher operating temperature will be adequate in bituminous coal
service in the U.S. due to high water vapor content of raw flue gas.
7-217
-------
Table 7
300 t*E CAPITAL REQUIREMENT AND TOTAL LEVELIZED COST (30 YEARS)
FOR MODERATELY DIFFICULT, HIGH (2.6*) StJLFUR RETROFIT FOR SQ2 RH43VAL
Case
1. Wet FGD
2. Spray Dryer (Semi-Dry)
3. Dry Injection
Technologies (U.S.)
4. Dry Injection
Technologies (FRG; HSH)
5. Multi-stage (FRG: HSH)
(Preliminary Estimate)
S02 Removal
(*)
90
90
4O-60
App. 60
85-90
Capital Cost
(S/fcW)
150-280
140-210
70-120
75
<120
Total Control Cost
(S/tcn S02)
350-600
360-540
420-750
550
<400
Note: Costs are expressed in constant 1990 dollars.
7-218
-------
INITIAL OPERATING EXPERIENCE
OF THE SNOX PROCESS
David J. Collins, P.E.
ABB Environmental Systems
31 Inverness Center Parkway
Birmingham, Alabama 35243
Riziero Ricci
Snamprogetti, S.p.A.
20097 S. Oonato Milanese
Italy
Christian H. Speth
Haldor Topsee A/S
Nymallevej 55
DK-2800 Lyngby
Denmark
Rita E. Bolli
Ohio Edison Company
76 South Main Street
Akron, Ohio 44308
7-219
-------
Intentionally Blank Page
7-220
-------
ABSTRACT
The SNOX Process catalytically removes greater than 95% of the S02 and the N0X from
utility coal-fired boiler flue gases and, with integration of recovered heat, has
significantly lower O&M costs than conventional technologies. Two SNOX Process
units have recently been commissioned in Europe - a 30 MW plant in Gela, Italy and a
305 HW plant in Vodskov, Denmark. Also, a 35 MW demonstration unit will startup in
the United States late this year.
The SNOX Process is a highly efficient pollution control technology which produces a
salable concentrated sulfuric acid byproduct and no waste streams, liquid or solid.
With integration into the furnace block, the SNOX Process can supply heated
combustion air which more than compensates for the energy requirements of the air
pollution control equipment. Ammonia is the only reagent required.
This paper describes the SNOX Process and presents the startup experiences and
initial operating data for two SNOX applications in Europe. The soon-to-be
commissioned U.S. demonstration of the process is also discussed.
7-221
-------
INTRODUCTION
The current global emphasis on environmental protection and pollution abatement has
fostered the development of several new technologies and processes focused on
pollution control in the power production industry. Many of these new technologies
and processes represent significant improvements over those currently employed.
These improvements are in the areas of increased pollutant removal efficiencies,
reduced reagent requirements, reduced waste streams, and reduced operating costs.
The U.S. Department of Energy (DOE), through its Clean Coal Technology Program
(CCTP), is presently providing funds for the development and demonstration of the
most promising of these new processes. One such process is a catalytic de-NOyde-
S0x process developed in Denmark by Haldor Topsee A/S and offered under license ir
North America by Asea Brown Boveri (ABB) Environmental Systems.
This highly efficient technology, known as the SNOX Process, has successfully been
demonstrated at a 6,200 SCFM (10,000 Nm3/h) pilot scale facility in Sloerbsk, Denmark
for over 20,000 hours. The exceptional performance of this pilot facility fostered
predictions of the successful application of this process to the utility power
production market and has provided the design data necessary for the full-scale
units which have come on-line this year. One of these units treats a 30 MW
equivalent of flue gas from a petroleum coke-fired furnace in Gela, Italy. The Gela
SNOX Plant was commissioned in April. The performance of this unit has exceeded its
design targets for sulfur dioxide (S02), nitrogen oxide (N0X) and particulate
removal. The sulfuric acid produced by this unit is used for fertilizer production
in an adjacent plant.
In August of this year a much larger SNOX installation went on-line in Vodskov,
Denmark to treat the flue gas from a 300 MW coal-fired power plant. While operating
data from this unit is just becoming available, initial indications are that the
SNOX Process has again exceeded its design performance targets. This Plant
represents the final step in introducing the SNOX Process to the utility power
generation market.
7-222
-------
Late this year the SNOX Process will be demonstrated in the United States as part of
the DOE CCTP on high sulfur Ohio coal at the Niles Plant of the Ohio Edison
Company. This SNOX facility will treat a 35 MW equivalent flue gas slipstream and
target 95% S02 removal, 90% N0x removal, and less than 0.0004 gr/SCF (1 mg/Nm3) of
particulate emissions. Operating data from this plant will be available in early
1992.
THE SNOX PROCESS
The SNOX Process consists of five key areas; particulate collection, N0x reduction,
S02 oxidation, sulfuric acid (H2S04) condensation and acid conditioning. The
integration of these individual steps is shown in Figure 1 which is a process flow
schematic for a typical full-scale application of the SNOX Process.
To briefly discuss Figure 1, flue gas leaving the air preheater is treated in a
particulate control device and passed through the cold side of a gas/gas heat
exchanger (GGH) which raises the gas temperature to above 700*F (370*C). An ammonia
and air mixture is then added to the gas prior to the selective catalytic reactor
(SCR) where nitrogen oxides are reduced to free nitrogen and water. The flue gas
leaves the SCR, its temperature is adjusted slightly, and enters the S02 converter
which oxidizes S0? to sulfur trioxide (S03). The S03 laden gas is passed through
the hot side of the GGH where it is cooled as the incoming flue gas is heated. The
processed flue gas then enters a falling film condenser (the WSA-Condenser) where it
is further cooled with ambient air to below the sulfuric acid dewpoint. Acid
condenses out of the gas phase on borosilicate glass tubes and is subsequently
collected, cooled, and stored. Cooling air leaves the WSA-Condenser at over 390*F
(200*C) and is used for furnace combustion air after collecting more heat through
the air preheater.
Particulate Collection
The degree of particulate collection upstream of the S02 Converter has a significant
effect on the operating costs of this process. This correlation is due to the
inherent characteristic of the S02 oxidation catalyst to collect and retain greater
than 95% of all particulate matter entrained in the flue gas. The collection of
this particulate matter, over time, increases the pressure drop across the S02
Converter. The initial pressure drop can, however, be restored through on-line
catalyst screening (described later). Higher dust loads therefore require more
frequent catalyst screening which provides an incentive to utilize a high efficiency
particulate collector upstream of the SNOX Process area. A target dust level of
7-223
-------
0.0004 gr/SCF (1 mg/Nm3) leaving the collector is used for design. A fabric filter
with GoreTex® membrane bags has been demonstrated to achieve this very low emission
level. As a consequence of both the high efficiency dust collector and the dust
retention characteristics of the S02 Converter, particulate emissions from the
system are significantly less than 0.0004 gr/SCF (1 mg/Nm3) which is far below any
current regulations or standards.
It should be noted that while a high efficiency particulate collector has benefits
related to system operating costs, the economics do not require it exclusively. The
SN0X plant currently operating in Italy, for instance, uses an electrostatic
precipitator (ESP) and is expected to require S02 Converter catalyst screening at
two week intervals. For comparison, the Niles Demonstration Project with a fabric
filter and GoreTex® bags is expected to require screening only once every year.
Nitrogen Oxide Reduction
After the particulate matter is collected and the temperature of the flue gas is
increased to over 700°F (370"C) through the GGH, an ammonia (NH3) and air mixture is
introduced to the gas stream via a nozzle grid located upstream of the SCR. A
slipstream of hot air from the WSA-Condenser is used to evaporate and dilute a
metered mass of ammonia. The resultant mixture is agitated with a static mixer and
supplied to the nozzle grid. The design of the nozzle grid allows for controlled
stoichiometric ratios of NH3 to N0x on a localized scale over the cross-section of
the SCR inlet duct. This is critical in order to optimize system NOx removal
efficiency since any unreacted NH3 which "slips" across the SCR will be oxidized in
the S02 Converter downstream to water (H20), nitrogen (N2), and NOx.
The flue gas/NH3 mixture enters the SCR and contacts the Hal dor Topsee DNX
monolithic catalyst which has been demonstrated to reduce 97%+ of the entering N0x
to N2 and H20. The reduction of nitrogen oxide (NO) follows Equation 1.
NO + NH3 + 0.25 02 - «2 + 1.5 Hz0 + 5,880 Btu/lb NO (13.7 HJ/kg NO) (1)
7-224
-------
The small amount of nitrogen dioxide (N02) present in the flue gas is reduced
similarly. The general arrangement of the SNOX Process equipment offers one
significant advantage over other SCR technologies in that those processes are
limited to NHj/NC^ molar ratios of less than 1.0. This is necessary to limit the
NH3 "slip" past the SCR to 5 ppm or less. Higher levels may result in ammonium
sulfate or bisulfate scaling in lower temperature areas downstream. The N0x removal
efficiency of these processes is thus limited. Any NH3 slip in the SNOX Process,
however, is oxidized as it contacts the S02 Converter catalyst downstream. This
allows stoichiometric ratios in excess of 1.0 and consequently higher N0x removal
efficiencies without adverse downstream effects. Ammonium "salting" does not occur
in the duct between the SCR and the S02 Converter because the temperature is well
above the dewpoints of aranonium sulfate and ammonium bisulfate. Excess NH3
slippage, however, must still be minimized in order to maximize system N0x removal
due to the N0x forming oxidation of NH3 in the S02 Converter.
Sulfur Dioxide Oxidation
The SCR effluent is heated slightly with natural gas, oil or steam to reach the
optimum S02 Converter inlet temperature of 770*F (410*C) and passed through beds of
Haldor Topsee sulfuric acid catalyst. This catalyst has seen wide use in the U.S.
sulfuric acid industry for the past decade with a high degree of success. The
efficiency of the Topsee catalyst is not affected by the presence of water vapor or
chlorides in concentrations up to 50% and several hundred ppm, respectively.
Without any reagents or additives, over 95% of the entering S02 is oxidized via
Equation 2.
S02 + 0.5 02 - S03 + 660 Btu/lb S0Z (1.5 MJ/kg S02) (2)
Due to surface fouling by flyash, the oxidation catalyst requires screening to
maintain a minimum converter pressure drop. The screening frequency is dependent on
the removal efficiency of the particulate collection device upstream. The required
screening frequency will range from once °very two weeks to once a year. Regardless
of the efficiency of the particulate collector, however, virtually all remaining
particulate is retained in the S02 Converter which results in the inherently minimal
particulate emissions of this process.
The screening procedure consists of the isolation of an individual catalyst bed,
removal and mechanical screening of the catalyst in that bed, and refilling the bed
7-225
-------
with the screened catalyst. Beds are sequentially screened in this manner. The
procedure can be automated and is performed while the process is on-line at full
load. This procedure separates and removes virtually all flyash and other
contaminants from the surface of the catalyst pellets and restores S02 oxidation
efficiency. Catalyst loss during screening is estimated at 2-3%.
An additional benefit of the sulfuric acid catalyst is that it oxidizes most of the
carbon monoxide (CO) and hydrocarbons present in the flue gas stream to carbon
dioxide and water. This is of importance since emissions of both these compounds
have come under increased scrutiny as evidenced by the recent legislative actions.
Sulfuric Acid Condensation
The S03 in the gas leaving the S02 Converter is hydrated and condensed in two steps.
First, the bulk of the S03 is hydrated to sulfuric acid vapor (Equation 3) as the
flue gas passes through the hot side of the GGH and the temperature drops
approximately 300°F (165"C).
S03 + H20 - H2S06{vapor) + 680 Btu/lb S02 (1.6 HJ/kg S02) (3)
At this point the flue gas is still well above the acid dewpoint thus avoiding acid
condensation and corrosion of the ductwork. The flue gas, at around 500*F (260"C),
is passed through the proprietary WSA-Condenser developed by Haldor Topsee. The
WSA-Condenser is a unique tube and shell falling film condenser with ambient air
used as the cooling medium on the shell side. Proprietary borosilicate glass tubes
are used to convey and cool the flue gas. There are several design features of
these tubes which make possible the virtually complete condensation and capture of
sulfuric acid at concentrations of 94 to 97 wt.%. The precipitation of the sulfuric
acid from the gaseous phase to the liquid phase is also an exothermic reaction which
releases 460 Btu/lb S02 recovered as H2S04 (1.1 MJ/kg S02). The flue gas is cooled
to about 210*F (100'C) at the outlet of the condenser which, combined with the
presence of approximately 5 ppm of uncollected sulfuric acid mist, will require the
downstream ductwork and stack to be lined. The condensed sulfuric acid product is
funnelled through an acid brick lined trough at the bottom of the WSA-Condenser into
the acid conditioning and storage system.
The WSA-Condenser's discharge cooling air represents the only other by-product of
the SNOX Process. In an integrated system, the bulk of this heated ambient air at
7-226
-------
about 400*F (200*C) will be passed through the furnace air preheater and used as
combustion air. A small percentage of the product air is used for system
auxiliaries such as ammonia evaporation and dilution, support burner combustion air,
and coal milling. The WSA-Condenser, in effect, collects the heat released from the
reactions in the SCR and S02 Converter, the hydration of S03, the condensation of
H2S04, the support burner, the booster fan compressiof' and the overall decrease in
flue gas temperature. This results in a considerab.c «_«ount of thermal energy which
can be easily utilized as preheated furnace combustion air to increase boiler
efficiency. Steam production can increase on the order of 1% per each percent of
sulfur in the fuel. At 2-3% sulfur, the SNOX Process energy requirements are
compensated. These heat recovery characteristics make the use of h.gh sulfur, low
cost fuels economically attractive with this process.
Acid Conditioning and Storage Systems
The hot concentrated sulfuric acid product at about 40CTF (200°C) is collected and
circulated through a fluoropolymer lined system consisting of a holding tank,
circulation pump:, and a water cooled heat exchanger. The purpose of this loop is
to cool the acid to more manageable temperatures (70-100*F or 20-40'C) and to allow
for dilution of the acid to a specified concentration, normally 93.2 wt.%. As acid
collects in the conditioning loop, it is metered off at a rate which maintains a
constant level in the holding tank. This product acid is stored in carbon steel
tanks prior to removal by tanker truck. It should be noted that the sulfuric acid
produced by the SNOX Process meets or exceeds U.S. Federal Specification 0-S-801E
Class I and is expected to be commercially tradeable without limitation.
Pilot Results
The performance characteristics of the SNOX Process as seen at the Haldor Topsae
3.5 MW equivalent pilot facility in Sksrbsk, Denmark over the past three years are
summarized below:
• FABRIC FILTER - less than 0.0004 gr/SCF (1 mg/Nm3) of particulate
leaving the filter which operates at about 4 in w.c. (10 mbar) of
pressure drop and an air to cloth ratio of 3.3 fpm (60 ra/h).
• SCR - 98% N0X removal with 400 ppm inlet NO , an NHj/NO^ ratio of 1.02
and a space velocity equal to that of a ful*l-scale plant. Only
minimal deactivation of the SCR catalyst has been observed.
• S02 CONVERTER - 97% and 95% S02 conversion at inlet S02
concentrations of 500 and 1,700 ppm, respectively, at 20% higher
7-227
-------
than full-scale space velocities. The S02 catalyst has been
screened successfully several times and only minimal
deactivation of the catalyst has been observed.
• WSA-CONDENSER - high S03 capture efficiency with an acid mist
carryover of 5 ppm normally and never greater than 10 ppm even
through an operating range of 30 to 120% of design gas flow. The
sulfuric acid by-product is consistently 95 wt.% with water clarity
and no heavy metals present at greater than 0.5 ppm.
It should also be noted that all design heat and material balances have been
substantiated and verified from Skaerbek data.
A 30 HW PETROLEUM COKE APPLICATION
The Gela Power Plant is owned by ENICHEM S.p.A. in Gela, Italy and contains a BREDA
type boiler that supplies 380 t/h of steam at 1550 PSIA to a turbine and the
adjacent petrochemical complex. The boiler was commissioned in 1960. The total
flue gas from the boiler is about 250,000 SCFM (400,000 Nm3/h), of which 25% is
treated in the SN0X Plant. The normal fuel supply is pet-coke with a sulfur content
of 6-7 wt.%. The boiler can also burn fuel-oil of 3.5 wt.% sulfur.
The Gela SNOX Plant
The Gela Plant represents the first industrial application of the SNOX technology
according to the so-called "hot scheme" as shown in Figure 2. The key aspects of
this type of application are as follows:
1. Flue gas is taken upstream of the boiler air preheater.
2. Particulate collection is accomplished with a hot side ESP at
730 to 750*F (390-400'C).
3. A gas/air heat exchanger is used instead of a gas/gas heat
exchanger for flue gas cooling prior to the WSA-Condenser.
4. Atmospheric air used to condense the product sulfuric acid is
heated to 645-660*F (340-350*C) and sent directly to the furnace
burners.
5. The flue gas fan is installed downstream of the WSA-Condenser,
thus all process equipment operates under negative pressure.
The design specifics of each major piece of process equipment are discussed below.
7-228
-------
Electrostatic Precipitator. As stated before, the ESP at the Gel a Plant is of the
hot-side design. Its design inlet flue gas conditions are:
• Flowrate 62,400 SCFM 100,000 Nm3/h
• Temperature 790*F 420*C
• Dust load (pet-coke) 2.9 gr/SCF 7.0 g/Nm3
Dust load (fuel oil) 0.5 gr/SCF 1.3 g/Nm"
,3
This ESP normally operates with five fields in service. The dimensions (LxWxH) of
the casing are 72x33x46 ft (22x10x14 m). This precipitator is designed to limit
outlet dust emissions to .004 gr/SCF (10 mg/Nm3) for all operating conditions.
SCR Reactor. Flue gas enters the reactor from the top and exits via the bottom
passing through two catalyst layers. Inside the reactor there are actually four
catalyst support grids that make it possible to change the arrangement of the layers
for experimental purposes. The reactor operates near 690*F (365*C) and utilizes a
type of the Topsee DNX catalyst known as "medium dust."
SO. Converter. The S0? Converter for this plant consists of six catalyst panels
built into one casing. The flue gas leaving the reactor is conveyed through three
ducts, each provided with a damper which allows the catalyst screening without
interrupting plant load. The catalyst cleaning is done by means of an automated
system. This converter also operates near 790*F (420*C) using Topsoe sulfuric acid
catalyst.
MSA-Condenser. In the WSA-Condenser the flue gas is cooled to about 210*F (100*C)
with air from the existing boiler air fan. This condenser contains 5800 glass
condensing tubes, each about 20 ft (6 m) long. The WSA-Condenser's casing is
constructed of carbon steel and has dimensions (LxWxH) of 23x23x20 ft (7x7x6 m).
The acid collecting bottom of the condenser is lined with a PTFE membrane and acid
brick.
Demonstrated Performance
The main objective of the Gel a Plant's test program is to demonstrate the
application of the "hot scheme" SNOX Process on a full-scale power station. During
the two-year test program, a parametric study of different operating conditions will
be performed on various fuels. The information and data collected will be mainly
used to identify the factors which affect process reliability.
7-229
-------
During the first operating period (March 1991-July 1991), the plant was tested at
the following conditions:
Fuel
Flue gas flowrate
Cooling air flowrate
Hot air to boiler
Process temperatures:
Inlet ESP
Inlet SCR
Inlet S02 Converter
Outlet WSA-Condenser
Hot air to boiler
Sulfuric acid production
Sulfuric acid concentration
Pet-coke
50,000 SCFM
62,000 SCFM
53,000 SCFM
735*F
690 *F
790 *F
225*F
645"F
Pet-coke
80,000 Nm /h
100,000 Nm3/h
85,000 Nm3/h
390*C
365"C
420*C
106*C
340*C
1800-2400 Ib/h 800-1100 kg/h
> 94 wt.%
> 94 wt.%
Dust Removal. The ESP performance has been evaluated by isokinetically measuring
outlet dust conditions and values less than .002 gr/SCF (5 mg/Nm3) have been found.
No inlet loadings have been measured thus far. The minimal pressure drop increase
across the S02 Converter confirms these very low outlet measurements.
N0^ Removal. The system N0x removal, calculated from process instrumentation, was
greater than 96% with SCR inlet and WSA-Condenser outlet N0x concentrations of 350
and 9 ppm, respectively. This data was collected at an NH3/N0x ratio of 1.0.
SO- Conversion. The system S02 conversion, calculated from the plant's Continuous
Emissions Monitoring System (CEMS) and confirmed by lab analyses, was greater than
96% throughout the operating period. The plant inlet S02 concentration varies
widely between 2400 and 4000 ppm depending on fuel sulfur content. Even through
this wide range of inlet sulfur loadings, outlet emissions are maintained between 70
and 140 ppm - greater than 96% S02 removal. The wide range of S02 inlet
concentrations caused no operating problems with the SN0X Plant.
Sulfuric Acid Condensation. Apart from a light turbidity for the first few days of
start-up, the sulfuric acid product has been perfectly colorless. The acid
concentration has remained between 94 and 94.5 wt.%. Based on lab analyses, the
7-230
-------
sulfuric acid is of good industrial quality with a very low heavy metals content and
is normally used inside the Gela petrochemical complex for fertilizer production.
The thermal exchange characteristics of the WSA-Condenser have been in accordance
with expected values.
Initial Operating Experience
The main problems during the first operating period of the test have not been
directly connected with the SNOX Process. For instance, during the final ESP
commissioning phase, it was necessary to replace certain electrical insulators.
This changeout resulted in a corresponding delay in the expected plant start-up
date.
A persistent mechanical problem, which has affected plant availability, has been
improper balance of the flue gas fan. This problem is being investigated by the fan
vendor.
It should be noted that the SNOX Plant start-up and shut-down procedures have not
caused any problems with the normal boiler operation. Also, SNOX Plant operation is
fully automatic.
A 305 HW COAL-FIRED APPLICATION
The power station of Vendsyssel is owned by I/S NEFO, the North Jutland Electricity
Supply Company, and is the northernmost power station within the ELSAM (Jutland-
Funen power pool) cooperation. This power station consists of 2 coal/oil-fired
units of 137 MW and 305 MW, respectively. The SNOX Plant is to clean 100% of the
flue gas from Unit 2, the larger, which is of the Benson type with a nominal steam
production of 910 t/h (235 kg/s) at 2900 PSIA and 1000*F.
The NEFO SNOX Plant
The equipment supplied at the NEFO Plant is shown as a simplified process flow
schematic in Figure 3. A discussion of each major piece of equipment is provided
below.
Fabric Filter. The fabric filter is arranged in sections with common inlet and
outlet manifolds. Thus, each section, or compartment, can be isolated and
maintained during full operation. Cleaning of the bags is done according to the
reverse air principle. The filter material was supplied by Gore-Tex® and consists
of a glass fiber backing laminate coated with a Teflon (PTFE) membrane. The
7-231
-------
filtration itself is accomplished by the so-called membrane filtration technique
which allows a maximum dust detention simultaneously with high filtration speed and
low pressure loss.
Start-Up Burner. For preheating the SNOX Plant to operating temperature before
treating flue gas, two light oil-fired start-up burners heat a partial flow of the
cooling air leaving the WSA-Condenser. The preheated air is introduced to the
system prior to the flue gas blowers. Preheating the SNOX Plant from a cold
condition requires six to eight hours.
Flue Gas Blowers. In order to overcome the pressure loss across the SNOX Plant, two
50% axial fans are provided - each with a motor rating of approximately 4 MW.
Gas/Gas Preheater. For heating the flue gas after the flue gas fans, a rotating
regenerative gas/gas preheater is installed. The preheater is designed to heat the
SCR inlet flue gas to approximately 725*F (385*C) while simultaneous cooling of the
stream exiting the S02 Converter. In order to reduce the temperature differences
which normally occur across the cross section of a rotating gas/gas preheater, the
speed of the rotor is increased from 0.6 rpm to 1.2 rpm in the high load area.
The thermal efficiency of this preheater is designed at 85% with a maximum leakage
of 2% at full load. The heating surface area totals approximately 915,000 ft2
(85,000 m2). This preheater, which follows the Ljungstrom principle, is provided
with a leakage minimizing system in order to maximize system S02 and N0X removals.
Any leakage across the preheater, in effect, bypasses the catalytic reactors and
results in a decrease in system removals.
SCR Reactor. The SCR reactor contains two layers or beds of catalyst and one blank
bed grid. When the catalyst activity has degraded to a certain limit, a third layer
is added. The catalyst is built up into box-shaped modules, each module consisting
of six cassettes. The catalyst design allows the passage of any dust particles.
For this reason the SCR reactor does not require cleaning equipment such as steam
lances. The construction of the reactor allows for easy exchange of the catalyst
modules by means of equipment specially designed for this purpose. A plant outage
is, however, required.
Steam/Gas Preheater. Having passed through the SCR reactor, the flue gas is heated
in order to obtain the reaction temperature necessary in the S02 Converter. This
heating is done by a steam/gas preheater which uses superheated steam taken after
7-232
-------
the second high pressure superheater of the boiler. Having been cooled by
approximately 210*F (100'C), the steam is returned to the boiler circuit prior to
the third high pressure superheater. The heat exchanger is designed as a finned
tube heat exchanger with a total heating surface area of approximately 44,000 ft2
(4,100 m2).
SO- Converter. The S02 Converter consists of 36 catalyst panels built into one
casing. The lateral faces of the panels are perforated allowing flue gas flow
horizontally through the panels from a central inlet chamber to the outlet chambers.
The panels can be isolated from the flue gas flow during the filling and discharge
of catalyst. The catalyst is a vanadium-containing oxidation type formed as rings.
Catalyst addition is performed through top mounted feed chutes on each panel.
Discharge is done at the bottom of the beds through discharge chutes.
Unless major leakage occurs in the bag filter, the catalyst is expected to be
cleaned only once a year - at the time of the annual outage. The cleaning of the
catalyst material in the S02 Converter is executed with a mobile screen system. The
screen system is connected to each panel sequentially after isolation from the flue
gas flow. The cleaned catalyst material flows to a heated container. The container
is then moved to the top of the panel where the catalyst material is deposited into
the panel.
WSA-Condenser. The WSA-Condenser consists of six sections, each containing twelve
modules. Each module consists of a bundle of vertically mounted glass tubes. As
the flue gas is cooled to about 210*F (100'C) in the WSA-Condenser, the sulfuric
acid vapor from the S02 Converter is condensed and concentrated on the glass tubes.
The condensed sulfuric acid is collected in lined vessels placed under the glass
tubes and cooled in an acid cooling circuit with a sea-water cooled heat exchanger.
Having been cooled to approximately 100*F (40*C), the sulfuric acid is conveyed to
the acid storage tanks by means of conveying pumps. These steel storage tanks
contain up to 3,750 tons of sulfuric acid, corresponding to approximately six weeks
of full load coal-fired operation with 1.6% sulfur fuel.
The cooling air to the WSA-Condenser is delivered by two existing secondary air
fans. Having been heated in the WSA-Condenser, the discharge cooling air passes
through a trim cooler and then to the existing air preheater to be used as furnace
7-233
-------
combustion air. The trim cooler is cooled by condensate taken after the first low
pressure preheater and returned to the feed tank. This brings about a gross
increase of the efficiency of the turbine generator of approximately 2.2 MW at 100%
load.
Chimney. At the WSA-Condenser outlet the flue gas temperature will be 190-230*F
(90-110*C) dependent on the ambient temperature. The flue gas will contain small
quantities (5-10 ppm) of sulfuric acid mist. Therefore, the flue gas duct after the
WSA-Condenser is provided with an acid-resistant coating and the new SNOX chimney is
provided with a chimney pipe constructed of acid-resistant brick. Technical and
economic estimates showed that the cheapest alternative was to build a new SNOX
chimney since the existing chimney would, among other things, have required the
application of a protective coating. Now, the pre-existing chimney is used only
during bypass operation and start-up of the SNOX Plant.
Consumption and Production Data
With a coal sulfur content of 1.6%, the full load consumption and production data
for this SNOX Plant will be as follows:
• Flue gas loading, SCFM (Nm3/h) 562,000 (900,000)
• Sulfur dioxide removal, % 93+
• Nitrogen oxide removal, % 90+
• Water consumption, gpm 0
• NH3 consumption, t/h 0.38
• Acid production (95% H2S04), t/h 5.2
• Reduced efficiency of power plant, % 0.2
THE U.S. SNOX DEMONSTRATION PROJECT
The U.S. SNOX Demonstration Project will be located at the Ohio Edison Niles Power
Plant near Niles, Ohio in Trumbull County. The plant is situated on 130 acres alone
the southern bank of the Mahoning River. The Niles Plant is part of the Ohio Edisor
System with over 6000 MW of capacity serving central and northeastern Ohio and
western Pennsylvania. This project is one of 10 clean coal projects Ohio Edison
currently has underway or has recently completed.
Ohio is a high sulfur coal state and clean coal research and development is
supported by government, business and labor. Ohio is one of the leading states in
7-234
-------
the consumption of coal. In 1989, the State of Ohio produced 31.4 million tons of
coal and has demonstrated reserves of 19 billion short tons. Coal will likely
continue to be the cornerstone of Ohio's energy supply. However, largely as a
result of the Clean Air Act, the demand for Ohio-produced coal, both domestic and
out-of-state, has fallen considerably since 1970.
One of eleven power plants in the Ohio Edison system, the Niles facility was
commissioned in 1954. The main power plant structure covers an area of
approximately 166 ft by 200 ft, and houses two cyclone coal-fired steam electricity-
generating units with a net demonstrated total capacity of 216 MW for both units.
The boiler units burn high-sulfur coal with a capacity factor of approximately 67
percent. Flue gases from both boiler units are dispersed into the atmosphere by a
single 393 ft tall dual flue stack. Each unit utilizes an ESP to control
particulate emissions.
The SNOX Demonstration Project will treat about one-third of the flue gas stream
from Unit 2 or approximately 16 percent of the total flue gas generated at the
plant. The flue gas is taken as a slipstream just upstream of the Unit 2 ESP. Flue
gas cleaned by the SNOX Plant will be returned directly to the existing st^ck
breeching. Project facilities will be installed on a 150 ft by 120 ft unoccupied
area southeast of the plant building.
Demonstration Objectives
The SNOX technology and all the mechanical components which are required for its
successful application have completed both the bench and pilot phases of
development. The 30 HW application in Italy and 300 MW application in Denmark are
evidence that this technology has also entered a commercialization phase. Thus, the
SNOX facility at the Niles Station is not a developmental project but a
demonstration of this technology's readiness for commercial use in the United
States.
The first and foremost objective of the project is to successfully apply the
technology and proprietary equipment to a North American power plant firing high
sulfur coal to confirm the capability to economically meet the pollution control
needs of the utility market. Several supporting objectives are in place to meet
this primary objective. These supporting objectives are to confirm the results
achieved at the Sksrbsk pilot facility, demonstrate the marketability and economic
credits of the sulfuric acid and heat energy by-products, confirm the estimated low
0&M costs, and define any limitations of each piece of major equipment with respect
7-235
-------
to a utility environment. This demonstration facility began operation in November
of this year.
SUMMARY
The SNOX Process has been developed with the objective of near complete removal of
sulfur oxides, nitrogen oxides, and particulate from flue gas with minimum
consumption of resources and maximum recovery of valuable by-products without the
creation of any secondary environmental problems. All indications to date are that
this ambitious objective has been achieved and that the full-scale capital and O&M
costs of this technology are comparable to or lower than those of currently utilized
technologies.
The SNOX Process has several distinct advantages over other de-N0x/de-S0jt processes
and conventional technologies which will rate it as a desirable and superior
technology in the coming decade of increased environmental concern. These
advantages include:
• High S02 removal
• No alkali reagent required
• High N0x removal
• Very low particulate emissions
• Low CO and hydrocarbon emissions
• Only by-product is salable sulfuric acid
• Operating costs decrease with increasing fuel sulfur content
• Furnace integration of recovered heat increases plant thermal
efficiency and compensates for backend energy requirements.
7-236
-------
SNOX INTEGRATED PROCESS FLOW
AH
aoiLOt
"a0
FIGURE 1
SNOX HOT SCHEME PROCESS FLOW
mausx
Ad
HEAT
FIGURE 2
7-237
-------
NEFO PLANT PROCESS FLOW
mxu
COOLER £
—ifcli-— t
DO5TXN0
Mil
CHll
FIGURE 3
7-238
-------
PROGRESS REPORT OF THE
NIPSCO - PURE AIR - DOE
CLEAN COAL O PROJECT
S. B. SATROM
Pure Air
7540 Windsor Drive
Allentown, PA 18195
CL. YEH
Pure Air
7540 Windsor Drive
Allentown, PA 18195
E. WROkil
Northern Indiana Public Service Company
Bailly Generating Station
246 Bailly Station Road
Chesterton, IN 46404
7-239
-------
Intentionally Blank Page
7-240
-------
ABSTRACT
The paper will describe the current status of the NIPSCO - Pure Air - DOE Clean Coal II
Project at the Northern Indiana Public Service Company's Bailly Generating Station. It
will describe the site features, layout, configuration, and current construction status. The
paper will also cover the results of pilot plant design confirmation testing done with
actual coal and limestone samples. The testing was performed at the Mitsubishi Heavy
Industries test facility in Hiroshima, Japan. These tests confirmed the design calculations
for the plant.
7-241
-------
Pure Air, a general partnership between Air Products and Chemicals, Inc., and Mitsubishi
Heavy Industries America, Inc., was established in 198S to market flue gas
desulfurization (FGD) systems and services in North America. MHLA is a wholly-owned
subsidiary of Mitsubishi Heavy Industries, Ltd. (MHT). MHI has sold more than 90 FGD
units worldwide, with a total of over 500 years of operating time on all the units
combined. The general partnership combines the Mitsubishi Advanced Flue Gas
Desulfurization (AFGD) technology with Air Products' plant design, construction, and
operating capability to form a company that can either sell FGD systems on a turnkey
basis or design, construct, finance, own operate and maintain FGD plants. Air Products
pioneered the "on-site" concept over 40 years ago, and currently owns and operates over
16S industrial gas, chemical, cogeneranon, and waste-to-energy plants worldwide.
In 1988, Pure Air was selected by the U.S. Department of Energy (DOE) under the
second solicitation of the Clean Coal Technology Program to install an AFGD system at
Northern Indiana Public Service Company's (Northern Indiana) Bailly Generating
Station. Pure Air will design, construct, finance, own, operate, and maintain the AFGD
facility, and has signed a contract to provide these services to Northern Indiana at Bailly
Station for a 20 year term. During the initial three years of operation, a demonstration
test plan will be conducted to confirm that the Pure Air AFGD system will perform at
high removal efficiencies, in an economic fashion over a wide range of midwestem high
sulfur coals. The economics of this system were previously presented at the 1990 SO2
Control Symposium.
7-242
-------
The Project will demonstrate the following novel features:
Single 600 MW Module
A single 600 megawatt module will treat flue gases from two boilers, Bailly Units #7 and
#8. No spare modules are necessary. Pure Air is providing an availability guarantee to
Northern Indiana for the entire life of the contract.
Co-current Module With In-Situ Oxidation
The facility is designed to produce high quality gypsum by-product while operating on a
range of coals. Oxidation is performed in the absorber tank rather than a separate vessel.
Own and Operate Concept
The FGD system supplier will own, operate, and maintain the SO2 removal, dewatering,
and wastewater system for a period of 20 or more years. Pure Air will provide ongoing
performance guarantees for the life of the plant, which reduces risk to Northern Indiana
and its customers.
Saleable Bv-Product Gypsum
The by-product gypsum produced by this facility will meet specifications established by
the wallboard industry for use in gypsum wallboaid, thus avoiding landfill costs
associated with disposable systems and providing an additional revenue service to
Northern Indiana to partially offset the costs of SO2 removal.
Drv Limestone Injection
Dry limestone powder will be directly injected into the absorber reaction tank below the
liquid level. This system is of particular benefit where space limitations restrict on-site
installation of wet grinding systems. For this project, limestone will be ground off site at
a nearby processing plant and transported to the site in pneumatic vehicles.
High SQ2 Removal Efficiency
Removal efficiencies of 95% or more will be demonstrated.
Wastewater Minimization
Wastewater from the facility is minimized by adjusting the pH of FGD system
blowdown, pressurizing a portion of the flow, and injecting it into the flue gas duct
7-243
-------
upstream of the electrostatic precipitator. The water is vaporized and the solids removed
with the fly ash (Figure 1). Note that the wastewater evaporation system was joindy
developed by Kansai Electric Power Company and MHL
A flowsheet of the project is shown in Figure 2.
Contracts
On 1 July 1989, Pure Air and Northern Indiana entered into a Flue Gas Processing
Agreement whereby Pure Air agreed to finance, design, construct, own, operate, and
maintain the FGD system at the Bailly Generating Station for a period of 20 years. In
exchange for a monthly fee, Pure Air guarantees operating costs, capital costs, and
availability to Northern Indiana on an ongoing basis.
Pure Air and the Department of Energy concluded negotiations and signed the
Cooperative Agreement for the project on 20 December 1989.
Bids for pulverized limestone delivery were solicited by Pure Air on behalf of Northern
Indiana. On 31 December 1990, a contract for supply of limestone was executed with
the J. M. Huber Company.
Pure Air assisted Northern Indiana in marketing the gypsum by-product from this
facility. On 1 June 1990, Northern Indiana concluded a gypsum sale agreement with
United States Gypsum, whereby USG will buy the by-product output from the FGD
facility for use at a nearby wall board facility.
BAILLY PROJECT STATUS
Environmental Annrnvals
The Environmental Assessment required by the National Environmental Policy Act
(NEPA) was prepared and reviewed by DOE officials. A finding of no significant
impact was issued to the project on 16 April 1990.
The Air Permit to Construct was applied for on 23 August 1989 and officially approved
by the Indiana Department of Environmental Management/Air Board on 7 November
1990.
7-244
-------
The National Pollutant Discharge Elimination System (NPDES) Permit Modification was
applied for on 24 August 1989 and officially approved by the Indiana Department of
Environmental Management/Water Board on 15 November 1990.
All required state and local building permits have been applied for and received.
Design Confirmation Test
A two week design confirmation test for the Bailly AFGD Project was conducted at
APC-200 test facility at MHTs Hiroshima Technical Institute. The APC-200 test facility
(Figure 3) is a state-of-the-art pilot unit for total analysis and design confirmation of flue
gas cleaning systems. A coal feed rate of 25 kg/hr. is used to generate 200 NN/b/hr. of
flue gas. The facility is capable of monitoring numerous flue gas process configurations
and of analyzing the inter-related effects of various fuel and absorbent conditions.
Predictions from the APC-200 unit have been confirmed against operating units. The
Bailly plant design confirmation tests were conducted using samples of the coals and
limestone that are expected to be used at Bailly. The purposes of the test were to confirm
calculated performance for SO2 removal efficiency and calcium utilization under design
conditions, verity the reactivity of the limestone source, and confirm by-product gypsum
quality.
Testing was performed using two grinds of limestone which previously had been
qualified based on purity and reactivity. The testing involved combusting coal samples
in the test unit furnace and treating the resultant flue gas in the test unit. The testing
determined the SO2 removal efficiency for different liquid to gas ratios and inlet SO2
concentrations, and at the same time producing a gypsum by-product of greater than
95 wt. % purity. The results of this testing are summarized in Table 1. These results
indicate that SO2 removal efficiencies greater than 97.5% can be achieved processing
flue gas from coals containing up to 4.5 wt. % sulfur.
Construction Status
Official ground breaking ceremonies for the project were held at the Bailly site on Earth
Day 1990 (20 April 1990). As of 1 September 1991, Northern Indiana has completed
necessary site fill and grading operations, the shell of the new chimney, and is in the
process of installing the chimney liners. Northern Indiana is also constructing the
7-245
-------
electrical and other utility feeders to the FGD system battery limits, and making
necessary control modifications to the power plants.
Pure Air has completed Equipment Procurement and Detailed Design Engineering for the
FGD facility. The absorber vessel has been erected and is being lined. Lining activities
are approximately 75% complete. The absoiber hold tank is complete; the recirculation
pumps have been set and piping work is underway. Electrical construction is in progress.
In July, the circulating water lines leading into and out of the power plant collapsed
causing significant damage to the power plant and lesser damage to the FGD facility.
Although the cause of the collapse is still not known, immediately after the collapse, Pure
Air and Northern Indiana formed a joint team to 1) complete efforts to ensure that safe
conditions exist at the accident site and assemble an engineering team, including outside
experts, to begin evaluating the cause of the accident and the degree of damage to all
facilities, 2) develop an integrated approach to begin repairs to the power plant's water
supply system and to resume operations as soon as possible, and 3) resume construction
of the scrubber facility and minimize any delays in its completion.
Pre-commissioning activities are expected to begin in the first quarter of 1991, with
commercial operation expected in July 1992.
7-246
-------
FIGURE 1
WASTEWATER EVAPORATION SYSTEM (WES)
7-247
-------
FIGURE 2
ADVANCED FLUE GAS DESULFURIZATION
PROCESS FLOW DIAGRAM
FLUE GAS DUCTING AND
FANS SECTION
AM
HEATER
UNfT 7
AM
UMTS
HEATER
EXISTING STACK
SO, REMOVAL UMT SECTION 1
1.0. FAN
IMTERUnTEMT
WASH WATER
J
REAGENT FEED
SECTION
GYPSUM 8Y-PR00UCT HANOUNQ
* WUTIWIDDI
owavLow
7-248
-------
FIGURE 3
APC-200
200m>N/h TEST fACILITY
7-249
-------
TABLE 1
RESULTS OF APC-200 DESIGN CONFIRMATION TEST
Limestone: NL-3
Inlet SO2,
ppmd
3,055
2317
Limestone: NNL-3
Inlet SO2,
npmd
3,040
2,330
Calcium Gypsum
L/G, SO2 Removal Utilization, Purity,
1/Nm3 Efficiency. % % wt. %
22.2 97.5 94.3 95.9
16.1 95.9 94.2 95.8
Calcium Gypsum
L/G, SO2 Removal Utilization, Purity,
1/Nm3 Efficiency. % % wt. %
22.2 97.7 93.8 95.5
16.1 96.1 94.1 95.7
7-250
-------
DEVELOPMENT OF A POST-COMBUSTION DRY S02
30NTR0L REACTOR FOR SMALL-SCALE COMBUSTION SYSTEMS
John C. Balsavich
Tecogen Inc.
45 First Avenue
P.O. Box 9046
Waltham, Massachusetts 02254-9046
7-251
-------
Intentionally Blank Page
7-252
-------
ABSTRACT
Under a contract with the U.S. Department of Energy, Pittsburgh Energy Technology Center (PETC),
Tecogen Inc. is developing an integrated emissions control system for use with small-scale, coal water
slurry (CWS) furnace systems. This program focuses on the reduction of NOx, S02, and particulate
emissions to levels of 02., 0.4. and 0.02 pounds per million Btu fired, respectively. With respect to
particulate emissions, the objective is to capture 99.9 percent of the particles in the size range of 0.5 to
10 microns.
As part of this work, a unique emissions control reactor is being developed for the post-combustion
reduction of S02 emissions. This reactor uses the principles of fluid mechanics and particle dynamics
in an innovative way to enhance the capture of S02 by various sorbents. The S02 emissions control
reactor is a vortical flow device designed to separate sorbent particles and then confine them within the
reactor. This separation of particles within the reactor substantially increases their residence time
relative to the gas and therefore provides an effective soibent-to-sulfur ratio within the reactor that is
many times greater them what was injected, providing a very high sorbent particle surface area
concentration within a very compact reactor.
This paper describes the significant features of the S02 emissions control reactor, presents a simplified
analytical model used to determine key operation parameters, and reports recent test results. These
results include reductions in S02 levels greater than 90 percent
7-253
-------
DEVELOPMENT OF A POST-COMBUSTION DRY S02
CONTROL REACTOR FOR SMALL-SCALE COMBUSTION SYSTEMS
INTRODUCTION
The current contract Tecogen Inc. has with PETC (Contract No. DE-AC22-89PC8904) deals with the
development of an integrated emissions control system for small-scale, CWS-fired furnace systems.
This program focuses on the reduction of NOx. S02, and particulate emissions to levels of 0.2, 0.4, and
0.02 pounds per million Btu fired, respectively. With respect to particulate emissions, the objective is to
capture 99.9 percent of the particles in the size range of 0.5 to 10 microns.
The control of NOx and particulate emissions is being achieved via staged combustion for NOx control
and the use of a high-efficiency baghouse for particulates. In order to control S02 emissions a dry
sorbent post-combustion strategy was adopted. To achieve the program's SOz emissions goal, a
separate high-temperature reactor is currently being developed.
The emissions control reactor currently being developed uses the principles of fluid mechanics and
particle dynamics in an innovative way to enhance the capture of S02 by various sorbents. The S02
control reactor is a vortical flow device in which the products of combustion and sorbent particles enter
tangentially at the reactor base and flow upward towards the discharge. Due to the vortical flow field
within the reactor, the sorbent particles are separated and then confined within the reactor. Separation
of the sorbent particles within the reactor substantially increases their residence time relative to the gas
and therefore provides an effective sorbent-to-sulfur ratio within the reactor that is many times greater
than what was injected, providing a very high sorbent particle surface area concentration within a very
compact reactor.
The S02 reduction reactor was designed on the principle of axial recirculation. In a vortical flow field,
the pressure at the periphery is greater than that at the centerline. By connecting one end of a tube
tangentially from the periphery at one axial location and the other end at the centerline. an external
axial recirculation flow path is established that operates only on the natural pressure difference present
in the vortical flow field without requiring external assistance.
When the principle of axial recirculation is used in a gas/solids system, the residence time of the solids
within the reactor can be significantly increased relative to that of the gas, and can provide good solids
distribution throughout the reactor as well. A schematic diagram of both the internal and external
particle recirculation paths is shown in Figure 1, and a photograph of the external particle recirculation
is shown in Figure 2.
7-254
-------
ANALYTICAL MODELING
In developing an analytical model of the S02 reactor, several complexities are encountered due to the
coupling of the flow field in the main body of the reactor to that of the axial recirculation loop. These
two highly coupled flow fields result in the need to solve a three-dimensional, highly nonisotropic
turbulent flow field with a dispersed second phase of particles that undergo chemical reactions.
Resolving the flow field, tracking the sorbent particles, and accounting for changes in chemical species
require an extremely high computational effort As an alternative to this approach, a simplified model of
the reactor was developed. The purpose of this model was to determine the dominant parameters that
govern the reduction of S02 within the reactor as well as to provide performance estimates under
various operating conditions.
One of the prime areas of concern is the ability of the reactor to detain particles of a given size for
given gas-flow conditions. Though the flow within the reactor is quite complex, except for the
recirculation line effect the reactor is quite similar to a cyclone separator. The ability of cyclones to
collect particles of a given size has been well documented; by Zenz. for example.1 For a given
geometry and gas flow conditions the theoretically smallest-size particle size that a cyclone can collect,
typically called the cut point diameter, can be found by:
where:
is the gas viscosity
Vjn is the inlet velocity
pp, pg are the particle density and gas density, respectively
Ns is the number of spiral turns, which is found from experiment
L^ is the cyclone inlet width
For application of this equation to the axial recirculation reactor, L^, is defined as the radial distance
between the reactor wall and the outside radius of the recirculation tube.
Once the cut point diameter is determined, the separation efficiency of any other size particle can be
determined from normalized separation efficiency data, as shown by Zenz.1 Knowing the separation
efficiency of a given-size particle allows one to determine its residence time in the reactor relative to the
gas. Under steady-state conditions and assuming that the reactor is capable of fluidizing all the
particles within it as one particle is added to the reactor, one particle t jst leave. However, the
particle that is added to the system will be recirculated N times before it leaves. The number of times a
particle of a specific diameter within the reactor circulates - N - is dependent on the separation
efficiency for this size particle. This relationship can be expressed as:
D
^ **0 2
0)
N=1 ->• TJ T|2 + T^3 +
(2)
where ti is the separation efficiency. This relation can be expressed in closed form as:
7-255
-------
N 2— (3)
1 - T)
Once N is determined, the amount of solid material in the reactor that can participate in the reduction of
S02 can be found since it is equal to N times the inlet solids flow rate.
When the amount of solids present in the reactor is known, one-dimensional species-continuity
equations for the concentration of S02 in the gas phase and CaO in the solid phase can be written as:
d [SOJ -6 (1 - e) [SOJ
dz (1 ^ (4)
e DpV8
[k * kj
d [CaO]
dz
-6 [SOJ
Dp V
f-L + ±)
(5)
where
[S02] is the concentration of S02 in the gas phase
[CaO] is the concentration of CaO in the solid phase
Vg, Vs are the gas and solid phase velocities, respectively
e is the void fraction (adjusted to account for particle recirculation)
kc, kd are the chemical kinetics and diffusion resistances, respectively
D is the particle diameter
The diffusion resistance, kd, can be found from the relationship:
i 1
Sh = 2 + 0.6 Re2 Sc3
3 (6)
where:
Sh is the Sherwood number defined by:
Sh-K.% P>
D is the diffusivity
Re is the Reynolds number based on the slip velocity
Sc is the Schmidt number
The chemical kinetics resistance is determined by using Borgwardfs2 model, which takes into account
the decrease in the rate of reaction due to pore plugging.
7-256
-------
k..A„«p(-|>£)«P (-^)
(8)
where:
A0 is the frequency factor
n' is the sulfate in the stone
W is the weight in the solid sample
Ft is the universal gas constant
T is the absolute temperature
E is the activation energy
P is the stone-dependent constant determined by experiment
Model Parametric Study
The model was used as part of a parametric study to determine the effect of such parameters as
reactor diameter, Ca/S, and particle diameter on the S02 capture. For this analysis, the reactor model
was operated with the combustion products of the CWS specified for this program. The parent coal in
this CWS is an Illinois No. 6 coal, which contains 2.63-percent sulfur and has a higher heating value of
13,579 Btu/lbm. This provides the potential of 3.87 Ibm/MMBtu of S02. When this CWS is burned with
20-percent excess air, this S02 potential translates into a concentration of 1,950 ppm on a dry basis.
The parameters used in this analysis are shown in Table 1.
The effect of sorbent particle diameter on reactor performance is shown in Figure 3 for reactor
diameters of 4, 6, and 8 in., operating with a Ca/S mole ratio of 3:1. The interesting feature of Figure 3
is that for each reactor there is an optimum particle diameter for S02 capture. This optimum diameter
is based on the reactor's ability to capture particles of a given size. Particles smaller than the optimum
diameter cannot be detained in the reactor and pass through without participating to a great extent in
the S02 capture process. Particles that are larger than the optimum diameter are able to be detained
in the reactor; however, since the sorbent particle surface area to volume ratio decreases as the
diameter increases, there is less available surface area in the reactor to participate in the S02 capture
process compared to the optimum.
The effect of Ca/S mole ratio on reactor performance is shown in Figure 4 for reactor diameters of 4, 6,
and 8 in. For each of these cases the model used the optimum particle diameter for these operating
conditions, as shown in Figure 3. For the soibent properties used in this analysis, the required 90-
percent reduction in SOz emissions is achieved at a Ca/S mole ratio of about 4.5 for the 8-in.-diameter
reactor.
The effect of input rate on reactor performance is shown in Figure 5 as a function of particle diameter.
For this case, only an 8-in.-diameter reactor is considered since at the higher flow rates, the pressure
drop becomes excessive in the smaller units. Figure 5 shows that as the input rate is increased,
reactor performance decreases. This is due to the reduction in gas phase residence time associated
with the fixed reactor height.
7-257
-------
An interesting feature of Figure 5 is that as the input rate is increased, the optimum particle diameter
for S02 capture decreases. This is due to the fact that at the higher tangential velocities associated
with the increased flow rates, the reactor has a greater ability of capturing smaller sized particles. This
provides a greater surface area concentration within the reactor.
Figure 5 indicates that if operation at higher input rates is required and comparable performance with
lower input rate operation is required, provisions must be made to compensate for the reduction in gas
phase residence time. This can be accomplished by increasing either the reactor length or diameter.
The specified coal for this program, when burned with 20-percent excess air, has the potential to
produce 1,950 ppm of S02 on a dry basis. Operation of the reactor with the products of combustion of
other slurries yields the potential for different SOz inlet concentrations. The performance of the reactor
as a function of inlet S02 concentration is shown in Figure 6 for a Ca/S mote ratio of 3:1 and reactor
diameters of 4. 6, and 8 in. Figure 6 shows that the percent reduction in S02 increases as an
exponential function of inlet concentration, approaching a limiting value of close to 100-percent
reduction.
The fact that the percent reduction in S02 approaches an exponential lim't at near 100 percent as inlet
concentration is increased indicates that in absolute terms, i.e., with respect to the actual outlet
concentration (ppm), the reactor would have a lower outlet S02 concentration when operated with
higher sulfur coals. To illustrate this, the percent reduction results shown in Figure 6 are plotted in
Figure 7 in terms of the actual outlet concentration of S02 (ppm on a dry basis). Figure 7 shows that
particularly for the 6-in.-diameter and 8-in.-diameter reactors, as the inlet concentration is increased, for
a fixed Ca/S ratio there is more sorbent within the reactor and the outlet S02 concentration decreases.
For example, consider the 8-in.-diameter reactor. Using this particular sorbent, an outlet concentration
of about 400 ppm would be expected with the reactor operating at the 1,950 ppm level expected for this
coal. This result is about 200 ppm higher than the level required to meet the 0.4 Ibm/MMBtu program
goal. However, if the reactor was operated with a coal that produced about 3,500 ppm of S02. the
outlet concentration would meet the goal of 0.4 Ibm/MMBtu (200 ppm).
S02 REACTOR TEST RESULTS
A schematic diagram of the SOz reduction test facility is shown in Figure 8. In order to have more
control over the operation of the SOz reactor, it was decided to operate the facility with flue gases
produced from an oil burner and inject S02 as required in order to obtain desired inlet concentrations.
Also, since oil burners can operate over a very wide range of input rates, typically from 68,500 to over
400.000 Btu/hr for residential scale burners, they provide a great deal of flexibility in reactor operation.
The reactor tested in this facility had the following geometry:
Inside Diameter: 8.329 in.
Outlet Diameter: 1.5 in.
Active Reactor Height: 43 in.
Inlet Area: 4 in.2
7-258
-------
Four different calcium-based sorbents were tested in the 8-in .-diameter reactor, and a summary of
these sorbents in terms of specific calcium compound, weight percent, and mean particle diameter is
shown in Table 2. The sorbent used "m this work represented a cross section of various products found
throughout the United States. Sorbent A-l was a dolomite from northwest Connecticut. The as-
received version of this sorbent. called Sorbent 1. had a mean diameter of 194 microns. In order to
obtain the as-tested size distribution, the as-received dolomite was sieved and then selectively blended.
Sorbent 2 was a high-calcium limestone from Pennsylvania that was received in pulverized form.
Sorbent 3 was a pulverized lime that was calcined from the same limestone used in Sorbent 2, and
Sorbent 4 was an as-received limestone from Illinois.
The S02 reduction test facility operates with the products of combustion from an oil burner using
Number 2 heating oil. Gaseous S02 is injected to achieve the desired reactor inlet concentration.
Tests conducted with the 8-in .-diameter reactor were designed to simulate operation with the specified
CWS for this program. The coal in this slurry contained 2.36-percent sulfur and had a heating value of
13,579 Btu/lbm, yielding a potential for the formation of 3.87 Ibm/MMBtu of S02.
The products of combustion of Number 2 oil differ in composition from those of CWS, and it becomes
necessary to put these results on a coal-equivalent basis to allow weight-based, not volume-based
results to be reported. In order to establish this equivalent basis, it was decided to equate the oxygen
content in the products of combustion of heating oil with that of CWS. With an equivalent oxygen level
for CWS combustion established, the stoichiometric ratio can be determined and the volumetric
measurements recorded by the analyzers can be converted to a weight basis.
Results of 8-in.-diameter reactor testing are shown in Figure 9 in terms of S02 content. Ibm/MMBtu. as
a functions of Ca/S mole ratio along with lines that represent a least squares method fit of the data
(exponential fit). Figure 9 shows that the S02 emissions goal for this program - 0.4 Ibm/MMBtu -
was reached at a Ca/S mole ratio of about 5:1 for Sorbent 1-A.
The data shown in Figure 9 bring up an interesting feature about the reactor and the S02 capture
process itself. Consider the results for Sorbents 2 and 3. which have essentially the same mean
particle diameter and distribution. Sorbent 3 is made from the calcination of the material used in
Sorbent 2, and since it is CaO that captures S02 directly and not CaC03, it was expected that better
SOz capture would be achieved with Sorbent 3 since the calcination step within the reactor was
eliminated. This was not the case. Sorbent 2, which undergoes the calcination process within the
reactor, exhfoited a much greater ability to capture S02 and at much lower Ca/S mole ratios than
Sorbent 3. In fact Sorbent 2 achieved an S02 emissions level of 0.54 Ibm/MMBtu, an 88.3-percent
reduction at a Ca/S mole ratio of 4.12 compared to 0.73 Ibm/MMBtu, and an 80-percent reduction at a
Ca/S mole ratio of 7.53.
In addition to reporting S02 reduction as a function of Ca/S mole ratio, it is also desirable to use as an
independent variable the sorbent mass loading, i.e., the mass of sorbent injected into the reactor for
each unit mass of S02. These results are shown in Figure 10. The main difference between Figure 9
and Figure 10 is with respect to Sorbent 1-A, which achieved 0.39 Ibm/MMBtu at a Ca/S mole ratio of
about 5:1. However, since Sorbent 1-A is a dolomite that contains approximately 40-percent MgC03, it
takes considerably more sorbent on a mass basis to achieve a given Ca/S mole ratio compared to any
7-259
-------
of the other high-calcium-based sorbents. Therefore in order to achieve the required reduction in SO^
over 13 pounds of Sorbent 1-A are required for every pound of S02 in the gas stream. For purposes of
comparison, consider achieving a reduction in SOz to 0.6 Ibm/MMBtu. For Sorbent 1-A, this would
require 12 pounds of sorbent for each pound of S02; however, for Sorbent 2, less than 6 pounds of
sorbent are required for each pound of SOg.
The results shown in Figures 9 and 10 are S02 content on a Ibm/MMBtu basis versus Ca/S and
sorbent loading ratio, respectively, which represent absolute values. In some instances, it is desirable
to present the data on a relative basis; i.e., percent reduction in SOg. This is shown in Figure 11, in
which percent reduction in S02 is plotted against Ca/S mole ratio. This plot shows in a very simple
manner how each sorbent performs relative to the others. For example, at a Ca/S mole ratio of 1:1,
Sorbent 2 exhibits an almost 60-percent reduction in SOz whereas Sorbent 4 at this Ca/S mole ratio
exhibits only about 15-percent reduction.
In comparing the data obtained in the S02 reduction test facility with the analytical model, three
constants are needed in order to characterize the sorbents' chemical kinetic behavior. These constants
are the activation energy, E; the frequency factor, A0; and the pore plugging parameter, beta. Based
on test results obtained with Sorbent 3, a high-CaO lime product taken at three temperatures, these
constants were determined to be:
Activation Energy: 11,750 cal/gmole
Frequency Factor: 243,100 l/s
beta: 0
which are well within agreement of published data.2
With an estimation of the sorbent properties, the computer model comparison was performed for two
different cases. These results are shown in Figure 12 in which percent reduction in SOz is plotted
against Ca/S mole ratio. Figure 12 shows that a fairly good agreement exists between the model and
data.
SUMMARY AND FUTURE WORK
Test results obtained with the 8-in.-diameter reactor using calcium-based sorbents have achieved over
a 90-percent reduction in S02 emissions. Emission levels of as low as 0.39 Ibm/MMBtu have been
obtained, which are lower than the 0.4 Ibm/MMBtu program goal.
Reactor performance was found to be a function of both the sorbent used in the capture process and
its mean particle diameter. A noticeable increase in reactor performance was observed when the
sorbent used in the S02 capture process was calcined within the reactor as opposed to using a
precaJcined product
In future work, parametric investigation of the S02 reactor will continue. This work will include testing
various other sorbents as well as baffles and other devices to aid in particle separation. In addition to
7-260
-------
parametric testing, the S02 reactor will be integrated into a complete small-scale CWS-fired furnace
and will be tested extensively.
ACKNOWLEDGEMENT
This work was funded by the U.S. Department of Energy, Pittsburgh Energy Technology Center under
Contract No. DE-AC22-89PC8904. The author would like to express his appreciation for the financial
support to accomplish this development effort The valuable guidance and encouragement given by
Mr. Thomas Brown, DOE'S Project Manager, is gratefully acknowledged.
REFERENCES
1. FA. Zenz, "State of the Art Review and Report on Critical Aspects and Scale Up Considerations in
the Design of Fluid Bed Reactors," Phase I Final Report, DOE/MC/14141-1158.
2. R.H. Borgwardt, "Kinetics of the Reaction of S02 with Calcined Limestone," Environmental Science
and Technology, Vol. 4, No. 2, 59, 1970.
7-261
-------
Figure 1. External and Internal Circulation
in the Emissions Control Reactor
Figure 2. Emissions Control Reactor Model (High Load)
Intensive External and Internal Circulation
4
V
i
S *
1
Figure 3. Effect of Particle Diameter on Reactor Performance
(Operating with a Ca/S Ratio of 3:1)
7-262
-------
M
m
y*
m
M
to
»
Figure 4. Effect of Ca/S Mole Ratio on Reactor Performance
to
w
to
to
30
H
to
Figure 5. Effect of Input Rate on Reactor Performance
(For an 8-ln.-Diameter Reactor Operating
with a Ca/S Ratio of 3:1)
m
to
TO
to
M
to
M
H
>Ci
Figure 6. Effect of S02 Concentration on Reactor Performance
(Operating at a Ga/S Ratio of 3:1)
7-263
-------
M*
M
iO
Figure 7. Effect of Inlet SO2 Concentration on Reactor Outlet Concentration
(Operating with a Ca/S Ratio of 3:1)
Figure 8. SO2 Removal Test Facility
Figure 9. Effect of Ca/S Mole Ratio on 8-In.-Diameter Reactor Performance
7-264
-------
Figure 10. Effect of Sorbent Loading (Weight Basis) on the Performance
of the 8-In. Reactor
Figure 11. Effect of Ca/S Mole Ratio on 8-In.-Diameter Reactor Performance
m
m
79
M
M
40
3D
19
Figure 12. Comparison of Model Prediction and Reactor Data for the 8-ln.-Diameter Reactor
Operating with Sorbent 3
7-265
-------
TABLE 1
MODELING PARAMETERS
Fuel:
Coal Content (weight %)
Higher Heating Value (Btu/lbm)
Sulfur Content (coal)
IL6-248-MCW-G
(Specified fuel for the program)
5322
7,227
2.63
Input Rate (Btu/hr):
Excess Alr(%)
S02 (ppm dry basis)
Reactor Operating T. mperature (F)
120,000 (nominal)
20
1,950
1,600
Reactor Diameter (In.)
Reactor Height (In.)
4, 6,8
45 (nominal)
Sorbent Properties:
Activation Energy (cal/g-mole)
Frequency Factor (1/s)
Beta (g/g-mole)
Density (Ibm/tt5)
10,000
335,000
333.3
150
TABLE?
SORBENTS USED IN TESTING THE 8-INCH-DIAMETER S02 REACTOR
Sorbent
Calcium
Compound
Percentage
Mean Diameter
(microns)
1-A
CaC03
60
117
2
CaC03
973
88.5
3
CaO
95.4
82.1
4
CaC03
983
172
7-266
-------
SCRUBBER REAGENT ADDITIVES FOR
OXIDATION INHIBITED SCRUBBING
Jeffery L. Thompson
Process Calx, Inc.
3059 Old Stone Drive
Birmingham, Alabama 35242
7-267
-------
Intentionally Blank Page
7-268
-------
ABSTRACT
Gas film limited scrubbing (maximum removal for a given L/G)
requires a high liquid phase alkalinity in the scrubbing
liquor, most readily obtained in the form of buffered
sulfite ions (SO33). The cations used as an ion pair with
the sulfite are most often Mg (the Thiosorbic process by
Dravo Lime) and Na (in dual alkali systems). Excessive
amounts of MgO as em additive allows the more soluble
sulfate to increase, thus depressing the sulfite solubility,
and total alkalinity.
Manufactured minerals such as CaAl2Si20g, and
CaAl2Si20s-Al2Si205(OH)4, have a much greater buffering
capacity for sulfite ions than Mg(0H)2/ and more nearly
match sulfite/sulfate solubility so that sulfite alkalinity
is not diminished by small excesses of additive.
Because these mineral compounds are manufactured, small
amount of sulfide can be included by reacting the minerals
with weak solutions of hydrogen sulfide, thus providing a
built-in oxidation inhibitor.
7-269
-------
BACKGROUND
Wet FGD scrubbing is controlled almost entirely by liquid
phase chemistry of the liquor being pumped to the absorber
module(s). In the simplest terms, the capture of a given
amount of sulfur dioxide requires a matching amount of
alkalinity in the liquor. The required amount of alkalinity
can be thought of as some number, A. A is the product of
the volume of liquor, VL, and the concentration of
alkalinity in the liquor, cB. Thus, A ¦ (cB) x (VL). If
the alkalinity concentration is high, VL can be small. As
the concentration, cB, diminishes, a corresponding increase
in VL is required to obtain the same A. In terms of a
scrubber, the L/G ratio must increase as the concentration
of alkalinity in the liquor declines in order to remove the
same amount of SO2 from the flue gas.
It is the liquor that reacts with sulfur dioxide in the
absorber module to provide SO2 removal. The role of
limestone and lime in wet FGD systems is not to capture
sulfur dioxide in the absorber, but rather to knock the
absorbed sulfites/sulfates out of the liquor as precipitates
in the recycle or mix tank. The liquor does the scrubbing
in the absorber, and the limestone and/or lime reagent
reconstitutes the liquor in the recycle tank. The soluble
additives in the reagent define the liquid phase chemistry,
and hence sulfur dioxide capture. The reagent itself
removes the captured sulfur from the liquor.
7-270
-------
If calcium la the only, or primary cation In the system,
gypsum scaling Is almost Inevitable, because the absorbtlon
and precipitation phases overlap. There Is too much Ca++ In
solution In the absorber, thus the addition of more sulfur
produces a gypsum precipitate. Additives are used to
provide additional cations, which act to depress the calcium
solubility, and lower the relative gypsum saturation.
Essentially, modern scrubbers are attempting to approach the
dual alkali concept with a single loop. The dual alkali
concept uses sodium as the cation to carry buffered sulfite
alkalinity. The soluble sodium bl-sulflte solution, NaHS03,
Is then reacted with calcium hydroxide to form calcium
sulfite solldes and reconstitute the sodium sulfite solution
(scrubbing liquor). The Dravo Thiosorbic process uses the
same idea, but with Mg rather than sodium, and the magnesium
is included in the lime as KgO. Organic acids are intended
to play a parallel role in limestone systems.
Oxidation of sulfite to sulfate in the absorber only serves
to lower the available alkalinity for sulfur dioxide
capture. Thus, ideally, oxidation only occurs in the
recycle or mix tank, not in the absorber. The liquid phase
should be oxidation inhibited in the absorber, and in the
case of forced oxidation, the reaction should be delayed
until the solution reaches the recycle tank.
Scrubbing systems have two exactly opposite demands. Lower
pH values favor increased solubility of the reagents,
limestone, or lime. But higher pH values provide greater
alkalinity, hence better removal rates with lower L/G
ratios.
7-271
-------
MANUFACTURED MINERAL ADDITIVES
Synthetic minerals which dissociate incongruently in a
scrubbing liquor can have the dual advantage of providing
higher buffered sulfite alkalinity over a wider range of pH
than single compounds, e.g., magnesium hydroxide, and can
act to inhibit sulfite oxidation in the absorber, while also
promoting large, simple crystal growth in the recycle tank.
At the present time, research indicates that as the mineral
becomes increasingly more complex with a larger total amount
of OB~ ions associated with its hydrated form, the buffering
capacity of sulfite alkalinity is improved. A series of
minerals have been investigated, starting with natural
Kaolinite: Al2Si205(0H)4. The sulfite buffering capacity
was improved by making a combined mineral of
CaAl2Si20g-Al2Si205(OH)4. still greater buffering capacity
has been provided by Ca4Al2Si2(OH)32-Al2s^2°5(OH)4•
Oxidation inhibition can be added by reacting the hydrated
mineral compound with a weak solution of hydrogen sulfide,
H2S. Oxidation inhibition in the absorber and enhanced
crystal growth in the recycle tank have been achieved by
treating the hydrated mineral compounds with a mixed
solution of hydrogen and iron sulfides.
It is unlikely that the ideal mix of calcium and magnesium
carbonate will occur near every scrubber. Moreover, as a
scrubber system becomes tighter (less and less make-up water
is added) the optimal amount of additive will change
(diminish). The two ingredients, calcite or calcium
carbonate, and dolomitic stone or calcium-magnesium
carbonate, can be viewed as two colors of paint: black and
white, with which one can produce any shade of gray. Thus,
reagent design can be defined in terms of high calcium lime,
or limestone, and an additive that provides optimal
7-272
-------
performance of a given combination of coal, boiler,
scrubber, and scrubber operation.
Process Calx, Inc., is presently designing and building a
pilot scrubber to provide utilities with a service that will
simulate any major scrubber configuration, along with boiler
quench rate, and use any coal, to determine the optimal mix
of limestone, and/or lime, and additives.
The manufactured minerals, both composition and methods of
production are the subject of a patent application by
Process Calx, Inc.
7-273
-------
Intentionally Blank Page
7-274
-------
RECOVERY OF SULFUR FROM CALCIUM SULFITE
AND SULFATE SCRUBBER SLUDGES
Jeffery L. Thompson
Process Calx, Inc.
3059 Old Stone Drive
Birmingham, Alabama 35242
blank
7-275
-------
Intentionally Blank Page
7-276
-------
ABSTRACT
Calcium and sulfur can be effectively extracted from scrubber
sludges so that the calcium can be re-used, and the sulfur sold
off, through a two step process of bio-chemical reaction and
smelting. Calcium sulfite/sulfate sludges are placed in a
digester containing anaerobic bacteria, desulfo-vibrio, to which
is added metal powders, e.g., electric arc furnace dust, or BOF
dust. The resulting mixed metal sulfides are split by hydro-
cycloning: calcium sulfide, and all other heavier sulfides,
i.e., iron, manganese, chrome, etc. The sulfides are then
smelted individually, calcium by itself, and the others. The
calcium is returned to the scrubber, the combined metals are
returned to the melt shop, and the sulfur is captured for
commercial sale.
7-277
-------
BACKGROUND
Reagent: and disposal costs are significant, factors in the
decision matrix of comparing wet FGD options of enhanced lime
(oxidation inhibited) vs. limestsne/forced oxidation. While
there is no debate over the chemical facts that higher pH
scrubbing with buffered sulfite provides more NTU (Number of
Transfer Units) at lower L/G ratios, the economics of the
comparison is anything but obvious.
Reagent costs and disposal volume can be reduced by recycling the
calcium used to precipitate the absorbed sulfur out of the
scrubbing liquor, and extracting the sulfur from the calcium
salt. Calcining calcium sulfate to recover calcium oxide and
sulfur has been in commercial practice for a number of years by
ICI in the UK, among others. This process requires a reducing
atmosphere, and is as energy intensive as calcining limestone
into lime. Moreover, the capital required and the expense of
operating the sulfur dioxide recovery operation is relatively
high compared to current sulfur prices. The commercial viability
of this avenue is very sensitive to the prices of fuel and
sulfur.
A much simpler method of calcium recycling and sulfur recovery
was devised by Process Calx, Inc., in the course of defining
methods for controlling the Hydrogen sulfide odor problem that is
common to many FGD systems. Almost all rivers and streams used
for process make-up water contain some concentrations of common
anaerobic, sulfur reducing bacteria: primarily, desulfo-vibrio.
The bacteria reduce the sulfites and/or sulfates (remove oxygen
and add electrons) to yield sulfides and release hydrogen sulfide
7-278
-------
(the source of the odor problem). While the objective of the
study was to control the bacteria and eliminate the odor, a
critical dimension was to define the conditions in which the
bacteria flourish.
Typical fly ash from a utility boiler contains sufficient
unbumed carbon to apparently act as a food supply for the
bacteria. Anaerobic digester conditions with a pH in the range
of 8 to 9, and temperatures of 30 to 45 °C produced the most
rapid growth of bacteria and reduction of the sulfite/sulfate
sludges. Adding powdered metal oxides to react with the H2S in
solution (to keep the pH around 8.5) and produce a metal sulfide
precipitate was more effective than adding additional hydrate,
calcium hydroxide.
Electric arc furnace dust, EAFD or K-061 in the EPA list of
Hazardous Wastes, was the most effective additive to remove the
hydrogen sulfide in solution in the digester. EAFD itself has
been the subject of investigation. The presently accepted method
of destruction is to vaporize the EAFD in a plasma arc and then
condense the metals from the vapor. Energy requirements for this
technology are in the range of 1700 KWh/ton of EAFD. There aire
several problems with EAFD as more and more scrap metal is
recycled. Recycling the EAFD has problems in that the melt shop
doesn't want zinc and lead returned to the melt, as these
concentrations build in the recycle dust. Reacting the EAFD, a
very fine powder of mixed metal oxides, with H2S in solution is
an effective method of removing the hydrogen sulfide, and
producing individual metal sulfides which are easily separated on
the basis of density.
7-279
-------
sulfide density
CaS 2.5
FeS, FeS2 5.0 to 4.8
CaS
PbS
ZnS
HiS
7.5
4.0
5.3 to 5.6
The commercial interest for stainless steel shops is the recovery
and recycle of the nickel along with the elimination of Pb, and
Zn. It isn't necessary to separate the iron and nickel. The
operation is set to yield four catagories of sulfides: the
lightest fraction (calcium, along with fly ash, etc.,); light
middle (zinc); heavy middle (Fe, the bulk of the metal oxides,
along with Hi for stainless shops); and the heaviest fraction
(Pb).
Roasting a sulfide to recover a metal with a nearly pure sulfur
atmosphere is much simpler system than calcining sulfites and/or
sulfates to recover sulfur dioxide.
PROCESS DESIGN
A typical 500 MWe boiler, 60% load factor, uses 1 million tons
per year of coal. With a 3% sulfur content in the coal, the
reagent requirements are approximately 58,000 tons per year of
lime (including a 4% MgO, or other additive), and the disposal of
almost 130,000 tons of scrubber solids. With this sulfur
extraction and calcium recycle system, the reagent requirement is
only 6,000 annual tons, and the disposal is 15,000 tons of dry
solids from the slaking operation. The essential dimension of
this technology is the rate at which the bacteria can process
(reduce) the sulfites and sulfates.
7-280
-------
10® ton Coal
3% S
500 MWa
BOILER
ESP
ABSORBER
fly ash
grit to
ASH POND
SLAKER
THICKENER
51 k
tn CaO
kiln
24 k tn
ANAEROBIC OIGESTER
ASH POND
HYDRO-CYCLONES
PbS F«S ZnS CaS
I ^2 , \
I NiS \
EAFD 12 k tn
from malt shop
6 k tn
8 ktn return
to melt shop
Fa, Ni, Mn.
Cr, Cu. Co
kiln
Roast
2 ktn
to metal recovery
Figure 1. Flow diagram for the extraction of sulfur from
sulfite and sulfate scrubber solids by means of a
biochemical process, including the recycle of calcium, and
the production of reduced metals.
7-281
-------
Bench scale work has maintained conversions of scrubber solids
sulfites to calcium and metal sulfides at a rate of 0.1 lbs/(hr
ft3) of active digester volume, or 1.6 g/(hr liter).
This process has two revenue streams: a fee charged to the melt
shop to recycle the metals from the electric arc furnace dust;
and the sulfur that is roasted from the sulfides.
A flow sheet for the process applied to a 500 MWe unit is shown
above in Figure 1.
CURRENT STATUS
Increasing the rate at which the bacteria reduce the sulfites and
sulfates to sulfides and producing cleaner separations of the
sulfides are the critical parameters. Present rates of sulfite
reduction are 0.1 lbs/(hr ft3), which does not increase with
bacteria count (8.5 avg pH, and 100 °F). Present recycle of
calcium can be maintained at 90+ % of total CaO in the system.
Cleaner iron sulfide precipitate (free of Zn and Pb) would
improve the value of the returned metals. A typical EAFD
contains about 20% zinc oxide, and 2% PbO. Presently, the
roasted iron/nickel metal contains less than 1% total Zn, and Pb.
The process as described has been studied at the laboratory, or
bench scale, and is the subject of a patent application by
Process Calx, Inc.
7-282
-------
MAGNESITE AND DOLOMITE
FGD TECHNOLOGIES
D. Najmr
S. Najmr
Ore Research Institute
25210 Mnisek pod Brdy
Czechoslovakia
7-283
-------
Intentionally Blank Page
7-284
-------
ABSTRACT
Solutions of calcined magnesite or dolomite are used as scrubbing
agents. Useful by-products are produced. In the magnesite proce
MgS03 or MgS04 are produced; in the dolomite process a mixture of
MgSOj and CaCOj is produced, all valuable agricultural fertilizers
Employing a unique purification based on the recrystallization of
MgS03 from a metastable solution of MgSOj.SHjO, 99.5% pure MgSO} cc
also be obtained. Both magnesite and dolomite technologies are
described. Results of tests conducted on a pilot plant at the OR
facilities are summarized. 75 t.h"1 steam boiler desulfurization
described. Attention is given to agrochemical properties of the
products.
7-285
-------
BACKGROUND
Ore Research Institute has developed magnesite FGD technology which
uses low quality calcined magnesites as scrubbing agents. It is an
open technology (as opposed to regenerable technology) which produces
MgSOj as by-product, valuable fertilizer. The technology has been
tested on a 2,000 Nm'.h"1 pilot plant at the Ore Research Institute
facilities in Mnisek pod Brdy, Czechoslovakia since 1989. Technical
studies for application on commercial size projects have been done.
One of them is being considered for 200 MW power plant at Tusimice,
Czechoslovakia, where a desulfurization plant was constructed by a
team from the Soviet Union in years 1983-1990, but was never brought
to operation. Ore Research Institute is proposing 10% MgS03 to be
taken out of regenerable cycle and to be used as fertilizer. This
will prevent build up of ash and other insolubles in the circulating
sorbent. Experience gained during pilot testing and at 200 MW
Tusimice power plant gives ORI desulfurization team confidence to
evaluate their open magnesite technology as one of FGD technologies
that is ready to be used on commercial scale projects.
DESULFURIZATION OF 75 t.h"1 STEAM BOILER
BY ORI MAGNESITE TECHNOLOGY
Data
Flue gas
SOz content
Fly ash content
SOz capture
Fly ash capture
MgO content
S02 captured
100 000 Nm5.h"1
8.42g.Nitf3
0.24g. Nm"3
92% minimum
90% minimum
75%
775 kg.h-1
7-286
-------
Process
Please see Figure 1.
In the process the flue gas passes through heat exchanger 1 (where it
exchanges heat with scrubbed exhaust) into absorber 2 where it is
scrubbed. Suspension flows into circulating tank 4 from which it is
pumped through hydrocyclones 6 into spraying floors of the absorber
2. The flow from hydrocyclones enters distributor 7 where it is
distributed either to circulation tank or to centrifugal separator
12. The cake is dried in drier 14. Dry product is sorted 15,
crushed 16, and transported into expedition tank 19.
MgO is transported from storage tank 8 into operational tank 10 from
which it is distributed into suspension tank 11 which overflows into
circulation tank 4.
Energy, labor, sorbent
Calcined magnesite (75% MgO)
HzO
Gas
Product (MgSOj)
Labor
720kg .h"1
800kg .h"1
120 Nm3 .h"1
200kg .h"1
2 workers per shift
Product
MgS03 . 3H20 82%
MgS04 7%
MgO 2%
Inerts 8%
Fly ash 1%
This product is an excellent fertilizer for Mg-deprived soils. Tests
done at Ministry of Agriculture and Ministry of Health in
Czechoslovakia proved that this fertilizer is comparable to or better
than other Mg fertilizers (Kieserit, MgS04) . During testing period
content of heavy metals have been found acceptable to Czechoslovak
regulations.
7-287
-------
Content.
Metal
Czech reoulations
MoSo3fertilizer
As
10
1.5
Cd
50
1.5
Cr
200
5
Hg
2
0.1
Pb
30
15
The open Magnesite technology doesn't produce any waste products, is
very flexible.- and production of fertilizers lowers the operational
cost.
7-288
-------
WATER
//W///W//
nm
Figure
WATER
-------
Intentionally Blank Page
7-290
-------
SO, AND PARTICULATE EMISSIONS REDUCTION IN A PULVERIZED COAL UTILITY BOILER
THROUGH NATURAL GAS COFIRING
K. J. Clark
Aptech Engineering Services, Inc.
Post Office Box 3440
Sunnyvale, California 94089
T. S. Torbov
R. J. Impey
T. D. Burnea
7-291
-------
Intentionally Blank Page
7-292
-------
ABSTRACT
A field test was conducted in the fall of 1989 to evaluate cofiring of gas is a 450 MW coal-fired utility boiler.
The boiler was commissioned by Public Service Company of Oklahoma in 1979 and is a tangentially-fired,
supercritical unit. It normally burns a blend of Western subbhuminous and Oklahoma bituminous coals. The
main purpose of the tests was through natural gas cofiring to improve the boiler operation - reduce slagging and
fouling and provide added life for critical boiler components. Stack emissions reduction was not the ""in goal,
but was also under consideration.
During the tests, substantial leveraged reductions of 30% for opacity and 15% for SO, were obtained for 10%
natural gas cofiring. The fuel gas was injected through existing comer gas burners, parallel to die main caol
flame. These gas burners were mid-level in the burner bank so that most of the flue gas from the coal flame
was either processed through or exposed to a higher temperature combustion zone.
A possible mechanism for the leveraged reduction of S02 and particulate emissions is suggested.
Comparison of tests with 100% coal firing and 90% coal with 10% natural gas cofiring showed a 280% increase
in the sulfate content in the flyash. This resulted in a leveraged reduction in flue gas SO.. The higher sulfate
concentration in the flue gas reduced flyash resistivity, thereby enhancing ESP performance.
It is recommended that additional long-term testing be pursued on several different types of boilers and coal to
determine bow to optimize the cofiring configuration for maximum site specific emissions reduction benefits.
The experience thus gained will facilitate future conversions to cofiring in utility boilers throughout die United
States wherever the overall economics are favorable.
7-293
-------
INTRODUCTION
In recent years the subject of cofiring gas with coal in boilers which usually operate on 100% coal has received
increasing attention. A major reason for this is that cofiring can be a simple, effective technique for reducing
stack emissions (including SO;, NO,, and particulates) by at least as much as the proportion of gas fired relative
to coal.
In late 1988, Public Service Company of Oklahoma (PSO) explored the use of cofiring mainly for operational
benefits to alleviate a slagging problem. Aptech Engineering Services, Inc. (APTECH) developed the project
and performed the management, test, data reduction and boiler performance analysis.
TEST SITE DESCRIPTION
Public Service Company of Oklahoma Northeastern Station (NES) is located in Oolagah, Oklahoma, SO miles
Northeast of Tulsa. Units 3 and 4 are two identical Combustion Engineering tangentially fired units rated at 450
MW, delivering 3,200,000 pounds of steam per hour at 3614 psi and 1005°F with single reheat. The boiler at
NES 4 possesses burners for a variety of fuels including coal, natural gas, No. 4 oil, and refuse. Since start-up
in 1979, however, both NES 3 and 4 have been operated exclusively on natural gas for start-up and warm-up
and a coal blend for base-load service. Number 4 oil and refuse have never been fired in these units. There are
five elevations of coal burners, with each set of four burners at a given elevation supplied by a separate mill.
There are four levels of gas burners, and each level is situated between a pair of coal burner levels. Also at
these four levels, but on the wall adjacent to the gas burners, are located four levels of gas ignhors. Finally,
one elevation of gas warm-up guns is provided at the bottom of the burner bank. Regarding emission control
equipment, these units are supplied with overfire air for NO, control, a cold precipitator for particulates, and
a single common stack. The coal blend consists of 90% Wyoming subbituminous and 10% Oklahoma bituminous
on an annualized basis. Throughout the year, however, the coal blend can vary from 0% to 30% Oklahoma coal.
Tables 1 and 2 summarize the typical properties of Western subbituminous coal and Oklahoma bituminous coal
used at NES 3 and 4, based upon coal analyses performed in the summer of 1989.
TEST RESULTS
Several different test series were conducted. The Baseline Performance Test was performed to characterize and
optimize the unit on 100% coal. The coal blend was 100% Western, 0% Oklahoma. The data from this test
were used for comparison and evaluation of further measurable improvement (or degradations) during subsequent
gas cofiring.
7-294
-------
The Exploratory Cofiring Test Series was a sequence of short tests to screen a variety of firing configurations.
The following optimum configuration was determined: 4.5% to 5.0% excess 65% overfxre air;
windbox/furaace pressure differential of 7 to 8 inches H.O; coal burner levels B-C-D-E in service with no mill
bias, level A (bottom level) out of service; gas cofiring through gas burners between levels C and D; burner tflts
on automatic. The gas cofired was tested in the range of 9% to 10% and only one level of four gas burners were
used to keep the gas injection momentum sufficiently high to avoid the formation of "lazy flame."
This optimum was based upon the following considerations:
• Avoid clinker formation on the ash hopper slopes
• Minimize and potentially reduce waterwall slagging and convection pass fouling
• Maintain a "tight fireball"
• Maintain bulk superheated and reheated steam temperatures
• Do not increase stack opacity
• Do not increase superheater and r eh eater tube metal temperatures
• Keep percentage of excess Oj at a minimum consistent with the above
Two long term test series were conducted. The Cofiring Test series consisted of four continuous days of
cofiring, when about 10% as a heat input natural gas was burned through the gas burners between Levels C
and D. A blend of 26% Oklahoma coal with 74% Western Coal was used.
The second long-term test series was the 100% Coal Test which also has a duration of four continuous days.
The same coal blend was used for this second long-term test.
STACK SULFUR DIOXIDE AND NITROGEN OXIDES MEASUREMENTS
Figure 1 shows the stack sulfur dioxide and nitrogen oxides measurements each hour throughout the 24-hour
period for November 5, and 16, 1989. Plotted are the actual hourly measurements, not the rolling three-hour
averages. During full load, 100% coal firing, the average concentration of SO. is approximately 1.0 lb/MMBtu,
while for full-load cofiring with 10% natural gas, it is approximately 0.82, representing an 18% reduction. This
is typical of the results for all long-term cofiring testing.
7-295
-------
Table 3 provides a more detailed analysis of the S02 data chained during the long-term testing. The daily
average S03 concentration for all times during the day for which the boiler was at full-load is tabulated for long-
term 100% coal firing and cofiring. As indicated, percent reductions in S02 as high as 28% were obtained with
no more than 13.3% gas cofiring.
Figure 2 graphically displays the leveraged S03 reductions obtained during the long-term cofiring testing.
STACK OPACITY MEASUREMENTS
Figure 3 shows the stack opacity measurements each hour throughout the 24-hour period for November 5 and
November 16, 1989. Each hourly measurement is actually an average of ten instantaneous measurements
recorded, one every six minutes, for that hour. As indicated, the mean during full-load 100% coal firing was
9.9% opacity, while during full-load cofiring, it was S.5% opacity.
This represents a 44% reduction. The average opacity reduction for the entire long-term cofiring test period was
35% relative to the average opacity for the long-term 100% coal firing test.
Table 4 provides a more detailed analysis of the opacity data obtained during the long-term tests. Three different
opacity values are shown for each day of the long-term cofiring tests. The first two opacity values are the
instantaneous values for 10:00 and 16:00. These different times of the day are of interest because the precipitator
performance is influenced by the ambient temperature level. The third opacity value is the average value for all
times during the day for which the boiler was at full-load (including both 10:00 and 16:00). The opacity
reductions resulting from cofiring for 10:00, 16:00, and full-load daily average are tabulated in Table 6.
Figure 4 graphically illustrates the leveraged opacity reductions achieved from cofiring for 10:00, 16:00, and
die full-load daily average. As indicated, opacity reductions of almost 50% were obtained with no more than
13.3% gas cofiring.
TEST DATA INTERPRETATION
Introduction
In conducting this data interpretation, selected long-term test data from both gas cofiring and 100% coal firing
have been considered. The emphasis has been to compare the results of the gas cofiring to those for 100% coal
firing. Specifically, an attempt has been made to explain the changes in furnace temperatures and stack emissions
that were observed as a result of going from 100% coal firing to gas cofiring while holding all other operational
variables constant.
7-2%
-------
Firebox Flue Gas Temperature Evaluation
As a first step in interpretation of the test data, the firebox flue gas temperature distribution was evaluated. A
one-dimensional heat transfer model was used to model the profile of bulk gas average temperature at different
firebox horizontal cross sections. This one-dimensional model gives integrated combustion and heat transfer data
at selected horizontal cross sections of the furnace which are relatively easy to interpret and relate to the coal
properties and NO» and SOm formation.
Calculations of the flue gas bulk temperatures at the top of each of the three zones were performed with die one-
dimensional model using die fundamental coal properties obtained from the analysis of coal samples taken in
September prior to the test program. Table 5 summarizes the results of the ultimate analysis for the Western
coal being used at NES 4 at that time. Table 6 presents an "equivalent" ultimate analysis for a fuel comprised
of 90% Western coal and 10% natural gas.
Figure 5 presents the firebox bulk gas temperature profile predicted by the one-dimensional model for 100% coal,
and that for 90% coal and 10% gas. The model predicts essentially no change in bulk gas temperature at the
furnace exit (top of Zone 3) and the bottom of the bull nose (top of Zone 2) due to gas cofiring. However, at
the top of the combustion zone (Zone 1) and in the combustion zone, the bulk gas temperature is S0°F to 100°F
higher when cofiring. Specifically, at the top of Zone 1 just above the burners, the bulk gas temperature is 54°F
higher with 10% cofiring, while down lower the flame temperature with heat transfer is 66°F higher while die
adiabatic flame temperature is 94°F higher. Of course, in the very localized region where the natural gas is
injected (Elevation CD gas burners), there are zones of much higher temperature through which the coal
combustion products are processed.
SO, Emissions
The organic and pyritic sulfur in coal oxidizes to S02 and SO, in the combustion process. Only a small portion
of this sulfur is converted to SO, during combustion. Usually, for pulverized coal boilers, less than 1% of the
sulfur is converted to S03 in the flue gas. However, in the flame (Zone 1) up to 5% to 12% of the available
sulfur in the coal is oxidized to SO}.
7-297
-------
The oxidation of SOj from molecular oxygen can be written as:
2SOz + 02-2S0j
Hie equilibrium constant of this reaction is:
log K„ = 1C^73 - 1.222
The equilibrium relationship between SO, and SO, is shown in Figure 6 for various conditions of temperature
and molecular oxygen availability.
This reaction takes place in a temperature range below 1600°F and could be promoted by a catalyst. Hie
chemical kinetics for SO, formation show that with a temperature increase, the role of molecular oxygen as an
oxidizing agent is reduced and that of atomic oxygen is increased. Hie most probable reaction is:
so2 * lo2-so3
Figure 7 shows the conversion rate of S02 oxidation to SO, under adiabatic conditions. It also shows the atomic
oxygen as a fraction of total oxygen as a function of excess air. This graph indicates that SO, concentration
increases as atomic oxygen concentration decreases. This is due to the effect of temperature and also explains
why after excess air reaches a certain value, the curve for SO. conversion rate begins to level off.
The effect of flame temperature on die SO, formation is shown in Figure 8. The amount of SO, formed increases
with the flame temperature and is attributed to the greater concentration of atomic oxygen at higher temperatures.
The curve leveling off at 3200°F is not well explained, but probably is due to die kinetic mechanisms of die
reaction involving SO: and 0: which also takes place. The equilibrium residence time of the reaction is about
200 msec. From experiments with different furnaces, it has been established that a shorter residence time of the
gases in a higher temperature zone is favorable for producdon of SO,.
When coal is burned, CaCO, and MgCO, are converted to CaO and MgO through a calcination process at high
temperature conditions. Simultaneously, in the furnace CaO and MgO react with SO: and SO, in the presencei
of 02 forming CaSO« and MgS04. The reactions with SOj are less probable than with SO, because of the lower
7-298
-------
free energy available after the reactions are completed. This has also been established in experimental and full
scale facilities for sorbent (CaCOj.MgCOJ injection with coal burning for SO2 reduction and oQ burning for SO,
reduction in the flue gas.
For the long-term 100% coal and cofiring tests, a sulfur balance was performed in order to determine the sulfur
distribution in the combustion products. The results of the sulfur chemical analysis are shown in Table 7. The
sulfur balance for both long-term tests is shown in Figures 9 and 10. From these results it is concluded that the
fly ash in the cofiring test is significantly (approximately 50%) enriched in sulfur.
The sulfur reduction in the flue gas predicted from the sulfur balance is comparable with the reduction indicated
from the direct flue gas SO: measurement data (Figure 11). The small difference is most likely a result of
limitations on sample collection. These samples are not averaged from continuous sampling, but from samples
extracted every four hours.
The results correspond well with the theory of SO, formation presented in the beginning of this section. For the
NES 4 cofiring configuration, the natural gas was burned at the center of the coal flame, thus allowing coal
combustion products to be processed in a higher temperature zone for a very short residence time. This promotes
SO, formation in Zone 1 and in parallel, a sulfation process with CaO and MgO in the ash particles. In
summary, this causes the SO, reduction in the flue gas to be higher than that expected due to the fuel sulfur
reduction because 10% of the coal is replaced by natural gas. With special process designs and combustion
process fine tuning, it is possible that this "leveraged" reduction in SO: can be even further increased.
PARTICULATE EMISSIONS AND STACK OPACITY
The flue gas opacity in the stack was measured during both 100% coal and cofiring tests with a Lear Sieglar
RM41 opacity meter. Calibration of the instrument was performed eight months before the tests, and after
adjustment, the maximum error was in the range of 2% to 3%.
Since direct measurement of the particulate mass concentration via EPA Method 5 was not performed, the
particle mass concentration was estimated based on experimental data relating mass concentration to light
attenuation in the stack gas of a coal-fired boiler with ESP control. The experimental data for two different
plants are shown in Figure 12. This graph shows significant differences between the mass concentrations for
the same measured opacity for Plants A and B. This is due to differences in the particulate properties as well
as possible errors in the determination of the mass concentration through the measurement technique.
7-299
-------
Table 8 shows the results from the tests conducted in this project where the particulate emissions level, as
indicated by opacity, is reduced over 30% when the particle loading input is reduced by only 10%. The
estimates for particle mass concentrations reflect the range of the test data in Figure 12.
Generally, the level of particulate emissions is direcdy related to die electrostatic precipitator efficiency. The
optimum ESP performance depends on the electrical resistivity of the flyash which should be high enough to
result in electrostatic pinning of particles to the collecting surface, but not so high that electric discharge through
the layer occurs as the corona current passes through it. The optimum flyash resistivity range measured in situ
is considered to be 10* to 10'° ohm cm.
Flue gas produced by combustion of coals with low sulfur content in the range of 0.5% to 1% and at
temperatures between 250°F to 300°F has a flyash resistivity higher than 10" ohm cm, which is significantly
above the desired 10'° ohm cm upper limit (Figure 13).
With the natural gas cofiring tests conducted here, the sulfate content in the flyash increased to 0.42% from
0.11% when 100% coal was fired. This results in lower electrical resistivity of the flyash and higher ESP
current which improves ESP collection efficiency. Thus, a highly leveraged opacity reduction results because
the flyash loading to the ESP is reduced by 10% when 10% of the coal is replaced with gas, and the collection
efficiency for this flue gas with already reduced particle concentration is significandy increased.
CONCLUSIONS AND RECOMMENDATIONS
Overall, the results of this test program indicate that gas cofiring offers significant technical promise as a
practical means for reducing stack SO. and particulate emissions. The circumstances surrounding this project
should be kept in mind. The field testing was performed under the utility's normal operating conditions (except
for gas cofiring) and only routine plant instrumentation was used. Also, the cofiring testing involved a specific
boiler design, a specific coal, and a specific cofiring configuration. Nevertheless, the very significant leveraged
reductions in S02 and opacity were obtained as reported here. The emissions data obtained in these field tests
were consistent and repeatable, and the leverage trends are outside the estimated data uncertainty bands.
7-300
-------
With the caveat that the majority of the cofiring testing was limited to a maximum of 13% gas at 100% load,
and the test was not designed specifically for stack emissions reduction, the following is the major technical
conclusion:
• Cofiring at NES showed substantial promise as an emissions reduction technique,
particularly for opacity and S02. Substantial leveraged reductions were obtained with 13%
cofiring: over 40% for opacity and over 20% for S02 were typical results. These leveraged
reductions were attributed to significantly greater sulfur retention in the flyash during
cofiring: sulfates content increased almost 400%, and sulfur content increased 16%. This
was speculated to be a result of slightly higher combustion temperatures during cofiring
(predicted not measured) which increased the conversion of SO: to SO„ the latter being
more reactive with flyash mineral matter. The increased sulfate content lowers the flyash
resistivity and improves the ESP collection efficiency. These conclusions are supported by
a detailed sulfur balance. The higher combustion temperatures promoted the formation of
thermal NO, which offset the inherently lower fuel NO,, so that total NO, concentrations
for 100% coal firing and cofiring were essentially the same. This suggests that it will be
difficult to simultaneously lower both SO: and NO, in basic cofiring configuration similar
to that available in the NES 4 boiler. Cofiring could be extremely cost beneficial in
reducing emissions to remain in compliance without incurring very large capital
expenditures for scrubbers and ESP or baghouses.
Concerning cofiring in coal-fired boilers in general, the following recommendations are made:
• Long-term cofiring tests should be conducted with a variety of boiler designs, loads, coal
properties, firing configurations, and gas percentages.
• The results of these tests can be used to develop a cofiring field test data base from which
site-specific technical guidelines can be developed for new users. These guidelines would
tell users how to optimize cofiring for their desired site-specific benefits.
7-301
-------
ACKNOWLEDGEMENTS
The financial support of the Gas Research Institute (GRI) and the assistance of Mr. John Piatapas of GRI
throughout the project are gratefully acknowledged. The support of Sensor and Simulation Products, Inc., and
Scientific Engineering Instruments, Inc., in providing their expertise in infrared videocameras and acoustic
pyrometers, respectively, is gratefully acknowledged. Last, but definitely not least, numerous people at Public
Service Company of Oklahoma, Northeastern Station, spent many hours supporting this project in the areas of
instrumentation and testing. Their efforts and cooperation are recognized to have made this project possible.
REFERENCES
1. Lange, H. B., R. C. Booth, and M. F. Szabe, "Cofiring Gas on Coal-fired Utility Boilers-
Summary of Benefits and Review of Operating Experiences," Topical Report, Gas Research
Institute Contract 5086-251-1232 (June 1986).
2. Booth, R. C., B. P. Breen, C. A. Gallaer, and R. W. Glickert, "Natural Gas/Pulverized
Coal Cofiring Performance Testing at an Electric Utility Boiler," Topical Report, Gas
Research Institute Contract 5086-251-1232 (June - October 1986).
3. Booth, R. C. and R. W. Glickert, "Extended Development of Gas Cofiring to Reduce
Sulfur Dioxide and Nitric Oxide Emissions From a Tangentially Coal-fired Utility Boiler,"
Topical Report, Gas Research Institute Contract 5086-251-1232 (August 1988).
4. Oglesby and Nichols, "A Manual of Electrostatic Precipitator Technology, Part n,"
Southern Research Institute (1970).
5 Crumley, P. H. and A. W. Fletcher, "The Formation of Sulfur Trioxide in Flue Gases,"
Journal Institute of Fuel, V.29, N.187 (1956).
7-302
-------
Table 1
COAL PROPERTIES. WESTERN COAL
Moisture
Ash
Volatile Matter
Fixed Cuboo
Sol fur
BTU/Ib
Ulnrrurr Analysts
Carbon
Hydrogen
Nitrogen
Sulfur
Ash
Oxygen
Total Moisture
As-Received
15 64*
5.47*
38 85*
40.04*
0.4S*
10,068
58.83*
4.27*
0.88
0.45*
5.47*
14.46*
15.64*
Moisture free
6.48*
46.05*
47.47*
0.53*
11.934*
69.74*
5 06*
1.05*
0_53*
6.48*
17.14*
Table 2
COAL PROPERTIES. OKLAHOMA COAL
Moisture
Ash
Volatile Maner
Fixed Carbon
Sulfur
BTU/lb
As-Received
10.38*
12.75*
28.57*
48.30*
0J5*
10.984
Moisture Free
14.22*
31.88*
53.90*
0.39*
12.256
Ultimate Analysis
Carbon
Hydrogen
Nitrogen
Sulfur
Ash
Oxygen
Total Moisture
63.81*
3.84*
1.45*
0.35*
12.75*
7.42*
10.38*
71.20*
4.29*
1.62*
0.39*
14.22*
8.28*
7-303
-------
Table 3
SO; DATA ANALYSIS
1. 100% Coal (Test 6, Long-Term 100% Coal Firing)
Date
11/3/89
11/4/89
11/5/89
11/6/89
Daily Average
2. Gas Cofiring (Test 4, Long-Term Cofiring)
Date
Percent Gas
Averape SO,
1/16/89
11.0
0.87
1/17/89
10.5
0.84
1/18/89
9.9
0.90
1/18/89
13.3
0.82
1/19/89
13.3
0.74
1/19/89
9.0
0.85
Average SO,
-------
Table 4
OPACITY DATA ANALYSIS
1. 100% Coal (Test 6, Long-Term 100% Coal Firing)
Average Opacity
Date
Ooacitv ffl 10:00
Ooacitv © 16:00
<® Full-Load
11/3/89
8.8
9.7
9.1
11/4/89
8.3
10.8
9.2
11/5/89
10.3
12.2
9.9
11/6/91
8.7
10.8
8.9
Daily Average
9.0
10.9
9.3
2. Gas Cofiring (Test 4, Long-Term Cofiring)
Date
Percent Gas
Ooacitv <® 10:00
Ooacitv © 16:00
Average Opacity
© Full-Load
11/16/89
11.0
5.8
6.2
5.9
(35.6)
(43.1)
(36.6)
11/17/89
10.5
6.0
7.5
6.2
(33.3)
(31.1)
(33.3)
11/18/89
9.9
5.1
—
6.3
(43.3)
(32.3)
11/18/89
13.3
—
5.6
5.4
(48.6)
(41.9)
11/19/89
13.3
6.8
—
4.9
(24.4)
(47.3)
11/19/89
9.0
—
7.4
7.3
(32.1)
(21.5)
() = Percent opacity reduction relative to daily average opacity at full-load on 10096 coal
Note: All opacity values are percent opacity
7-305
-------
TableS
FUEL CHARACTERISTICS - 100% WESTERN COAL
ULTIMATE ANALYSIS FROM COAL
SAMPLE TAKEN SEPTEMBER 1989
C 58.83%
H 4.27%
S 0.45%
N 0.88%
O 14.46%
A 5.47%
W 15.64%
Total 100%
LHV - 10068 Btu/lb
Stoichiometric combustion air = 93.56 scft/lb
Stoichiometric flue gas = 103.81 scft/lb
7-306
-------
Table 6
FUEL CHARACTERISTICS - 90% WESTERN COAL, 10% NATURAL GAS
Natural Gas
CH. - 98.9%
CjH. - 0.3%
011,-0.1%
CJI„,-0.1%
Na - 0.4%
CO,-0.2%
Total 100%
LHV - 897 Btu/scft
S - 0.0445/Ib/cft
EQUIVALENT ULTIMATE ANALYSIS
C - 58.90%
H - 5.96%
S- 0.43%
N - 0.88%
O - 14.76%
A - 4.92%
W - 14.18%
Total 100%
LHV = 11227 Btu/lb
Stoichiometric combustion air - 100.77 scft/lb
Stoichiometric flue gas - 1663.04 scft/lb
Equivalent in Mass %
C-74.31%
H - 24.62%
O - 0.39%
N - 0.70%
Total 100%
7-307
-------
Table 7
FORMS OF SULFUR IN THE COAL AND COMBUSTION PRODUCTS
1st
100% Coal
11/14/89
Simple
(As-
Rcceived^
Coal
Fly ash
Bottom ash
Pvritie
0.11
0.37
0.09
Forms of Sulfur *
Sulfas
0.01
0.11
0.06
Organic
0.39
0.0S
0.12
Tool Sulfaf Untamed
i *
0.51
0.56
0.27
0.08
0.15
Cofiring
11/16/89
Coal 0.11 0.01 0.39 0.51 —
Fly ash 0.17 0.42 0.06 0.65 0.08
Bottom ash 0.09 0.17 0.01 0.27 0.70
Table 8
COMPARISON OF PARTICULATE EMISSIONS
Iss
100% Coal
Cofiring - 90% Coal
10% Natural Gas
Reduction Percent
Permit Qwog
9.6
6.5
32.3
Fgimawwl Panicle Mass
CoDcsaraboo mg/m1
Average/Rang^
103
176-30
76
134- 19
26.2
23.9 - 36.7
7-308
-------
rr-4 *4 - £^.ss.cns (SO JNOX)
000
Figure 1A - Sack Sal far Dioxide and Nitrogen Oxides Emissions. Long-Tenn 100% Coal Firing
re\ 4* - E^.asions (SOxMOX)
tl/U/ft
O 90
OSO
0 60
Figure IB * Sack Sulfur Dioxide and Nitrogen Oxides Emissions, Lotlfc-Tenn Co firing
7-309
-------
30-
% Gas Cofiring
Figure 2 - Leveraged SO, Reductions Achieved From Long-Term Cofiring.
7-310
-------
P.S.O. N.U. •A UNIT OPACITY
li-S'W
T
1
—
1 T
** * •
•
:
, sm of* i
7
j
i 1 ;
1 y
r • m ;
i 1 j
1 i ^
i ¦ \ 'i/~
r-»
i ^ r'
• I !
1 »"*
Vi 1
1
1 :
1
•
1
ru iae
• «
1
i i
' ' i I !
1 ' 1 : 1 !
-1 1 1 1 1 1 1 1 1
O 4 I 12 11 S y
NMorvr
Figure 3A - Sock Oparicy, Long-Tenn 100% Coal Firing.
P.S.O. N.C.S. UNIT OPACITY
ii/u/n
1 i ! 1 1 1
! «M \
1 5TO. oc*. 0
_L .
i ! 1 i
\
i
I
i
i
i
•
.
1
r" ,"
1
j
K J
r"
V i
mi lm
i
—i—
1
1 1
• 4 • tt m a u
MVMT
Figure 3B - Sack Opacity. Long-Term Cofiriag.
7-311
-------
% Opacity
Reduction
® 10:00
% Opacity
Reduction
@ 16:00
% Opacity
Reduction, Ave.
For All Times
During Day At
Full Load
40-
20-'
40--
20-
O
<§>
o
Proportional
10
Proportional
15
c#
o
Proportional
% Gas Cofiring
Figure 4 - Leveraged Opacity Reductions Achieved From Long-Term Cofiring.
7-312
-------
Pyrosonics 2000
ygg J Measurement
778 I
666"
flane
1001 Coal
2404
2467
90' Coal
10- Gas
2401
2465
2606
2660
radiabat1c 3650
2853
3744
2919
All Temperatures 1n °F
Figure 5 - Firebox Temperature Profile.
7-313
-------
100
80
e
«>
w
&
60
2 40
E
S
20
N
\
..
w
r-nm2S0
—«¦ 100
ml
\
\ P*>J
\
-00008 itr
-a0004 atr
11
n
0
400
800
1200 1600
Temperature, *F
2000
2400
Figure 6 - Equilibrium Yields of SO, for Ibe Eqtuboa 2SOj •+• O, ¦ 2SOh-
*
8'
10"'
ia*
no*
10-
10*
a
X
K
c
i
/"
/
X
/
/
X
/
\
N
/
1
0.6 0.8 1.0 1.2 1.4
StOKNomctnc Ratio Sft
Figure 1 - Equilibrium Conversion of SO, to SO, and Atomic Oxygen Conceatrsaoo at a Function
of (be Stoichiometric Rjbo.
7-314
-------
4)
U
S>
O.
s
CO
0.008
0.006
0.004
0.002
0
•-
* 4
¦
X
X
9 J
Kjp
~
• Ktroa
x Distill
tne; sutfi
tte fud o
ir contant 2%. *6
il; natural sulfur c
dcd asa
ontent 3
trbon dis
%
ulfidt
2800 2900 3000 3100 3200 3300 3400 3500 3600 3700
Flame temperature, *F
Figure 8 - Variation in SO, Content of Flue Gases Containing 12% C02 With Flame Temperature.
Flue gas SOx
S—-0.00478 lb
93.7%
Fuel Input
0.0051 lb-
100%
Fly ash
"0.0003 lb
5.9%
Bottom ash
S—0.00002 lb
0.4%
Figure 9 - Sulfur Balance (Per Pound Fuel) 100% Coal (November 4, 1989).
7-315
-------
Flue gas SO,
S—0.00413 lb
89.96
Fuel Input
0.00459 lb
100%
Fly ash
~0.00044 lb
9.6%
Bottom ash
S —0.00002 lb
0.44%
Figure 10 - Sulfur Balance (Per Pound Fuel) 90% Coal, 10% Gas (November 16, 1989).
FROM SOx MEASUREMENT
100% coal (November 4,1989) - SO, = 1.04 Ib/MBtu
90% coal, 10% natural gas (November 16,1989) - SO,
= 0.88 Ib/MBtu
SO. reduction -15.4%
FROM SULFUR BALANCE (Assumptions tor coal ash distribution, 90% as a fly ash
and 100% as a bottom ash)
SO, reduction -13.6%
Figure 11 - SO. Emissions Reduction.
7-316
-------
I
i
•/.
• /'
'•r
« 10 20
1 M|*(» IT*;* ,v
Figure 12 ¦
Mm CooccBrarion Versus LigbtAseaioBof Particulate Emissions s Coal-Fired Power
Plans Wsb ESP Controls
10"
10"°
«0*
200
;/
1 1
1 1
V
\
—T
.0.5-1%
/
/
/ ^
1 5 -2%
. I" *•*« ^
m*oit* r*mt«n
~ 1 2% S
o 0.8% S
A 2.3% 5
• 2 9% 5
x 2.5% S
\\
2 5
\ >
\ \
\ N
% \
v \ N
-3%-^'
\
%
250
500 550
T#^W0Tyff. t
«00
•50
Figure 13 ¦
Trends in Resistivity of Ftyxsh With Variations in Flue Gas Temperatures and Coil
Sulfur Content.
7-317
-------
Intentionally Blank Page
7-318
-------
Design, Installation, and Operation of the
First Wet FGD for a Lignite-Fired Boiler in Europe at
330 MW P/S Voitsberg 3 in Austria
H. Kropfitsch
ODK
A-9020 Klagenfurt
Austria
M. Weitzer
W. Pursch
Waagner-Biro
A-8021 Graz
Austria
A. Saleem
GE Environmental Systems
Lebanon, PA
7-319
-------
Intentionally Blank Page
7-320
-------
ABSTRACT
The FGD has been designed and constructed by Messrs. Waagner-Biro and incorporates
the IFO-process of GEESI. The main features of the IFO-process (= Insitu Forced
Oxidation) are the simultaneous removal of SO2 and production of gypsum in the same
vessel as well as the sophisticated limestone feed control which is essential to cope with
the fluctuating S02-load in the range of 4000 to 7000 mg/m3 typical for lignite. The
Voitsberg installation was the first wet FGD appplication for lignite in Europe.
The FGD is designed in two lines each rating 50 % of the total flue gas volume without
spare vessels. Approx. 50,000 tons of gypsum have been produced annually and had
been dumped in the exhausted strip mine nearby. Since 1990 100 % of the gypsum
produced has been made use of in the cement industry. Further utilization possibilities are
investigated and tested at present (e.g. high-grade sealing of disposals by residues from
lignite fired P/S). Due to the specific site conditions no waste water is produced.
The desulphurization plant has been in operation for approx. 20,000 hours and has fulfilled
all expectations regarding SO2 reduction and reliability. Typical operating data are shown
in the paper. The power plant availability never was influenced by the desulphurization
plant Operation and maintenance experience will be presented. The total project costs for
the desulphurization plant were approx. 70 million US-$.
7-321
-------
1.
General
2. Furnace Limestone Injection (FLI)
3. Wet Limestone Scrubbing (FGD)
3.1 Process Design and Arrangement
3.2 Operating Experience
4. NOx Reduction (DeNOx-plant)
5. Conclusion
6. Figures
7-322
-------
1. GENERAL
The Osterreichische Draukraftwerke AG (short ODK), founded in 1947 with the
main duty of developing the hydraulic power of the river Drau, decided to build the
330 MW Thermal Power Station of Voitsberg 3 (Fig. 1) on the basis of a long-term
delivery contract for local lignite in July 1977.
The proven resources of 31 million tons of lignite have am ?. v^ge calorific value c<
approx. 10 500 kJ/kg and a sulphur content in the range of 0.5 to 1.5 % and is
open-pit mined. This quantity is sufficient to guarantee the fuel supply of this unit-
type power station for the duration of its actual life service (Fig. 1).
Voitsberg 3 has a once-through boiler with a steam generating capacity of ?30 t/h
and a triple-cased 330-MW-turbine.
The civil engineering and erection works for the plant started in autumn 1977 and
were fundamentally completed by the commissioning date of the unit in Apri.'1983.
Voitsberg 3 Power Plant is designed for approx. 4000 design load hours/year. A
substantial portion of the total investment costs, i.e. approx. 1,6 billion ATS, has
been spent for environmental precautions (ESP, LSI, FGD. DeNOx) Also see
fig. 2.
2. FURNACE LIMESTONE INJECTON (FLf)
With this process limestone powder is injected into the furnace where it is calcined
to CaO which reacts with the SO2 to CaS04 (Fig. 4). The desulphurization product
and the non-reacted additive are collected in the precipitator together with the fly
ash. Testing numerous different injection points at different lignite fired plants, the
limestone injection together with the combustion air brought the best results as the
limestone is injected into a favourable temperature region and there is enough
pulse with the overfire air of the burner to distribute the additive.
The plant was initially designed without an FGD. In order to fulfill emission
requirements FLI was applied from the very beginning of commercial operation in
April 1983 as an intermediate solution. The process was in operation for approx.
13,000 hours (Fig. 5). Depending on the boiler load a removal efficiency of 50 to
70 % was achieved. To reach this efficiency at design load a molar ratio Ca/S of
approx. 3.5 was required. There were no reductions in plant availability when using
the FLI process.
7-323
-------
Advantages (+) and disadvantages (-) of this process are:
+ simple process, quick realization
+ low investment costs
+ low energy consumption
- limited removal of SO2
- tendency of slagging in the boiler
- more difficult handling and disposal of ashes
3. WET LIMESTONE SCRUBBING (FGD)
To meet the new emission limits a wet limestone scrubbing process was retrofitted
to Voitsberg 3. The FU is now a stand-by to reduce S02-peaks from sudden
changes of the sulphur content in the coal. Up to now this measure has not been
used due to the excellent removal capacity of the wet-type FGD.
3.1 Process Design and Arrangement
The FGD has been designed and constructed by Messrs. Waagner-Biro and
incorporates the GEESI-IFO-Process (Fig 3). The main features of the IFO-
process (= Insitu Forced Oxidation) are the simultaneous removal of SO2 and
production of gypsum in the same vessel as well as the sophisticated limestone
feed control which is essential to cope with the fluctuating S02-load typical for
lignite. The Voitsberg installation was the first wet FGD appplication for lignite in
Europe.
The FGD is designed in two lines each rating 50 % of the total flue gas volume. The
flue gases discharged from the stack are led via two booster fans into the spray
towers. Upstream of spray tower 1 a gas/gas heat exchanger is installed which
cools down the flue gas before entering the spray tower and heats up the cleaned
gas to approx. 100°C. In line 2 the flue gas is directly led into the spray tower. The
cleaned and heated gas from spray tower 1 is mixed with the cleaned and unhealed
gas from spray tower 2. The gases leave the stack at a mixing temperature of
approx. 80°C (Fig. 7a, 7b).
-------
The FGD product (gypsum, Fig. 8) is thickened up to 30 % solid content and
subsequently dewatered to 15 % moisture in a vacuum drum filter. Thickener
overflow and filtrate are completely recycled to the process. No waste water is
produced. Up to now approx. 25 % of the gypsum is utilized in the cement industry,
the remaining part has been dumped in the exhausted coal mines (surface mining)
together with the fly ash. Since 1990 100 % of the gypsum cake has been made
use of in the cement industry.
3-2 Operating experience (Fig. 6)
In the following the experiences made during commissioning and the acceptance
tests as well as during commercial operation will be presented.
The acceptance test measurements earned out by an officially authorized
measuring institute showed that all guarantee values for the first large scale FGD
working on the basis of limestone for a lignite fired power station in Europe could be
fulfilled. Only in case of spray tower 2 there were slight excess values in mist carry-
over.
In all details the problems that occurred during the operation of the plant and that
mainly were due to the method of boiler operation and the use of new materials of
construction as well as the failure of less important components will be presented
below together with the solutions.
Mist eliminator
As can be seen in fig. 9 the mist eliminator was designed as a 2-stage-unit where
the uppermost blades (at the clean gas outlet of the spray tower) initially was set in
opposite direction to the flow. However, as in accordance with the contract the mist
eliminator was designed for a wet flue gas capacity of 2 x 625,000 stm3/h, wet and
as due to a variety of modifications in the operation of the boiler the steam
generator produced an approximate amount of 2 x 685,000 stm3/h wet flue gas or
even more the mist eliminator clearly had to operate at the design limit
This means that, as the flue gas velocity over the mist eliminator surface never has
been constant there were local overloads that resulted in the following carry-over in
the flue gas mainly but not only during the washing of the mist eliminator. The
measurements were made by TUV Hamburg acc. to the wet soot particles method.
7-325
-------
Spray tower 1, design load
Spray tower 1, overload
Spray tower 2, design load
Spray tower 2, overload
Guarantee value, design load
M.e. carry-over
mg/m3
greater than 34
greater than 100
greater than 110
greater than 1000
max. 100
The first possible means of improvement consisted in turning the mist eliminator
packages so that the end blade of the uppermost mist eliminator was oriented
toward the flue gas outlet of the spray tower. In addition catchment pockets were
installed in the clean gas ducts.
Acceptance test measurements made then by Prof. Wurz showed the following
results. At the design value the flue gas capacity upstream the critical components
(gas reheater, clean gas damper, etc.) clearly was below the guarantee value and
even in case of am overload of 10 % the amount of carry-over was acceptable.
M.e. carry-over
mg/m3
Spray tower 1, design load
greater than 5
Spray tower 1, overload
greater than 28
Spray tower 2, design load
greater than 94
Spray tower 2, overload
greater than 153
Unlna of the flue pas ducts
The flue gas ducts have been protected from the chemical attacks by clean gas by
means of a fiber glass reinforced lining.
7-326
-------
As Voitsberg 3 is a peak load power plant and frequently is stopped at weekends
(when energy consumption is low in all of Austria) the components containing flue
gases are subject to a very great number of changes from hot to cold states
entailing considerable variations in the lengths and the resulting stresses.
After an approximate 3,000 operating hours the lining manifested cracks reaching
deep down to the carbon steel surface. As these cracks occurred almost exclusively
in the comer and edge zones the reason of these defects soon became very clear,
unfortunately, only after the defect had occurred.
The lining had been applied by spraying and in the comer zones its thickness wets 4
to 5 mm instead of the usual thickness of 1 to 1.5 mm upon plane surfaces. The
glass fiber reinforced lining of that thickness could not bear the great number of
changes in the expansion coefficients and the permanent heat stresses without
cracking.
The defects were repaired by sandblasting the comers once again and applying the
lining by means of a spatula, resulting in a thickness of about 1.2 mm.
For the plants being presently under construction we use flue gas ducts of slightly
rounded corners (r = approx. 20 mm) with which we have made the best
experiences considering manufacturing, strength, and lining. This design permits
treating all flue gas duct areas by means of the faster and less expensive spraying
process.
For the flue gas ducts from the raw gas damper to the spray tower inlet a lining had
to be retrofitted. Initially that section had been made of carbon steel only and
because of the backflow of the wet and chemically aggressive vapours of the spray
tower into the flue gas ducts (during the plant outages) that section began corroding
within short.
During the plant outage this section incl. the gets reheater is dried by means of an
air dehumidifier in order to minimize the production of sulphurous and sulphuric
acid.
Oxidation air
As the process used for the present plant is the IFO-process where air - necessary
for the reaction of calcium sulphite into calcium sulphate - is added in the same
tank where SO2 is removed and calcium sulphite is formed. Hence each spray
7-327
-------
tower had been equipped with a rotary piston fan (Roots compressor} and a
distribution level of glass fiber reinforced pipes.
Due to their function these compressors caused several problems, mentioned
below.
All of the system vibrates (the frequency is twice the speed) and at all
'weak" points strong vibrations are produced.
Like organ pipes the vibrating column of air excites the water injection
cooler required for generating optimum conditions of reactions, i.e. a
temperature of approx. 50°C, for the oxidation air, the compression
temperature of which originally is at some 120°C. As a result of the
excitation of the injection cooler the maximum admissible inhouse
noise level of 85 dB(A) is markedly exceeded.
The air distribution system in the spray tower which is equally vibrating
manifests a substantial amount of wear due to the permanent
movements and the considerably "abrasive" slurry.
In order to solve these problems fixtures were installed at short distances and the
cooler (especially the one of spray tower 1) modified externally (i.e. shortened,
special design of the water evacuation system).
It is true that these measures eliminated the vibrations but the sound radiation of
the cooler had been improved only slightly. This problem was overcome by an
absorption silencer that shifted the vertical air column out of the resonance range.
For the attachments of the air distribution systems within the spray towers a variety
of solutions had been tried:
Rigid connections by means of attachments of glass fiber reinforced
material.
- Flexible connections by means of corrosion resistant rubber rings.
Rigid connections by means of a combination of glass fiber reinforced
bearings and attachment brackets of austenitic material.
We did not achieve a sufficient amount of continuous running hours using the first
two of the variants.
Up to now the last variant has proven to be rather effective. However, it is to be
underlined that the air pipes of glass fiber reinforced material themselves manifest
7-328
-------
signs of material fatigue after approx. 15,000 running hours due to the permanent
vibrations.
Hence, the conclusion to be drawn from the remarks made up to now is, that a
lasting solution, i.e. 8 to 10 years, can be a distribution system of austenitic material
(1.4539 or even better 1.4529).
Sound Insulation
As the local hospital is only 420 m from the power plant particular attention had to
be paid to sound reduction. Guarantee values of 50 dB(A) in 1 m from the contours
of the building are required.
Considering the fact that during a normal conversation sound power levels of up to
60 dB(A) may be registered this demand illustrates the high requirements to the
different sound insulating components and their detailed design in view of sound
reduction.
Some detailed calculations and check measurements, e.g. sound damping over the
complete stack height, were necessary in order to be able to install the proper and
exhaustive sound reduction systems without incum'ng expensive modifications
afterwards.
FGD operation without oxidation air
Due to a variety of experiences from other plants the taking out of service of the
oxidation air seemed possible considering the flue gas and the ash composition.
The following are the important parameters that had to be considered in this
context:
Oxygen content in the flue gas
pH-value: It determines the ratio of sulphite and bi-sulphite. As only
dissolved bi-sulphites can form oxides they finally have a decisive
influence on the oxidation rate.
Content of solids in gypsum slurry: The existence of superficially
active fine particles and the concentration of dissolved and
catalyticalty effective components like manganese or copper ions. The
results can already be found in the specialized literature.
7-329
-------
Spray tower sump level and number and/or position of the spray levels
in operation (retention time).
Droplet size distribution: specific exchange surface.
L/G = liquid/gas ratio.
Temperature conditions: Influence on thermodynamic equilibrums,
diffusion coefficients, kinetic processes, etc.
Fluid dynamics: They determine remixing effects and residence time
behaviour, above all the
oxygen mass transfer from the gas into the liquid phase. This transfer
is the slowest process phase and hence it determines the overall
process velocity.
For the degree of oxidation the transport of oxygen is the most decisive factor.
As these are considerations mainly derived from theory but have not been studied
completely on large scale a field test of 215 operating hours was conducted with
spray tower 1 from Nov. 22 to Dec. 2,1988.
For fined and definitive statements the phases of the steady operation certainly were
too short yet Nevertheless, the preliminary results of this first large scale
experiment at Voitsberg are rather encouraging in view of maintaining the proper
operation of the spray tower without making use of the oxidation air fan.
However, the difficulties in measuring the pH-vaiue turned out to be an important
obstacle to the fast and exact dosing of limestone.
The pH-measurement determines only the liquid phase equilibrums. It is a complex
system of liquid reactions, e.g. the absorption of HSO3, gas/liquid mass transfer,
e.g. the absorption of SO2 or O2. and interphase reactions, e.g. the dissolution of
CaC03 and/or the precipitation of gypsum.
In case of instationary operating conditions like varying contents of SO2 in the raw
gas, as it is typical for lignite the common feed forward control of the limestone
supply has to be improved by implementing sophisticated data collection and
computer calculations to control the limestone feed accordingly.
7-330
-------
Hence, for the optimum operation of the spray tower information on the exact
content of residual carbonates and their reactivities are of decisive importance. The
direct measurement of the content of carbonates and their reactivities permit the
spray tower control by means of the limestone side. The quality of the produced
gypsum as well as the SO2 removal efficiency of the spray tower can be monitored
and improved. So an appropriate measuring system had been designed and
developed by Waagner-Biro Energy & Environment and shortly will be applied in
Voitsberg.
The thickening function
Already at the first trial operation it clearly could be seen that brown sludge is
carried over into the overflow tank.
During a plant inspection after some 4,000 operating hours it turned out that the
overflow tank was partly filled with slurry.
So a dissolution and dosing unit for an organic flocculent (i.e. a liquid polymer
substance) had to be retrofitted. The amount of dosing is appr. 2 to 4 g/ton of
gypsum.
However, attention is to be paid to having the temperature of the mixing water of not
more than 20°C as already at 30°C the undesired flocculation would result in the
mixing with the concentrate.
Apart from finally having a clear overflow liquid the rake performance of the rake
within the thickener also could be homogenized. This means that the concentration
of the input liquid is more homogenous and the solid particles are moved better and
more easily by the rake vanes as a certain amount of crystal glideability is given by
the addition of the flocculent
As during heating, resulting from the calcination of the gypsum, the liquid polymer is
split up into CO2 and H2O there is no negative influence on the quality of the
gypsum.
Special lining with Arbosol
In order to protect the spray levels of the two spray towers a new lining material had
been used instead of the conventional rubber lining.
7-331
-------
Arbosol is a PVC plasitsol that forms a homogenous and chemically highly
endurable lining that is resistant to abrasion and corrosion.
Depending on the wall thickness of the pipe to be protected layer thicknesses of up
to 6 mm may be created by means of a submerging process.
Since 1984 Waagner-Biro has been a licensee of Arbonite Corp., USA. A wide
range of application is given above all in FGDs. Especially in case of the spray
levels the seamless lining with Arbosol provides high resistance to abrasion
compared to a conventional rubber lining. The areas of direct contact with droplets
from the upper level have to be covered by wear layers of butyl rubber.
As has been learned by experience the service life of Arbosol is a little bit shorter
than the standard rubber lining. Considering the considerably lower application cost
there is an advantage in the total operating cost
Agitators
Each of the two spray towers has been equipped with four horizontal agitators. The
parts in contact with the medium were made of austenitic material (1.4539).
During a revision phase of approx. 7,000 operating hours slight erosion marks at
the vanes were welded out and ground flush.
After the installation the agitators were running for another approx. 100 hours when
strong vibrations occurred which were due to cracks in the agitator vanes close to
the shaft
The reason was that this material has no distinctive creep strength values for the
given gypsum slurry data, i.e. pH-values of about 5.2 to 5.6, chloride ions of
approx. 5000 mg/m3. This means the creep strength will decrease proportionally to
the number of load changes.
The first damages occurred after an approximate 7,000 hours what corresponds to
10 to 100 million load changes. These were manifest as cracks in the weld seam
area (transition from the shaft to the vane) along the grain boundaries of the
microscopic structures.
7-332
-------
As a first measure the thickness of the dampers was changed from 15 to 20 mm
what corresponds to a theoretical life of apprux. 1 billion load changes and/or a life
of more than 6 years.
For small tanks vertical agitators were used the dampers of which are rubber lined.
There also was one inexpected incident with one of these agitators.
In case of the limestone tank agitator used for homogenizing the limestone slurry
with water (a content of solid particles of 20 %) there were signs of considerable
wear at the soft rubber lining after approx. 4,000 operating hours despite the use of
the double lining design, i.e. soft rubber lining applied to a base of hard rubber.
As a measure of improvement a highly wear resistant vane of cast fine steel was
used as in case of high concentrations of solid particles the soft rubber lining
literally is ground off. The metal propeller now used will permit a continuous running
time of more than 25,000 operating hours.
Full metal pumps
Centrifugal pumps were used as recycle pumps as well as much smaller slurry and
water recycle pumps.
The recycle pump is connected directly to the drive motor by means of an elastic
coupling and mounted upon a common base frame (see fig. 10).
The parts of the centrifugal pump that are in contact with the medium are of material
G-X3CrNiMoCu N 25 7 (1.4517 acc. to SEW-410). This material is characterized
by its high resistance to corrosion and abrasion. Continuous running times of
30,000 operating hours are the best evidence of these qualities.
The impeller of the MFA centrifugal pump which is not completely covered and the
weldability of the material permit a repair of wear spots and hence a further
increase in the continuous running time. The robust design of the pump, i.e. heavily
dimensioned and oil lubricated bearings with a shaft of stainless material, guarantee
high availability and efficiency.
The installation of a single mechanical seal with a connection to the flushing water
has proven to be completely effective and has been running for more than 20.000
hours up to now.
7-333
-------
The full metal pump in FGDs has taken the lead in the European wet-type FGD
technology.
Raw and clean gas ducts
For the flue gas ducts leaving the boiler steel structure 617 tons of structural
steelwork had to be used. The design values were a temperature of 100 to 160°C
and a design pressure of ± 50 mbar. The material used was RSt 37-2. In order to
have an architecturally impressing design especially the clean gas duct was
executed as a statically and structurally demanding component compared to the
usual design of duct systems. In order to avoid thermal stresses due to the changes
in the temperature a certain kind of static bearing had been chosen. Two bearing
support points in the stack and one truss-type pendulum support close to the
building of the FGD assured the transfer of the vertical forces into the ground. The
branching of the big duct before the stack and the installation of a Y-piece required
the welding in of torsional bracings in the flue gas flow.
Special dampers
For most of the FGD- and DeNOx-refits louvre and guillotine dampers were used.
Later more and more large damper toggle arm dampers were preferred to shutoff
and diverter dampers.
A typical example of application are the toggle-arm dampers which were developed
for Voitsberg (see fig. 11).
For the operation of the flue gas by-pass three double louvre dampers of common
design were necessary for each FGD-train.
Two toggle-arm dampers with an actuating toggle-arm system have been adopted
instead of the double louvre or the guillotine damper.
The operating experience has proven that the large damper damper with toggle-
arm systems are suited for the tightly sealing mansafe system. They meet all the
requirements of the FGD, i.e.
sealing absolutely tightly,
- minimum loss of pressure,
- corrosion resistance,
fast switching time from FGD to by-pass operation,
i
7-334
-------
high availability, and
minimum maintenance requirements.
Due to the toggle-arm principle used in the diverter strong forces are exerted just
when the diverter is closing and in the middle of the damper where these forces are
particularly necessary. The toggle-arm principle leads to independence from
pressure or the direction of the gas flow.
In the beginning there were problems with the Viton sealings at the diverter damper.
Those sealings had not been designed for loads sufficient for the contact pressure
and the burst load. Then we have made use of Viton sealings with a wire mesh on
the inside. These sealings have proven to be the best and up to now they did not
have to be exchanged.
There also were some difficulties with the clean gas damper in the sealing air area
where moisture may penetrate into the sealing air duct Therefore sealing air valves
have been developed that are arranged on the perimeter of the sealing seat and
that are opened only when required by a separate actuator. The sealing air valves
are designed so that the sealing air can be blown directly into the sealing air duct
from the sealing air ring on the outside between the sealing elements of a double
gasket. The outer ring duct can be opened when the sealing air valves are closed
and can be inspected during operation.
The outer ring duct also has another advantage, i.e. that it is not a part of the frame
and therefore no corrosion can occur when the sealing air is not blown in. The
sealing air valves are sealed against humidity and dust.
Experience of Operation with the Gas-Gas Heater (GGH)
The fluegas from the boiler Voitsberg 3 is divided into two streams (2 x 50%). One
stream passes the regenerative GGH with vertical shaft, cold side below.
There are two layers of heating elements - one double enameled sheet of 650 mm
height and one plastic layer (Noryl) of 150 mm height
The normal rotor speed is 1.5 rpm, during the cleaning the speed is reduced to
0.75 rpm.
The inlet temperature (untreated gas) is about 150°, the outlet temperature
(untreated gas) about 100°C. The cold treated gas enters the GGH with about 63°C
and leaves it with about 105°C.
The max. guaranteed pressure drop (sum of untreated and treated gas, untreated
gas quantity is 625,000 st.m3/hr, untreated gas temperature 140°C) is 11.38 mbar.
7-335
-------
If the pressure drop on treated gas side exceeds 10.5 mbar a high pressure
washing program (40 bar cold side and 80 bar hot side) will be started by hand.
For the first 3000 operation hours every 500 hours such a cleaning cycle of 8 hours
duration is intended.
For the next period of operation the GGH has to be cleaned evry 400 hours.
Besides this an everyday cleaning is intended by compressed air. The hot side
elements (enamel) are blown with an air pressure of 5 bar and the cold side
elements (plastic) 2 bar air pressure.
Duration of the blowing cycle 4.5 hours.
In the yearly revision time the scaling on the heating modules is cleaned by setting
under freshwater till it is saturated, then it could be easily removed by flushing with
low pressure water.
In Fig. 12 and 13 the chemical analysis from the scaling dated March 9, 1991 is
given.
The low pH-value (sour character of the eluate] The *~.igh salt loads, especially of
aluminium-, iron-, calcium-, and magnesium-sulpn^u- ~,nd chlorides resp., have to
be pointed out, e.g. Al2(S04)3-
The high Al-value is said to result mainly from the partly not intact enamel layer of
the GGH, because between 13 and 28 % by wt. feldspar is used for the processing
of enamel, that consists of complex aluminiumsilicate, composed of Me(AISi308);
Me = metal
The high iron values point to rust formations on the material below.
Due to the high salt load and the sour character of possible flushing waters the
waste disposal without conditioning (at least dilution and neutralisation, resp.) is not
possible.
The backwash in the spray tower during operation is basically possible as an
alternative (high dilution and at the same time neutralisation by the absorber).
4. NOx-REDUCTION (Fig. 14)
SCR-Process
Selective catalytic reduction (SCR) is a well-known technique especially when firing
hard coal. For lack of sufficient experience regarding SCR techniques at lignite fired
power plants, ODK tested this process at two pilot plants from 1985 to 1988.
7-336
-------
The aim was to test catalyst behaviour in high dust arrangement
The investigations showed that the SCR technology can also be used with high ash
and high sulfur lignites. The large scale DeNOx-plant Voitsberg 3 has been
commissioned in November 1990. It will be one of the first in Europe installed after
a lignite fired boiler.
Primary measures
Parallel to the SCR tests the installation of primary measures (excess air, flue gas
recirculation, overfire air) was investigated.
With investment costs of approx. 25 million AS (2 million US-$) the NOx emission
at design load could be reduced from approx. 500 to 550 to 300 to 350 (mg/sLm3,
dry, 6 % O2. as NO2). As side effect the costs for the SCR plant could be
decreased due to a reduction of catalysts.
SNCR-Process
Seeing that the retrofitting of old plants with SCR is too expensive the SNCR
(selective non catalytic reduction) process has been investigated very intensively
since 1986.
Pure ammonia (NH3), urea ((NH2)2CO) and ammonium water (NH4OH) were
tested. Comparable results (50-60% NOx-reduction at a stoechiometric ratio of
1.5 - 2.0) could be achieved with all three additives.
5. CONCLUSION
The FGD has been in operation for approx. 20,000 hours and has fulfilled all
expectations regarding SO2 reduction and reliability. Typical operating data are
shown in Fig. 6. Power plant availability was never influenced by the FGD.
7-337
-------
e! output at MCR
330 MW
steam output (MCR)
980 t/h
operating pressure
184 bar
steam/reheat temp.
535/535° C
flue gas quantity
125 -10s Nm3/h
lower calorific value
9.2-11 MJ/kg
carbon content
23 - 40 %
moisture
32 - 38 %
ash
15-30 %
sulphur content
o
cn
en
£
ODK
1991
Fig. 1
Thermal Power Plant Voitsberg 3
Waagner-Biro
Energy & Environment GmbH
7-338
-------
DENOX
_ SCR since 1990
electric
precipitator
wet FGD
§
2nd Step
3rd Step
Furnace Limestone Injection
Commissioned 1983
S02 Reduction >50%
SCR
Commissioned 1990
NOx Reduction>80%
Wet FGD
Commissioned 1986
S02 Reduction>90%
ODK
Fig. 2
Voltsberg 3 FGD
Waagner-Biro
Energy & Environment GmbH
7-339
-------
Clean Gas 2
Gear* Gas 1
Mixing Device
Water
Circulation
Crude Gas
Reaction Product
Ground Limestone
Ct&an Gas 1 — Line with
r> , ReHe3'!r%lRe^.ter) Ufnestone"Tank
Clean Gas 2 — Line without Reheating
ODK
1991
Fig. 3
FGO Process Scheme
Waagner-Biro
Energy & Environment GmbH
7-340
-------
Reactions:
CaC03 CaO + C02
S02 + CaO +i02-*-CaS04
Limestone
Ash + Gypsum
Fig. 4
Furnace Limestone Injection (FLI)
Waagner-Biro
7-341
-------
\[°/°]
80
60
40 "
20
o
>
~
~
t
1 ~
o
t
¦o
C
n
a
o
VJ
o
1 c
k
CQ
.c
E
E
*
i load
! 1
t
I
100
180
250
330 [MW]
ODK
1991
Fig. 5
S07-Reduction by FLI
Waagner-Biro
Energy & Environment GmbH
7-342
-------
so2
mg/Nrrf
5000
4000 -
3000 -
2000
1000 -
- 100
8/86 9/86 10/86 11/86
S02 before FGD
12/86
1/87 2/87 3/87
SO, after FGD
Fig. 6
FGD Operating experience
Waagner-Biro
Energy & Environment GmbH
7-343
-------
WAAGNER-BIRO Y
Entity * Environment Q—jnJbJH. Gr&x
Fig- 7a
FGD-Arrangement (Train 1)
7-344
-------
WAAGNER-BIRO K
Energy * Environment GvsjtiJxM. Graz
Fig. 7b
FGD - Arrangement (Train 2)
>-¦
7-345
-------
CaSOt. 2 HzO
97,4 weight %
CaSO,. 1/2 H20
0,32
CaC03
¦0,64
MgC03
0,21
Fe203
0,20
CL~
0,02
Acid insofubles
1,21
weight percentages related to dry substances
enlargement 300 fold
ODK
1991
Fig. 8
FGD gypsum
Waagner-Biro
Energy fi Enviromaant GmbH
7-346
-------
flue gas
outlet
~
w <
—/ ' t: ,
' s • * x / t V '
' * ' ' \ ^ % /N *
V v ;c >¦
» \ » '« * «
'' X V'
Vv'" V V" /'
- "V''~ v ' VV
¦
• 1 6 O— 3-
g | ^ • a g ,
m.e. 2nd stage
spray bank
m.e. 1st stage
spray bank
absorption spray
levels
ODK
1991
Fig. 9
M. E. - Arrangement
Waaener-Biro
Energy & Environment GmbH
7-347
-------
Fig. 10
Recycle Pump
WAAGNER-BIRO t
Emrgy* Environment CatmAKCrai
V-i riy
ikiiPvr r. ii
010
a
KS 500-750
7-3^8
-------
Fig. 11
Flue Gas Damper
rs
WAAGNER-BIRO Y
Eiteiyy < Environment CMJitmCwr
\
T
CLEAN GAS
FROM FGO
TO STACK
BYPASS I
OPERATION !
FLUE GAS
TO FGO
c
Hac^ «
-------
Fig. 12 GGH-Scaling
Data in wt-% referred to the immediate sample:
GGH-plastic elements GGH-beam
Annealing loss at 800°C 18.9 27.8
(incl.) 11.8% S03) (incl.) 15.6%S03
F8203
7.56
6.23
AI2O3
12.8
9.2
MgO
1.76
1.54
CaO
2.36
2.00
Si02
32.7
23.8
T1O2
0.75
0.62
K2O
1.67
1.07
Na20
0.34
0.26
S03
19.4
26.1
98.2
98.6
7-350
-------
Fig. 13 GGH-Scaling Elut'on Test
In order to check up all possibilities of cleaning the GGH from above mentioned scales a
sample according to DEV S 4 was eluted and investigated:
Results:
pH-value (-log H-) 1.3
conductivity (fiS/cm) 23700
CP (mg/l) 90
SO4" " > 39000
Al ¦ 4253
Fe * 2228
Ca ' 825
Mg * 647
K ' 401
71 ' 136
Na a 82
Zn " 51
As " 22
Mn " 18
Cr " 10
V " 9
Cu " 9
Ni " 7
Pb " 0.2
Cd " 0.1
-------
700
600
9 500
0 400
CO
1 300
PrimBry measures_
d> 200
too
270
180
240
270
300
330
BOILER LOAD (MW)
ODK
1991
Fig. 14
NOx emissions at Voitsberg 3
Waagner-Biro
Energy 6 Environment GmbH
7-352
------- |