EPA-650/2-75-027-C
September 1975
Environmental Protection Technology Series
FLUIDIZED BED COMBUSTION
PROCESS EVALUATION
PHASE II • PRESSURIZED FLUIDIZED
BED COAL COMBUSTION DEVELOPMENT
U S Environmental Protection Agency
OHice of Research and Development
Washington DC 20460
-------
EPA-650/2-75-027-C
FLUIDIZED BED COMBUSTION
PROCESS EVALUATION
PHASE II - PRESSURIZED FLUIDIZED
BED COAL COMBUSTION DEVELOPMENT
by
D.L. Kenirns, D.H. Archer, J.R. Ilanim, S.A. Jansson,
B.W. Lancaster, E.P. O'Neill. C.H. Peterson, C.C. Sun.
LL.F. Svcrdrup, E.I. Viclt, and W.C. Yang
Westmghouse Research Laboratories
Beulah Road, Churchill Borough,
Pittsburgh, Pennsylvania 15235
Contract No. 68-02-0605
ROAP No. 21ADB-009
Program Element No. IAB013
EPA Project Officer: P.P. Turner
Industrial Environmental Research Laboratory
Office of Energy , Minerals , and Industry
Research Triangle Park, North Carolina 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D. C. 20460
September 1975
-------
EPA REVIEW NOTICE
This report has been reviewed by the II .S. Environmental Protection
Agency and approved for publication. Approval does not signify that
the contents necessarily reflect the views and policies of the Environ-
mental Protection Agency, nor does mention of trade names or commer-
cial products constitute endorsement or recommendation for use.
RESEARCH REPORTING SERIES
Research reports of the Office ol Research and Development, U.S. Environ-
mental Protection Agency, have been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
m related fields. These series arc:
1. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2. ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5. SOCIOECONOMIC ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to
develop and demonstrate instrumentation, equipment and methodology
to repair or prevent environmental degradation from point and non-
point sources of pollution. This work provides the new or improved
technology required for the control and treatment of pollution sources
to meet environmental quality standards.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
Publication No. EPA-650/2-75-027-C
11
-------
ABSTRACT
Pressurized fluidized bed combustion for electric power
generation provides a direct combustion process for coal and low-grade
fuels, with the potential for improved thermal conversion efficiency,
reduced costs, and acceptable environmental impact. The present work
extends the previous studies to include the collection and analysis of
data on critical system parameters such as sulfur removal, spent sorbent
disposition, and trace element release; development of process options
such as particulate control equipment; assessment of power plant cycles
and component designs such as low-temperature gas cleaning, alternative
cycles and turbine corrosion/erosion; and updated commercial plant design
and cost estimates. No problems have been identified which preclude
commercialization. Available data support the basic pressurized fluidized
bed combustion boiler and adiabatic combustor design and operating
parameters previously selected. Updated system economics show energy
costs for first-generation plants ranging from 7 to 20 percent less than a
conventional plant with stack-gas cleaning. Westinghouse recommends that
the development effort be accelerated; a commercial-scale integrated test
facility be constructed; and experimental studies on gas-turbine tolerance
to particulates and trace contaminants, gas-cleaning technology, and
environmental control be expedited.
iii
-------
PREFACE
The Office of Energy, Minerals, and Industry (OEMI) of the
United States Environmental Protection Agency (EPA) is sponsoring work
on fluidized bed fuel processing, the purpose of which is to develop
and demonstrate new methods for utilizing fossil fuels to produce
electrical energy from utility power plants, to produce process steam,
or to produce process heat that meets environmental standards. These
methods should:
• Meet environmental goals for sulfur dioxide (S0_),
nitrogen oxide (NO ), ash smoke emissions, trace ele-
X
ment emissions, and wastes
• Utilize fuel resources efficiently
• Compete economically with alternative means for meet-
ing environmental goals -
Westinghouse Research Laboratories, under contract to OEMI, have carried
out a program to evaluate, develop, and demonstrate fluidized bed
combustion. This report describes work, performed from June 1973
£
through December 1974 under contract No. 68-02-0605, based on tasks set
forth by EPA which Westinghouse has completed under previous contracts
and on work completed by other investigators and government contractors.
The results from the prior tasks carried out by Westinghouse
on fluidized bed combustion were published in a three-volume report,
"Evaluation of the Fluidized Bed Combustion Process," in November 1971
under contract No. CPA 70-9 and as three volumes of a four-volume report
published in December 1973 under contract No. 68-02-0217: Volumes I and II,
"Pressurized Fluidized-Bed Combustion Process Development and Evaluation >
and Volume III, "Pressurized Fluid Bed Boiler Development Plant Design."
This previous work includes:
-------
• Assimilation of available data on fluidized bed com-
bustion, including sulfur dioxide removal, sorbent
regeneration, nitrogen oxide minimization, combustion
efficiency, heat transfer, particle carry-over,
boiler tube corrosion/erosion fouling, and gas-
turbine erosion/corrosion
• Assessment of markets for industrial boilers and uti-
lity power systems
• Development of designs for fluidisied bed industrial
boilers
• Development of designs for fluidissed bed combustion
utility power systems: atmospheric-pressure flui-
dized bed combustion boilers, pressurized fluidized
bed combustion boiler-combined cycle power systems,
adiabatic fluidized bed combustion-combined cycle
power systems—including first- and second-
generation concepts
• Preparation of a preliminary design and cost estimate
for a 30 MW (equivalent) pressurized fluidized bed
combustion boiler development plant
• Assessment of the sensitivity of operating and design
parameters selected for the base power plant design
on plant economics
• Preparation of cost and performance estimates for
once-through and regenerative sulfur removal systems
• Development of a plant operation and control
philosophy
• Collection of experimental data on sulfur removal and
sorbent regeneration using limestone and dolomites
• Provision of technical consultation and assistance on
the ORD fluidized bed fuel processing program.
The results of these surveys, designs, evaluations,and ex-
perimental programs provided the basis for the work presented in this
vi
-------
report. The scope of this work is directed toward the development of
pressurized fluidized bed combustion systems and includes:
• Collection and analysis of additional data on the
calcium-based sulfur removal system—sorbent selec-
tion, high utilization, regeneration
• Collection and analysis of data on spent sorbent
disposition—utilization and environmental impact of
disposal
• Projection and analysis of trace emissions from flui-
dized bed combustion systems and their impact on gas-
turbine performance
• Analysis of particulate removal requirements and
development of a particulate control system for high-
temperature, high-pressure fluidized bed combustion
systems
• Assessment of gas-turbine design and operating con-
straints
• Construction of a corrosion/erosion test facility for
the 0.63 MW Exxon minlplant
• Continued assessment of fluidized bed combustion
power plant cycles and component designs.
A summary of the work is presented here, with details
and supplemental information in the appendices.
vii
-------
TABLE OF CONTENTS
Page
I. INTRODUCTION 1
II. ASSESSMENT 7
System Economics 9
System Design and Performance 10
System Development 15
III. CONCLUSIONS AND RECOMMENDATIONS 17
IV. SYSTEM ECONOMICS 22
V. SYSTEM DESIGN AND PERFORMANCE • 29
Cycle Evaluation 29
Fluid!zed Bed Combustor 49
Sulfur Removal System1 52
. Sulfur Removal 56
Process Modeling 60
Sorbent Regeneration 62
Sorbent Attrition 66
Trace Element Emission Characteristics 66
Spent Sorbent Disposition 67
Nitrogen Oxide Emissions 72
Particulate Control 74
Fluid Bed Boiler Carry-over 76
Gas Turbine Specifications 76
Environmental Standards 78
Process Operating Range 80
Alternative Particulate Removal Systems 80
System Selection 81
Assessment 83
ix
-------
Table of Contents (Continued)
Minor Element and Trace Element Release 84
Trace Element Quantities Input to the System 85
Experimental Data on Emissions 86
Emission Chemistry and Turbine Tolerance 87
Control gg
Assessment QQ
Gas-Turbine Operation 91
Blade Erosion 91
Blade Corrosion 94
Deposition 96
Turbine Design 97
VI. SYSTEM DEVELOPMENT 99
Fluidized Bed Combustion Test Facility 99
Gas-Turbine Corrosion/Erosion Pilot Plant Test Rig 102
VII. REFERENCES 106
APPENDICES
A. Performance Analysis of High-Pressure Fluidized
Bed Boiler Systems 113
B. Boiler Design Evaluation 137
C. Sorbent Selection and Alternative Sorbents 159
D. Thermogravimetric Studies of the Sulfation of
Limestones and Dolomites. 175
E. Projections of Sorbent Utilization and Sulfur
Removal Efficiency Using Thermogravimetric Data 227
F. Spent Stone Disposition 253
G. Spent Stone Disposal-Assessment of Environmental
Impact 273
H. Trace Emissions Affecting Gas-Turbine Performance 289
I. Particulate Control 3^5
J. Gas-Turbine Design and Operation 337
K. Gas-Turbine Corrosion/Erosion Pilot-Plant
Test Program 3^3
L. Potential for Advanced Steam Conditions with
Fluidized Bed Combustion Boilers 413
M. Fluidized Bed Combustion Test Facility 421
-------
LIST OF FIGURES
Page
SUMMARY
1. Pressurized Fluidized Bed Boiler-Power Plant Schematic 2
2. Pressurized Fluidized Bed Adiabatic Combustor; Combined-
Cycle Power Plant Schematic 3
3. Electrical Energy Costs 23
4. Capacity Factor vs Energy Cost 27
5. High-Pressure Fluidized Bed Boiler Power System with
Intermediate Temperature Particulate Removal 35
6. Pressurized Fluidized Bed Combustion Power Plant with
Cold Cleanup of Combustion Products 37
7. Recuperative Supercharged Reheater Cycle 40
8. Superreheat with Vapor Phase Recuperative Cycle with
High-Pressure Fluidized Bed Boilers Added 43
9. Schematic for Pressurized Furnace Potassium Topping Cycle 44
10. Pressurized Fired Heater Subsystem 45
11. Closed-Cycle Combined Flow Diagram 46
12. Pressurized Fluidized Bed Combustion Power Plant with
Secondary Combustor 48
13. Sulfur Removal System Process Concepts 53
14. Comparison of Pressurized Sulfation of Limestone and
Dolomite 57
15. The Effect of Temperature on CaO Utilization in Sulfation 59
16. The Effect of Ca/S Mole Ratio on Sulfur Retention 61
17. Cyclic Regeneration of CaC03 in Sulfided Dolomite 65
18. Sulfur Removal System Process Alternatives 69
19. Composite Plot of Data for NO Emissions from Fluidized
Combustion of Coal x 75
20. Particle Size Distribution for Different Gas Streams 77
21. Erosion Test Rig Concept 104
xi
-------
List of Figures (Continued)
APPENDIX A
1. Performance of Pressurized Fluidized Bed Power Plant 116
2. Performance of Pressurized Fluidized Bed Power Plant 116
3. Performance of Pressurized Fluidized Bed Power Plant vs
Cycle Pressure Ratio 116
4. High-Pressure Fluidized Bed Boiler Power System with
Intermediate Temperature Particulate Removal 118
5. Plant Capacity vs Gas-Turbine Inlet Temperature for
High-Pressure Fluidized Bed Boiler with Intermediate
Temperature Particulate Removal 120
6. Plant Heat Rate vs Gas-Turbine Inlet Temperature for
High-Pressure Fluidized Bed Boiler with Intermediate
Temperature Particulate Removal 121
7. Pressurized Fluidized Bed Combustion Power Plant with Cold
Cleanup of Combustion Products 122
8. Plant Heat Rate vs Recuperator Effectiveness 126
9. Plant Heat Rate vs Gas-Turbine Pressure Ratio 126
10. Plant Heat Rate vs Boiler Outlet Temperature 126
11. Plant Heat Rate vs Air Equivalence Ratio 128
12. Intermediate Temperature Gas Cleaning for PFBB
(nominal 635 MW size) 130
13. Performance of Pressurized Fluidized Bed Power Plant with
Advanced Steam Conditions 133
APPENDIX B
1. Effect of Bed Temperature on Combustion Efficiency with
17% Excess Air 141
2. Effect of Time on Corrosion (Previous Results in Fluidized
Beds) 143
3. Fluidized Bed Combustion Miniplant Combustor Temperature 147
APPENDIX C
1. Temperature and Pressure Conditions for Stability of the
Sulfur Sorbent as Half-Calcined or Fully Calcined Dolomite
at Projected Combustor Outlet Gas Compositions 161
2. Maximum S02 Retention in Fluidized Beds of CuO •
(Thermodynamic Limit) 168
xii
-------
List of Figures (Continued) Page
APPENDIX D
1. Temperature and Pressure Conditions for Stability of the
Sulfur Sorbent as Half-Calcined or Fully Calcined Dolomite
at Projected Combustor Outlet Gas Compositions 177
2. Atmospheric Pressure Sulfation of Limestone 1359 - The
Effect of Calcination History 184
3. Sulfation of Limestone and Dolomite; Effect of Calcination
Conditions 185
4. Comparison of Pressurized Sulfation of Limestone and
Dolomite 186
5. Effect of Temperature on Sulfation of Limestone 1359 189
6. The Effect of Temperature on Limestone Sulfation 190
7. The Effect of Temperature on CaO Utilization in Sulfation 191
8. Comparison of Rates for Identical Sulfation Runs 193
9. Comparison of Repeat Experiments-Fast Phase of Sulfation 194
10. The Effect of Temperature on CaO Utilization in Sulfation
at Pressure 196
11. Sulfation of Half-Calcined Dolomite 199
12a. The Decline in Rate of Reaction (Calcium Fraction Reacting
per Minute) as Calcined Dolomite Sulfates 201
b. The Effect of Pressure on Sulfation of Calcined Dolomite
1337 (420 um; 0.5% S02/4% 02; 871°C) 201
13. The Effect of Temperature on the Course of the Reaction
between SO. and CaO 208
APPENDIX E
1. The Effect of CO2 Pressure during Calcination on S02
Emissions from a Fluid Bed 232
2. TG Data Predictions Compared to Fluidized Bed Data
Atmospheric Pressure, Limestone 1359 233
3. The Effect of Temperature on Limestone Sulfation 235
4. The Rate of Sulfation of Limestone 1359 236
5. Determination of the Stone Utilization at which the Rate of
Reaction Satisfies the Fluidized Bed Rate Criteria 240
6. The Effect of Temperature on Sulfur Retention in a
Fluidized Bed of Limestone 1359 241
7. Comparison of TG Data with Fluidized Bed Results 241
8. The Effect of CaS Mole Ratio on Sulfur Retention 242
xiii
-------
List of Figures (Continued) page
9. Mole Predictions of the Effect of Superficial Velocity
on Sulfur Retention 242
10. The Reaction Rate Criteria for Desulfurization 245
APPENDIX F
1. Sulfur Removal System Process Alternatives 254
2. Pressurized Fluidized Bed Combustion Spent Stone Disposal
Process 265
APPENDIX G
1. Leachate Characteristics as Functions of Batch Mixing
Time for Argonne Spent Stone and Iowa Gypsum #114 282
2. Leachate Characteristics as Functions of Stone Loading for
Argonne Spent Stone and Iowa Gypsum #114 283
3. Leachate Characteristics of the Argonne Spent Stone
Leachates Induced by the Run-off Tests 284
APPENDIX H
1. Projection of Liquidus Surface of NaSO^I^SO^-KCl-NaCl
System Showing Temperature Contours 293
2. Thermochemical Equilibria in the Sodium-Silicon-Oxygen-
Sulfur System at 1093°C 296
3. Thermochemical Equilibria in the Sodium-Silicon-Oxygen-
Sulfur System at 871°C 299
4. Vapor Pressures of NaOH, KOH, NaCl, KC1, NaSO^and K2SOA 300
5. Equilibrium Pressures of NaOH(g) and NaCl(g) over
Na2S04(s) Exposed to Steam and HCl(g) under Fluidized Bed
Boiler Gas Phase Conditions 301
6. Thermochemical Equilibria in the Potassium-Oxygen-Sulfur-
Carbon System at 1093°C 302
7. K-0-Si-C-S System at 1093°C 303
8. Schematic Flow Diagram for Important Gas and Solids Streams
in One of Four Fluidized Bed Boiler Modules in a 318 MW
Power Plant 305
9. Flow Diagram Showing Transport of Sodium and Potassium in
One of Four Fluidized Bed Boiler Modules in a 318 MW
Power Plant 306
xiv
-------
List of Figures (Continued)
10. Flow Diagram Showing Transport of Chlorine in One of Four
Fluidized Bed Boiler Modules in a 318 MW Power Plant 307
11. Summary of Gas Chemistry and Corrosive Contaminant Levels
in the Combustion Gas Stream Directed to the Gas Turbine 310
12. Pressure-Metal Temperature Relations for First- and Second-
Stage Components of a Large Industrial Gas Turbine 311
13. Conditions of Sodium Sulfate Stability 312
APPENDIX I
1. Particle Size Distribution for Different Gas Streams 316
2. Multicyclone Arrangements 319
3. Cyclone Grade Efficiency Curves 321
4. Operation of Aerodyne Particulate Separator 322
5. Time vs Efficiency - 0.033 m Deep Granular Bed Filtering
Fly Ash 325
6. Ducon Sand Bed Filter 328
7. Squires Panel Bed Filter 330
8. Cross-Flow Filter Arrangement - Westinghouse Gasification
Plant 332
9. Anticipated Solids Flow and "Dead" Zones for a Conventional
Cross-Flow Filter 333
APPENDIX J
1. Turbine Tolerance for Sodium as a Function of the
Concentration of Chlorine and Oxides of Sulfur (Sodium
Sulfate Melt Model) 338
2. Turbine Tolerance for Sodium as a Function of Chlorine
and Oxides of Sulfur Concentration (Na2SO//KoSO/ Eutectic
Melt Model) 339
3. Particle Size Distribution Measured at Entry and Exhaust of
Australian Direct Coal-Fired Gas Turbine 343
4. 9 urn Particle Trajectories in Full-Scale and 3/4-Scale
Stator Passages 346
5. 9 urn Particle Trajectories in 1/2-Scale and 1/4-Scale
Stator Passages 347
6. Effect of Scale Factor on Capture Efficiency 348
7. Effect of Scale Factor on Particle Impact Velocities 349
xv
-------
List of Figures (Continued)
8. Effect of Scale Factor on Particle Impact Angles 350
9. Erosion Results Obtained for Alumina Particles Impacting
2024 Aluminum Alloy 351
10. Effect of Scale Factor on Erosion Rate in Westinghouse-
501 First-Stage Stator 352
11. Trajectory of a 12 urn Particle through First Turbine Stage 356
12. Trajectory of a 6 ym Particle through First Turbine Stage 357
13. Trajectory of a 6 urn Particle Injected Nearer to Midspan
and Midpitch of the Blade Passage 358
APPENDIX K
1. Predicted Size Distribution of Coal Ash Particles Escaping
from Secondary Cyclones in the Exxon Miniplant 365
'2. Cascade Type Erosion Test Rig 370
3. Erosion Test Rig General Assembly 371
4. Erosion Test Rig Test Section 373
5. Erosion Test Rig 2nd Stage Separator 377
6. Erosion Test Rig Elbow Welding and Casting Assembly 379
•7. Erosion Test Rig Elbow Welding and Casting Assembly 380
8. Erosion Test Rig Straight Run Welding and Casting Assembly 381
9. Erosion Test Rig Liner Details 383
10. Erosion Test Rig Liner Details 385
11. Erosion Test Rig Test Specimen Details 387
12. Erosion Test Rig Test Specimen 388
13. Erosion Test Rig Test Specimen Assembly 389
14. Probe Arrangement 394
15. Anderson Impactor Assembly 396
16. Brink Impactor Assembly 397
17. Threshold Kinetic Energy vs Particle Size 403
18. Predicted Size Distribution of Coal Ash Particles Escaping
from Secondary Cyclones in Exxon Miniplant 405
19. Analytical Model for Particle Impaction 406
xvi
-------
List of Figures (Continued)
Page
20. Free Stream Position of Particle vs Point of Impact 408
21. Weighting Factor
22. Average Angle of Impact
23. Mass Impacting Per Unit Area per Unit Time vs @ 413
APPENDIX M
1. Overall Schematic of the Flexible Test Facility 423
2. Pressurized Fluidized Bed Steam Generator for Combined-
Cycle Plant 426
3. Westinghouse-Foster Wheeler Fluidized Bed Boiler 429
4. Material Balance for the Fluidized Bed Combustion Boiler 432
5. Deep Recirculating Fluidized Bed Boiler 436
6. Material Balance for the Recirculating Bed Concept
(Preevaporator) 438
7. Adiabatic Combustor Designs 441
8. Material Balance for the Adiabatic Fluid Bed Combustor 443
xvii
-------
LIST OF TABLES.
Page
SUMMARY
•1. Pressurized Fluidized Bed Combustion Systems Reference
Design Parameters and Performance 4
2. Economic Comparison of 600 MM Plants 24
3. 600 MW Plant Economics for Various Ca/S Ratios 25
4. Steam Conditions - Plant Heat Rate as a Function of Air
Equivalence Ratio 33
5. Energy Cost as a Function of Steam Conditions, Plant Heat
Rate as a Function of Pressure Ratio 33
6. Projections of the Calcium to Sulfur Feed Ratios 54
7. Acceptable Dust Loadings in Expansion Gas Based on Current
Westinghouse Specifications 79
8. Alternative Particulate Removal Systems 82
9. Estimated Dust Collection High-Efficiency Multicyclone
Secondary Collector 83
10. Erosion of Turbine Blades 92
11. Fluidized Bed Combustion System Design Parameters 101
APPENDIX A
1. Parametric Combinations Used in Performance Calculations 123
2. Temperatures for Ten Stations 124
3. Plant Power Outputs and Heat Rates 125
APPENDIX B
•1. Comparison of Operating Conditions 139
2. Summary of Weight Loss Measurements 142
3. Boiler Tube Materials 150
4. Chemical Compositions of Selected Boiler Tube Materials 151
5. Typical Analyses of Metal Specimens 152
6. Comparison of Design Parameters between the Basic Design
and the Pilot-Scale Experimental Units 154
xix
-------
List of Tables (Continued)
Page
APPENDIX C
•1. Exit Gas Conditions for Fluid Bed Combustors 169
•2. Decomposition Temperatures of Copper Sulfate and Copper
Oxysulfate 170
APPENDIX D
1. Sorbent Stability in the Pressurized Fluidized Bed
Combustor 1?9
2. Sorbents Used to Study the Sulfation Reaction 181
3. Limestone 1359 Sulfation Runs 183
4. TG Sulfation of Limestone 1359 Calcined in the 50 mm
Fluidized Bed 187
5. TG Runs on Sulfation of Calcined Limestone, Calcined
Dolomite, and Half-Calcined Dolomite 209
6. TG Data for Run 202 213
7. TG Data for Run 215 214
8. TG Data for Run 221 215
9. TG Data for Run 257 216
10. TG Data for Run 258 217
11. TG Data for Run 259 218
12. TG Data for Run 260 219
13. TG Data for Run 261 220
14. TG Data for Run 300 221
15. TG Data for Run 301 222
APPENDIX E
1. Physical Parameters of the TG System 234
2. Sulfur Retention Projections Compared with Experimental
Data (ANL-AR1) 237
3. Fluid Bed Sorbent Utilization Predicted from
Thermogravimetric Data 244
APPENDIX F
1. Spent Stone Disposition 256
2. Typical Composition of Calcium Compounds in Spent Sorbent 258
xx
-------
List of Tables (Continued)
Page
3. Solubility of Calcium Compounds in Water 259
4. Solubility of Magnesium Compounds in Water 260
5. Domestic Usage of Magnesium Compounds, 1968 262
APPENDIX G
1. Chemical Compositions of Spent Stone from ANL and Exxon
Pressurized Fluid Bed Combustion Pilot Plants and Iowa
Ground Gypsum #114 277
•2. Summary of Leachate Characteristics 279
3. Summary of Leachate Characteristics 280
4. Summary of Stone Activity Tests 287
APPENDIX I
1. Data for Estimating Tolerance of Gas Turbines towards Dust 318
APPENDIX J
1. Extrapolated Life of Blading in Redesigned, Low-Velocity
Australian Gas-Turbine Tests 344
APPENDIX K
1. Comparison of Particulate Concentrations at Gas-Turbine
Inlet 367
•2. Impingement Criteria 407
APPENDIX M
1. Design Basis for Pressurized Fluidized Bed Boiler 427
2. Fluid Bed Operating Conditions in 320 MW Commercial Design 428
3. Material Balance for the Fluidized Bed Combustion Boiler 433
4. Comparison of Design Parameters between the Basic Design
and the Pilot-Scale Experimental Units 434
5. Material Balance for the Recirculating Bed Concept 439
6. Adiabatic Combustor Designs 442
7. Material Balance for the Adiabatic Fluid-Bed Combustor 444
8. Evaluation of Test Facility Size 448
9. Evaluation of Test Facility Size 449
xxi
-------
List of Tables (Continued)
Page
10. Turbine Test Program Components 454
11. Investment Estimate for Test Facility 459
xxii
-------
LIST OF ABBREVIATIONS
A/F
A1(S04)3
ANL
BCURA
CaC03
CAFB
CaO
Ca(OH2)
CaS
Ca/S
CaS04
CH4
co2
CuO
CuS04
EPA
EPRI
ERDA
Esso
Exxon
GJ
HPFBB
J
kg
kJ
kPa
kw
kWh
air/fuel ratio
aluminum sulfate
Argonne National Laboratory
British Coal Utilization Research Association
calcium carbonate
chemically active fluidized bed
calcium oxide
calcium hydroxide
calcium sulfide
calcium/sulfur ratio
calcium sulfate
methane
carbon dioxide
copper oxide
copper sulfate
Environmental Protection Agency
Electrical Power Research Institute
Energy Research and Development Agency
Esso Research Centre, Abingdon, England
Exxon Engineering Corporation, Linden, N.J.
9
giga joule (10 ) S.I. unit of energy
high-pressure fluidized bed boiler
joule, S.I. unit of energy, 1055 J = 1 Btu
potassium sulfate
kilogram, S.I. unit of weight
kilojoule
kilopascal, S.I. unit of pressure, 101.3 kPa = 1 atm.
kilowatt
kilowatt hour
xxiii
-------
m - meter
Mg - megagram, S.I. unit of weight, Q.972 Mg = 1 short ton
MgCO, - magnesium carbonate
MgO - magnesium oxide
ml - milliliter
mm - millimeter
urn - micron
MnSO, - manganese sulfate
m/s - meters per second
MW - megawatt
Na - sodium
NALCO - alumina-based copper oxide catalyst
NAPCA - National Air Pollution Control Administration (now EPA)
NaSO, - sodium sulfate
NLLS - nonlinear least squares
NiSO^ - nickel sulfate
NCB - National Coal Board, U.K.
N02 - nitrogen dioxide
NO - nitrogen oxide
02 - oxygen
OEMI - Office of Energy, Minerals and Industry, EPA
ORNL - Oak Ridge National Laboratory
ppb - parts per billion
ppm - parts per million
PER - Pope, Evans and Robbins
PFBB - pressurized fluidized bed boiler
nns - root mean square
SO, - sulfur dioxide
S0_ - sulfur trioxide
TG - thermogravimetric
TVA - Tennessee Valley Authority
xxiv
-------
LIST OF ABBREVIATIONS
A/F
A1(S04)3
ANL
BCURA
CaC03
CAFB
CaO
Ca(OH2)
CaS
Ca/S
CaS04
CH4
co2
CuO
CuS04
EPA
EFRI
ERDA
Esso
Exxon
GJ
HPFBB
J
kJ
kPa
kw
kWh
air/fuel ratio
aluminum sulfate
Argonne National Laboratory
British Coal Utilization Research Association
calcium carbonate
chemically active fluidized bed
calcium oxide
calcium hydroxide
calcium sulfide
calcium/sulfur ratio
calcium sulfate
methane
carbon dioxide
copper oxide
copper sulfate
Environmental Protection Agency
Electrical Power Research Institute
Energy Research and Development Agency
Esso Research Centre, Abingdon, England
Exxon Engineering Corporation, Linden, N.J.
q
giga joule (10 ) S.I. unit of energy
high-pressure fluidized bed boiler
joule, S.I. unit of energy, 1055 J = 1 Btu
potassium sulfate
kilogram, S.I. unit of weight
kilojoule
kllopascal, S.I. unit of pressure, 101.3 kPa = 1 atm.
kilowatt
kilowatt hour
xxiii
-------
m - meter
Mg - megagram, S.I. unit of weight, 0.972 Mg = 1 short ton
MgCO« - magnesium carbonate
MgO - magnesium oxide
ml - milliliter
mm - millimeter
um - micron
MnSO, - manganese sulfate
m/s - meters per second
MW - megawatt
Na - sodium
NALCO - alumina-based copper oxide catalyst
NAPCA - National Air Pollution Control Administration (now EPA)
NaSO, - sodium sulfate
NLLS - nonlinear least squares
NiSO, - nickel sulfate
NCB - National Coal Board, U.K.
NO* - nitrogen dioxide
NO - nitrogen oxide
02 - oxygen
OEMI - Office of Energy, Minerals and Industry, EPA
ORNL - Oak Ridge National Laboratory
ppb - parts per billion
ppm - parts per million
PER - Pope, Evans and Robbins
PFBB - pressurized fluidized bed boiler
rms - root mean square
S0» - sulfur dioxide
SO, - sulfur trioxide
TG - thermogravimetric
TVA - Tennessee Valley Authority
xxiv
-------
ACKNOWLEDGMENTS
The results, conclusions, and recommendations presented in this
report represent the combined work and thought of many persons at
Westinghouse and OEMI. Other government contractors have freely shared
with us the results of their research and development effort.
In particular, we want here to express our high regard for and
acknowledge the contribution of personnel at Westinghouse Research
Laboratories and at OEMI who have directed the fluidlzed bed combustion
program and who have defined, monitored, and supported the efforts of
Westinghouse and others on the program. Mr. P. P. Turner, Chief of
Advanced Processes Branch, has served as EPA project officer on our work.
Numerous enlightening and helpful discussions have been held with
Mr. Turner; with branch members D. B. Henschel and S. L. Rakes; and with
R. P. Hangebrauck, director Energy Assessment and Control Division.
Personnel from the Westinghouse Research Laboratories have made signifi-
cant contributions. Drs. M. Menguturk and R. W. Hornbeck participated
in the gas turbine assessment work. Dr. L. N. Yannopoulos, Dr. C. Y. Lin,
and M. A. Alvin contributed to the work on trace element release and gas
turbine tolerance to trace elements. Messrs. R. E. Brinza, W. F. Kittle,
W. J. Petlevich, H. W. Sherwin, and L. Toth participated in the collection
of data. Westinghouse division personnel were consulted and participated
in the evaluation. Steam Turbine Division personnel contributed to the
assessment of advanced steam conditions.
XXV
-------
I. INTRODUCTION
Coal represents approximately 85 percent of the U. S. fossil
fuel resources and will play an important role in achieving greater energy
independence. Pressurized fluidized bed combustion for electric power
generation provides a direct combustion process for coal, with the
potential for improved thermal conversion efficiency, reduced costs,
and acceptable environmental impact. This technology can be used to
substitute coal and other low-grade fuels for oil and gas and to provide
an improved option for conventional coal-fired systems. Contributions
from this technology to the energy system have been projected to be up
to 40,000 MW of installed capacity by the year 2000, if the development
is successful. The basic pressurized fluidized bed combustion technology
is available from laboratory- and pilot-plant scale systems. Further
engineering and development is required, however, to demonstrate pressurized
fluidized bed combustion concepts on a commercial scale.
Westinghouse,under contract to EPA, has carried out a program
1 2
to evaluate and develop fluidized bed combustion processes. ' The
historical, technical, and economic aspects of fluidized bed combustion
systems have been reviewed, systems analyses performed, commercial plant
design and cost estimates prepared, and experimental data on the sulfur
removal system obtained. Two pressurized fluidized bed combustion power
plant systems, shown in Figures 1 and 2, have provided the basis for the
work on system design, performance, economics, and development. The
1 2
reference designs and related plant designs have been reported previously. '
The basic design and performance parameters for these two systems are
presented in Table 1. Current work extends the previous work to include
collection and analysis of data on critical system parameters (sulfur
removal, spent sorbent disposition, trace element release);development
of process options (particulate control); assessment of power plant cycles
-------
Dwg. 1676B91
Compressor
Gas Turbine
Electric
Generator
Secondary
Collector
Primary Participate
Collection System
Pressurized -
Fluidized Bed
Boiler Module
Tertiary
Collector
Heat Recovery Unit
(Boiler Feedwater)
Ash and Spent Sorbent
Reheated Steam
N Steam
Turbine
Superheated Steam
•*
Coal& Sorbent*
Feed System
Spent Sorbent for
Disposal/Utilization
or Regeneration
(from each bed)
Feedwater
Electric
Generator
Stack
Heat Recovery Unit
(Flue Gas)
Boiler Feed-
water Pump
Circulating
Water
•e.g. dolomite^
••e.g. spent dolomite (Ca$04-MgO,
CaCCy MgO)
Figure 1-Pressurized fluidized bed boiler - power plant schematic
-------
Owg. 1676890
Compressor
Coal and
Sorbent* c
Feed System
Gas Turbine
Electric
Generator
Secondary
Collector
Stack
Primary Particulate^
Collection System f=->
Heat Recovery
Boiler
Spent Sorbent for
Disposal/Utilization
or Regeneration
Electric
Generator
Circulating
Water
Boiler Feed-
water Pump
Fluidized Bed
Adiabatic Combustor
•e.g. dolomite (CaCCy
••e.g. spent dolomite (CaSCy MgO, CaCC^- MgO)
Figure 2-Pressurized fluidized bed adiabatic combustor - combined-cycle power plant schematic
-------
Table 1
PRESSURIZED FLUIDIZED BED COMBUSTION SYSTEMS
REFERENCE DESIGN PARAMETERS AND PERFORMANCE
Pressurized & fluidized bed
boiler power plant
Adiabatic combustor
combined-cycle power plant
Plant Capacity (nominal)
Cycle Parameters
Steam Conditions
Condenser Pressure
Gas-Turbine Expander
Pressure ratio (100% load)
Inlet temperature (100% load)
Compressor Expander Pressure
Loss (through fuel processing
system), % compressor outlet
' Number of Combustor Modules
Combustor Modules/Gas-Turbine
Module
Excess Air (range for 100%
load and turndown)
Fluidized Bed Combustor Design
Fluid Bed Units/Combustor Module
Vessel Diameter, m(ft)
Bed Area
Heat Transfer Surface
Vessel Height, m(ft)
Fluidized Bed Combustor
Operating Conditions
Bed Temperature
100% load
Range for turndown
300 MW
16.547 MPa/538°C/538°C
(2400 Psi/10000F/1000°F)
5.07 kPa (1-1/2 in Hg)
10:1
871°C(1600°F)
7.5
4
2
10 to 100%
3.7 (12)
1.5 x 2.1 m; 3.26 in
(5x7 ft; 35 ft2)
(a)
32.94 (108)
954°C (1750°F)
760-983'C (1400-1800'F)
Gas Velocity (100% load), m/s (ft/sec) 1.71-2.74 (5.6-9.0)
200 MW
4.823 MPa/427°C
(700 psi/800°F)
8.45 kPa (2-1/2 in Hg)
10:1
927°C(1700°F)
7.5
2
1
300 to 360%
9.15 (30)
64.64 m2 (695
None
4.88 (16)
954°C (1750°F)
760-983'C (1400-1800°F)
1.8 (6)
-------
Table 1 (Continued)
.
Bed Depth, m(ft)
Freeboard, m(ft)
Particle Size (coal & sorbent) ,
mm(ln)
Ca/S Molar Feed Ratio for
Pressurized & fluid! zed bed
boiler power plant
3.35-4.36 (11.0-14.3)
1.53-2.31 (5-7.6)
up to 6.35 (1/4)
2b
Adlabatlc combust or
combined-cycle power plane
2.0 (6.5)
2.44 (8)
up to 6.35 (1/4)
2b
Once-through Operation
Sorbent
Performance
Power Generation Split (100% load)
Z Gas turbine
Z Steam turbine
Plant Heat Rate (HHV)
Turndown Capability
(Z of full load)
Rate
Environmental Impact
SO,, kg/GJ (Ib/KT Btu)
HCT, kg/GJ (lb/106 Btu)
Particulate kg/GJ (lb/10° Btu)
Minor and trace element release
Heat rejection (% less than
conventional steam plant)
Waste liquids
Waste solids
Characteristics
Quantity Mg/hr/MW (tons/hr/MW)
Resources (multifuel capability)
Dolomite
18ZC
82ZC
9040 kJ/kWh (8570)Btu/kWh)c
12-1/2Z
5Z/mlnute
Dolomite
70Z
30Z
9600 kJ/kWh (9100 Btu/kWl
<10Z
5Z/mlnute
<0.3(<0.7)d
<0.04(<0.1)
To be determined
<0.3(<0.7)
<0.04(<0.1)
To be determined
None
Dry, granular, sulfated
limes tone/dolomlte
0.18b C\.0.2)
Tes
None
Dry, granular, sulfated
limes tone/dolomlte
0.18b
Tes
a2 in O.D. tubes on 3-1/2 in welded wall spacing.
1-1/2 in O.D. tubes in preevaporator and superheater; 2 in O.D. in reheater; heat transfer coefficient.
283.5 W/m2-°K (50 Btu/hr-ft2-°F); tubes represent 17 to 22.5Z of the bed volume.
Subsequent data indicate that Ca/S ratios from 1.0 to 1.2 can be used to achieve 90% sulfur removal.
This reduces the waste solids production to ^0.07 Mg/hr/MW or M).067 ton/hr/MW.
Parameters for the 100% excess air base case are: power generation: 29% gas turbine, 71% steam
turbine; plant heat rate 8940 kJ/kWh (8470 Btu/kWh)i. ^25% less heat rejection.
^Available data indicate W>2 emissions <0.17 kg/GJ (0.4 lb/10
Btu).
-------
and component designs (use of low-temperature gas cleaning, alternative
cycles); and provision of support to other contractors (gas-turbine
corrosion/erosion test rig design and construction).
-------
II. ASSESSMENT
The results from experimental pressurized fluidized bed
combustion test facilities (units operated by Exxon, Argonne National
Laboratories[ANL], Combustion Power, and British Coal Utilization Research
Association [BCURA]) during the past 18 months confirm the projected
performance for the reference power plant designs previously carried out
by Westinghouse under contract to EPA. The results support the pressurized
fluidized bed combustor and power plant reference design parameters
previously utilized. However, these tests have not been carried out under
all of the proposed operating conditions, have not investigated all the
plant components under projected operating conditions, and have not been
carried out on commercial-scale equipment, additional work is required
to confirm commercial plant designs. The results indicate:
• High sorbent utilization can be achieved in a once-through
system. A calcium/sulfur molar ratio less than 2, with
dolomite, will achieve 90 percent removal. This is consistent
with the projections reported in 1973 on the basis of
)
1
2
thermogravimetric (TG) data and lower than the original
projections of 3 to 6.
• Nitrogen oxide emission standards can be met. The mechanism
for the nitrogen oxide emissions, however, is not yet
understood. Data are typically less than half the emission
standard of 0.3 kg/GJ (0.7 lb/10 Btu) . For example, 0.13
to 0.17 kg/GJ (0.3 to 0.4 lb/10 Btu) have been reported
3
for tests with excess air from 15 to 100 percent.
• Combustion efficiency greater than 98 percent at 100 percent
excess air can be achieved. The data at 100 percent excess
air, with bed depths significantly less than the commercial
-------
design values , indicate that the complete cor^ustion loss of
1.5 percent assumed in the reference design with 100% excess
air will be achieved without requiring further processing
of the ash carry-over.
• Heat transfer coefficients of 284 W/m2-°K (50 Btu/hr-ft2-8F)
or greater can be achieved.
• Bed temperatures can be used over the range projected —
760 to 954°C (1400 to 1750°F) — and still meet environmental
standards.
Experimental and analytical studies performed under this contract show
that critical parameters in the reference design concepts are:
• Selection criteria for calcium-based sorbents
• Control of alkali-metal compounds to prevent corrosion/
deposition damage to gas turbines.
• Specification for particulate removal equipment to prevent
erosion damage to gas turbines
• Disposition of spent sorbent.
The available results from experimental and analytical studies
indicate that no problems have been identified which preclude either
commercialization or meeting environmental standards. The results also
indicate the processes are less sensitive to fuel variations, further
advanced in development, less demanding of the gas turbine (as the result
of lower turbine inlet temperatures), and superior to or comparable in
cost and efficiency with any competitive,advanced fossil fuel processing
system under development utilizing state-of-the-art power generation
equipment.
The work reported in this contract report focuses on providing
additional support data and analyses to understand specific process
components — sulfur removal, particulate control, and gas turbine systems:
on evaluating the current development status of pressurized fluidized bed
-------
combustion systems; and on providing support to on-going development work
being carried out by other contractors.
An assessment of the system economics, system design and
performance, and system development follows.
SYSTEM ECONOMICS
Energy costs for pressurized fluidized bed combustion combined-
cycle power plant systems are estimated to be up to 20 percent less than*
those for a conventional plant with stack-gas cleaning at around $100/kw.
These savings are based on a 3 percent sulfur fuel cost of 80c/GJ(10 Btu).
An energy cost reduction of 17 percent is projected for a pressurized
fluidized bed boiler (PFBB) operating at 17.5 percent excess air, with
three stages of high-temperature particulate removal utilizing a calcium/
sulfur ratio of 1.2 for a once-through sulfur removal system. The same
system with low-temperature particulate removal (e.g., water scrubbing)
would be approximately 15 percent lower in energy costs than the conven-
tional plant. This is a particularly significant result of the present
analysis since it indicates that low-temperature particulate control,
which may offer improved power plant reliability, is also economically
attractive. The pressurized fluidized bed boiler system, operating at
100 percent excess air to achieve higher carbon utilization in the
primary combustors and greater turndown flexibility, is approximately 13
percent lower in cost but must utilize high-temperature particulate
removal to be economical. Energy costs for a pressurized fluidized
bed adiabatic combustor combined-cycle power plant with a once-through
sulfur removal system using a calcium/sulfur ratio of 1.2 and three
stages of particulate control are estimated to be approximately 7 percent
lower than those of a conventional plant. Plant costs projected for
regenerative fluidized bed combustion systems indicate regeneration is
not cost competitive with the conventional plant until sorbent costs
(including disposal) exceed $10 to 20/Mg. Since the regenerative calcium-
based sorbent processes proposed are not yet well defined, however, a
definitive conclusion is not possible. A regenerative process is clearly
attractive environmentally and, perhaps, will be economically if a
-------
regenerative process can be developed to operate with calcium/sulfur
makeup ratios significantly less than 1.
SYSTEM DESIGN AND PERFORMANCE
Cvcle Evaluation
The national average heat rates for fossil fuel steam electric
plants have leveled off at 11.1 MJ/kWh (10,500 Btu/kWh). The most
efficient plants average around 9.4 MJ/kWh (8900 Btu/kWh) and do not
include sulfur removal or nitrogen oxide minimization. Conventional
plants which meet environmental standards have heat rates near 9.6 MJ/kWh
(9100 Btu/kWh). Coal-fired units now being built or on order are all
12.411 MPa/538°C/538°C (1800 psi/10008F/1000°F), 16.547 MPa/538cC/538°C
(2400 psi/1000°F/1000°F), and 24.132 MPa/538°C/538°C (3500 psi/1000°F/
1000°F). The first-generation pressurized fluidized bed boiler or
adiabatic combustor combined-cycle power plants utilizing state-of-the-
art power generation equipment are projected to match the most efficient
current plants at reduced cost and to achieve environmental control
requirements at the same time. Plant heat rates for the base designs
are 8.94 MJ/kWh. (8470 3tu/kWh) and 9.04 MJ/kWh (8570 Btu/kWh) for the
pressurized fluidized bed boiler plant operating at 100 percent and 17.5
percent excess air .respectively. The adiabatic combustor base plant
heat rate is 9.6 MJ/kWh (9100 Btu/kWh).
The two base systems (the pressurized boiler and adiabatic
combustor) remain the most attractive. Closed-cycle gas-turbine, air-
cooled heat transfer surface and related applications offer potential
advantages by avoiding or minimizing possible problem areas in the
reference designs and should thus be investigated. However, they do
not offer reduced capital cost or improved energy efficiency when
compared with the reference systems. The performance and energy cost
of a pressurized fluidized bed boiler power plant operating at low
excess air (e.g., 17.5 percent) with a low-temperature particulate
removal system is competitive with the low excess air pressurized boiler
with three stages of high-temperature particulate removal. This concept
offers the potential for improved plant reliability by achieving higher
10
-------
particulate removal efficiencies and condensing alkali metals prior to
entering the gas turbine. The gas heat exchanger performance and life
must be demonstrated to achieve these advantages. The use of advanced
steam conditions is not attractive because of the projected cost increase
which would be required for the steam turbine.
Combustor Design
The available data do not indicate any problems which preclude
the development of pressurized fluidized bed combustors. Further
development work is required in a number of key areas, however, including
performance with large-scale heat transfer surface, testing boiler tube
life, and operation over the full range of commercial plant conditions.
The results from the Exxon experimental test unit have indicated
one area of concern, the potential for nonuniform temperature gradient
in a deep fluidized bed combustion boiler. Further tests on this unit
and data from large-scale combustors operating at the design pressure,
temperature, velocity, and bed height are required to verify this aspect
of the reference boiler design. Modifications of the base designs, such
as alteration of the tube packing arrangement, may be desired.
Sulfur Removal System
The sulfur removal system must be compatible with the total
process. In order to achieve this compatibility, factors such as trace
element release and spent sorbent disposition must be investigated
along with the sulfur removal characteristics. The work on sulfur
removal shows that the conditions under which the calcium-based sorbent
is calcined is critical in establishing the calcium utilization. The
results show that calcium/sulfur ratios from 1.0 to 1.2 can be achieved.
Pilot-scale tests at ANL have demonstrated these results. A model of
the desulfurization process was developed which successfully predicts
sorbent performance on fluidized bed combustors,using TG data. Investi-
gation of the poor regenerability of calcium sulfide during regeneration
continues; analysis of the results to date has not revealed the mechanism
for deactivation. Tests on sorbent attrition show that attrition
11
-------
resistance will be a primary factor in selecting sorbents. Sorbent
attrition rates from laboratory screening tests show variations from 1
to 40 percent of the initial bed inventory and indicate that most of
the attrition occurs during calcination. Direct disposal and utilization
of the spent sorbent were investigated, and both leaching experiments
and activity tests indicate that direct disposal of the spent sorbent
will not cause water or heat pollution. Process options for alternative
disposition of the spent sorbent are identified. There is no indication
at this time that the sorbent will contribute a significant or hazardous
concentration of trace elements to the environment. The release of
sodium and potassium, however, is a primary concern because of gas-
turbine corrosion. This aspect will be addressed further.
Nitrogen Oxide Emissions
Nitrogen oxide emissions from fluidized bed combustion
processes result from the bound nitrogen in the fuel. Experimental
investigations show that the conversion of bound nitrogen is substantially
less than 100 percent. Data from pressurized fluidlzed bed boiler test
units and adiabatic combustor test units using coal show that the nitrogen
oxide emissions are less than approximately 0.173 kg nitrogen oxide/GJ
(0.4 lb/10 Btu) for excess air values up to around 300 percent. This
is approximately half the EPA standard of 0.302 kg NO^/GJ (0.7 Ib NO./
106 Btu).
The mechanism which produces the low nitrogen oxide emission
is not yet understood. Those parameters which have been identified as
having a significant effect on nitrogen oxide emissions are pressure
level, excess air, calcium/sulfur feed ratio, carbon monoxide, and
bed hydrodynamics. Sufficient data are not available to assess the
effect of the nitrogen content and character of the fuel. Some or all
of these factors may determine the nitrogen oxide emission level.
Further work is required to understand nitrogen oxide emissions from
pressurized fluidized bed combustion systems and to assure selection of
design and opr>r.;ting r-ir-'"'ierers to such emissions. The nitrogen t-xide
emissions., tiuwevtu , ale well wo.chin the current emission standard.
12
-------
Particulate Control
Particulate removal is critical for the successful operation
of pressurized fluidized bed combustion combined-cycle power plants.
Criteria for particulate control are ill defined, and there is a need
for more definitive data. The allowable limit for gas-turbine operation
is expected to be lower than the present allowable emissions (0.043
gm/MJ (0.1 lb/106 Btu) equivalent to 0.115 gm/m3 (0.05 gr/scf)).
Simple mechanical collectors, such as cyclones, are inadequate
for meeting the dust collection requirements. Instead, three stages of
cleaning are projected for particulate control:
• Primary collectors to collect coarse material
• Secondary collectors to remove the bulk of the remaining
fine material
• Tertiary collectors to reduce the level of ultrafine
particles to acceptable levels.
The use of three stages of particulate removal equipment represents a
1 2
significant change in the original reference designs, ' which assumed
two stages. This change is the result of a further assessment of the
gas-turbine tolerance to particulates, which projects a lower dust-
loading criteria, and of recent experimental tests on tornado cyclones,
which indicate the efficiency will not be as high as originally projected.
Experimental testing on large-scale, hot, pressurized equipment is required
to establish the operating performance of the dust collection equipment.
A test facility capable of handling 14 to 28 actual m3/min (500 to 1000
acftn) of gas flow is envisioned.
Minor Element and Trace Element Release
Trace emissions are important because of their effect on the
environment and on the operability of the plant. Preliminary experiments
to investigate the fate of potential environmental pollutants have been
initiated in pilot-scale combustors by ANL and by Exxon. This work is
still in the early stages and does not yet provide a basis for assessing
13
-------
the potential environmental impact. The experimental data on alkali-
metal emissions from pilot-scale combustors are also limited. The work
carried out under this contract was directed toward developing an under-
standing of the alkali-metal emission chemistry of the gas turbine and
its tolerance of alkali metals, relating this understanding to available
data and identifying critical concerns. The results indicate:
• That, on the basis of the release of sodium and potassium
into the gaseous effluent from the fluidized bed combus tor,
the equilibrium turbine tolerance for alkali metals may be
exceeded in normal operation without some control. It
must be noted that the available data are not sufficient
to permit a definitive conclusion, and further tests are
required.
• That, control options to prevent damage to turbine hardware
are available. These include the use of high-purity sorbents;
selection of combustor operating conditions to reduce release;
prevention of sulfate deposition by controlling chlorine
concentration or increasing the sulfur dioxide removal
(«200 ppm); selection of the low-temperature particulate
removal option which would condense the alkali metals; use
of additives to trap alkali and trace elements.
Gas-Turbine Operation
Erosion, corrosion, and deposition in the gas turbine were
investigated, and design features needed to duct the hot gas into the
turbine and to protect the turbine from localized contaminants were
assessed. Blade erosion data are extrapolated to indicate permissible
particulate loading. This projection shows that particulate loadings
of 0.012 to 0.093 gm/m3 (5 x 10~3 to 4 x 10~2 gr/scf) will yield about
25,000 hr blade life when the particle size distribution has a mean
size between 4 and 8 pm in diameter (see Appendix J). These estimates
must be considered preliminary until data on the erosivity of particulates
from fluidized bed combustors are available. The estimated tolerance
-------
expansion gas. The tolerance for Che fluidized bed boiler system is
significantly higher, since the chlorine concentration is greater with
the coal-fired system. If the alkali-metal compound tolerances are met,
deposits cemented by alkali-metal compounds should not form when the gas
turbine is operating at the temperatures specified. Deposits resulting
from the impaction and dry sintering of fine particles can occur.
Tolerances required to limit these deposits are not yet established.
The turbine tolerance model must be modified to allow for the presence
of calcium oxide fines which will increase the turbine tolerance to
alkalis. Further turbine design work is required to develop reliable
gas transfer pipe designs and to investigate design alternatives to
increase turbine life.
SYSTEM DEVELOPMENT
Experimental and systems work is required to investigate
potential problem areas, establish commercial-scale design criteria,
develop operation and control procedures, test plant components, and
investigate environmental impact. Various test facilities have been
proposed by different organizations to advance the development of
pressurized fluidized bed combustion systems. Three development aspects
are critical for the expeditious application of pressurized fluidized
bed combustion systems:
• Experimental tests on a commercial-scale pressurized
fluidized bed comb ustor test facility with integrated
auxiliary components
• Experimental and analytical studies to establish gas-turbine
performance
• Experimental and analytical programs to ensure compliance
with environmental regulations (air emissions and spent
solids disposition).
A commercial-scale fluidized bed combustion test facility was proposed
in 1971 and a preliminary design completed in 1973. An extension of the
previous concept is presented in this report. The experimental and
analytical work to establish gas-turbine performance must include work
15
-------
on combustor emissions (particulates and trace elements), particulate
control, and gas-turbine unit analysis and tests. Three types of
facilities are envisioned to obtain the necessary information:
• Test passages and supporting data from pilot-scale pressurized
fluidized bed combustion units
• Operation of an integrated commercial-scale fluidized bed
combustor, particulate collection system, and gas-turbine
test equipment
• Operation of independent test facilities to study gas
cleaning and turbine tolerance.
Under this contract Westinghouse designed and constructed a gas-turbine
corrosion/erosion pilot-plant test rig constructed by Westinghouse for
the Exxon miniplant and established a test program. Gas-turbine test
equipment is also identified for incorporation in a flexible pressurized
fluidized bed combustion test facility. Environmental studies should
be continued on the laboratory- and pilot-plant scale, and the commercial-
scale facility used to determine the environmental impact of pressurized
fluidized bed combustion systems.
Additional support studies are recommended to supplement these
critical program tests. These support studies should include, but not be
limited to, work on boiler tube wastage, instrumentation, operation
and control procedures, and solids feeding. Work should also be directed
toward the development of regenerative sulfur removal systems—utilizing
both calcium-based and alternative sorbents—for second-generation
pressurized fluidized bed combustion processes.
16
-------
III. CONCLUSIONS AND RECOMMENDATIONS
CONCLUSIONS
The primary conclusions from this work are that:
• Pressurized fluidized bed boiler and adiabatic combustor
combined-cycle power plants represent an opportunity to
improve fuel economy and reduce energy costs within
environmental constraints.
• No problems have been identified which preclude commercial-
ization; no problems have been identified which preclude
meeting environmental standards. The processes are less
sensitive to fuel variations and further advanced in
development than competitive advanced fossil fuel pro-
cessing technology using state-of-the-art power genera-
tion equipment, and they are comparable in cost and
efficiency.
• Available experimental data support the basic commercial
fluidized bed combustion boiler and adiabatic combustor
design and operating parameters previously selected.
• Updated system economics show energy costs for first-
generation pressurized fluidized bed combustion combined-
cycle power plants ranging from 7 to 20 percent less
than a conventional plant with stack-gas cleaning.
• Development priorities for pressurized fluidized bed
combustion systems are:
1. Operation of a commercial-scale pressurized fluidized
bed combustion test facility with integrated auxiliary
components (in other words, solids feed, combustor,
particulate control, gas-turbine test system)
17
-------
2. Experimental and analytical studies to establish gas-
turbine tolerance to particulates and trace elements
3. Investigation of high-temperature and low-temperature
particulate control systems
4. Experimental and analytical programs to assure compliance
with environmental regulations.
Specific conclusions from the experimental and analytical
studies are:
• The pressurized fluidized bed boiler power plant operating
at low excess air has the lowest projected energy cost for
first-generation pressurized fluidized bed combustion systems.
• A pressurized fluidized bed combustion boiler power plant
operating at low excess air (e.g. 17.5 percent) is economically
attractive(energy cost 15 percent lower than in a conventional
plant with stack-gas cleaning) with low-temperature (e.g.
water scrubbing) particulate control equipment before the
gas turbine.
• Limestone/dolomite sorbent (or alternate sorbent) selection
criteria must consider attrition characteristics, trace
element content and release characteristics, and spent sorbent
disposition factors in addition to the sulfur removal behavior
and regenerability.
• Calcination of calcium-based sorbents is critical in
establishing high sorbent utilization. Calcium/sulfur ratios
from 1.0 to 1.2 can be achieved with dolomite.
• Sorbent performance for sulfur removal in the fluidized bed
combustion processes can be successfully predicted from
thermogravimetric data.
• The technical or economical feasibility of sorbent
regeneration has yet to be demonstrated. A regenerative
system with makeup calcium/sulfur ratios significantly
less than one would be advantageous to minimize the
environmental impact from sorbent procurement and disposition.
18
-------
Direct disposal of spent dolomite from a once-through sulfur
removal system can be achieved without water or thermal
pollution.
Sufficient data are not yet available to assess the
environmental impact from minor and trace element release.
Available data and modeling of the alkali-metal emission
chemistry and gas-turbine tolerance to alkali-metal corrosion/
deposition indicate sodium and potassium release may exceed
the turbine tolerance unless control techniques are applied.
This is based on pilot-plant data indicating sodium-plus-
potassium levels greater than 1000 ppb using calcium-based
sorbents and on projected criteria for sodium plus
potassium in the expansion gas less than 300 ppb.
Control techniques to prevent turbine damage from
alkali-metal compounds are available; preferred options
include establishing high purity requirements for the
sulfur sorbent, selecting combustor operating condi-
tions to reduce release, preventing sulfate deposition
by controlling chlorine concentration utilizing low-
temperature particulate removal, or increasing sulfur
dioxide removal to achieve « 200 ppm.
Particulate control requirements for the gas turbine
are more restrictive than environmental regulations of
0.043 gm/MJ (0.1 lb/106 Btu) equivalent to 0.115
o
gm/m (0.05 gr/scf).
Particulate loading permissible for gas turbine opera-
3
tion is projected to be less than 0.012 to 0.093 gm/m
(0.005 to 0.04 gr/scf), based on turbine modeling studies
(> 25,000 hr life).
Three stages of high-temperature particulate removal will
be required to achieve the allowable limits projected
for gas-turbine operation.
Energy costs can be reduced an additional 8 percent if
turbine particulate tolerance is such that only two
stages of high-temperature particulate removal equipment
are required.
19
-------
RECOMMENDATIONS
On the basis of the experimental, analytical, and economic
studies, and on assessments carried out, Westinghouse recommends that:
• Development of pressurized fluidized bed combustion
and adiabatic combustor combined-cycle power plants
be accelerated
• A commercial-scale, pressurized fluidized bed combustion
test facility with integrated auxiliary components be
constructed
• Experimental studies to establish gas-turbine tolerance
to particulates and trace contaminants be expedited
• Large-scale [14 to 28 m3/min (500 to 1000 acfm) gas flow],
separate test facilities be operated to develop gas-
cleaning technology - particularly particulate removal
components
• Environmental control studies be continued to assure
that processes meet environmental impact requirements.
Recommendations for specific investigations include:
• Extending analyses of modified power plant cycles,
particularly the use of low-temperature particulate
removal with the low excess air pressurized boiler
concept
• Establishing sorbent selection criteria for sulfur
removal, attrition, trace element content, and spent
sorbent disposition
• Initiating work to assess the potential of achieving very
low sulfur dioxide concentrations in the product gas
(1 to 200 ppm) as a means to control gas-turbine
corrosion from alkali metals
• Continuing laboratory and pilot-scale work to further
understanding of sorbent calcination requirements to
achieve high utilization
20
-------
• Reassessing the use of fine particle calcium-based
sorbents to achieve high utilization
• Continuing laboratory-scale work to develop regenerative
calcium-based sorbent process(es) and regenerative sulfur
removal processes utilizing alternative sorbent materials
• Continuing experimental studies to investigate utiliza-
tion of spent sorbent and ash from once-through operation
• Continuing work to determine the environmental impact
of minor and trace elements
• Extending analyses and continuing experimental studies
on alkali-metal emission chemistry and gas-turbine
tolerance
• Developing control techniques to prevent damage to gas
turbines from alkali-metal compounds
• Carrying out large-scale tests to demonstrate alterna-
tive secondary and tertiary particulate control equipment
• Carrying out support work to further understand basic
component phenomena such as boiler tube wastage, nitrogen
oxide minimization, bed mixing, and heat transfer; and
to develop auxiliary systems such as coal feeding and
instrumentation.
21
-------
IV. SYSTEM ECONOMICS
Based on the economic comparison presented in the December
2
1973 report and on the cost updating discussed in Appendix A of this
report, the late-1974 minimum cost for a 600 MW steam power plant with
limestone scrubbing is about $465/kw and could range up to $530/kw.
Similarly, the late-1974 minimum cost for a once-through dolomite
pressurized fluidized bed boiler (PFBB) with 17.5 percent excess air and
two stages of particulate removal is $355/kw. The cost with three
stages of particulate removal is $394/kw. The adiabatic combustor system
cost with three particulate removal stages is $468/kw. The FFBB system
operating at 100 percent excess air to provide greater operating flexi-
bility and simplicity costs $426/kw. The capital cost of the low excess
air PFBB system using water scrubbing for particulate removal is
$395/kw — essentially the same as the cost estimated with three stages
of hot gas particulate control. The PFBB and adiabatic combustion system
costs are predicated on a calcium/sulfur molar ratio of 2.0. The lime-
stone scrubbing costs are based on a calcium/sulfur molar ratio of 1.27.
Figure 3 and Tables 2 and 3 are used to illustrate the energy
costs for the PFBB with 100 percent excess air and adiabatic combustor
plants compared to those of conventional plants. Variations in energy
cost are shown for various calcium/sulfur ratios and for a range of
conventional plant stack-scrubber costs of $50 to 115/kw ($465 to 530/kw
total plant cost).
It is evident from Figure 3 that the PFBB system with 100 percent
excess air can cost from 1 to 3 mills less than steam power plants with
stack-gas scrubbing, depending on the calcium/sulfur molar ratio used for
the once-through PFBB design and the cost of the scrubbing system. The
adiabatic combustion system is competitive on a once-through sorbent
design with a steam power plant plus a limestone stack-gas scrubbing
svstem costine more than about $60/kw. If the mode of PFBB operation
22
-------
Curve 681704-A
23
22
I 21
o
o
>,
p>
|20
UJ
19
18
.Conventional Plant with
$115/kw Stack Scrubbing
__ _ Conventional Plant with
~" $50/kw Stack Scrubbing
600 MW Plant Sorbent at $10/Mg
1
1.0 1.2 1.5 2.0
Ca/S Ratio for Dolomite or Limestone Sorbent
Figure 3-Electrical energy costs
-------
Table 2
ECONOMIC COMPARISON OF 600 MW PLANTS
(3% Sulfur Coal, 902 Sulfur Removal)
Pressurized fluid bed boiler
combined-cycle once-through
dolomite sulfur removal system
Adiabatic combustor
combined-cycle once-
through dolomite
sulfur removal svstem
Conventional steam power
plant with limestone
scrubbing
Plant Cost, $/kwc
Dolomite or Limestone
Supply & Disposal Cost
$/Mg '
Energy Cost, mills/kWh
426
10
a!5%/year
70% capacity factor
Coal @ SOc/GJ (106 Btu)
1977 operation of PFBB and adiabatic combustor
1978 operation of conventional plant.
100% excess air, three scages particulate removal.
CBased on total costs for November 1974 start of
project, and including allowances for 6% contin-
gency, 7.5% interest rate during construction,
•7.5% escalation during construction, all D]US 2%
468
10
4651
10
530"
10
Fixed charges
O&M
Fuel
S sorbent
Total Cost Base Ca/S Ratios
10.42
1.15
7.07
1.26
19.90£
11.45
1.26
7.35
1.31
21.37
11.37
1.39
7.42
0.45
20.63
12.96
1.58
7.42
0.45
22.41
$50/kw for limestone scrubbing.
6$115/kw for limestone scrubbing.
The energy cost estimate with two stages of
high-temperature particulate removal is
18.30 mills/kWh. The energy cost for the low
excess air case with two stages of particulate
removal is 17.90 mills/kWh.
-------
ro
Ui
Table 3
600 MW PLANT ECONOMICS FOR VARIOUS Ca/S RATIOS3
[PFBB (100% excess air) and Adiabatic Combustor]
Sorbent Cost, mills /kWh
PFBB
Adiabatic comb us tor
Total Energy Cost, mills /kWh
FFBB
Adiabatic comb us tor
Ca/S Ratio
1 1 i.2 1 1
0.63 0.76 0
0.66 0.79 0
19.27 19.4 19
20.72 20.85 21
.5
.94
.98
.58
.04
1 ,
1.26
1.31
19.90
21.37
Based on economics developed for PFBB in Table -2,
-------
were to be changed from once-through to regenerative, the operating cost
addition for the regeneration system could approach 3 mills/kWh and still
maintain a total cost competitive with that of a conventional steam power
plant with $100/kw limestone scrubbing. Further consideration would
indicate that the regenerative PFBB system could be preferable to a con-
ventional plant with a once-through scrubber, even at the same energy
cost, since both longer construction time and slurry discharge are
inherent intangible detriments for the conventional plant installation.
The comparison of regeneration processes made in Volume I of the 1973
2
report, Table 36, showed that the most expensive regeneration system
added 2.37 mills/kWh to $10/Mg once-through dolomite system energy costs.
Adjusting the 2.37 mills/kWh differential for today's inflated investment
costs and for the 80
-------
44
42
40
38
36
34
32
Curve 681706-A
30
o
28
26
24
22
20
18
- PFBB@ 100% Excess
Air + 2-Step Regeneration
Conventional Plants Scrubber
Adiabatic Combustor
0
Basis.- 600 MW Plants
3% Sulfur Coal;
90% Sulfur Removal
Coal@80*/GJ(MMBtu)
Sorbent@$10/Mg
Ca/S = 2forFluidized Bed-
Combustion Systems
(Once-through)
PFBB@ 100% Excess Air
PFBB@17.5%ExcessAir
I
10
20
60
30 40 50
% Capacity Factor
Figure 4-Energy cost vs capacity factor
70
80
-------
total costs for supply and disposal are $20/Mg or higher. The adiabatic
combustor once-through sorbent system can be competitive with a steam
power plant with a stack-gas scrubber system costing over $60/kw. Both
the dry and the wet designs of particulate removal applied to a low
excess air PFBB system can save about $30/kw compared to the 100 percent
excess air PFBB system.
This savings must be balanced against the higher carbon utili-
zation achieved in the primary beds and the flexibility in turndown
achieved with the 100 percent excess air case. If the three-stage
particulate removal system projected is required, however, then the low
excess air, low-temperature particulate removal system offers economic
advantages and may offer greater gas-turbine reliability. Also, the fuel
cost assumed is lower than typical costs projected for future coal
contracts, and the once-through sorbent requirement used for the base
case (calcium/sulfur = 2) is expected to be conservative since lower
calcium/sulfur ratios (Ca/S = 1.2 to 1.5) have been demonstrated on
laboratory- and pilot-scale equipment to achieve the specified sulfur
removal. These factors favor the pressurized fluidized bed combustion
systems and increase the energy cost differential with conventional
plants.
28
-------
V. SYSTEM DESIGN AND PERFORMANCE
CYCLE EVALUATION
Fluidized bed combustion technology can be applied with many
different configurations and can be utilized in many different power
cycles. A brief summary is presented here of the proposed pressurized
fluidized bed combustion boiler and adiabatic combustor designs. Alter-
native pressurized fluidized bed combustion power plant cycles are
reviewed and their status assessed. Operation, control, and performance
of these systems are evaluated.
Base Systems
Pressurized Fluidized Bed Boilers
Under EPA Contract No. CPA 70-9, and on the basis of an
evaluation of available data, Westinghouse prepared a preliminary design
for a commercial-scale, pressurized fluidized bed boiler power plant.
A 320 MW plant required four modules of 3.7 m (12 ft) diameter, and a
635 MW plant required four modules of 5.2 m (17 ft) diameter. Each
pressure vessel housed four separate fluidized beds of cross section
1.5 by 2.1 m (5 by 7 ft). The bottom bed was for preevaporation, the
top bed for reheating, and the other two beds for superheating. Evapora-
tion of water was accomplished in the water walls which enclosed the
beds. The design was characterized by vertical modular components,
horizontal steam tubes immersed in the bed, bed depths of 3.3 to 4.3 m
(11 to 14 ft), gas velocities of 1.8 to 2.7 m/sec (6 to 9 ft/sec),
pressure of 1013 kPa (10 atm), and limestone/dolomite for sulfur removal.
The effect of the change in fluidized bed operating conditions and
design parameters on economics and performance was analyzed and found
to be essentially invariant with the projected design basis. The effects
of steam pressure and gas-turbine pressure ratio were also evaluated.
29
-------
A pressurized fluid bed combustion boiler combined-cycle power plant
using state-of-the-art power generation equipment with steam conditions
of 16,548 kPa/538°C/538°C (2400 psi/iOOO°F/1000°F), gas-turbine pressure
ratio of 10:1, and gas-turbine temperature of 871°C (1600°F) was the
preferred design. This plant has a calculated heat rate of about
9040 kJ/kWh (8570 Btu/kWh). (See Appendix A.)
British Coal Utilization Research Association (BCURA)
developed a 140 MW pressurized fluidized bed boiler design for a
combined-cycle power plant. The design was characterized by a hori-
zontal 4.3 m (14 ft) diameter, 30.5 m (100 ft) long pressure shell,
horizontal steam tubes immersed in the bed, gas velocity of 0.6 m (2 ft)
/sec, bed depth of 1.2 m (4 ft), pressure of 810 kPa (8 atm), and no
provision for sulfur removal. Subsequent to this initial design, a
70 MW design was proposed. Design features of the 70 MW plant are
similar to the Westinghouse-Foster Wheeler design except that the BCURA
design operates at a lower fluidizing velocity, 0.8 m/s (2.5 ft/sec),
and at a higher pressure, 1611 kPa (16 atm). A pressurized fluid bed
combustor pilot plant has been successfully operated at BCURA to obtain
7-9
process data.
Pressurized Fluidized Bed Adiabatic Combustors
The requirement for an internal heat transfer surface in
fluidized bed combustion can be eliminated by increasing the excess air
to a bed operating at 927°C (1700°F) to approximately 300 percent. An
adiabatic coal-fired fluidized bed combustor is applicable to gas-turbine
or combined gas-turbine/steam-turbine cycles. Combustion Power Company
has operated a pilot plant of a gas turbine with a fluidized bed
combustor on prepared solid waste. The unit has been modified, under
contract to OCR, to test coal. The pilot plant combustor is 2.7 m
(9 ft) in diameter operating at 414 kPa (60 psi) with gas velocities up
to 2.1 m (7 ft)/sec and 0.6 m (2 ft) bed depths. The product gas goes
to a Rustin TA 1500 gas turbine. Westinghouse, under contract to EPA,
previously carried out conceptual designs and performance and economic
30
-------
studies on adiabatic fluidized bed combustion systems for coal-fired
combined-cycle power plants. Two combustor design concepts were
studied - a single fluid bed module and a stacked fluid bed module.
The fluidized bed combustors were designed to operate at gas velocities
of about 1.8 m (6 ft)/sec, bed depths of 1.8 to 2.1 m (6 to 7 ft), and
bed temperature of 954°C (1750°F). With a turbine inlet temperature of
927°C (1700°F), the calculated plant heat rate is about 9600 kJ/kWh
(9100 Btu/kWh).
Alternative Pressurized Fluidized Bed Combustion Power Cycles
Advanced Steam Conditions
In the late 1950s, the trend toward higher power plant effi-
ciency with little regard for economics culminated in the building of
the Eddystone Unit No. 1,with throttle steam conditions of 34,475 kPa
/649°C/566°C/566°C (5000 psi/12000F/10500F/1050°F).1:L Because of
serious operating problems with this unit, and with American Electric
Power's Philo No. 6,designed for 31,028 kPa (4500 psi), subsequent
supercritical units have been limited to pressures of 24,133 kPa
(3500 psi). More recently, the trend has been away from supercritical
cycles because they were relatively unreliable and uneconomical, and
unable to attain their predicted performance in practice.
During the early 1960s, a number of steam plants were con-
structed with superheat temperatures of 593°C (1100°F). Operating
experience with these plants showed that there was no economic advantage
in using steam temperatures above 538°C (1000°F), and since that time
most new units have been designed for 538°C (1000°F) superheat and
reheat.
A description of the typical coal-fired power plant of the
early 1970s is as follows:
o 800 MWe, 3600 rpm, tandem compound, four 31-in low-
pressure ends
o Seven feed heaters plus gland, generator, and oil
coolers, 5 degree approach
31
-------
• 16,548 kPa/538°C/538°C (2400 psi/1000°F/1000°F)
• Condensing at 8.4 kPa (2.5 in Hg), natural draft
wet towers
• Turbine cycle heat efficiency of 44 percent [8271
kJ/kWh (7840 Btu/kWh)]
• Unit net heat rate 9622 kJ/kWh (9120 Btu/kWh)
• Pulverized coal fired
3
• Consumptive water use 1.7 kg/kWh/50,006 ra per day
(3.75 lb/kWh/11,000,000 gpd)
During the initial phases of the preliminary design of the
high-pressure fluidized bed boiler for utility applications under
Contract No. CPA 70-9, a parametric study was made to compare the
estimated performance and energy costs of high-pressure fluidized bed
boilers operating at 16,548 (2400) and 24,133 kPa (3500 psi). For both
pressure levels the superheat and reheat temperatures were 538°C (1000°F).
The results of this parametric study indicated that there was no
economic advantage in using supercritical steam conditions in the high-
pressure fluidized bed boiler with superheat and reheat temperatures of
538°C (1000°F). Experience with conventional pulverized-coal plants
has shown that there is little economic advantage to using 24,133 kPa
(3500 psi) steam pressure instead of 16,548 kPa (2400 psi).
There is reason to think that the hot-side corrosion problems
in the fluidized bed boiler will be less severe than in the conventional
pulverized-coal boilers and, therefore, that super.leat and reheat
temperatures greater than 538°C (1000°F) would be technically feasible.
In view of this a series of cycle performance calculations was made to
see how much the heat rate of a plant with high-pressure fluidized bed
boilers could be improved with higher steam superheat and reheat tempera-
tures (and correspondingly higher pressures). The results of these
calculations for a gas turbine with a pressure ratio of 10, a turbine
inlet temperature of 871°C (1600°F)>and an air equivalence ratio of 1.1
2
are shown in Table 4.
32
-------
Table 4
PLANT HEAT RATE AS A FUNCTION OF STEAM CONDITIONS
Throttle
pressure,
kPa.(psi)
Superheat
temperature ,
Reheat
temperature ,
No. of
heaters
Heat
rate
kJ/kWh
(Btu/kWh)
Ratio of
heat rate to
base case
16,548
(2400)
24,133
(3500)
31,028
(4500)
34,475
(5000)
538 (1000)
538 (1000)
649 (1200)
760 (1400)
538 (1000)
538 (1000)
649 (1200)
760 (1400)
7
7
8
8
9040~~
(8570)
8900
(8440)
8190
(7770)
7840
(7430)
. 1.000
0.985
0.907
0.868
Cost estimates of high-pressure fluidized bed boiler systems
with advanced steam conditions made under NASA Contract NAS 3-19404 gave
the results shown in Table 5 for a gas turbine with a pressure ratio of
10:1, a turbine inlet temperature of 871°C (1600°F), and an air equiva-
lence of 1.1. The increased cost of the high-pressure steam turbine for
steam temperatures higher than 538°C (1000°F) with corresponding
pressures more than offsets the effect of the improved heat rate on the
cost of energy.
Table 5
ENERGY COST AS A FUNCTION OF STEAM CONDITIONS
PLANT HEAT RATE AS A FUNCTION OF PRESSURE RATIO
Throttle
pressure,
kPa (psi)
Superheat
temperature,
°C(°F)
Reheat
temperature,
°C(°F)
Specific
Cost ,
$/kw
Cost of
energy ,
$/kw
24,133
(3500)
31,028
(4500)
34,475
(5000)
538 (1000)
649 (1200)
960 (1400)
538 (1000)
649 (1200)
760 (1400)
374
474
534
21.28
23.70
25 - 19
33
-------
Oxygen-Blown System
Westinghouse has made an evaluation of an oxygen-blown
12 13
atmospheric-pressure fluldized bed boiler ' under EPA Contract
68-02-0605, which led to the conclusion that the high cost of oxygen
prohibits the use of an oxygen-blown atmospheric-pressure fluidized
bed combustion system for economical steam or power generation. The
bare cost of 95 percent oxygen, using depreciation rates allowable for
utilities, and direct costs only for fuel and labor, is $6.81/Mg. The
additional charges incurred by an industrial producer of oxygen such as
Linde or Air Products add at least $7 or $8 more per Mg of oxygen and
give over-the-fence costs of about $14. Since there would be no sig-
nificant cost reduction associated with an oxygen-fired plant, an
increase in energy costs of up to 10 mills/kWh would result.
Low-Temperature Cleanup Techniques for High-Pressure
Fluidized Bed Boilers
A study has been carried out to determine the performance
penalty which accompanies reduced-temperature techniques for removing
particulates from the combustion products of a high-pressure fluidized
bed boiler.
Two alternatives have been investigated:
• Cooling the combustion products by the use of a
convection-type boiler and removing particulates by
cyclone separators, tornado separators, electro-
static precipitators, or combinations thereof, at
intermediate temperatures
• Cooling the combustion products with a recuperator
followed by a scrubber-cooler.
The arrangement for the high-pressure fluidized bed boiler
system with intermediate temperature particulate removal is shown in
Figure 5. The convection-type boiler for cooling the products of com-
bustion to a temperature well below the bed operating temperature is in
series/parallel* with the fluidized bed boiler. This permits the
*Series for combustion products-parallel for working fluid.
34
-------
Intermediate Temperature
Particulate Removal
:.wg.
Convection-
Type
Boiler
Fluidized
Bed
Boiler
Sorbent —
Coal —
Spent
Sorbent
Solids
Auxiliary
^Combustor
(Start-up)
**•
ST
6
Condenser
^^H
_/VA>^
» Tn
Stack-Gas
Cooler
Figure 5 -High-pressure fluidized bed boiler power system with intermediate temperature
particulate removal
-------
temperature of the combustion products to be reduced from 870°C (1600°F)
to the gas-turbine idle temperature which is in the range of 482 to 538°C
(900 to 1000°F). Analysis of the performance of this system shows that
the plant capacity decreases about 1.8 percent, and the plant heat rate
increases about 158 kJ/kWh (150 Btu/kWh) for each 38°C (100°F) drop in
temperature.
The arrangement of the high-pressure fluidized bed boiler with
low-temperature particulate removal is shown in Figure 6. The hot com-
bustion products are passed through one stage of cyclone separation to
recover the larger fraction of the char particles which are elutriated
from the bed so that the carbon losses will be reduced to a minimum value.
The effluent from the char separator will retain the finer ash particles
with relatively low carbon content. This stream will be cooled to a
temperature in the range of from 93 to 204°C (200 to 400°F), depending
on the recuperator effectiveness and the temperature of the cold products
of combustion out of the scrubber-cooler. In the scrubber-cooler the
products are evaporatively cooled to the saturation line and then further
cooled along the saturation line until the water vapor content of the
gas mixture is equal to the initial value out of the fluidized bed boiler.
Performance calculations were made to determine the effects of
air equivalence ratio, boiler outlet temperature, system pressure ratio,
and recuperator effectiveness. The results of this parametric analysis
were as follows:
• The plant heat rate increases about six percentage
points for each ten percentage points decrease in
recuperator effectiveness.
• Plant heat is a rather weak function of the
pressure ratio, and the optimum pressure ratio is
about 8.5:1.
• The plant heat rate varies only about one percent
over the range of boiler outlet temperatures from
704 to 926°C (1300 to 1700°F).
36
-------
Recuperator ScrubberiCgojer ,-Waler
n
Sorbent out
F. B. Boiler
Sorbent in
Coal
f
Boiler
VT
ASH
I J
« JL _A , I
£ _ &AQOI 1JOO/1000 ^} |
Lpru.
Air
Auxiliary
Combustor
(Start-up)
Dwg.6364A33
Air
Comb. Prod.
Steam and Water
Solids
»• To Stack
Figure 6-Pressurized fluidized bed combustion power plant with cold cleanup of combustion products
-------
• Plant heat rate is a rather strong function of the
air equivalence ratio, with the heat rate increasing
as the air equivalance ratio increases.
Economics for the options considered for the FFBB include the
use of hot pressurized electrostatic precipitators for gas cleanup
ahead of the turbine, and heat exchange of the boiler off-gases down
to water-scrubbing temperature prior to injection into the turbine.
The results of these studies are presented in Appendix A and are
reviewed in Section IV.
Gas Turbine Cycle with Indirect Air-Cooled Fluidized Bed Combustor
One variation of the adiabatic combustor concept is the utili-
zation of the excess air required to maintain the bed temperature in
air-cooled heat transfer surface in the bed. This open-cycle gas
turbine with a pressurized fluidized bed combustor/air heater has been
14
described by Harboe. This permits part of the air (about 70%) from
the compressor to by-pass the combustor gas flow, pass through the
heat transfer surface immersed in the fluidized bed,where it approaches
the bed temperature, and be mixed with the products of combustion after
they have been cleaned of particulates. This alternative is less
demanding of the particulate control system than the pressurized boiler
and adiabatic combustor base designs. The cycle performance will be
slightly lower than the adiabatic combustor system for a given bed
temperature, since the excess air will only approach the bed temperature
and result in a slightly lower gas-turbine inlet temperature. The
capital cost, compared with the adiabatic combustor, will depend on a
trade-off between an increase in the combustor cost due to the air-
cooled heat transfer surface and a decrease in the particulate removal
system cost due to reduced volumetric gas flow. The plant reliability
may be increased if the particulate loading to the gas turbine can be
significantly reduced over that achieved with the adiabatic combustor.
This may result in a lower development risk.
38
-------
Balance Pressure Reheater Cycle
Progress in the art of steam power generation depends on
innovation as well as on an analysis and extension of existing concepts.
Westinghouse Research Laboratories has conceived a new,
improved, combined cycle. This cycle is called "superreheat with
vapor phase recuperation." As shown in Figure 7, this cycle consists of:
1. High subcritical or supercritical vapor generation
in a boiler which has been fired with gas-turbine
exhaust and fuel which need not be "clean"
2. Superheating the vapor to the 538°C (1000°F) range
in an orthodox superheater
3. Turbine expansion to reheat pressure near the
saturation line
4. Reheating in a steam-steam recuperator to the
538°C (1000°F) range
5. Reheat in "Balanced Pressure Superreheater" to
816 °C (1500°F) or higher
6. Turbine expansion of steam from 816°C (1500°F) to
First Law balance point
7. Expanded hot reheat steam recuperating Item A
above to pinch point
8. Pinch point steam expanding to condenser.
This apparatus includes a reheat exchange and a high-pressure
ratio gas turbine in "Velox" arrangement. The reheat exchanger is a
carbon-steel, refractory-lined drum containing a radiant-convection
steam reheat exchanger made of thin-walled, high-alloy tubes. Gas-side
pressure is approximately equated to steam-side pressure. Clean fuel
is fired to the drum and again at the drum exit. Reheated gas is
expanded in the gas turbine. Gas-turbine exhaust is cooled against
'feed heating, steam generation, and (possibly) steam generation at cold
reheat pressure. The gas turbine may beneficially consist of a free
gas generator and a reheated power turbine.
39
-------
Curve682083-B
o>
3
TO
g.
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
Steam Enthalpy - kJ/kg
465 930 1396 1861 2326 2792 3257 3722 4187 4653 5118
I I I I I
I I I
i r
T
i r
ToReheaterat 1093° C
(2000° F)
Gas
Condition
Curve
L P. Extraction
Feed Heating
\
I
J I
I
I
I
J I
926.7
871.1
815.6
760.0
704.4
648.9
593.3
537.8
482.2
u
o
I __
200 400 600 800 1000 1200 1400 1600 1800 2000 2200
Steam Enthalpy - Btu / Ib
426.7
-------
The strategy of this unusual arrangement is to maximize the
temperature of heat addition by means which do not require material
breakthroughs in terms of cost, corrosion resistance, and stress-rupture
properties. The superreheat turbine is at reasonable pressure for
thermal stress minimization, and its steam path may be cooled with
saturated (or near) cold reheat steam.
The main features are:
• Use of gas-turbine cycle compressed air to permit a
very high-temperature, unstressed steam reheater
• Use of steam-steam recuperator to suppress the
addition of low-temperature heat in the steam reheater
• Use of gas-turbine reject heat for feed heating
to maximize the steam flow in the high-temperature
reheater
• Optimization of gas-turbine compressor intercooler
pressure to emphasize either (a) the minimum heat
rate by raising the compressor outlet temperature
or (b) the maximum gas-turbine net power by reducing
the compressor work
• Minimization of steam extraction by maximizing the
gas-turbine/boiler-exhaust feed heating. This step
results in a power split which leans toward the
gas-turbine shaft. Optionally, boiler exhaust heat
may be used for additional steam generation at cold
reheat pressure.
The cycle chosen for a rough test of these principles used
24,133 kPa/538°C/816°C (3500 psi/1000°F/1500°F) steam coupled to a
34/1 pressure ratio, 941°C (1725°F) reheated gas turbine. A unit heat
rate of approximately 6858 kJ/kWh (6500 Btu/kWh) appears to be
possible for gas firing. This relationship implies about 7913 kJ/kWh
(7500 Btu/kWh).
The superreheat cycle with vapor-phase recuperation appears
to be a good opportunity for application of high-pressure fluidized bed
boiler technology to an advanced steam system.
41
-------
Figure 8 shows the superreheat with vapor-phase recuperation
cycle modified to operate with high-pressure fluidized bed combustion.
There are two high-pressure fluidized bed combustors. The first gener-
ates and superheats steam at supercritical pressure and operates at an
intermediate pressure level. The second fluidized bed combustor is
used to reheat the steam up to a temperature of about 816°C (1500°F)
with the boiler pressure level approximately equal to that of the
reheat steam pressure. The bed temperatures and, consequently, the
gas-turbine inlet temperatures would most likely be equal.
Liquid-Metal Topping Cycles with Fluidized Bed Combustion
Investigations of liquid-metal topping cycles coupled with
pressurized fluidized bed fired heaters (see Figure 9) have been made
by Oak Ridge National Laboratory and by General Electric under joint
NASA/OCR funding. The conclusions of these studies were that overall
plant efficiencies of 50 to 60 percent were attainable with this system.
A recent evaluation of this system made by Westinghouse under
NASA Contract NAS 3-19404 concluded that maximum overall plant effi-
ciency would be in the order of 48 percent and that the cost of energy
would be about 20 percent greater than that for the pressurized fluidized
bed boiler steam plant with 24,133 kPa/538°C/538°C (3500 psi/1000°F/1000°F)
steam conditions.
Closed-Cycle Gas-Turbine Applications
Recently Westinghouse carried out an evaluation of closed
cycle gas-turbine systems under NASA Contract NAS 3-19404 in which
fluidized bed fired heaters were applied. Figure 10 shows a pressurized
fired heater with fluidized bed combustion. When applied to the helium,
closed-cycle, gas-turbine system with steam bottoming shown on
Figure 11, the calculated plant heat rate is about 8750 and the
estimated capital cost for a plant with construction starting in mid-
1974 is about $635/kW. This gives an estimated energy cost which is
approximately 33 percent greater than that for a pressurized fluidized
42
-------
Dwg. 62U6A93
Low Excess Air
fry.
Reheater
Recuperator H /vs/vs* H
Regenerative
Feedwater
--! Heaters
To Stack
Stack Gas
Cooler
Figure 8 - Superreheat with vapor phase recuperative cycle with high
pressure fluidized bed boilers added
-------
Gas-Feedwater Heater
Air In
Refers to Cycle State
Exhaust Gas
Figure 9 - Schematic for Pressurized Furnace Potassium Topping Cycle
-------
Owg 6364A32
Pressurized
Fluidized Bed
Combustor
Sorbent
Recuperator
V-Stack-Gas
Cooler
Air
© Ambient
Turbine
Cooling
Air
Electric
Power
Figure 10 -Pressurized fired heater subsystem
45
-------
Dwg. 6251A08
Combustioh Air
Pump-up Set
PrU.
Turb.
PrU. .
Vapor
Generator
Pressurized
Furnace
Helium
G.T. Set
Helium
Turb.
Helium
Comp.
Helium
Vapor Generator
4
Steam
Turbine
Cold Reheat Line
Condenser
Figure 11 -Closed-cycle combined flow diagram
-------
Owg 6364A32
Pressurized
Fluidized Bed
l(l(
'U*
ipe
orl
™
te
mmmt
Cc
ip
*"i
j
itor-.
»IV/I ^
mt
bent
jnK
— t
ir
erafc
\
U^
k
\
mm m
r
X
X
i
<
<
1
^ ^
^
^
x
X
t
>
>
M ^W«H
i
<
s
<
^
b
p
ql
>
>
>
\
^
A
n
/
"i
i
\
*
t
f
1 — i
"N
T
1
A
i
/
\j
?
r
'
A q2
1 — ^AA^4-'
^ t
>
\-Stack-Gas
Cooler
, «
^-^^ u
III
Air
© Ambient
Turbine
Cooling
Air
Electric
Power
Figure 10 -Pressurized fired heater subsystem
-------
Dwg. 6251A08
Combustion Air
Pump-up Set
PrU.
Turb.
PrU. t
Vapor
Generator
Pressurized
Furnace
Helium
G.T, Set
Helium
Turb.
Helium
Comp.
Helium
Vapor Generator
Steam
Turbine
Cold Reheat Line
Condenser
Figure 11 -Closed-cycle combined flow diagram
46
-------
bed boiler with steam conditions of 24,133 kPa/538°C/538°C (3500 psi/
1000°F/1000°F).
Technique for Increasing Turbine Inlet Temperature in High-
Pressure Fluidized Bed Boilers
Turbine inlet temperatures for industrial and commercial gas
turbines have historically increased about 14°C (25°F) per year over
the past two decades. Current levels are 1038 to 1093°C (1900 to 2000°F)
for intermediate load units and 1093 to 11498C (2000 to 2100°F) for
peaking units. These levels significantly exceed the value of 871 to
927°C (1600 to 1700°F) imposed on the gas turbine in the high-pressure
fluidized bed boilers by the sulfur sorbent. The maximum allowable bed
temperature for effective, in-bed desulfurization is thought to be in
the range of 954 to 1010°C (1750 to 1850°F), and the turbine inlet
temperature is estimated to be about 66°C (150°F) lower than the bed
temperature.
Westinghouse has made cycle calculations to determine the
effect of turbine temperature on the performance of the high-pressure
2
fluidized bed boiler. The results of these calculations indicate that
the heat rate of this system improves about 0.5 percent for every 56°C
(100°F) increase in turbine inlet temperature.
One possible technique for obtaining turbine inlet temperatures
which exceed the limits imposed by in-bed desulfurization is shown in
Figure 12. Here char which is elutriated from the primary beds is used
as a feedstock for a gasifier. The fuel gas produced in the gasifier
is burned in a separate combustor to give state-of-the-art gas-turbine
inlet temperatures.
The separate gas-turbine combustor for burning the fuel gas
from the gasifier would probably be a conventional, integrated type,
which has a definite design and cost advantage over an external
combustor, especially for high turbine inlet temperatures.
47
-------
CO
Sorbent out
Dwg 617IA66
Air
Comb. Prod.
Fuel Gas
Steam and Water
Solids
T~~r 1
Fn
• D •
«—I [ f Gasifier
*-%•• -.I- i
rfv
Ash
*• To Stack
Air
Figure 12- Pressurized Fluidized Bed Combustion Power Plant with Secondary C-
-------
Modular Integrated Utility System
Oak Ridge National Laboratory (ORNL), under funding from the
Departments of Housing and Urban Development and Interior, has made an
evaluation of small coal-burning gas turbines with fluidized bed com-
bustion. This modular integrated utility system would use a small
on-site utility plant located adjacent to a new housing development to
provide electricity, hot water for building heating, domestic hot water,
and chilled water for air conditioning. Solid waste from the housing
complex would be used as a supplemental fuel in the energy system.
These fundamental gas-turbine cycle configurations were con-
sidered in this study:
• Direct-fired open cycle
• Exhaust-fired open cycle
• Indirect-fired closed cycle
An indirect-fired closed cycle with an air preheater was
selected as the preferred system for this application because of its
superior part-load performance. The distribution of input energy in
the preferred system was 30 percent converted into electricity and
60 percent into steam at 120°C (250°F) and 202.6 kPa (2 atm) or hot
water at 65°C (150°F). It was estimated that the overall installed
cost of the coal-fired fluidized bed combustion system and gas turbine
would be $350 to 400/kw(e) for production units.
FLUIDIZED BED COMBUSTOR
Atmospheric and pressurized fluidized bed combustion boilers
and adiabatic fluidized bed combustors have been demonstrated on pilot-
scale units. Atmospheric and pressurized fluidized bed combustor
379 18—27
tests ' * ' have demonstrated control of sulfur dioxide, nitrogen
oxide, and particulates; acceptable combustion efficiency; boiler tube
materials; and coal-feed systems.
2 5
Economic studies ' on operating and design parameters show
that the uncertainties for successful application of the pressurized
boiler are minimized, since cost and performance are essentially
49
-------
invariant with the projected design basis. Back-up design and operating
alternatives, however, have been specified in those problem areas which
have been identified. An illustration of the problem areas, proposed
solutions, and back-up alternatives has been presented. The proposed
design for a pressurized fluidized bed combustion boiler may be con-
servative in some areas. If deep beds are practical, for example, the
number of fluid bed units can be reduced; and the carbon burn-up cell,
which is desirable in an atmospheric pressure system, may not be
required with deep beds and high excess air.
Results from operating units can be used to construct first-
generation prototype plants. In order to take advantage of the full
potential of fluidized bed combustors, large-scale tests will also be
required with advanced designs at advanced operating conditions, for
example, deep bed [(4.57 to 9.14 m [15 to 30 ft)] and high velocity
(> 3.04 m/sec [10 ft]) operation for pressurized combustors, heat
transfer surface material, and configuration tests at advanced steam
conditions [31,028 kPa, 649°C (4,500 psi, 1200°F)], advanced combustor
designs (for example, circulating bed), and advanced control systems
to achieve improved load-follow capabilities (circulating beds, use of
excess air).
Assessment
The operational conditions and the design parameters used for
2 5
the basic design ' were evaluated on the basis of recent data avail-
able from the pressurized BCURA pilot-scale unit and the Exxon mini-
plant (Appendix B). High combustion efficiencies were obtained with
excess air from 10 to 17 percent. This may allow the carbon burn-up
cell in the basic design to be eliminated and considerably simplify the
operation and control of a pressurized fluid bed combustor. Operation
with excess air up to 100 percent offers additional operating advantages
and would assure high combustion efficiency. Based on the data avail-
284 W/m2-°K (50 Btu/ft2-hr-°F) in the basic design appeared to be
able so far, the assumption of a bed-tube heat transfer coefficient of
284 W/m2-°K (50 Btu/ft2-
reasonably conservative.
50
-------
Long-term boiler tube wastages were estimated from the avail-
able experimental data. Weight loss usually reached a reliable, steady
state after long duration. The steady-state weight loss was considerably
lower than the initial weight-loss rate. The preliminary evaluation
with the available data indicated that conventional boiler tube
materials could be used in the fluidized bed boiler application if the
long-duration runs showed that the rate of weight loss did, indeed,
level off with an increase in operation time. Higher alloy materials
can be used with minimum cost penalty (a cost increase of less than one
percent of the total plant cost). A final assessment can be carried
out only after more data are accumulated at higher fluidized velocity,
higher operating pressure, high temperature, and longer duration. The
effect of adding limestone and of burning different coals with different
impurities, especially the high-sulfur coals, should also be studied.
Conflicting observations have been made with regard to
temperature distribution in a deep pressurized coal combustor. The
difference was due to the operating conditions and the design parameters.
From the existing evidence it was clear that a reasonably uniform
temperature profile could be obtained in a pressurized fluidized bed
combustor if the operating conditions and the design parameters were
properly set. More data from large combustors operating at high
pressure, high temperature, high velocity, and large bed height are
required to verify the basic design conditions. These appear still
reasonable on the basis of available data.
No particular difficulty was experienced in multiple feeding
of coal. Since in a commercial unit coal will be fed at more points
and at higher rates, more experience in these respects is required.
No general operational problems were encountered. No data, however,
are yet available on controllability of a pressurized fluidized bed
boiler. The response of the boiler to load-follow and the turndown
capability are important in plant operation and cannot be effectively
studied in smaller combustors.
51
-------
SULFUR REMOVAL SYSTEM
The sulfur removal system in fluidized bed combustion is based
on the principle that a solid sorbent can trap the fuel sulfur in solid
form as the coal is burned and prevent its release to the environment
as gaseous sulfur dioxide (SO-)• Thermodynamic analysis shows which
solids will react with sulfur dioxide under process conditions and,
therefore, defines those sorbents which must be considered for use.
The calcium-based sulfur removal process has been developed more
extensively than have those using alternative sorbents. Experimental
work to date has used limestone or dolomite sorbents as sources of
calcium carbonate (CaCO,) or calcium oxide (CaO) in the processes out-
lined in Figure 13. Primary consideration has been given to optimizing
operation of the fuel processing/sulfur removal system module.
More than 90 percent of sulfur dioxide emissions can be pre-
vented and the fuel sulfur captured in a dry solid using limestone or
dolomite as sorbents. The environmental standard of 0.54 kg (1.2 Ib)
sulfur dioxide/1.055 GJ (10 Btu) can be readily obtained. Maintenance
of this standard of sulfur dioxide pollution abatement for the combustor
operating conditions, while minimizing solid waste accumulation, has
required excess sorbent or calcium/sulfur molar feed ratios of about
2/1 for dolomite and 3/1 for limestone.
The TG analysis studies reported here and in previous reports
may be used to project the calcium/sulfur feed ratios necessary for
adequate desulfurization. The projections are summarized in Table 6.
The reduction of these ratios is a major goal in the development of
the fluidized bed combustion system.
The primary objectives of experimental programs carried out
thus far have been to determine sulfur dioxide removal capability with
calcium-based sorbents and to investigate the effect of operating con-
ditions and choice of sorbent on sulfur removal. To a limited extent
models have been proposed to represent the available data.
The most important results established, which include those
performed in parallel with those reported here, related to the effects
•52
-------
Dwg. 6229A56
U)
Fresh
Limestone *•
(or Dolomite)
High
Sulfur ^
Fossil
Fuel
Fluidized Bed
Combustor/
Sulfur Removal
System
bpent stone
Spent Stone
Snpnt Stnnp «—
Regenerated
Stone
Wasl
Stor
(Ca/S
Soent Stone ~-
DA/I An AV ^rfrrtH ••
r\eyenerdiea *—
Stone
Ca/S>l
Spent Stone Ca/S>l
Processing
Regeneration
^ Disposal/
Utilization
Disposal/Utilization
~™^^ finrliirlinn culfur
111 \\rl UUIIIIJ 9UIIUI
recovery options)
Sulfur Rich ^c..,,...!
Gas
fp Spent Stone
ie * Processing
<^1)
Regeneration
Waste
Stone
(Ca/S~l)
Sulfuric Acid
^Disposal/Utilization
(including sulfur
recovery)
Sulfur Rich
Gas
1 1
Stone
Processing
•
Solids Disposal/
"Utilization
Figure 13-Sulfur removal system process concepts
-------
Table 6
PROJECTIONS OF THE CALCIUM TO SULFUR FEED RATIOS
Pressure, 101.3 kPa (atm)
Sorbent
816°C(1500°F)
954°C(1750°F)
Pressurized, 1013 kPa (10 atm)
816°C(1500°F)
954°C(1750°F)
Limestone
Uncalcined
Calcined
2.8/1
2.2/1
2/1
in
Dolomite
Half-calcined
Calcined
2.2/1
1.6/1
1.2/1
1.6/1
1.2/1
Mean particle size of 500
Gas residence time = 1 second
»80% sulfur removal efficiency.
-------
of calcium/sulfur molar feed ratio, pressure, temperature, and gas
velocity on sulfur removal. The data have been reviewed previously.
The calcium/sulfur feed ratio at atmospheric pressure, which
achieved over 90 percent sulfur dioxide removal, was found by several
groups of investigators to be 2.5/1 for dolomite, and about 3/1 for
28 29
limestone. ' At- temperatures above 843°C (1550°F), however, sulfur
30
removal efficiency was drastically reduced. Reduction of the sorbent
particle size from 600 urn to 40 ym did not improve desulfurization
28
efficiency or calcium utilization, largely because of reduced stone
residence times. Fines recycling, however, made an important contri-
9
bution in improving desulfurization efficiency. The effect of gas
velocity was to reduce desulfurization efficiency drastically where
desulfurization efficiency was already low; but when desulfurization
efficiency was high, doubling the gas velocity had a minor effect in
reducing the fraction of sulfur dioxide captured in the fluidized bed.
The initial results on pressure at 506.5 kPa (5 atm) showed
that limestone was marginally less efficient in desulfurization than
g
at atmospheric pressure. (This effect has been confirmed; it is in
agreement with projections from TG data). Later fluidized bed results
with dolomite showed that TG analysis results demonstrating high calcium
utilization (90 percent) were reflected in calcium sulfur molar feed
31
ratios of 1/1 at high desulfurization efficiency. The decline in
desulfurizing performance noted at atmospheric pressure above 843°C
(1550°F) was not noted at 1013 kPa (10 atm) pressure, and excellent
31 8
sulfur removal performance was noted up to 954°C (1750°F). '
Laboratory studies in support of the sulfur removal system
development were carried out under this contract in three areas:
• Development of techniques for lowering the calcium/
sulfur molar ratio required for desulfurization,
thereby reducing sorbent requirements and spent
sorbent disposal problems (Appendix D).
• Development of a model to predict quantitatively the
sulfur removal in a fluidized bed combustor,
(Appendix E), and
55
-------
• Development of sorbent selection criteria
32
(Appendices C and E).
In addition, work has continued on the low-temperature
regeneration step of converting calcium sulfide to calcium carbonate
under contract to Energy Research and Development Administration (EKDA).
A brief summary of this work as it pertains to fluidized bed combustion
is given below.
Sulfur Removal
Experimental work to develop an understanding of the processes
which limit the calcium/sulfur molar ratio has continued, using the
1 2
thermogravimetric apparatus described previously ' over a range of
temperatures, 750 to 950eC (1382 to 1742°F); pressure, 101.3 to
1013 kPa (1 to 10 atm); stone particle sizes (2,000 to 500 um diameter);
stone sorbents (limestone, dolomite, alumina-based copper oxide); and
calcination conditions (half-calcined dolomites, limestone and dolomite
calcined in atmospheres containing variable partial pressures of carbon
dioxide). The thermogravimetric data suitably interpreted provide a
record of the extent of sulfation which can be achieved and the rate
of reaction as a function of sulfur loading on the stone.
The conditions under which the calcium-based sorbent has been
calcined was shown to be a critical factor in establishing the calcium
oxide utilization in subsequent sulfation of the stone, both at 101.3
33
(atmospheric pressure) and at 1013 kPa (10 atm), as shown in Figure 14.
The change from rapid to slow reaction can occur at 5 percent or 30
percent utilization of the calcium in the stone. For dolomite, calci-
nation under a relatively high partial pressure of carbon dioxide
(60 percent of the equilibrium partial pressure over calcium carbonate/
calcium oxide) increases the utilization achievable in fast sulfation
from ^40 percent of the calcium reacted to 90 percent reacted. This
effect is independent of total system pressure. The importance of this
finding is evident when the desulfurization process takes place in a
pressurized fluidized bed combustor where the carbon dioxide partial
56
-------
Curve 655535-B
.40 -
02
420-SOOpm
S.OOOppm S02,
4% 02 in No
871°C (1600° F)
Calcined in CC?
Calcined in N2
Calcined in COg
Calcined in N2
100
80
60S
•8
o
3
40
20
12 14 16 18 20
Time/Minutes
28
Figure 14 4-Comparison of pressurized sulfation of limestone and dolomite
57
-------
pressure may rise to 152 kPa (1.5 atm) at the freeboard as a result of
fuel combustion. This creates the correct conditions for forming a
calcined dolomite which may be 90 percent sulfated at a fast rate of
reaction. This, in turn, should result in excellent sulfur removal at
calcium/sulfur molar ratios close to 1.2 to 1.0 - performance which has
been demonstrated by Argonne National Laboratories (ANL) in their
31
pressurized fluidized bed combustor. Additional evidence for the
impact of calcination is presented in Appendix D.
In combination with the effect of temperature on the sulfation,
the effect of prior calcination conditions was shown to reproduce the
maximum in calcium utilization at 843°C (1550°F) at atmospheric
pressure which has been noted in fluidized bed combustion, as shown
in Figure 15. At 1013 kPa (10 atm) the results were not as clean, but
it was shown that high calcium utilization (35 percent sulfated) in
limestone can be attained at 950°C (1742°F) and 1013 kPa (10 atm)
pressure. This extends the upper limit for desulfurization with calcium
oxide from limestone at pressures to temperatures above the atmospheric
pressure limit of 850°C (1560°F). (It had previously been demonstrated
that Tymochtee dolomite remained active at temperatures up to 1010°C
[1850°F]).2
For conditions under which calcium carbonate is the stable
form of the unsulfated sorbent, it was confirmed that limestone 1359
cannot be directly sulfated and that no activation of the sorbent by
calcination and recarbonation occurs. As previously noted, however,
half-calcined dolomite is an excellent sulfur dioxide sorbent in
general (dolomite 1337, Tymochtee dolomite, Salamonie dolomite); a
half-calcined mountain dolomite (large grains) (Canaan dolomite),
however, did not show kinetic activity in sulfation and, like limestone,
it could not be activated by calcination and recarbonation. Previous
results were extended to show that over 50 percent sulfation by large
particles (2,000 urn) of half-calcined dolomite could be attained.
Half-calcined dolomite should be utilized at calcium/sulfur molar ratios
intermediate between those for activated calcined dolomite (1/1) and
normally calcined dolomite (2.5/1).
58
-------
Curve 655^57-A
0.50
^ 0.45
5 0.40
0>
fe 0.35
i °-3°
"co
| 0-25
| 0-20
cu
| 0.15
£ o. 10
o
0.05
71
CaO + S02 +!/ 2 02 — • CaS04
420 - 500 Mm
- Limestone 1359
1.01 x!05 N/m2
- 5,000ppm S02
4% 02 in N2
20 mg Stone Calcined at 900°C (1652°F) in 60% C02 in N2
- Stone Brought to 900°C (1652°F) in Pure C02 at 10 k/JV\in
TG 258, 259, 260, 261, 262
-*— -a^
,
-------
Process Modeling
In order to apply thermogravimetric data to the prediction of
sorbent performance in fluidized bed combustion, a first-order model of
the desulfurization process was developed. In the model described in
Appendix E.. , it is assumed that sulfur dioxide is liberated from coal
uniformly throughout the bed and the fraction of sulfur dioxide avail-
able for capture in each element of bed height is calculated. The
fraction of sulfur dioxide captured is then computed for a mean first-
order reaction rate between sulfur dioxide and lime in the bed. The
result is plotted in the form of mean reaction rate as a function of
calcium/sulfur molar ratio fed to the bed, or calcium utilization of
the stone. Experimental thermogravimetric data on the rate of reaction
as a function of calcium utilization intersects this curve to produce
a steady-state value for the calcium utilization and, hence, the degree
of sulfur dioxide removal. If a functional relationship (empirical or
theoretical) describes the rate of reaction as a function of calcium
utilization, then the sulfur retention can be solved directly at a
given calcium/sulfur molar feed ratio, calcium density, and gas
residence time in the combustor. Agreement between the predictions of
the model using Westinghouse thermogravimetric data and experimental
data from fluidized bed combustors was excellent.
In addition to overcoming anomalies resulting from previous
modeling studies, the modeling results correctly showed:
• The maximum temperature for sulfur dioxide removal
at atmospheric pressure
• The variation in sulfur dioxide removal with calcium/
sulfur molar ratio in the feedstock, as shown in
Figure 16
• The effects of superficial gas velocity on sulfur
dioxide retention at two different calcium/sulfur
molar ratios.
The model can be used to specify the rate of reaction required
in a given fluidized bed combustor design to achieve a desulfurization
60
-------
6S5729-A
100
c
.o
'•+Z
c
-------
target. The thermogravimetric apparatus may then be used, under the
appropriate conditions (with particular emphasis on carbon dioxide
pressure), to establish the stone utilization at which the sulfation
rate falls below that specified by the model. The required operating
conditions of calcium/sulfur feed rate is thus established for the
particular sorbent.
Further investigation of the use of limestones and dolomites
for fluidized bed desulfurization can be carried out within the frame-
work of the model developed. The use of thermogravimetric experiments
coupled with application of the model should be extended beyond the
range of sorbents so far tested (one limestone, four dolomites). The
selection criteria for reaction rate at a given utilization should be
most useful in comparative evaluations of different sorbents available
at a particular plant location.
Additional work should be carried out on the extension of the
temperature interval for desulfurization above 950°C (1742°F) at 1013
kPa (10 a tin) pressure. Results previously published showed that sulfa-
tion of Tymochtee dolomite at pressure is rapid up to temperatures of
1010°C (1850°F) (depending on the calcium/sulfur feed ratio required).
Sorbent Regeneration
Regeneration of the sulfur sorbent is a potentially attractive
option for fluidized bed combustion systems because it will:
• Reduce the sorbent requirements
• Reduce the spent sorbent to be handled by a
disposal system
• Recover the fuel sulfur
• Perhaps reduce the contribution to trace-element
emission from the sulfur-removal system (while this
might be an important advantage of regeneration over
once-through systems, it is only a postulate at
this time ).
62
-------
The disadvantages of regeneration arise from:
• The increased plant costs of the regeneration
system
• The energy penalty associated with reduction of the
stable compound calcium sulfate
• The additional complexity introduced into plant
operations.
The increased costs of regenerative as opposed to once-through
2
systems have been evaluated, and an overall economic assessment is
presented in Section IV. For high-temperature reduction of calcium
sulfate, the one-step regenerative process energy costs were approxi-
mately 13 percent higher than those for the once-through system. The
2
optimum case for once-through sulfur removal considered by Newby
(Ca/S = 1.2), which was based on projections from TG data, is now
reasonably established by fluidized bed combustion experiments. The
status of calcium/sulfur feed ratio for the regenerative system,
34
however, remains at 1/1: the advantages of using a regenerative
system in this instance are extremely dubious.
For the two-step regenerative process, in which calcium
sulfate is first reduced to calcium sulfide and then reacted with steam
and carbon dioxide, according to the equation:
CaS + H20 + C02 -»• CaC03 + H2S ,
energy cost was found to be similar to that of one-step, high-temperature
regeneration; and the capital investment, assuming a constant load design,
was also similar.
In this case there is experimental evidence that the calcium/
sulfur makeup ratio assumed (1/1) can be lowered. Laboratory work at
35
Westinghouse has shown that the reaction cycle
CaC03 + H2S ^ CaS + H20 + C02
63
-------
can be continued for thirty cycles. For a dolomite which is fully
sulfided after each regeneration, the extent of regeneration decreases
after each cycle ,as shown in Figure 17. These results agree with those
reported by Conoco Coal Development Co. for the decline in regener-
ability of calcium sulfide. If similar results can be obtained for the
two-stage process in which the calcium sulfide is produced by direct
reduction of calcium sulfate, then the calcium/sulfur makeup rate would
fall to 1/3 for the 20 percent sulfur differential between combustor
and regenerator taken as the base case.
Thermogravimetric experiments to probe the decline in regen-
erability of calcium sulfide on repeated cycling have been continued.
In these experiments dolomites from a five-state area - Ohio, Indiana,
Illinois, Kentucky, and Michigan - have been tested. None of the
stones tested gave more efficient regeneration than did the dolomite
1337 shown. Variation in the partial pressure of steam and carbon
dioxide in the regeneration case altered the initial rates of regen-
eration but not the cut-off value at which regeneration virtually ceases.
A structural study of the formation of crystallites of
sulfide and carbonate in regenerated and sulfided dolomite is continuing.
To date, none of the results indicate the mechanism of deactivation.
The sole means of achieving complete conversion of calcium sulfide to
calcium carbonate is to raise the temperature of the regeneration
reaction; at 870°C (1600°F) it was found that total regeneration is
possible. The equilibrium concentration of recovered hydrogen sulfide,
however, decreases rapidly with increasing temperatures. Temperatures
close to 650°C (1200°F) must be maintained if hydrogen sulfide yields
more than 3 percent are to be expected. Since detailed economic studies
of this regeneration scheme are being prepared, the potential for
improving the overall economics by adjusting reaction kinetics should
become clear, and further experimental work can be focused on the
sensitive areas.
64
-------
"S
t3
CD
CO
CO
O
o>
0>
Q_
. Regeneration
Sulfidation
lOAtm
704°C
10% H20
10% CO
lOAtm
871°C
0.5%H2S
10% HoO
1200-1000Mm Dolomite (1337)
TG No 307
MgO-CaS + HoO + COo- MgO-CaC03+ H2S
10
Cycle Number
Figure 17 -Cyclic regeneration of CaCO, in sulfided dolomite
-------
Sorbent Attrition
Sorbent attrition is discussed in Appendix C, and in more
detail in Volume II of Fluidized Bed Combustion Process Evaluation:
Phase 1 - Residual Oil Gasification/Desulfurization Demonstration at
Atmospheric Pressure (March 1975).
The results available from fluidized bed combustors show a
wide range of sorbent attrition rates and indicate that attrition
resistance will be a primary factor in selecting sorbents for the
fluidized bed combustor. For regenerative systems the loss rate by
attrition, rather than the decline in kinetic activity on regeneration,
may limit the lifetime of the sorbent in the system, particularly at
high sorbent recirculation rates between a combustor and regenerator
with a low-sulfur differential. These are some results which suggest
that sorbent attrition is process dependent: the rate of fines pro-
duction depends on whether the chemical environment is that of cal-
cination, combustion, or sulfur removal in a reducing atmosphere.
38
Work to date on attrition resistance has consisted of comparative
bed weight-loss measurements for a series of limestones and dolomites
during calcination and half-calcination in a 2 cm (1.3 in) fluidized
39
bed. Comparison of the data with results obtained on the Esso 1 MW
pilot-plant oil gasifier showed that stones suffering high attrition
or very low attrition behaved similarly on both units. The attrition
loss from some stones with intermediate loss rates is not predictable.
Additional work has demonstrated that the water content in
some dolomites contributes appreciably to sorbent decrepitation.
Further work on identifying the properties of sorbents, and the process
conditions to which they can be subjected to make them attrition
resistant, is necessary.
Trace Element Emission Characteristics
The emission of trace elements from the sorbent into the
effluent gas from the combustor raises environmental and corrosion
questions (as discussed in Appendix H).
66
-------
From the process viewpoint the major concern is that sufficient
sodium and potassium will be liberated into the gas stream to induce hot
corrosion in the turbine. The best approach is to use the dolomite
with the lowest alkali content: samples of dolomite from different
deposits contain from 100 to 6000 ppm (by weight) of potassium.
One alternative for preventing the deposition of sulfates is
to increase the efficiency of the sulfur removal system. If the sulfur
dioxide level in the gas is sufficiently reduced, the deposition of
liquid sodium and potassium sulfate films in the turbine can be pre-
vented. The projected sulfur dioxide concentration required to
achieve this goal is less than 200 ppm. This option is discussed in
greater detail in the section on minor-element and trace-element
emissions and in Appendix H.
There is no indication as yet that the sorbent will contri-
bute a significant, environmentally hazardous concentration of trace
elements to the atmosphere.
Spent Sorbent Disposition
The pressurized fluidized bed combustion process, as presently
conceived, results in the production of dry, partially utilized
dolomite or limestone particles up to 6 mm in size. In addition, fine
particles of sorbent and ash will be collected in the particle removal
system. The sorbent material may be either regenerated for recycling
to the fluid bed boiler for repeated sulfur dioxide removal or disposed
of in its partially utilized form in a once-through system. The former
process has the potential advantage of producing less solid waste for
disposition, but much uncertainty still exists about the regenerative
processes. Conceptually, it is possible in a regenerative process to
recycle all of the sorbent. This might involve a synthetic calcium-
based material, reconstitution of the spent sorbent, or an alternative
sorbent material. The composition of the spent sorbent depends on the
characteristics of the original stone, the coal feed, the variation in
operating temperature and pressure, as well as on once-through or
67
-------
regenerative modes of operation. The major compounds in the waste
stone to be disposed of are calcium sulfate (CaSO,), calcium oxide (or
calcium carbonate), and magnesium oxide (MgO when dolomite is used, and
calcium sulfate and calcium oxide or calcium carbonate when limestone
is used. Trace elements arising from impurities in the coal and
dolomite will also be present.
A summary of the general process options for disposition of
the spent sorbent is presented in Figure 18. Two methods of dealing
with the spent sorbent are being considered:
• Disposition of the spent stone being discharged
directly from the fluidized bed combustion process
(once-through or regenerative, Options 1 and 3)
• Disposition of the spent stone after further
processing (Options 2 and 4).
Disposition without processing includes direct disposal or utilization
of the material in soil stabilization, for example. This is the pre-
ferred option,since it does not require additional processing. Dis-
position of the spent stone after further processing is being considered
in order to develop methods for rendering the stone environmentally
acceptable for disposal, if direct disposal is not universally per-
mitted, and to investigate alternative markets for the spent stone,
such as in the form of refractory brick.
Recovery of sulfur from the spent stone for disposal or
utilization has not been considered attractive, although the waste
stone could be processed to recover the sulfur. Sulfur or sulfuric
acid (H_SO.) would be recovered from the sulfur-rich gas (SO- or H-S)
from the regeneration processes. A third general option for processing
the waste stone (Option 5) would be to react the off-gas from the
regenerator with the waste sorbent to produce a solid material for
disposal or utilization. This option does not appear to offer advan-
tages over the once-through process options.
Among the factors that will affect the disposition of the
spent sorbent are, for example, the quantity of spent sorbent, its
68
-------
Dwg. 16726*4 2
vo
rresn Lirnesione
or Dolomite
High-Sulfur
Fossil Fuel
Fluidized
DnH
ueo
Combustor
Spent Sorbent
Spent Sorbent
Spent Sorbent
Rea en P. rated
Sorbent
Cnant CArhont
bpeni Doruciu
P ono n P ratpri
Sorbent
Ca/S^l
^
Processing for
Sulfur Recovery ^.
. Spent Sorbent
Prnr.P«ing Ca/ S = 1 m
Ca/S^l
Spent Sorbent ^
Processing
.Sulfur-Rich Gas
• Sulfur -Rich Gas
M
m Sorbent m
Soent Sorbent Processing
Ca/S~ 1
Figure 18 -Sulfur removal system process alternatives
Disposal/ Utilization
Sulfur Recovery Option
Disposal/ Utilization
Disposal/ Utilization
Disposal/ Utilization
Sulfur/Sulfuric Acid
Option
(1)
<2)
<3>
(4)
ids Disposal/Utilization
-------
chemical characteristics, regulations, geographical location, and the
size of the market for respective applications. These factors are dis-
cussed in Appendix F.
Typical quantities of spent sorbent projected for disposal
from a 500 MW power plant burning a 3 wt % sulfur fuel with 95 percent
removal would range from approximately 41 Mg/hr (45 tons/hr) for a
once-through system using dolomite (Ca/S = 1.5) to approximately
19 Mg/hr (21 tons/hr) for a regenerative system utilizing a calcium/
sulfur makeup rate of 0.75. Specific quantities will depend on
operating conditions and sorbent characteristics.
The disposition of unprocessed spent sorbent is represented
by Options 1 and 3 in Figure 18. Westinghouse obtained samples of
spent dolomite from the Argonne and Exxon pressurized fluidized bed
combustion pilot plants and carried out preliminary leaching experiments
and activity tests.
Results from these preliminary tests indicate:
• Natural gypsum leachates contain approximately the
same amounts of dissolved calcium and sulfate ions
as the fluidized bed combustion spent sorbent
leachates. Both agreed relatively well with the
calcium sulfate solubility, and both exceeded the
water quality standards, 75 mg/1 for calcium and
250 mg/1 for sulfate.
• There was negligible dissolution of magnesium ions.
• Insignificant amounts of heavy metal ions were
found in the leachates.
• Fluidized bed combustion spent sorbent leachates
were alkaline, with pH = 10.6 to 12.1. It is
interesting to note, however, that the run-off
leachates showed a gradual decrease in pH with the
amount of leachates passing through.
• Temperature increase of the spent sorbent will be
negligible when subjected to the environment.
70
-------
The experimental data and an assessment of the environmental impact of
direct disposal are presented in Appendix G. The chemical composition
of the spent sorbent from once-through and regenerative processes and
its likely environmental impact require additional comprehensive
leaching tests and activity tests to determine the chemical fate of
the constituent compounds (calcium, magnesium, sulfate ion, and so on)
and trace elements which may occur in the raw sorbent or which may
accumulate during the combustion of the coal.
The direct disposal or utilization of spent sorbent may not
be possible or permitted in all cases. Thus, alternatives for spent
stone disposition must be developed to permit utilization of the
fluidized bed combustion process. It may also be possible to develop
a more attractive use for the spent sorbent through some processing
technology. Several proposed processes are presented in Appendix F.
Several potential applications for the processed spent
sorbent and for the unprocessed spent sorbent were identified:
• Soil stabilization
• Landfill
• Concrete
• Refractory brick
• Gypsum
• Municipal waste treatment
• Acid mine drainage.
Both high-temperature stone processing, including spent sorbent/fly ash
and spent sorbent/clay sintering, and low-temperature pozzolanic
activity require further study. Materials to be tested should include
the spent stone from the once-through process, spent stone from pilot
plant and prototype plant tests, and blends of these materials with fly
ash, clay, and soil. The soil stabilization tests will require deter-
mination of unconfined compressive strength, Atterberg limits, and
direct and triaxial shear strengths. The influence on agricultural
soils of surface dumping of the spent sorbent should also be investigated.
Spent sorbent should be evaluated for suitability as aggregate material
71
-------
by testing it in concrete mixes for compressive strength, splitting
tensile strength, flexural strength, and modulus of elasticity. The
interfacial behavior of the concrete and foundation soil should be
examined.
In summary, environmental problems associated with disposal
of the spent sorbents from fluidized bed combustion systems differ
favorably from those associated with disposal of lime sludges, in that
they are solids and they do not possess great water solubility. Data
indicate the spent dolomite (or limestone) can be used as dry landfill
with known civil engineering practices for controlling structural
rigidity and ground water flows. Alternatives are also available for
utilization of the spent stone. In addition, the advanced sulfur
removal systems being developed would minimize the quantity of spent
stone available and, thus, could minimize the problem.
NITROGEN OXIDE EMISSIONS
In the fluidized bed combustion of coal with in-bed desulfuri-
zation using limestone and dolomite sorbents, the operating temperature
range for the bed is from 704 to 1010°C (1300 to 1850°F). Since 1010°C
(1850°F) is well below the temperature level at which fixation of free
40
nitrogen (thermal nitric oxide) through the Zeldovich mechanism
becomes negligible [about 1538°C (2800CF)], the only significant source
of nitric oxide is from bound nitrogen in the coal.
In diffusion flames the conversion of bound nitrogen to
nitric oxide which takes place during the carbon-hydrogen-oxygen
reactions, has been found to be nearly 100 percent. At flame tempera-
tures greater than about 1538°C (2800°F), where thermal nitric oxide
formation occurs, the net conversion of bound nitrogen is less than
100 percent, since the presence of nitric oxide from bound nitrogen
suppresses the formation of thermal nitric oxide. This indicates that
the conversion of bound nitrogen in fluidized bed combustion at bed
temperatures less than 1010°C (1850°F) should be nearly 100 percent.
Experimental investigations, however, have shown that, on the contrary,
72
-------
the conversion of bound nitrogen in fluidized bed combustion is sub-
stantially less. ' '
Various investigators have attempted to determine the effect
of the fluidized bed combustion design parameters on nitrogen oxide
emission. Those parameters which have been identified as having a
significant effect on nitrogen oxide emissions are pressure level,
excess air, calcium-sulfur ratio, water vapor, carbon monoxide concen-
tration, and constraint on mixing in the bed.
Argonne National Laboratories (ANL) have measured nitrogen
31
oxide emissions levels at 810.4 kPa (8 atm), considerably less than
42
those previously measured by them at atmospheric pressure. Combustion
power test data show that emissions at 405.2 kPa (4 atm) are substan-
tially lower than those at low pressures. Similar effects of
pressure on nitrogen oxide emissions were reported at the Third Inter-
43
national Conference on Fluidized Bed Combustion.
44
Test results obtained by Exxon from both their batch unit
41
and their miniplant show that nitrogen oxide emissions are a strong
function of the percent excess air for excess air levels up to about
50 percent. Above 50 percent excess air, the nitrogen oxide emissions
are indicated to be nearly independent of excess air. The National
Research Development Corporation (NRDC) data agree well with the Exxon
data.
ANL has conducted tests to determine the effect of calcium
sulfur ratio on the nitric oxide concentration in flue gas. The
results of these tests show that increasing the calcium sulfur ratio
above a value of 1 causes a substantial increase in the nitrogen oxide
emissions from fluidized combustion of coal when operating at 810.4
kPa (8 atm), 15 percent excess air, and 788 to 899°C (1450 to 1650°F)
bed temperature. The explanation for this effect is that the
presence of sulfur dioxide in the bed either suppresses the formation
or promotes the decomposition of nitric oxide. The reduction of
nitrogen oxide to elemental nitrogen by carbon monoxide has also been
27
postulated. Water vapor may have a catalytic effect.
73
-------
Results of tests conducted by Exxon in their batch unit
indicate that constraints on mixing within the bed tend to increase
44
the nitrogen oxide emissions. When horizontal coils were replaced
by vertical ones which were more open, the nitrogen oxide emission
decreased significantly.
No data were found in the literature from which could be
discerned the effect of the nitrogen content of the coal on the nitro-
gen oxide emission level. Tests in gas-turbine combustors using
simulated bound nitrogen showed that the nitrogen oxide concentration
in the products of combustion are less than proportional to the nitro-
45
gen content of the No. 2 distillate.
Figure 19 is a composite plot of the nitrogen oxide emission
data from Exxon and NRDC, which shows that the nitrogen oxide emissions
from fluidized bed combustion of coal is well below the current EPA
emission standard over a range of excess air levels up to about
50 percent.
Extrapolation of the Exxon and NRDC data indicates that the
nitrogen oxide emissions for adiabatic fluidLzed bed combustion would
also be below the EPA emission standard of 0.302 kg NO-/GJ (0.7 Ib NO,/
6
10 Btu) input. Data from the Combustion Power Company's process
development unit indicate that the nitrogen oxide emission level
from an adiabatic combustor (excess air 200 to 300 percent) is in the
order of 0.173 kg N02/GJ (0.4 Ib N02/106 Btu) input.
In summary, the data indicate that nitrogen oxide emissions
from pressurized fluidized bed boilers and adiabatic combustors will
be well below the EPA emission standards.
PARTICULATE CONTROL
Particulate removal is critical for the successful operation
of pressurized combined-cycle fluidized bed combustion power plants.
The particulate removal system must be capable of reducing the partic-
ulate loading in the combustion off-gas to levels compatible with gas-
turbine operating conditions and environmental standards. Analyses of
74
-------
Curve 681860-A
1.0
0.8
10
Approximate Percent Excess Air
18 25 33 42
53
65
00
.a
i 0.4
c
o
• ^B
I 0.2
I
Current EPA Emission Standard
Pressure: 303.9 to 1013 k Pa (3 to 10 Atm)
1
I
I
0
3456
Percent 02 in Flue Gas
8
Figure 19-Composite plot of data for N0x emissions from fluidized
combustion of coal
75
-------
particulate control systems have been based on estimates of carry-over
from fluidized beds, of turbine tolerance for particulates, and of dust
collector performance. Unfortunately, there is no reliable estimate of
turbine tolerance for particulates or the dust loading from a commercial-
scale fluidized bed combustor. As a result, criteria for the particulate
removal system are ill defined, and more definitive experimental data
are needed.
Fluid Bed Boiler Carry-over
Estimates of carry-over have been published in earlier
O / £. O
reports. ' The basic design assumes a carry-over of 15 gm/m
(6.7 gr/scf), with particle size distributions as indicated in
Figure 20. Dust loadings of up to 67 gm/m (30 gr/scf) have been
assumed in parametric studies. Particulate emissions will include
fuel ash, unburned carbon, and sulfur sorbent. The dust loading from
the combustor will depend on the fluidized bed combustion design (e.g.
bed internals, freeboard) and operating conditions (e.g. gas velocity,
bed depth, temperature); the solids feed preparation (e.g. washed and
unwashed coal, fresh or regenerated sorbent); the solids feed charac-
teristics (e.g. coal size and ash content, sulfur sorbent size and
attrition characteristics); and sulfur removal system (e.g. sorbent
makeup rate).
Gas-Turbine Specifications
The turbine tolerance will depend on particle size, particle
physical properties, impact velocity, impact angles, and turbine
materials. Thus, the turbine tolerance will depend on the particles,
the gas-turbine design, and the operating conditions. Current estimates,
based on model studies and on extrapolation of experimental data,
indicate that gas turbines will tolerate dust loadings two to one
hundred times greater than present specifications allow. [Presently
-4 3 -4
allowed are 4.6 x 10 gm/m (2 x 10 gr/scf) in the turbine expansion
gas.] Thus, the allowable limit may be estimated to be somewhere
76
-------
Curve 656022-A
.1
1
5
15
25
35
£ 45
.2* 55
a, ??
* 65
£ 75
i85
«/>
V.
o>
a 93
96
98
99
ii
i i i i i 1
Curves 1, 2, & 3 Represent Projected
Particle Size Distributions
i i i i 1
I
I
nl
.5
3
10
20
30
40
50
60
70
80
90
95
97
98.5
10 100
Particle Size-Microns
1000
Figure 20 -Particle size distribution assumed for dust elutriated from gasifier
-------
/ O *3 / O
between 9.2 x 10 and 4.6 x 10~ gm/m (4 x 10 and 2 x 10 gr/scf).
A preliminary estimate of the turbine tolerance for particulates is
presented in Appendix J. This estimate projects a tolerance from
3 -3 -2
0.012 to 0.093 gm/m (5 x 10 to 4 x 10 gr/scf). Further work is
required to establish more precise specifications. For the present
assessment of particulate control requirements, however, it has been
-3 3 -3
assumed that 4.6 x 10 gm/m (2 x 10 gr/scf) is the allowable limit.
This limit must be associated with an allowable particle size
distribution. Using current specifications, an acceptable size distri-
bution has been projected (Table 7). Clearly, smaller particles will
be less erosive than large particles; and, consequently, heavier
loadings than these allowable limits may be accommodated, provided the
size distribution is finer.
A technique for estimating the erosive power of various size
fractions (Appendix 1) has been used to produce particle loading/
particle size distribution combinations which are considered to have
erosive characteristics equivalent to those of the allowable one. It
can be shown that a fine particle size distribution may have a loading
ten times higher than the allowable specification yet have the same
erosive characteristics (Table 7).
Environmental Standards
Allowable emission levels, 0.042 gm/MJ (0.1 lb/10 Btu-
o
equivalent to 0.115 gm/m [0.05 gr/scf] for the pressurized boiler
with 15 percent excess air) from fossil fuel plants are higher than
the limit of 2 x 10~ gr/scf imposed by gas-turbine erosion. Conse-
quently, in meeting these turbine demands the legislative requirements
on particulates will also be met.
It has been found that health hazards and haze problems are
most acute with submicron particles which contribute little to the
mass of emissions and cause no turbine problems. Future legislation
may require better control of these emissions — and this could require
a revision of the particulate control system design. On the basis of
78
-------
Table 7
ACCEPTABLE DUST LOADINGS IN EXPANSION GAS BASED
ON CURRENT WESTINGHOUSE SPECIFICATIONS
Allowable particle loading and size
distribution estimated from current
gas- turbine specifications
Size (urn)
Expansion gas
gm/ni3 (gr/scf)
Particle loading with fine size
distribution equivalent in ero-
sive nature to allowable case
Size (vim)
Expansion gas
gm/m^ (gr/scf)
0-2
2-3
3-4
4-5
5-6
6-10
10-20
20 +
0.0018 (0.0008)
0.00045 (0.0002)
0.00036 (0.00016)
0.00022 (0.0001)
0.00018 (0.00008)
0.00055 (0.00024)
0.00045 (0.00020)
0.00050 (0.00022)
0-2
2-3
3-4
4-5
5-6
6-10
10-20
20 +
0.045 (0.20)
0.016 (0.007)
0.0036 (0.0016)
0.0009 (0.0004)
0.0003 (0.00012)
0.0002 (0.00006)
TOTAL
0.00451 (0.00200)
TOTAL
0.066 (0.0292)
79
-------
current technology, however, collectors for these ultrafine particles
could be operated more effectively at low temperature and pressure.
Consequently, major revision of the high-pressure equipment design to
meet such emission requirements is not anticipated.
Process Operating Range
Gas cleaning equipment employed in fluid bed boiler facilities
will probably rely on inertial mechanisms for dust collection. Equip-
ment of this nature shows a decline in operating performance when gas
velocity through the apparatus is reduced. As a consequence, it is
necessary to examine how start-up, shutdown, and turndown will affect
gas-flow rates and, hence, dust collector performance.
The current concept for turndown involves an adjustment of
gas temperature and pressure which only slightly alters the volumetric
gas flow through the collectors. Thus, turndown is unlikely to have a
significant affect on collector performance.
Shutdown procedures will involve slumping the bed. While a
reduced gas flow may be maintained during this period, particle
elutriation rates would be low, and collector efficiency, although
reduced, will be adequate.
Start-up may involve a significant period during which gas-
flow rates are reduced to minimize heat loss from the bed. Reduced
collector performance may be anticipated during this period. If this
is not matched by an equivalent drop in elutriation, the gas cleaning
system must be installed in a modular arrangement which would allow
individual units to be shut down, maintaining design flows to the
remaining equipment.
Alternative Particulate Removal Systems
An initial stage of conventional high-temperature cyclones
will be used to collect coarse material carried over from the fluidized
beds. This will be followed by one or more stages of high-efficiency
dust collection which will reduce particulate loadings to levels com-
patible with the requirements of the expansion turbines.
80
-------
Several alternative collectors have been considered for this
duty (Table 8). In all cases, however, the available operating data
cannot be directly related to fluid bed combustion processes. As a
consequence, various systems have been selected for further experimental
evaluation. This selection is based on availability, operability, and
performance potential.
System Selection
The gas cleaning system consists of three stages:
• Primary cyclones to collect coarse material and
return material to the beds if the carbon content
is high
• Secondary collectors to remove the bulk of the
remaining fine material
• A tertiary collector to reduce the level of
ultrafine particulates to acceptable limits.
Primary Collectors
Primary collectors will be cyclones of conventional design.
Secondary Collectors
The gas stream passing to the secondary collectors will
3
contain approximately 2.3 gm/m (1 gr/scf) of dust. Most of the dust
particles will be smaller than 5 urn. Consequently, the collector must
be able to handle large quantities of dust routinely and still have
useful efficiency down to 2 ym.
Of the equipment considered, cyclones offer the most appro-
priate characteristics. Cyclones of conventional design, manifolded
as multicyclone units, are preferred,as they do not require clean
secondary gas streams to maintain efficiency.
Table 9 lists the estimated dust loading and particle size
distribution at the inlet and outlet of this collector. The overall
efficiency is approximately 50 percent with 80 percent collection of
the +2 ym fraction.
81
-------
Table 8
ALTERNATIVE PARTICULATE REMOVAL SYSTEMS
Item Type Comments
Cyclones Conventional Commercially available, performance available
Aerodyne Commercially available, under test
Shell Commercially available, performance available
Donaldson Commercially available
Filters Granular bed Pilot-plant testing complete
Porous metal Commercially available
oo
Porous ceramic
Electrostatic Research Cottrell Laboratory investigations show feasibility
-------
Table 9
ESTIMATED DUST COLLECTION HIGH-EFFICIENCY MULTI-
CYCLONE SECONDARY COLLECTOR (Stairmand Design)
Dust size
(um)
0-2
2-3
3-4
4-5
5-6
6-10
10-20
20 +
TOTAL
Inlet Loading
gm/m3(gr/scf)
0.500 (0.220)
0.114 (0.050)
0.079 (0.035)
0.054 (0.024)
0.041 (0.018)
0.057 (0.025)
0.045 (0.020)
0.014 (0.006)
0.904 (0.398)
Outlet Loading
gm/m3 (gr/scf)
0.363 (0.16)
0.045 (0.02)
0.018 (0.008)
0.009 (0.004)
0.007 (0.003)
0.007 (0.003)
0.0015 (0.0007)
0.0005 (0.0002)
0.451 (0.1989)
Tertiary Collector
The final filter stage must reduce the loading of dust from
0.46 gm/m (0.2 gr/scf) to a level acceptable in turbine expansion gas.
Westinghouse projections show that this will require a collection effi-
ciency of at least 90 percent and possibly as high as 99 percent on a
particle size distribution which is essentially all below 10 urn.
Granular bed filters show the greatest potential for meeting
this requirement, and experimental development of these filter systems
is recommended. Preliminary testing in the laboratory has shown
adequate efficiency, but more data are required.
Assessment
Simple mechanical collectors, such as cyclones, are inadequate
for meeting the dust collection requirements. Specialized high-
efficiency cyclones which require clean flows of high-pressure secondary
gas are difficult to incorporate into a power plant scheme without
seriously reducing cycle performance.
83
-------
Granular bed filters have shown acceptable efficiency. A
more detailed study of mechanical design, however, and of cleanup
cycles is required to establish their utility.
The requirement for particulate removal equipment primarily
remains dependent on the gas turbine tolerance. This tolerance depends,
in turn, on particle characteristics as size distribution and density;
on the turbine operating conditions; and on the turbine design. The
function of each particulate removal stage must also be assessed; for
example, a first-stage cyclone may not be required if solids are not
returned to the combustor. Thus, simplifying the particulate removal
system may be achieved by increasing turbine tolerance, improving
particulate removal capability, or combining component functions to
reduce the number of stages , and may result in significant cost reductions
and improved reliability.
Intensive experimental testing on large-scale, hot,
pressurized equipment is required to establish more effectively the
operating performance of dust collection equipment.
MINOR-ELEMENT AND TRACE-ELEMENT EMISSIONS
Trace emissions from the fluid bed combustion process are
important from two standpoints:
• Their effect upon the environment
• Their effect on the operability of the power plant.
The primary environmental concern is with the toxic effects
of the compounds of beryllium, mercury, fluorine, lead, cadmium,
arsenic, nickel, copper, zinc, barium, tin, phosphorous, lithium,
vanadium, manganese, chromium, and selenium. The primary technical
concern is with the corrosion of turbine metal induced by the deposi-
tion of liquid films of minor elements such as alkali-metal sulfates.
In order to predict the impact of trace element emissions
on the operability and environmental acceptability of the fluidized
bed combustion process, it is necessary to know:
-------
• The quantity and chemical form in which each trace
element of concern enters the fluidized bed
combustion process
e The chemistry of the processes by which the elements
are partitioned between the solid and gaseous
effluents from the process
• The tolerance of the system operations to the
release of trace elements within the plant
• The release of trace elements within the plant
• The emission levels acceptable for environmental
protection
• The control measures which can be taken to alter
the chemistry of the release process or the quan-
tities of trace elements released, and the
tolerance to trace elements.
The chemistry of occurrence and release of minor and trace
elements could develop into an interminable task; a prudent approach,
however, will require reference to experimental data obtained on
fluidized bed combustion and analysis of the behavior of those elements
which clearly raise environmental or technical questions.
Trace Element Quantities Input to the System
The distribution of trace elements in U.S. coals has been
47
explored by the Illinois State Geological Survey. Studies at
Westinghouse on the quantities of trace elements present in some
western coals agreed very closely with Gluskoter's work. The form in
which these elements occur is not well characterized, so that, at
present, estimates of their ultimate form are best obtained from
analysis of the solid residues and gaseous effluents from the fuel
processing system. For the dolomitic sorbent, the trace element data
are sparse and of unknown value. Initial studies of the distribution
of trace elements in dolomites are in progress at the Ohio Geological
Survey, and work on the distribution of alkali metals in dolomites
85
-------
is in progress at Westinghouse under contract to the Energy Research
49
and Development Administration. An acceleration of effort in this
area is essential in order to define sorbent selection criteria and to
assess spent sorbent and ash disposal processes.
Experimental Data on Emissions
Preliminary experiments on the fate of mercury, lead,
beryllium, and fluorine in a fluidized bed combustor have been reported
by the Argonne National Laboratory. This work is still in the
developmental stage, but it indicates that fluorine and lead were
retained by the solid products, but the quantities of mercury and
beryllium recovered were about one-third and two-thirds of the input.
44
The estimation of trace element emissions is also in progress at Exxon.
Their experimental results indicate an 86 percent retention of arsenic
and 96 percent retention of manganese. The uncertainties involved in
the analyses indicate that trace element analyses will be necessary
on condensate obtained from flue gas.
The experimental data on alkali-metal emissions are sparse
Q
but consistent. The BCURA-NRDC report indicates that ^1.5 ppm sodium
and potassium are liberated into the gas phase beyond the test cascade
from a continuous fluid bed combustor at 607.3 kPa (6 a tin) pressure
and 899°C (1650°F). The Exxon data obtained by material balance after
solids analysis indicate that 10 percent of the sodium and 20 percent
of the potassium input to their pressurized combustor escaped from the
bed in the gas phase. Even at the low sulfur dioxide levels (<25 ppm)
achieved in the BCURA combustor, the sodium concentration in the off-
gas exceeds the calculated turbine tolerance for sodium as given by
the pure sodium sulfate melt model or by the assumption that a
eutectic melt of potassium and sodium sulfate forms. The chlorine
content of the coal used was about one-third of the mean value for
U.S. coals, but even had there been a three-fold increase in chlorine
content, the turbine tolerance for pure sodium sulfate would have been
exceeded.
86
-------
Emission Chemistry and Turbine Tolerance
The components of gas turbines exposed to the hot gas
stream are made of materials that form oxide scales to protect them-
selves from oxidation. High-strength nickel-based superalloys are used
for the highly stressed rotating components. More oxidation-resistant,
but lower-strength, cobalt-based alloys are used for the stationary
components. Inlet gas temperatures over 1093°C (2000°F) are currently
used in industrial turbines. Air cooling is used to maintain metal
temperatures lower than 899°C (1650°F) so that the alloys retain suffi-
cient mechanical strength.
Metal recession rates due to oxidation alone are of the
order of 0.10 mm (0.004 in) per year on the hottest components of
current Westinghouse turbines. In the presence of alkali-metal com-
pounds, which react with sulfur dioxide and sulfur trioxide from the
combusted fuel gas, liquid films of sulfate and sulfate-chloride
mixtures can be deposited on the turbine hardware. In the sodium-
potassium-sulfur-oxygen-chlorine ' (Na-K-S-0-Cl) system, liquid films
are possible over the temperature range of 514 to 1069°C (957 to 1956°F),
which coincides with the range of temperatures encountered in the
turbine flow path. (Gases are exhausted from turbines at about
400°C [752°F].)
Molten alkali-metal compound films are dangerous because
under some conditions they attack the protective oxide scale, allowing
accelerated or catastrophic oxidation (hot corrosion) of the turbine
components. The films can also accumulate particulates that allow the
build-up of thick deposits on the metal surfaces. Aerodynamic
performance of the turbine can thus be seriously impaired.
Sodium and potassium compounds emitted into the gas stream
are potentially hazardous to the operation of the gas turbine.
Chlorides and hydroxides are volatile species and can transport sodium
and potassium from the combustor to the turbine. At hydrogen chloride
(HC1) levels exceeding 0.4 ppm by volume in the combustor gas, solid
or liquid sodium sulfate (Na-SO.) will convert to gaseous sodium
87
-------
chloride (NaCl). The hydrogen chloride level in the combustion gas
resulting from the complete release of chlorine from a low-chlorine
coal (100 ppm chlorine) exceeds this level by over a factor of ten and
is 5 ppm. In a fluid bed combustion process the predominant transport
should be by the chlorides, as discussed in Appendix H.
In the gas turbine, reactions between the chloride and the
sulfur oxides in the combustion gas will form liquid sulfate-chloride
melts on the turbine hardware if the sodium and potassium levels are
sufficiently high. These melts must be prevented because they
initiate hot-corrosion and deposit formation. The gaseous concentra-
tions of sodium, chlorine, sulfur, oxygen, and steam in the turbine at
the turbine tolerance level are interrelated and may not be considered
independently of one another. Hydrogen chloride in the turbine acts
to prevent sulfate deposits from forming or, once formed, acts to
remove them. A hydrogen chloride level of 40 ppm by volume in the
combustion gas is sufficient to prevent a liquid sodium sulfate melt
from being stable at sodium concentrations in the gas up to 0.2 ppm by
volume when present in 100 percent excess air. (Forty ppm hydrogen
chloride corresponds to complete release of chlorine from coal con-
taining 800 ppm by weight of chlorine; 0.2 ppm of sodium corresponds
to a one percent release of sodium from a coal containing 130 ppm by
weight of sodium.) The concentration both of hydrogen chloride and of
sulfur oxides (SC^ and SO.) have a strong influence on the stability of
the melt. If the sulfur dioxide level in the gas is dropped from 200
ppm to 100 ppm by volume, the concentration of hydrogen chloride
required to prevent deposition of a liquid sodium sulfate film would
drop to about 25 ppm by volume. Three additional factors must be
understood before it will be possible to define the sodium and
potassium tolerances which will prevent hot-corrosion attack in the
turbine. These are:
• The influence of the interaction between sodium and
potassium to form complex melts on the hardware.
In such melts the activities of the sodium and
88
-------
potassium are reduced, and the equilibrium con-
centrations of sodium and potassium species that
can exist in gas above the melt are also lowered.
Interaction tends to reduce the tolerable concen-
trations of sodium and potassium in the turbine
expansion gas. Westinghouse is working to
establish the magnitude of this effect and also
to establish the influence of the relative sodium
and potassium levels on the composition and melting
point of stable deposits.
• The ability of turbine stator vane and rotor blade
alloys to withstand a combustion gas containing up
to 200 ppm sulfur dioxide and up to 40 ppm
hydrogen chloride
• The shifts in the turbine tolerance which will occur
if equilibrium levels of sulfur trioxide are not
achieved and the degree to which kinetic factors
such as these influence the tolerable concentration
of sodium and potassium chlorides.
Control
Our present knowledge of the system suggests that control of
the impact of trace element release may be achieved in four ways:
• The concentrations of minor elements input to the
combustor may be controlled by using high-purity
N
dolomites.
• The concentration of trace elements may be
minimized by selecting combustor operating condi-
tions which prohibit or reduce release from the
coals and sorbents.
• The deposition of sulfates may be prevented by
- Removing chlorine in coal during pretreatment
(to minimize formation of sodium chloride during
fuel processing)
89
-------
- Using high-chlorine concentration coals (to
minimize sodium sulfate formation on turbine
components)
- Addition of hydrogen chloride to the turbine
inlet gas
- Improving the efficiency of the sulfur removal
system
- Combinations of the above.
• Additives may be introduced into the fluidized bed
to trap the alkali and trace elements as they are
released from coal and dolomite.
• Additives may be introduced into the turbine
combustor to retain the alkalis in the gas phase
as they pass through the turbine.
Assessment
• Current data on the release of sodium and potassium
into the gaseous effluent from the fluidized bed
combustor (>1000 ppb sodium plus potassium)
indicate that the projected turbine tolerance for
alkali metals (<300 ppb sodium plus potassium) may
be exceeded in normal operation without some
control techniques.
• The equilibrium turbine tolerance model currently
in use overestimates the turbine tolerance to alkali
metals because it does not account for the lowered
activities which result from eutectic melt formation.
The model, however, may underestimate the turbine
tolerance because the kinetics of the conversion to
sodium sulfate on the turbine hardware may not
permit the reactions to proceed to equilibrium
within the relevant gas residence times. The pre-
sence of calcium oxide fines has not been considered:
90
-------
their effect would be to increase the tolerance
to alkalis by lowering the sulfur dioxide partial
pressure (Appendix H).
• A variety of control options to prevent damage to
turbine hardware is available.
GAS-TURBINE OPERATION
There are four areas of concern for reliable gas-turbine
performance with pressurized fluidized bed combustion systems:
• Blade erosion
• Blade corrosion
• Relation of deposits in the turbines
• Special design features needed to duct air out
and hot gas into the turbine and to protect the
turbine from localized containment concentration.
Blade Erosion
In turbine stages in which all particulates may be safely
considered to act as solids and where low-melting liquid films cannot
form, turbine vane and blade erosion may limit turbine life. Empiri-
cally, turbines operated at turbine inlet temperatures below 680°C
(1250°F) experienced erosion (as opposed to corrosion) damage.
Several gas turbines have been built and tested which utilize
dust-laden gases. These results are summarized in Table 10.
Table 10 does not permit a design correlation of erosion with
dust loading for the fluidized bed system. The tolerance of the
turbine from the standpoint of erosion depends on:
• The physical characteristics of the particles -
which depend heavily on the fuel processing system
involved
• The velocities of particulate impact - which depend
on turbine size and design
• The size distribution of the particles escaping
the particulate cleanup system.
91
-------
Table 10
EROSION OF TURBINE BLADES
Facility
Dust Loading, gm/nH
(gr/103 ft3)
Comment
Locomotive Gas Turbine
>2.8
Bad erosion-4000 hr of
operation with 3 sets
of blades
USBM Tests,
Locomotive GT
2.8
(120)
Erosion-limited blade
life estimated to be
5000 to 7000 hr
Australian Joint Coal
Board-Ruston Hornsby GT
German Blast-Furnace GT
Westinghouse/USS(Steel)
Blast-Furnace GT
Donaldson Air Cleaners
fit to small GT
Helicopters
0.042
(18)
0.0016
(0.7)
0.0028
(1.2)
0.0023 to 0.012
(1 to 5)
Some erosion-several
thousand hr of HP blade
life estimated
No erosion
No erosion
No erosion
Hovercraft
0.016 to 0.069
(7~£5 30)
No erosion
Trucks
BCURA Fluid Bed
Combustor Test
<0.0037
0.16 to 3.7
(70 to 160)
No erosion
No erosion
With 99% particulate removal of predominantly large particles, 30 to
170 um.
92
-------
Both the locomotive gas turbine and the BCURA fluid!zed bed
tests had about the same average particle size - the locomotive gas
turbine had particles somewhat smaller than 5 ym diameter, and the
average particle size of the BCURA particles ran between 6 and 8 urn.
With respect to particle structure, the fly ash in the locomotive tests
was a fused glassy particle resulting from exposure of ash with about a
1204°C (2200°F) softening temperature to a fire zone in the combustor
s
of 2204°C (4000°F), and the particles in the BCURA tests were somewhat
friable platelets that were exposed to peak temperatures well below the
softening point of the fly ash [a 816°C (1500°F) bed temperature
compared to a 1204°C (2200°F) softening temperature]. The BCURA parti-
cles contained both coal-ash mineral matter and denser sulfated dolomite
particles, the comparative erosivity of which is not yet well known.
Particle velocities at impact differ between the turbines tested and
what would be expected in large machines installed in a fluid bed com-
bustion power plant. Modern gas turbines in electrical utility services
have gas velocities leaving first-stage stator vanes in excess of
609.6 tn/s (2000 ft/sec). In the older or smaller machines gas velocities
less than 457.2 m/s (1500 ft/sec) are more typical. The particle
velocities on impact depend on both the gas velocity and the turbine
flow-path design, which influences the ability of the particulates to
follow the changes in gas velocity. The lag of particle velocities
behind the gas velocity variation gives rise to large particle velocities
with respect to blading in the later stages of the turbine.
In Appendix J blade erosion data from the Australian Direct
Coal-Fired Turbine Project are extrapolated to indicate permissible
particulate loadings to prevent excessive erosion damage. A rule of
thumb dependence of erosion damage on gas velocity exiting from the
stator vanes is used to account for the differences in gas velocities.
The results of a geometric scaling study are employed to account for the
decrease in the number of impacting particles, the increase of impact
velocity, and the grazing angles of particulate impact that occur in a
large-sized machine. No allowance for the differences in the erosivity
93
-------
of the particulates is made to provide a safety factor in these esti-
mates. In other words, particulates from a pressurized fluidized bed
combustor system are projected to be less erosive than the fused ash
from the Australian experience. The estimates show that a modern 60 MW
turbine should be able to operate with an acceptable level of erosion
damage on turbine expansion gas containing particulate loadings between
9 x 10 and 6 x 10 kg dust/kg of expansion gas. This corresponds to
3 -3 -2
a particulate loading of 0.012 to 0.093 gm/m (5 x 10 to 4 x 10
gr/scf) of expansion gas. These particulate levels are 10 to 100 times
the current Westinghouse standard for particulate matter when the fuel
gas particulate standard is converted to the exhaust gas by including
the diluting effect of the excess air used to obtain a satisfactory
turbine inlet temperature. The allowable particulate loadings are less
than the air pollution environmental protection particulate limits,
0.043 kg dust/GJ fuel (0.1 Ib dust/10 Btu of fuel input to a stationary
power source) which for a fluid bed combustion system is equivalent to
3 -2
a particulate loading of 0.115 gm/m (5 x 10 gr/scf) of particulate
matter . A turbine erosion-deposition damage model using calculated
particle impact velocities and angles and measured particle impact data
is being developed to define better the turbine particulate tolerance.
Blade Corrosion
The first tests (1951 through 1959) on coal-fired gas turbines
were on small rating turbines aimed at locomotive usage. The problem
of blade erosion dominates these investigations. The gas-turbine
temperatures were considerably below those of today's machines, and
corrosion was of secondary consequence.
Work on corrosion directly applicable to fluidized bed com-
bustion is in progress at BCURA on pressurized fluidized bed combustion
operating at 788 to 949°C (1450 to 1740°F). BCURA cites an exhaust gas
with contaminant levels of 0.4 to 0.7 ppm of sodium-potassium alkali at
bed temperatures of 788 to 816°C (1450 to 1500°F) and contrasts this
with levels of 10 to 40 ppm of sodium-potassium alkali in the exhaust
gas from conventional pulverized-coal burning plants operating with
94
-------
flame temperatures in excess of 1649°C (3000°F). The large difference
in alkali vapor present in the exhaust gases from the fluidized bed and
from the pulverized-coal burning is attributable to the much lower
release of alkali at the lower temperatures in the fluidized bed.
Hot corrosion is essentially prevented if liquid sulfate-
chloride melts are prevented from forming on the components so that
reactions to penetrate and remove the protective oxide or to prevent
the oxide from reforming cannot occur.
Appendix J reports the results of chemical equilibrium and
calculations to establish the maximum concentrations of sodium compounds
that can exist in the turbine expansion gas without forming liquid
sodium sulfate films. The tolerable concentrations are shown to be
strong functions of the sulfur dioxide and hydrogen chloride levels in
the gas stream. If sodium were the only alkali-metal compound present,
the turbine expansion gas could contain between 0.05 and 0.2 ppm of
sodium compounds by volume with 90 percent sulfur removal, depending on
the hydrogen chloride level in the gas. The higher tolerance could
correspond to a hydrogen chloride concentration in the coal gas of
about 40 ppm by volume that obtained by burning a high (0.8 percent)
chlorine content U.S. coal.
The presence of potassium compounds in the' turbine expansion
gas lowers the allowable tolerance further, as discussed in Appendix H.
The formation of the eutectic melt between potassium and sodium sulfate
[melting point 832°C (153°F)] would reduce the tolerance for sodium by
a factor of four to between 0.01 and 0.05 ppm with 90 percent sulfur
removal. The addition of chlorides to the melt further depresses the
melting point and the allowable sodium compound concentrations. Further
work is required to establish:
• The range of melt composition that can exist in each
turbine stage and the corresponding gas phase
composition tolerances
• The degree of supersaturation that can exist in the
turbine and the effect of supersaturation on the
turbine tolerance
95
-------
• The effect of process residence time and process
temperatures on the alkali-metal content of the
gases leaving fluid bed combustion processes in
order to design a process capable of meeting the
turbine tolerance
• The technical feasibility of limiting alkali release
in the fluid bed combustors through hydrogen chloride
level control
• The technical and economic feasibility of using
higher hydrogen chloride levels and lower sulfur
dioxide levels in the turbine to enable the turbine
tolerance to meet process capability.
Deposition
If the alkali-metal compound tolerances are met, one is
reasonably assured, because of the low-combustion temperatures in a
fluid bed combustion system, that liquid films will not be present on
either the turbine hardware or on the surface of the particles.
Deposits cemented by alkali-metal compounds should not form. Sintering
of deposits will not be accelerated by the presence of the liquid.
Deposits resulting from the impaction and dry sintering of
the mix of fine particulates can still occur, especially in the first
stages of the turbine where blade-metal temperatures may exceed 870°C
(1600°F) in some places. We are not yet in a position to define par-
ticulate tolerances to limit deposit growth to acceptable rates. We
need to understand the influence of blade temperature and particulate
arrival rates on the growth and sintering of deposits and to establish
rates of erosion of deposits so that deposit formation and removal
processes are quantified.
Deposit formation is less of a problem at low-turbine inlet
temperatures. Turbine operating experience had indicated that below
677°C (1250°F) erosion rather than deposition may become the factor
that limits turbine life.
96
-------
Turbine Design
To preserve high efficiency in a pressurized fluid bed corn-
bus tor system, hot gases [1013 to 1520° kPa (10 to 15 atm), 871 to
927°C (1600 to 1700°F)] leaving the dust collection system must be
transferred directly to the gas turbine for expansion. Various design
configurations to accommodate thermal expansion, to control leakage,
and to provide a uniform distribution of particles over the flow
channel of the turbine have been suggested. Operating experience is
available from European compound-cycle power plants utilizing one of
two high-pressure connections to the turbine. These installations have
generally delivered hot gas at turbine inlet temperatures between 700
and 760°C (1300 and 1400°F); in other words, about 300°C (500°F) below
the turbine inlet temperatures currently used in electrical utility
gas turbines.
Further work is needed to develop reliable and economical
transfer pipe designs. Tests on an integrated fluid bed combustor-
turbine system at an appropriately large scale will be needed.
A turbine design is needed that:
• Provides for uniform distribution of the dust-laden
gas over the inlet flow channel
• Directs particle flows in blade and vane wakes to
avoid raising the erosion and deposition potential
of dust at blade and vane roots
• Uses stepped sidewalls, carbide wear-resisting
inserts, and/or cooling air injection as appropriate
to protect blade and vane roots from erosion damage
• Appropriately thickens and hard-faces blade tips
to resist erosion damage
• Incorporates spray systems and drains, and provides
for injection and removal of milled nut shells (or
equivalent) for washing and cleaning of blade and
vane surfaces without the need to open the turbine
97
-------
• Lowers the velocity of gases in the turbine, if required,
to achieve satisfactory erosion life.
In order to realize the full potential of pressurized fluidized
bed combustion systems, work must continue to:
• Develop commercial gas-turbine designs
• Carry out analytical and laboratory tests to under-
stand turbine tolerance to corrosion, erosion, and
deposition
• Obtain data on large-scale integrated fluidized bed
combustor, particulate control, and turbine test
systems.
98
-------
VI. SYSTEM DEVELOPMENT
Pressurized fluidized bed combustion systems have the poten-
tial to meet environmental requirements at energy costs lower than any
other competitive system. This assessment is based on the available
experimental data and on systems studies which have been carried out by
Westinghouse and other investigators in the U.S. and abroad. In order
to realize this potential further experimental work is required to:
• Investigate potential problem areas
• Establish design criteria for commercial fluidized
bed combustion plants
• Develop operation and control procedures
• Test plant components - for example, fluidized bed
combustor designs, coal feeding, particulate
removal, gas turbine, instrumentation, materials
• Investigate environmental impact.
A variety of test facilities has been proposed by different
organizations to advance the development of pressurized fluidized bed
combustion systems. Development work is required in three primary
areas: commercial-scale fluidized bed combustor/component test facility,
gas-turbine tolerance and performance, and environmental impact. A
fluidized bed combustion test facility and a gas-turbine corrosion/
erosion pilot-plant test rig and program are addressed in this report.
FLUIDIZED BED COMBUSTION TEST FACILITY
A pressurized fluidized bed boiler development plant was con-.
ceived to demonstrate pressurized fluidized bed boiler operation under
2
a previous contract to the EPA. Preliminary designs, cost estimates,
experimental programs, schedule, and program alternatives were reported.
During the period of this contract, Westinghouse was invited to submit
99
-------
a proposal for a program to design, construct, and operate a multi-
purpose environmental test facility. A proposal was submitted in
November 1974. The work in this proposal to extend the original develop-
ment plant concept was not carried out as part of this Westinghouse-EPA
contract; since such a facility is considered important for the efficient
development of pressurized fluidized bed combustion, the basic philosophy
and plant concept is presented in Appendix M.
The test facility includes two pressurized fluidized bed
combustor modules served by one common set of auxiliaries. One module
is designed as a fluidized bed boiler, and one is designed to operate
as a recirculating bed boiler or an adiabatic combustor. The utiliza-
tion of two modules permits the real-time test capability to be increased,
with only an incremental increase in plant cost, and the added capability
of simulating multibed operation.
The size of a test facility is always a critical decision,
both from a technical and an economic view. The size of a test facility
should be determined on the basis of the problems to be investigated,
an assessment of what phenomena are critical to understanding the
problem areas, and the risk permitted, before proceeding to the next
stage of development. The philosophy behind building the test facility
is to provide information which will permit direct application of
pressurized fluidized bed combustion systems on a commercial scale. The
primary problems identified include performance of a commercial-size
fluid bed unit (e.g., assure resolution of tube configuration to avoid
operating problems such as temperature distribution and tube vibration),
performance of the gas turbine (e.g., understand particle erosion and
alkali-metal deposition phenomena), and performance of particulate
removal equipment (particularly the final stage of cleanup). The size
selected for the proposed test facility is a commercial size fluidized
bed, with a coal feed rate up to 10,000 kg/hr (approximately 22,000 Ib/hr).
The basis for this selection is discussed in Appendix M. The fluidized
bed combustor operating conditions for the proposed facility, the prior
development plant, and a commercial-scale pressurized boiler and adia-
batic combustor are summarized in Table 11.
100
-------
Table 11
FLUIDIZED BED COMBUSTION SYSTEM DESIGN PARAMETERS
Commercial bed designs
Pressurized fluid bed
boiler <«. »)
Adiabatic comb us tor 0>)
Development plant
design (c)
Fluidized bed combustion tesc facilicv
Boiler module
Adiabatic combustor
module
Operating Conditions
Pressure. kPa(atm)
Temperature, "C(°F)
Gas Velocity, m/s(ft/s)
Excess Air, % of
1013(10)
760-954(1400-1750)
1.71-2.74(5.6-9.0)
1013(10)
760-954(1400-1750)
1.8(6)
101.3-2026(1-20)
704-1093(1300-2000)
(6-15)
101.3-2026(1-20)
704-1093(1300-2000)
1.8-4.6(6-15)
1013(10)
704-954(1300-1750)
1.8(6.0)
stolen iometric
Bed Depth, m(ft)
Freeboard, m(ft)
Coal Size, mm (in)
Sorbent Size, mm (in)
Ca/S Molar Feed
Coal Feed Rate kg/hr
(Ib/hr)
Design Parameters
Bed Area, m2(ft2)
Bed Height /Diameter ,m(ft)
Heat Transfer Coefficient,
U/m2-0K(Btu/ft2-hr-°F)
Tube Packing , Z bed
volume
10-100
3.35-4.36(11.0-14.3)
1.53-2.31(5-7.6)
6.35 x 0(1/4 x 0)
up to 6.35(1/4)
top size
3-6
9.78 x 10*(2.16 x 105)
(300 MW plant)
3.25(35)
2-2.5(6.6-8.2)
283.5(50)
17-22.5
300-360
2.0(6.6)
2.44(8)
3.18 x 0(1/8 x 0)
up to 3.18(1/8)
top size
3
6.36 x 10* (1.40 x 103)
(200 MW plant)
10.5-64.5(113-695)
0.2-0.5(0.65-1.6)
up to 100
1.22-9.15(4-30)
1.22-12.80(4-42)
6.35 x 0(1/4 x 0).
up to 6.35(1/4)
top size
up to 6
1.0 x 10* (2. 2 x 10*)
3.25(35)
1.5-3.0(4.9-9.8)
2.83.5(50)
17-22.5
up to 160
1.22-9.15(4-30)
1.22-12.80(4-42)
6.35 x 0(1/4 x 0)
up to 6.35(1/4)
top size
up to 6
up to 1.0 x 10
(2.2 x 10*)
3.25(35)
1.5-3.0(4.9-9.8)
283.5(50)
10-30
300-360
2.0(6.6)
up to 7.3(24)
3.18 x 0(1/8 x 0)
up to 3.18(1/8)
top size
1-6
4.8 x 10-» (1.06 x 10")
10.5(113)
0.5(1.6)
Archer, D. H., et al. Evaluation of the Fluidized Bed Combustion Process. Office of Air Program. Westinghouse Research Laboratories.
Pittsburgh, Pa. OTIS PB 211-494, PB 212-916, PB 213-152. November 1971. (318 MWe plant capacity).
Keairns. D. L., et al. Evaluation of the Fluidized Bed Combustion Process. Office of Research and Development. Environmental
Protection Agency. Westinghouse Research Laboratories. Pittsburgh, Pa. EPA-650/2-3-73-048a. Contract 68-02-0217.
December 1973. Vol. I. (200 MWC plant capacity).
clbid. Vol.' III. (10 to 30 MWe equivalent capacity).
-------
The test facility also provides for tests on:
• Commercial-scale regenerative processes, once-through
processes, and waste sorbent processing for by-product
utilization or disposal with limestone/dolomite
sorbents or alternative sorbents
• Primary, secondary, and tertiary particulate removal
equipment
• Gas-turbine performance, utilizing a rotating, multi-
stage turbine and several stationary test passages
• Auxiliary systems such as solids feeding, on-line
gas composition monitoring, etc.
• Operation and control philosophies
• Advanced system and component concepts.
The estimated installed cost (December 1974) for the facility
is $21 million.
GAS TURBINE CORROSION/EROSION PILOT-PLANT TEST RIG
Exxon Research and Engineering is under contract to EPA for
the design, construction, and operation of a high-pressure, fluidized
bed boiler, 0.63 MW-equivalent miniplant to obtain information for the
design of a high-pressure fluidized bed boiler demonstration plant.
Westinghouse was responsible for designing and constructing an erosion/
corrosion test rig for installation in the discharge line from the mini-
plant. The test rig has been constructed and shipped to Exxon and a
test program developed. The technical basis for the design of the
erosion/corrosion test rig, the detailed design of the test rig, the
apparatus and procedures for sampling particulates, a plan for tests to
be conducted in the test rig, and an analytical procedure for inter-
preting the test results are presented in Appendix K.
The plant operates at 1010 kPa (10 atm) pressure with a
product gas flow rate of 0.739 kg/s (1.63 Ib/sec). Two stages of
cyclone separators are located in the discharge line from the combustor.
102
-------
This is not expected to give dust loadings as low as the projected
turbine requirements. Another final particulate collection stage is
scheduled to be included which should provide an improved simulation of
the expected particle loading.
Two concepts were considered for the test rig: a cascade-type
unit patterned after the design used by BCURA and a straight-through
test unit (see Figure 21). Since the results obtained on a cascade rig
for the Exxon miniplant would be of questionable value because of the
reduced scale, and simultaneous testing of several vane and blade
materials would not be possible, it was concluded that a cascade-type
erosion test rig was not feasible for this application. A straight-
through erosion/corrosion test rig, therefore, was designed for the
Exxon miniplant conditions.
The straight-through erosion test rig consists of a bell-mouth
nozzle followed by a length of straight duct to provide adequate time
for the particles to accelerate to near gas-stream velocity ahead of
the target. Here, again, a 6.35 mm (1/4 in) diameter cylinder is con-
sidered to be the minimum practical size for the erosion target.
A 6.35 mm (1/4 in) diameter target located in the straight section would
limit the velocity to less than 457 m/s (1500 ft/sec) because of choking
in the plane of the target. Since this velocity is too low, the erosion
target is located in the diffuser section, where the gas stream velocity
will reach about 579 m/s (1900 ft/sec) at the design conditions for the
Exxon miniplant. '
Two types of erosion targets will be utilized — the cylindrical
rod and the wedge target. The cylindrical target has the advantage,
theoretically, of giving erosion as a function of impact angle over a
full spectrum of impact angles (from 0 to 90 degrees), whereas the wedge-
type target will give erosion data for a relatively narrow range of
impact angles. This is particularly significant since it is anticipated
that erosion will be taking place in the oxide layer, and there is no
information available on the impact angle for the maximum erosion rate
of the oxide layer. With the wedge-type probe, tests with a minimum of
103
-------
Dwg. 6358A82
122-183m/s
(400-600 ft/sec)
Cascade-type erosion test rig
396-610 m/s
(1300 - 2000 ft/sec)
Straight-through erosion test rig
Figure 21 -Erosion test rig concept
104
-------
three angles would be necessary to determine the impact angle for the
maximum erosion rate. An advantage of the wedge target is that it gives
a more precise measurement of the erosion rate at the impact angle of
maximum erosion rate. Measurement techniques are reviewed in Appendix K.
A high-pressure, high-temperature particulate sampling system
for use in the Exxon miniplant erosion/corrosion tests has been designed
and constructed. A sampling probe and sampling system were designed and
shipped to Exxon for use in the erosion test program.
The proposed test plan utilizes both cylindrical and wedge
targets, cooled and uncooled target specimens, and four different
materials. The initial test plan is for 12 test runs. A procedure was
also developed to analyze results from these tests.
In summary, the straight-through erosion/corrosion test rig
designed for installation in the Exxon Research and Engineering high-
pressure fluidized bed boiler miniplant will simulate the most severe
conditions of impact velocity, impact angle, and temperature to be found
in a gas turbine in the first generation -high-pressure fluidized bed
boiler. However, the predicted volumetric concentration of particulates
in the discharge stream from the miniplant may be more than 25 times
greater than the level anticipated in the fuel-scale plant with the
present particle collection system. Current plans to add another collec-
tion unit will help alleviate this problem.
105
-------
VII. REFERENCES
1. Archer, D. H.,et al. Evaluation of the Fluidized Bed Combustion
Process. Vols. I-III. Final Report. Office of Air Pollution.
Westinghouse Research Laboratories. Pittsburgh, Pennsylvania.
November, 1971. Contract 70-7. NTIS PB 211-494, PB 212-916, PB 213-152.
2. Keairns, D. L. et al. Evaluation of the Fluidized Bed Combustion
Process — Vol. I and II: Pressurized Fluidized Bed Combustion
Process Development and Evaluation, and Volume III: Pressurized
Fluidized Bed Boiler Development Plant Design. Environmental
Protection Agency. Westinghouse Research Laboratories. Pittsburgh,
Pennsylvania. December, 1973. Publication No. EPA-650/2-73-048a,
b and c. NTIS PB 231-162, PB 231-163, and PB 232-433.
3. Miniplant Tests. Under contract to Environmental Protection Agency.
Exxon Research and Engineering Company. Linden, N.J. Contracts
68-02-1451 and 68-02-1312. 1975.
4. Keairns, D.L., R.A. Newby, B.W. Lancaster, and D.H. Archer. Sulfur
and Particulate Emissions Control from Pressurized Fluidized Bed Com-
bustion Systems. Proceedings of Fluidized Combustion Conference.
Vol. I. Institute of Fuel Symposium Series No.l. 1975.
5. Keairns, D. L., W. C. Yang, J. R. Hanun, and D. H. Archer. Fluidized
Bed Combustion Utility Power Plants - Effect of Operating and Design
Parameters on Performance and Economics. Proceedings of the Third
International Conference on Fluidized Bed Combustion. Hueston Woods,
Ohio. 1972.
6. National Air Pollution Control Assocation - National Coal Board.
Information Exchange Meeting in United Kingdom. April 20-24, 1970.
7. Hoy, H. R.,and J. E. Stantan. Amer. Chem. Soc. Div., Fuel Chem.
Prepr. 14(2):59, 1970.
106
-------
8. Pressurized Fluidized Bed Combustion. Report No. 85. Interim No. 1.
Office of Coal Research. National Research Development Corporation.
London, England. 1974.
9. Reduction of Atmospheric Pollution. Final Report. Vols. 1-3.
Office of Air Programs. National Coal Board. London, England.
September 1971. PB 210-673, PB 210-674, PB 210-675.
10. Energy Conversion from Coal Utilizing CPU 400 Technology. Interim
Report No. 1. Office of Coal Research. Combustion Power Company•
November 1974.
11. Rincliffe, R. G. The Eddystone Story. Electrical World. March 11, 1963.
12. Keairns, D. L. Private Communication to P. P. Turner. July 15, 1974.
13. Keairns, D. L. Private Communication to R. R. Hangebrauck.
July 31, 1974.
14. Harboe, H. Coal for Peak Power. Stal-Laval, Ltd. (Presented at ACS
National Meeting, Chicago, August 1973). Div. of Fuel Chem. Preprints.
18 (4).
15. Fraas, A. P. Potassium-Steam Binary Vapor Cycle with Fluidized-Bed
Combustion. (Presented at Annual AIChE Meeting, New York,
November 1972).
16. Rossbach, R. S. Final Report of Joint NASA/OCR Study of Potassium
Topping Cycles for Stationary Power. GESP 741. NASA Lewis. General
Electric Company. Cincinnati, Ohio. November 13, 1973.
17. Fraas, A. P. Concept Preliminary Evaluation of Small Coal-Burning
Gas Turbine for Modular Integrated Utility System. Atomic Energy
Commission. Oak Ridge National Laboratory. Department of Housing
and Urban Development; Office of Coal Research. May-Sept. 1974.
NTIS PB 239-109.
18. Bagnulo, A. H. et al. Development of Coal-Fired Fluidized Bed
Boilers. Final Report 14-01-0001-478. Office of Coal Research.
Pope, Evans and Robbins. Alexandria, Virginia. 1970.
19. Multi-Cell Fluidized-Bed Boiler. Contract No. 14-32-0001-1237 .
Office of Coal Research, Department of the Interior. Pope, Evans
and Robbins. Alexandria, Virginia.
107
-------
20. Reduction of Atmospheric Pollution by the Application of Fluidized
Bed Combustion and Reduction of Sulfur-Containing Additives. Annual
Report. Environmental Protection Agency. Argonne National Laboratory.
June, 1973. Publication No. EPA-R2-73-253. Interagency Agreement
EPA-1AG-0020.
21. Robison, E. G. et al. Study of Characterization and Control of Air
Pollutants from a Fluidized-Bed Combustion Unit. Environmental
Protection Agency. Pope, Evans and Robbins. Contract No. CPA 70-10.
1972.
22. Gordon, J. S. et al. Study of Characterization and Control of Air
Pollutants from a Fluidized-Bed Boiler — The SO- Acceptor Process.
Final Report. Environmental Protection Agency. Pope, Evans and
Robbins. Alexandria, Virginia. Contract No. CPA 70-10. July 1972.
NTIS PB 229-242.
23. Jonke, A. A. et al. Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion. Argonne National Laboratory
Annual Report ANL-ES-CEM-1001. July 1968 to June 1969.
24. Jonke, A. A., et al. Reduction of Atmospheric Pollution by the
Application of Fluidized Bed Combustion. Report ANL/ES-CEN-1004.
Environmental Protection Agency. Argonne National Laboratory.
June 1971.
25. Skopp, A. et al. Studies of the Fluidized Lime-Bed Coal Combustion
Desulfurization System. Final Report. Environmental Protection
Agency. Esso Research and Engineering. Linden, N.J. 1971.
NTIS PB 210-246.
26. Mammons, G. A., and A. Skopp. A Regenerative Limestone Process for
Fluidized-Bed Coal Combustion and Desulfurization. Final Report.
Environmental Protection Agency. Esso Research and Engineering.
Linden, N.J. February 28, 1971. PB 198-822.
27. Hoke, R. C. et al. A Regenerative Limestone Process for Fluidized
Bed Coal Combustion and Desulfurization. Environmental Protection
Agency. Esso Research and Engineering. Linden, N.J. EPA-650/2-74-001.
January 1974.
108
-------
28. Jonke, A. A. et al. Reduction of Atmospheric Pollution by the
Application of Fluidized Bed Combustion. Environmental Protection
Agency. Argonne National Laboratory. ANL/ES-CEN-1002, 1970.
29. Skopp, A. et al. Fluid Bed Studies of the Limestone Based Flue Gas
Desulfurization Process. Final Report to NAPCA. Esso Research and
Engineering. Linden, N.J. PH 86-67-130. 1969.
30. Moss, G. The Fluidized Bed Desulfurizing Gasifier. Proceedings of
the 2nd International Conference on Fluidized Bed Combustion. Hueston
Woods, Ohio. 1970. Published by EPA as AP-109.
31. Vogel, G. J. et al. Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion. Environmental Protection
Agency. Argonne National Laboratories. EPA-650/2-74-057. 1974.
32. O'Neill, E. P. and D. L. Keairns. Selection of Calcium-Based
Sorbents for High-Temperature Fossil Fuel Desulfurization. Westinghouse
Research Laboratories. (Paper presented at 80th National AIChE
Meeting, Boston, September 7-10, 1975).
33. O'Neill, E. P., D. L. Keairns and W. F. Kittle. A Thermogravimetric
Study of the Sulfation of Limestone and Dolomite - The Effect of
Calcination Conditions. Proceedings of the annual meeting of the
North American Thermal Analysis Society. Trent University.
Peterborough, Ontario. June 8-12, 1975.
34. Robinson, E. B. et al. Characterization and Control of Gaseous
Emissions from Coal-Fired Fluidized Bed Boilers. Interim Report.
NAPCA. Pope, Evans and Robbins. October 1970.
35. Keairns, D. L., E. P. O'Neill, D. H. Archer. Sulfur Emission Control
with Limestone/Dolomite in Advanced Fossil Fuel Processing Systems.
Proceedings of Symposium: Environmental Aspects of Fuel Conversion
Technology. May 1974. St. Louis. EPA 650/2-74-118. 1974.
36. Curran, G. P. et al. Production of Clean Fuel Gas From Bituminous
Coal. Environmental Protection Agency. Consolidation Coal Co.
EPA 650/2-73-049. December 1973.
37. Advanced Coal Gasification System for Electric Power Generation.
R&D Report No. 81. Interim Report No. 2. Energy Research and
Development Administration. Westinghouse Electric Corporation.
July 1, 1973 - June 30, 1974.
109
-------
38. Keairns, D. L. et al. Fluidized Bed Residual Oil Gasification/
Desulfurization at Atmospheric Pressure. Vols. I and II. Environmental
Protection Agency. Pittsburgh, Pa. Westinghouse Research Laboratories.
EPA 650/2-75-027a and b. March 1975. NTIS PB 241-834 and PB 241-835.
39. Craig, J.W.T. et al. Chemically Active Fluid Bed Process for Sulfur
Removal during Gasification of Heavy Fuel Oil — Second Phase.
Environmental Protection Agency. Esso Research Centre. Abingdon,
England. EPA 650/2-73-039. November 1973.
40. Zeldovich, J. The Oxidation of Nitrogen in Combustion and Explosions.
Acta Physicochira. URSS, 2J.(4) 577 (1946).
41. Nutkis, M. P. Pressurized Fluidized Bed Coal Combustion. Proceedings
of International Fluidization Conference, (Engineering Foundation
Conference). Asilomar Conference Grounds. California. June 15-20, 1975.
42. Jonke, A. A. et al. Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion. Annual Report.
July 1971-June 1972, ANL/ES/CEN-1004 (1972).
43. Proceedings of Third International Conference on Fluidized Bed
Combustion. Environmental Protection Agency. EPA 650/2-73-053.
December 1973.
44. Ruth, L. A. Combustion and Desulfurization of Coal in a Fluidized
Bed of Limestone. Exxon Research and Engineering Co. Proceedings
of International Fluidization Conference (Engineering Foundation
Conference). Asilomar Conference Grounds. California. June 15-20, 1975.
45. Dilmore, J. A. Nitric Oxide Formation in the Combustion of Fuels
Containing Nitrogen in a Gas Turbine Combustor. ASME Paper
74-GT-37.
46. Clean Power Generation from Coal. Final Technical Report. Office
of Coal Research. Westinghouse Research Laboratories. Pittsburgh, Fa.
Contract 14-32-0001-1223. April 1974. NTIS PB 234-188-1.
47. Ruch, R. R., H. J. Gluskoter, et al. Occurrence and Distribution
of Potentially Volatile Trace Elements from Coal. Environmental
Protection Agency. Illinois State Geological Survey. EPA 650/2-74-054.
July 1974.
48. Streib, D. Ohio Geological Survey. Private Communication.
110
-------
49. Advanced Coal Gasification System for Electric Power Generation.
Interim Report No. 3. Energy Research and Development Administration.
Westinghouse Electric Corporation. July 1, 1974 - June 30, 1975.
50. Vogel, G. J., A. A. Jonke et al. Reduction of Atmospheric Pollution
by the Application of Fluidized Bed Combustion and Regeneration of
Sulfur-Containing Additives. Environmental Protection Agency.
Argonne National Laboratories. EPA 650/2-74-104. 1975.
Ill
-------
APPENDIX A
PERFORMANCE ANALYSIS OF HIGH-PRESSURE
FLUIDIZED BED BOILER SYSTEMS
-------
APPENDIX A
PERFORMANCE ANALYSIS OF HIGH-PRESSURE
FLUIDIZED BED BOILER SYSTEMS
MODIFICATION IN PERFORMANCE CALCULATION PROCEDURES
The procedure used for calculating the performance of high-
pressure fluidized bed boiler systems as reported in References 1 and
2 has been revised. The heat rates given in these references were
calculated using the conventional boiler efficiency expression. For
the parametric studies reported in Reference 1, an assumed boiler
efficiency value of 89.6 percent was used.
The conventional loss factors for Ohio Pittsburgh No. 8
seam coal with 3 percent moisture, a stack-gas temperature of 135°C
(275°F) and 17.5 percent excess air, and regenerative desulfurization
were calculated to be as follows:
• Sensible heat loss of dry stack gas 3.88%
• Sensible and latent heat losses of moisture
in coal and water from hydrogen in fuel 4.14
• Sensible heat loss of moisture in air 0.08
• Radiation loss 0.15
• Incomplete combustion 1.50
• Sensible heat loss of solids 0.11
• Unaccounted-for losses and manufacturer's
margin 1.50
TOTAL 11.36%
This gives a boiler efficiency of 88.6 percent,which was used to
calculate the heat rates used for energy cost estimates in Reference 1
and all of the heat rates in Reference 2.
It has been determined that the conventional definition of
boiler efficiency is not applicable to the high-pressure fluidized bed
113
-------
boiler system, and a corrected boiler efficiency has been developed
for use instead. The correct procedure for calculating the plant heat
rate for a high-pressure fluidized bed boiler system with in-bed desul-
furization is as follows:
e Ideal fuel rate:
(Wf) ideal = (heat to steam + heat to gas turbine)/LHV
• Plant heat rate (HHV):
Plant heat rate = ((Wf) ideal x HHV)/[(net plant power)(combustion efficiency)]
(heat to steam + heat to gas turbine] ( (HHV/LHV) ]
net plant power j [combustion efficiency I
The loss factors for Ohio Pittsburgh seam No. 8 coal with
3 percent moisture and once-through desulfurization are as follows:
e Incomplete combustion 1.50%
o Radiation losses 0.04
• Sensible heat in solid wastes 1.50
• Heat absorbed by desulfurization reactions (-0.80)
• Unaccounted-for losses and manufacturer's
margin 1.50
TOTAL 4'10*
If the difference between the higher and lower heating values
of this coal is included as a boiler loss factor (as it is in the conven-
tional boiler efficiency expression), there is an additional loss factor
of 3.9 percent and the total losses are 8.0 percent. The combustion
efficiency for the high-pressure fluidized bed boiler is, therefore, 92.0
percent rather than the calculated conventional boiler efficiency value
of 88.6 percent.
Those heat rates calculated using a boiler efficiency of 88.6
percent are 3.8 percent higher than they should have been; and the heat
rates for the parametric studies in Reference 1, which were based
In this case the ratio HHV/LHV becomes part of the boiler efficiency.
114
-------
on an assumed boiler efficiency of 89.6 percent, were 2.7 percent higher
than they should have been. The corrected heat rate for the base case
wherein the condenser pressure is 1-1/2 in Hg.the boiler efficiency is
92.0 percent, and the gas-turbine pressure loss is 7.5 percent is,
therefore, 9040 (8568) rather than 9381 kJ/kWh (8892 Btu/kHh),
EFFECT OF GAS-TURBINE DISCHARGE TEMPERATURE LIMIT ON OPERATION
The Westinghouse Gas Turbine Engine Division limits the dis-
charge temperature of the gas turbines which it manufacturers to 538°C
(1000°F). This limit imposes constraints on the operation of the high-
pressure fluidized bed boiler systems which use these gas turbines.
These constraints can be defined in terms of minimum allowable pressure
ratios for given values of turbine inlet temperature and air equivalence
ratios. Minimum allowable cycle pressure ratios for turbine inlet
temperatures of 871, 927, and 982°C (1600, 1700, and 1800°F) and for an
air equivalence ratio of 1.1 are shown on Figure A-l to be 5.7, 7.2, and
8.9 respectively.
Figure A-2 shows that the minimum allowable cycle pressure
ratio imposed by the 538°C (1000°F) limit on turbine discharge tempera-
ture is a weak function of the air equivalence ratio.
The cycle pressure ratio at which the plant heat rates of high-
pressure fluidized bed boilers are minimum is about 8:1 (see Figure A-3).
For the range of turbine inlet temperatures which are of interest in
fluidized bed combustion, the minimum allowable cycle pressure ratio im-
posed by the 538°C (1000°F) limit on gas-turbine discharge temperature
is, therefore, generally lower than the value for minimum plant heat
rate. It can be concluded, therefore, that a 538°C (1000°F) limit on the
discharge temperature of gas turbines used in high-pressure fluidized bed
boilers does not impose a significant constraint on the operation of the
boiler system unless cue turbine inlet temperature exceeds 982°C (1800°F).
For a gas-turbine cycle with an adiabatic combustor, the optimum
pressure ratio is greater than for that of the pressurized boiler cycle.
The gas-turbine discharge temperature limit, therefore, imposes even less
of a constraint on its operation.
* •
Ratio of actual to stoichiometric air/fuel ratios.
115
-------
075153-»
Curve 679660-A
649 593
(1200)
593
(1100)
538
(1000)
u.
O
1 371
5 (700)
o
316
1600)
260
(5001
204
i i i i i i i i i
Pressure Ratio = 5 1
Maximum Allowable Value
n =8.1
Gl InlelTemp = 871"C I1600°FI
-
i i i i i i i i i
(4001 LO 11 12 13 14 15 16 17 18 19 i
Air Equivalence Ratio
,400) i 3 5 7 9 u 13 15 17 19 21 figure A-c-rertormance oi pressurized iiuiaizeaoeo
power plant
Cyde Pressure Ratio
Figure A-1-Performance of pressurized fluidlzed bed
power plant
C .r,. c S80J-
9706
(9200)
9495
19000)
£ 9284
^ 18800)
"5
CO
5 9073
' 186001
Ol
O=
| " 8862
18400)
8651
18200)
8440
Air Equivalence Ratio
o L8-1
o 1 !• 1
G-t Temn
N. f 871°C
^X. ^^^^(1600"?)
^^5^*° o — "^^ 927°C
O *^^^^_ ,^^^^^ *^F^ ^11 7f¥l®P 1
\<5\ X-0 3-- 0*"*'
V ^*vi. ^^^cr-'^982cC
\ •»>— ^ ^xllBOO'F)
\ jy^
V ^
>3 '— — '
Steam Plant- 24. 133 kPa/538" C/538" Cl 11 82 kPa
(3500psi/l0000F/10DO<>/3.5inHgl
- G-T Row = 647 kg/sec
(1427 Ib/ sect
! 1 1 1
18000) i 5 a 11 17 71
Gas-Turbine Cycle Pressure Ratio
Figure A-3-Perlormance of pressurized lluidized bed
power plant vs cycle pressure ratio
116
-------
EVALUATION OF LOW-TEMPERATURE CLEANUP OF COMBUSTION PRODUCTS
One of the design premises for the high-pressure fluidized bed
boiler design concept has been particulate removal at boiler outlet tem-
perature. Since the cost of particulate removal equipment is a rather
strong function of operating temperature level, there is a possibility
that the minimum cost for energy produced by the high-pressure fluidized
bed boiler would occur when particulate removal is carried out at a reduced
temperature level. There is also a possibility that the only practical
technique for particulate removal which would give the required degree of
removal for reasonable turbine life is scrubbing. A study has been
carried out, therefore, to determine the performance penalty which accom-
panies reduced temperature techniques for removing particulates from the
combustion products of a high-pressure fluidized bed boiler.
Two alternatives have been investigated:
• Cooling the combustion products by the use of a
convection-type boiler and removing particulates by
cyclone separators, tornado separators, electro-
static precipitators, or combinations thereof at
intermediate temperatures
• Cooling the combustion products with a recuperator
followed by a scrubber-cooler.
The arrangement for the high-pressure fluidized bed boiler
system with intermediate temperature particulate removal is shown in
Figure A-4. The convection-type boiler for cooling the products of com-
bustion to a temperature well below the bed operating temperature is in
series/parallel with the fluidized bed boiler. This permits the tempera-
ture of the combustion products to be reduced from 871°C (16008F) to the
gas-turbine idle temperature which is in the range of 482 to 538°C (900 to
1000°F) and independent of air equivalence ratio.
If the plant were operated with the gas-turbine inlet temperature
at or near the idle valve, plant turndown (by reducing bed temperature)
Series for combustion products-parallel for working fluid.
117
-------
00
Intermediate Temperature
Particulate Removal
Convection-
Type
Boiler
Fluidized
Bed
Boiler
Sorbent —
Coal —
Spent
Sorbent
Solids
Auxiliary
•^Combustor
(Start-up)
*-*
ST
o
Condenser
-xAAA-
•WV*-
Stack-Gas
Cooler
Figure A-4 -High-pressure fluidized bed boiler power system with intermediate temperature
particulate removal
-------
would require by-passing of a fraction of the products of combustion around
the convection boiler in order to maintain the turbine inlet temperature
at or above the minimum value. This minimum value will increase as bed
temperature decreases, because of the reduction in the quantity of combustion
products to the turbine.
The effect of gas-cleaning temperature on plant capacity is
shown in Figure A-5. This shows that plant capacity decreases about 8 MW
per 56°C (100°F) drop in temperature.
The effect of the gas-cleaning temperature on the plant heat
rate is shown in Figure A-6. This indicates that the plant heat rate in-
creases about 158 kJ (150 Btu)/kWh for each 56°c(100°F) drop in temperature.
The arrangement of the high-pressure fluidized bed boiler with
low-temperature particulate removal is shown in Figure A-7. The hot com-
bustion products are*passed through one stage of cyclone separation to
recover the larger fraction of the char particles which are elutriated
from the bed so that the carbon losses will be reduced to a minimum value.
The effluent from the char separator will retain the finer ash particles
with relatively low carbon content. This stream will be cooled to a tem-
perature in the range of from 93 to 204°C (200 to 400°F), depending on the
recuperator effectiveness and the temperature of the cold products of com-
bustion out of the scrubber-cooler.
In the scrubber-cooler the products are evaporatively cooled to
the saturation line and then further cooled along the saturation line
until the water vapor content of the gas mixture is equal to the initial
value out of the fluidized bed boiler. The initial water vapor content
of the combustion products is a function of the coal.composition and the
percent excess air in the fluidized bed combustor, and the final tempera-
ture of the products out of the scrubber-cooler is a function of the
water vapor content and the system pressure ratio. The temperature of
the clean fuel gas leaving the scrubber-cooler and entering the recupera-
tor is, therefore, a function of the excess air in the combustion process
and the system pressure ratio.
Westinghouse made performance calculations for the parametric
combinations shown in Table A-l to determine the effect of air equiva-
lence ratio, boiler outlet temperature, system pressure ratio, and recup-
erator effectiveness.
119
-------
482
700
650
'£600
8-
N>
O
to
Q-
550
500
538
Gas-Turbine Inlet Temperature. °C
593 649 704 760
Air Equivalence Ratio = 1.175
Bed Temperature - 954°C I1750°FI
Convection-Type Boiler in Series/
Parallel with F B B
900 1000 1100 1200 1300 1400
Gas-Turbine Inlet Temperature, .°F
Curve 679729-B
816
871
1500
1600
Figure A-5-Plant capacity vs gas-turbine inlet temperature for high-pressure
fluidized bed boiler with intermediate temperature participate removal
-------
Curve 679730-B
ro
10200
(10761)
- 10000
i (10550)
9800
[10339)
01
2 9600
~ (10128)
£ 9400
(9917)
9200
(9706)
9000
427
Gas-Turbine Inlet Temperature, °C
482 538 593 649 704 760 816 871
.Minimum
Air Equivalence Ratio = 1.175
Bed Temperature -954°C (1750°F)
Convection-Type Boiler in Series/
Parallel with F B B
(9495) 800 900 1000 1100 1200 . 1300 1400
Gas-Turbine Inlet Temperature, °F
1500
10761
10550
10339
10128
9917 2.
Q_
9706
9495
1600
Figure A-6-Plant heat rate vs gas-turbine inlet temperature for high-pressure fluidized
bed boiler with intermediate temperature particulate removal
-------
Scrubber-Cooler ,-Water
—
Dug 6236M'*
Sorbent out
•^ —
F R Rnilpr
1 . U. DUIICI
Sorbent in
Cn-t\ »
l/Udl *
,--:> « -
v J;fi
~l
Boiler
i
i
1
i
i
t
m
*
*
\
\
i
i
i
I
I
1 — ,*-- -
/ir'TN
]®
L L
r r*^
;
^. •)
s. k.t4UU/ 1UUU/
^" 7~
1
i
\ %
f J 4
1 |
i i
i i
| Air
"^ , _ pomh PrnH
| 1 — r.tr.im and Wafpr
1 ACL] ' JICCIMI CIIIU VVCJIOI
1 AjM I
i i ... -Solids
*1
1
1
i nnn A
1000 * I
^ II
1 I 1
1 ' 1
1 I 1""~ ~^T
1 I 11
1 1 • 1
1 II
1 1 1 1
4^"*TlDTl. -*^***TPT fc — *^7pT r>
I
Air
Auxiliary
Combustor
(Start-up)
®
qOut
- -* To Stack
(?)
Figure A-7-Pressurized fluidized bed combustion power plant with cold cleanup of combustion products
-------
Table A-l
PARAMETRIC COMBINATIONS USED IN PERFORMANCE CALCULATIONS
a
Case
1
2
3
4
5
6
7
8
9
*air
1.175
1.175
1.175
1.50
2.00
1.175
1.175
1.175
1.175
Boiler outlet
temp., °C (°F)
871 (1600)
871 (1600)
871 (1600)
871 (1600)
871 (1600)
927 (1700)
816 (1500)
871 (1600)
871 (1600)
System
press, ratio
10
10
10
10
10
10
10
15
5
Recuperator
effectiveness c
0.860
0.930
0.789
0.860
0.860
0.860
0.860
0.860
0.860
A gas-turbine compressor air flow of 635 kg/s (1400 Ib/sec)
was used for all cases.
b
Air equivalence ratio.
C Recuperative effectiveness = (t3 - t^)/(t3/t5). See Figure A-7.
The temperatures at each of the ten stations shown in Figure A-7
for the nine cases are shown in Table A-2:
123
-------
NJ
Table A-2
TEMPERATURES FOR TEN STATIONS
Station
number
1
2
3
4
5
6
7
8
9
10
Temperature °C (°F)
Case 1
15 (59)
313 (596)
871 (1600)
191 (376)
90 (176)
760 (1400)
376 (709)
274 (525)
135 (275)
286 (547)
1 Case 2
15
313
871
136
80
816
412
274
135
298
(59)
(596)
(1600)
(276)
(176)
(1500)
(773)
(525)
(275)
(568)
Case 3
15 (59)
313 (596)
871 (1600)
247 (476)
80 (176)
704 (1300)
340 (644)
274 (525)
135 (275)
274 (525)
1 Case 4 |
15
313
871
189
77
759.
371
274
135
296
(59)
(596)
(1600)
(372)
(171)
5 (1399)
(700)
(525)
(275)
(565)
Case 5
15 (59)
313 (596)
871 (1600)
183 (362)
71 (160)
759 (1398)
365 (689)
274 (525)
135 (275)
314 (598)
Case 6
15 (59)
313 (596)
927 (1700)
198 (389)
80 (176)
808 (1487)
407 (764)
274 (525)
135 (275)
298 (569)
1 Case 7
15 (59)
313 (596)
816 (1500)
182.5 (361;
80 (176)
713 (1315)
345 (653)
274 (525)
135 (275)
274.5 (526)
Case 8
15 (59)
387 (728)
871 (1600)
200 (392)
90 (195)
762 (1403)
326 (618)
274 (525)
135 (275)
268 (515)
1 Case 9
15 (59)
204 (400)
871 (1600)
177 (350)
64 (147)
758 (1397)
474 (885)
274 (525)
135 (275)
318 (604)
-------
The calculated plant power outputs and plant heat ratio are
shown in Table A-3:
Table A-3
PLANT POWER OUTPUTS AND HEAT RATES
Case
1
2
3
4
5
6
7
8
9
Plant power - kw
612410
630150
596594
469945
343700
606535
626762
608780
610404
Plant heat rate
kJ/kWh (Btu/kWh)
10058 (9534)
9650 (9147)
10485 (9938)
10254 (9719)
10553 (10003)
10007 (9485)
10078 (9553)
10139 (9610)
10113 (9586)
Plant heat rate as a function of recuperator effectiveness is
shown plotted on Figure A-8. The heat rate is shown to increase about
six percentage points for each ten percentage points decrease in recupera-
i
tor effectiveness.
Plant heat rate is plotted as a function of gas-turbine pressure
ratio on Figure A-9. This shows that plant heat rate is a rather weak
function of pressure ratio and that the optimum pressure ratio is about
8.5:1.
Plant heat rate is plotted as a function of boiler outlet tem-
perature on Figure A-10. This indicates that over the range of boiler
outlet temperatures which correspond to upper and lower limits on flui-
dized bed temperatures with in-bed desulfurization, in other words, 926
125
-------
1106001
11183
110400)
10972
1102001
10761
IIOQOOI
10550
3 08001
«f 10339
(9(00)
10128
(94001
9917
192001
9906
Pressure Ratio - ID
Boiler Oullet Temp • 871°C (16000FI
Air Equivalence Ratio • 1 175
(9000) ' i
9495 75
i
>Q
oe
i
1106001
11183
1104001
10972
1102001
10761
• 100001
10550
198001
103)9
19600)
10128
194001
9917
192001
9706
5
Boiler Outlet Temp • 871°C I1600°FI
Air Equivalence Ratio - 1 175
Recuperator Eflcctiveness - 86%
10
Gas-Turbine Pressure Ratio
Fiqure A-9-Plant heal rate vs gas-turbine pressure ratio
80 90 95
Recuperator Effectiveness -»
100
Figure A-8-Planl heal rate vs recuperator effectiveness
1106001
11183
110400)
10972
H0200I
10761
£ (100001
10550
S 198001
3 10339
19600)
10128
19400)
9917
(92001 ,_
Pressure Ratio - 10
Air Equivalence Rdlio - I 175
Recuperator Effectiveness • 86*
816
' 15001
871
116001
927
117001
982
118001
9706 704 760
(1300) i14001
bc.ilcr OI.HPI Idiinerdlu'p • "C i°l I
I iijure A-10 -Plant liedl rdlp vs boiler outlet temperature
1038
11900)
126
-------
and 704°C (1700 and 1300°F) respectively, the heat rate varies only
about one percent.
Plant heat rate is plotted as a function of air equivalence
ratio on Figure A-ll This shows that the heat rate is a rather strong
function of air equivalence ratio, with the heat rate increasing with an
increasing air equivalence ratio. This results from the increase in the
fraction of gas-turbine power as the air equivalence ratio increases and
indicates that the optimum air equivalence ratio is the minimum level
which will give good carbon burn-up—in other words, 1.15 to 1.20.
COST ESTIMATE
Order of magnitude cost estimates have been made to evaluate
the intermediate temperature and water scrubbing options for gas
3
cleaning. First, the PFBB investment cost of $265/kw for a 635 MW
installation was escalated to late 1974, using the Chemical Engineering
construction cost index for process facilities, to give a November 1974
cost of $355/kw. Then, the cost for a third stage of granular bed
filter particulate removal was added to the PFBB investment. The
installed subcontract cost for the eight pairs of granular bed filters
required to clean up the effluent from the eight tornado cyclones
constituting the second particulate removal stage in a 600 MW unit in
late-1974 is $16,128,000. When interest during construction, escalation
during construction, and so on are added to this, the total Installed
cost of the granular bed filtration third-stage particulate removal is
$23,452,000, or $39/kw. Thus $355 plus $39,or $394/kw,represents the
late 1974 cost of a PFBB installation based on 17.5 percent excess air.
If 100 percent excess air is used in the PFBB, the boiler effluent gas
quantity increases and so does the cost of removing particles from it.
The $39/kw derived above, plus $25/kw for the first two stages of
particulate removal, total $64/kw. This, when increased to accommodate
the larger gas volume at 100 percent excess air design, becomes $96/kw.
The total PFBB cost, therefore, becomes $355-25+96 = $426/kw. This
amount does not account for the incremental change in the boiler and
power generation equipment cost which will result from higher excess
127
-------
00
i
o>
"ro
O>
c
re
(10600)
11183
(10400)
10972
(10200)
10761
(10000)
10550
(9800)
10339
(9600)
10128
(9400)
9917
Curve 679657-A
(9200)
9706
1.0
1.2
Pressure Ratio -10
Boiler Outlet Temp. - 871°C (1600°F)
Recuperator Effectiveness - 86%
1.4 1.6 1.8
Air Equivalence Ratio
2.0
2.2
Figure A-11-Plant heat rate vs. air equivalence ratio
128
-------
air. The boiler cost is reduced as the result of changes in the shell
and heat transfer surface costs and elimination of a carbon burn-up
cell. The power generation equipment cost increases slightly due to
the shift in the percentage of gas turbine power generated. Finally,
the $340/kw cost for a conventional stream power plant with a stack-
gas scrubber was updated. Removing $30/kw for the scrubbing unit,
and escalating by the above referenced cost index to late 1974, gives
4
$415/kw. If the recent TVA-evaluated • limestone scrubbing system
minimum costs of $50/kw are added, the minimum cost for a steam power
plant with limestone scrubbing is $465/kw, and that cost, according
to TVA, could range as high as $530/kw.
With cooling of the 871°C (1600°F) PFBB products of combustion
in a waste heat boiler to 482°C (900°F) so as to permit metal such as
low chrome-molybdenum steel alloy to be substituted for refractory
lining, it can be seen from Figure A-12 that total costs rise (for a
PFBB with conventional cyclones in series with tornado cyclones and
granular bed filters) so that the plant cost at 427°C (800°F) gas
cleanup temperature is $458/kw. Most of the rise in cost is due to
decreased plant output. Investment costs for refractory-lined and low-
alloy steel particulate removal equipment goes down as temperature (and
gas volume) go down, while waste heat boiler cost goes up as cleaning
temperature goes down. These two cost trends tend to cancel each
other. The net $/kw increase, then, is due entirely to the loss in
plant output shown earlier in Figure A-5.
If an electrostatic precipitator could replace the tornado
cyclones and granular bed filters at 538°C (10008F), approximately
$65/kw or $39 million would be saved in particulate removal installed-
system costs. The cost of a pressurized, 538°C (1000°F), low-sulfur-
content gas precipitator has been investigated to determine if cost
savings would be possible from its use. A precipitator design based
on 1.5 m (3.5 ft)/sec velocity adds roughly $25/kw for installed-system
costs. Thus, it is apparent that precipitators may be substantially
129
-------
Curve 681705-A
«/>
'
C.
-1
E
+10
0
-10
^S -20
c
O
-30
470
460
.^
^ 450
5 440
"K
to
430
420
410
W. H. Boiler Cost
Particulate Removal System Cost
560 MW
I
595 MW
I
635 MW
- Low -Alloy Particulate Removal System
_ Refractory
Lined
- Particulate
Removal
- System
427 538
649
761
871
(800) (1000) (1200) (1400) (1600)
Intermediate Temperature, °C(°F)
Figure A-12-Intermediate temperature gas
cleaning for PFBB (Nominal
635 MW size)
130
-------
less expensive than tornado cyclone and granular bed filter combinations
at 538°C (1000°F).
An inquiry to the EPA staff at Research Triangle Park
revealed that Mr. J. Abbott at that location is instigating work on hot
pressurized electrostatic precipitators. Mr. Abbott explained during
a subsequent phone conservation that Research Cottrell will soon receive
an EPA contract to study pressurized precipitators. Mr. Abbott pointed
out, though, that present indications are not very optimistic regarding
the performance of the pressurized units. Collection efficiency, gas
density, charge potential, and other factors indicate that pressurized
precipitation will be difficult. Also, 538°C (1000°F) is possibly
too high a temperature for effective precipitator operation. A 316°C
(600°F) temperature level may be much better for electrostatic preci-
pitators at 1013 kPa (10 atm). In any case, Research Cottrell will
quantify the problems involved in hot pressurized electrostatic preci-
pitation in their forthcoming contract.
Costs were obtained for the cases in Table A-2 dealing with
wet-scrubbing alternative systems. It immediately became apparent that
for Case 4 (50 percent excess air) and Case 5 (100 percent excess ari)
wet scrubbing was uneconomical. The losses in cycle efficiency, plus
the larger size wet scrubbers and recuperators required, added more
than $150/kw for Case 4 and $350/kw for Case 5, making total plant
costs $505/kw and $705/kw respectively. They are both costs that exceed
the conservatively estimated dry particulate removal cost of $426/kw
for a 100 percent excess air PFBB installation.
It is apparent from the cost increases incurred by Cases 4
and 5 that a wet-scrubbing gas cleanup system would also be much too
costly for use vith adiabatic combustion to be competitive with conven-
tional boiler/scrubber power systems.
Order-of-magnitude costs for the 17.5 percent excess air cases
showed Case 7 to have the least added cost for wet scrubbing, and Case 2
to have the highest added cost. All of these eight cases showed between
131
-------
§40 and $75/kw added costs for wet scrubbing, with about $55/kw a
representative average added cost. Considering the very approximate type
of cost estimating used, and until more becomes known about actual costs
for the Incoloy 617 heat exchange material assumed for the very large
(46,450 to 139,350 m2/500,000 to 1,500,000 ft2 of surface) recuperative
heat exchangers used in the cost estimates, no firm conclusions can
be drawn regarding the relative costs of the seven cases studied with
17.5 percent excess air.
It may be of future interest to compare cold sulfur dioxide
removal, once the gas was cooled for particulate removal in a 17.5%
excess air design, with the cost of hot limestone pickup of sulfur
dioxide. The pressurized low-temperature sulfur dioxide scrubber may
possibly be less costly than once-through limestone for sulfur dioxide
removal.
PERFORMANCE OF THE HIGH-PRESSURE FLUIDIZED BED BOILER WITH ADVANCED
STEAM CONDITIONS
Reference 5 contains performance data on high-pressure f lui-
dized bed boiler systems with steam conditions of 22,754 kPa/5930C/593°C
(3300 psi/11000F/1100°F) and 31,028 kPa/649°C/593°C (4500 psi/1200°F/
1100°F). Performance calculations have now been extended to even higher
steam conditions of 31,028 kPa/649°C/649°C (4500 psi/1200°F/1200°F) and
34,475 kPa/760°C/760°C (5000 psi/1400°F/1400°F). Figure A-13 is a plot of
the plant heat rate as a function of the air equivalence ratio with steam
condition parameters for a gas-turbine inlet temperature of 871°C (1600°F),
10:1 pressure ratio, and a steam condenser pressure of 11.82 kPa (3-1/2 in.
Hg). This shows that the plant heat rate is a weak function of the air
equivalence ratio for all supercritical steam conditions.
Steam conditions of 31,028 kPa/649°C/649°C (4500 psi/1200°F/
1200°F) have a potential for plant heat rates which are 8 to 10 percent
better than for those steam conditions which currently predominate in con-
ventional steam plants—in other words, 16,548 kPa/538°C/538°C (2400 psi/
1000°F/1000°F) and 24,133 kPa/538°C/538°C (3500 psi/1000°F/10008F). These
steam conditions are similar to the design conditions for Eddystone No. 1,
132
-------
Curve 679801-A
ra
o»
9495
(9000)
9284
(8800)
9073
(8600)
8862
(8400)
8651
(8200)
8440
(8000)
8229
(7800)
8018
(7600)
16,548 k Pa/5380C/538°C/11.82 k Pa
0 (2400psi/1000°F/1000°F/3.5 in Hg)
24,133 k Pa/538°C/538°C/ll. 82 k Pa
1.0
(3500psi/1000°F/lOOO°F/3.5 in Hg)
G-T Inlet Temp. = 871 ° C (1600° F)
G-T Pressure Ratio = 10:1
31,028 k Pa/649°C/649°C/ll. 82 k Pa
(4500 psi/1200°F/1200°F/3.5 in Hg)
34,475 k Pa/7600C/760°C/ll. 82 k Pa
(5000psi/1400°F/1400°F/3.5 in Hg)
1.2
1.4
1.6
1.8
Air Equivalence Ratio -0..
2.0
Figure A-13-Performance of pressurized fluidized bed power
plant with advanced steam conditions
133
-------
34,475 kPa/649°C/566°C/566°C (5000 psi/1200eF/10500F/1050°F). The Eddy-
stone plant has had only limited operation at design conditions but ex-
tensive operation at a superheat temperature of 613°C (1135°F). Operation
is limited to superheat temperatures less than 649°C (1200°F) by low-cycle
fatigue problems on turbine control and stop valves. The operation of
the boiler and steam turbine at 649°C (1200°F) superheat temperature is
reported to be technically and economically feasible. The equivalent
plant efficiency level for these steam conditions is about 44 percent.
Steam conditions of 34,475 kPa/760°C/760°C (5000 Psi/1400°F/
1400°F) can potentially further improve plant heat rate by 4 to 5 percent.
These steam conditions, however, are well beyond the state-of-the-art
technology for steam superheaters and steam turbines. (See discussion
of advanced steam cycles technology in Appendix L.)
134
-------
REFERENCES
1. Archer, D. H. et al. Evaluation of the Fluidized Bed Combustion
Process. Office of Air Programs. Environmental Protection Agency.
Westinghouse Research Laboratories. Pittsburgh, Pennsylvania.
Contract 70-9. NTIS PB 211494 and PB 212916. November 1971. Vols.
I and II.
2. Keairns, D. L., et al. Evaluation of the Fluidized Bed Combustion
Process. Office of Research and Development. Environmental Protec-
tion Agency. Westinghouse Research Laboratories. Pittsburgh,
Pennsylvania. EPA-650/2-3-73-048 a, b, c, d. Contract 68-02-0217.
December 1973. Volumes I to IV. NTIS PB 231 162/9, 231 163/7,
232 439/3, and 233 1U1/5.
3. Ibid. Volume I. p. 25.
4. McGlamery, G. G., et al. Detailed Cost Estimates for Advanced
Effluent Desulfurization Processes. Tennessee Valley Authority.
Muscle Shoals, Alabama. January 1975. TVA-Bull-Y-90, EPA/600/2-75-006,
PB-242 541/1WP 429p.
5. Reference 2. Volume I. p. 97.
135
-------
B
APPENDIX B
BOILER DESIGN EVALUATION
-------
APPENDIX B
BOILER DESIGN EVALUATION
Problems are invariably associated with a new development such
as fluidized bed combustion boilers. The potential technical uncertain-
ties have been discussed previously. The evaluation here will be confined
to those problems pertaining to the fluidized bed boiler design alone.
In order to optimize boiler design, deep fluidized beds on the
order of 3.35 to 4.27 m (11 to 14 ft) deep were selected. Fluidizing
velocities of 1.83 to 2.74 m/sec (6 to 9 ft/sec) were proposed in the
basic design. Both the bed depth and the operating velocity are outside
the range of experimental conditions employed in the pilot-scale units
of other contractors. The recent completion of a pressurized fluidized
bed combustion experimental series in the British Coal Utilization
Research Association (BCURA) pilot unit and the start-up of the Exxon
Research and Engineering Company, Linden, New Jersey (Exxon) miniplant have
contributed significantly to the understanding of coal combustion in a
pressurized fluidized bed combustor. We are now in a better position to
assess the operating conditions and design parameters set for the basic
boiler design.
REVIEW OF EXPERIMENTAL DATA
BCURA Pilot-Scale Unit
2
Combustor Design
The pressurized combustor was a 0.91 m by 0.61 m (3 ft by 2 ft)
fluidized bed housed in a refractory-lined pressure shell of 1.83 m (6 ft)
diameter. Heat transfer surfaces in the form of tube bundles were located
at different levels. The lowest tube bundle, which was always immersed
137
-------
in the fluidized bed during normal operation, was referred to as the
"bed cooling coils." In the freeboard there were two sets of uncooled
baffle tubes which helped to reduce solids carry-over. The top tube
bundle, situated above the baffle tubes, was referred to as the "freeboard
cooling coils." The bed cooling coils were of 25.4 mm (1 in) outside diameter,
stainless steel tubing. There were 25 rows of tubes linked into 10 separate
cooling circuits located between 394 mm (15.5 In) and 1410 mm (55.5 in)
from the distributor plate. The 25 rows were arranged on a staggered
configuration with 152 mm (6 in) horizontal pitch and 38 mm (1.5 in)
vertical pitch.
Air to fluidize the bed was supplied through a distributor elate
with 185 stainless steel tubes on a 52.4mm by 95.3mm (2-1/16 by
3-3/4 in) square pitch. The tubes were closed off at the upper ends.
Four holes were drilled through the tube walls 25.4mm (1 in) above the
upper surface of the supporting mild-steel plate. The unfluidized bed
below the holes acted as an Insulating layer between the hot fluidized
bed and the supporting mild-steel plate.
Four coal-feed nozzles, two on each side of the casing along
the 0.91 m (3 ft) dimension, were situated 88.9 mm (3.5 in) above the
distributor plate. The water-cooled nozzles were 19.1mm (3/4 in) inside
diameter and extended 10.2 mm (4 in) horizontally into the bed.
Two natural-gas-fired burners with a nominal maximum heat output
of 422 million J/hr (0.4 million Btu/hr) each were provided for initial
bed heating and for Igniting propane vapor supplied via the distributor
plate.
2
Experimental Conditions
The operating conditions are compared with those of the basic
design in Table B-l. The main changes in operating variables during the
experiments were:
• Bed temperature - 899°C and 954°C (1650°F and 1750°F)
• Coal source - Pittsburgh and Illinois coals
• Type of additive - Dolomite and limestone
• Ratio of calcium/sulfur incut - 1.4 to 2.2 mole/mole.
138
-------
OJ
VO
Table B-l
COMPARISON OF OPERATING CONDITIONS
Basic design BCUEU
Exxon miniplant
Test conditions
I unit to date Design conditions
Fluidizing velocity, m/sec (ft/sec) 1.83-2.74 0.67-0.76 1.83-3.05 3.05
(6-9) (2.2-2.5) (6-10) (10)
Bed depth (expanded) m (ft) 3.35-4.27 1.22-1.43 0.91-4.57 up to 4.57
(11-14) (4.0-4.7) (3-15) (up to 15)
Combustor pressure, kPa (atm) 1013 (10) 456 405-1013 1013 (10)
(4.5-5.0) (4-10)
•
Excess air, % 17.5 14-20 25-130
Bed temperature, °C (°F) 954 884-952 815-982 982
(1750) (1630-1745) (1500-1800) (1800)
-------
2
Experimental Results
The exnerimental results are discussed >elow. Onlv results
relevant to the boiler design are included here.
Combustion Efficiency. The combustion efficiency increased
significantly with an increase in bed temperature. The results from
these experiments at bed temperatures of 899°C (1650°F) and 954°C (1750°F)
are compared with earlier data obtained at 799°C (1470°F) and 17 percent
excess air in Figure B-l. There was no detectable combustion in the free-
board at either 899°C (1650°F) or 954°C(1750°F) bed temperatures, while
2 to 3 percent heat input was released in the freeboard during combustion
at 799°C (1470°F). At a 799°C (1470°F) bed temperature, the combustion
efficiency was also sensitive to the amount of excess air, increasing
from about 96 percent at 10 percent excess air to 97.5 percent at 17 percent
and 99 percent at 25 percent.
Heat Transfer Coefficient. The average bed-tube heat transfer
coefficients were 380 W/m2-°K (67 Btu/ft2-hr-°F) at 899°C
(1650°F) bed temperature and 403 W/m2-°K (71 Btu/ft2-hr-°F) at
954°C (1750°F). The presence of the tube bank did not affect the heat
transfer coefficient significantly. The heat transfer coefficients in the
lower 3048 mm (12 in) of the bed where there was no tube bank present were
on]y about five percent higher.
Corrosion of Metal Specimens in the Bed. Metal specimens
were examined metallographically after 100 hours' exposure. The duration
of exposure was too short to make any generalized conclusion regarding the
corrosion. For 1 percent, 2-1/4 percent, and 12 percent chrome steel
specimens, the attack was slight, with no evidence of intergranular
penetration at temperatures up to their respective recommended maximum
operating limits for use in steam boilers. The weight losses are summarized
in Table B-2. Comparison with the previously lonper duration runs is shown
graphically in Figure B-2. Apparently, weight loss reaches a reliable, steady
140
-------
100
800
.22
"o
o
.o
o
o
99
98
97
Bed Temperature: °C
850 900
950
o Average Result from Previous Work
• Results from Recent Tests
i
I
1500
1600 1700
Bed Temperature: °F
1800
Figure B-1-Effect of bed temperature on combustion
efficiency with 17% excess air
141
-------
to
Table B-2
SUMMARY OF WEIGHT LOSS MEASUREMENTS
(100 hours' duration)
Alloy
1% Chrome
2-1/4% Chrome
12% Chrome
18% Chrome
(AISI 321)
21% Chrome
(Incoloy 800)
Tube
temperature,
°C (°F)
199-371
(390-700)
(a) 371-538
(700-1000)
(b) 260-677
(500-1250)
(a) 521-660
(970-1220)
(b) 671-777
(1240-1430)
(a) 621-749
(1150-1380)
(b) 777-832
(1430-1530)
749-827
(1380-1520)
Weight
loss
ug/cm^h
100
260
400
160
170
70
40
45
Previous measurements
in fluidized beds3
(500 hours & 1000 hours' duration)
C
\
c
I
No data
j*
180 to 230 jig/on h @ 499°C [930°F]
570 to 970 ug/cm2h @ 599°C riHO°F]
7 to 14 yg/cm2h @ *99°C [1290°F]
r\
790 to 10CO ug/an h @ «40°C [1560°F]
10 - 20 yg/cm2h @ 699°C [1290°F] For
, AISI
10 - 70 yg/cm h @ 840°C [1560°F] 316
No data
National Coal
Environmental
Board. Reduction of Atmospheric Pollution. Final Report.
Protection Agency. September 1971.
-------
100
(urve (.73HO-A
10
• l^
o
1 1 1 1
Present Results with 321 Steel
Type 316 Steel
~— T
Q
^* ^^* ^^» ^^ O
o
i I
1 |
1
Tube ~
• Temp.
(1560°F)
o ~
°(1290°F)~
I
CM
4
1000
01
CO
100
10
12% Chrome Steel
g*=- Present
Results
Tube
Temp.
•-849°C
(1560°F)
JL
200
400 600
Time: hours
—§-699°C
(1290°F)
_J
800
1000
Figure B-2-Effect of time on corrosion (previous results in
fluidizedbeds)2
143
-------
state onlv after long duration — a few hundred hours for high-alloy
materials and at least 1000 hours for low-alloy materials. On the basis of
exoerimental results, the wastage was not excessive for alloys operating
up to their recommended maximum design temperatures for use in steam
boilers. With Incoloy 800, which is a possible material for air heater
applications, attacks at 816°C (1500°F) appeared to be minimal.
Temperature Distribution in the Bed. Temperature distribution
throughout the bed was reasonably uniform. In tests at 899°C (1650°F) the
bed temperature measured by 14 thermocouples from a point 76 mm (3 in)
above the distributor air nozzles to the top of the bed varied only by
about 5.5°C (10°F). The corresponding variation at 954°C (1750°F) bed
temperature was 19°C (66°F).
General Operation. No difficulty was experienced In feeding
coal into the bed at any of the operating conditions. No substantial
operational difference was observed between Pittsburgh and Illinois coals.
Although no clinker or accumulation formed in the bed or on the tubes in
the bed, extensive accumulations formed on the cooling coils In the free-
board at 954°C (1750°F) bed temperature, but not at 899°C (1650°F).
The deposits were extremely powdery and crumbled away on touching.
A refractorynke deposit accumulated on the walls of the
primary cyclone but not on the secondary cyclone and interconnecting
ducts. The dense and hard deposit contained mainly calcium and magnesium
compounds. The deposit was more prevalent at 954°C (1750°F) than at
R9t>°C (1650°F). The deposit did not affect the cyclone efficiency. A
refractory-lined cyclone with hotter inside-wall temperature may prevent
this condensation.
Exxon Minlplant
3 4
Combustor nesign '
The combustor was a 610 mm (24 in) steel sheet lined with
refractory to an actual internal diameter of 318 mm (12.5 in). The
144
-------
reactor was constructed in flanged sections with an overall height of
9.76 m (32 ft). The heat transfer surface in the bed was provided with
19 nun (3/4 in) outside diameter stainless steel serpentine tubes on a
57 mm (2-1/4 in) horizontal pitch. Each looo occupied 457 mm (18 in) of
2 2
bed height, and approximately 0.7 m~ (7.5 ft ) of surface area.
Altogether, ten individual loops were available for controlling the
bed temperature. The bottom of the first coil was located 686 mm (27 in)
above the grid ^nd the top of the last coil, 5 m (16.5 ft) above the grid.
The overall heat transfer coefficients were determined during experiments
by measuring the cooling-water flow and temperature. The coolant
(demineralized water) entered and left the combustor through five coolant
distributor plates between flanges at 0.9 m (3 ft) vertical intervals.
Fluidlzing air was supplied by a stationary compressor at
operating pressure up to 862 kPa (125 psig). The air passed through the
distributor plate and out through two stages of cyclones for solids
removal before it was cooled in a heat exchanger. The distributor plate
was a 10 mm (3/8 in) stainless steel plate with 137 air nozzles on a
23.8mm (15/16 in) square pitch. The nozzles were 16 mm (5/8 in) in
diameter and contained eight equally spaced horizontal holes of 2.0 mm
(5/64 in) diameter. The nominal pressure drop through the grid was 483 mm
(19 in) of water. One coal feedpoint was provided at about 279 mm (11 in)
above the grid.
4
Experimental Conditions
The miniplant is in the shakedown phase of operation. The test
conditions to date are summarized and compared with those of the basic
design in Table B-l. The design conditions of the miniplant are similar
to those of the basic design.
Experimental Results
No comprehensive experimental results are available so far;
some important preliminary results, however, can be briefly discussed
here.
145
-------
Temperature Distribution in the Bed. Contrary to the
observation in the BCURA unit, where a small temperature gradient 19°C
(66°F) was observed in a 1.3 m (4.4 ft) bed depth at 927°C (1700°F) bed
temperature, the temperature gradient in the miniplant during shakedown
runs (curves 1 and 2, Figure B-3) was 56 to 83"C (100 to 150°F) for a
0.9 m (3 ft) static bed and 167 to 222°C (300 to 400°F) for a 1.5 m
(5 ft) static bed at 927°C (1700CF) bed temperature. The highest tempera-
ture invariably occurred at the vicinity of the coal feedpoint.
Substantial improvements in fluidization and heat dispersion
throughout the fluidized bed combustor have been accomplished recently
by modifications to the coil configuration, orientation, and distribution
along the bed height. The bed temperature variations are currently less
than 67°C (120°F) within the first 3 m (10 ft) of the expanded bed at a
combustor operating temperature of 927°C (1700°F) (see curves 3 and 4 in
Figure B-3.) Vertical tube orientation was found to be superior to hori-
zontal tube orientation in smoothing out the temperature gradient in the
combustor.
Corrosion of Boiler Tubes in the Bed. Samples from the
bottom-most cooling coil were examined after exposure to the high-
temperature coal combustion conditions for approximately 50 hours. The
temperature in the combustion zone near this coil ranged from 816 to
1010°C (1500 to 1850°F). Short intervals of temperatures exceeding 1066°C
(1950°F) were also experienced. The water temperature inside the coil
tube varied from 13ft to 160°C (280 to 320°F). The results of the metal-
lurgical analysis showed no evidence of corrosion or deterioration of
the coil.
A section of the 3 mm (1/8 in) stainless steel supporting rod
was also subjected to metallurgical analysis. The rod was not cooled
and ,therefore ,was exposed to the high-temperature environment existing
in the combustor. The results showed carbide precipitation in the grain
boundaries and signs of corrosion.
General Operation. No great difficulty was encountered in
operating the combustor. The pressurized coal-feeding system was
146
-------
100 200
300
Height Above Grid, cm
400 500 600 700
800
Curve 681107-8
900 1000
2000
1800
1600
1400
1200
1000
800
600
Bed Depth, m (ft)
Combustor Press., kPa (atm)
Fluidizing Velocity, m/sec (ft/sec)
Coal Feed Rate, kg/hr (Ib/hr)
Expanded Bed, m (ft)
-0.61 (-2.0)
901.6 (8.9)
1.89 (6.2)
68.6(151)
0.91(3.0)
901.6(8.9)
2.07(6.8)
66.7(147)
1.78(5.8)
911.7(9.0)
1.83(6.0)
149.4(329)
1.52(5.0)
921.8(9.1)
3.20(10.5)
142.0(313)
-1.52 (-5.0) 1.52(5.0) >3.00(>10) 3.29(10.8)
•-4
I I
1000
o
o
900 £
1
800 I
700
600
500
400
20 60 100 140 180 220 260
Height Above Grid, inches
300 ' 340 380 420
Figure B-3-Fluidized bed combustion miniplantcombustor temperature
-------
adequate but still could be improved to provide more reliable service.
Distortion of the bottom-most cooling coil was extensive, but all other
coils were in satisfactorv condition. The bottom coil was compressed and
pushed to one side. Two vertical bends were flattened, and most of the
3.0 mm (1/8 in) stainless steel supporting rods were either bent or broken
loose. The cause of this distortion is still unknown.
Evaluation of the Basic Design
The most comprehensive pressurized fluidized bed coal combustion
data were obtained from the BCURA unit. Unfortunately, the operating
conditions were at considerably lower fluidizing velocities and lower bed
depths. Even with this handicap extrapolation of the results to higher
fluidizing velocities and higher bed depths mav be possible with a mini-
mum of error. Exxon's miniplant, which was designed to operate at
essentially the basic design conditions, will provide data directly
applicable to the basic design. The basic design conditions are reevalu-
ated below on the basis of information presently available.
Combustion Efficiency
At low velocities (^ 0.76 m/sec/'v- 2.5 ft/sec) and low bed depths
(^ 1.37 m/^ 4.5 ft), the BCURA unit obtained a combustion efficiency of
96 percent at 10 percent excess air, 97.5 percent at 17 percent, and
99 percent at 25 percent. The BCURA unit also operated at a lower
pressure (506.5 kPa/5 atm). Increasing the fluidizing velocity to 3.0 m
(10 ft/sec), bed depth to 3 to 3.7 m (10 to 12 ft), and pressure to 1013 kPa
(10 atm) may give similar combustion efficiency or higher because an
increase in both the bed height and the operating pressure will increase
the combustion efficiency; the coal residence time in the bed increases
with increasing bed height and decreases with increasing fluidizing
velocity. In a 114 mm (4.5 in) inside diameter pressurized batch
combustor, Exxon observed a 98.1 percent combustion efficiency operating
at 628 kPa (6.2 atm), 927°C (1700°F), 58 percent excess air, 1.9 m/sec (6.2
ft/sec), and 0.6 m (2 ft) bed depth without carbon recycle. If the observed
148
-------
combustion efficiencies are typical, a carbon burn-up cell may not be needed
in the basic boiler design. This will considerably simplify the operation
and control of a pressurized fluid bed combustor, although the corresponding
saving in reactor cost is small.
Heat Transfer Coefficient
The basic design assumed an average bed-tube heat transfer
coefficient of 284 W/m2-°K (50 Btu/ft2-hr-°F). The bed-tube heat transfer
coefficient depends on physical properties of the bed material and fluidizing
gas—such as particle size, thermal conductivity—and on operating conditions-
such as temperature, pressure, and fluidizing velocity. The BCURA unit
2
attained an average bed-tube heat transfer coefficient of 380 W/m -°K
(67 Btu/ft2-hr-°F) at 8998C (1650°F) bed temperature and 403 W/m2-°K
(71 Btu/ft2-hr-°F) at 954°C (1750°F). The presence of the tube bank did
not affect the heat transfer coefficient significantly. The overall heat
transfer coefficient with Exxon miniplant have been measured to be 284 to
312 W/m2-°K (50 to 55 Btu/ft2-hr-°F) with the bed temperature at 927°C
(1700°F) velocity at 1.83 m/s (6 ft/sec), and pressures at 911.7 kPa (9 atm).
Corrosion/Erosion Heat Transfer Tubes in the Bed
Again, the most comprehensive data were from the BCURA unit.
The corrosion/erosion data for the boiler tubes In the bed were reviewed
earlier. The more recent data are reviewed here. The boiler tube
materials proposed for use in the basic design and the conditions they
will be subjected to are summarized in Table B-3. Chemical compositions
of the selected boiler tube materials are presented in Table B-4. The
chemical composition of the material tested in the BCURA unit is pre-
sented in Table B-5 and the corrosion results in Table B-2 and Figure B-2.
Note that the current tests are of short duration. Weight loss usually
reached a reliable, steady state after long duration (see Figure B-2 and
Reference 7). The steady-stage weight loss was considerably lower than
the initial weight loss. On the basis of these experimental results,
the wastage was not excessive for alloys operating up to their recommended
maximum design temperatures for use in steam boilers.
149
-------
Table B-3
BOILER TUBE MATERIALS
Tube locations
Pressurized Boiler
Proposed
material
Design temp.
°C (°F)
Water Walls
Preevaporator
Evaporator
Superheater
Reheater
SA-213-T22
SA-210-A1
—
SA-213-T2
SA-213-T22
SA-213-TP-304H
SA-213-T22
524 (975)
389 (732)
—
482 (900)
570 (1058)
621 (1150)
606 (1122)
A cost study was performed on the assumption that higher
alloys were required for the boiler tubes. The originally proposed
tube materials and the assumed tube materials for cost study are compared
in Table B-5. The metal wastages due to corrosion/erosion projected for
30 years are included. Weight losses were calculated on the basis of long-
duration tests performed in the BCURA unit on comparable tube materials.
The average weight losses in 500 hours' operation were taken to be typical
wastages for long-term projection. This was a conservative assumption
because weight losses level off considerably after about 100 hours'
operation. If the steady-state weight losses after 500 hours' operation
are taken to be the typical long-term wastages, the wastages in Table B-5
could presumably be reduced by about a factor of three. At these weight
losses, the proposed 2-1/4 percent chromium steel (SA-213-T22) for
reheater and upper-superheater in the basic design was probably a little
bit too optimistic. A new cost study was performed by assuming that the
18 percent chromium alloy (SA-213-TP304H) would be used for all beds except
the boiler tubes in the preevaporator, where the 2-1/4 percent chromium
150
-------
Table B-4
CHEMICAL COMPOSITIONS OF SELECTED BOILER TUBE MATERIALS
( in wt %)
1
SA-210-A1
SA-213-T2
SA-213-T22
SA-213-TP304H
Welded SA-178-A
tubing (low-carbon steel)
Seamless SA-192
tubing
Carbon |
0.27 max
0.10-0.20
0.15 max
0.04-0.10
0.06-0.18
0.06-0.18
E Manganese I
0.93 max
0.30-0.61
0.30-0.60
2.00 max
0.27-0.63
0.27-0.63
Phosphorus I
0.048 max
0.045 max
0.030 max
0.04 max
0.05 max
0.048 max
Sulfur |
0.058 max
0.045 max
0.030 max
0.03 max
0.06 max
0.058 max
Silicon
0.10 min
0.10-0.30
0.50 max
0.75 max
-
0.25 max
I Chromium I Molybdenum | Nickel
_
0.50-0.81 0.44-0.65
1.90-2.60 0.87-1.13
18.00-20.00 - 8.00-11.00
-
- - -
V/i
-------
steel would be sufficient. The cost for the boiler tubes (including
fabrication) increased from the original $1.3/kw to about S4.1/kw, a
sizable Increase for boiler tube cost; however, it represented less than
a one percent increase in total plant cost. This cost estimation was
conducted with 1972 dollars so that a comparison with cost figures in the
earlier reports ' would be valid.
Table B-5
TYPICAL ANALYSES OF METAL SPECIMENS
Designation
12% Cr
Rf 36
SF 316
Esshete 1250
P.E. 16
1% Cr 1/2% Mo
2-1/4% Cr 1% Mo
Nominal
Cr | Ni | Mo
12
18 12
17 12 2.5
15 10 1
18 37 5
1 0.5
2.25 1
composition %
I Mn | Ti | Al | Nb
1
2
6 1
1.2 1.2
0.5
0.5
1 Fe
88
69
66.5
67
37
98
96.25
To evaluate the present fluidized bed boiler designs critically,
longer-duration runs of up to 2000 hours should be performed at a higher
velocity (up to 4.57 m/sec/15 ft/sec), and a higher pressure (up to
1013 kPa/10 atm). The effect of adding limestone and of burning dif-
ferent coals with different impurities, especially high-sulfur coals,
should be studied.
The preliminary evaluation with the available data indicates
that conventional boiler tube materials mav be used in the fluidized
bed boiler if the long-duration runs show that the rate of weight loss
does, indeed, level off with an increase in operation time. Otherwise,
higher alloy materials can be used with minimal cost oenalty. Final
assessment can be carried out only after more data are accumulated at
higher velocity, higher pressure, high temperature, and longer duration.
152
-------
Temperature Distribution in the Bed
There were two conflicting observations on temperature distri-
bution in two different pressurized fluidized bed combustors. BCURA
obtained a reasonably uniform temperature profile in experiments —
6°C at 899°C (10°F at 1650°F) bed temperature and 19°C at 954°C (35°P
at 1750°F). Exxon, however, had difficulty in reducing the bed tempera-
ture gradients to less than 56°C (100°F) at 0.9 m (3 ft) static bed depth
and to less than 167°C (300°F) at 1.5 m (5 ft} static bed depth in the
minlplant. This difference in observation probably stems from the
difference in operating conditions and in design parameters. A few
factors can be elaborated here.
• The miniplant was operated at higher pressure—911.7 kPa
(9 atm) versus 506.5 kPa (5 atm) for the BCURA unit. Higher
pressure resulted in a higher combustion rate. This tended
to promote combustion in the vicinity of the coal-feeding
nozzles. The dissipation of heat was then dependent on the
solid mixing rate in the bed, which is slow when compared to
the combustion rate.
• Coal-particle size distribution will also affect the
temperature distribution. Too large a quantity of fines
will promote rapid combustion near the coal-feeding nozzles
and increase the temperature gradient in the bed. The fines
in the coal fed to the BCURA unit were screened out, which
helped the temperature distribution.
• The designs of the BCURA unit and the Exxon miniplant are
different in bed area, cooling coils arrangement, and
coal feed-rate. The design parameters for the basic design,
the BCURA unit, and the Exxon miniplant are compared in
Table. B-6. Heat release rates per unit bed area and per
unit bed volume are essentially similar for the basic design
and the Exxon miniplant. These heat release rates are five
times larger than those in the BCURA unit, based on unit bed
area, and about twice as large, based on unit bed volume.•
153
-------
Table B-6
COMPARISON OF DESIGN PARAMETERS BETWEEN THE BASIC DESIGN
AND THE PILOT-SCALE EXPERIMENTAL UNITS
_L
Basic design
J_
BCURA unit
Exxon miniplant
Ln
Bed Cross-Section
Bed Height
Bed Height/Diameter Ratio
Tube Packing, Z bed orobs-sectlnn
Z bud volume
Heat Release R,ice/lied Arcj
Volume Meat; Relu.ise R.itc
Bed Area/Coal Feeding Nuzzle
1.52 n x 2.13 m (3.25 m )
(5 ft x 7 ft (35 ft2|)
3.35 m-4.27 m (11-14 ft)
2-2.5
21.S-2B.5
17-22.5
1.2-2.0 x 107 J/m2-si-c
(3.8-6.3 x 106 Blu/rt2-lir)
3.6-4.7 x 10ft l/m3-si-.(
(3.5-4.5 x 10"' lltu/lt J-hr)
0.81 m2
(8.75 ft2)
0.61 m x 0.91 m (0.56 m )
(2 ft x 3 ft |6.ft2|)
l.'J7 m (4.5 fl)
2
17
8
J.2 x 106 J/m2-scc
(1.0 x 10h Btu/ft2-hr)
i.\ x I06 J/m3-si-c
(i.} x 10** Btii/flJ-lir)
0.14 m2
(1-5 fl2)
0.3175 m I.D. (0.08 m )
(12.5 in I.D. |0.85 ft2])
3.05 m-4.6 m (10-15 ft)
9.5-14.5
28
9
1 .4h x 107 J/m2-sec
(4.6 x 106 Htu/ft2-hr)
J.7 x 10h .l/ra3-sec
C\.b x I05 Mtu/ft^lir)
O.OH m2
(O.H5 ft2)
For beds of rectangular cross-section, the hydrnullc diameters .ire usud.
Basic design has four fluldlzed beds of slightly different designs pur module for prccvaporntion, superheatiiiK, and reheating.
-------
These differences in design are primarily due to differences
in bed depth. The high heat release rate will certainly contri-
bute to the poor temperature distribution. The poor temperature
profile may also result from the size difference. It is
significant that the bed height/diameter ratio for the
BCURA unit and the basic design is around 2 while that
for the miniplant is 10 to 15. The solid mixing rate has
been found to depend strongly on the bed depth/bed diameter
Q
ratio. Thus, it is dangerous to extrapolate the result of
the bed-temperature profile obtained in the miniplant directly
to that of the basic design.
Other factors which will affect the temperature distribution
are the tube arrangement and bed area per unit coal-feeding nozzle. The tube
packings are compared In terms of percentage bed volume in Table B-6. The
tube packing will affect the rate of solid mixing and the rate of heat
removal from the bed. The effect of tube packing on the bed temperature
profile is not clearly known. The effect of bed area on the coal-feeding
nozzle is also not known at this moment. These all point to the necessity
of a large, flexible, fluidized bed combustion facility to study the factors
identified here.
From the existing evidence It is clear that a reasonably uniform
temperature profile can be obtained in a pressurized fluidized bed corn-
bus tor if the operating conditions and the design parameters are properly
set. More data from large combustors operating at high pressure, high
temperature, high velocity, and large bed height are required to verify
the basic design conditions. The conditions still appear reasonable on
the basis of available data.
Coal Feeding
No difficulty was experienced in multiple feeding of coal
into the pressurized BCURA unit at any operating conditions up to 506.5 kPa
(5 atm) pressure. Exxon did encounter some difficulties in trying to
maintain a steady flow of coal to the combustor. In view of the shakedown
155
-------
nature of the miniplant, this difficulty should not constitute a major
problem. In fact, in some more recent runs coal up to more than 136 kg/hr
(300 Ib/hr) was successfully fed into the corabustor. In a commercial unit
coal will be fed at more points and at higher rates. On the basis of
current experience, no serious trouble is anticipated.
General Operation
Pittsburgh and Illinois coals were combusted in the BCURA unit
without substantial difficulty. No clinker or accumulation formed in the
bed or on the tubes in the bed. Exxon's miniplant shakedown operation also
uncovered no serious flaws. BCURA did observe extensive accumulations
on the cooling coils in the freeboard at 954°C (1750°F) bed temperature,
but not at 899°C (1650°F); the deposits were extremely powdery and
crumbled away on touching. The accumulations may also relate to the
particular coal used. Thus, it does not seem to be a serious problem.
A refractory-like deposit, mainly of calcium and magnesium
compounds, also appeared on the walls of the primary cyclone primarily due
to condensation. A refractory-lined cyclone with hotter inside-wall
temperature may avoid this problem.
Another important facet which cannot be studied in smaller
combustors is the controllability of a pressurized fluidized bed boiler.
The response of the boiler to load-follow and the turndown capability are
important in plant operation. No data are yet available.
156
-------
REFERENCES
1. Keairns, D. L. et al. Evaluation of the Fluldized-Bed Combustion
Process - Vol. I, Pressurized Fluid!zed-Bed Combustion Process Develop-
ment and Evaluation. Environmental Protection Agency. Westinghouse
Research Laboratories. Pittsburgh, Pennsylvania. December 1973.
Publication No. EPA 650/2-73-048a. NTIS PB 23] 162/9.
2. Pressurized Fluidized Bed Combustion. Final report. Contract 14-32-
0001-1511. The Office of Coal Research, Department of the Interior.
National Research Development Corporation. London, England.
November 1973.
3. Nutkls, M. S. and A. Skopp. Design of Fluidized Bed Miniplant.
Proceedings of the Third International Conference on Fluidized Bed
Combustion. Hueston Woods Lodge, Ohio. October 29 - November 1, 1972.
4. Exxon Research and Engineering Program. (Presented at FBC Contractors
Meeting. Argonne National Laboratories. September 11-12, 1974.)
5. Nutkis, M. S. Pressurized Fluidized Bed Coal Combustion. Proceedings
of the International Fluidization Conference. Asilomen Conference
Grounds. Pacific Grove, California. June 15-20, 1975.
6. National Coal Board. Reduction of Atmospheric Pollution. Final Report.
Environmental Protection Agency. September 1971.
7. Archer, D. H. et al. Evaluation of the Fluidized Bed Combustion Process-
Vol. II, Technical Evaluation. Environmental Agency.
Westinghouse Research Laboratories. Pittsburgh, Pennsylvania.
November 1971. NTIS PB 212 960/9.
8. Kunii, D. and 0. Levensplel. Fluidization Engineering. New York.
John Wiley & Sons, Inc., 1969.
157
-------
APPENDIX C
SORBENT SELECTION AND ALTERNATIVE SORBENTS
-------
APPENDIX C
SORBENT SELECTION AND ALTERNATIVE SORBENTS
SORBENT SELECTION
The use of calcium-based sorbents (limestone or dolomite) to
trap in solid form the sulfur released from "dirty" fuels during combus-
tion is based on the thermodynamic stability of calcium sulfate under
fluidized bed combustion conditions;and on the kinetic efficiency, struc-
tural integrity, and economical availability of the sorbent, which permit
application of the basic idea in a technically sound process.
Calcium carbonate is found as both limestone and dolomite in
the eastern and midwestern states which produce high-sulfur coal. Both
pure and impure limestones and dolomites have been tested as sorbents:
the sulfur removal efficiency has generally been well within EPA limits.
In order to select the best available material as sorbent, sorbent selec-
tion criteria must be developed. Establishing selection criteria will
minimize the cost and time involved in assessing the usefulness of rock
quarried near a particular plant site.
The criteria (which require further development) for choosing
a stone are based on:
• Acceptor properties of the stone for sulfur removal
• Attrition resistance of the stone
• Trace element emission characteristics
• Regeneration characteristics
• Suitability of spent sorbent for final processing for
disposal
• Economic availability of the stone.
159
-------
Acceptor Properties of the Stone for Sulfur Removal
The acceptor properties of the stone depend on the stable form
of the sorbent in the particular process, and on the kinetics of the sul-
fation reaction.
Both limestone and dolomite may be used as sorbents at atmos-
pheric pressure. At higher pressures and under exit gas conditions as
Figure C-l shows, calcium carbonate is the stable form of the sorbent.
Calcium carbonate in limestone does not react rapidly with sulfur
dioxide: the sorbent probably depends for its desulfurizing action on
the extent of calcination which occurs at or near the bed air inlet.
Both sulfur dioxide and carbon dioxide will then compete for the calcined
2
lime. Although Exxon Research and Engineering (Exxon) has been suc-
cessful in using limestone as a sorbent at pressure, it is uncertain if
there are high-pressure and low-temperature limits to its use. Dolomite,
on the other hand, is normally reactive in the half-calcined state and
may be used as sorbent, irrespective of the carbon dioxide pressure in
the system. The calcium-to-sulfur molar ratio fed to the bed, however,
will be higher with half-calcined than with fully-calcined dolomite.
The kinetics of the sulfation reaction are discussed in
Appendix D; the application of the data to estimating fluidized bed de-
sulfurization performance is described in Appendix E. To date, most of
the data have been concerned with the effects of process variables on a
few stones. For those stones for which thermogravimetric (TG) data are
available, an estimate of performance can be made, given the proposed
operating conditions. For any proposed system it is also possible to
carry out TG tests on a particular stone to estimate its suitability as a
sorbent under the system operating conditions. Extension of the data
base on the kinetics of dolomite and limestone sulfation to cover the
major variations in stone type encountered in the eastern United States
is desirable.
160
-------
990
(1814)
950
£ (1742)
p 910
J1670)
§ 870
£ (1598)
I 830
£(1526)
790
(1454)
The Boundaries Shown Have Been Calculated
from Data for the Reaction CaC03 z CC>2
Given by Curranl
i
I
I
10% Excess Air
T
300% Excess
Air
§*
u o
•s J
fc c
Q. OJ
E cD
-------
Attrition Resistance
The term attrition is used to cover all aspects of loss of sor-
bent from the fluidized bed by elutriation, irrespective of the mechanism.
Decrepitation, thermal shock shattering, bursting apart during calcina-
tion, fragmentation during sulfation, abrasion between sorbent and ash
and sorbent and refractory or internals, distributor jet impingement, and
sorbent circulation through ducts may all contribute to loss of sorbent
from the bed. It is impossible to establish criteria for attrition re-
sistance which must be met without reference to the system design, since
the system design itself may take into account the attrition loss ex-
pected from available sorbents. In general, however, a minimal amount of
attrition loss is desirable, if only to reduce the load on dust removal
equipment. Additionally, for regenerative systems where long cumulative
residence times in the desulfurizing system are required, the rate of
calcium makeup feed to the bed may be governed by attrition losses. At
the other end of the scale, continuous recirculation of fines assists in
attaining higher stone utilization in once-through systems, or in systems
with a high calcium-to-sulfur molar feed rate.
The prediction of attrition losses a priori for any particular
stone is not yet possible. In stone selection tests for the chemically active
fluidized bed (CAFB) oil gasification process, initial laboratory tests
screening candidate stones met with some success. The extremes of high
attrition rate and low attrition rate correlated well with results ob-
3
tained on Esso Research Centre, Abingdon, England (Esso) batch units. An
intermediate range of attrition loss in the laboratory unit, however, yielded
no information on the suitability of the stone as a sorbent in that parti-
cular process. At the moment, laboratory tests can give a relative list-
ing of the expected order of attrition losses from a series of stones.
The evidence, however, is that suitable adjustment of the process vari-
ables may alter attrition loss sufficiently to accommodate a particular
stone to a particular process. For example, Pope, Evans and Robbins
4
(PER) successfully modified initial calcination conditions in their
fluidized bed combustor to permit operation with a local, impure limestone.
162
-------
The original conditions, standardized for operation with a high-purity
limestone, had resulted in such a high loss rate that the bed could not
be maintained. Another approach is evident in the work of Pell et al.
at Conoco Coal Development Company. Their procedure for hardening the
stone sorbent has increased the number of cycles for which the sorbent
can be circulated between desulfurization (under reducing conditions) and
regenerator, permitting a low sulfur differential between the two
vessels.
The development of relative measures of attrition rates depends
on further research into the mechanism of particle fracture, in addition
to laboratory measurements of decrepitation and fluidized bed attrition
during calcination and sulfation. Experiments to generate sufficient
data,under identical process conditions, which compare attrition losses
for a range of dolomites and limestones of different grain structure to
permit empirical evaluation of stone sorbents should proceed in order to
assist in the screening of candidate stones.
Trace Element Emission Characteristics
The possibility of the emission of trace elements from the sor-
bent into the effluent gas from the fluidized bed combustor raises envi-
ronmental and corrosion questions. These considerations are discussed in
Appendix H. The present state of knowledge as it affects stone selection
is summarized here.
The major concern is that sodium and potassium liberated into
the gas stream will form liquid deposits on turbine components, thereby
inducing hot corrosion. Preliminary results to date i'ndicate that a very
small fraction of the alkali-metal content will be liberated from dolo-
mites during fluidized bed combustion. The major alkali-metal species
emitted are potassium compounds. Since it is not yet known what potas-
sium species in the original sorbent liberates its potassium to the
gas stream, the best approach is to use the dolomite with the lowest
alkali-metal content. Analysis of a range of dolomites from Ohio,
Indiana, Illinois, and Michigan indicates that the potassium content
varies from 100 to 6000 ppm (by weight), while the sodium content
163
-------
varies from 150 to 350 ppm (by weight). Occasional large deviations from
these values may be encountered: for example, a Bahaman aragonite dredged
from the sea contains 4,000 ppm sodium. To date, no correlation of
alkali-metal content with geological formation has been found. The
sparse data available show that the alkali content may vary significantly
through a series of rock strata, and within one stratum from quarry to
quarry. It is extremely doubtful whether reported values in the older
literature should be used,other than for internal comparison with other
reported data: details of the analytical procedures adopted have rarely
been reported.
In the current state of knowledge, the best recommendation is
to choose as sorbent from those available the dolomite or limestone which
is lowest in potassium content.
Regeneration Characteristics
Data at the moment are insufficient to distinguish between dif-
ferent stones on the basis of regenerability. Optimization of regenerabi-
lity will in all likelihood include a study of the influence of stone
structural type on the loss of kinetic activity in desulfurization, both
for the high-temperature and low-temperature (two-step) regeneration
schemes.
Suitability of Spent Sorbent for Final Processing for Disposal
Any final processing of the spent sorbent required before dis-
posal will probably be a local problem—dependent on the particular flui-
dized bed combustion system operated ,on regulations governing disposal
of solid wastes, or on local marketing opportunities. The major influ-
ence these factors may have is in requiring a choice between limestone or
dolomite. Investigation of this aspect of stone selection is dependent
on the production of tonnage quantities of spent sorbent so that its prop-
erties may be characterized. (See Appendices F and G.)
164
-------
Economic Availability of the Stone
The cost of dolomite in the eastern and midwestern states was
in the range of $2.30to 5.00/Mg (ton) at the plant (late 1974). Signifi-
cant transport costs are likely if the sorbent must be hauled for long
distances. For this reason, operation with local sources of stones will
be preferable. While systems studies can reveal the impact of sorbent
cost on plant economics, additional work to determine the design condi-
tions needed to operate (desulfurize) with a wide variety of stone types
is equally important. Successful operation of the Rivesville plant with
a local, impure limestone Indicates the potential for this approach. TG
work to modify the properties of the stone sorbents by special calcina-
tion treatment in order to develop suitable stone porosity for desulfuri-
zation was successful with pure and impure dolomites and with a pure
limes tone.
Conclusion
The establishment of stone selection criteria for choosing
limestones and dolomites suitable for use as fluidized bed desulfurizing
agents is an important part of the technical support studies needed to
optimize the fluidized bed combustion process.
The stone selection criteria developed will not be rigid speci-
fications but will depend on the particular system design. The availabi-
lity of particular sorbents will influence optimization of a particular
design with respect to local conditions.
ALTERNATIVE SORBENTS
Because of the low cost and widespread geographical availabi-
lity of calcium carbonate as limestone or dolomite rocks and its excel-
lent performance as a sorbent in preventing sulfur dioxide emission, a
compelling reason must exist before an alternative sorbent is considered
as a substitute. Apart from cases where an alternative to calcium car-
bonate might be considered because of local availability, the most gene-
ral ground for assessing alternative sorbents is that of regenerability.
165
-------
Calcium sulfate is extremely stable and requires the expenditure of con-
siderable energy before it will release sulfur, either as sulfur dioxide
or as hydrogen sulfide, in concentrated form. The desirability of regene-
ration rests on two factors—the recovery of sulfur as a valuable re-
source and a reduction in both the material requirements for sulfur
sorption and the quantities of spent sorbent solid which are produced as
a by-product of the sulfur removal process. The ideal alternative sor-
bent is one which is mechanically resistant to attrition in the solids
circulation system of the fluidized bed process; is an efficient sulfur
getter over the range of coal combustion conditions in the fluidized bed
process; can be regenerated under mild reducing conditions (or by thermal
decomposition alone); and yields a solid oxide which retains the capacity
and kinetic activity of the original sorbent on repeated recycling. In
addition, it should contribute no fine particulate matter or trace ele-
ments to the effluent gases and should trap these materials as they are
released from coal and coal ash during combustion.
A large number of metal oxides have been screened for their
suitability in removing sulfur dioxide from flu gases. The thermodynamic
data tabulated for the stability of metal sulfates rule out most of these
oxides for consideration as sorbents in the fluidized bed combustion pro-
cess.
The free-energy diagram for metal sulfate stability shown by
Bartlett, indicates that at 727°C (1341°F), the order of thermal stabi-
lity of the common metal sulfates is:
• Potassium sulfate (K.SO.)
• Sodium sulfate (NaSO.)
• Calcium sulfate (CaSO.)
• Manganese sulfate (MnSO.)
• Nickel sulfate (NiSO.)
• Copper sulfate (CuSO,)
• Aluminum sulfate
Calcium sulfate requires severe treatment in reducing gas (in other words,
high temperature) to release the sulfur, indicating that the more stable
alkali-metal sulfates may be excluded.
166
-------
At the request of EPA, a special alumina-based copper oxide
catalyst, HALCO 471, was evaluated as a potential sulfur removal sorbent.
Both the thermodynamics and kinetics of the reaction of sulfur dioxide
with the catalyst were examined.
Thermodynamic Evaluation
Copper sulfate (CuSO.) decomposes thermally in two stages as
shown by the equations:
2 CuS04 2 CuO • CuSO^ + S02 + y QZ (C-l)
CuO • CuS04 J 2 CuO + S02 + j 02 (C-2)
At a given temperature in the range of interest for fluidized
bed combustion the equilibrium pressure of sulfur dioxide is lower for
reaction 2 than for reaction 1. Therefore, reaction 2 should be con-
sidered as the process which will thermodynamically limit the sulfur re-
tention in a fluidized bed of copper oxide. The degree of dispersion of
copper oxide on the alumina base, however, may either prevent formation
of the oxysulfate or, indeed, stabilize it. Mixtures of sodium sulfate
and copper oxide are more effective sulfur dioxide sorbents than copper
oxide alone; the sulfation reaction is faster than with copper oxide; and
thermal decomposition requires heating to a higher temperature.
The limitations which reaction 2 places on sulfur dioxide re-
moval are shown in Figure C-2 for: (A) the 101.3 kPa (1 atm) pressure
combustor, (B) the 1013 kPa (10 atm) combustor with 10 percent excess air,
(C) the 1520 kPa (15 atm) with 10 percent excess air, and (D) the 1520 kPa
(15 atm) adiabatic combustor. Table C-l shows the assumed exit gas condi-
tions .
o
Exxon reported that their calculations indicated that sulfur
retention would fall off from more than 90 percent to 20 percent in the
range 627 to 727°C (1161 to 1341°F).
167
-------
CO
100
90
Curve 658089-A
Temperature, °F
1022 1112 1202 1292 1382 1472 1562 1652 1742 1832
oo
•s
C£
CM
o
co
70
60
50
40
30
20
10
Proposed Operating
Temperatures for
Fluidized Combustor
10% Excess Air
.. 101.3kPa
(1 Atm) Combustor
R 1013 kPa
°" (10 Atm) Combustor
c_1520kPa
(15 Atm) Combustor
300% Excess Air
1520 kPa
D-Adiabatic Combustor
(15 Atm)
550
600 650
700 750 800
Temperature, °C
850
900 950
1000
Figure C-2~ Maxim urn S02 retention in fluidized beds of CuO- Cu S04
(thermodynamic limit)
-------
Table C-l
EXIT GAS CONDITIONS FOR FLUID BED COMBUSTORS
System
A.
B.
C.
D.
101.3 kPa (1 atm) fluidized bed
with 10 percent excess air
1013 kPa (10 atm) pressure with
10 percent excess air
1520 kPa (15 atm) pressure with
10 percent excess air
Adiabatic combustor 1520 kPa
(15 atm) pressure
PS02 with
no sorbent
(atm)
.005
.05
.075
.0188
P02
(atm)
.04
.17
.255
2.265
Experimental Results
TG experiments were carried out at atmospheric pressure using
the copper-impregnated alumina catalyst and an oxide prepared by thermal
decomposition of copper sulfate pentahydrate.
The copper sulfate pentahydrate decomposed to yield a 32.1 per-
cent weight loss for dehydration (theoretical value = 32.3 percent )and a
36.1 percent weight loss for decomposition of sulfate to oxide (theoreti-
cal value =36.1 percent). The oxide was then exposed to 12 percent
sulfur dioxide/nitrogen/4 percent oxygen and the temperature varied in
the range 600 to 800°C (1112 to 1472°F). Equilibrium temperatures were
noted at the transition from weight gain to weight loss.
As expected, two equilibrium temperatures were noted,and the
temperatures were in relatively good agreement with those predicted from
9
the experimental thermodynamic data in the literature, as shown in
Table C-2.
This is not to suggest that the data were taken under the
precisely-controlled conditions Ingraham reported: his values are noted
to indicate that the TG values are fairly close to the true values and
are a guide in interpreting the behavior of the sorbent.
169
-------
Table C-2
DECOMPOSITION TEMPERATURES OF COPPER SULFATE AND
COPPER OXYSULFATE: COMPARISON OF © TG DATA AND
THERMODYNAMIC PREDICTIONS FROM INGRAHAM'S STUDY
TG temperature I&M temj
°C (°F) °C 1
720
773
782
>erature
[°F) Reaction
(1328) 740 (1364) 2 CuSO^ •*• CuO • CuSO^
(1422)
(1438) } 778 (1432) 795 (1/i63> Cu° ' CuS0^ ->• 2 CuO +
+ so2 + I o2
so2 + 1 o2
T.R. Ingraham. Thermodynamics of the Thermal Decomposition of Cupric
Sulfate and Cupric Oxysulfate. Trans. Met. AIME 233; 359, 1965.
-------
The NALCO 471 catalyst, impregnated with copper oxide (as sup-
plied by Exxon) was then tested in the1 same sulfur dioxide/nitrogen/
oxygen mixture. For runs with two samples, only one equilibrium tempera
ture was found at 684°C ± 5°C (1263°F ± 41°F) (9 readings). This appar-
ently anomalous behavior results from the reaction
3 S02 +
Calculation of the temperature for equilibrium at the experimental gas
10
pressures, using the thermodynamic tabulation of Stern and Weise, gave
680. 5°C (1257°F). In addition, the total quantity of sulfur trioxide (SO,)
reacted was about three times stoichiometric for formation of copper sul-
fate (on a 5 wt % copper basis). It was decided to reverse the
experimental procedure and start the absorption/temperature run at high
temperatures, cool the solid, and determine the temperature for formation
of either copper sulfate or copper oxysulfate free from interference from
the alumina.
Copper Oxide Sorbent
Pressurized sulfation runs (1013 kPa/10 atm) were carried out
on the NALCO catalyst. A trace reaction was noted at 900°C (1652°F) . On
cooling the sample in the gas flow (0.5 percent .sulfur dioxide; 4 percent
oxygen in nitrogen) , the sorbent gained weight at the temperature where
aluminum sulfate becomes thermodynamically stable. The sample lost
weight on recycling through this temperature (650°C [1202°F]). A blank
run on colloidal alumina showed a trace reaction at 900°C (1652°F), simi-
lar to that observed for the NALCO catalyst, a weight gain corresponding
to 1 percent of the original sample weight.
It was concluded that the copper-impregnated catalyst is not a
suitable sorbent for sulfur dioxide removal in fluidized bed combustion:
the sulfur dioxide adsorption step takes place below the practical tem-
perature range for operation of the fluidized bed combustor.
171
-------
General Assessment
The chief interest in alternative sorbents lies in finding a
sorbent which can undergo multiple regeneration. For this reason an
analysis of the minimum acceptable performance of a sorbent should be
carried out. This analysis would require definition of
1. A range of acceptable hydrogen sulfide or sulfur di-
oxide concentrations produced in the regenerator
2. Acceptable fuel consumption in the regenerator
3. Minimum sulfur loading on the sorbent and acceptable
stone recirculation rates
A. Thermodynamic screening of the sorbents in light of
1 and 2.
5. Kinetic tests for sorbent activity In the light of 4.
The criteria developed would be system dependent: the atmos-
pheric pressure, pressurized boiler, and adiabatic combustor cases would
each be treated separately.
172
-------
REFERENCES
1. Curran, G.P., C.E. Fink, and E. Gorin. GQ Acceptor Gasification
Process in Fuel Gasification. Consolidation Coal Co. Advances in
Chemistry Series 69. American Chemical Society. Washington, D.C.
1967. p. 141.
2. Hoke, R. C., L. A. Ruth, and H. Shaw. Combustion and Desulfurization
of Coal in a Fluidlzed Bed of Limestone. Exxon Research and
Engineering Co. Linden, N.J. (Presented at IEEE-ASME Joint Power
Generation Conference, Miami Beach, Fla. Sept. 15-19, 1974.)
3. Keairns, D. L., et al. Fluidized Bed Residual Oil Gasification/
Desulfurization at Atmospheric Pressure. Vols. I and II. United
State Environmental Protection Agency. Westinghouse Research
Laboratories. Pittsburgh, Pennsylvania. Contract No. 68-02-065.
December 1974.
4. Mesko, J. E., S. Ehrlich, and R. A. Gamble. Multicell Fluidized-Bed
Boiler Design Construction and Test Program. Office of Coal Research.
Pope, Evans and Robbins. New York, N.Y. NTIS PB 236-254. August 1974.
PB 236-254. August 1974.
5. Pell, M. Conoco Coal Development Co. Private Communication. 1974.
6. Bartlett, R. W. Sulfation Kinetics in SO, Absorption from Stack
Gases. Environmental Protection Agency. Stanford University.
Palo Alto, California. Grant No. AP00876. June 1972.
7. Bumazhmov, F. T. Sulfation of Copper Oxide by Mixtures of Air and
Sulfur Dioxide. Izv. Vyssh. Ucheb. Zaved., Tsvet. Met. 22-6, 1973.
(Chemical Asbtracts 79, 55489C).
8. Exxon Research and Engineering. Private Communication. 1974.
9. Ingraham, T. R. Thermodynamics of the Thermal Decomposition of
Cupric Sulfate and Cupric Oxysulfate. Trans. Met. AIME 233; 359,
1965.
10. Stern, K. H. and E. L. Weise. High Temperature Properties and
Decomposition of Inorganic Salts. Part 1 - Sulfates. NSRDS-NBS7.
National Bureau of Standards. Washington, D. C. 1966.
173
-------
APPENDIX D
THERMOGRAVIMETRIC STUDIES OF THE SULFATION OF
LIMESTONES AND DOLOMITES
-------
SUMMARY
APPENDIX D
THERMOGRAVIMETRIC STUDIES OF THE SULFATION OF
LIMESTONES AND DOLOMITES
• The sulfation of limestone proceeds to higher calcium
utilization at a fast rate of reaction at atmospheric
pressure and at 1013 kPa (10 atm) if the pressure of
carbon dioxide (CO.) has been high (^ 0.6 x Peq)
during the limestone calcination step (where Peq is
the equilibrium pressure of carbon dioxide over cal-
cium carbonate at the reaction temperature).
• The effect of temperature on sulfur dioxide (S02) re-
tention in fluidized beds of limestone is governed by
the above calcination effect and by the effect of tem-
perature on the reaction of lime with sulfur dioxide
and oxygen not primarily by the oxidation/reduction
cycle in the fluidized bed.
• Pressurized reaction may require the use of half-
calcined dolomite or uncalcined limestone. Most half-
calcined dolomites are excellent sulfur dioxide
sorbents, intermediate in reactivity between dolomite
calcined under low carbon dioxide pressure and dolomite
calcined at high carbon dioxide pressure. Uncalcined
limestone is an extremely poor sorbent for sulfur di-
oxide; calcination and recarbonation does not improve
its sorbent properties.
175
-------
INTRODUCTION
The use of limestone and dolomite to maintain sulfur dioxide
emissions below the levels required by the Environmental Protection
Agency (EPA) during the fluidized bed combustion of coals has been suc-
cessfully demonstrated by several teams of investigators whose work has
been summarized in previous contract reports. Because of the enormous
range of conditions which must be considered in designing fluidized bed
combustors for electric power generation, there is an urgent need for
models of the desulfurization process which will permit simulation of the
effects of changes in operating conditions on the efficiency of sulfur
removal. Such models require data on the kinetics of lime and dolomite
sulfation taken under conditions which isolate the important rate-
influencing variables in the fluidized bed combustion process. The field
of operation of proposed systems — excess oxygen, total pressure, combus-
tion, conditions, and temperature — dictates the form in which the dolo-
mite sorbent will be used, as shown in Figure D-l. (The variation in
carbon dioxide pressure along the bed height means that parts of the bed
may exist in a different stability area, a complication which will be
considered later.) The same stability field applies to limestone, where
half -calcined dolomite and uncalcined limestone exist under the same con-
ditions, as do fully-calcined dolomite and calcined lime. Desulfuriza-
tion may be affected using any of these four forms of calcium sorbent
either singly or in combination.
The thermodynamics of the reactions
CaO + S02 +
CaC03
show a strong driving force for desulfurization by removal of sulfur di-
oxide from the gas stream. The kinetics which govern the rate of sulfur
dioxide removal dictate the size of the fluidized bed reactor (gas and
solid residence times) , the excess calcium required (extent of reaction) ,
176
-------
Curve 656793-A
990
(1814)
950
£ (1742)
Q_
o 910
J1670)
S 870
£ (1598)
I 830
£(1526)
790
(1454)
The Boundaries Shown Have Been Calculated
from Data for the Reaction CaC03 2 C02
Given by Curran2
10% Excess Air
T
O>
00
300% Excess
Air
Q. O
£ ft
i_ Q
o >4_
3 «
to
ce
I
I
I
I
I
(10) ttl) Q2) tt3) (14) (15)
1013 1114 1216 1317 1418 1520
Total Pressure, KPa (atm)
Figure D-l -Temperature and pressure conditions for stability of
the sulfur sorbent as half-calcined or fully -calcined
dolomite at projected combustor outlet gas compositions.
177
-------
particle size of the sorbent, the temperature and pressure of the process,
and the stone pretreatment. The ranges of these variables which have
been considered are shown in Table D-l, which illustrates the scope of
the investigation of kinetics of sulfation. It should be noted that fac-
tors not directly concerned with the desulfurization reaction may limit
the choice of sorbent. These may be listed as:
• The economic availability of limestone or dolomite
at a power plant site
• The attrition behavior of the sorbent in the flui-
dized bed
• The emission of trace elements from the sorbent into
the effluent gas (or turbine feed streams)
• The method chosen for disposal of spent sorbent
• The requirements of a regeneration process.
The work described here is a continuation of investigations de-
scribed in an earlier report, in which thermogravimetric (TG) studies of
the sulfation of calcium-based sorbents showed a close correspondence
with fluidized bed work.
In particular, the work showed that calcium utilization in the
sulfation of dolomite at pressure was highly dependent on the conditions
of carbon dioxide partial pressure which prevail during stone calcina-
tion. This work was extended to consider both the sulfation of limestone
and dolomite at atmospheric pressure, and the sulfation of half-calcined
dolomite at pressure.
Summary of Previous Laboratory Studies
The results of several laboratory studies on the sulfation
3-9
of lime and calcined dolomite are summarized here.
• Sulfation is first order with respect to sulfur dioxide.
• Sulfation is zero order with respect to oxygen.
• The rate of reaction increases with temperature.
Various values of activation energy have been derived,
but there is no unambiguous determination of activation
energy.
178
-------
Table D-l
SORBENT STABILITY IN THE PRESSURIZED
FLUIUIZED BED COMBUSTOR
% Excess Air
Pressure, kPa (atm)
CaO/CaCC-3 Boundary
Temperature
Sorbent
a
Reaction
Relevant TG Data
1
1013
C < 950
F < 1740
Dolomite
1
A, B
10
(10)
> 950
> 1740
Dolomite
Limestone
2
A, B
| 300
1013
< 850
< 1562
Dolomite
1
A
(10)
> 850
> 1562
Dolomite
Limestone
1, 2
A, B
Reaction 1 MgO-CaC03 + S02 + ^ QZ -»• MgO-CaSO, + CO,
Reaction 2 MgO-CaO + S02 + y QZ -»• MgO-CaSO,
b
A - This report.
B - Keairns, D.L., et al. Evaluation of the Fluidized Bed
Combustion Process. Environmental Protection Agency.
Westinghouse Research Laboratories. Pittsburgh, Pa.
Vol. 1. NTIS PB-231 162/9. December 1973.
179
-------
• Reactivity depends upon the pore volume formed during
calcination:
- Small pores give a high initial reaction rate but, if
the particles are large, a low overall capacity results.
- Large por^s give lower rates but increased capacities.
- For small particles, capacity is determined by the
pore volume available for product accumulation.
- Other properties which influence the pore structure
formed on calcination may serve as general indices to
the capacity, in other words, temperature and conditions
of calcination, sodium content of the stone.
- The distribution of sulfur through a sulfated stone
depends on the stone: a coarse limestone shows total
permeation of the stone by sulfate, while Iceland spar
forms a rim of sulfate.
EXPERIMENTAL RESULTS
The Sulfation of Limestone and Dolomite
Experiments were carried out in the following areas:
• The effect of calcination conditions on the sulfation
of limestone and dolomite
• The effect of temperature on limestone sulfation
• The effect of pressure and temperature on limestone
sulfation
• The effect of temperature on the reaction between
sulfur dioxide and calcium oxide.
The materials used in the experiments and their composition are listed in
Table D-2. The thermogravimetric equipment used has been described in an
earlier contract report.
180
-------
Table D-2
SORBENTS USED TO STUDY THE SULFATION REACTION
oo
Sorbent Ca Mg
Ignition
wt. loss Principal impurities
Limestone 1359 38.4 .04 43.4 Silica
Stephens City, Va.
Tymochtee Dolomite 20.5 11.9
Huntsville, Oh.
Glasshouse Dolomite 21.5 12.5
(Dolomite 1337)
Gibsonberg, Oh.
Salamonie Dolomite 21.7 12.9
Portland, Ind.
Canaan Dolomite 22.2 12.8
New Canaan, Conn.
44.4 Silica, pyrites, alumino silicates
47.7 Silica, pyrites, alumino silicates
47.9
46.0 Traces of amphibole
-------
The Effect of Calcination Conditions
Experimental runs aimed at determining the effect of calcina-
tion conditions on the sulfation of limestone 1359 are listed in
Table D-3.
In the first set of experiments, at 1013 kPa (10 atm) pressure,
TG 196-199, the effect of carbon dioxide partial pressure during calcina-
tion was probed. By calcining the limestone in nitrogen, either rapidly
at 900°C (1814°F), or slowly in the temperature range 680 to 870°C
(1256 to 1598°F), a lime was formed which was relatively inert and ceased
to sulfate at a rapid rate after 14 percent utilization of the calcium.
When calcination was retarded, however, by maintaining 101.3 kPa (1 atm)
of carbon dioxide over the solid and heating it to 930°C (1706°F), a more
active lime was produced which yielded 32 and 37 percent calcium utili-
zation in successive experiments.
A second set of runs at atmospheric pressure, TG 215, 216, 220,
221, demonstrated that the activation of the lime is not due to pressur-
ized operation: Figure D-2 shows the effect of calcination history on
the course of the subsequent sulfation reaction.
Since the carbon dioxide partial pressure during calcination
controls the activity of the product lime during sulfation, even at at-
mospheric pressure, tests were run using dolomite 1337 as sorbent at at-
mospheric pressure. It was evident that dolomite could be activated at
atmospheric pressure by calcining it under a high partial pressure of
carbon dioxide. The comparative results for limestone and dolomite are
shown in Figures D-3 and D-4, using some TG results reported earlier.
At the two pressures studied dolomite is a superior sorbent, both on the
basis of the weight of raw sorbent used and on the basis of calcium uti-
lization. The possibility remains, however, of greatly increasing the
capacity of the limestone, but over 80 percent of the calcium in the
dolomite is sulfated, leaving only a small margin for further improvement.
Fluidized Bed Calcine
A set of experiments, TG 262-265, was run on the sulfation of
lime calcined in the 50 mm fluidized bed unit. The calcined sample was
182
-------
Table D-3
LIMESTONE 1359 SULFATION RUNS
TG no.
Pressure,
kPa (atm)
Particle size,
pm
Calcination
Temperature,
°C (°F)
Atmosphere
Time/min.
Sulfation
Temperature,
°C (°F)
Utilization,
% Ca
00
to
196
197
198
199
215
216
220
221
229
230
231
1013 (10.0)
1013 (10.0)
1013 (10.0)
1013 (10.0)
101.3 (1.0)
101.3 (1.0)
101.3 (1.0)
101.3 (1.0)
101.3 (1.0)
101.3 (1.0)
101.3 (1.0)
420-500
420-500
420-500
420-500
420-500
420-500
420-500
420-500
420-500
420-500
420-500
900 (1652)
680-870
(1256-1598)
930 (1706)
930 (1706)
900 (1652)
900 (1652)
871 (1600)
900 (1652)
954 (1749)
843 (1549)
899 (1650)
N2
N2
£,
10% C02/N2
10% C02/N2
N2
30% C02 in N2
55-30% C02 in NZ
60% C02 in N2
15% CO 2 in N2
15% C02 in N2
15% CO. in N.
4.0
^ 12.0
8.0
8.0
2.5
5.0
80.0
30.0
^ 1.2
34.0
3.0
871 (1600)
871 (1600)
871 (1600)
871 (1600)
871 (1600)
871 (1600)
871 (1600)
871 (1600)
954 (1749)
843 (1549)
900 (1652)
14.0
14.0
32.0
^ 37.0
9.0
14.0
34.5
42.0
12.0
14.0
11.0
-------
Curve 65^300-B
.40
.36
"S 3?
>
CTJ
C
.28
-24
.20
.16
.12
.08
.04
hS02+l/202-> CaS04
Limestone 1359
420 - 500 Mm
871° C (1600° F)
101.3 K Pa (lAtm)
5,000ppm S02; 4% 0^, in N2
Pretreatment: Calcination at 900°C (1652°F)
Calcined in N,
o Calcined in 63% CO^N^ TG221
a Calcined in30%C02/N2: TG216
TG215
i i I
j i i
0.1
Time/Minutes
10
100
Figure D-2-Atmospheric pressure sulfation of limestone 1359 - the effect of calcination history
-------
Curve 65^599-B
0.40
Dolomite Calcined
inC02
TG223
Limestone Calcined
in C02 TG 222
Dolomite Calcined in Nitrogen
TG108
Calcined in 60% C(>2 in N2
at 900° C (1652° F)
Calcined in N?
Limestone 1359
Calcined in 60% C02 900°C (1652°F)
Calcined in
Dolomite 1337
Limestone Calcined in N? TG 215
0.34
0.30
0.26
CaO + S02 + i 02 -»
420-500 Mm
5fOOOppmS02
4% 02 in N2
871 °C (1600° F)
Atmospheric Pressure
CaS04
12 16
Time/Minutes
24
28
Figure D-3-Sulfation of limestone and dolomite; effect of calcination conditions
185
-------
Curve 65553 5-6
I I
\ Oo—CaS04
Limestone 1359
Dolomite 1337
Calcined in C02
Calcined in N2
Calcined in
Calcined in N2
100
80
_
60S
"S
5
"5
tO
o
ro
U
40
20
10 12 14 16 18 20
Time/Minutes
22 24 26 28
Figure D-4-Comparison of pressurized sulfation of limestone and dolomite
186
-------
obtained in the test on attrition behavior of limestone 1359 in the
screening process for candidate stones for testing in the Esso Research
Centre, Abingdon, England (Esso) oil gasification pilot plant. Since the
sample was allowed to calcine in a stream of nitrogen in the temperature
range 650 to 750°C (1202 to 1382°F), the stone produced have been rela-
tively inactive. The extent of sulfation achieved in TG experiments from
101.3 to 1013 kPa (1 to 10 atm) lay in the range 13.5 to 17.5 percent,as
shown in Table D-4. These experiments confirmed that calcination at low
partial pressures of carbon dioxide do produce an inactive stone. The experi-
mental schedule on the 50 mm unit has not yet permitted examination of
the case where calcination takes place under a relatively high partial
pressure of carbon dioxide.
Table D-4
TG SULFATION OF LIMESTONE 1359 CALCINED
IN THE 50 MM FLUIDIZED BED3
TG no.
P/(1.03 x 105N/m2)
% CaO utilization
After 5 min | After 1 hr
263
264
265
266
1.0
10.0
10.0
5.0
11.7
13.9
11.9
10.7
17.5
17.8
15.1
13.5
alnitial particle size (1,000-420) urn; calcined in nitrogen
650 to 750°C (1202 to 1382CF). Sulfation at 5,000 ppm S02,
4 percent 02 in N2 at 870°C (1598°F).
Temperature Effect
Previous studies on the sulfation of limestone and dolomite
have demonstrated that the effect of temperature on the reaction is a
complex phenomenon. Activation energies obtained at low calcium uti-
lization show values in the range of 5 to 15 kcal-mole, 20 to '62 kJ, in-
dicative of a mixture of mass transport and chemical reaction control.
In fluidized beds, however, where sulfur dioxide removal is normally
187
-------
carried out at 30 to 40 percent utilization, a maximum in the extent of
sulfur dioxide removal at a fixed calcium utilization is observed at a
temperature of about 843°C (1550°F) at atmospheric pressure, implying
that the rate of the overall reaction decreases above 843°C (1550°F).
In three experiments, TG 229, 230, and 231 at 843, 899, and
954°C (1550, 1650, and 1749°F), limestone was calcined and sulfated. The
stones were inactive, leading to utilization in the range 11 to
14 percent. The results, illustrated in Figure D-5, showed that maximum
reactivity was observed at 843°C (1550°F): the calcium utilization, how-
ever, was not typical of fluidized bed conditions. It was decided,
therefore, to carry out sulfation runs at different temperatures on sam-
ples which had all been calcined at the same conditions of temperature
and carbon dioxide partial pressure.
Calcination of the samples was effected at 930°C (1706°F) in
60 percent carbon dioxide in nitrogen, and the samples were then sulfated
at temperatures from 750 to 950°C (1382 to 1742°F) at 50°C (122°F) inter-
vals, in random order. The results are shown in Figure D-6 and indicate
that although the initial rates were scarcely distinguishable, the course
of reaction was different for each temperature after 20 percent utiliza-
tion of the calcium. The importance of these differences becomes clear
when the extent of reaction after a fixed time interval, one hour, is
plotted as a function of sulfating temperature, Figure D-7. The utiliza-
tion of the stone peaks at a maximum value of 43 percent at about 860°C
(1580°F), reproducing the sulfur dioxide retention efficiency pattern ob-
served in fluidized bed results. The data derived for the rate of reac-
tion as a function of calcium utilization in these experiments are used in
Appendix E to project fluidized bed sulfur retention.
The rate of sulfation derived here may be compared with that
noted in experiments at TVA. ^ Using limestone 1359, 707 to 595 microns,
soft-calcined at 900°C (1652°F) in nitrogen, these experiments were
carried out by sulfating^SO mg of calcine in 4 percent sulfur dioxide,
4 percent oxygen isothermally at 930°C (1706°F). The maximum rate of
sulfation noted was 3.99 mg/minute, which may be compared with rates of
2.64 and 3.08 mg/minute noted in the .Westinghouse experiments. When
188
-------
Curve 65^597-B
CO
VO
843° C (1550° F)
o 899° C (1650° F)
n 954° C (1750° F)
420-500 Mm Limestone 1359
5,OOOppmS02
4% €2 in N2
Atmospheric Pressure
Calcination inl5%C02/N2
at Sulfation Temperature
5 6
Time/ Minutes
Figure D-5 -Effect of temperature on sulfation of limestona 1359
-------
Curve
60
50
•I 40
5
^
0)
?B
-------
Curve 655^57-A
0.50
^ 0.45
f 0.40
c
I 0.35
•g 0.30
0.25
••g 0.20
C
o
2
o.is
% 0.10
o
0.05
CaO + S02 +1/ 2 02
420 - 500 urn
Limestone 1359
1.01 x 105 N/m2
5,000ppm S02
CaSCty
4% 02 in N2
20 mg Stone Calcined at 900°C (1652°F) in 60% C02 in N2
Stone Brought to 900°C (1652°R in Pure C02 at 10 k/Min
TG 258, 259, 260, 261, 262
kAx
1
1
700
(1292)
750
(1382)
800 . ,850 900 950 1000
(1472) (1562) (1652) (1742) (1832)
Temperature °C(°F)
Figure D-7-The effect of temperature on CaO utilization in sulfation
191
-------
these races are adjusted to unit gas concentration, the rate noted here
is almost six times as rapid as that recorded by TVA. Probably the chief
reason for this difference is the small size of the Uestinghouse sample
(10 mg of calcine). Previous work at Westinghouse has illustrated the
effect of sample size on the initial rate of sulfation.
The rate of sulfation was determined for each of the five runs
at 10 percent utilization, and an apparent activation energy, Ea, of
5 kcal. mole derived using the Arrhenius equation. Such low values of
Ea are typical for reactions in regimes where the rate is predominantly
mass-transport controlled.
The data obtained in experiments TG 258 through 262 were ob-
tained on samples calcined after heating to the temperature of calcina-
tion at 10°C (18°F) per minute. Two runs at a different heating rate,
20°C (36°F) per minute, carried out before and after the above-mentioned
runs,gave a reactivity in sulfation different from the 10°C (18°F) per
minute experiments, as Figure D-8 shows. The utilization after one hour
in each of these experiments was 48 percent calcium, as opposed to
40 percent in the 10°C (18°F) per minute run. The initial rates of sul-
fation in runs on samples heated at 10°C/min (l8°F/min) and 20°C/min
( 36°F/min) were identical, as Figure D-9 shows: possibly it is during the
later phase of sulfation that the structure developed in the stone affects
the reaction rate. Application of the data to fluidized bed modeling,
however, as described in Appendix £, showed that the latter data, taken at
the higher heating rate and indicating higher reactivity, more accurately
predicted fluidized bed performance. This suggests that the data in
TG 262 was anomalous and that the sample heat-up rate is not important at
the level of 10 to 20°C (18 to 36°F)/minute.
The Effect of Temperature on the Sulfation of Limestone 1359
at Pressure
A set of experiments parallel to those described above was
carried out at a pressure of 1013 kPa (10 atm). The results obtained
were far less clear-cut, principally because of the difficulty in repro-
ducing controlled conditions of calcination. At the higher volumetric
192
-------
10.0
Curve 655^55-A
5.0
4.0
3.0
I 2.0
- o
o>
o
CO
o>
c
OJ
o
O
(D
O
1.0
0.5
0.2
III II
CaO + S02 + I/ 2 02 -*. CaS04
Limestone 1359 420-500 Mm
800 °C 1.03xl05N/m2
5,000ppm S02; 4%02 in N2
a Precalcined at900°C (1652°F) in60% CO
o* AHeated to 900°C (1652°F) in C02 at
> 20K/Min
O
^ TG 257
o TG268
1
0 10 20 30 40
CaO Percent Utilization
50
60
Figure D -8-Comparison of rates for identical sulf-
ation runs
193
-------
v£>
§
"ro
13
c
^g
"u
0)
•o
s
O
ro
O
.30
.28
.26
.24
. 22
.20
.18
.16
.14
.12
.10
.08
.06
.04
.02
Curve 655^56-A
I [III
CaO + S02 + I/ 2 02 —- CaSCfy
Limestone 1359
420 - 500 u m
800 °C (1472° F)
1.03xin4|\l/m2
Calcined 900°C (1652°F) in 60% C02 in N2
Sulfated in
5,000 ppm S02, 4% 02 in N2
T
0
TG262 (10k/Min from 20° C)
TG257 ( 20k/Min from 20° C)
4 5 6
Time/Minutes
8
10
Figure D-9-Comparison of repeat experiments - fast phase of sulfation
-------
flow rates reproduction of the temperature and gas composition was poor,
as evidenced by the variation in calcination time recorded.
The extent of sulfation recorded in these experiments after
30 minutes is shown in Figure D-10 as a function of temperature. The
extent of sulfation increases up to 900°C (1652°F). At 950°F (1742°F) it
has decreased; but, as shown, significant differences in the extent of
the decrease were noted. The conclusion that may be drawn is that cal-
cined limestone should be an effective sorbent at pressures up to a tem-
perature between 900 and 950°C (1652 and 1742°F). Reexamination of the
sulfation kinetics in the region 900 to 950°C (1652 to 1742°F) at pres-
sure is required.
The Sulfation of Dolomite
The activation process of calcining under a high partial pres-
sure of carbon dioxide yields a dolomite with greater sulfur dioxide
sorption capacity at a fast rate of reaction. Since the data reported
earlier referred to 420 to 500 ym particles, the effect of an increase in
particle size to 1,000 im particles was studied. In TG 201 1200/
1000 ym particles of dolomite 1337 were calcined in 10 percent carbon
dioxide in nitrogen at 750°C (1382°F) to calcine the magnesium carbonate
fraction. The sample was then heated to 925°C (1695°F) and the carbon
dioxide partial pressure dropped to 81.04 kPa (0.8 atm). The sample was
calcined and sulfated at 871°C (1600°F) in 0.5 percent sulfur dioxide in
4 percent oxygen in nitrogen. A comparison sample, TG 202, was similarly
sulfated after calcination under nitrogen by heating to 880°C (1616°F).
Comparison of the rates of reaction showed that the calcination treatment
in carbon dioxide had increased the stone capacity. If 2 percent calcium/
minute is taken as the criterion for an acceptably fast rate of reaction
with sulfur dioxide under the given experimental conditions, then the
rate of reaction fell below this value at 50 percent utilization in
TG 201 and 35 percent utilization in TG 202. This can be regarded as an
improvement of the calcium/sulfur molar feed ratio from 2.6 to 1.8 at a
mean particle size of 1100 ytn. The activation process for increasing the
utilization' of dolomite can be used with large particles of sorbent.
195
-------
0.50
J2 0.45
3
2 0.40
£
| 0.35
i 0.30
•§
S 0.25
| 0.20
2 0.15
.2 0.10
u
(D
O
0.05
0.
— CaS0
420 -500um
Limestone 1359
l.Olx 106 N/m2
5,000ppm S02
20 mg Stone Calcined at 900°C (1652°F) in 60% C02 in NZ
Stone Brought to 900°C (1652°F) in 10% C02 at 10 K/min
I
) 750
(1292) (1382)
850 900
(1472) (1562) (1652)
Temperature, °C (°F)
950
(1742)
1000
(1832)
Figure D-10-The effect of temperature on CaO utilization in sulfation
at pressure
196
-------
Attempted Sulfation of Limestone 1359 (Uncalcined)
A 450-to-500 urn sample of limestone 1359 was heated to
850°C (1562°F) under carbon dioxide and sulfated in 0.18 percent sulfur
dioxide, 4 percent oxygen, 60 percent carbon dioxide in nitrogen. The
total weight gain noted corresponded to sulfation of 3.4 percent of the
calcium carbonate content of the stone in 100 minutes. The thermody-
namic equilibrium for the reaction
CaC03 + S02 + | 02 -> CaS04 + C02
lies far to the right: at 816°C (1500°F) and 871°C (1600°F) the equili-
brium pressure of sulfur dioxide is less than 1 ppb and 1 ppb in 4
percent oxygen and 10 percent carbon dioxide at 1013 kPa (10 atm). The
limitation on desulfurization is kinetic.
In order to test the possibility of activating limestone as a
sorbent, a sample was calcined, recarbonated, and exposed to sulfur di-
oxide and oxygen. A slight weight increase occurred at a slow rate indi-
cating that residual calcium oxide which had not recarbonated was reacting
to form the sulfate.
Sulfation of Half-Calcined Dolomite
As Figure D-l illustrates, half-calcined dolomite is stable
over much of the operating range of the low excess-air fluidized bed corn-
bus tor. The stability line shown in the figure is for the exit gas
carbon dioxide composition; at the bed entrance, where the fluidized gas
is mainly air, some calcination will take place, yielding fully-calcined
dolomite; and sulfur dioxide and carbon dioxide will compete for the
calcium oxide generated. Previous work at Westinghouse has shown that
half-calcined dolomite is a reactive sorbent for sulfur dioxide removal. '
The runs carried out, however, used larger samples than have been found
optimal in subsequent studies, and the kinetics of the reaction could not
be assessed relative to the reactivity of calcined dolomite.
Three experiments were carried out on the sulfation of half-
197
-------
calcined dolomite 1337. The first two experiments (280, 281) were tests
with 1680 to 2000 ym diameter particles. (It should be noted that only
two particles are used in a run with this size stone.)
In the first experiment the temperature was cycled between 850
and 690°C (1562 and 1274°F) as the reaction proceeded. After 23 minutes
33 percent utilization was noted, during which time the sample had been
cooled to 690°C (1274°F) and heated back to 820°C (1508°F). In the fol-
lowing experiment isothermal sulfation at 800°C (1472°F) gave 50 percent
utilization in 23 minutes. The decline in reaction was not as abrupt as
has been universally noted with calcined dolomite after two hours of re-
action; 87 percent utilization was noted in the isothermal experiment.
A third experiment on (420 to 500 pm) 1337 half-calcined dolo-
mite showed greater initial reactivity—61 percent utilization in
23 minutes. The subsequent activity, however, was slow, yielding
68 percent utilization after two hours. (See Figure D-ll.)
The reactivity of half-calcined dolomite appears to be inter-
mediate between nitrogen-calcined dolomite and dolomite which has been
calcined under a high partial pressure of carbon dioxide relative to the
equilibrium pressure. In addition, the slowing down of reaction is less
abrupt than with calcined dolomite or limestone. This latter fact means
that increased gas residence times in the fluidized bed should improve
(lower) the calcium/sulfur molar ratios required for any given sulfur re-
moval criterion.
Discussion of the Temperature Effect
The phenomenon of an optimum temperature for sulfur dioxide ab-
sorption has been discussed by Moss. He pointed out that when a fresh
bed of limestone was used, absorption efficiency was 90 to 95 percent in
the Esso experiments, and the efficiency was unaffected by temperature in
the range recorded (presumably around 800 to 925°C/1472 to 16978F).
This phase of reaction corresponds to the initial phase of reaction in
the Westinghouse TG data and other kinetic curves: in this range of re-
action (5 to 10 percent calcium oxide reacted) a slight increase in reac-
tion rate is noted corresponding to an activation energy of 5 to 15 kcal.
198
-------
TG 281
Dolomite 1337
1680-2000 ym
Half-Calcined
0.39% S02/10% 02/12% S02
In Nitrogen
800°C (1472°F) 1013 kPa, (10 Atm)
"20 4oto fo inn120 i4n
time/minutes
Figure D-ll. Sulfation of half-calcined dolomite
199
-------
mole —insufficient to cause any appreciable change in the extent of
sulfur dioxide absorption as noted in the fluidized bed experiments.
Moss has noted, however, that when 31 percent of the lime has
reacted, absorption efficiency is markedly temperature-dependent with a
maximum at 865°C (1589°F). He argues that since this optimum temperature
"only becomes apparent after the lime is appreciably reacted, it follows
that chemical reaction rate cannot be the controlling factor, it would
operate on used and fresh lime alike." This argument is excessively
universal—while it could be true it need not be true, and the overall
evidence of the TG work reported here weighs against it. The primary
flaw in the argument lies in the inherent assumption that if the chemical
rate is controlling at 32 percent calcium oxide utilization, then it must
also control at lower utilization. A typical TG rate curve, as in
Figure D-12, may be conveniently considered as a three-stage process.
Initially, region A, the rate of reaction, is controlled by mass transfer
to the surface of the solid from the gas stream. The evidence that this
is true is provided by
• The insensitivity of the rate to temperature (apparent
Ea at 10 percent sulfation of calcium oxide is
5.3 kcal. mole'1)
• The insensitivity of the rate to source of calcium
oxide—both limestone and dolomite of different cal-
cium content, 40 to 18 percent, have similar rates of
reaction.
• The dependence of initial rate on pressure, is accurately
predicted by consideration of the mass transfer rate
to ideal spherical particles under TG experimental
conditions
The final phase of reaction, C, is controlled by the rate of dif-
fusion of the reactant or reactants across the solid product accumulated
in the pores of the solid. This interpretation is supported by
• The insensitivity to temperature of final rate of re-
action
200
-------
4)
4J
a
o
•H
4J
o
(d
o>
05
TG 106
420/500 ym
Calcined Dolomite
1337
871°C/101 kPa
0.5% S02/4% 02
.2
rs
Calcium Fraction Reacted
Figure D-12a:
The decline in rate of
reaction (calcium fraction
reacting per minute) as calcined
dolomite sulfates
4J
cd
H
4J
CO
0)
4J
CO
M
s
CO
<0
0)
15
10
71 2 4
Figure D-12b: The effect of pressure on
sulfation of calcined dolomite
1337 (420 ym; 0.5% S02/4% 025
871°C).
201
-------
• The insensitivity to system pressure of the final rate
of reaction. Since the diffusion coefficient, D, varies
inversely with P (the system pressure), the effect of
increasing the partial pressure of sulfur dioxide by
pressurized operation is negated.
• The successful prediction of the rate of reaction
to form calcium carbonate when carbon dioxide is sub-
stituted for sulfur dioxide as the reactant because of
the different diffusivities of carbon dioxide and
sulfur dioxide.
The intermediate range I, in which the rate falls through seve-
ral orders of magnitude is a region of mixed kinetic control—and it is
difficult to devise tests which will separate the controlling factors.
The mechanism thought to apply will be described, and the evidence will
be discussed qualitatively.
For the initial stages of reaction, the flow of sulfur dioxide
gas through the external surface of the solid is immediately reacted with
calcium oxide—so that an effective sulfur dioxide sink exists, and flow
is governed solely by the diffusional resistance at the surface. As the
walls of the pores in the calcium oxide react, however, the rate at which
sulfur dioxide is fixed decreases, and the pressure of sulfur dioxide in
the pores increases. The calcium oxide available for direct reaction as
the gas collides with the solid has decreased; the resistance to reaction
rises; and as it becomes rate controlling, the flow of sulfur dioxide
into the solid is controlled by
• The constant boundary diffusional resistance at the
surface
• The in-series resistance within the stone, which in-
creases as reaction proceeds.
As reaction inside the pores gets slower, owing to the depletion of solid
reactant on the internal surface, the available supply of sulfur dioxide in-
creases until the pressure in the pores equals the partial pressure of
sulfur dioxide outside the stone. This effect should be noticeable the
202
-------
slower the overall rate, while there is still unreacted solid which can
be reached by the gas without diffusion across the solid product layer
on the pore wall (which, of course, makes a continuous minor contribution
to the overall reaction rate). The first critical test can now be ap-
plied to this model. If the mechanism described is correct, the effect
of increasing total pressure at constant sulfur dioxide concentration on
the rate of reaction expressed as the ratio of the overall rates should
increase from about 2 (the ratio for mass transfer control) to about 10
(the ratio for chemical control) and then fall to 1 (as diffusion through
the product layer takes over). Figures D-12 a and b show that the effect
is observed for kinetic experiments at 101.3 and 1013 kPa (1 and 10 atm)
total system pressure. The comparison was made for calcium oxide
sorbents calcined under the same conditions of carbon dioxide partial
pressure to ensure as far as possible that the same pore structure exists
in both samples. Each of the comparisons made demonstrated the effect
noted.
The second indicator of a change from boundary layer mass trans-
fer to chemical control in region D lies in the effect of temperature on
the rate of reaction in this region. In TG 303 the rate of reaction was
determined as a function of utilization in the range 33 to 44 percent and
found to follow closely the relation:
log rate = -k (utilization) + constant.
The temperature of the sample was then raised to 950°C (1742°F) and cooled
again. The rate data (uncorrected for the effect of utilization) showed
an activation energy of 60.7 kcal. mole . Correction for the increasing
stone utilization, however, yielded the value 51 kcal. mole over the
range 817 to 950°C (1503 to 1742°F). The only value found in the lit-
erature which does not reflect diffusional limitations is 57 kcal.
mole obtained by Jungten and van Heek, in the lower temperature
range.
203
-------
In order to probe further the mechanism of lime sulfation,
experiments were carried out on the reaction
I 1 3
CaO + S02 -*• [CaSOj] * ^ CaS + ^ CaS04
which may be the first step in the lime sulfation reaction.
Experiments on the Reaction Between Sulfur Dioxide and Calcined Lime
The equilibrium between calcium oxide, sulfur dioxide, and
calcium sulfite was briefly examined in order to explore the reaction
between calcined limestone and sulfur dioxide and to find the source
of the decreasing utilization as temperature is increased.
The ability of the TG apparatus to determine such an equili-
brium was checked in a preliminary experiment by determining one
equilibrium point for the reaction
CaC03 2 CaO + C02
The solid used was obtained by calcining calcium hydroxide (Ca[OH]_)
reagent. Pure carbon dioxide was passed over the sample and the
temperature raised and lowered. The TG balance output was connected
so that weight was recorded as a function of temperature, and the
sensitivity of the system adjusted so that a one percent change in the
solid weight yielded a 25 .4 mm (1 in) deflection on che chart paper.
The tracing can be read to within 1.3 mm (1/20 in). The result
obtained, averaged over four cycles of decomposition and recarbonation,
is compared with some literature values.
POA , kPa (atm) T°C (°F)
C02 _
TG© 96.8 (0.956) 892.1 (1638)
Hill8 96.8 (0.956) 901.0 (1654)
Curran2 et al. 96.8 (0.956) 891.3 (1636)
Chemical Rubber Company9 96.8 (0.956) 895.0 (1643)
204
-------
There was a 7°C (13°F) difference between the recarbonation
and regeneration temperatures, but this precision and accuracy should
be adequate.
A sample of limestone 1359 was then tested. The first
calcination temperature under 96.6 kPa (0.954 atm) carbon dioxide was
obsured by the decomposition of the magnesium carbonate(MgCO-) content
of the stone. The recarbonation and subsequent cycling gave equili-
brium temperatures of 896.5, 892.0, and 895.5°C (1645, 1637.6, and
1643.9°F), compared with literature values of 900.8°C (1653.3°F)
(Hill)9and 890.5 (1634.9°F) (Curran).2
The sample was heated and cooled in sulfur dioxide (.005
atm) to form calcium sulfite, reheated in nitrogen to drive off
residual carbon dioxide, and cooled in sulfur dioxide. The sample was
then cycled through weight gain and weight loss, and the average
equilibrium temperature was 930.7°C + 25°C (1707.1°F ± 77°F). The
temperature span was much greater here than in the case of the calcium
carbonate decomposition.
The sample was heated and cooled further, and an interesting
anomaly was apparent. ' The sample was heated in 0.005 atm sulfur
dioxide until it lost weight at 938°C (1720.4°F) and on cooling down
it gained weight at 968°C (1774.4°F). Reheating caused a weight loss
to begin at 956°C (1752.8°F) and further heating caused a weight gain
at 1013°C (1855.4°F). On cooling, the sample gained weight continuously.
Reheating caused a weight loss at 900°C (1652°F) followed by a reversal
to weight gain at 979°C (1794.4°F).
In summary, there was an equilibrium pressure of sulfur dioxide
calcium sulfite (CaSO_)/calcium oxide at 900 to 950°C (1652 to 1742°F) and a
second equilibrium point at 979 to 1013°C (1794.4 to 1855.'4°F). This
would require that the solid phase pass through a transformation which
lowered its free-energy by about 2800 cal. mole . Possibly, liquid
formation in the system calcium sulfide(CaS)/calcium sulfate caused by
disproportionation of the calcium sulfite would suffice to stabilize
the calcium sulfite or calcium sulfite/calcium sulfate mixture.
205
-------
Apart from the possible use of this method to determine the
liquid eutectic in the calcium sulfide/calcium sulfate, the apparent
equilibrium temperatures observed are surprising. Since calcium sulfite
is thermodynamically unstable with respect to disproportionation to
calcium sulfide and calcium sulfate, and the latter react to form sulfur
dioxide according to the equation
•| CaS + | CaS04 -»• CaO + S02,
the equilibrium pressure of sulfur dioxide should be that noted by
2
Curran. At the partial pressure of sulfur dioxide used, the equilibrium
temperature should be MJ40°C (1544°F) according to Curran1s data for
calcium sulfide/calcium sulfate, and Esso obtained the same equilibrium
pressures as Curran starting with calcium oxide and sulfur dioxide.
The Radian data calculated using estimated values for the thermodynamic
properties of calcium sulfite give 760°C (1400°F) as the temperature
for equilibrium by sulfur dioxide over calcium oxide at .005 atm sulfur
dioxide. It was thought necessary to repeat the experiments with pure
calcium oxide rather than with lime derived from limestone containing
impurities such as magnesium oxide (MgO) or ferrous oxide (Fe.Q ).TG
experiments were carried out in which a flow of sulfur dioxide (0.5
percent) in nitrogen over calcined calcium carbonate caused a weight
charge, as the temperature was varied, at a constant heating rate.
• The rate of sulfur dioxide uptake increases with increasing
temperature, up to 815°C (1499°F) at 0.5 percent sulfur
dioxide. The rate decreases on heating pajt 815°C (1499°F).
• At above 890"C (1634°F) the product begins to decompose,
as the solid is heated. Complete reversal of the weight
gain was never noted.
• No substantial oxidation of calcium sulfide takes place
during the experiment, since reaction ceased entirely
when 100.8 percent of stoichiometric calcium sulfite (as
determined by weight gain) has formed.
206
-------
• The second equilibrium point was noted at 976°C (1789°F),
in other words, readsorption began again on heating
through this temperature.
• The samples may have suffered surface melting during the
experiments. Figure D-13 shows the weight changes
observed during a typical heating run.
It can be concluded from these experiments that
• The rate of reaction between sulfur dioxide and calcium
oxide reaches a maximum between 800 and 850°C (1472 and
1562°F) closely corresponding to the optimum temperature
for lime sulfation.
• Desorption of sulfur dioxide occurs at 900°C (1652°F),
so that formation of calcium sulfate will be a competitive
process between oxidation and desorption.
• Complex equilibria involving solid solutions or liquid
eutectics occur at temperatures close to 950°C.
• Further work, including differential thermal analysis
and microscopic examination of sample structure should
be carried out to characterize the effects observed at
950°C (1742°F).
Kinetic data, representative of the experiments represented
in Table D-5, are shown in Tables D-6 through D-15.
207
-------
Curve 65702U-A
N>
O
00
11
10
9
8
"5»
•£ 7
c
'Jo
0 6
m
4
3
2
I I I \ I
_StoichjpmetricJ/Velght Gain for CaSOj Formation
T
10.3mg. CaOfrom CaCC^ Reagent
S02: 0.5%inN2
Sample Heated at 10k min"1
Atmospheric Pressure
1
I
I
I
700 750 800 850 900 950 1000 1050
(1292) (1382) (1472) (1562) (1652) (1742) (1832) (1922)
Temperature, °C (°F)
Figure D-13-The effect of temperature on the course of the reaction between S02 and CaO
-------
Table D-5
TG RUNS ON SULFATION OF CALCINED LD1ESTONE, CALCINED DOLOMITE,
AND HALF-CALCINED DOLOMITE
o
VO
TG no.
196
197
198
199
200
201
202
204
Stone
Pretreatment
Calcination at 871°C
L1359 (1000°F) retarded by
420-500 micron C02
Calcination in N2 at
L1359 348C/min to 871eC
420-500 micron (1600°F)
Calcined in 5% C02
L1359 at 930°C (1706°F)
LI 359
Slow MgC03 calcination
at 730°C(1346°F) slow
D1337 CaC03 calcination at
420-500 micron 930°C (1706°F)
D1337
1200-1000 micron
D1337
1200-1000 micron
it
Calcined in N2
Calcined in
D1337 two stages using
1200-1000 micron CC>2 suppression
Sulfation
P.kPa(atm)
103(10)
103(10)
103(10)
103(10)
103(10)
103(10)
103(10)
103(10)
T,0C(°F)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
Result
molar
%S04/mins
12.7/71
12.5/60
29/60
41.8/220
86/10
69/60
62/60
69/60
Comment or puroose
Calcination effect
Calcination effect
Calcination effect
Calcination effect
Slow calcium calcination
62% utilization in
N*2 calcination has
one hour-
effect
65% utilization in one hour-
max. temp, in calcination wa
1100°C (2012°F)
-------
Table D-5 (Continued)
TG no.
210
' 214
215
216
220
NJ
° 221
222
-
226
229
230
231
257
Stone
Tymochtee
D1337
LI 359
LI 359
L1359
LI 359
D1337
D1337
2000-1680 micron
Limestone 1359
420-500 micron
it
L1359
Sulfation Result
Pretreatment P.kPa(atm) T,°C(°F) %S04/mins Comments or purpose
80/63 Leaching effect
Calcined in two stages-
CaC03 at 900°C(1652eF) 54/114 Calcination effect
Heated to 900°C(1652°F)- 871
calcined in N2 10 (1.0) (1600) 9/114 Calcination effect
Heated to 900eC(1652°F)
in C02-calcined in 30% 871
C02/N2 10 (1.0) (1600) 14/72 Calcination effect
Heated to 871°C
(1600°F)-calcined in , 871 ,. .., _ . . . ,, _
30% C02/N2 102(1.0) (1600) 34/46 Calcination effect
Heated to 901°C(1654°F) Calcination effect at
in C02-calcined in 30% , 871 1 atm
C02/N2 10 (1.0) (1600) 42/52
Calcined in two _ 871 High utilization at
stages ' 10 (1.0) (1600) 86/60 atmospheric pressure
Tvo-stage calcina- _ 871
tion 10J(10) (1600) 86.9/60
Fast calcination _ 954
in nitrogen 10 (1.0) (1749) 13/165
0 843
2
10 (1.0) (1549) 14
9 899
10 (1.0) (1650) 11/120
Procedure A
Heated to 900°C
(1652°F) in C02 at
20°C/rain calc. in 800
60% C02 10Z(1.0) (1472)
-------
Table D-5 (Continued)
TG no.
258
259
260
261
262
268
263
264
265
266
267
279
280
Su
Stone Pre treatment P,kPa(a
Ifation Result
cm) T,°C(°F) %S06/mins Comments or purpose
Heated to 900°C(1652°F)
in C02 at 10°C/min calc. „ 950 Effect of temperature
L1359 in 60% O>2 10 (1.0) (1742) 34/60 on reaction
, 750
L1359 B 10 (1.0) (1382) 32/60 "
- 900
L1359 B 10 (1.0) (1652) 39/60
, ' 850
L1359 B 10 (1.0) (1562) 43.2/60
-, 800
L1359 B 10 (1.0) (1472) 39.7/60
, 871
L1359 A 10 (1.0) (1600)
L1359 Fluid bed calcined ,
420-500 micron in nitrogen 10 (10)
LI 359 ,
420-500 micron " 10 (10)
L1359 ,
420-500 micron " 10 (10)
L1359
420-500 micron " 507(5)
D1337 Half-calcined in ,
2000-1680 micron l.S.atm C02 10 (10)
L1359 103(10)
D1337
2000-1680 micron Half -calcination 10 (10)
870
(1598) 17.5/60 Low calcium utilization
870 System pressure does not
(1598) 17.8/60 affect utilization significantly
870
(1598) 15.1/60
870
(1598) 13.5/60
893-903 Rate does not decline
(1639-1657) 40/60 abruptly
5.6/100
900-700-
900
(1650-1382-
1652) Temperature scan
-------
Table D-5 (Continued)
TG no.
281
282
283
298
299
ISJ
G 300
301
302
303
304
305
320
330
331
Su
Stone Fretreatment P,kPa(a
D1337 -
2000-1680 micron Half -calcination 10 (10)
D1337 _
420-500 micron Half-calcination 10 (10)
Tymochtee
L1359 103(10)
L1359 103(10)
L1359 103(10)
L1359 103(10)'
L1359 103(10)
L1359 103(10)
L1359 103(10)
L1359 103(10)
Lfation Result
tm) T,°C(°F) %SO£/mins Comments or purpose
800
(1472) 87/120
800
(1472) 68
850 Temperature effect at
(1562) pressure
Temperature1 effect at
pressure (utilization
950 low run repeated in
(1742) 23/ TG304)
750 Temperature effect at
(1382) pressure
900
(1652)
850
(1562) 34/37
800
(1472) "
950
(1742) "
37/60 "
Powdered CaC03 Calcined in ^
(Fisher) nitrogen .10 (1.0) scan
Limestone
N.B.S. 2.0 wt% 2
standard Nad added 10 (1.0) scan
ii it
scan 750-
1000(1382-1832)
-------
Table D-6
TG DATA FOR RUN 2023
(DOL 1337 1200/1000 pm; SULFATION)
% REACTED
5.28
9.72
13.56
17.28
20.52
21.28
25.80
27.84
29.76
31.80
36.48
41.52
47.04
52.92
56.64
59.52
61.56
63.84
66.12
68.16
69.12
69.48
TIME/MIN.
0.38
0.63
0.88
1,
1.
1,
1,
,13
,38
,63
,88
2.13
2.38
2.63
38
38
,88
3,
4,
5,
8.38
10.88
14. 38
18.38
25.88
40.88
55.88
70.38
80.66
% /MIN.
13.62
17.76
15.36
14.88
12.96
11.04
10.08
8.16
7.68
8.16
6.24
5.04
3.68
2.35
1.48
0.82
0.51
0.30
0.15
0.13
0.06
0.03
7, REACTF.n
2.64
7.50
11.64
15.42
18.90
21.90
24.54
26.82
?8.80
30.78
34.14
39.00
44.28
49.98
54.78
58.08
60.54
62.70
64.98
67.14
68.64
69.30
UNIT RATE
27252.91
35521.70
30721.47
2976].&2
25921.24
22^81.05
20160.96
16320.78
15360.73
16320.78
12480.59
10080.48
7360.35
4704.22
2976.14
1645.7<»
1020.04
608.0?
304.01
272.01
128.00
73.66
Conditions on Table D-5 .
-------
Table D-7
TG DATA FOR RUN 2153
(LIMESTONE 1359; SULFATION)
ro
% REACTED
2.00
2.73
3.33
3.73
4.20
,06
,73
6.00
6.40
,00
,33
.50
7.67
8.20
8.33
8.50
5,
5,
7,
7,
7
TIME/TUN.
0.41
0.66
0.91
1.16
1.41
2.41
3.91
4.91
7.A1
14.91
22.41
29.91
37.41
74.36
94.36
114.36
%/MIH.
4.88
2.93
2.40
1.60
1.86
0.86
0.44
0.26
0.16
0.08
0.04
0.02
0.02
0.01
0.00
0.00
% REACTED
1.00
2.36
3.03
3.53
3,°6
4.63
5,
5,
,40
,86
6.20
6.70
7.16
7.41
7.5R
7.13
8.27
8.42
UNIT RATE
9760.42
5869.26
4802.12
3201.41
3734.98
1734.10
889.28
533.56
320.14
160.07
88.02
44. A6
44.46
28.RR
13.33
16.67
Conditions on Table D-5.
-------
Table D-8
TG DATA FOR RUN 221a
(LIMESTONE 1359 420/500 ym; SULFATION)
ro
t-'
Ui
7, REACTED
1.44
2.75
3.93
5.11
6.15
9.43
12.19
14.74
16.97
19.14
20.97
22.35
24.71
27.33
29.82
30.80
33.42
35.13
36.64
37.75
33.41
39.72
40.70
41.29
41.49
TIME/Hin,
0.27
0.52
0.77
1.02
1.27
2.02
2.77
3.52
4.27
5.02
5.77
6.52
7.52
9.02
10.52
11.52
14.0?.
16.52
19.02
21.52
24.02
29.02
36.5?
44.02
51.97
%/MIN.
5.24
5.24
4.71
4.71
4.98
4.10
3.67
3.40
2.97
2.88
2.44
1.83
2.35
1.74
1.66
0.98
1.04
0.68
0.60
0.44
0.26
0.26
0.13
0.07
0.02
% REACTED
0.72
2.09
3.34
4.52
5.73
7.89
10.81
13.47
15.86
18.05
20.05
21.66
23.53
26.02
28.57
30.31
32.11
34.28
35.«8
37.1?
38.OR
30.06
40.21
41.00
41.39
UNIT RATE
10^87.68
10487.68
9438.91
9438.91
9963.30
8215.35
7341.37
6816.99
5943.02
5768.22
4894.25
3670.68
4719.45
3495.89
3321.10
1966.44
2097.53
1363.TO
1206.08
891.45
524.38
524.3P
262.10
157.31
4 Q . A 7
Conditions on Table D-5.
-------
Table D-9
TG DATA FOR RUN 257a
(LIMESTONE 1359 SULFATION)
% REACTED
2.30
3.31
6.30
8.14
11.43
13.01
14.58
15.90
18.59
21.02
23.19
25.16
26.94
28.45
30.75
32.59
34.82
35.88
30.16
41.40
44.68
46.78
48.23
49.54
50.07
TIME/MIN.
0.55
1.05
1.55
2.05
3.05
3.55
4.05
4.55
5.55
6.55
7.55
8.55
9.55
10.55
12.05
13.55
15.55
17.05
21.76
26.76
36.76
46.76
56.76
66.76
74.26
%/MIN.
4.13
3.02
4.99
3
3
68
28
3.15
3.15
2.62
2.69
2.43
2.16
1.97
1.77
1.51
1.53
1.22
1.11
0.70
0.69
0.44
0.32
0.21
0.14
0.13
0.07
REACTED
1.15
3.05
5.06
7.22
9.79
12.22
13.80
15.24
17.25
19.81
22.11
24.18
26.05
27.69
29.60
31.67
33.71
35.35
37.52
40.28
43.04
45.73
47.51
48.89
49.81
UNIT RATE
8363.65
6045.72
9988. 5R
7360.01
6571.43
6308.58
6308.58
5257.15
5388.58
4862.86
4337.15
3942.86
3548.57
3022.86
3066.67
2453.33
2234.28
1401.90
1393.73
893.71
657.14
420.57
289.14
262.85
140.19
Conditions on Table D-5.
-------
Table D-10
TG DATA FOR RUN 258a
(LIMESTONE 1359 SULFATION)
REACTED
1.30
A.05
6.41
3.50
10.20
11.90
13.60
14.91
16.35
17.53
19.88
21.71
23.28
25.25
26.82
23.65
29.96
30.74
31.66
32.18
33.23
33.42
34.01
34.80
35.12
TIME/HIM,
0.25
0.75
1.25
1.75
2.25
2.75
3.25
3.75
4.25
4.75
5.75
6.75
7.75
9.75
10.75
12.75
14.75
17.25
19.75
22.25
29.75
37.25
57.45
107.45
147.45
%/MIN.
5.23
5.49
4.70
4.18
3,
3,
3,
40
40
,40
2.61
2.87
2.35
2.35
1.83
1.56
0.98
1.56
0.91
0.65
0.31
0.36
0.20
0.13
0.02
0.02
0.01
0.01
% REACTED
0.65
2.68
5.73
7.45
Q.35
11.05
12.75
14.26
15.63
16.94
18.70
20.80
22.50
24.26
26.03
27.73
29.30
30.35
31.20
31.92
32.70
33.32
33.72
34.40
34.96
UNIT RATE
10466.48
10989.81
OA1Q.83
8373.18
6803.21
6803.21
6803.21
5233.24
5756.56
4709.01
4709.91
3663.27
3139.94
1962.46
3139.94
1831.63
1308.31
627.98
732.65
418.65
279.10
52.33
58.29
31.39
16.35
Conditions on Table D-5.
-------
Table D-ll
TG DATA FOR RUN 259*
(LIMESTONE 1359 SULFATION)
oo
% REACTED
1.05
2.91
4.63
6.22
7.81
9.27
10.46
11.65
12.84
13.97
15.89
17.48
19.73
20.52
21.85
23.04
23.83
24.50
25.16
27.54
29.40
30.99
32.18
33.37
35.22
37.08
37.81
TII1E/HIN,
0.27
0.77
1.27
1.77
2.27
2.77
3.27
3.77
4.27
ft.77
5.77
6.77
8.77
9.77
12.27
14.77
17.27
10.77
22.27
32.27
42.27
52.27
62.27
72.27
92.27
112.27
123.77
% /MIH
3.85
3.70
3.44
3.17
3.17
2.91
2.28
2.38
2.38
2.25
1.P2
1.58
1.12
0.79
0.52
O.u7
0.31
0.2fi
0.26
0.23
0.18
0.15
0.11
0.11
0.09
0.09
0.06
REACTED
0.52
1.98
3.77
5.42
7.01
8.54
9.86
11.05
12.25
13.40
14.93
16.68
18.60
20.13
21.18
22.44
23.44
24.16
24. R 3
.? 6 . 3 5
28.47
30.10
31.58
32.77
34.30
36.15
37.44
UNIT RATE
7705.43
7416.48
6886.73
6356.98
6356.98
5827.23
4767.73
4767.73
4767.73
4502.86
3P40.67
3178.49
2251.43
1589.24
1059.^9
953.54
635.69
52°.7A
529.74
A76.77
370.82
317.84
238.38
23*.38
185.41
185.41
126.67
a
Conditions on Table D-5.
-------
Table D-12
TG DATA FOR RUN 260*
(LIMESTONE 1359 SULFATION)
VD
% REACTED
2.11
4. 36
6.33
3.44
10.01
11.65
13.08
14.58
17.23
19.75
21.87
23.84
25.27
26.77
20.97
32.15
33.79
34.88
35.97
36.52
33.15
33.83
39.51
39.72
TIME/MI JJ.
0.43
0.93
1.43
1.93
2.43
2.93
3.43
3.93
4.93
5.93
6.93
7.9T
8.93
9.93
12.43
14.93
17.43
19.03
22.43
24.93
39.51
59.51
99.51
122.01
4.91
4.40
3.95
4.22
3.13
3.27
2.86
2.99
2.65
2.52
2.11
1.97
1.43
1.40
1.2R
0.87
0.65
0.43
0.43
0.21
0.11
0.03
0.01
0.01
% REACTED
1.05
3.23
5.34
7.30
9.23
10. 33
12.36
13.83
15.™
18.49
2fi.O?
2 a . 3 7
11. nf
3^. 97
3A.3/,
3 5 . A T
36. 7.U
37.33
38.4«
3". 17
39.62
UNIT RATE
0*24.18
71
79
46
00
37
92
7903
8448
6268
6541
57?3
5995
5114.56
5042.02
3051.85
7861.68
?097.96
7561
1744
1308
872
«72
436
224
68
34
89
26
20
13
13
06
23
13
06
18.16
Conditions on Table D-5.
-------
Table D-13
TG DATA FOR RUN 261*
(LIMESTONE 1359 SULFATION)
10
O
REACTED
2.IS
4.65
6.56
9.71
10.25
11.89
14.97
17.78
20.51
23.79
28.31
31.18
33.23
34.74
36.45
37.74
39.25
39.93
41.30
42.12
43.08
43.76
44.86
45.81
46.16
TIME/MIN,
0.4P
0.90
1.40
1.90
2.40
2.90
3,
4
5
7,
.90
.90
.90
.40
9.90
11.90
13.90
15.90
18.90
22.40
27.40
29.10
37.40
44.90
54.90
64.90
94.90
144.90
164.90
% /MIN,
5.47
4.02
3.32
6.29
09
28
07
80
2.73
2.18
80
43
02
1,
1
1,
0.75
0.56
0.37
0.30
0.27
0.13
0 10
0.00
C.06
0.03
0.01
0.01
7, PEACTED
1.09
3.41
5.60
8.13
9.98
11.07
13.43
16.37
19.14
22.15
?6.05
29.74
32.21
33.98
35.59
37.10
3S.50
30.50
40.62
41.71
42.60
43.4?
4/-.31
45.3/1
45. oo
UNIT RATE
10941.98
9847. 7*
7659.38
12583.28
2188.3?
6565.19
6154.86
5607.76
5470.99
4376.79
3610.85
2872.27
2051.62
1504.52
1139.79
742.49
601.80
547.09
364.73
218.83
191.4R
136.77
72.04
38.20
34.10
a
Conditions on Table D-5.
-------
Table D-14
TG DATA FOR RUN 300®
(LIMESTONE 1359 SULFATION)
ro
to
REACTED
3.75
6.03
3.04
9.91
14.07
17.15
20.37
21.71
22.78
24.45
27.40
29.48
31.09
32.49
33.63
35.64
38.19
38.72
39.66
TIME/KIN.
0.41
0.66
0.91
1.16
1.91
2.66
4.16
5.66
7.16
10.16
17.66
25.16
32.66
40.16
47.66
62.66
77.66
92.66
97.53
9.09
9.11
8.04
7,
5,
,50
53
4.10
2.14
0.89
0.71
0.55
0.39
0.27
0.21
0.18
0.15
0.13
0.16
0.03
0.19
7. REACTED
1.87
4.80
7.03
8.97
11.99
15.61
18.76
21.04
22.24
23.61
25.93
28.44
30.28
31.79
33.06
34.64
36.92
38.46
39.19
UNIT RATE
1R193.39
18225.88
16081.66
15009.54
11078.47
8219.51
4288.44
1786.85
1429.48
1116.78
786.21
553.92
428.84
375.23
303.76
268.02
339.50
71.47
384.86
Conditions on Table D-5.
-------
Table D-15
TG DATA FOR RUN 301*
(LIMESTONE 1359)
N>
M
NJ
% REACTED
1.27
2.25
3.52
4.94
7.05
9.17
10.65
12.28
15.52
19.05
21.17
24.42
27.10
30.14
32.32
34.30
35.29
36.98
37.76
38.47
38.32
39.38
TIME/MIN,
0.12
0.22
0.32
0.47
0.72
0.97
,22
.47
,97
.72
,22
4.22
5.22
6.72
8.22
10.22
12.22
18.39
23.39
33.39
43.39
59.39
%/MIN.
10.16
9.88
12.70
9.41
8.47
8.47
5.92
6.49
6.49
4.70
4.23
3.24
2.68
2.02
1.45
0.98
0.49
0.27
0.15
0.07
0.03
0.03
% REACTED
0.63
1.76
2.89
4.23
5.99
8.11
9.91
11.47
13.90
17.29
20.11
22.79
25.76
28.62
31.23
33.31
34.79
36.14
37.37
38.11
38.64
39.10
UNIT RATE
20329.18
19764.48
25411.47
18823.31
16940.98
16940.98
11858.68
12988.08
12938.08
9411.65
8470.49
6494.04
5364.64
4047.01
2917.61
1976.44
988.22
549.58
310.58
141.17
70.58
70.58
Conditions on Table D-5.
-------
The Calcination of Limestone and Dolomite
The TG curves for the calcination of limestone and dolomite
obtained in preparing the samples for sulfation tests generated enormous
quantities of data. For the most part the detailed kinetics have not
been analyzed; when the data acquisition/data reduction system has been
commissioned, these data will be reduced as a matter of course. Several
features of the curves were examined, however, as time permitted.
Sample Consistency. The normal sample weight for limestone
1359 was approximately 19 mg. It is possible that the choice of a
small sample (essential if the maximum rates of reaction are to be
recorded) may lead to gross variations in sample content and give
results unrepresentative of the particular stone. One check on this
possibility is to examine the variation in carbon dioxide loss on
calcination in the TG apparatus.
For samples calcined in nitrogen the carbon dioxide weight
loss corresponded to 43.59 ± 0.20 percent (16 samples). This may be
compared to the weight loss noted on ignition to 1000°C (1932°F):
43.32 percent, (Westinghouse), 43.6 percent (TVA) .. For samples on the
TG apparatus which were calcined after annealing in carbon dioxide
the weight loss was 44.15 percent ±0.15 percent (seven samples). It
seems unlikely that errors were introduced by wide impurity variations
in the small samples.
i
The Rate of Limestone Calcination
For TG 215, in which 420-to-500 micron limestone 1359 was
held in carbon dioxide, heated to 901°C(1652°F) and then calcined in
nitrogen, the contracting cube equation
3kt
adequately described the course of reaction, yielding a k value of
-2 -1
2.2 x 10 sec (-calcination completed in 3.9 minutes). This value
223
-------
can be compared with a value obtained by extrapolating k values for
isothermal experiments at different temperatures (550 to 870°C/
-2 -1
1022 to 1598°F)which is 1.2 x 10 sec . The rate constant obtained
from nonisothermal TG measurements, however, using the nonlinear least
-2 -1
squares (NLLS) method is 2.6 x 10 sec , in excellent agreement with
the observed value. Since the isothermal experiments were carried
out by plunging the sample into the furnace at the temperature of the
experiment (probably fracturing some of the particles), it is not
surprising that the nonisothermal method in which the sample was
heated slowly (10 K/min) yields the closest value. The NLLS data
should be valuable as a predictor of the calcination behavior of lime-
stone.
The Rate of Dolomite Calcination
The nonisothermal data for the calcination of dolomite 1337
(1200 to 1000 micron) in nitrogen in TG 205 was analyzed by the method
( 12
of Coates and Redfern. The data yielded an activation energy of
214.9 kJ/mole (51.4 kcal/mole) which is close to the value of 54.5 for
an Ohian dolomite (of almost identical analysis) given by Jungten and
van Heek in their monograph on nonisothermal kinetics.
224
-------
REFERENCES
1. Keairns, D. L., D. H. Archer, R. A. Newby, E. P. O'Neill, E. J. Vidt.
Evaluation of the Fluidized Bed Combustion Process. Vol. I.
Environmental Protection Agency. Westinghouse Research Laboratories.
Pittsburgh, PA. EPA-650/2-73-048 a. NTIS PB-231 162/9. December
1973.
•2. Curran, G. P., C. E. Fink, and E. Gorin. CO. Acceptor Gasification
Process in Fuel Gasification. Consolidation Coal Co. Advances in
Chemistry Series 69. American Chemical Society. Washington, D.C.
1967. p. 141.
3. Hartman, M. and R. W. Coughlin. Reaction of Sulfur Dioxide with
Limestone and the Influence of Pore Structure. Ind. Eng. Chem.
Process Des. Develop. 13; 248, 1974.
4. Hatfield, J. D., Y. K. Kim, R. C. Mullins, and G. H. McClellan.
Investigation of the Reactivities of Limestone to Remove Sulfur
Dioxide from Flue Gas. Office of Air Pollution. Tennessee Valley
Authority. 1971.
5. O'Neill, E. P., D. L. Keairns, and W. F, Kittle. (Proceedings of
the Third International Conference on Fluidized Bed Combustion.
Hueston Woods, Ohio, 1972). EPA 650/2-73-053. December 1973.
6. Moss, G. Sulphur Removal During Fluidized Bed Combustion. Esso
Research Center. Abington, England. The Chemical Engineer
(Birmingham University) 23: 24, 1972.
7. Jungten, H. and K. H. van Heek. Topics in Current Chemistry. 13,
3/4, 601-699, Springer-Verlag, Berlin, 1970.
8. Hills, A. W. D. Transactions/Section C of the Institution of Mining
and Metallurgy, _76_: C241, 1967.
9. Chemical Rubber Co. Handbook, ed. Weast. 53rd Edition, F66, 1972.
225
-------
10. Schwitzgebel, K. and P. S. Lowell. Thermodynamic Basis for Existing
Experimental Data in MgO-S02-02 and Ca-SCL-O- Systems. (Radian
Corporation, Texas1). Env. Sci. Technol. ]_: 1147, 1973.
11. Handman, L. M. and E. P. O'Neill. Unpublished. Westinghouse
Research Laboratories. Pittsburgh, Pa. 1972.
12. Coats, A. W. and J. P. Redfern. Kinetic Parameters from Thermo-
gravimetric Data. Nature, 201: 68, 1964.
226
-------
APPENDIX E
E
PROJECTIONS OF SORBENT UTILIZATION AND SULFUR REMOVAL
EFFICIENCY USING THERMOGRAVIMETRIC DATA
APPENDIX E
DERIVATION OF THE MODEL EQUATIONS
-------
APPENDIX E
PROJECTIONS OF SORBENT UTILIZATION AND SULFUR REMOVAL
EFFICIENCY USING THERMOGRAVIMETRIC DATA
INTRODUCTION
The object of the thermogravimetric (TG) studies on the sulfa-
tion of limestone and dolomite is to obtain design parameters for flui-
dized bed desulfurization systems. At the outset of this study it was
postulated that the rate of sulfur sorption by particles of limestone or
dolomite by the overall Reaction 1:
CaO + S02 + 1/2 02 •*• CaSO^
and the total capacity of the stone to absorb sulfur are primary factors
in determining the usefulness of a particular sorbent, granted the thermo-
dynamic conditions for sulfur dioxide (S0_) removal are favorable. TG
studies can be assumed, a priori, to provide an estimate of the relative
impact of particle size, temperature, pressure, gas composition, and
choice of sorbent on desulfurization in a fluidized bed. The TG data
should, therefore, permit a qualitative assessment of the effect of these
parameters on the design and operation of fluidized beds. The possibility
of enhancing the value of TG data, however, by using it to make quantita-
tive predictions of sulfur removal in fluidized beds, is worth analysis
because it would:
• Permit a critical assessment of advanced designs
• Permit optimization of a given system
• Permit development of a meaningful laboratory screening
technique for choosing particular sorbents
• Provide a valuable tool for trouble-shooting in cases where
poor desulfurization may be the result of a number of factors
• Justify development work by TG analysis to improve the sulfur
removal system, as opposed to total reliance on more expensive
fluidized bed tests as the sole experimental tool.
227
-------
LABORATORY STUDIES OF THE KINETICS OF REACTION
Numerous studies of the kinetics of Reaction 1 have been pub-
lished. Justification for further work on the system rests on the need
for data under closer simulation of proposed fluidized bed operating con-
ditions (1013 to 1520 kPa; 10 to 15 atm pressure) and also on the failure
of currently available data to explain fluidized bed results. As exam-
1*
pies of the latter, work at Exxon Research and Engineering, Tennessee
2 3
Valley Authority (TVA), and the National Coal Board, U.K. (NCB), may be
cited.
In the TVA study kinetic measurements on a particular limestone
2
(limestone 1359), the maximum degree of sulfation observed was 18 percent.
Using the same type of stone, Exxon workers found that 32 percent of the
stone was sulfated in fluidized beds before significant escape of sulfur
dioxide from the bed was observed. Similar particle sizes and oxygen
pressures were used at atmospheric pressure in both tests. The results
4
are significant , because Borgwardt has established that the limit on
fast sulfation kinetics is determined by the space available in the stone
pores to accommodate the bulky sulfate ion. This limit should not depend
on whether the reaction takes place in a TG apparatus or a fluidized bed.
In the NCB study ' laboratory results from a differential reactor were
used to model extensive fluid bed results. Excellent agreement was ob-
tained for one limestone and for one dolomite, but with a second lime-
stone (limestone 1359) significantly better retention was observed in the
fluid bed than predicted by the model. A successful resolution of these
apparent anomalies will be necessary before the utility of models can be
accepted.
Projections of Fluidized Bed Performance
Application of kinetic studies to projections of fluidized bed
desulfurization has been carried out by at least two groups working at
At the time this report was written, the corporation was named Esso. To
avoid confusion with Esso (U.K.) the current name, Exxon, will be used.
228
-------
Argonne National Laboratory (ANL) and NCB. The results these workers
obtained represent important milestones, but they were not carried through
to the point of assessing what performance projections could be made
using laboratory kinetic data, and they did not have data at pressure,
a parameter of considerable importance to the practical application of
fluidized bed desulfurization.
In addition to providing a resolution of the anomalies noted, a
successful model which uses laboratory kinetic data should predict the
effect of:
• Calcium/sulfur (Ca/S) ratio on desulfurization
• Temperature on desulfurization
• Bed height and gas velocity on desulfurization
• Pressure on desulfurization.
Fluidized Bed Predictions'
The model developed at ANL simulates the sulfation process as
a first-order reaction which proceeds to an equilibrium loading on the
stone, Ye, which is less than stoichiometric, as experiments show. Sulfur
retention in the bed is calculated by mass balancing this stone reaction
rate with the removal of sulfur dioxide as it passes through the bed
and the generation of sulfur dioxide in the bed by coal combustion. While
application and development of the model was clearly hampered by the
dearth of suitable kinetic data, estimated data from experiments conducted
by the National Air Pollution Control Administration (NAPCA)* were used
to illustrate the model. The predictions of 90 to 99 percent sulfur
removal compared with the value of 80 percent observed in a fluidized bed
test (BCR-8). In order to apply the model the equilibrium loading and the
calcium/sulfur molar ratio in the feed are specified. This model has not
undergone subsequent development but was used by Westinghouse to apply TG
data to predicting a fluidized bed test result. The chief difficulty lay
in defining Ye, the equilibrium loading, on the sorbent, and arbitrary
*
Now the EPA.
229
-------
allocation of Ye (by inspection of the TG kinetic curve) for a represen-
tative TG result predicted a sulfur removal efficiency of 87 percent as
opposed to 81 percent in a fluidized bed test. Had the equilibrium sulfur
loading been chosen as 37 percent (the actual stone loading) in the ex-
periment the model would have yielded the correct answer, 81 percent,
since the reaction rate of the stone up to 37 percent was sufficiently
high.
A more recent model has been published by the NCB investigators,
who were, once again, hampered by the lack of appropriate kinetic data to
use in their model. They achieved excellent agreement between laboratory
kinetic data and their fluidized bed results, with the conspicuous excep-
tion of fluidized bed sorbent tests with the widely used fluidized bed
*
sorbent, limestone 1359, which did not agree with the model predictions.
The reasons for this disagreement are essential to an understanding of
the processes controlling fluidized bed desulfurization.
It is accepted that the pore structure formed in the sorbent
during calcination controls the kinetics of sulfation.^ The Westinghouse
work on calcination under different partial pressures of carbon dioxide
(see Appendix J) shows that when calcination occurs under carbon dioxide
with Pc02/Pe > 0.6 (where Pe = equilibrium partial pressure of CO, over
CaCO./CaO at the temperature of calcination), the capacity of the stone
for sulfation at a high rate of reaction is increased; inspection of the
fluidized bed literature on sulfur dioxide retention shows powerful evi-
dence confirming this phenomenon.
In the course of experiments designed to determine the optimum
temperature for sulfur dioxide capture by limestone, the NCB workers
operated a fluidized bed of coal and limestone at a series of temperatures
and recorded the steady-state sulfur dioxide retention as a function of
temperature. They determined that the optimum temperature for sulfur di-
oxide removal was approximately 830°C (1526°F). Their data, however, can
* 5
Not stated in their paper but can be inferred from this paper and the
overall project report.^
230
-------
be examined from the point of view of the carbon dioxide particle pres-
sure condition in the bed. (The carbon dioxide is contributed both by
limestone and by coal combustion.) The sulfur dioxide escaping from the
bed has been plotted here as a function of PcQ./Pe, and, as illustrated in
Figure E-l,it shows that maximum sulfur dioxide retention in the bed
occurs when P(;o«/Pe is about 0.6.
In order to see if the TG results for sulfation after calcina-
tion under high relative partial pressures of carbon dioxide explained
the relatively high utilization obtained by the NCB, the data for a typi-
cal run TG 199 were approximated by first-order equation. The data were
then used to predict fluidized bed experimental results as NCB noted.
A comparison of predicted and observed results is shown in Figure E-2.
Since the NCB investigators found about 11 percent utilization in their
laboratory tests, it is clear that their data could not predict 80 percent
sulfur retention at a calcium/sulfur molar ratio of 3/1 within any imagi-
nable gas residence time in the fluidized bed. It can be concluded that
the carbon dioxide partial pressure during calcination is a factor of
crucial importance for fluidized bed desulfurization. Accordingly, a
simple model of fluidized bed desulfurization, which permits direct use
of the TG data without arbitrary allocation of an equilibrium sulfur
loading, was developed. A set of sulfation experiments using limestone
1359 was carried out under carefully monitored calcination conditions,
the data were applied to the model, and the results were compared with
fluidized bed experimental results.
Experimental
Limestone 1359 (M. J. Grove Lime Co., Stephens City, Va.) con-
taining 96.7 percent calcium carbonate was ground and the fraction 420 to
500 vim used in the experiments. Calcination and sulfation were effected
using 20 mg samples of limestone in a Pt gauze basket in a DuPont 951
7 8
thermobalance modified for work with corrosive gases . ' The samples were
brought to 900°C (1652°F) in pure carbon dioxide, then calcined in
60 percent carbon dioxide in nitrogen at 900°C (1652°F). Sulfation at
231
-------
"8
CQ
•4—
o
to
1000
900
800
700
600
500
400
300
200
100
0
1 I ' I i I '
i I
Welbeck Coal
U.K. Limestone<3175um
Ca/SMoleRatio2.8
Coal Feed Rate 137 kg/hr
Velocity 2.5-2.3m/sec
Bed Height 0.67m
799-888° C (1470-1630° F)
Atmospheric Pressure
i I i I i I i I i
D .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0
Fraction of Equilibrium C02 in Effluent,
Figure E-l-The effect of C02 pressure during calcination
on SOo emissions from a fluid bed. (Data taken
from National Coal Board (U. K.) Study - Tempera-
ture Survey-Table A. 1. 3.19, Page Al. 95)
232
-------
Curve 679225-A
90
C
o
•as
on
CO
70
60
50
40
30
20
10
0
a NCB
Fluidized Bed Results,
6" Combustor,
Illinois Coal,
Atmospheric Pressure, 800°C
(1470°F)0.9 m/s (3ft/sec)
0.61m (2ft.); 0.9m (3ft.)
Median Diameter, ~520um
o TG 199-Limestone 1359
after Slow Calcination.
Using ANL Model
0
1234
Calcium/Sulfur Mole Ratio
Figure E-2-TG data predictions compared to fluidized bed data
atmospheric pressure limestone 1359
233
-------
each temperature was in 5,000 ppm sulfur dioxide, 4 percent oxygen in
nitrogen, flowing at 400 ml min through the 20 mm diameter reactor, at
atmospheric pressure. The primary data recorded, weight as a function of
time, indicate the course of sulfation, as shown in Figure E-3. The de-
rived rate of reaction is shown as a function of the extent of reaction
or calcium oxide utilization in Figure E-4 for two temperatures. Ana-
lytic expressions for these data can only be written in terms of
9 10
coupled reaction and diffusion rates requiring arbitrary parameters. '
The data, however, can be used in the graphic form shown to project
fluidized bed performance.
Calculation of Kinetic Limits to Sulfur Retention
To apply the data from the TG runs (257 to 262) to "predictions"
of fluidized bed sulfur retention, the following procedure was adopted:
• The fluidized bed conditions were those described by
ANL in their tests AR-1 and HUMP as shown in
Tables E-l and E-2.
• A first-order model with respect to gas concentration
was assumed, and uniform generation of sulfur dioxide
in the bed was considered, according to the equations
which follow.
Table E-l
PHYSICAL PARAMETERS OF THE TG SYSTEM
ANL/fluid bed
Westinghouse/TG
Stone
Particle Size
Temperature Range
Oxygen
S00
Limestone 1359
(98 wt % CaC03)
615 urn
788 to 871°C
(1450 to 1560°F)
3.5%
3,000 to 420 ppm
Limestone 1359
(96.75% CaC03)
460 ym
750 to 950°C
(1382 to 1742°F)
5,000 ppm
234
-------
Curve 65553^-B
N>
OJ
Ul
CaO + Sty + 02
1.03 x 10* N/m2
~10mg in 400 ml. min
Limestone 1359
420-500 urn
750°C <1382°F)
800° C (1472° F)
850°C <1562°F)
900°C (1652°F)
950°C U742°F)
10
20
50
Time/Minutes
Figure E-3-The effect of temperature on limestone sulfation
-------
Curve 655732-A
.1
c
o
:= 01
T» •u*
10
u>
LO
-5 .001
I
n 1 i
0950°C(1742°F)
A750°C(I382°F)
Limestone 1359
420-500 urn
5,000ppm SO 2
in N2
5.0 10.0 15.0 20.0 25.0 30.0
Calcium Oxide, % Utilization
35.0 40.0 45.0
Figure E-4-The rate of sulfation of limestone 1359
-------
Table E-2
SULFUR RETENTION PROJECTIONS COMPARED
WITH EXPERIMENTAL DATA (ANL-AR1)
Experiment
TG/W
FB/ANL
FB/ANL
TG/W
FB/ANL
FB/ANL
TG/W
FB/ANL
TG/W
TG/W
Temperature
°C (°F)
750 (1382)
760 (1400)
787 (1448.6)
800 (1472)
816 (1500.8)
843 (1549.4)
850 (1562)
871 (1599.8)
900 (1652)
950 (1742)
Utilization, %
14.7
18
26
23.6
36
36
33.2
34
32.4
26.2
% sulfur
retention
36.3
44
65
59
91
91
83
86
80
65.5
237
-------
1 \7 /i v \
R - 100 U - frr (1 ~ e )
- kH
(E-l)
(E-2)
where
U = CaO fraction sulfated ("the utilization")
R = % sulfur retained in the fluid]zed bed
v = actual gas velocity in fluidized bed, mm/sec
H = bed height, mm
k = first-order rate constant for the reaction
CaO + S0 + 1/2
where k = f fcj, p)
F = Ca/S mole ratio as supplied to the bed
p =• molar density of calcium in the fluid bed.
(Equations E-l and E-2 are derived in Appendix E...)
• The values of F 'Ca/S mole ratio), v (gas velocity),
H (bed height) are taken from the ANL description of
their experiments. The value of v was corrected for
the effect of solid volume in the bed. The value of
p, the molar density of calcium in the bed, was esti-
mated.
• The values of k, the rate constant, necessary to
achieve the sulfur retention implicit in fixed Ca/S
feed rate ratios and stone utilization were calculated
as a function of U, the utilization.
• The values of k were converted into the rate, k', as
usually expressed for the TG data where pk'/C. = k
(described in Appendix E..).
238
-------
• The values of k = f (u) from the model equations were
compared with graphic plots of the experimental data. The
coincidence of k and the experimental rate gives the utilization
U and, hence, the sulfur retention, as shown in
Figure E-5.
Results
A comparison of the limits predicted by the TG results and
fluidized bed experimental results for sulfur retention with limestone
1359 is shown in Figures E-6 and E-7. At a calcium/sulfur ratio of 2.5/1,
the sulfur retention predicted is within 10 percent of that observed up
to 871°C (1600°F) and is predicted to fall off in the range 870 to 950°C
(1598 to 1742°F). The apparent temperature for maximum sulfur retention
is about 30°C (86°F) higher for the TG data than for the fluidized bed
experiments at ANL but is in close agreement with the optimum temperature
Esso noted.
The experiments with Humphrey coal, at a calcium/sulfur ratio
of 4.5/1, are also relatively well predicted (within 10 percent) by the
TG data, up to 850°C (1562°F). The 9008C (1652°F) temperature result at
this calcium/sulfur ratio shows that in practice the sulfur retention is
much poorer than the predicted value.
The good agreement at the lower temperatures is further demon-
strated by the comparison of TG projections and experiments in which the
calcium/sulfur ratio was varied (from 1 to 4) at 788°C (1450°F), as shown
in Figure E-8. This agreement covers a range of sulfur retention in the
bed from 38 to 95 percent.
The model equations were used to project the effect of superfi-
cial gas velocity on sulfur retention at two calcium/sulfur ratios; 4/1
and 1/1, as shown in Figure E-9. Qualitatively,the projections are in
12
agreement with the experimental data from fluidized beds, which show:
• Increasing the velocity four-fold (up to about
2.5 m/sec [8 ft/sec]) decreases sulfur retention from
over 90 percent to about 80 percent.
239
-------
Curve 655733-A
32
28
o
CO
O
*O
20
re
| 16
o
CO
CD
o
0>
"TO
12
8
Data from TG at
750° C (1382 °F)
Fluidized Bed
Model
Predicted
Utilization
2
6 10 14 18
Calcium Oxide % Utilization
22 26
Figure E-5- Determination of the stone utilization atwhich
the rate of reaction satisfies the fluidized bed
rate criteria
240
-------
Curve 655730-A
100
.1 80
**
"55
^
"3
60
20
Experimental Data From Argonne
National Laboratories 150 mm
Diameter Coal Combustor Using.
Limestone 1359.
Bed Height 610 mm
Velocity ~ 800 mm/sec
Ca/S Mole Ratio 2.5/1
o Fluidized Bed Behavior
Predicted by First-Order
Model Using ^ TG Data
for Reaction:
CaO+S02+ 1/2 02»CaS04-
with Limestone 1359
750 800 850 900 950
(1382) (1472) (1562) (1652) (1742)
Temperature, °C (°F)
Figure E-6-The effect of temperature on sulfur retention in a fluidized bed of
limestone 1359
Curve 655731-A
o>
at
100
90
80
70
60
50
40
30
20
10
iii i
° 0
- °*
~ &
-
A
_ AANL Fluidized Bed Experiments. 150 mm Coal
Ca/S =4. 1-4. 3
Limestone 1359
610 mm Bed Height
850 mm/ sec Velocity
_ o Limits Derived from TG Data
ill i
750 800 850 900
(1382) (1472) 11562) (1652)
Temperature, °C (°F)
0
—
-
-
-
Combustor_
-
-
950
(1742)
Figure E-7-Comparisen of TG data with fluidized bed results
241
-------
*s 655729-A
TG Data At 800°C (1472° F)
in First-Order Model
AANLData
Experiments Hump ID
2A
2B
3
788°C (1450° F)
793 mm/sec Velocity
610 mm Height
Limestone 1359
152 mm Diameter
Coal Combustor
Atmospheric Pressure
23456
Ca/S Mole Ratio
Figure E-8-The effect of Ca/S mole ratio on sulfur retention
Curve 65573^-A
100
90
80
70
I 50
40
30
20
10
Ca/S = 4/1
Ca/S = 1/1 -
1000 2000 3000
Velocity/mm/sec
Figure E-9-Model predictions of the effect of
superficial velocity on sulfur
retention
242
-------
• The decrease in sulfur retention follows a steeper
curve at low calcium/sulfur ratios (1/1).
Both of these features are predicted by the model, as shown in
Figure E-9.
The decrease in sulfur retention at temperatures above 870°C
(1598°F) is more severe in practice than the simple first-order model
predicts. The TG data used here, however, do not faithfully reproduce
the conditions of the fluidized bed experiments; calcination in the
fluidized bed will be much more rapid in this temperature range than
occurred in the TG experiments, since the equilibrium pressure of carbon
dioxide rises rapidly in this temperature range. Thermogravimetric ex-
periments show that the time/temperature history during calcination
affects the capacity of the calcined stone for reaction with sulfur di-
oxide.7'13
Projections Using the First-Order Model
Because the model shows a satisfactory relation between sulfur
removal in fluidized beds and the TG data obtained under careful simula-
tion of fluidized bed conditions, a number of general curves were gene-
rated from Equation E-2 to show the average reaction rate required to
desulfurize the gases generated in the fluidized bed combustion of coal.
The curves in Figure E-10 illustrate the effect of gas residence
time (actual) in the bed on sulfur removal. For 80 percent sulfur removal
with a residence time of one second, a rate of reaction in the TG tests
_2
at (0.5 percent sulfur dioxide, 101.3 kPa [1 atm ]) of 2.2 x 10 cal-
cium reacting per minute is required.
Table E-3 shows the utilization of sorbent at which the reaction
rate has fallen to the critical value for maintenance of the 80 percent
sulfur dioxide retention level.
While the parametric plot of Figure E-10 can be used to estimate
the performance of a sorbent in a given configuration of fluidized bed
combustor using TG data from Appendix D, the data used must be selected
to reflect the calcination conditions in the fluidized bed.
243
-------
Table E-3
FLUID BED SORBENT UTILIZATION PREDICTED
FROM THERMOGRAVIMETRIC DATA
NJ
TG Sorbent Ca]
% Ca utilization for 80% S02
Pressure removal at 1.0 sec
.cination kPa,(atm) gas residence time
215 Limestone 1359 Calcined in N, 101.3(1) 7
(420-500 urn)
221 " Calcined in C02 " 37
257
258
259 "
260
261
300
301
202 Dolomite 1337
50
33
34
it ii 38
42
1013(10) 21
1013(10) 34
1013(10) 54
(1200-1000 pm)
-------
Curve 679226-A
-1.0
-2.0
E
e
o
I— I
SP
-4.0
'
.
as Residence Time
a 0. 125sec.~l
° * cn°!eC'
o 0. 50 sec.
A 1.0 sec
Ca Density in Bed = 7. 75 x io"3 Mole/cm"3
(Dolomite) Stone Reaction Rate at
0.5%SOX; 4%02
,(SOX =
S03)
1
1
r
i
-1.0
TO
0.
-2.0
C
O
03
O
"""o
• 1
f
0 10 20 30 40 50 60 70 80 90 100
% SCL Retained
Figure E-10-The reaction rate criteria for desulfurization
245
-------
The model developed, in spite of its simple approach, shows
that fluidized bed results on desulfurization are accurately paralleled
by TG data and that quantitative conclusions can be drawn from TG data.
Specifically, the model showed:
• A temperature maximum for desulfurization at atmos-
pheric pressure
• The effect of calcium/sulfur molar ratio on desul-
furization
• The effect of gas velocity on desulfurization
• The effect of the ambient carbon dioxide partial pres-
sure on desulfurization by virtue of the effect of
PpQ2 on calcination
• The effect of pressure which raises the partial pres-
sure of carbon dioxide thereby causing higher utili-
zation through increased stone capacity rather than
by directly influencing the rate of reaction.
The overall conclusion from the application of TG data to the
model is that:
• Ca/S mole ratios of ^ 1/1 can achieve 90 percent de-
sulfurization if the dolomite is not calcined under
low partial pressures of carbon dioxide. Pressurized
operation, therefore, which maintains a high PCQO in
the bed should tend towards the 1/1 calcium/sulfur
molar ratio.
• Half-calcined dolomite should be used as a sorbent in
the range intermediate between 1/1 and 2.5/1 calcium/
sulfur molar feed ratio. Improvement in desulfuriza-
tion can be affected by modest increases in bed height
or gas residence time.
• Further experimental and theoretical development of
the model should consider
- The effect of particle size distribution
- The effect of high-temperature operation above
1750°C (3182°F)
246
-------
The effect of grain structure of the raw dolo-
mite on its capability to be calcined so that
it can be used at low calcium/sulfur molar
ratios.
247
-------
REFERENCES
1. Hammons, 6. A. and A. Skopp. A Regenerative Limestone Process for
Fluidized Bed Coal Combustion and Desulfurization. Environmental
Protection Agency. Report to A.P.C.O. Exxon Research and Engineer-
ing Co. Linden, N.J. February 1971.
2. Hatfield, J. D., Y. K. Kim, R. C. Mullins, and G. H. McClellan.
Investigation of the Reactivities of Limestone to Remove Sulfur
Dioxide from Flue Gas. Air Pollution Control Office. Tennessee
Valley Authority. 1971.
3. Cox, D. G., J. Highley, et al. Reduction of Atmospheric Pollution.
Environmental Protection Agency. National Coal Board. London,
England. September 1971.
4. Borgwardt, R. H. and R. D. Harvey. Env. Sci. Tech. 16: 350, 1972.
5. Bethell, F. V.,D. W. Gill, and B. B. Morgan. Mathematical Modeling
of the Limestone-Sulfur Dioxide Reaction in a Fluidized Bed Com-
bustor. Fuel. 52, 1973.
6. Jonke, A. A., G. J. Vogel, et al. Reduction of Atmospheric Pollu-
tion by the Application of Fluidized Bed Combustion. Environmental
Protection Agency. Argonne National Laboratory. Argonne, Illinois.
ANL/ES-CEN-1002. June 1970.
7. Keairns, D. L., D. H. Archer, et al. Evaluation of the Fluidized
Bed Combustion Process, Volume 1. Environmental Protection Agency.
Westinghouse Research Laboratories. Pittsburgh, Pennsylvania.
EPA-650/2-73-048a. December 1973. NTIS PB 231 162/9.
8. Ruth, L. A. Ph.D. Dissertation. The Reaction of Hydrogen Sulfide
with Half-Calcined Dolomite. City University of New York. New York.
N.Y. (1972).
9. Wen, C. Y. and M. Ishida. Reaction Rate of Sulfur Dioxide with
Particles Containing Calcium Oxide. Env. Sci. Technol. 1: 703, 1973,
248
-------
10. Pigford, R. L. and G. Sliger. Ind. Eng. Chem. Process Des. Develop.
1£: 85, 1973.
11. Moss, G. The Fluidized Bed Desulfurizing Gasifier. (Proceedings of
the 2nd International Conference on Fluidized Bed Combustion.
Hueston Woods. 1970.)
12. Archer, D. H., D. L. Realms, et al. Evaluation of the Fluidized
Bed Combustion Process. Environmental Protection Agency.
Westinghouse Research Laboratories. Pittsburgh, Pennsylvania.
Contract CPA 70-9. November 1971. NT1S PB 211 494, 212 968/9, 213 152/2.
13. Appendix D—this report.
249
-------
APPENDIX E
DERIVATION OF THE MODEL EQUATIONS
The fluidized -bed is assumed to behave as a series of plug-flow
reactors in each of which a first-order reaction with respect to gas con-
centration reduces the concentration of sulfur dioxide from Co to c> so that
-V'r
C = C e k T
o
where
C = SO. concentration entering an element of bed height
C = SO. concentration leaving an element of bed height
T = residence time of the gas in the bed element
k1 = first-order rate constant for the reaction
CaO + S02 + 1/2 02 = CaSO^; k1 = f (u, p, C)
u = mean utilization of the stone in the bed
p = molar density of CaO in the bed
v = gas velocity through the bed
If the sulfur is liberated from the coal uniformly throughout
the bed, the fraction of the total sulfur released in the element of bed
height, db, is given by db/H.
Since this sulfur travels through a bed height, b, the fraction
of the sulfur released in the element db which is reacted and retained in
-k't1
the bed is 1 - e , where T' is the residence time of the gas in the
bed height b above the element db. The sulfur retention due to sulfur
generation in the element db is
-k--(H-b)
(1 - e v )db
250
-------
The fraction of the total sulfur retained in the bed is given by
H -k'(H-b)
11 - e V ]db
-k'H
-k'H
V
R = 100 [ 1 - 7 (1 - e ) J .
Using the definitions
_ moles of sulfur reacted with CaO , ....
R = - : - s - =-= - 3 - - - x 100
moles of sulfur input
p _ moles of calcium input
moles of sulfur input
„ _ moles of sulfur reacted with CaO 100
moles of Ca input 1
R = FU
k1 is obtained from T6 data as follows: At any given utilization of the
calcium oxide in the stone, the rate of reaction of the stone at sulfur
dioxide concentration C^ is
-T— = k; a = molar fraction
at
If p = molar density of the stone in the fluidized bed, the rate of reac-
tion (moles m sec ) is
251
-------
At unit gas concentration, the rate of reaction (sec ) is
£_da = Jffi. = k.
C±dt C±
252
-------
APPENDIX F
SPENT STONE DISPOSITION
-------
APPENDIX F
SPENT STONE DISPOSITION
INTRODUCTION
The pressurized fluidized bed combustion process, as presently
conceived, results in the production of dry, partially utilized dolomite
or limestone particles from 0 to 6 mm in size. In addition, fine parti-
cles of sorbent and ash will be collected in the particle removal
system. The sorbent material may be either regenerated for recycling to
1 2
the fluid bed boiler for repeated sulfur dioxide (S0_) removal ' or dis-
2 1
posed of in its partially utilized form in a once-through system. The
former process has the potential advantage of producing less solid waste
for disposition, but much uncertainty still exists about the regenerative
processes. Conceptually, it is possible in a regenerative process to
recycle all of the sorbent. This might involve a synthetic calcium-based
material, reconstitution of the spent sorbent, or an alternative sorbent
material. The composition of the spent sorbent depends on the characte-
ristics of the original stone, the coal feed, the variation in operating
temperature and pressure, as well as on once-through or regenerative
modes of operation. The major compounds in the waste stone to be dis-
posed of are calcium sulfate (CaSO,), calcium oxide (CaO)(or calcium car-
bonate [CaCO,]), and magnesium oxide (MgO) when dolomite is used, and
calcium sulfate and calcium oxide or calcium carbonate when limestone is
used. Trace elements arising from impurities in the coal and dolomite
will also be present.
A summary of the general process options for disposition of the
spent sorbent is presented in Figure F-l. Two methods of dealing with
the spent sorbent are being considered:
• Disposition of the spent stone being discharged directly
from the fluidized bed combustion process (once-through
or regenerative, Options 1 and 3)
253
-------
Fresh Limestone
or Dolomite
N)
ui
High-Sulfur
Fossil Fuel
Fluidized
Bed
Combustor
Spent Sorbent
Ca/S^l
Spent Sorbent
Spent Sorbent
Regenerated
Sorbent
Spent Sorbent
Regenerated
Sorbent
Spent Sorbent
Processing
Processing for
Sulfur Recovery
Ca/S^l
Ca/S^l
Spent Sorbent
Processing
Regeneration
Sulfur-Rich Gas
Regeneration
Sulfur-Rich Gas
Sorbent
Processing
Spent Sorbent
Ca/S~ 1
Figure F-l-Sulfur removal system process alternatives
Dwg.
Disposal/ Utilization
-»• Sulfur Recovery Option
-»• Disposal/Utilization
-*• Disposal/ Utilization
•*• Disposal/ Utilization
•*• Sulfur/Sulfuric Acid
Option
(1)
(2)
(3)
(4)
Solids Disposal/Utilization (5)
Figure 18
-------
• Disposition of the spent stone after further pro-
cessing (Options 2 and 4).
Disposition without processing includes direct disposal or utilization of
the material in soil stabilization, for example. This is the preferred
option since it does not require additional processing. Disposition of
the spent stone after further processing is being considered in order to
develop methods for rendering the stone environmentally acceptable for
disposal, if direct disposal is not universally permitted, and to inves-
tigate alternative markets for the spent stone, such as in the form of
refractory brick.
Recovery of sulfur from the spent stone for disposal or utili-
zation has not been considered attractive, although the waste stone could
be processed to recover the sulfur. Sulfur or sulfuric acid (H-SO.)
would be recovered from the sulfur-rich gas (SO. or H.S) from the regene-
ration processes. A third general option for processing the waste stone
(Option 5) would be to react the off-gas from the regenerator with the
waste sorbent to produce a solid material for disposal or utilization.
This option does not appear to offer advantages over the once-through
process options.
FACTORS AFFECTING SPENT STONE DISPOSITION
Among the factors that will affect the disposition of the spent
sorbent are, for example, the quantity of spent sorbent, its chemical
characteristics, regulations, geographical location, and the size of the
market for respective applications.
Quantity
Representative quantities of spent material are illustrated in
Table F-l for application in fluidized bed combustion systems. Compari-
sons with regenerative systems and stack-gas cleaning effluents were pre-
pared by Westinghouse and presented in previous reports. Typical
quantities of spent sorbent projected for disposal from a 500 MW power
plant burning a 3 wt % sulfur fuel with 95 percent removal would range
255
-------
Table F-l
SPENT STONE DISPOSITION
10
in
Total Utility Consumption
Coal, Tg (106 ton)
Sulfur3, Tg (106 ton)
Fluidized Bed Installation Factor
Spent Dolomite0, Tg (10 ton)
• Once- through
• Regenerative
Landfill, (acre-ft)
• Once-through
• Regenerative
1970 | 1975
296 (326) 369 (407)
8.9 (9.8) 11 (12.2)
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
| 1985
485 (535)
U.6 (16.1)
0.09 (0.10)
9.5 (10.5)
4.9 (5.4)
4,560
2,350
2000
708 (780)
21.2 (23.4)
0.41 (0.45)
62.6 (69.0)
31.8 (35.0)
30,000
15,200
a
Sulfur in coal at 3 percent.
t>
Ratio of electrical energy produced in coal-fired plants with fluid-bed units to electrical
energy produced in coal-fired plants with other combustion units.
C
Assumes dolomite Ca/S molar ratio of 1.5 for once-through and 0.75 for regeneration.
-------
from approximately 41 Mg/hr (45 tons/hr ) for a once-through system
using dolomite to approximately 19 Mg/hr. (21 tons/hr ) for a regenerative
system utilizing a calcium/sulfur makeup rate of 0.75. Specific quanti-
ties will depend on operating conditions and sorbent characteristics.
Chemical Characteristics
The chemical characteristics of the spent sorbent will depend
on a number of operating and design parameters—for example:
• The operating conditions
• The specific stone utilized in the process
• The location of the effluent stream from the process
(for example, in a regenerative process the stone
for disposal can be removed before or after the re-
generation process).
The primary constituents are calcium sulfate, calcium oxide (or calcium
carbonate),and magnesium oxide when dolomite is used; and calcium sulfate
and calcium oxide (or calcium carbonate) when limestone is used. The
amount of magnesium oxide present is determined by the magnesium carbo-
nate (MgCO.) composition of the fresh stone. The distribution of the
calcium-based compounds in the spent sorbent is summarized in Table F-2.
The chemical characteristic of the spent sorbent are assessed in
Appendix G, Spent Stone Disposal—Assessment of Environmental Impact.
The solubility of calcium and magnesium compounds are of parti-
cular importance in determining chemical activity. A compilation of
readily available solubility data on calcium and magnesium compounds was
made. The solubility data are presented in Tables F-3 and F-4 in order of
decreasing solubility. The oxide form, as well as carbonates, phosphates,
and silicates, of magnesium have minimal environmental impact. Magnesium
sulfites and sulfates would not be inert. The concentration of these
compounds, however, is negligible in the spent sorbents tested
(Appendix G). This raises the possibility, however, of achieving a sepa-
ration of dolomitic sorbents into calcium and magnesium compounds which
257
-------
separately might have market potential. One problem with phosphate end
forms is that phosphoric acid is relatively expensive, and there does not
appear to be a cheap source of the phosphate ion for spent stone process-
ing.
Table F-2
TYPICAL COMPOSITION OF CALCIUM COMPOUNDS IN SPENT SORBENT
Once-through
Dolomite
Limestone
Regenerative
Dolomite/limestone
CaO or CaC03a
mole %
10
50
10-70
CaS04
mole %
90
50
30-90b
xhe oxide or carbonate form is determined by the fluidized bed
combustor operating temperature, pressure, and excess air.
The CaSO, composition is determined by the sulfur loading
for the sorbent to the regenerator.
Regulations
An important aspect in the development of a viable stone dis-
posal process is the constraint of existing regulations relative to en-
vironmental impact. Each site will require examination of both local and
federal regulations.
The Solid Waste Disposal Act was set forth as Title II of
Pub'xc Law 89-272 of the 89th Congress, S. 306, approved October 20, 1965.
This was amended by Public Law 91-512, H.R. 11833, approved October 26,
1970. In essence, these laws recognized that the problems of waste dis-
posal had become national in scope and required federal action to promote
the development of solid waste management and resource recovery systems
and, also, an overall national research and development program. Legal
documents and technical investigations which have been reported to carry
out the provisions of these acts will provide background information and
criteria for disposition of spent sorbent from the fluidized bed combus-
tion process.
258
-------
Table F-3
SOLUBILITY OF CALCIUM COMPOUNDS IN WATER
Formula
Ca(HC03)2
Ca(HS03)2
Ca(HS)2 •
Form or
natural mineral
6H20
CaS04 • 1/2 H20 a (stable form)
CaS04 III
II
Ca(OH)2
B (unstable form)
Soluble anhydrite
Insoluble anhydrite
CaO + 2Si02
2CaO + Si02
CaO • SiO- Rseudowollastonite
CaS
CaS03 • 2H
Ca3(P04)2
CaC03
Oldhamite
2o
Calcite
Aragonite
Hydrolysis studies; 164 ppm free SiO-
containing 0.103 moles CaO/mole SiO^.
Solubility , ppm
166,000
sol.
70,000-250,000
3,000-8,000
6,600
7,000
2,100-3,000
1,300-1,400
236a
212b
95
212C
120-210d
43
25
13-14
12-15
Temperature ,
20 (68)
20 (68)
20 (68)
30 (86)
20 (68)
20 (68)
0 (32)
30 (86)
30 (86)
17 (63)
15 (59)
18 (64)
25 (77)
25 (77)
also in solution at equilibrium
Hydrolysis studies; 83 ppm free CaO also in solution at
containing 1.13 moles CaO/mole Si02>
GHydrolysis studies; solid phase 1.07 moles CaO/mole Si02
d
°C(°F)
with solid
equilibrium with solid
•
259
-------
Table F-4
SOLUBILITY OF MAGNESIUM COMPOUNDS IN WATER
Formula
MgS04.7H20
MgSCy6H20
MgHP04-7H20
MgC03-5H20
3MgCO -Mg(OH) -3H00
Natural
Mineral
Epsomite
Nesquehonite
Landsfordite
Hydromagnesite
Solubility,
ppm
710,000
660,000
3,100
1,790
1,760
400
Temperature,
°C (°F)
20 (68)
25 (77)
16 (61)
7 (45)
MgC03-Mg(OH)2-3H20
CaO-MgO-2C02
MgC03
Mg(OH)2
MgO
MgS
MgO-Si02
7MgO-8Si02-H20
3MgO-2Si02'2H20
2MgO-Si02
CaO'MgO-Si02
CaO-MgO-2Si02
2CaO-MgO-2Si02
3CaO-MgO-S102
3CaO.Mg0.2Si02
Ca0.3Mg0.4C02
2CaO-5MgO-8Si02-H20
Artinite
Dolomite
Newberyite
Magnesite
Brucite
Periclase
Bobierite
Clinoenstatite
Anthophyllite
Talc
Serpentine
Forsterite
Monticellite
Diopside
Akermanite
Nervinite
Merwinite
Huntite
Tremolite
260-440
250
200
106
9
6.2
Dec.
200
200
25 (77)
18 (64)
T)ec. = decomposes.
b
I = insoluble.
260
-------
Location and Market
The geographical location of each plant will determine the sor-
bent utilized and the disposal options. Limestones and/or dolomites vary
geographically. Factors which affect sorbent selection and will be de-
pendent on location are discussed in Appendix C. The question of market
compatibility with by-products must be considered for each case. Illus-
trations of possible applications are presented in the following sections.
One aspect of the market picture is to assess the total market available.
For example, the domestic consumption of magnesium compounds is of inter-
est. Domestic usage of magnesium compounds in 1968 is presented in
Table F-5.
The two largest uses were high-temperature refractory applica-
tions—dead-burned dolomite and refractory magnesia. This is not a
likely market for the spent stone sorbent from the fluidized bed combus-
tion process In the nonregenerative versions. The effluent sorbent would
in these cases have a high sulfate content which could not be retained
at refractory use-temperatures. The regenerative versions with low cal-
cium sulfate loading, however, might be able to supply stone to this mar-
ket. At 330 on-stream days per year, a 500 MW plant would produce on the
order of 25,000 Mg (27,557 tons) of magnesium per year in various forms
(fully-calcined dolomite, half-calcined dolomite, and so on) in the re-
generative process, and twice that quantity in the nonregenerative ver-
sions. It appears that the potential market for refractory forms of mag-
nesia and dolomite could be large enough to accommodate the output from
a substantial number of 500 MW plants.
If the spent sorbent cannot be used as dead-burned dolomite,
the market size shrinks substantially. In this case new, large-scale
uses would have to be developed.
Other potential markets-exist for the spent sorbent. A com-
prehensive assessment of the alternatives is required, followed by tests
with actual spent sorbent for the preferred options.
261
-------
TABLE F-5
DOMESTIC USAGE OF MAGNESIUM COMPOUNDS, 19683
[End use/% | Mg | M$ | $/Mg
Dead-burned Dolomite 1,664,000 31.63 19
Refractory Magnesia 600,000 44.54 74
Other Magnesia 122,000 12.23 100
Pulp and paper 12
Chemical processing 11
Oxychloride and oxysulfate cement 10
Rayon 10
Rubber 8
Fertilizer 6
Insulation - 80% MgO 1
Otherb 42
100
Magnesium Hydroxide 60,800 2.48 41
Magnesium Chlorides 357,000 27.5 76
aU. S. Bureau of Mines. Minerals Yearbook. 1968. p. 667-681.
Including electrical and medicinal uses; and in flux, ceramic, glass,
sugar, animal feed, fuel additives, water treatment, and uranium process-
ing.
262
-------
DISPOSITION OF UNPROCESSED SPENT SORBENT
The disposition of unprocessed spent sorbent is represented by
options 1 and 3 in Figure F-l. Westinghouse recently obtained samples of
spent dolomite from the Argonne and Exxon fluidized bed combustion pilot
plants and carried out preliminary leaching experiments and activity
tests.
Results from these preliminary tests indicate:
• In both the leaching-time experiment and the stone-load-
ing experiment calcium and sulfate dissolution plateaued
at concentrations limited to the calcium sulfate solu-
bility.
• The equilibrium calcium and sulfate concentrations
were high, exceeding the water quality criteria.
Since calcium sulfate occurs abundantly in nature as
gypsum, leachates induced from a natural gypsum
offers a good reference for Its calcium and sulfate
concentrations. Iowa ground gypsum No. 114 was
selected to undergo parallel leaching tests with the
ANL waste stone. Results indicated that gypsum
leachates contained approximately the same amounts of
dissolved calcium and sulfate ions as the ANL leach-
ates. Both agreed relatively well with the calcium
sulfate solubility, and both exceeded the water
quality standards, 75 mg/1 for calcium and 250 mg/1
for sulfate.
• There was negligible dissolution of magnesium ions.
e Insignificant amounts of heavy metal ions were found
in the leachates.
• ANL leachates were alkaline, with pH = 10.6 to 12.1.
It is interesting to note, however, that the run-off
leachates showed a gradual decrease in pH with the
amount of leachates passing through.
The experimental data and an assessment of the environmental impact of
direct disposal are presented in Appendix G.
263
-------
The spent sorbent characteristics (size and composition) may
make the stone attractive for utilization. Areas of particular interest
include soil stabilization and landfill. In a general sense, soil stabi-
lization includes any treatment of soil whereby it is made more stable.
It is well established for conventional soils that such properties as
strength, stiffness, compressibility, permeability, workability.suscepti-
bility to frost and sensitivity to changes in moisture content may be
altered by various methods. Such methods range from simple compaction to
expansive techniques for grouting, drainage, waterproofing, and
strengthening of material by thermal means. Locally available road-bed
soils are stabilized commonly by the addition of agents such as Portland
cement, lime, and.lime/fly ash mixtures. They act by forming cementiti-
ous compounds that more or less permanently bond together individual par-
ticles or aggregates of soil.
Tests are required to determine what the characteristics of
various types of soils are when they are blended with the spent sorbent.
Specifically, tests to determine the unconflned compressive strength, the
Atterberg limits, and direct and triaxial shear strength of soils blended
with various amounts of the waste materials should be performed. These
tests would provide sufficient information to enable specification of
soil-waste material mixtures which can be used as load-bearing materials
in highway foundation and embankment structures. Wherever necessary,
these mixtures of soil and waste material may be further blended with
conventional cementitious materials such as Type 1 Portland cement.
DISPOSITION OF PROCESSED SPENT SORBENT
The direct disposal or utilization of spent sorbent may not be
possible or permitted in all cases. Thus, alternatives for spent stone
disposition must be developed to permit utilization of the fluidized bed
combustion process. It may also be possible to develop a more attractive
use for the spent sorbent through some processing technology.
Several processes are proposed, as summarized in Figure F-2.
The direct disposal options discussed in the previous section and the
final disposition of the sulfur in the regenerative processes are also
264
-------
Regenerative
Fully Calcined
Stone
Half-Cak
Stone
High -Temperature
Regeneration
2 Acid
Low -Ten
Regen
Plant
npprature 2 m Sulfur
eration Plant
( Low Sulfate Content)
Direct
Disposal
RDF-1
I
CaO- MgO
Non regenerative
Direct
Disposal
NDF-1
Pressurized Dust
Combustion Removal
V
Regenerative \
Silica Dead-burninc
Sintering
RDF-2 RDF-3
/ * *
4caOln- Si02 MgO CaO- MgO
1 High Sulfate Content)
Nonregenerative
•J Cooling
__ High-Temperature
Regeneration
^_
Silica Dead-burning
Sintering
NDF-2 NDF-3
^ e. .!«..,
Water
Quench
I
Steam Disposal Wet
Recarbonation Carbonation
RDF-4 RDF-1 RWF-2
* * 1
CaC03- MgO Cat OH) 2- MgO CaCOj • MgO
. Water
Quench
t
Steam Disposal
Recarbonation
NDF-4 NWF-1
Wet
Carbonation
NWF-2
Cool
nn
M Low Temperature „ Aci
Regeneration Plar
Sulf
Pla
Dkpotal
NDH-1
L Dkpo«;al
RDH-1
d
it
jr
High Su
' (CaC03
fc Low Sul
(CaC03
^ Su If uric
^ <-..n.._
nt
Acid
Wet
Sulfation
RWF-3
Sulfite
Leaching
RWF^J
i f
CaS04 • MgO CaCOj &
Mg ( HSo, -
[
Wet
Sulfation
NWF-3
sposal
fur Stone
• MgO)
ur Stone
• MgO)
:Acid
Figure F-2-Pressurized fluidizedbed combustion: spent stone disposal processes
-------
shown. The basic option for dolomite is to operate either half-calcined
or fully-calcined. In each of these modes it is possible to operate re-
generatively or nonregeneratively. Further, there are two options on
regeneration: a one-step, high-temperature process, and a two-step, low-
temperature process. The former leads to sulfur dioxide, which is prob-
ably best routed to an acid plant but can be converted to sulfur. The
latter releases hydrogen sulfide, which is convertible to elemental
sulfur or acid. In reading the figure, for each of the options on the
respective regenerative branches one must include one of the two sulfur
processing alternatives.
The figure is expressed in terms of dolomite, but similar con-
siderations apply if limestone is used. This figure focuses on the pro-
cessing of the unsulfated fraction of the stone.
In all cases coal ash is assumed to be taken overhead and re-
moved from the pressurized hot flue gas stream by particulate removal
system before sending the gas on to power recovery turbines. The various
process options shown on Figure F-2 are given a letter code for conven-
ience .
First letter - R for regenerative
N for nonregenerative (once-through)
Second letter - D for dry process
W for wet process
Third letter - H for half-calcination (retention of CaCO )
F for full-calcination (formation of CaO)
Where compositions are shown, only the form of the residual excess calcium
compound is given. In every case calcium sulfate is also present as cal-
cium sulfate/magnesium oxide. Brief descriptions of the process options
follow.
Disposal: OptionHDH-1
This is a simple method for stone disposal, involving only cool-
ing of the stone, since the composition calcium sulfate, calcium carbonate,
and inerts is expected to be environmentally inert. While landfill is the
most likely end use, it is possible that special cements may be made from
266
-------
it. As a nonregenerative.process, all the sulfur captured is removed from
the process as calcium sulfate. No sulfur recovery facilities are re-
quired .
Disposal; Option RJH-1
This is also a simple stone disposal process, needing only
stone cooling. Sulfur processing facilities are needed since it is a re-
generative process. The type depends on which regeneration process is
used. The final spent stone may have a low sulfur content. This depends
on the concentration of calcium sulfate specified for the regeneration
process. Spent stone having a low calcium sulfate/magnesium oxide con-
centration, say 5 wt %, may be a better candidate for use in the cement
industry than the stone from NDH-1.
Direct Disposal; Options RDF-1 and NDF-1
The bulk of the processes proposed deal with fully calcined
stone. They fall into two groups according to whether they are regenera-
tive or once-through. It is convenient to discuss the counterparts to-
gether.
Both RDF-1 and NDF-1 will have stone cooling. The ability to
use this material as landfill will depend on the content and the
reactivity of the stone. RDF-1 product might be marketable where an in-
expensive alkali is needed, as in municipal waste treatment or acid mine
drainage treatment. NDF-1 product has only 33 percent calcium oxide/
magnesium oxide, which probably makes this process less attractive than
RDF-1.
Silica Sintering; Options RDF-2 and NDF-2
To take advantage of the free calcium oxide in the spent stone,
recovered coal ash, which is expected to be high in silica, is blended
with the stone and sintered at temperatures perhaps as high as 1500°C
(2732°F). The conditions would be a compromise between what is required
for sintering and what must be avoided to prevent loss of sulfur dioxide
267
-------
from the sulfate present. It is likely that a cooling and a grinding
step for the spent stone prior to blending will be required.
If cements cannot be prepared from the spent stone, it would
appear that whatever pozzolanic activity is present would result in a
material that would behave as a stable soil in a landfill operation.
Dead-Burning; Options RDF-3 and NDF-3
Exposure of materials like calcium oxide and calcium sulfate
for sufficient lengths of time to elevated temperatures reduces their
surface area and, hence, their general chemical reactivity. Because of
the presence of calcium sulfate, the dead-burned product probably could
not be used as a refractory. The material, however, is expected to have
a minimal impact on the environment if used for landfill because of the
reduced activity of calcium and magnesium compounds.
Steam Recarbonation; Options RDF-4 and NDF-4
In both processes the stone is cooled to an intermediate tem-
perature, say 500 to 700°C (932 to 1292°F), pulverized with air in a jet
mill, and then contacted with flue gas and steam to convert calcium oxide
to calcium carbonate. The steam increases the reaction rate. This
provides a method for rending the sorbent inactive if the calcium oxide
activity prohibits direct disposal.
Disposal; Options RHF-1 and NWF-1
These are simple wet processes in which the calcium oxide is
merely hydrated to calcium hydroxide to eliminate any problems in handling
calcium oxide. The treated stone may be usable as an inexpensive alkali,
Wet Carbonation; Options RWF-2 and NWF-2
Spent stone is quenched with water and subjected to a three-
stage carbonation using recycled flue gas. The object is to convert the
free calcium oxide to the less active carbonate form to facilitate dis-
posal.
268
-------
Wet Sulfation: Options SWF-3 and NWP-3
An alternative to carbonation is to use sulfuric acid to
convert the free calcium oxide to calcium sulfate. Since magnesium
oxide is soluble in acids, it may not he oossible to avoid forming
magnesium sulfate. Gynsum and eosom salts are oossible end nroducts.
Sulfite Leaching; Option RWF-4
A Canadian oatent (639,443) covers a process being used by the
Aluminum Company of Canada to recover magnesium from dolomitic ores.
It features the use of carbon dioxide as well as sulfur dioxide to
achieve a 92 percent recovery of magnesia, while holding the solution
of calcium to less than 3 percent in the magnesium solution. Sulfur
dioxide from the regenerator, fortified with carbon dioxide, might
be used to produce magnesium bisulfite, usable in the naper industry.
Alternatively, the bisulfite could be calcined to recycle the sulfur
dioxide and produce pure magnesia. Here, sulfur recovery facilities
would also be required. A nonregenerative version is not included
because a ready source of sulfur dioxide is lacking.
SUMMARY
Solids from fluidlzed bed combustion plants will vary, depending
on how the svstem is operated. Spent solids include fuel ash and dry,
granular, spent sulfur sorbent. Spent sorbent composition with limestone
or dolomite mav consist of
« CaSO, and CaCO.
• CaSO, and CaO
• MgO • CaSO, and MgO • CaCO^
• MgO • CaSO, and MgO • CaO.
The sorbent material may be disposed of in its partially utilized
form in a once-through sorbent system or regenerated for reuse in the
fluidized bed combustor. Available pilot-plant test data show that cal-
cium to sulfur (Ca/S) ratios of approximately 2 or greater are required
269
-------
to achieve 90 percent sulfur removal for a once-through dolomite system.
Recent tests carried out at Westinghouse indicate the stone requirement
for a once-through system can be significantly reduced—1.2 Ca/S ratio
for 90 percent sulfur removal. Regeneration of the spent stone has the
potential to reduce stone requirements further and, if the stone (or al-
ternate sorbent) can be reconstituted, virtually to eliminate the need
for makeup sorbent.
Direct disposal of the spent sorbent is attractive for first-
generation plants. Results from preliminary tests indicate the disposi-
tion of the spent sorbent from fluidized bed combustors will not cause
water pollution problems. Preliminary activity tests indicate the tem-
perature increase of the spent stone will be negligible when subjected to
the environment. Extensive tests must be carried out, however. The
chemical composition of the spent sorbent from once-through and regenera-
tive processes and its likely environmental impact require comprehensive
leaching tests and activity tests to determine the chemical fate of the
constituent compounds (calcium, magnesium, sulfate ion, and so on) and
trace elements which may occur in the raw sorbent or which may accumulate
during the combustion of the coal.
Several potential applications for the processed spent sorbent
were identified:
• Soil stabilization
• Landfill
• Concrete
• Refractory brick
• Gypsum
• Municipal waste treatment
• Acid mine drainage.
Both high-temperature stone processing, including spent sorbent/fly ash
and spent sorbent/clay sintering, and low-temperature pozzolanic
A pozzolan is defined as a siliceous or siliceous and aluminous material
which in itself possesses little or no cementitious value but which will,
in the finely divided form and in the presence of moisture, chemically
react with calcium hydroxide (lime) at ordinary temperature to form com-
pounds possessing cementitious properties.
270
-------
activity require further study. Materials to be tested should include
the spent stone from the once-through process, spent stone from pilot
plant and prototype plant tests, and blends of these materials with
fly ash, clay, and soil. The soil stabilization tests will require
determination of unconfined compressive strength, Atterberg limits, and
direct and triaxial shear strengths. The influence on agricultural
soils of surface dumping of the spent sorbent should also be
investigated. For suitability as aggregate material, spent sorbent
should be evaluated by testing it in concrete mixes for compressive
strength, splitting tensile strength, flexural strength, and modulus of
elasticity. The interfacial behavior of the concrete and foundation
soil should be examined.
In summary, environmental problems associated with disposal of
the spent sorbents from fluidized bed combustion systems differ
favorably from those associated with disposal of lime sludges, in that
they are solids and they do not possess great water solubility. Data
indicate the spent dolomite (or limestone) can be used as dry landfill
with known civil engineering practices for controlling structural
rigidity and ground water flows. Alternatives are also available for
utilization of the spent stone. In addition, the advanced sulfur
removal systems being developed would minimize the quantity of spent
stone available and, thus, could minimize the problem.
271
-------
REFERENCES
1. Keairns, D. L. et al. Evaluation of the Fluidized Bed Combustion
Process, Vols. I and II. Environmental Protection Agency.
Westinghouse Research Laboratories. Pittsburgh, Pennsylvania.
EPA-650/2-73-048a and b. December 1973. NTIS PB 231 162/9, 231 163/70.
2. Archer, D. H. et al. Evaluation of the Fluidized Bed Combustion
Process. Office of Air Pollution. WestLnghouse Research
Laboratories. Pittsburgh, Pennsylvania. NTIS PB 211-494, 212-916,
213-152. November 1971.
3. Consolidated Coal Company.
4. Brunner, D. R.,and D. J. Keller. Sanitary Landfill Design and
Operation. Environmental Protection Agency. 1972.
5. U. S. Bureau of Mines. Minerals Yearbook. Vols. I and II. 1968.
p. 669-681.
6. Pressurized Fluidized Bed Combustion. Office of Coal Research.
National Research Development Corporation. England. Report No. 85,
Interim No. 1. 1974.
272
-------
APPENDIX G
SPENT STONE DISPOSAL -
ASSESSMENT OF ENVIRONMENTAL IMPACT
-------
APPENDIX G
SPENT STONE DISPOSAL-ASSESSMENT OF ENVIRONMENTAL IMPACT
The oressurized fluid bed combustion process results in the
production of partially utilized dolomite or limestone material in the
form of calcium sulfate (CaSO.). This sorbent material may be regenerated
for recycling to the fluid bed boiler for repeated sulfur dioxide (SO.)
removal or disposed of in its oartially utilized form in a once-through
system. Although the regenerative process has the advantage of decreased
solid waste for disposition, there is still much uncertainty about it.
In either case, the spent stone is composed of magnesium oxide (MgO),
calcium sulfate, calcium oxide (CaO) (or calcium carbonate [CaCO.]) when
dolomite is used, and calcium sulfate, calcium oxide (or calcium carbonate)
when limestone is used. The waste stone has particle size ranges of
6.4 mm (1/4 in) and down. Factors affecting the final composition include
composition of the fuel, composition of the limestone or dolomite, and
operating conditions. The spent sorbent may be utilized as road-base
aggregates, concrete and block fillers, neutralizing agents for acid mine
drainage, and so on; or marketed for chemical recovery — sulfur recovery
or magnesium oxide extraction. Appendix F deals more extensively with
this subject. In this section, the environmental impact of the spent stone
when it is dumped or used as a landfill is assessed.
CHEMICAL STABILITY2'3
Of the above compounds, calcium sulfate and calcium carbonate
are most likely to be environmentally stable and suitable for direct
disposal without further processing, as the abundance of naturally
occurring limestone and gypsum deposits will attest. Calcium oxide will
hydrate readily to form calcium hydroxide (CatOH].) with release of heat
(64.041 J/gm-mole [15,300 cal/gm-mole]) on contact with water unless it
273
-------
is dead-burned. Recarbonation of calcium oxide will also occur with
carbon dioxide in the ambient air in the presence of water or water vapor.
Magnesium oxide is virtually insoluble (0.0086 gin) and does
not hydrate under atmospheric pressure except in the case of commercially
prepared reactive grades. When most types of dolomite quicklimes are
hydrated under atmospheric conditions, all calcium oxide components readily
hydrate, but very little of magnesium oxide slakes. As a result, high-
calcium quicklime .slakes much more readily than does dolomitic, which
usually requires pressure or long retention periods for complete hydration
because of its hard-burned magnesium oxide component.
Magnesium oxide occurs infrequently in nature as the mineral
periclase; it is also the end product of the thermal decomposition of
numerous magnesium compounds. The physical and chemical properties of
magnesium oxides vary greatly with the nature of the initial material,
time and temperature of calcination, and trace impurities. The increase
in density which results from increasing calcination temperature is
paralleled by a decrease in reactivity. Magnesium oxide, prepared in
the temoerature range of 400 to 900°C (752 to 1652°F) from magnesium
hydroxide or basic magnesium carbonate,is readily soluble in dilute acids
and hydrates rapidly, even in cold water. Fused magnesium oxides are
virtually insoluble in concentrated acids and are indifferent to water
unless very finely pulverized. The oxides prepared below 900°C (1652°F),
which are easily h>drated with water, are known as caustic-burned magnesias.
The unreactive magnesias, prepared at higher temperatures, are called dead-
burned or sintered magnesias. The hydration rate of various magnesium
oxides is determined by the active surface area and may vary from a few
hours, in the case of the reactive oxides obtained at low temperatures,
to months or even years for the dead-burned grades.
It is generally recognized that the dissociation temperature
for calcite is 898°C (1648°F) under atmospheric pressure in a 100 percent
carbon dioxide atmosphere. The temperature of dolomite, however, is not
nearly as explicit. Magnesium carbonate (MgCO,) dissociates at a much
lower temperature — 402 to 480°C (756 to 896°P). Since the proportion
274
-------
of magnesium carbonate to calcium carbonate differs in the many species
of dolomites, the dissociation temperature also varies. Differences in
the crystallinity of stone also appear to add to the disparity of data.
The magnesium carbonate component of dolomite decomposes at higher tem-
peratures than do natural magnesites. A good average value for complete
dissociation of magnesium carbonate in dolomite at 760 mm pressure in a
100 percent carbon dioxide atmosphere is 725°C (1337°F); the calcium car-
bonate component of dolomite would, of course, adhere to the above higher
value representing dual-stage decomposition. As a result of these dif-
ferences in dissociation points, the magnesium oxide is usually hard-
burned before the calcium oxide is formed.
From the above discussion it follows that there may be some
question about the environmental stability of the magnesium oxide compo-
nent of the spent sorbent mixtures calcium carbonate/calcium sulfate/
magnesium oxide. With the fluid bed boiler conditions approaching 955°C
(1750°F) and 1013 kPa (10 atm), it is expected that the magnesium oxide
component is hard-burned and, therefore, suitable for disposal without
further processing. Activity and leaching tests have been performed on
the waste stone from the Argonne National Laboratory (ANL) 152.4 mm
(6 in) fluid bed boiler and the Exxon Research and Engineering (Exxon)
batch unit and miniplant and will be discussed in a later section of this
appendix.
ENVIRONMENTAL IMPACT
The environmental impact of any disposed material is a function
of its physical and chemical properties and of the quantity involved.
Potential water pollution problems can, in many cases, be predicted by
chemical properties such as solubility, the presence of toxic metals,
and the pH of leachates. Disposal of the spent stone from the fluid bed
combustion system may create air pollution or an odor nuisance such as
hydrogen sulfide, depending on the amount of calcium sulfide present,
although it is not expected in significant quantities. Heat may be
275
-------
released on hydration of calcium oxide when the calcium in the spent stone
exists in the calcined state if the combustion temperature is high enough
to produce fully calcined dolomite.
The first consideration when looking at potential water
pollution from the solid waste disposal is the volume of leachate that
will be produced. This is a direct function of the amount of water
reaching the landfill. There are two possible sources of this water:
rainfall and naturally occurring subsurface flow through the landfill
site. Subsurface flow is a natural phenomenon which can seriously inter-
fere with safe operation of landfills in two ways. First, it is a source
qf additional, potentially harmful leachate. Second, it can serve as a
direct means of groundwater contamination. Prevention can be effected
by a thorough geological study of the site beforehand and, if needed,
installation of rerouting devices for the groundwater flow. In a similar
vein, coverage of the landfill area when complete will greatly reduce, if
not eliminate, the amount of leachate produced.
In order to predict leachate characteristics of a landfill, it
is first necessary to describe the general features of water movement and
geological considerations for this disposal method. Due to the recent
surge of ecological interest in sanitary landfills for solid waste disposal,
there is an abundance of information available. Emrich's review of research
4 5
in this field presents an overall perspective of progress on this subject. '
Research is being conducted to define and solve this problem, but re-suits
6
available to date are not sufficient to assess it fully.
EXPERIMENTS
Leaching experiments and activity tests were performed in order
to assess the potential environmental impact of the spent stone from the
pressurized fluid bed combustion process and its suitability for disposal as
a landfill material. The samples used in these experiments were the spent
sorbents from the ANL and Exxon pressurized fluid bed combustion pilot
plants (partially sulfated Tymochtee dolomite from ANL run C2 and C3 and
0
partially sulfated Grove limestone 1359 from Exxon run 8.4 ). As calcium
276
-------
sulfate was a major constituent of the waste stone from the pressurized
fluid bed combustion processes, a naturally occurring calcium sulfate
(ground gypsum 114 of -20 mesh from Fort Dodge, Iowa) was selected
to undergo similar leaching conditions for comparative purposes. Table G-l
summarizes the chemical compositions of the ANL and Exxon spent stones
as well as the Iowa gypsum.
Table G-l
CHEMICAL COMPOSITIONS OF SPENT STONE FROM ANL AND EXXON
PRESSURIZED FLUID BED COMBUSTION PILOT PLANTS
AND IOWA GROUND GYPSUM 114
Composition. % | ANL spent stone | Exxon spent stone I Gypsum 114
57 26 74.0
CaC03 9 58 1.8
CaO 2 7.6
CaS <0.05 <0.05
MgO 20 0.8 0.2
H20 (combined) — — 19.0
Others 12 7.6 5.0
Leaching Tests
Procedures
A series of leaching experiments was designed to study leachate
characteristics as functions of the varying parameters and procedures to
induce leachates, as follows:
1. Mixing time - 250 ml of deionized water was mixed with
25 gm of waste stone in a 500 ml Erlenmeyer flask. The
mixture was agitated for various lengths of time using an
automatic shaker (Eberback) at 70 excursions per minute
and room temperature. The supernatant resulting from this
operation was passed through a Whatman No. 42 filter. The
filtrate was used for determination of pH, specific
277
-------
Results
conductance, calcium, magnesium, sulfate, sulfide, and
trace metal concentrations.
2. Stone load - 250 ml of deionized water was mixed with
different amounts of spent stone and shaken for 24 hours.
The supernatant was filtered and analyzed.
3. Mixing mode - shaking versus nonshaking: 250 ml of de-
ionized water was mixed with 25 gm of waste stone. The
mixture was allowed to sit for 24 hours at room tempera-
ture with and without shaking. The supernatant was fil-
tered and analyzed.
4. Sample compaction - 25 gm of sample stone, either in its
original size or ground to fine powder, was isostatically
pressed at 68,950 to 344,750 kPa (10,000 to 50,000 psi)
into pellet form and then immersed in 250 ml for 24 hours
for nonshaking leaching time. The supernatant was fil-
tered and analyzed.
5. Run-off tests - Deionized water was dripped at a constant
rate onto 25 gm ANL waste stone which was packed manually
in a cylindrical column of 11 mm (0.43 in) diameter and
213 mm (8.375 in) height. Successive 250 ml leachates
were collected and monitored for pH, specific conduc-
tance, calcium, magnesium, and sulfate.
Table G-2 summarizes the chemical characteristics (pH, specific
conductance, calcium, magnesium, sulfate, sulfide, and trace metal concen-
trations) of leachates induced from the ANL waste stone under conditions
corresponding to the severest cases and compares them with leachates from
a natural gypsum and with water quality standards set by the Commonwealth
9
of Pennsylvania, the U. S. Public Health Service, and the World Health
10
Organization.
Table G-3 presents results from leaching tests on the ANL waste
stone and Iowa gypsum No. 114 using the procedures described previously.
278
-------
Table G-2
SUMMARY OF LEACHATE CHARACTERISTICS
Species3
Stone composition (Z)
ANL
waste
stone
Iowa
gypsum
No. 114
Leachates (mg/1)
ANL waste stone
50 em/25O ml/26 hr
Iowa gypsum Ho. 114
50 gm/250 ml/24 hr
Water standards (mg/1)
U. S. Public Health
Service Drinking
l.'ater Standards^
World Health
Organization
Standards for
Potable Water*
Commonwealth of
Pennsylvania Water
Quality Standards0
pH
Specific
Conductance
micromhos/cm
SOi,"
5
C03-
Ca
«g
Al
Ag
B
Bi
Be
Co
Cr
Cu
Fe
K
Li
Mn
Ho
Na
Hi
Pb
Sb
Si
Sn
Tl
V
Zn
Zr
Sr
40
< 0.05
5.1
22.5
11.7
5
< 0.001
0.03
< 0.002
0.005
0.001
3
0.2
0.01
0.05
0.002
0.1
0.002
< 0.01
< 0.005
> 10
< 0.005
0.08
0.002
< 0.02
< 0.01
52
1.1
22.2
0.14
0.1
< 0.01
< 0.03
< 0.01
< 0.001
< 0.01
< 0.01
< 0.01
0.3
0.1
< 0.01
< 0.01
< 0.01
0.1
< 0.01
< 0.03
3
< 0.01
0.03
< 0.01
< 0.01
< 0.03
0.1
12.1
3850
1280
744
. 0.03
NDd< 0.1
ND < 0.05
0.03
ND < 0.1
ND < 0.01
ND < 0.1
ND < 0.1
ND < 0.1
< 0.1
0.3
0.2
ND < 0.05
ND < 0.03
6
ND < 0.1
ND < 0.1
ND < 0.3
0.2
ND < 0.1
ND < 0.05
ND < 0.05
ND < 0.4
ND < 0.05
0.3
7.4
2140
1465
607
0.5
ND < 0.1
ND < 0.05
0.3
ND 0.1
ND 0.01
ND 0.1
ND 0.1
< 0.1
< 0.1
0.5
< 0.1
< 0.05
ND < 0.1
0.5
ND < 0.1
ND < 0.2
0.4
ND < 0.1
ND < 0.1
ND < 0.05
ND < 0.4
ND < 0.05
1
7.0 to 8.5 6.0 to 8.5
250 200 250
75
50
0.05
1.0
1.0
1.0 1.0 0.1
0.3 0.3 0.3
0.05 0.1 1.0
2.0
0.05 0.1
1.0
5.0 5.0 0.05
All cations except Ca and Mg are determined by emission spectrochemical method, 1/3 to 3x estimates.
b
Lund. H. F. Ed. Industrial Pollution Control Handbook. New York. McGraw-Hill Book Co. 1971.
Pennsylvania Department of Environmental Resources. Water Quality Standards Summary.
Harrisburg, Pennsylvania 17120. Document No. 42-006.
TJD - Not detected.
-------
Table G-3
SUMMARY OF LEACHATE CHARACTERISTICS
10
g
Experimental
Stone
weight
(gm)
1
2.5
5
10
25
50
25
25
25
25
25
25
25
25
25
258
25
25
...
H20
volume
(ml)
250
250
250
250
250
250
250
250
250
250
250
250
250
250
250
250
250
250
250
250
250
250
250
250
50
50
Mixing
time
(hr)
24
24
24
24
24
24
1
3
6
17
24
48
96
24
24
24
1st run-off
2nd run-off
3rd run-off
4th run-off
5th run-off
6th run-off
7th run-off
8th run-off
1st run-off
2nd run-off
parameters
Shaking
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Ho
No
Rate
5 ml/mln
5 ml/mln
5 ml/mln
5 ml/mln
5 ml/mln
5 ml/mln
5 ml/mln
5 ml/mln
0.24 ml/mln
0.24 ml/min
Procedure
(referred
to text
description)
2
2
2
2
2
2
1
1
1
1
1
1
1
3
3
4
5
5
5
5
5
5
5
5
5
5
Leachate results
ANL spent stone
PH
11.2
11.4
11.8
11.9
12.1
11.8
11.9
11.9
12.0
11.9
11.9
11.9
11.9
12.1
11.9
11.8
11.5
11.3
11.3
11.2
11.2
11.2
10.6
™
Specific
conductance
(micromho/cm)
1-260
1590
2210
3100
3850
1290
1810
1860
2280
3100
3475
3980
3100
1940
2300
2180
1960
1920
1980
1870
1780
1760
1640
-—-
Ca M
(mg/1) (mg
g so,,
/I) (mg/1) P
Iowa P.VDS
Specific
H conductance
(micromho/cm)
urn No. 114
Ca M
(mg/1) (mg
g so,,
/I) (mg/1)
304 < 0.2 690
397 < 0.2 883
7
508 < 0.2 1048 7
870 < 0.2 1725 7
744 < 0.2 1280 7
226
326
330
456
870
908
916
870
306
305
491
510
545
500
483
463
460
~
422 -
576 -
576 7
865 -
1725 7
1840 7
1830 7
1725 7
480 7
_.
1055 -
1130 -
1240 -
1240 -
1130 -
1110 -
1110 -
1055 -
" 7
7
.4 2000
.4 2100
.4 2150
.4 2140
. 3 2040
.4 2150
.4 2150
.4 2150
.4 2150
.3 1910
—
._ _ —
— ___
._ -_
— —
4 2250
1 2160
570
603
615
607
590
615
610
615
615
530
___
__
— __
— __
__ —
— —
620
615
1380
1450
1470
1465
— —
-
1500
_ ___
1470
1470
1470
1470
1270
-
— — «_
— — _
_ _ __
_
— — ._
— •••
— «_
1490
1440
Pellet pressed at 10,000 psi.
-------
Calcium and sulfate concentration, pH, and specific conductance of the
leachates were shown as functions of batch mixing time, between 1 and
96 hours, in Figure G-l. Leaching Procedure 1 was used for inducing the
leachates. In Figure G-2 calcium, sulfate, pH, and specific conductance
were plotted as functions of stone loading from 1 to 50 gm of sample
stone in 250 ml deionized water using Procedure 2. Characteristics of
leachates from a natural gypsum induced under identical conditions are
also shown in Figures G-l and G-2 for comparison. A shaking versus non-
shaking mode of leaching was studied using Procedure 3 and results sum-
marized in Table G-3. Procedure 4 was used in an attempt to compact
sample stone into dense pellets so that water permeability might be re-
duced, but the pellets crumbled on contact with water and resulted in
leachates similar to those from the unpressed samples. Finally, a run-
off test was carried out on ANL waste stone using Procedure 5. Figure G-3
shows the calcium concentration, sulfate, pH, and specific conductance
for successive 250 ml leachates collected. Such data would be useful in
cases where stone washing before disposal or leachate treatment was
needed.
Conclusions
Results from leaching tests on ANL spent stone from the pres-
surized fluid bed combustion process indicated that:
• In both the leaching-time experiment (Figure G-l) and the
stone-loading experiment (Figure G-2), calcium and sul-
fate dissolution plateaued at concentrations limited by
the calcium sulfate solubility.
• The equilibrium calcium and sulfate concentrations were
high, exceeding the water quality criteria. Since cal-
cium sulfate occurs abundantly in nature as gypsum,
leachates induced from a natural gypsum offers a good
reference for its calcium and sulfate concentrations.
Iowa ground gypsum No. 114 was selected to undergo paral-
lel leaching tests with the ANL waste stone. Results
281
-------
cium Concentration,
i— i >— '
3808
J^
1 1 1 1 1 1 1 1 1 1
m
0
|,2000
c-1500
0
|iooo
1 500
^
&
o* 0
00
12
11
10
9
8
7
6
—
-
l-l m i-*
—
1 1 1 1 1 1 I 1 1 1
4000
| _ 3000
|-B2000
o -c
o E
o S1000
tte* O
^j C
a\ Cl
a."" n
oo U
I
I
I
I
0 10 20 30 40 50 60 70 80 90 100
Mixing Time (hours)
Figure G-1-Leachate characteristics as functions of batch mixing time for
o Argonne spent stone
o Iowa gypsum #114
282
-------
I I I I I I I I I I
92000
iiooo
•s
I 500
o
<3 0
i i i i i i i i i i
£
It.
11
10
9
8
A
—
—
—
1 1 1 1 1 1 1 1 1 1
c
•§
o
_o
H
&
|3000
12000
11000
I I I I I I I I
0 5 10 15 20 25 30 35 40 45 50
Stone Loading (grams in 250and H20)
Figure G-2-Leachate characteristics as functions of stone loading for
o Artjonne spent stone
o Iowa gypsum #114
283
-------
fisoo
a Concentration
" S
So
o
—
—
0^^"^ ^3 — 0
1 1 1 1 1 1 i
,2000
§1500
C
•£1000
S
J 500
o
«/>
1 1
1 1
12
11
10
9
8
7
i i
12 e
|| 2000
o E
" S 1000
.ii — n
1
1
1
I
234567
Numbers of 250 ml Samples
Figure G-3-Leachate characteristics of the Argonne spent
stone leachates induced by the run-off tests
284
-------
indicated that gypsum leachates contained approximately
the same amounts of dissolved calcium and sulfate ions
as the ANL leachates. Both agreed relatively well with
the calcium sulfate solubility,and both exceeded the
water quality standards, 75 mg/1 for calcium and 250 mg/1
for sulfate.
• There was negligible dissolution of magnesium ions.
• Insignificant amounts of heavy metal ions were found in
the leachates.
• ANL leachates were alkaline, with pH = 10.6 to 12.1. It
is interesting to note, however, that the run-off leach-
ates showed a gradual decrease in pH with the amount of
leachates passing through.
Results from the leaching experiments on ANL spent stone agreed
well with those reported by the British Coal Utilization Research
11 12
Association (BCURA) and Pope, Evans and Robbins (PER) in that the
leachates had high pH, calcium, and sulfate, and negligible magnesium
ions. The difference between the PER and BCURA findings in the extent of
calcium and sulfate extraction was, in the light of the Westinghouse
stone-loading and mixing-time studies, largely due to the difference in
the stone-to-water ratio and the leaching time of their experiments. The
Westinghouse results also indicated that it is unlikely that the heavy
metal ions in the leachates from pressurized fluid bed combustion pro-
cesses will cause water pollution. The solutions were alkaline, but the
run-off tests showed a gradual decrease in pH with the amount of water
passing through. Superficially, the calcium and sulfate concentrations
in the leachates might suggest the possibility of a water pollution
hazard from these ions, but it must be emphasized that the conditions in
these experiments were much more extreme than those that would exist in
actual cases, where water percolation is minimized. The fact that
equally high calcium sulfate dissolution was found in gypsum leachates
under identical conditions offers a useful comparison. Further tests
using much larger quantities of material and taking into consideration
geocriteria such as topography, geology, hydrology, and soil conditions
285
-------
of the actual landfill site are needed to determine fully the environmen-
tal impact of spent stone disposal.
Activity Test
Heat release experiments were carried out on the following sam-
ples to determine their reactivity toward water:
1. Spent sorbent from ANL pressurized fluidized bed com-
bustion process of run C2/C3 —sulfated Tymochtee dolo-
mite, -14 mesh
2. Tymochtee dolomite, -16 +18 mesh
3. Spent sorbent from Exxon pressurized fluidized bed com-
o
bustion process of run No. 8.4 —sulfated limestone
1359, -8 +80 mesh
4. Limestone 1359, -16 +18 mesh
5. Limestone 1359, -18 +35 mesh
6. Calcined limestone 1359 at 960°C (1760°F) -18 +35 mesh.
In each of the above tests, 3 gm of stone was added to 20 ml
of deionized water in a Dewar flask which had been thermally equilibrated.
Iron-constantan thermocouples were vsed to monitor the temperature rise
in the stone/water system with an Omega cold junction compensator and a
digital voltmeter readout. Table G-4 summarizes the maximum temperature
rise and time required for reaching the temperature. Less than 0.2°C
(0.36°F) temperature rise was found for the above samples 1 to 5 but a
temperature rise of 55°C (99°F) was found for sample 6. The great con-
trast between the calcined limestone and the other samples indicated the
validity of the experiment as well as the lack of reactivity of the ANL
and Exxon spent stones with water. Although it can be safely assumed
that no heat pollution will result if these particular batches of spent
stone are subjected to rainfall after disposal, it must be pointed out
that the activity and heat release properties of the spent sorbent are
functions of the operating conditions of the fluid bed combustion process.
286
-------
Table G-4
SUMMARY OF STONE ACTIVITY TESTS
Samples
Stone/H00TAT ,
L max
max
ANL spent stone
-14 mesh
Tymochtee Dolomite
-16 +18 mesh
Exxon Spent Stone
-8 +80 mesh
Limestone 1359
-16 +18 mesh
Limestone 1359
-18 +35 mesh
Calcined Limestone 1359
-18 +35 mesh
3 gm/20 ml 0.1 (0.18)
3 gm/20 ml 0.1 (0.18)
3 gm/20 ml 0.2 (0.36)
3 gm/20 ml 0.1 (0.18)
3 gm/20 ml 0.1 (0.18)
3 gm/20 ml
55 (99)
5 sec
4 to 20 sec
CONCLUSIONS AND RECOMMENDATIONS
The pressurized fluid bed combustion process results in the
production of spent solids in the form of partially sulfated limestone or
dolomite. The long-term stability and the suitability for its disposal
into the environment have been discussed. Laboratory leaching and acti-
vity tests have been initiated. Preliminary results indicate that it
would probably not cause water and heat pollution. It must be remembered,
however, that the physical and chemical properties of the spent sorbent
are functions of the operating conditions of the pressurized fluid bed
process and that the physical characteristics of the specific disposal
site must be judged individually in evaluating the leaching properties of
the spent stone.
To assess fully the environmental impact of spent sorbent, fur-
ther tests on stone analyses, leaching properties, heat release proper-
ties, landfill properties, and air emission should be performed.
287
-------
REFERENCES
1. Keairns, D. L., et al. Evaluation of the Fluidlzed Bed Combustion
Process. Vol. IV. Office of Research and Develonment. Environmental
Protection Agency. Westlnghouse Research Laboratories. Pittsburgh,
Pa. EPA-650/2-3-73-048d. Contract 68-02-0217. NTIS PB 233 101/15.
December 1973. 322 pages.
2. Standen, A., ed. Kirk-Othmer Encyclopedia of Chemical Technology
John Wiley & Sons, Inc. New York 1967.
3. Boynton, R. S. Chemistry and Technology of Lime and Limestone
John Wiley and Sons, Inc. New York. 1966.
4. Emrich, G. H. "Guidelines for Sanitary Landfills — Ground Water
and Percolation. Compost Science. May, 1972. pp. 12-15.
5. Emrich, G. H., Merrit, G. L., and Rhindress, R. D. Geocriteria for
Solid Waste Disposal Sites (Program for the 5th Annual Meeting,
Northeastern Section, Geological Society of America) 2(1), 1970, p. 17.
6. Phillips, N. P. and Wells, R.M. Solid Waste Disposal. Environmental
Protection Agency. Radian Corporation EPA-650/2-74-003, May 1974.
7. Argonne National Laboratory. Monthly Progress Reports Nos. 70 and 71
ANL-ES/CEM-FO-70 and 71. August and September, 1974.
8. Exxon Research and Engineering Comoany. Monthly Progress Reoort
No. 57, Environmental Protection Agency. Contracts 68-02-1451 and
68-02-1312. November 1974.
9. Pennsylvania Department of Environmental Resources. Water Quality
Standards Summary. Harrisburg, Pennsylvania, 17120. Document No. 42-006.
10. Lund, H. F., ed. Industrial Pollution Control Handbook McGraw-Hill
Book Company. New York. 1971.
11. Pressurized Fluidized Bed Combustion. Office of Coal Research. National
Research and Development Corporation. Contract No. 14-32-0001-1511.
November 1973.
12. Pope Evans and Robbins, Inc. Multicell Fluidized-Bed Boiler Design
Construction and Test Program. Interim Report No. 1. Report PER-570-74.
Contract No. 14-32-0001-1237. August 1974.
288
-------
APPENDIX II
TRACE EMISSIONS AFFECTING GAS-TURBINE PERFORMANCE
H
-------
APPENDIX H
TRACE EMISSIONS AFFECTING GAS-TURBINE PERFORMANCE
SUMMARY
Sodium and potassium compounds are potentially hazardous to the
operation of the gas turbine. Chlorides and hydroxides are volatile
species and can transport sodium and potassium from the combustor to the
turbine. At hydrogen chloride (HC1) levels exceeding 0.4 ppm by volume
in the combustor gas, solid or liquid sodium sulfate (Na.SO.) will con-
vert to gaseous sodium chloride (NaCl). The hydrogen chloride level in
the combustion gas resulting from the complete release of chlorine from a
low-chlorine coal (100 ppm Cl) exceeds this level by over a factor of ten
and is 5 ppm. In a fluid bed combustion process the predominant trans-
port should be by the chlorides.
In the gas turbine, reactions between the chloride and the
sulfur oxides in the combustion gas will form liquid sulfate-chloride
melts on the turbine hardware if the sodium and potassium levels are suf-
ficiently high. These melts must be prevented because they initiate hot-
corrosion and deposit formation. Hydrogen chloride in the turbine acts
to prevent sulfate deposits from forming or, once formed, acts to remove
them. A hydrogen chloride level of 40 ppm by volume in the combustion
gas is sufficient to prevent a liquid sodium sulfate melt from being
stable at sodium concentrations in the gas up to 0.2 ppm by volume.
(Forty ppm hydrogen chloride corresponds to complete release of chlorine
from coal containing 800 ppm by weight of chlorine; 0.2 ppm of sodium
corresponds to a one percent release of sodium from a coal containing
130 ppm by weight of sodium.) The concentration both of hydrogen chloride
and of sulfur oxides (SO. and SO.) have a strong influence on the stabi-
lity of the melt. If the sulfur dioxide level in the gas is dropped from
200 ppm to 100 ppm by volume, the concentration of hydrogen chloride
289
-------
required to prevent deposition of a liquid sodium sulfate film would drop
to about 25 ppm by volume. Three additional factors must be understood
before it will be possible to define the sodium and potassium tolerances
which will prevent hot corrosion attack in the turbine. These are:
• The influence of the interaction between sodium and
potassium to form complex melts on the hardware. In
such melts the activities of the sodium and potassium
are reduced,and the equilibrium concentrations of
sodium and potassium species that can exist in gas
above the melt are also lowered. Interaction tends
to reduce the tolerable concentrations of sodium and
potassium in the turbine expansion gas. Westinghouse
is working to establish the magnitude of this effect
and also to establish the influence of the relative
sodium and potassium levels on the composition and
melting point of stable deposits.
• The ability of turbine stator vane and rotor blade
alloys to withstand a combustion gas containing up to
200 ppm sulfur i.ioxid**. ar.d -ip tc /'O pp»n hydrogen
chloride. Careful experimental studies are needed.
• The shifts in the turbine tolerance which will occur
if equilibrium levels of sulfur trioxide are not
achieved and the degree to which kinetic factors such
as these influence the tolerable concentration of
sodium and potassium chlorides.
In the combustor itself a competition takes place for the sodium
and the potassium. Hydrogen chloride attacks the sodium and potassium
compounds attempting to form the volatile chlorides. On the other hand,
in the burning char particle, silica reacts with sodium to form stable
silicates. Unfortunately, potassium silicates are not stable, and potas-
sium clay mineral compounds are breaking down and converting to either
polysulfides or chlorides, depending on the composition of the local at-
mosphere. In the oxidizing atmospheres outside of combusting particles,
290
-------
the sulfates are stable at low levels of hydrogen chloride ; but at levels
above 0.4 ppm, reconversion of condensed sulfates to chlorides will occur.
Westinghouse is attempting to establish the feasibility of controlling
the alkali-metal content of the gas through:
• Control of the level of hydrogen chloride in the corn-
bus tor. Reducing hydrogen chloride levels in the"corn-
bus tor below 0.4 ppm by volume would lower the sodium
chloride content of the transport gas by slowing or
preventing attack of ash minerals and by promoting
conversion to low-volatility condensed sulfate.
• Conversion of high-volatility sodium and potassium
chlorides (KC1) in coal and dolomites to sulfates or
silicates,with simultaneous removal of hydrogen
chloride by low-temperature pretreating in the coal
and stone dryers and preheaters.
Experimental measurement of the actual release of alkali that
can be translated to gas compositions projections for power plants are
needed. Because release fractions near one percent are anticipated,
measurements of the material released are required for accuracy.
TRACE EMISSIONS
Trace emissions from the fluid bed combustion process are im-
portant from two standpoints:
• Their effect upon the environment
• Their effect on the operability of the power plant.
The primary environmental concern is with the toxic effects of
the compounds of beryllium, mercury, fluorine, lead; cadmium, arsenic,
nickel; copper, zinc, barium, tin, phosphorous, lithium, vanadium,
manganese, chromium, and selenium.
From the standpoint of preventing hot corrosion of turbine
alloys, growth of deposits within the turbine, and corrosion and heat trans-
fer problems in heat recovery boilers, Westinghouse is concerned with the
291
-------
compounds of sodium, potassium, lead, vanadium, calcium, and magnesium.
The latter are of concern primarily as particulates that can contribute
to deposition problems.
Argonne National Laboratory (ANL) working in conjunction with
the Office of Coal Research (OCR), has been studying the release during
fluid bed combustion of those trace contaminants dangerous to the envi-
2
ronment. The Illinois Geological Survey has continued and expanded its
extensive work on the mineralogy of coals and the distribution of trace
3
elements in coal. Westinghouse under this contract with OCR has
been concerned primarily with the release of contaminants potentially
dangerous to the operability of power plants. Other work has been
directed initially to sodium and potassium release since these are con-
sidered the most hazardous. Vanadium release is potentially dangerous,
and the experimentation measuring release studies the composition of
deposits to detect all components transported in the combustion gas.
The Turbine-Corrosion/Deposition Problem
The components of gas turbines exposed to the hot gas stream
are made of materials that form oxide scales to protect themselves from
oxidation. High-strength nickel-based superalloys such as Inconel 738
or Incolloy 700, are used for the highly stressed rotating components.
More oxidation-resistant, but lower-strength, cobalt-based alloys, such as
X-45 or MAR M509, are used for the stationary components. Air cooling is
used to maintain metal temperatures lower than 899°C (1650°F) so that the
alloys retain sufficient mechanical strengths. Inlet gas temperatures
over 1093°C (2000°F) are currently used.
Metal recession rates due to oxidation alone are of the order
of 0.10 mm (0.004 in) per year on the hottest components of current
Westinghouse turbines. In the presence of alkali-metal compounds, which
react with sulfur dioxide and sulfur trioxide from the combusted fuel
gas, liquid films of sulfate and sulfate-chloride mixtures can be depos-
3
ited on the turbine hardware. Figure H-l shows the temperatures at
which liquid deposits form in the sodium-potassium-sulfur-oxygen-chlorine
(Na-K-S-0-Cl) system. The figure shows that liquid films are possible
292
-------
Curve 652781-A
to
VO
OJ
(1623°F) 884°C
628°C
(1472°F) 800°C
Nad
~ 832°C
Nv/
V,
%
\
~ 658°C
1069°C (1956°F)
Phase (X) = Compound
of Na2$04and K2$04 in
ratio of either 3:1 or 2:1
690°C (1274° F)
774°C (1425°F)
KCI
Figure H -1-Projection of liquid us surface of Na2S04- K2S04- KCI -NaCI system showing
temperature contours (after E. K. .Akopov and A. G. Bergman) .3
-------
over the temperature range of 1069 to 514°C (1956 to 957°F), which coin-
cides with the range of temperatures encountered in the turbine flow
path. (Gases are exhausted from turbines at about 400°C [752°F].)
Molten alkali-metal compound films are dangerous because under
some conditions they attack the protective oxide scale, allowing accele-
rated or catastrophic oxidation (hot corrosion) of the turbine compo-
nents. The films can also accumulate particulates that allow the build-up
of thick deposits on the metal surfaces. Aerodynamic performance of the
turbine can be seriously impaired.
Alkali-Metal Components in Coals and Dolomites
Sodium and potassium exist in similar chemical form in coal and
in dolomite, as clay minerals (primarily illite and montmorillonite which
contain both elements) and as sodium and potassium chlorides in saline
ground water filling pores or cracks in coal and rock beds. The propor-
tions of different mineral constituents in coal vary from one site to
another. The degree of saline water entrainment. and the resulting sodium
chloride content of coal and rock depend on the porosity and crack struc-
ture of the material and on the depth from which it is mined. The sali-
nity of the ground water increases with depth. As a rule, strip-mined
coal and rock will, therefore, have a lower content of sodium chloride
than deep-mined material. The high volatility of sodium chloride makes
it likely that coal and rock containing entrained saline water will re-
lease sodium (as the chloride) to the coal gases. Volatilities of
mineral-type alkali compounds are low, lower than those of sodium sul-
fate. From the point of view of alkali-release, reactions of alkali-
4 5
containing clay minerals at coal gasification temperatures ' are
important. When illite, the principle clay mineral contaminant, is
heated in air, dehydration occurs below 600°C (1112°F) and dehydroxyla-
tion is complete at about 800°C (1472°F). ' The illite anhydride struc-
ture formed by these processes begins to decompose at 850°C (1562°F),
forming a spinel-type crystal phase if the illite is not associated with
kaolinite. The amount of spinel phase increases with firing temperature
294
-------
to 1200°C (2192°F) and then disappears gradually above this temperature.
This behavior is attributed to the presence of iron and magnesium and to
a slightly lowered aluminum content in the unassociated (pure) illite
aluminosilicate structure. Mullite (3 Al?^1} " ^ SiO.) appears as a cry-
stal phase only above 1100°C (2012°F). If the illite is associated with
kaolinite (illite fire-clays), mullite begins to crystallize between 900
and 950°C (1652 and 1742°F). Amorphous glassy phases also form. The
formation of mullite below 1000°C (1832°F) is attributed to the potassium
and other alkali and alkali-earth cations present in the illite structure.
The potassium ions are thought to move in to prop the structure open and
prevent the reformation of broken silicon-oxygen bonds.
The glasses that form at the highest temperatures should be re-
sistant to attack by hydrogen chloride and retain alkali-metal compounds
in the ash.
Factors Affecting Release of Alkali-Metal Compounds to the Expansion Gas
In order to estimate the extent of alkali release from mineral
constituents, it is assumed that the glassy phase can be approximated by
sodium disilicate (Na.Si.Oe) • This represents a composition in the
sodium oxide-silicon dioxide (Na.O-SiOO system which is stable in the
presence of an excess of silica, a condition which is typical for coal
ash minerals.
During combustion of char in a dolomite/char/ash or limestone/
char/ash fluidized bed with an average temperature (and exit gas tempera-
ture) of 871°C (1600°F), the temperature of individual burning char par-
ticles will exceed the bed temperature by several hundreds of degrees.
As combustion of char particles proceeds, the surface zones of
these particles convert to a fine powder of mineral matter and residual
carbon which is carried away with the gas stream. This picture is valid
for temperatures below the fusion temperature of the ash.
Figure H-2 is a thermochemical phase diagram for the sodium-
silicon-oxygen-sulfur system at 1093°C (2000°F). The different scales
show interrelated chemical potentials for oxygen and sulfur. Slanted,
295
-------
Curve uS-...: ,
-4
-8
-12
-16
J-zo
a.
f-24
-28
-32
-36
-40
-44
0
I
i
0
-2 -4 -6
-2-4-6
109
log P^
-28 -24 -20 -16 -12 -8 -4 0
I I I I
I I
-2
0
2
8
-------
thin, solid or broken lines show equilibrium sulfur dioxide and sulfur
trioxide pressures. These lines, and the information contained in the
different scales, allow immediate definition of the most important gas-
pnase components at any point of the diagram. Heavy, solid, and broken
lines separate the regions of stability of condensed phases in the
system. The large filled circle indicates the estimated chemistry at the
surface of a burning char particle. The position of this circle in the
diagram corresponds to local equilibrium at a carbon/gas interface. It
takes into account a counterflow mass balance in a boundary diffusion
layer where oxygen, carbon dioxide, and water vapor move towards the par-
ticle to form carbon monoxide and hydrogen at the proper pressures and
pressure ratios, and where the latter reaction products diffuse away from
the char particle before reacting any further. As shown in the diagram,
the filled circle falls in the liquid sodium disilicate field. Thermo-
chemical calculations show that the most important gaseous sodium species
that will be released from liquid sodium disilicate in this environment
are gaseous sodium hydroxide and sodium chloride. The equilibrium gaseous
sodium hydroxide pressure for the reaction
Na2Si20500 + H20(g) J 2 NaOH(g) + 2 Si02(s)
at relevant water vapor pressure is ^ 10 , log P n % -6.4. This
NaOH
corresponds to a concentration of 40 ppb by volume. The formation of
silica as a reaction product may separate the silicate phase from the
gaseous environment. This would tend to slow down the continued reaction
and,thus, the release of sodium from the silicates. Gaseous hydrogen
chloride can react with liquid sodium disilicate to form gaseous sodium
chloride volatile species. The reaction is:
Na2Si205(£) + 2 HCl(g) •* 2 NaCl(g) + 2 Si02(s) + H20(g).
It is thermochemically strongly favored even at trace levels of gaseous
hydrogen chloride and becomes more important as the level increases. The
297
-------
formation of solid silicon dioxide reaction product may again be an im-
portant factor in limiting the extent and rate of this reaction.
As the volatile alkali compounds or species reach the outside
of the diffusion layer and enter the main gas stream, and as fine parti-
cles are carried in that gas stream, further opportunities for particle/
gas interactions occur. Figure H-3 shows a thermochemical phase diagram
for the sodium-silicon-oxygen-sulfur system at 871°C (1600°F). The filled
circle represents the approximate chemistry of the gas stream that leaves
the boiler. The circle is in the sodium sulfate field. Consequently,
liquid sodium disilicate from mineral matter is unstable and reacts to form
silicon oxides and either sodium sulfate or gaseous sodium compounds or
species, if sufficiently volatile and thermochemically favored. For the
calculations of alkali reactions under these conditions, which also are
valid for gas entering the gas turbine, the boiler gas composition is
assumed to be 74 mole % nitrogen, 15 mole % carbon dioxide, 2 mole 7,
_2
oxygen, 9 mole % water, and 1.7 x 10 mole % sulfur dioxide/sulfur tri-
oxide. The volatility of sodium sulfate is low, or log P^a^SO/ ^ ~'*^»
(see Figure H-4) . Figure H-5 shows the equilibrium pressures of gaseous
sodium hydroxide and gaseous sodium chloride over solid sodium sulfate
under boiler conditions. The figure shows that the reaction of solid
sodium sulfate to form volatile gaseous sodium hydroxide is not important,
the vapor pressure of sodium sulfate exceeding the equilibrium pressure
of gaseous sodium. The critical level of hydrogen chloride above which
the driving force for release of sodium by hydrogen chloride attack ex-
ceeds the driving force for sodium release by sulfate vaporization is
-4 -6
log P.,-. ^ -5.4 or ?„„- = 4.05 x 10 kPa (4 x 10 atm). This corres-
HLJ.
ponds to a gaseous hydrogen chloride level of 400 ppb by volume.
Although in gas- or oil-fired gas turbines sodium is the pre-
dominant alkali contaminant, coal and dolomite contain both sodium and
potassium. Thermochemical equilibrium calculations have been made for
several systems containing potassium. Figures H-6 and H-7 show thermo-
chemical diagrams for the potassium-oxygen-sulfur-carbon and potassium-
silicon-oxygen-sulfur systems at 1093°C (2000°F), representing the
conditions inside or in the diffusion layer around a burning char
298
-------
Curve
'°9PH2/PH20
-2 -4 -6 -8'
-28 -24 -20 -16 -12 -8 -4
-44
Figure H-3 -Thermochemical equilibria in the sodium-silicon-oxygen-sulfur system
at 871°C (1600°F». Filled large circle indicates chemistry of gas phase
at exit end of fluidized bed boiler
299
-------
Temperature, °F
1200 1300 1400 1500 1600 1700 1800 1900 2000
10 l -
Temperature, K
Figure H-4-Vapor pressures of NaOH, KOH, NaCI, KCI,
K2S04. NaOH „, KOHe«, etc., are effective equilibrium
pressures which represent the sums of the pressures of
monomer and (twice the) dimer species.
300
-------
-6
E
S _7
O
ns
O
O
Na2S04(s) + 2 HCI (g) * 2NaCI (g)4-H20(g)
— O7lO
871° C (1600° F)
PT=1013KPa(10atm)
Sulfate
Stable
Na
Sulfate
Reacts To
Gaseous Chloride
Oppb
Na(OH)
or
NaCI
-6 -5.4 -5
(400ppbHCI)
log pHC( (atm)
Figure H-5-Equilibrium pressures of NaOH (g) and NaCI (g) over Na? $04 (s)
exposed to steam and HCI (g) underfluidized bed boiler gas phase
conditions
301
-------
log PH2/ PH20
20-2
Curve 6S878I-B
-4 -6
i i
-12
- -16
CM
I/I
O.
- -20
-24
-28
-32
-36
-40
-44
2 0
iogpco/P(
-2
C02
log pn (atm)
u2
-28 -24 -20 -16 -12
'S it}
(2)
-2
6 "k,
8
10
12
14
16
18
20
CM
X
16 14 12 10 8 64 2
log pso2'Pso3
Figure H-6-Thermochemical equilibria in the potassium-oxygen-sulfur-
carbon system at 1093°C (2000°FI. Filled circle indicates local
chemistry at surface of char particle.
302
-------
-2
Curv,- 658782-B
-6
i
81
-40 -
-44
16 14
10
"*psoz/pso3
Figure H-7-K-O-Si-C- S system at 1093°C (2000°F)
303
-------
particle. The large filled circles show the estimated chemical composi-
tion of the gas phase at the surface of the particle. The circle falls
within the field of stability for liquid potassium sulfide [KS (&)].
Potassium silicate and potassium carbonate formation are not favored.
Estimated of Levels of Chlorine Compounds and Alkali-Metal Compounds
Based on Mass Balance Considerations
The concentrations of alkali metals appearing in the gas will
be limited by the feed rate of coal and dolomite to the combustion system
and the rate at which alkalis can be liberated from the stone and char
particles during their stay in the reactors.
Figures H-8 through H-10 summarize material balances and the
estimated levels of alkali and chlorine contaminants in one of four flui-
dized bed boiler modules in a 318 MW power plant equipped with a W 501
gas turbine.
The contaminant estimates are based on assumed release frac-
tions for coal and dolomite. These are at this time strictly working
estimates, the validity of which are being examined through laboratory
experiments. For the calculations it was assumed that sodium and potas-
sium are released to the gas stream as volatile chlorides and hydroxides
at the one percent level. All chlorine that enters the fluidized bed
boiler was assumed to be released to the gas stream as hydrogen chloride
gas. This assumes that all but one percent of the sodium and potassium
entering the system as chlorides converts to silicates or sulfates, most
of which are retained in the bed ash. The chlorine is released as hydro-
gen chloride and finds its way to the turbine. These are oversimplified
assumptions, but they provide a starting point.
0
It should be noted that Pope, Evans and Robbins have measured
hydrogen chloride emissions from an atmospheric pressure fluidized bed
combus tor under conditions where a solution of sodium chloride is added
to the bed. They found about three percent of the chlorine as hydrogen
chloride in the flue gas. It is important to note differences in the
chemical environment experienced by sodium chloride in a burning coal
304
-------
Owg. 1670881
t
Exhaust
Gas to
Heat
Recovery
Boiler
Ash, Dolomite, &Char
Sulfated Dolomite
jidized Bed Boiler | »•
Coal
24,454 kg/ hour
(53,910lbs/hour)
p=
|7c
(15.5
Steam to Turbine
Dolomite
••^•B
7.061 kg/hour
493.059 kg/hour
(1.086,990 Ibs/hour)
Figure H-8- Schematic flow diagram for important gas and solids streams in one of four fluidized bed
boiler modules in a 318 MW power plant
305
-------
0.078 kg/h r (0. 171 Ibs/h r Sodi urn ]
.464 kg/hr(l. 022 Ibs/hr Potassium j
As Volatile
Chlorides
Gas-Turbine Expander |
0.1 -1.6% Sodium and Potassium in
53.3 kg/h r (117.6 Ib/hr of Particulates (0.05 gr. /cu ft. I
t
Exhaust
Gas to
Heat
Recovery
Boiler
i Ash, Dolomite. & Char
V
Sulfated Dolomite
I Flu id ized Bed Boiler
3. 09 kg/hour
(6. Sib/hour)
Sodium
31.0 kg/h r
(68.4lb/hr)
Potassium
Dolomite
4.7 kg/hr
(10.3lb/hr) Sodium
15.33 kg/hr 03.8 Ib/hr) Potassium
Air
Figure H-9-Flow diagram showing transport of sodium and potassium in one of four fluidized bed
boiler modules in a 318 MW power plant. One percent of the alkali content in Pitts-
burgh No. 8 coal and Tymochtee dolomite was assumed to produce volatile alkali com-
pounds that reach the gas turbine
306
-------
Dug. 1670883
3.18-25.2ku(7.0 -55.6 Ibs)hour Chlorine
I Gas-Turtine Expander |
Exhaust
Gas to
Heat
Recovery
Boiler
2.5 -19.6 kg/hour
(5.4-43.1)—
Ibs/hour
Chlorine
0.7-5.7 kg/hour
.6-12.5 Ibs/hour)
Chlorine
Figure H-10-Flow diagram showing transport of chlorine in one of four fluidized bed boiler modules in a
318 MW power plant. All chlorine, estimated to range between 100 and 800 ppm hy weight
was assumed to produce hydrochloric acid vapors that reach the gas turbine.
307
-------
particle as treated here and by sodium chloride injected in a solution.
Additionally, the recovery of hydrogen chloride from flue gas depends on
the extent to which the gas had equilibrated at the sampling temperature.
Typically, there is a lowering of equilibrium gaseous hydrogen chloride
concentration greater than one order of magnitude in cooling from 700°K to
300 °K: measurement of hydrogen chloride levels at 300 °K will then be ap-
propriate for assessment of the environmental impact, but they may be
irrelevant as an indication of the hydrogen chlorine concentration in the
gas entering the turbine.
The calculations were made for Pittsburgh No. 8 coal and for
Tymochtee dolomite. Although the alkali levels of this coal are typical
for eastern United States coals, the Tymochtee dolomite has alkali levels
which are near the higher end for rock beds. Lower alkali dolomites
typically have alkali contents which are three to ten times lower.
These calculations project a turbine expansion gas that could
contain as much as 5 to 40 ppm hydrogen chloride with complete chlorine
release from coals containing 100 to 800 ppm chlorine. Sodium chloride
at the 200 ppb and potassium chloride at the 700 ppb level would result
from a one percent release of these species from the mineral constitu-
ents.
Stability of Sodium Sulfate Melts on Turbine Hardware
Calculations of the conditions for alkali sulfate stability on
first- and second-stage components in a large industrial gas turbine as a
function of temperature, system pressure, and gas composition have been
made. Figure H-12 shows typical pressure-temperature relations for the
first- and second-stage vanes and rotating blades in such a turbine.
Sulfate deposit stability is calculated for vaporization and reactive de-
composition of sodium sulfate by any of the following reaction processes:
NaS0(s,A) t (NaS0)g (H-l)
H20(g) Z 2NaOH(g) + S02(g) + l/202(g) (H-2)
308
-------
04(s,£) + H20(g) ? 2NaOH(g) + S03(g) (H-3)
2HCl(g) J 2NaCl(g) + H20(g) + S02(g) + l/202
-------
Dwg. 1670882
KPa
Gas: <(147psia)
871° C (1600° F)
507.353 kg (1,118,502 Ibs/hr
08,318 Ib moles/hour)
On
2
N2
CO.,
9.59 mole%
76.21
&57
5.63
Corrosive Contaminants
(volume fractions)
S02,
185 ppm
HCI 5. 0-40.0 ppm
NaCI 189 ppb
KCI 666 PPb
Trace NaOH and KOH
Solids: 53. 3 kg (117.6 Ib) hr
(0.05gr/cu. ft), containing
Na &K at 0.1-L6* Level
Sulfated Dolomite
Fluidized Bed Boiler
Coal
Gas-Turbina Expander]
^
Exhaust
Gas to
Heat
Recovery
Boiler
t
, Dolomite, &Char
* «•
Air
t
Steam to Turbine
-4- Iteter
Dolomite
Figure H-ll'Summary of gas chemistry and corrosive contaminant levels in the combustion
gas stream directed to the gas turbine
310
-------
Curve 677869-A
2200
12
10
8
LO
O)
to
2
1200
0
1500
2000
1100
i I
Temperature, °F
1800 1600
1400
1st Vane
1SSl
ade
2nd Vane
2nd
Blade
1200
~ni2i6
1013
810
608
Temperature, °C
1000 900 800 700
i I i I i I i I
1400 1300 1200 1100
Temperature, K
1000
CO
a.
ZJ
in
in
O)
405 £
203
900
Figure H -12 -Pressure -metal temperature relations for first- and second
stage components of a large industrial gas turbine
311
-------
Curve C75343-A
2200
2000
Temperature, °F
1800 1600
1400
1200
12
10
j=
09 O
k.
E
O>
4
2
luppb
1200
1100
i I
Temperature, °C ^'Working estimate"
1000 900 800 700
1216
1013
810
608
405
203
ra
Q.
£
E
o>
en
1500 1400 1300 1200 1100
Temperature, K
1000
900
Figure H-13- Conditions of sodium sulfate stability (to the right of the
different concentration lines) on the first and second stages of a
large industrial gas turbine driven by coal combustion gases from
2 fluidized bed boiler power plant. These gases contain 40 ppm by
volume hydrochloric acid vapor and also sodium chloride at the
indicated volume concentration levels. The heavy broken line rep-
resents a working estimate for sodium release (1% of that in
coal and dolomite).
312
-------
dioxide would fall because of the formation of calcium sulfate. As the
sulfur dioxide concentration falls, the turbine tolerance to alkali metals
rises. The extent to which particulate calcium oxide can react with sulfur
dioxide is subject to a large number of variables, which means that the
turbine tolerance would vary with the number and reactivity of particles
with significant residence time in the turbine. At the limit - where
reaction of calcium oxide is complete and excess oxide is present — there
is about a 30-fold increase in the tolerable concentration of alkali. While
it is extremely unlikely that the limit can be approached, the influence
of particulates containing calcium oxide will have to be considered in a
kinetic model of the turbine tolerance.
CONCLUSIONS
The chemical reaction governing the release of alkali-metal
compounds to the gas-turbine expansion gas have been analyzed. Experi-
mental determination of the kinetics of alkali release over the range of
hydrogen chloride and sulfur dioxide concentration from 0 to 200 ppm are
needed.
Estimates of the tolerance of the turbine to prevent sodium
sulfate formation have been made. These show that the tolerable sodium
level depends on the hydrogen chloride and the sulfur oxide levels. The
predicted sodium tolerance of 0.2 ppm sodium in fuel oil for the case of
negligible hydrogen chloride level agrees well with the empirically de-
rived limit of 0.5 ppm sodium in fuel oil. The calculations show that a
sodium level of 0.2 ppm in the turbine expansion gas can be tolerated if
the hydrogen chloride level in the gas exceeds 40 ppm at a sulfur dioxide
level of 180 ppm, provided that turbine materials have a satisfactory life
in this atmosphere. Extensions of the turbine tolerance calculations to
include the effects of interacting sodium and potassium condensed com-
pounds are needed, along with experimental determination of the sensiti-
vity of the tolerance to kinetic factors.
313
-------
REFERENCES
1. Jonke, A. A. Reduction of Atmospheric Pollution by the Application
of Fluidized Bed Combustion. Office of Coal Research. Environ-
mental Protection Agency. Argonne National Laboratory. Argonne,
Illinois. Contracts EPA-1AG-0199 and OCR 14-32-001-1543. 1974.
2. Ruch, R. R., H. J. Gluskoter, and N. F. Shimp. Occurrence and
Distribution of Potentially Volatile Trace Elements for Coal.
Environmental Protection Agency. Illinois State Geological Survey.
EPA-650/2-74-054. July 1974.
3. Akapov, E. K. and A. G. Bergman. Reversible Adiagonal System of
Sodium and Potassium Chlorides and Sulfates. Zhur. Obshchei Khim
(Moscow) 24: 1524, 1954.
4. Grim, R. E. and W. F. Bradley. J. Am. Cer. Soc. 23: 242-248, 1940.
5. Slaughter, M. and W. D. Keller. Ceramic Bulletin _38: 703-707, 1959.
6. Graf, D. L. The American Mineralogist 37; (162), p. 1-2.
7. Graf, D. L. The American Mineralogist 37: (162), p.706.
8. Mesko, J. E., S. Ehrlich, and R. A. Gamble. Multicell Fluidized
Bed Boiler Design Construction and Test Program. Office of Coal
Research. Pope, Evans and Robbins. New York, N.Y. NTIS PB-236 254.
August 1974.
9. Durgin, G. A. and P. C. Holden. Westinghouse Gas Turbine Systems
Division Engineering. Lester, Pa. Private communication. October
1974.
314
-------
APPENDIX I
PARTICULATE CONTROL
-------
APPENDIX I
PARTICULATE CONTROL
TURBINE TOLERANCE FOR PARTICULATES
Previous estimates of the particle size distribution from
combustors have produced the distribution shown in Figure 1-1 (curve 1).
In addition, two progressively finer distributions (curves 2 and 3) have
previously been used in economic sensitivity analyses.
Erosivity
As a first approximation, the turbine erosion damage due to
the impact of a single particle is proportional to the kinetic energy of
the particle. The erosivity of a dust particle may thus be defined by
the product of its kinetic energy and its probability of impacting. To
obtain a measure of the erosivity of a dust suspended in a gas stream
it is necessary to total the erosivity of all the particles contained in
a unit volume of gas. This erosivity may be used to estimate acceptable
levels of partlculates in gas-turbine expansion gas.
Turbine Specifications
At present, typical specifications for gaseous fuels for gas
3
turbines require that dust loadings be kept below 0.023 gm/m (0.01 gr/scf);
that 95 percent of the dust be less than 20 urn; and that no particles be
greater than 100 urn. These specifications apply to high heating-value fuels
and must be adjusted for expansion gas, as the important consideration in
turbine operation is the particulate concentration in the expansion gas.
The adjustment of specifications must account for the dilution of the
fuel gas by excess air during combustion. High heating-value gas requires
315
-------
Curve 656022-A
.1
1
5
15
25
35
45
55
65
o>
93
96
98
99
Curves 1, 2, & 3 Represent Projected
Particle Size Distributions
1 Mill
I
t i i 11 M
I
i i
.5
3
10
20
30
40
50
60
70
80
90
95
97
98.5
10 100
Particle Size-Microns
1000
Figure I -1 Particle size distribution assumed for dust elutriated from gasifier
-------
approximately a fifty-times dilution. Consequently, the particulate
loading in the expansion gas on this basis should be restricted to one-
fiftieth of that specified for high heating-value gas, in other words,
0.0005 gm/m3 (0.0002 gr/scf).
An estimate of the particle size distribution of a dust which
would comply with this specification is shown in Table 1-1. The erosive
nature of this dust is dependent on the loading, the particle size
distribution, and its velocity relative to turbine components.
For example, we might assume that the dust suspension of
Table 1.1 is carried by a gas stream which passes over turbine components
at 300 m/s (1000 ft/s). If all particles are assumed to travel at the
same velocity as the gas stream, then the product of the dust concen-
tration with the probability of impaction (catch efficiency) gives a
direct measure of the erosive loading of the dust stream. This is shown
to be 133 x 10"6 gm/m3 (58 x 10~6 gr/scf) at 300 m/s(1000 ft/s).
Any other dust-laden low-heating-value gas which has an erosive
loading of less than 133 x 10~ gm/m (58 x 10~ gr/scf) at 300 m/s would
be acceptable as a fuel for gas turbines operating with this gas velocity.
It should be noted that particles below two microns are not
considered erosive in this case, as their efficiency of impaction is
negligible. These comments illustrate the turbine tolerance considerations
and the requirements which the particulate removal system must consider.
Turbine blade erosion is discussed further in Appendix J.
REVIEW OF DUST COLLECTION EQUIPMENT
Cyclones
Cyclones are able to handle heavy dust loading and high gas
throughput at reasonable cost and low-pressure drop. They are not prone
to plugging, but their Internal surfaces are subject to erosion.
Cyclones of conventional design lose efficiency as the cyclone
diameter is increased. Thus, for high efficiency with high gas flow,
banks of small cyclones are manifolded together in a parallel flow
arrangement (Figure 1-2). If the cyclones are connected directly to a
317
-------
Table 1-1
DATA FOR ESTIMATING TOLERANCE OF GAS TURBINES TOWARDS DUST
Particle size (urn)
0-2
2-3
3-4
4-5
5-6
6-10
M 10-20
CO
20 +
TOTAL
Particle loading,
emAn3 (gr/scf )
18.2 x 10~5 (8.0 x 10~5)
4.5 x 10"5 (2.0 x 10~5)
3.6 x 10~5 (1.6 x 10~5)
2.3 x 10~5 (1.0 x 10~5)
1.8 x 10~5 (0.8 x 10~5)
5.5 x 10"5 (2.4 x 10~5)
4.5 x 10"5 (2.0 x 10~5)
5.0 x 10~5 (2.2 x 10~5)
45.4 x 10~5 (20.0 x 10~5)
Catch efficiency,
(@ 300 m/s)
0
O.Q3
0.12
0.21
0.30
0.50
0.90
0.95
Erosive loading,
@ 300 m/s ((3 1000 fos)
( 0.6 x 10~6)
1.3 x 10"6 ( 0.6 x 10~6)
4.3 x 10~6 ( 1.9 x 10~6)
4.8 x 10~6 ( 2.0 x 10~6)
5.4 x 10~6 ( 2.4 x 10~6)
27.5 x 10~6 (12.0 x 10~6)
40.5 x 10~6 (18.0 x 10~6)
47.5 x 10~6 (21.0 x 10~6)
131.3 x 10~6 (57.9 x 10~6)
-------
Dwg. I67I860
CO
I-"
MS
Gas Outlet
-— Inlet
Inlet
Tube Sheet
Figure I-2-Multicyclone arrangements
a. Swirl vane inlet
b. Tangential inlet
-------
common dust hopper, secondary flows between individual units reduce the
operational efficiency of these units. Consequently, performance may be
improved by withdrawing the dust individually from each unit cyclone.
High-efficiency cyclones which employ secondary gas flows to
improve collection efficiency are also available both in large individual
units and in multicyclone configurations.
Conventional Cyclones
A conventional design high-efficiency cyclone, 203.2 nun (8 in)
diameter and with a capacity of 0.07 m Is (150 cfm) will have a grade
efficiency curve as shown in Figure 1-3 (at fluidized bed combustor
conditions). Units of this size could be incorporated into a multicyclone
design with little loss in efficiency.
Shell Collector
Shell oil uses high-temperature multicyclones to collect par-
ticulates from cat cracker off-gases before they are expanded through
gas turbines. The performance of these units is similar to that of
small conventional cyclones.
Rotational Flow Cyclones (e.g. Aerodyne)
Rotational flow cyclones employ a secondary gas flow as shown
in Figure 1-4. The vendor's data for fractional efficiency shows
extremely high collection efficiency for particles smaller than 5 urn
when a clean secondary gas flow is used (Figure 1-3). It is difficult,
however, to visualize such an arrangement in a fluid bed boiler plant.
It has been assumed that in such a plant the secondary gas flow would be
provided by using a portion of the dirty gas stream.
Tests of rotational flow cyclones with a dirty secondary gas
flow have shown poor collection efficiency for particles 5 ym
and smaller. Unless this performance can be radically improved the
units could not be considered for use in gas-turbine power plants.
320
-------
Curve 678095-A
NJ
•-/
T
Grade Efficiency Curves
Dust SG = 2.5
Gas Temp. =870°C (1600°F)
1. Convention Cyclone (0. 2 m Dial
2. Aerodyne Cyclone (Vendors -
Curve for aII Sizes)
3. Donaldson 'Tan Jet" Multi-
Cyclone (Individual Cyclones -
(0. 21mDiam.)
Figure I-3-Cyclone grade efficiency curves
-------
Exhaust (Clean Gas)
Dwg. 616^12
Secondary
Gas Inlet
Inlet (Dirty Gas)
Secondary Air Pressure
Maintains High
Centrifugal Action
Secondary Airflow
Creates Downward
Spiral of Dust and
Protects Outer Walls
From Abrasion
Dust is Separated From
Gas By Centrifugal
Force, is Thrown
Toward Outer Wall and
into Downward Spiral
Falling Dust is
Deposited in Hopper
Figure 1-4 - Operation of Aerodyne particulate separator
322
-------
Donaldson "Tan Jet" Cyclone
The Donaldson Company has recently released a new multicyclone
system which is significantly more efficient than conventional cyclones
in collecting particles in the O-to-5 ym range (Figure 1-3). This
unit requires a clean secondary gas flow equivalent to 15 percent of the
primary flow. The secondary flow should be at the same temperature as
the primary flow to prevent thermal stressing of the assembly and to
maintain the flow field at its design condition.
Filters
Three filtration systems have been considered:
• Granular bids
• Porous metals
• Porous ceramics.
Granular Bed Filters
Beds of granules should, in princiole, be able to filter dust
particles from gas streams in much the same way as do beds of fibers.
The literature on fiber filtration is extensive and shows that very high
efficiencies can be achieved over a broad range of particle size.
Unfortunately, fiber-bed filters cannot easily be cleaned and this limits
their utility.
Granular beds are somewhat easier to clean, may be operated at
high temperature, and have the potential to achieve the same high
efficiencies as fiber filters.
2
Squires reviewed the literature in 1970. His review indicates
that several filters have been operated with collection efficiencies
better than 90 percent on dusts down to 2 urn particle size. The
data, however, are soarse and inconclusive.
3
Since 1970, Taub has studied the transient behavior of granular-
bed filters while collecting dispersed fly ash. His results show high
efficiencies are possible with clean filters, but performance deteriorates
323
-------
as the dust content of the filter increases. His analysis of the filter
performance fails to explain his results adequately) some of them
appear spurious.
Ducon has attempted to commercialize a continuously operating
granular-bed filter and has published the results of its performance on
4
cat cracker regenerator off-gas. The results are encouraging but are
difficult to interpret. The latest information implies that the manu-
facture of this unit is too costly for it to compete with alternative
collectors.
Squires claims encouraging results for his panel-bed filter,
but limited data are currently available on its performance.
Mode of Operation. Although it appears that impaction collec-
tion will dominate in determining the efficiency of a granular filter,
two operating regimes exist: one in which impaction predominates and
particles are collected in the interstices of the filter, the second
in which the initial collection at the filter surface produces a filter
cake. This cake acts as a filter aid, retaining essentially all of the
collected solids at the filter surface.
The first operating mode is normally associated with high gas
velocities and, consequently, with high-throughput filters. When a
filter operates for an extended period under these conditions, the
operating efficiency will initially remain constant with time. The bed
will become progressively saturated with dust, however, and as the
saturation zone extends through the bed, the collection efficiency wil]
3
decline. Some results of Taub's work show this trend (Figure 1-5).
Saturation apparently occurs when dust deposits on impaction sites
profile the gas flow, avoiding sharp changes in direction and minimizing
the effects of inertlal deposition. It is most likely that there is also
a dynamic equilibrium between deposition and reentrairment in saturated
zones.
Analysis of the transient behavior of beds is, thus, complex
and is further confused by a lack of fundamental information. The only
data available are from Taub and appear to be unreliable. His grade
324
-------
OJ
M
Ul
100
90
80
* 70
C? 60
22
| 50
UJ
40
30
20
10
0
I I
0.033m bed, 1.49mm. spheres, 60cm./sec.
0
10
20
Time, mins
30
40
Figure I-5-Time vs efficiency -0.033 m deep granular bed filtering fly ash
-------
efficiency curves show no significant difference in collection efficiency
between particles of 10 urn diameter and particles of 70 pm
diameter, which is contrary to general experience. This could result
from poor dispersion of the test dust employed. From Taub's description
of his equipment, the dust disperser was crude and could not be relied
upon for adequate dispersion.
The second operating mode — formation of a filtering layer —
is essentially the same as is encountered in bag filters and in some porous
metal filter installations. Although some dust leakage may be expected
through the clean filter, once the filter cake is established filtration
is almost absolute. As the cake builds upon the filter surface, there
is a consequent increase in pressure drop. At some point, the pressure
drop becomes too great; it is necessary to remove the filter cake and
repeat the cycle.
Dust Removal. When dust is deposited throughout the filter,
it may be removed by one of two methods:
• Fluidizing the bed with a reverse flow of gas. This is
limited to horizontal beds and requires some secondary
collector for cleaning the flushing gas.
e Removing both the granules and the dust from the filter and
repacking the filter with fresh granules. This is normally
limited to vertical-panel filters. The packing may be
intermittently dumped or continuously removed in a cross-
flow arrangement.
When the dust accumulates in a cake on the filter surface, the
cake may be removed by a reverse flow of gas. If the filter is a verti-
cal panel, the flushing gas should be supplied as a sharp puff at high
pressure. This will lift the cake from the surface and deposit it in a
secondary collector. If the filter is arranged horizontally, the
reverse flow of gas is used to fluidize the bed and elutriate the dust
deposits.
326
-------
Capacity. The work of Taub shows that a 0.033 m (1.3 in) deep
bed of 1500 pm particles has a capacity of approximately 2000 Rm/m2
(3000 gr/ft ) before the collection efficiency begins to fall. This
figure may be used as the safe operating limit for the deeper beds which
would be used in a fluidized system.
Ducon Filter
Operation. The Ducon filter employs screens to retain the
granular bed while permitting removal of the collected dust by blow-
back techniques. The arrangement and operation of the unit are
illustrated in Figure 1-6.
When being filtered, the dirty gas passes through the outer
screen and down through the granular sand bed. When the bed has accumu-
lated sufficient dust, a short blast of high-pressure gas is used to
reverse the flow. The bed flexes, is fluidized, and the dust is carried
from the bed.
Performance. This unit normally operates with a 0.063 m
(2-1/2 in) deep bed of 760- pm sand. The usual superficial gas
velocity is between 0.15 and 0.45 m/s (0.5 and 1.5 ft/s ).
4
Performance correlations published by Ducon show that the
instantaneous efficiency of the filter improves as the quantity of dust
in the bed Increases. This implies that collection is enhanced by the
formation of a filter cake on the bed surface. The overall efficiency,
however, is considerably lower than would be expected if a coherent
cake were present. For example, when operating on cat cracker emissions,
the efficiency was normally around 95 percent. By using a finer grade of
sand, this performance can be improved. Available data on sintered
metal filters indicate that 100 percent collection down to 1 pm can
be achieved with granules of 100 pm diameter and velocities around
0.03 m/s (1 ft/s).
No grade efficiency curves for the filter have been made
available by Ducon.
327
-------
Dwg. 6224A12
Outer
Screen
Catalyst-
Laden
Dust
,***
Granular
Sand Bed
Inner
Screen
Filter Element Internals
Collected
Catalyst
—Fixed
Sand Bed
4
HQH
I Clean Gas
Operating Cycle
--Fluidized
Sand Bed
Blow-back
Gas
Cleaning Cycle
Figure I-6-Ducon sand bed filter
328
-------
Operating Problems
• Leakage - Ducon has noted that dust leaks through the filter
beds during the blow-back phase of the operating cycle. This
markedly reduces the efficiency of the unit.
• Plugging - If finer granules are used in the Ducon filter,
finer retaining screens will be required. Experience has
shown that fine screens have a tendency to accumulate dust
deposits and become plugged.
Squires Panel-Bed Filter
Operation. The Squires panel-bed filter consists of a vertical
bed of granules held in place by louvered walls which resemble Venetian
blinds. The unit operates with superficial gas velocities up to 0.15 m/s
(0.5 ft/s). Recent information suggests that collection efficiency in-
creases with superficial velocity. v
The filter is cleaned intermittently by a sharp puff-back of
gas which lifts the filter cake, along with a small quantity of granules,
from the surface and deposits it in a collecting hopper. Fresh granules
are supplied from the top of the panel to make up for losses during the
puff-back.
The operation of the unit is illustrated in Figure 1-7.
Performance. There is little data available on the efficiency
of the Squires Filter. However, filtration efficiencies of better than
2
99 percent have been recorded on redispersed fly ash, with a filter con-
sisting of 0.025 m (1 in} sand.
Operating Problems.
• Size - If the equipment operates at low gas velocities, a
large filter area is needed for treating commercial quanti-
ties of high-temperature gas.
• Solids handling - Distributing and collecting the filter
granules will produce mechanical difficulties.
329
-------
Fresh
Solid
I
Dirty-Gas
Inlet
Clean-Gas
Outlet
Quick
Acting
Valve
Compressed
Gas for
Puff-Back
Dust and Solids out
Figure I-?-Squires panel bed filter
330
-------
• Plugging - The cleanup cycle for this filter will only
remove surface accumulations of dust. It is possible that
with long operating schedules dust will penetrate the
filter panel and block the interstices.
Cross Flow Filters
Operation. Panel bed filters may be operated with a continu-
ous downward movement of the column of solids, clean granules being
introduced at the top and a mixture of dust and granules being removed
from the bottom. This is consistent with high-gas velocity/impaction
collection operation.
For the fluidized bed boiler, the spent limestone granules
could be used as the filter medium (Figure 1-8), providing a continuous
supply of clean, hot, filter granules. As the stone is to be dumped
anyway, this arrangement does not produce any secondary disposal problems.
Q
Performance. Dorfan has built cross-flow filters which use
0.013 m (1/2 in) to 0.038 m (1-1/2 in) granules and claim 98 percent
collection efficiency for 2 to 10 urn dust carried at superficial
velocity of 1.8 m/s (6 ft/s).
Q
Zahradnik et al. report some preliminary results for a
cross-flow bed of 1600 um (1/46 in) alkalized alumina. They show
99 percent collection efficiency for redispersed fly ash carried at
0.15 m/s (0.5 ft/s).
Operating Problems
• Solids flow — Solids flow in a conventional panel bed
retained by louvers will create dead zones near the filter
surfaces. These zones would eventually saturate with dust
and possibly plug (Figure 1-9).
• Reentrainment - Relative motion of the filter granules will
cause dust reentrainment in the gas stream. Unless the bed
is sufficiently wide in the direction of gas flow, poor
performance will result.
331
-------
Stone From
Reactor
8700C(1600°F)f15ATS
Gas From Reactor
870°C(16000F),15ATS
*v ^
v .•".••'. •'•-' ''
V-. V/l
Clean Gas
to PACE
•Filter Panel
Containing Vessel
Stone
Dust
to Waste
Figure I-8-Cross-flowfilter arrangement -©gasification plant
-------
Gas
Dead Zones
••V:--.:--/-:">.-.-J/-:
:///,.';.'-'Velocity Profile ;':-V;
'..'•'••: for Granules .-'••;
Louvres
Figure I-9-Anticipated solids flow and "dead" zones
for a conventional cross-flow filter
333
-------
• Solids handling - Solids distribution to the filter panels
presents mechanical difficulties.
• Erosion - Solids retaining elements will be subject to
erosion by the moving granular bed.
Porous Metal Filters
Porous metal filters have been used commercially for 15 years
to filter catalyst fines from effluent gases in fluidized bed reactors.
They can be operated with virtually 100 percent efficiency on particles
down to one micron on a continual, cyclic basis.
Porous metals have two serious operating problems:
• They are limited to operation at temperatures below 538°C
(1000°F)
• They require a clean, high-temperature gas supply for the
cleanup portion of the operating cycle.
Porous metals are also costly: a budget estimate is $1000 per
square meter of filter surface.
Porous Ceramic Filters
Porous ceramic filters may in principle operate in a way
similar to that of porous metal filters. At high gas flows, however,
vibrational strains lead to cracking in the ceramic elements. In
addition, sealing ceramic elements in a manifold is difficult when
high-temperature operation is required.
334
-------
REFERENCES
1. Keairns, D. L. et al. Evaluation of the Fluidized Bed Combustion
Process. Office of Research and Development. Environmental
Protection Agency. Westinghouse Research Laboratories. Pittsburgh,
Pa. EPA-650-2-3073-048a. Contract 68-02-0217. NTIS PB-231162/9
December 1973 /
2. Squires, A. M. and R. Pfeffer. Panel Bed Filters for Simultaneous
Removal of Fly Ash and Sulfur Dioxide. J. A. P. C. A. 20: 534, 1970.
3. Taub, R. Filtration Phenomena in a Packed Bed Filter. Ph.D.
Dissertation. Carnegie-Mellon University. 1970.
4. Kalen, B. and F. A. Zenz. Pollution Control Operations: Filtering
Effluent from a Cat-Cracker. C. E. P. 69; 67, 1973.
5. Tesorio, A., Ducon Company, Inc. Private Communication.
6. Squires, A. M. Private Communication.
7. Pall, D. B. Filtration of Fluid Catalyst Drives from Effluent Gases.
I. E. C. 45: 1197, 1953.
8. Perry, R. H., and C. H. Chilton, eds. Chemical Engineering Handbook.
Fifth Edition. New York. McGraw-Hill Book Company. 1973.
9. Zahradnik, R. L., J. Anyigbo, R. A. Steinberg, and H. L. Toor.
Simultaneous Removal of Fly Ash and Sulfur Dioxide from Gas Streams
by Shaft Filter Sorber. Env. Sci. Tech. 4^ 665, 1970.
335
-------
APPENDIX J
GAS-TURBINE DESIGN AND OPERATION
J
-------
APPENDIX J
GAS-TURBINE OPERATION
To successfully operate a gas turbine on the exit gases of
a pressurized fluid bed combustion process, it is necessary to limit
the concentration of alkali-metal compounds and of particulates to
levels low enough that excessive deposition, hot corrosion, and erosion
of turbine components does not occur.
ALKALI-METAL COMPOUND TOLERANCE
Alkali-metal compound concentrations must be low enough so
that oxygen-excluding liquid films of sulfates, or mixed sulfates and
chlorides do not form to initiate hot corrosion of turbine components. The
simplified sodium sulfate deposition model discussed in Appendix H has
shown that the tolerance for alkali metals is a strong function of both
the sulfur dioxide (SO.) and the hydrogen chloride (HC1) levels in the
expansion gas. If sodium was the only alkali metal present, transport
between the fluid bed combustion system would be primarily by sodium
chloride (NaCl), and the tolerance of the turbine would vary, as shown by
Figure J-l, between 50 and 200 ppb by volume at a sulfur dioxide level
of 200 ppm for hydrogen chloride levels between 5 and 40 ppm (those
expected in burning low- and high-chlorine coals). Better sulfur
removal could increase the sodium tolerance greatly. As indicated in
Appendix H, the presence of potassium compounds reduces the allowable
alkali-metal concentration.
Figure J-2, which presents the equilibrium case for a deposition
model which permits the sodium sulfate-potassium sulfate (NaSO,-K0SO,)
eutectic melt (melting point 832°C [1530°F])to form on the blading, shows
that the maximum alkali-metal tolerance for sodium is reduced by a factor of
337
-------
u>
I
"e
o
•H
4J
CO
I
4000
3500
3000
2500
2000
1500
1000
500
Fluid Bed Combustor
System Pressure: 10 Atm
3.125 ppm SO
40 61
Chlorine Concentration/ppm
6.25 ppm SO
12.5 ppm SO
x
25 ppm SO
x
.50 ppm SO
A
100 ppm SO
200 ppm SO
100
Figure J-l - Turbine tolerance for sodium as a function of the concentration
of chlorine and oxides of sulfur (sodium sulfate melt model)
Basis: 2NaCl + H20 + SO., + 1/202 * Na^ + 2HC1
Na2S04 melting point, 884°C (1623°F)
-------
LJ
VO
•a
ex
d
01
u
§
4000
3500
3000
2500
2000
1500
*§ 1000
500
Fluid Bed Combustor
System Pressure: 10 Atm
02: 1.7%
Hn> ft iv
_U. O.3&
Basis: 2NaCl +
+ 1/20
2
+ 2HC1
eutectic temperature, 832°C (1530°F)
3.13 ppm SO,
6.25
u
ii
Chlorine Concentration/ppm
Figure J-2 - Turbine tolerance for sodium as a function of chlorine
and oxides of sulfur concentration
(Na2SO,/K2SO, eutectic melt model)
-------
about 4 by the presence of this melt. This is not the worst case. The
addition of potassium and sodium chlorides to the melt permits a four-
component eutectic of even lower melting point —51A°C (957°F)— and the
permissible concentrations of sodium and potassium to prevent the
formation of this liquid will be even lower.
Further work is required to quantify the tolerance
for the ranges of sodium and potassium compounds likely to be encountered.
In addition to understanding the turbine tolerance, work is needed to
understand how best to meet the tolerance. This work includes:
• Developing the effects of bed combustion temperature and
residence time on the alkali-metal concentration in the
coal gas
• Evaluating the technical and economic effectiveness of
additives in the beds or in granular bed filters in lowering
concentrations of trace chemical contaminants to the turbine
tolerance
• Establishing the technical feasibility of limiting hydrogen
chloride concentrations in the combustion systems to minimize
alkali release and of increasing hydrogen chloride concen-
trations in the turbine to prevent deposition
• Evaluating the effectiveness of pretreatment in the coal
and stone dryers of the coal and dolomite feed to convert
sodium and potassium chlorides to more stable and less
volatile compounds before they enter the higher temperature
combustion system.
Particulate Tolerances
If the alkali-metal compound tolerances are met, one is
reasonably assured .because of the low-combustion temperatures in a fluid
bed combustion system, that liquid films will not be present on either
the turbine hardware or on the surface of the particules. Deposits
340
-------
cemented by alkali-metal compounds should not form. Sintering of
deposits will not be accelerated by the presence of the liquid -phase.
Deposits resulting from the impaction and dry sintering of the
mix of fine particles can still occur, especially in the first staees
of the turbine where blade-metal temperatures may exceed 870°C (1600°F)
in some places. Information is not yet available to define particulate
tolerances to limit deposit growth to acceptable rates. We need to
understand the influence of blade temperature and particulate arrival
rates on the growth and sintering of deposits; we need to establish
rates of erosion of deposits so that deposit formation and removal
processes are quantified.
Deposit formation is less of a problem at low turbine inlet
temperatures. Turbine operating experience has indicated that below
677°C (1250°F) erosion rather than deposition may become the factor
that limits turbine life.
A preliminary estimate of the turbine tolerance of particulates
for control of erosion damage can be made by utilizing the experimental
results of the Australian Direct Coal-Fired Gas Turbine Program.
Summarizing their experimental results.
• The turbine inlet temperature was dropped to 677°C (1250°F)
so that erosion rather than deposition would limit turbine
life.
• Direct firing of pulverized coal through a "can"-type
combustor was used so coal ash particulates can experience
temperatures above 1650°C (3000°F) in the combustor transit.
The melted cenospheres resulting from this flame transit
are more erosive than fluid bed combustion ash,which even
in the combustion zone does not see temperatures above
the softening point of most coal ash; in other words, fluid
bed combustor combustion-zone temperatures are near 1100°C
(2000CF).
341
-------
• The turbine was redesigned to lower gas velocities
leaving the turbine nozzles from about 370 m/s (1200 ft/sec)
to 240 m/s (800 ft/sec) in order to lower erosion damage.
-4
• The redesigned turbine expanded gas contained 7 x 10
kg ash/kg gas.
• The mean particle size of the ash entering the turbine
ranged between 4 and 8 ym in diameter. The particle size
distribution was such that essentially no particles were
larger than 22 urn diameter:
- 10 percent of the particle mass exceeded 10 pm diameter
- 60 percent of the particle mass exceeded 5 urn diameter
- 90 percent of the particle mass exceeded 3 um diameter.
Figure J-3 shows the measured size distributions.
• The rate of erosion of both stator and rotor blades
decreased with time for the first 50 hours of turbine
operation and remained stable for the remaining 73 hours
of testing.
• Extrapolation of the wear data indicated that rotor blades
and stator vanes would require replacement,due to thinning
of their trailing edges. (The protection of blade roots
from notching afforded by the stepped side-wall construction
and trailing edge fences to control dust concentrations
appeared to be effective.)' Blade life was calculated on
the basis of the time required to reduce the trailing edge
blade thickness from its design value of 1.47 mm (0.060 in )
to zero thickness. Table J-l shows the life expectancy of
the blading in the modified turbine.
• The Australian turbine expanded 9 kg/s (20 Ib/sec) of hot
gas compared to the nearly 300 kg/s (700>Ib/sec) of hot
gas expanded by a 60 MW electrical utility gas turbine.
The blade height and blade pitch in the first-stage rotor
Australian turbine are about, half those of a large, 60 MW
turbine.
342
-------
Curve 68l2Mf-A
5 10 15 20 25 30 35
Particle Size, urn
Figure J-3- Particle size distribution measured at
entry and exhaust of Australian
direct coal-fired gas turbine. Greta
Seam Coal Test. Redesigned low gas
velocity turbine-( Reference 1,
Figure 122)
343
-------
Table J-l
EXTRAPOLATED LIFE OF BLADING IN REDESIGNED LOW-VELOCITY
AUSTRALIAN GAS TURBINE TESTS3.b,c,d
First
Second
Third
stage
stage
stage
I Stators.hr
51,000
25,000
27,000
Rotors.hr
31,000
52,000
51,000
aCobalt-based Stellite^ Blades HS31.
Design life of original Ruston Hornsby turbine,100,000 hr.
Because of nonuniform ash distribution at entrance to the turbine,
blade life is a function of position in the stator row: 5 blades
in the second stator row had an estimated life between 10,000
and 15,000 hr; 2 blades in the first-stage stator row, 7 in the
second, and 10 in the third had an estimated life between 15,000
and 20,000 hr.
Interdepartmental Steering Committee. The Coal-Burning
Gas-Turbine Project. Department of Minerals and Energy. Aeronautical
Research Laboratories. Department of Supply. Australian Government
Printing Service. Canberra, Australia. 1973. (Table 44).
344
-------
To a first approximation erosion damage would be expected
to be directly proportional to the particulate concentration in the gas
being expanded by the turbine, and proportional to some power of the
gas velocity leaving the turbine vanes. Empirical evidence has led to
estimating this power dependence to be between the gas velocity cubed
and the gas velocity to the fifth power. In addition,erosion damage
would be expected to be a function of turbine rating because the
particle size distribution of the ash is fixed; but as the turbine rating
increases, the physical size of the gas passage becomes larger. Figure J-4
and J-5 illustrate this point, showing the trajectories of 9 urn diameter
particles through geometrically similar first-stage stator vanes where
all linear dimensions have been decreased by the scale factor "k". As
the passage size shrinks, the fraction of the particles entering the
channel which impact the pressure surface of the passage increases from
about 35 percent in the full-scale stator vane to about 60
percent in half-scale passages (see also Figure J-6). The impact
velocities and impact angles also change. As Figure J-7 demonstrates,
impact velocities drop slightly as the passage is made smaller. This
would tend to lower erosion damage in small turbines. Impact angles,
however, increase, as seen in Figure J-8. In large turbines, trailing
edge impacts tend to occur at grazing angles - about 6 degrees for
the 9 urn diameter particles - as shown in Figure J-8- As the passage
shrinks, these impact angles become greater, and in half-scale passages,
9 pm particles are now striking at incidence angles near 10 degrees.
Erosion damage is a function of impact angle (as the data of Figure J-9
show), and damage rises rapidly as the impact angle changes from 0 toward
20 degrees. The net effect of the larger number of impacts, the slightly
lower impact velocities, and the impact at more damaging angles of impact
in the smaller passage would be to cause the full-scale machine to erode
at about 80 percent of the rate of a half-linear-scale machine (Figure J-10).
To a first-order approximation, the erosion .damage yielding
acceptable life for the Australian machine was proportional to :
345
-------
Curve 681266-A
je
X
u>
*-
ON
•CO
I-10
CD
UJ
Ul
IE
U)
o.
.
k = 1.0
-2 -1 0
x
•CO
I- I/I
C9
U
Of
CCM.
3
U
Q£
O —
k = 0.75
— 100%
--0%
1 2 3 4 S
RXIRL LENGTH.It), x k
-2 -I
0 I Z 3 4 5
RXIRL LENGTH.IN.x k
FIGURE J-4- 9 qm particle trajectories in full-scale and
3/4-scale stator passages
-------
Curve 681265-A
a.
•CO.
KM.
O
:<"».
0£
kJ
U.
:CM.
o.
k = 0.5
-100%
— 0%
-2 -1
t-M ,
ta
zco,
u
O£
SCSI.
u
oe.
i .
(M
k = 0.25
--0%
0123456
RXIRL LENGTH.IN.x k
-2 -1 0
1 2 3 4,
AXIRL LENGTH.IN.x k
FIGURE J-5- 9 4m particle trajectories inl/2-scale and
lA-scale stator passages
-------
1.0
0.8
.2 0.6
o
2 0.4
Q. •
O
0.2
0.0
0.0
I I I T
Particle Diameter = 9um
J I
I I 1
0.5
Scale Factor
1.0
Figure J-6- Effect of scale factor on capture efficiency.
(Capture efficiency = number of particles
that hit blade surface/total number of
particles at inlet)
348
-------
OJ
*>
\o
1600
1400
1200
«/>
^ 1000
^
I 800
o
(Q
600
400
200
0
0.0
^ r
A Full Scale
a 3/4 Scale
o 1/2 Scale
o 1/4 Scale
Particle Diameter =9iim
1.0
2.0
3.0
4.0
5.0
Axial Distance, in x scale factor
Figure J-7-Effect of scale factor on particle impact velocities
-------
f i.r\o 'I-'••'•-' >-
ui
in
O
I/I
O>
CD
CD
80
70
60
50
40
£30
E
F—I
20
10
0
0.0
a Full Scale
n 3/4 Scale
o 1/2 Scale
o 1/4 Scale
Particle Diameter =
i
I
1.0 2.0 3.0 4.0
Axial Length, in x scale factor
5.0
Figure J-8-Effect of scale factor on particle impact angles
-------
c
o
3.0
2.5
Particle Velocity:
171 m/s
(560fpsJ
-/ //118 m/s
(390 fpsJ
30 60
Impact Angle, degrees
Figure J-9-Erosion results obtained for alumina
particles impacting 2024 aluminum
alloy2
351
-------
Curve 680606-A
a. K =
b. K = 3/4
c. K =
d. K =
Particle Diameter = 9um
1.0 2.0 3.0 4.0
Axial Distance, in x K
Figure J-10-Effect of scale factor on erosion rate in
®-501 first-stage stator
352
-------
Turbine Particulate
Effect of
Erosion Damage j (Machine Scale
I Dust 1
\Concentration]
I Nozzle r
[Velocity/
Australian Direct Coal-
Fired Turbine Damage j
And for the same life of blading in a 60 MW turbine:
Large Electric Utility n QA /Acceptable Particulate 1 /cnn , »;
Gas-Turbine Damage a °'8° \ Concentration I (61° m/s)
So that:
[Acceptable Particulate
\ Concentration
1 (7 x 10"4) (240)3
0.8 (610)3
4J (e.A x 10~2J
= 6 x 10
7 x 10 16.4 x 10
-5 kg of dust
kg of expansion gas
And since a kilogram of turbine expansion gas occupies about 0.74 m ,
this corresponds to a permissible particulate loading in the turbine
-5 kg -2
expansion gas of 8 x 10 —^3 ( 4 x 10 gr/scf).
sm
If we had considered erosion damage proportional to the fifth
power of velocity, the allowable particulate concentration would have
been:
Acceptable Particulate V
Concentration
0-8
7 x
kg gas
(240 m/s)
(610 m/s~)5
9 x 10
'6
kg gas
This is almost an order of magnitude lower.
353
-------
These rule-of-thumb extrapolations have indicated that large
electrical utility gas turbines might be operated with acceptable blade
life at 10 to 100 times the current allowable Particulate loadings in
-4 3 "-4
the expansion gas- 5 x 10 gm/m (2 x 10 gr/scf).
To assess the reliability of these estimates the following
factors must be considered',
• Erosivity of the coal ash from the fluid bed combustion
process - friable platelets of alumino silicate ash mineral
matter - is expected to be much less than that of the direct
coal-fired ash. The erosivity of the sulfated, haIf-calcined,
dolomite sulfur sorbent is not known. The first factor
would make the estimate conservative. Hard data on the
erosivities of the expected particulates impacting turbine
materials at the correct particle and metal temperatures are
needed.
• The rule of thumb that turbine erosion damage is proportional
to gas velocity raised to a power between 3 and 5 is based
on limited experience with small turbines. Particle velocities
at impact deviate substantially from the gas velocities, and
secondary flows in the turbine passage act to change the local
particle concentrations in the flow path. To replace this
rule-of-thumb estimate with more reliable data, we are
calculating erosion damage from a model that calculates the
particle trajectories through the turbine, accounting for:
- The development of the fluid boundary layers on the vane
and blade surfaces and on the hub and outer casing end-
walls of the turbine
- The development of the passage vortices
- The radial flows which occur in the wakes of the vane
and blades due to the higher gas pressures at the outer
casing wall as compared to the hub of the machine
354
-------
- The decrease in gas velocity in the blade and vane wakes.
Figure J-ll shows a typical trajectory produced by the
program. Top and side views of a 60 MW turbine first-stage stator and
rotor are given. In addition to the hardware boundaries, the computer
has printed out the boundaries of the regions in which boundary-layer
models and blade-wake models are used to calculate the gas flow. The
passage vortex flow, which is a function of the turning of the gas flow
in the blade passages and the thickness of the end-wall boundary layers,
is acting on the particles in both the vane and blade passage. The
particle trajectory is shown both in fixed and rotating coordinate systems.
Printed program output (not shown) provides the angle and velocity of the
particles before and after impact. The program is written to accept
experimental data on impact characteristics. Figures J-12 and J-13 show
the trajectories of a 6 urn particle through the first stage of the turbine
for two differing inlet positions of the particle. These trajectories
show the effect of the radially inward gas flow in the blade wakes on
the particles. Westinghouse is now adding an erosion damage prediction
subroutine and is improving its main gas-flow models to provide a
particle trajectory model capable of predicting erosion and deposition
damage rates. This model will be used to establish more reliable
specifications for the turbine particulate tolerance. It should also
be useful in investigating the effectiveness of deposition and erosion
control techniques.
TURBINE DESIGN
To preserve high efficiency in a pressurized fluid bed
combustor system it is necessary to transfer the 1013 to 1520° kPa
(10 to 15 atm), 871 to 927°C (1600 to 1700°F) hot gases leaving the dust
collection system directly to the gas turbine for expansion. Various
design configurations to accommodate thermal expansion, to control
leakage, and to provide a uniform distribution of particulates over the
flow channel of the turbine have been suggested. Operating experience
is available from European compound-cycle power plants utilizing one or
355
-------
to
m
V)
m
m •
Z
i— m
. co
s«
cr
o: -
at
CM
ui
-Casing Wall
/ "*
Casing Boundary Layer
Hub Boundary Layer
Hub Wall
Particle Trajectories
(absolute)
CM
Blade Wake Radial
Flow Model
Z c-
Ld
Blade Boundary
Layers
(O
Q£.
CC
o •
Blade Wake
Model Region
articile Trajectories
(relative to
blade)
IDENT. NQi 1203; 0=12.0
Passage Vortex
Model
-1 0 1
10 11 12
Z, INCHES
Figure J-ll - Trajectory of a 12 ym diameter particle through
first-turbine stage
356
-------
o>
t!
to
o>
•&
36
35
34
33
32
31
30
29
15
14
13
12
11
10
9
8
7
6
5
4
3
2
I
0
-1
Curve fc7S
I I I I I I I 1 I I I I
i '• .
T •„ .
Outer casing boundary layer
Hub boundary layer
' 1 ' '-'^
J I
I I
J I
j l
Side View
i i
i i i i i i i i
Particle trajectories - -.
fixed coordinate system
Particle trajectories-
elative to rotating blade
i i i i i i i i i i i i
-1 0 1 2 3 4 5 6 7 8 9 10 11 12
x , Inches
Top View
Figure J-12-Trajectory of a 6pm particle through the first-turbine stage
357
-------
Curve 678336-B
JG
I
u
c
o»
u
35
34
33
32
31
30
29
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
-1
i r i i i i i r
Outer casing boundary layer
Radial flow effects
I I
I I I I
T \
Particle trajectories- -
fixed coordinate system
Particle trajectories-.
relative to rotating blade
i i
i i i i
-1 0 1 2 3 4 5 6 7 8 9 10 11 12
x , Inches
Top View
Figure J -13 -Trajectory of a 6 urn particle injected nearer to
mid-span and mid-pitch of the blade passage.
Note the decreased effect of the radial flow at the
trailing edges of both the stator vane and blade on the
radial displacement of the particle as it leaves.
358
-------
two high-pressure connections to the turbine. These installations have
generally delivered hot gas at turbine inlet temperatures between 700
and 760°C (1300 and 14008F),in other words, about 300°C (500°F) below
the turbine inlet temperatures currently used in Westinghouse industrial
gas turbines. Westinghouse Gas Turbine Division engineers have prepared
a preliminary design using external manifolding to distribute hot gases
3 4
around the annulus of our W501 (60 MW electric) turbine. ' This
design had the objective of avoiding distortion of the turbine casing by
nonuniform temperature distributions which they feared would be associated
with a single hot-gas distributor.
Further design work to assess the technical problems associated
with these designs, improving them, and developing reliability and economic
estimates of alternative constructions are required. The initial and
advanced designs will then require tests on an integrated combustor-turbine
system.
5-7
Past experience with gas turbines expanding dust-containing
gases indicate design modifications that may be helpful in avoiding life-
limiting erosion of turbine hardware. Of special concern are design
features in the turbine which may cause localized concentrations of the
dust in regions susceptible to erosion attack. A turbine design is
needed that:
• Provides for uniform distribution of the dust-laden gas
over the inlet flow channel
• Directs particle flows in blade and vane wakes to avoid
raising the erosion potential of dust at blade and vane
roots
• Uses stepped sidewalls, carbide wear-resisting inserts, and/or
cooling air injection as appropriate to protect blade and
vane roots from erosion damage
• Appropriately thickens and hard-faces blade tips to resist
erosion damage
359
-------
• Incorporates spray systems and drains, and provides for
injection and removal of milled nut shells for washing and
cleaning of blade and vane surfaces without the need to
open the turbine
• Lowers velocity of gases in the turbine, if required, to
achieve satisfactory erosion life.
In order to realize the full potential of pressurized fluidized
bed combustion systems, work must continue to:
• Develop commercial gas-turbine designs
• Carry out analytical and laboratory tests to understand
turbine tolerance to corrosion, erosion, and deposition
• Obtain data on large-scale integrated fluidized bed
combustor, particulate control, and turbine test systems.
360
-------
REFERENCES
Interdepartmental Steering Committee. The Coal-Burning Gas Turbine
Project. Department of Minerals and Energy. Aeronautical Research
Laboratories. Department of Supply. Australian Government Printing
Service. Canberra, Australia. 1973.
Grant,G.,and W. Tabakoff. Erosion Prediction in Turbomachinery
Due to Environmental Solid Particles. AIAA Paper No. 74-16.
AIAA 12th Aerospace Sciences Meeting. Washington, D. C. January
1974.
Archer, D. H., et al. Evaluation of the Fluidized Bed Combustion
Process. Office of Air Programs. Westinghouse Research
Laboratories. Pittsburgh, Pa. November 1971. NTIS PB 211-494,
212-916, 213-152.
Keairns, D. L., J. R. Hamm, and D. H. Archer. Design of a
Pressurized Fluidized Bed Boiler Power Plant. AIChE Symposium.
Series 126, 68: (1972).
Smith, J., R. W. Cargill, D. C. Strimbeck, W. M. Nabors, and
J. P. McGee. Bureau of Mines Coal-Fired Gas Trubine Research
Project - Test of New Turbine Blade Design. Bureau of Mines.
U. S. Department of Interior. RI 6920. 1967.
Stettenbenz, L. M. Minimizing Erosion and Afterburn in the Power
Recovery Gas Turbine. Oil and Gas Journal. 6_8:65-70, 1970.
Atkin, M. L. Australian Coal-Burning Unit. Gas Turbine International.
September-October 1969. pp. 32-36.
361
-------
APPENDIX K
GAS TURBINE CORROSION/EROSION PILOT-PLANT TEST PROGRAM
K
-------
APPENDIX K
GAS TURBINE CORROSION/EROSION PILOT-PLANT TEST PROGRAM
INTRODUCTION
Exxon Research and Engineering is under contract to EPA for
the design, construction, and operation of a small-scale, high-nressure,
fluidized-bed boiler, 0.63 MW-equivalent miniplant to obtain information
for the design of a high-nressure fluidized bed boiler demonstration
plant. Westinghouse was responsible for designing and constructing an
erosion/corrosion test rig for installation in the discharge line from
the miniolant. The test rig was shinned in October 1973; minor modifica-
tions have been made at Exxon. In addition, Westinghouse is responsible
for formulating a plan for the erosion/corrosion rig tests.
This aooendix
• Discusses the technical basis for the design of the erosion/
corrosion test rig
• Describes the test rig in detail
• Describes the apparatus and orocedures for samoling
oarticulates
• Presents a plan for tests to be conducted in the test rig
• Describes an analytical orocedure for interpreting the test
results.
BACKGROUND INFORMATION
The design conditions for the Exxon high-oressure fluidized bed
boiler miniplant are as follows:
e Pressure - 1013 kPa (lO atnO
• Products temperature - *70 to 927°C
(1600 to 1700°F)
363
-------
• Airflow rate - 0.680 kg/s
(1.5 Ib/sec)
• Coal feed rate - 0.059 kg/s
(0.13 Ib/sec)
• Products flow rate - 0.739 kg/s
(1.63 Ib/sec)
• Gaseous products molecular weight - 29.0
Two stages of cyclone separators are located in the discharge
line from the fluid bed combustion miniplant. Exxon has estimated that
the concentration of fly ash in the product gases from the second-stage
cyclone will be 0.0224 kg/m3 (1.40 x 10"3 lb/ft3) and that the size dis-
tribution of these particles will be as shown in Figure K-l. No esti-
mate has been made of the concentration or size distribution of the
sorbent particles which will probably be present in the combustion pro-
ducts because of the attrition of the bed material.
The target for maximum solids concentration in the gas stream
2
at the inlet to a Westinghouse gas turbine is as follows:
Total Solids - 3.43 E-04 kg/sm3 (< 0.15 gr/scf)
3
Particles > 2 pm - 2.88 E-05 kg/sm (< 0.01 gr/scf)
Experience with erosion in gas turbines indicates that the erosion caused
by particles with diameters less than 2 pm is negligible and that the allow-
able concentration of particles with diameters greater than 2 ym should be
less than about 2.288 x 10~ kg/sm (0.01 gr/scf). At a pressure of
1013 kPa (10 atm) and a temperature of 870°C (1600°F), the above concen-
-33 -33
trations are equivalent to 0.881 x 10 kg/m (0.055 x 10 lb/ft ) total
solids and 0.0593 x 10~3 kg/m3 (0.0037 x 10~3 lb/ft3) solids over 2 pm
diameter.
For the high-pressure fluidized bed boiler system preliminary
design presented in Reference 3, the estimated concentration of total
solids in the gas-turbine inlet was 0.737 x 10 kg/m (0.046 x 10~ lb/
ft ) with the size^distribution given in Figure K-l. This gives a con-
centration of 0.0737 x 10"3 kg/m3 (0.0046 x 10"3 lb/ft3) solids over 2ym
364
-------
in
.1
1
10
15
25
«- 35
i>45
o>
5,55
1*65
I
Curve 679183-1)
93
96
98
99
Projected Particle Size
Distribution Leaving Secondary
Collectors in Full-scale High -
Pressure Fluid Bed Boiler Plant
I I I I I I I II
Note: Elutriation of attrited sorbent particles
was not considered in this prediction.
I
1 2 3 4 5 6 7 8910
Particle Size, urn
100
1000
Figure K-1-Predicted size distribution of coal ash particles escaping from secondary cyclones
in the Exxon miniplant
-------
diameter. These values were subsequently revised to values approximately
4
three times those originally estimated.
Table K-l compares the specified preliminary design and revised
preliminary design concentrations with the predicted concentration of ash
particles in the Exxon miniplant. The original predicted concentrations
for the high-pressure fluidized bed boiler are essentially equal to the
specified values, and the revised values are about three times the ori-
ginal. The estimated total ash concentration for the Exxon miniplant is
about 25 times the specified level and 10 times the revised design level.
For those particles greater than 2 ym in diameter the predicted concen-
tration for the Exxon miniplant is over 300 times the specified value and
almost 100 times the revised design value. The expected addition of at-
trited sorbent particles to the ash particles in the Exxon miniplant will
increase these ratios to even higher values.
One possible method for reducing the solids concentrations in
the Exxon miniplant is the addition of a third-stage separator. In order
to reduce the total solids concentration to the revised system design
value, this separator would have to have an efficiency of 90 percent. A
96 percent efficient third-stage separator would be required to reduce
the concentration to the specification level.
A second possible method of reducing the solids concentration
in the gas stream going to the erosion test rig is by diluting it with
products of combustion of a clean fuel at the same temperature level.
The use of a sorbent for in-bed sulfur removal in the fluidized
bed combustion of coal limits the temperature of the bed to 927 to 982°C
(1700 to 1800°F). With bed temperatures of this level in the high-
pressure boiler, the temperature of the products of combustion going to
the gas turbine will be in the range of 871 to 927°C (1600 to 1700°F).
This temperature level coincides with the maximum temperature for uncooled
first-stage turbine vanes. In this application, therefore, the first-
stage turbine vanes would most likely be uncooled, and the vane and blade
materials would both be operating at near gas-stream temperatures.
With the conditions which prevail at the discharge of the mini-
plant (871°C/1600°F, > 15 percent excess air), all turbine vane and blade
366
-------
u>
Table K-l
COMPARISON OF PARTICIPATE CONCENTRATIONS AT GAS-TURBINE INLET4
Target
for HPFBB"
HPFBB
desiznc
Revised
HPFBB
design"
Exxon
miniplante
Total, kg/m3(lb/ft3)
Particles > 2 urn,
0.83xlO~ (0.055xlO~3)
[1.0]
0.74xlO~3(0.046xlO~3)
[0.83]
2.24xlO"3(0.14xlO~3)
[2.5]
[1.0]
22.4x10 (1.40x10" )
[25.5]
[10.1]
kg/m3 (lb/ft3)
0.059xlO"3(0.0037xlO~3)
[1.0]
0.074xlO"3(0.0046xlO~3)
[1.24]
0 . 22x10" (0 . 014x10 )
[3.8]
[1.0]
19.2xlO~3(1.20xlO"3)
[325]
[186]
Conditions - 1013 kPa (10 atm) and 1127°C (2060°F) .
Evaluation of Fluldlzed Bed Combustion Process. Office of Research and Monitoring. Environmental Protection Agency.
Westlnghouse Research Laboratories. Pittsburgh, Pa. November 1971. Vol. II, p. 156.
CIbid, p. 285.
Yang, W. C., and D. L. Realms. Partlculate Removal Studies from High-Temperature, High-Pressure Gases. Westinghouse
Laboratories. Pittsburgh, Pa. Report 73-9E3-COCLN-R1 , April 25, 1975.
Letter from H. S. Nutkls to J. R. Hamm dated August 8, 1973.
-------
materials will be oxidized. If no corrosive contaminants are present in
the gas stream and the velocity is low, a stable oxide film of 50 pm or
more thickness will be formed in a period of about 50 hours, and the
metal recession rate will approach zero. The metal oxidation rate is
parabolic with time because of the diffusion characteristics of the oxide
films.
If contaminants are present which cause the oxide film to spall,
the oxide film will not stabilize and the recession of base metal will
continue indefinitely. High-velocity gas streams can have similar
effects on the oxide film. NASA-Lewis has conducted oxidation tests,
wherein sonic jets of clean gas impinged on the metal surface, which
showed that metal recession rates did not diminish with time.
The presence of particulates in high-temperature gas streams
will have a similar effect on the oxide films on turbine vanes and blade
materials. If the erosion rate is high, so that the metal oxide Is removed
as soon as it is formed, erosion of the base metal will occur, and the
metal recession will be due to the combined effects of the base metal
oxidation rate and the particle erosion. If the erosion rate is low
enough so that a minimal oxide film can be maintained, the film thickness
will range between near zero and the stable film thickness. Metal
recession will be due to oxidation, and erosion will take place in the
oxide film rather than in the base metal. The rate of metal recession
will be a function of the oxide film thickness under equilibrium conditions.
Erosion of solids by ^articulates in gas streams is primarily
a function of the impact velocity and the physical characteristics of both
the particles and the solid. The Impact angle for maximum wear rate is
a function of the target material type. Brittle materials have maximum
rates at impact angles near 90 degrees and ductile materials at near
30 degrees. The turbine vane and blade materials which will be used in
the erosion tests are typically ductile materials. The metal oxides
which form on these materials are brittle at low temperatures. Their
characteristics at operating temperatures, however, are not well known.
368
-------
DESIGN OF THE EROSION/CORROSION TEST RIG
The initial concent of the static erosion test rip; was patterned
after the design used by the British Coal Utilization Research
Association (BCTIRA) and is shown in Figure K-2. A cascade of turbine
vanes is installed in a duct with typical vane inlet conditions. Targets
are placed in the high-velocity gas stream downstream of the vane cascade
which simulates the conditions at the leading edge of the turbine blades.
These targets are placed normal to the vane axes so that a representative
sample of the average conditions for the moving blades can be obtained.
The use of multiple vanes and targets permits the simultaneous testing
of several turbine vane and blade materials.
A test rig of this type was designed for the conditions in
the Exxon miniplant (see Figure K-3). The complete drawing list for this
rig is as follows:
• General assembly 4879D29
• Outlet assembly 256C008
• Outlet welding assembly 4879D32
• Outlet elbow assembly 255C987
• Inlet assembly 255C986
• Blade pack detail and assembly 4879D31
• Insert details 255C988
• Line details 4879D31
R. W. Hornbeck made studies of particle dynamics in a turbine
stage, which showed that the size of turbine vanes and blades has a sub-
stantial effect on the trajectory of particles within the turbine stage
and, therefore, would affect the factors which control erosion. A cas-
cade of the type shown in Figure K-3 with four free-standing vanes has
vanes which are considerably smaller than those in current utility-type
gas turbines when designed for Exxon test conditions. Even if the number
of free-standing vanes were reduced to one or zero, the vane size would
be unrealistically small.
369
-------
122-183 m/s
(400-400 ft/ sec)
Figure K-2- Cascade type erosion test rig
370
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
The minimum practical diameter for a cylindrical target is con-
sidered to be about 6.35 mm (1/4 in). With four free-standing vanes only
one target of this diameter can be used without causing choking at the
target; with no free-standing vanes two 6.35 mm (1/4 in) targets could be
used.
Since the results obtained on a cascade rig for the Exxon mini-
plant would be of questionable value because of the reduced scale, and
simultaneous testing of several vane and blade materials would not be
possible, it was concluded that a cascade-type erosion test rig was not
feasible for this application. A straight-through erosion/corrosion test
rig, therefore, was designed for the Exxon miniplant conditions (see
Figure K-4). The complete list of drawings for this test rig is as fol-
lows:
Dwg. No. Figure No.
• Test section assembly 256C112 K-4
• Test rig assembly 256C111 K-5
• Inlet elbow assembly 256C113 K-6
• Outlet elbow assembly 256C114 K-7
• Straight run assembly 724B684 K-8
• Liner details 4880D54 K-9
• Liner details 4879D31 K-10
• Cylindrical target detail 6208A13 K-ll
• Wedge target detail 6213A46 K-12
• Cooled target detail 6213A44 K-13
The straight-through erosion test rig consists of a bell-mouth
nozzle followed by a length of straight duct to provide adequate time for
the particles to accelerate to near gas-stream velocity ahead of the
target. Here, again, a 6.35 mm (1/4 in) diameter cylinder is considered
to be the minimum practical size for the erosion target. A 6.35 mm
(1/4 in) diameter target located in the straight section would limit the
velocity to less than 457 m/s (1500 ft/sec) because of choking in the
plane of the target. Since this velocity is too low, the erosion target
373
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
iriinni .•!> • nnn r-r i rrr /
*» mi fit* ^tanmtei t
LJ
•vl
VO
Figure K-6
t» *.J
-------
u>
CO
o
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
IK
Westinghouse Electric Corporation
TITLE
RESEARCH LABORATORIES CHUHCHILL »o»o FITT»»U»IOM. ft. H118 U*A
DESCRIPTION - MATERIAL
DIMENSIONS IN INCHES
REF DWG
on
PART NO
SVM
OR
.0000$
Figure K-ll
TOLERANCES
IJNLCtf OTHJRWI8E •PECIFIIO
.02
'I QIC
,OO3~
iO.9'
FINISH TO BE \x^ UNLESS
OTHERWISE SPECIFIED
RB. ', i I RM-3
387
-------
1
6
it
•
3
X
IN!
0
-
1
^N
West!
TITL
s
in
u
i-
/
JL
w
r
' _S
1
nghouse Ei
.E £'/2-0g'/
7~£~ST £S=>£~C/ss)e-sJ
DESCRIPTION -MATERIAL. REF DWG SYM
DIMENSIONS IN INCHES PARING. OR 1 2 3 4 3
MA7t. r& Se s/zecs^/£Z> £r
-------
' J!
Westlnghouse Electric Corporation
TITLE £ftOS/OfiJ
RESEARCH LABORATORIES CHUHCMILL lo»o FITTCIUHGM P» ittlf U«A
T£ST
DESCRIPTION - MATERIAL
DIMENSIONS IN INCHES
REP DWG
on
PART NO
SYM
CR
TtsBf
Figure K-13
TOLERANCES
UNLCM OTHIMWKt IPCCIPIIO
02
± 009
±0.9'
FINISH TO BE v' UNLESS
OTHERWISE SPECIFIED
62I3A44
389
-------
is located in the diffuser section where the gas stream velocity will
reach about 579 m/s (1900 ft/sec) at the design conditions for the Exxon
miniplant.
The piping between the exit of the second-stage separator and
the test section and the test section itself,is lined with oxidation-
resistant material to prevent particles of refractory lining from getting
into the gas stream.
The maximum allowable material recession for the turbine vanes
and blades in the large Westinghouse utility-type gas turbines is about
0.762 mm (0.030 in). For a 25,000-hour target turbine life, the allow-
able metal recession rate is therefore 0.0305 ym/hr (1.2y in/hr). A
study has been carried out to learn what combination of erosion target
design and metal recession measuring technique will determine whether the
above metal recession rate is exceeded in the least possible test time.
Two types of erosion targets were considered—the cylindrical
rod (Figure K-ll) and the wedge target (Figure K-12). The cylindrical
target has the advantage, theoretically, of giving erosion as a function
of impact angle over a full spectrum of impact angles (from 0 to
90 degrees), whereas the wedge-type target will give erosion data for a
relatively narrow range of impact angles. This is particularly signifi-
cant since it is anticipated that erosion will be taking place in the
oxide layer, and there is no information available on the impact angle
for the maximum erosion rate of the oxide layer.
With the wedge-type probe, tests with a minimum of three angles
would be necessary to determine the impact angle for the maximum erosion
rate. An advantage of the wedge target is that it would give a more pre-
cise measurement of the erosion rate at the impact angle of maximum ero-
sion rate.
Three techniques for measurement of erosion effects were con-
sidered. These are:
• Frofilometer
• Optical comparator
• Weight loss.
390
-------
The lubrication laboratory of Westinghouse Research Laboratories
has a Gould Model 360 Surfanalyzer (profilometer). This instrument can
measure surface roughness and profile on flat or cylindrical surfaces down
to a minimum of 0.0254 pm (1 p in) per chart division. A datum surface,
however, is required in order to measure material removal. It is not possible
to maintain a datum surface on a wedge-type target. The use of a profilo-
meter, therefore, is not feasible with this type of target.
Two simulated wear tests have been conducted to determine the
resolution possible with the profilometer on a cylindrical erosion target.
Sample targets were prepared with a surface finish of 0.025 to 0.075 pm
(1 to 3 p in), and a cylindricity of about 1 pm (40 p in). An attempt was
made to simulate from 0.25 to 0.75 pm (10 to 30 p in) of erosion over a
180 degree arc by liquid honing. It was found that a minimum application
resulted in the removal of about 9 pm (350 p in) of material. A second
sample target had a nominal erosion of 0.75 pm (30 u in) simulated by
electroetching. Before and after profiles taken on these two sample
+12
targets indicate that 1.5 + 0.3 pm (60- p in) of metal removal can be
measured with a 95 percent confidence level by this technique. At the
maximum allowable metal recession rate, this amount of material would be
removed in about 50 hours of operation.
The metallurgical laboratory of the Westinghouse Research
Laboratories has an optical comparator which they use to measure metal
recession of corrosion specimens from the corrosion test rig in the
combustion laboratory. This instrument has a resolution of about 0.025 mm
(0.001 in). Use of this technique requires sectioning the specimen. At
the maximum allowable metal recession rate, this amount of material would
be removed in about 800 hours of operation.
Precision balances are available with capacities adequate for
measuring 6.35 mm (1/4 in) erosion targets with a resolution of 1 mg
or better. Since weighing gives a measurement of total weight change,
this technique is only applicable to the wedge-type target where the weight
change would be roughly uniform over the exposed surface area. A resolution
of 1 mg is equivalent to a metal recession of 0.4 to 0.5 pm (15 to 20 p in).
391
-------
At the maximum allowable metal recession rate, this amount of material would
be removed in about 15 hours of operation.
As stated earlier the predicted concentration of ash particles
in the exit gases from the second-stage cyclone separator of the Exxon
miniplant is about 25 times the estimated concentration of total particulates
in the full-scale, high-pressure fluidized bed boiler plant. In addition,
the predicted size distribution of the particles in the miniplant exhaust
is substantially larger than that projected for the commercial-scale plant.
Because of the higher impaction factor of the larger ash particles on the
target, the ratio of ash erosion effects will be greater than the concentra-
tion ratio. When the probable effect of attrited sorbent particles is
added in, it is indicated that the erosion tests in the miniplant will be
accelerated by a factor substantially greater than 25. In view of this,
significant erosion effects would probably be obtainable with the cylindrical
targets in test times of several hours. With the wedge-type target, test
periods required for significant results would be of the order of one hour.
Test periods of several hours are substantially less than those
required for a stable oxide film to form on the high-temperature turbine
vane and blade materials. Because of the characteristic hyperbolic oxidation
rate, however, the amount of oxidation which will occur during test periods
of this magnitude is significant. This will result in a weight gain over
the unexposed (to erosion) surface of the wedge-type target and a change
in elevation of the unexposed surface of the cylindrical-type target. To
prevent these changes, the unexposed surfaces of both types of targets will
-3
be coated with a thin layer (minimum of 1 x 10 um) of oxidation-resistant
material such as platinum.
The coals which will be used in the Exxon miniplant tests will
contain corrosive contaminants such as sulfur, alkali metals, and chlorine.
Some of these contaminants will be present in the combustion products as
condensible vapors. Therefore, the amount of these contaminants which will
be deposited on the erosion/corrosion targets by condensation is a function
of the target temperature. An analysis of the heat transfer along the target
shank (see Figure K-ll) showed that the exposed surfaces of the target will
be at near gas-stream temperature. In cases where the turbine vanes and/or
392
-------
t'lades are cooled, the metal surface temperatures are, of course, below
the gas-stream temperature. Tests of cooled cylindrical targets can
be carried out in the erosion test rig by the use of the tubular target
design shown in Figure K-13.
The ultimate measure of vane and blade life is the metal
recession rate. Measurements of both the wedge- and cylindrical-type
probes will be made, therefore, after removal of the oxide scale. The
condition of the oxide surface, however, will also be of interest, so
measurements will be made prior to descaling.
As noted earlier, the metal recession rate is dependent upon the
combined effects of erosion and corrosion. Therefore, tests in which the
erosion is accelerated by increased particulate concentration (as will
be the case in the Exxon miniplant) may not give realistic metal
recession rates. In view of this, tests should also be made with targets
which have been preoxidized in order to provide some insight into the
situation where the erosion effect is relatively small and a nearly stable
oxide film exists.
HIGH-PRESSURE HIGH-TEMPERATURE PARTICULATE SAMPLING SYSTEM
Westinghouse has designed and constructed a high-pressure, high-
temperature particulate sampling system for use in the Exxon miniplant
erosion/corrosion tests.
Probe Assembly
A sampling probe — 6.35 mm (1/4 in) OD, 4.57 mm (0.18 in) ID
Inconel 600 tube — is to be located in the straight pipe section which
follows the secondary cyclones of the Exxon miniplant.
The orobe subassembly will be mounted on a blank flange attached
to a 76 mm (3 in) nipple, as shown in Figure K-14. The probe is held by
a fitting which allows it to be retracted and rotated as required.
Connected to the probe is a tee to provide a pressure tap; a heat ex-
changer to quench the gas sample; and an isolating ball valve.
393
-------
- 15"
Location
Probe
Pitot
Pitot Assembly
Dwg. 625^26
AP Taps
from Pitot
AP Taps
from Flow Meter
AP Indicator
Selection Valves
UJ
Refractory
"Plug"
3" Nipple
Inconer
Lining
1/4" Ball Valve
— 1/4" Conax
Standard
3" Flanges
Probe
Pitot Static
Figure K-14-Probe arrangement
-------
A second port is provided in the flange for a static pressure
probe which can be used in conjunction with the sampling probe to deter-
mine gas velocity and, thus, allow isokinetic conditions to be estab-
lished.
High Flow Rate Sampling System
When the high flow rate Andersen impactor is used, the total sam-
pling time available is of the order of 30 seconds. For accuracy it is
thus necessary to preset the sampling flow carefully before sending it
to the impactor. Serious errors would result if flow adjustments were
made during the brief sampling period.
The sampling system thus incorporates two flow loops of
equivalent pressure drops: one used when setting the sampling flow to
isokinetic conditions, the other used while sampling. The arrangement of
these circuits is shown in Figure K-15.
The sampling system incorporates a flat bottomed "scalping"
cyclone which will collect essentially all particles above 20um (prevent-
ing them from interfering with impactor operation). This is followed by
a pressurized Anderson impactor, a flow meter, and a flow control valve.
The alternative loop comprises a pressure drop equivalent to that of the
sampling equipment and is connected to the flow meter and control valve
which are common to both circuits.
Low Flow Rate Sampling System
The low flow rate Brink impactor assembly is shown schemati-
cally in Figure K-16. The gas stream from the sampling probe passes
through a scalping cyclone and then splits into two streams. One stream
is sent to the impactor, its flow rate controlled by a critical orifice;
the other passes through a flow meter and control valve which are used to
establish isokinetic conditions.
The impactor used in this assembly has been modified from the
standard Brink design. Stages A and 5 (which classify submicron particles
395
-------
Dwg. 625^25
r
Flow
-* Meter
Control
Valve
U
I
~i
Andersen
Impactor
Ball Valve
Figure K-15-Andersen impactor assembly
396
-------
Dwg.
"Brink"
Impactor
Flow Meter
Control
Valve
Critical Orifice
Figure K-16-Brink impactor assembly
397
-------
under normal operation) have been removed and two additional steps added
which extend the size range of the instrument into the 5-to-10 ym
range.
Experience with High Flow Rate Impactors
Cascade impactors have been selected for sampling and sizing
particulate dispersion, as they give a more direct measurement of size
distribution than instruments which rely on collection-redispersion
techniques.
The initial work on particle size determination was carried
out with Andersen cascade impactors. To overcome problems associated
with particle "bounce" from collection plate surfaces, the plates were
covered with specially slotted glass fiber filters to retain impacted
deposits. Experience showed that significant weight losses were
encountered when these filters were handled. The exact cause of the
weight loss was not determined but was probably due to loss of filter
material where the filters were cut by the Andersen interplate gaskets.
In subsequent experiments Andersen instruments were used without the
glass collection mats. Results were unsatisfactory.
A comparative study of particle size determinations, using
both Andersen and Sierra impactors, showed that the Sierra instrument
consistently gave a coarser distribution than did the Andersen. In
addition, the Sierra distribution corresponded closely to the particle
size distribution of the initial test dust.
Scanning electron microscope examination of dust deposits on
the collection stages of both instruments indicated that the Sierra
impactor was measuring close to the real size distribution, whereas
the Andersen impactor was indicating a distribution finer than that of
the test dust.
398
-------
TEST PLAN FOR EROSION/CORROSION TEST RIG IN EXXON MINIPLANT
The following test plan is proposed for the erosion/corrosion
tests in the Exxon miniplant.
1. Target Preparation (Westinghouse Research Laboratories)
a. Cylindrical target
Diameter - 6.35±0'025 mm (0.25±0-°01 in)
Surface finish - 0.025 ym (In in) rms
Cylindricity - 0.75 ym (30y in)
Plate unexposed surfaces with oxidation resistance material
t 1.3 ym (50u in) thick.
Make profile measurements at three stations.
Weigh.
*
Preoxidize.
*
Make profile measurements at three stations.
Weigh?
b. Wedge target
Diameter - 6.35±0'025 mm (0.25±0-°01 in)
Surface finish - 0.13 ym (5y in) rms
Cylindricity - 1.3 ym (50y in)
Plate with oxidation resistance material ^ 1.3 ym (50y in) thick.
Grind flats at specific angle to 0.2 ym (8y in) rms finish.
Weigh.
*
Preoxidize.
Weigh.
2. Nominal Test Conditions
Pressure - 1013 kPa (10 a tin)
Temperature - 871°C (1600°F)
Gas flow rate 0.739 kg/s (1.63 Ib/sec).
In the event that operation at 1013 kPa (10 atm) is not possible, the
flow rate should be adjusted to maintain a constant value of W /~T~/P.
8
Only for preoxidized specimen.
399
-------
3. Gas Stream Measurements
The dust concentrations in the gas stream leading to the test
section shall be determined by procedures outlined in ASME Power Test
Code 27-1957 or the equivalent. The sodium vapor and chlorine content of
the gas stream should also be determined.
4. Properties of Particulate Matter
The size distribution of the partlculate matter collected
from the gas stream at the inlet to the test section shall be determined
by procedures outlined in Section 5 of ASME Power Test Code 28-1965 or
the equivalent.
5. Test Procedure
a. Install test target in passage so that target does not extend
into gas stream, and secure pressure fitting.
b. Start up fluid bed combustor and operate at near atmospheric
pressure until conditions are stabilized.
c. Loosen compression fitting(s) and move target into gas stream
until stop is reached. Align probe accurately and tighten
compression fitting. Increase pressure level as rapidly as
practicable.
d. Take sample of particulates from gas stream.
e. Operate at steady-state conditions for three hours with cylindrical
targets (both uncooled and cooled) and for one hour with wedge-
type target.
f. Take sample of particulates from gas stream.
g. Reduce system pressure quickly to 101.3 kPa (1 atm).
h. Loosen compression fitting(s) and retract exposed section of
target into well. Secure compression fittings.
i. Shut down fluidized bed boiler so that target can be removed
from passage. (In removing target from passage, compression
fitting should be dissembled so that target can be removed
without damage to exposed section.)
400
-------
6. Target Measurements (to be made at Westinghouse Research Laboratories)
a. Cylindrical targets
Weigh.
Make profile measurements at three stations.
Descale specimen.
Make profile measurements at three stations.
Weigh.
b. Wedge targets
Weigh.
Descale specimen.
Weigh.
Test Runs
Run
1
2
3
A
5
6
7
8
9
10
11
12
Material
X-45
MARM-509
U710
IN738
it
Co
Ni
Target
Uncooled cylinder
Uncooled cylinder preoxidized
Cooled cylinder
Wedge*
Uncooled cylinder
Wedge*
Uncooled cylinder
Uncooled cylinder preoxidized
Cooled cylinder
Wedge*
Uncooled cylinder
Wedge*
K
Wedge angles will be determined from tests on cylindrical
targets. Wedge targets may or may not be preoxidized.
401
-------
ANALYTICAL PROCEDURE FOR EROSION TEST RESULTS FROM EXXON MINIPLANT
The erosion test rig which has been fabricated for use with the
Exxon fluidized bed boiler miniplant consists of a 28.19 mm (1.11 in)
diameter passage with a 6.35 mm (1/4 in) diameter target placed normal
to the direction of flow. After each erosion test, the amount of material
removed from the upstream surface of the cylindrical target will be
measured by a profilometer. A procedure has been developed for analyzing
the results of these erosion tests in the Exxon miniplant.
The rate of erosion of a target placed in a flowing gas stream
with entrained particles is a function of the following factors:
• The concentration of particles in the stream
• The size distribution of the particles
• The angle of impingement of the particles on the surface of
the target
• The velocity of the particles at impact
• The physical properties of the particles
• The physical properties of the target material.
C. E. Smeltzer et al. have shown that, for a given angle of
impingement and for impact velocities greater than the threshold values,
the amount of material removed from a target is proportional to the mass
of the particle and independent of the particle size. Figure K-17 shows a
2
plot of the threshold energies MV as a function of particle size based
on test results from Reference 7.
The range of gas-stream velocity which will be used in the
miniplant erosion tests extends well beyond the threshold levels shown
in Figure K-17. The erosion effect, therefore, can be assumed to be inde-
pendent of particle size for a given impingement angle. The trajectory
of entrained particles in a gas stream flowing around a cylinder, however,
is a function of particle size. Thus, the impingement angles and particle
concentrations at specific points over the upstream surface of the
cylindrical target will be a function of the particle size distribution
in the free stream.
402
-------
Curve
CM
O
-------
The procedure which has been developed for analyzing the erosion
test results from the miniplant includes:
• A program designed to calculate particle trajectories
• A preliminary model for calculation of a weighted average
angle of impact with respect to a polar coordinate 6 (to
designate a point on the target surface).
• A preliminary model for calculation of the total mass striking
a unit area per unit time with respect to 8.
This procedure has been applied to the ash particle size dis-
tribution and concentration at the discharge from the miniplant as esti-
mated by Exxon (See Figure K-18). This estimate does not consider
particles of attrited sorbent which will be present.
Particle Trajectories
A particle carried by a stream of gas will tend to cross stream
lines when approaching a cylinder normal to the flow. The extent to
which this occurs is, for a given velocity, a function of the size of the
particle. For instance, a small particle will be deflected more than a
large particle, given the same point of origin in the free stream.
Using the size distribution given in Figure K-18, ten average
sizes were selected at the midpoint of each ten percent of the total mass
of particles in the flow. A point of origin was chosen at a distance of
five cylinder radii upstream of the center of the target. Here, the flow
can be assumed to be undisturbed by the cylinder (See Figure K-19).
For each particle size, a number of free stream positions were
chosen on the y-axis. A trajectory was computed for each particle size
at these positions. If the particle struck the target surface, the point
of impact and the angle of impact (with respect to the x-axis) was
computed. By varying the value of (r/r0) it was possible to determine
for each particle size the point in the free stream outside of which the
particle would pass by the cylinder.
404
-------
.1
1
15
25
,-45
fas
a>
Nl
'£85
-------
Particleljrajecto
X = - 5 To
-r0
Figure K-19 -Analytical model for particle impaction
406
-------
Table K-2
IMPINGEMENT CRITERIA
Particle Size
(ym)
1.4
2.1
2.7
3.3
3.7
4.4
5.1
6.3
8.0
12.5
(r'/ro) = criterical point of origin
C1125
Q280
a 390
a 485
0.535
OL605
•OL660
0,725
a 795
a 880
Within these limits, a plot of the point of impact for each
particle size as a function of the free-stream point of origin was made.
(See Figure K-20.) On this plot the arrow designates both the limit in
the free stream beyond which the particles will miss the target and the
point on the target surface beyond which no particle of that size will
strike. This occurs where the angle of impact approaches zero.
Concentration of Particles
The estimated concentration of particles in the gas stream at
_3
the outlet of the secondary cyclone in the Exxon miniplant is 1.40 x 10
kg/kg (1.40 x 10~ Ib/lb) gas. The total free stream mass flux through
a plane at x = -5 is 2.53 kg/m ro.519 Ih/ft2)/s. TMs 1« used to
determine the mass of particles actually striking the target surface per
unit time. This is an important parameter in the determination of
erosion loss.
407
-------
•o
o
oo
Q}
on
-------
Angle of Impact Distribution
The data provided by the computer program include:
• The trajectory of the particle with respect to time
• The particle velocity (x- + y-components) with respect to time
• In the event of impact, the coordinates of the impact point
(x,y) and the angle 6 of the particle's trajectory with
respect to the x-axis. (See Figure K-19'.)
Although these quantities define the event of impact, they do
not represent the simplest means of dealing with the phenomenon of
erosion. Erosion is a function of the angle of impact. In this model
this is the angle a, or the angle from the tangent line at the point of
Impact, at which the particle strikes. It is also more convenient to
define the point of impact by the polar coordinate 6 = arctan y/x. There-
fore, a is given by a = 90 - 6 - 6 and can easily be computed.
For each increment of surface on the cylinder, there will be a
range of particle sizes striking at angles dependent upon the size and
point of origin. It should be possible, therefore, to compute an average
angle of impact for each point on the cylinder surface (defined by 3)•
Since the distribution of impacts over the surface from 6 = 0° to 8 = 90°
is not the same for every particle size, the average must be weighted by
some factor related to size.
Since each particle size used in the computation represents ten
percent of the total mass flow of particles, the free-stream concentration
for each size is equal: C0. = C0/10.
The concentration at the surface of the cylinder is lower than
the free-stream concentration by a factor of r'/r0, the critical point of
origin for a particular particle size. Small particles will fan out more
than large particles. A weighting factor was derived which took into
account this expansion. The free-stream interval from y = 0 to y = r'/r0
was divided into ten equal segments. The concentration of particles was
assumed to decrease by a factor of r'/r0 as it moved to the target surface.
The weighting factor was calculated for each A3 where AfJ. is the segment
of the cylinder surface corresponding to the i free-stream interval.
409
-------
The weighting factor is given by:
18
w _
fi Co
These factors were computed for 0<8<90 and plotted, assigning Wf for each
surface interval to a gm at the midpoint of AB.. This developed a curve,
with particle size as a parameter, of W. versus B. (See Figure K-21.) The
weighted average angle of impact was then computed by the formula
for ten points on the surface of the cylinder. This, in turn, generated
a plot of a versus B. (See Figure K-22.)
Concentration Distribution
The mass striking the cylinder per unit time is another factor
in erosion and is easily calculated. The mass flow of each particle size
has been shown to be ten percent of the total particle concentration.
Since Wf . = C. /C0. , the mass flow at the surface is given by
Ci - Wfi C'i
where C8. = C0/10.
Thus, the total mass striking at a certain point is given by
10 10
*r (-T-) - E "fi AH - AI± E wfi
ft sec 1=1 i^l
where
QJ.O (2.53 kg/m2sec)(0.10 (Q519) Ib/ft2sec^
.253 kg/m2secX0.0519 Ib/ft2sec)
410
-------
i iirvo B.S.-J!»-A
P(rad)
0 .2 .4 .6 .8 1.0 1.2 1.4 1.6
i i i r i i i
i l i i i
0 10 20 30 40 50 60 70 80 90
Figure K-21-Weighting factor Wf = 18(r/ro> vs p
71 Ap
.2 .4
P
-------
Ib
This generates a plot of Mj (ft;2 gec) versus 8. See Figure K-23.
This concentration distribution factor will be used to normalize
the metal recession quantities measured on the erosion targets after a
test run on the basis of particle mass concentration per unit of surface
area. This will then permit a correlation of normalized metal recession
as a function of a.
Except for the trajectory computations, all of the procedures
described were carried through by hand and are thus subject to inaccuracies,
The next phase will involve a modification of the trajectory computations
to produce more convenient quantities and the writing of a program
that will arrive at a more accurate analysis of the model.
CONCLUSIONS
The straight-through erosion/corrosion test rig designed for
installation in the Exxon Research and Engineering high-pressure fluidized
bed boiler miniplant will simulate the most severe conditions of impact
velocity, impact angle, and temperature to be found in a gas turbine in
the first generation high-pressure fluidized bed boiler. However, the
predicted volumetric concentration of particulat.es in the discharge
stream from the miniplant may be more than 25 times greater than the
level anticipated in the full-scale plant with the present two stages of
particle collection. In addition, the particle sizes in the miniplant may
be substantially greater than those expected in a commercial plant. The
addition of a third-stage unit to investigate fine particle removal
should minimize the dust concentration.
The high concentration and large size of the particulates in the
miniplant may pose a nroblem in the interpretation of the erosion/cor-
rosion test results because of the unrealistic time relationship between
erosion and corrosion. Consideration may have to be given to modification
of the miniplant to attain particulate concentration and sizes which are
more representative of those values expected in the commercial-scale plants.
412
-------
Curve 678529-A
0 .2 .4
in
CM
.1
0
p(rad)
.8 1.0 1.2 1.4 1.6
C= Ib/ft sec Striking the
Cylinder Surface
P= Polar Coordinate on Cylinder
Surface
I I
1.1
1.0
.9
.8
.7
CM
.3
.2
.1
0 10 20 30 40 50 60 70 80 90
Figure K-23-Mass impacting per unit area per unit
time vs
413
-------
REFERENCES
1. Letter from M. S. Nutkis to J. R. Haram dated August 8, 1973.
2. Archer, D. H., et al. Evaluation of Fluldized Bed Combustion
Process. Office of Research and Monitoring. Environmental
Protection Agency. Westinghouse Research Laboratories. Pittsburgh,
Pa. November 1971. NTIS PB 212 916. Vol. II. p. 156.
3. Ibid. p. 285.
4. Yang, W. C., and D. L. Keairns. Particulate Removal Studies from
High-Temperature, High-Pressure Gases. Issued as part of Contract
report to Office of Coal Research. Clean Power Generation from Coal.
Contract OCR-14-32-0001-1223. Westinghouse Research Laboratories.
Pittsburgh, PA. Report 73-9E3-COCLN-R1, NTIS PB-234 188.
•t>ril 25, 1973.
5. Information obtained by W. E. Young during visit to NASA-Lewis
Laboratories.
6. Analysis done under OCR Contract 14-32-0001-1514 and reported in
monthly progress reports.
7. C. E. Smeltzer et al. Mechanisms of Metal Removal by Impacting
Dust Particles. Journal of Basic Engineering. Sept. 1970.
414
-------
APPENDIX L
POTENTIAL FOR ADVANCED STEAM CONDITIONS
WITH FLUIDIZED BED COMBUSTION BOILERS
L
-------
APPENDIX L
POTENTIAL FOR ADVANCED STEAM CONDITIONS
WITH FLUIDIZED BED COMBUSTION BOILERS
Fluidized bed combustion boiler power plants can potentially
increase overall power plant efficiency by increasing the steam tempera-
ture and pressure. The effect of higher steam temperatures on the per-
' 1 2
fonnance of the plant has been reported previously. ' An increase of
38°C (100°F) in both superheat and reheat temperatures will give a reduc-
tion of about 422 kJ/kWh (400 Btu/kWh) in plant heat rate. The use of
advanced steam conditions may be attractive should the potential reduc-
tion in boiler tube material corrosion/deposition be realized. The com-
mercial realization of advanced steam conditions in a fluidized bed
combustion system will depend on the trade-offs between increased cycle
performance, fuel costs, Increased component costs, and plant reliability.
STEAM-TURBINE DESIGN AND OPERATION
Currently, the most prevalent design steam conditions for con-
ventional steam plants are 16,548 kPa/538°C/538°C (2400 psig/1000°F/1000°F)
and 24,130 kPa/538°C/538°C (3500 psig/1000°F/1000°F). In the past, 566°C
(1050°F) and 593°C (1100°F) steam temperatures were common, but experience
to date has shown that the economics of steam plants with steam tempera-
tures greater than 538°C (1000°F) are not competitive with those at 538°C
(1000°F).
The plants in the United States which were designed with super-
critical cycles and steam temperatures above 593°C (1100"F) are now being
3 4
operated at reduced steam conditions. The Eddystone No. 1 unit ' was
designed for 34,475 kPa/649°C/5660C/566°C (5000 psig/12000F/10500F/1050°F)
and is now being run with a pressure somewhat less than 34,475 kPa
(5000 psig) and a superheat temperature of 613°C (1135°F). American
415
-------
Electric Power's Philo No. 6 was designed for 31,028 kPa/621°C (4500 psig/
1150°F) and is now operating at 31,028 kPa/566°C (4500 psig/1050°F).
Avon 8, Cleveland Electric Illuminating Company, was designed for inlet
conditions of 24,133 kPa/593°C (3500 psig/1100°F) and is now operating at
24,133 kPa/566°C (3500 psig/1050°F).
The primary technical limitations on operating steam turbines
at steam temperatures above 593°C (1100°F) are:
• High-temperature creep of steam turbine rotating and
stationary parts. The use of austenitic materials in
place of ferritic materials can extend the steam tem-
perature range above 566°C (1050°F), but these mate-
rials have the undesirable characteristics of low
yield point, low thermal conductivity, and large
thermal expansion coefficient. These have led to
thermal distortion of turbine housings.
• Low-cycle thermal fatigue of steam-turbine parts.
The use of the heavier sections associated with high-
pressure/high-temperature equipment tends to increase
the severity of this problem. Also, the expected
change from base-load to intermediate-load operation
for fossil-fired steam plants as nuclear plants
become more numerous will increase the degree of tem-
perature cycling. Low reliability compounds this
problem because of the increased number of shutdowns.
A steam power plant workshop sponsored by the Electric Power
Research Institute (EPRI) in 1974 reached the following conclusions re-
garding the most advanced, highly reliable plant which could be built
without research and development:
• A reliable 1000 MW single reheat unit could be built
with steam conditions of 24,133 kPa/538°C/538°C
(3500 psig/1000°F/1000°F).
416
-------
• A reliable double reheat unit could be built for
steam conditions of either 24,133 kPa/538°C/552°C/
566°C (3500 psig/1000°F/10250F/1050°F) or 24,133 kPa/
566°C/566°C/566°C (3500 psig/1050°F/1050°F/1050°F) in
maximum sizes ranging from 750 to 1000 MW.
• A reliable 24,133 kPa/566°C/593°C (3500 psig/1050°F/
1100°F) unit could be built with a rating in the 600-
to-800 MW range.
The limit on maximum unit size for the above steam conditions
is imposed by the unavailability of forging equipment for the higher
strength materials.
The consensus of those attending the EPRI workshop was that al-
though the technology exists for construction of reliable plants at the
above listed steam conditions, the increased costs of the steam genera-
tors and steam turbine for these higher steam conditions balance out the
fuel savings associated with the lower heat rates. There is thus no
economic advantage to using the higher steam conditions.
BOILER TUBE MATERIALS
The technical limit on operating the boiler at steam conditions
above 593°C (1100°F) is the superheater and reheater tube wastage due to
liquid-phase ash corrosion. In conventional pulverized coal-fired
boilers this problem increases in severity as the steam temperature in-
creases above 538°C (1000°F).
The use of fluldized bed boilers with in-bed desulfurization
1 5—8
can potentially overcome this constraint on advanced steam conditions. '
The ash corrosion problem for the boiler superheater and reheater tubes
is substantially less in the fluidized bed boiler with in-bed desulfuri-
zation than in the conventional pulverized coal-fired boiler. In addi-
tion, the increased hot-side heat transfer coefficient in the fluidized
bed boiler results in a substantial reduction in the boiler tube surface
requirements, which will tend to make the use of steam temperatures in
the 538-to-593°C (1000-to-1100°F) range economically viable.
417
-------
This is particularly true for the pressurized fluidlzed bed
combustion boiler-combined cycle power plant, where the heat transfer
surface is reduced 60 to 70 percent over that of a conventional
pulverized-fuel boiler.
ASSESSMENT
Short-term experimental data have been obtained on boiler tubes
in fluidized beds of limestone and dolomite burning various coals. These
data indicate that the boiler tube material limitations may be overcome
by utilizing fluidized bed combustion boilers. Long duration tests—
2000 hours or longer—are required to confirm these data. Additional
materials tests are required to investigate advanced steam conditions
with alternate materials.
Large expenditures on materials research and development pro-
grams also are required in order to attain the technology necessary for
the design and construction of reliable steam turbines with steam tempera-
tures in excess of 649°C (1200°F) with or without fluidized bed combus-
tion.
Alternative steam cycles and apparatus should also be investi-
gated. For example, the temperature limitation on boiler tube material
may be reduced if boiler tubes unstressed because of balanced pressure
are used (the super-reheat concept, Appendix A). Trade-offs between
cycle performance, power plant component costs, fuel costs, and compo-
nent reliability must be identified and assessed. Energy costs must be
projected, operating characteristics reviewed, and control requirements
identified to provide a basis for comparison with other advanced power
generation concepts.
418
-------
REFERENCES
1. Keairns, D. L., D. H. Archer, R. A. Newby, E. P. O'Neill, and
E. J. Vidt. Evaluation of the Fluidized Bed Combustion Process.
Volume I. Environmental Protection Agency. Westinghouse Research
Laboratories. Pittsburgh, Pennsylvania. EPA-650-2-73-048a
NTIS PB-231 162/9. December 1973.
2. Elliott, D. E. ,and E. M. Healey. Some Economic Aspects of High
Temperature Steam Cycles. NTIS 214 750. EPA AP-109. (Proceedings
of Second International Conference on Fluidized Bed Combustion.
1970.)
3. Campbell, C. B., C. C. Frank, Sr., and J. C. Spahr. The Eddystone
Superpressure Unit. Westinghouse Electric Corporation. Lester,
Pennsylvania. 1957.
4. Williamson, R. B. Turbine Generator Unit Leads Heat-Rate Parade.
Electrical World. March 11, 1963.
5. Archer, D. H., et al. Evaluation of the Fluidized Bed Combustion
Process. Office of Air Programs. Westinghouse Research Laboratories.
Pittsburgh, Pennsylvania. Contract 70-7, NTIS PB 211-494, PB 212-916,
PB 213-152. November 1971.
6. Reduction of Atmospheric Pollution. Volumes 1, 2, and 3. Environmen-
tal Protection Agency. National Coal Board. London, England.
Contract CPA 70-97. NTIS 210-673, 210-674, 210-675. September 1971.
7. Pressurized Fluidized Bed Combustion. R&D Report No. 85. Interim
No. 1. Office of Coal Research. National Research Development
Corporation. London, England. Contract No. 14-32-0001-1511.
NTIS PB 235-591. 1974.
8. Dainton, A. D.,and D. E. Elliott. (Presented at Seventh World Power
Conference. Moscow. 1968.)
419
-------
APPENDIX M
FLUIDIZED BED COMBUSTION TEST FACILITY
M
-------
APPENDIX M
FLUIDIZED BED COMBUSTION TEST FACILITY
INTRODUCTION
A pressurized fluid bed boiler development plant was
conceived to demonstrate pressurized fluidized bed boiler operation
under a previous contract to the Environmental Protection Agency (EPA).
Preliminary designs, cost estimates, experimental program, schedule,
and program alternatives Jiave been reported. The development plant
design incorporated the pressurized fluidized bed boiler design concept
2
proposed for commercial plants. The plant objectives were to provide
the capability for studying fluid bed combustion and heat transfer,
steam generation, solids feed and handling, particulate removal, gas-
turbine performance, boiler control, and sulfur dioxide and nitrogen
oxide emission control.
EPA extended the concept for the development plant during
1974 to incorporate greater flexibility. Goals for the test facility
included provision of:
• Complete environmental data on fluidized bed combustion
processes — including conditions with first-and second-
generation concepts
• Necessary environmental data on an experimental scale
such that data can be directly applied to demonstration
and commercial plants
• Necessary data on a schedule compatible with commerciali-
zation of fluidized bed combustion technology
• Direct dissemination of information to U.S. industry,
government, and other interested institutions involved or
interested in developing fluidized bed combustion technology.
421
-------
Westinghouse was invited to submit a proposal for a program
to design, construct and operate a multipurpose, environmental test
facility. A proposal was submitted in November 1974. While this work
was not carried out as part of the Westinghouse-EPA contract, it is
related. The essence of the proposed work is included to provide
perspective and to document some of the thinking subsequent to the
development plant design reported in December 1973.
SCOPE
The test facility would have the ability to investigate
• Fluidized bed combustor designs — fluidized bed boiler,
recirculating bed boiler, and adiabatic combustor concepts
• Sulfur removal systems — high utilization once-through
limestone/dolomite, regenerative limestone/dolomite processes,
alternative sorbent processes
• Particulate removal systems — cyclonic devices, granular
bed filters
• Auxiliary equipment — solids feed systems, instrumentation
• Gas-turbine performance — corrosion, erosion, deposition
tolerance
• Alternative fuels — coals, coal chars, low-grade petroleum
fractions, refuse
• Advanced cycle concepts — supercritical steam cycles,
liquid metal-topping cycles.
To meet these objectives, a test facility was proposed which
includes the systems indicated in Figure M-l.
Two fundamental design philosophies should be considered
for the flexible test facility during the conceptual design study:
• The utilization of a single pressurized fluidized bed
combustor module capable of investigating fluidized bed
422
-------
Owg. 1671811*
Sorbents Handling
Systems
Sorbent
Fuel
Fluid Bed Combustion
Systems
Flue Gas
Particulate Removal
Systems
Fines
NX
Fuel Handling Systems
I
Turbine
Cascades
Regenerated Sorbents
*-Working Fluid
—»-To Scrubbing
Facility
Is^
SS-Sf
&r
Gas Monitoring
Systems
To Stack or
Quench Drum
Gas-Turbine
Test
Systems
- ». To Stack or
Quench Drum
Sorbent Regeneration
Systems
T
Reducing Gas
—- Solid Stream
— Gas Stream
...*Jfa
Spent Sorbent
Utilization & Disposal
Figure M-l -Overall schematic of the flexible test facility
-------
boiler, recirculating bed boiler, and adiabatic
combustor concepts
• The utilization of two pressurized fluidized bed combustor
modules served by one common set of auxiliaries — one
vessel designed as a fluidized bed boiler and the second
designed as a recirculating bed boiler that is capable of
conversion to an adiabatic combustor. The utilization
of two modules with a common set of auxiliaries has the
advantages of
- Increasing the real-time test capability of the
facility with only an incremental increase in plant
cost
- Providing the added capability of simulating multibed
operation.
Since the two-module design provides greater flexibility, it
was selected for the conceptual design which follows.
CONCEPTUAL DESIGN
Fluidized Bed Combustion Systems
A two-module test facility would consist primarily of two
pressurized vessels nominally with an inside diameter of 3.7 m (12 ft).
Vessel 1 is 23.8 m (78 ft) high and is designed as a fluidized bed
boiler. Vessel 2 is 12.2 m (40 ft) high and is designed as a recirculating
bed boiler, with the capability of subsequent conversion to an adiabatic
combustor. Addition of the second vessel without the internals is
estimated to cost approximately $150,000, or about 1 percent of the
projected plant cost. With the internals for the recirculating bed
boiler, the cost will increase to about 5 percent of the plant cost;
this, however, doubles the real-time test capability. The three boiler
design concepts initially considered are discussed in the paragraphs
that follow.
424
-------
The facilities would be designed for maximum flexibility of
operation. The two vessels would be sited adjacent to one another and
served by a common set of auxiliaries. Upstream would be units for
receiving, handling, and feeding fuels and dolomite (or an alternative
sorbent) to the vessels, and a compressor for supplying air for
combustion and fluidization. Downstream would be a spent dolomite
treatment and disposal system, a particulate removal system for
cleaning the vessel fuel gas, and a gas-turbine test facility and by-
pass dump quench drain for handling the product gas.
Fluidized Bed Boiler
Conceptual designs for fluidized bed steam generators for
electric power generation using a combined cycle plant have been
2
carried out for 320 and 635 MW stations. These each consist of four
sets of fluidized bed boilers, and each set contains four stacked
fluidized beds as illustrated in Figure M-2. The beds are in parallel
with respect to the airflow and in series with respect to the steam/water
flow, and are contained in an unllned pressure vessel. The walls of the
beds are water cooled, and banks of boiler tubes are immersed in the
beds. The four beds carry out the functions of preevaporator, first
superheater, second superheater, and reheater respectively.
Vessel 1 in the test facility would be capable of simulating
the operation of any one of these fluidized bed boiler units at the
actual size envisaged for the 320 MW station. In the 3.7 m (12 ft)
inside diameter pressure shell there would be one 1.5 by 2.1 m (5 by
7 ft) fluidized bed unit similar to the previous Westinghouse-Foster
Wheeler test plant design snown in figure M-3. The range of operating
parameters which could be accommodated are outlined in Table M-l. The
operating conditions in the 320 MW commercial design are presented in
Table M-2 for comparison. The 24 m (78 ft), vessel height allows for
fluidized bed depths up to 9.1 m (30 ft) which is twice the maximum
bed depth allowed for in the conceptual commercial design.
425
-------
Reheated Steam
Reheatet Bed
Superheated Steam
Superheater Bed
][ ][ I
I Gas
Superheater Bed
Pre evaporator Bed
Feed Water
110 Ft.
t
Provision
.for Carbon
Burn-up
Cell*
PUNT VESSEL
JIZE OjAMETER, D
320 mw 12 Ft.
635 mw 17 Ft.
Dtpendeil in opintiii ttnlitlgns:
IIMI1 lir. III dpIS |is iilotil)
Grade Elevation
ELEVATION
PRESSURIZED FLUIDIZED BED STEAM GENERATOR
FOR COMBINED CYCLE PLANT
FOUR (4| REQUIRED
Figure M-2—Pressurized fluidized bed steam generator
for combined cycle plant
426
RM-59131
-------
Table M-l
DESIGN BASIS FOR PRESSURIZED FLUIDIZED BED BOILER
Pressure
Temperature
Gas Velocity
Bed Area
Bed Depth
Ca/S
Heat Transfer Coefficient
Particulate Carry-over
Particle Size Coal
Particle Size Dolomite
Site
Air supply
Excess air
Air preheat
Feedwater temperature
Heat transfer surface
Coal feed
Sorbent feed
101.3-2026 kPa (1-20 atm)
816-1093°C (1500-2000°F)
1.8-4.6 m/s (6-15 f/s)
^3.25 m2 (-\<35 ft2)
1.22-9.15 m (4-30 ft)
1-6 for once-through
2-10 for regeneration
285 W/m2/°K (50 Btu/hr-ft2-°F)
22.9-68.7 gm/m3 (10-30 gr/scf)
-6.35 mm x 0 (-1/4 in x 0)
-6.35 mm x 0 or 6.35 mm x 0.6 mm
(-1/4 in x 0, or 1/4 in x 30 mesh)
Existing power plant
Separate air compressor, motor drive
>160% capability
To 316-454°C (600-850°F)
To 110-260°C (230-500°F)
Capability for testing water walls,
preheat, evaporation, and superheat
tube bundles
Up to 9979 kg (22,000 lb)/hr
Up to 16,194 kg (35,700 lb)/hr
427
-------
Table M-2
FLUID BED OPERATING CONDITIONS IN 320 MW COMMERCIAL DESIGN
10
00
Fuel flow, Air
ke(lb)/min ke (lb
Preevaporator 8,020 87
(17,680) (192
Superficial
Flow Flue gas, velocity,
)/min ke(lb)/min mm/sec (ft/sec)
,137 93,532
,100) (206,200) 2.74(9.0)
First superheater 5,942 64,729 69,718
(13,100) (142,700) (153,700) 2.07(6.8)
Second superheater 4,872 53,026 57,108
(10,740) (116,900) (125,900) 1.71(5.6)
Reheater 5,620 61
(12,390) (134
,145 65,863
,800) (145,200) 1.95(6.4)
Bed
depth,
m(ft)
3.90
(12.8)
3.35
(11.0)
3.35
(11.0)
4.36
(14.3)
Bed
temp . ,
954
(1750)
954
(1750)
954
(1750)
954
(1750)
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
The simplified process flow diagram and material balance for
the fluidized bed boiler, Vessel 1, in the test facility are given in
Figure M-4 and Table M-3. The same auxiliaries will be used for
operation of Vessel 2.
Commercial boiler design operating conditions (Table M-2)
require that the bed be operated at higher gas velocity, higher bed
pressure, higher temperature, and with a deeper bed than the test
facilities currently available(for example, the British Coal Utilization
Research Association (BCURA) unit and the Exxon miniplant ' ). A
comparison of the parameters of the commercial design and those of the
existing pressurized pilot-scale experimental units is presented in
Table M-4. in addition to the difference in physical size, spme design
parameters, such as bed height/bed diameter ratio and bed area/coal—
feeding nozzle, are difficult to duplicate in small-scale combustors.
There are also differences in the experimental results obtained from
the small-scale combustors. For example, Exxon found a large temperature
gradient in the order of 167°C/m (300°F/ft) bed depth in their initial
bed design, but BCURA observed only a 17°C (30°F) temperature difference
in their 1.37 m (4.5 ft) bed. All the design parameters listed in
Table M-4 may contribute to this discrepancy. This points to the
importance of providing a large-scale test facility to investigate the
design, construction, and performance of the proposed commercial boiler
plant,at the proposed operating conditions,so that commercial feasibility
may be assessed. The flexibility features provided in Vessel 1 are:
• Capability of operating the fluidized bed boiler up to a
9.1 m (30 ft) bed depth. This changes the bed height/
diameter ratio.
• Capability of changing the number of coal-feeding nozzles.
This alters the variable bed area/coal-feeding nozzle.
• Capability of varying the coal feed rate, which changes
the heat release rate per unit bed area and the heat
release rate per unit bed volume.
431
-------
Dwg.
Fluidized Bed
Combustor
Return to Utility
Saturated Water or
Steam from Utility
_ ®_
t
I
n
-*-To Quench
Drum
Figure M-4 - Material balance for the fluidized bed combustion boiler
-------
Table M-3
MATERIAL BALANCE FOR THE FLUIDIZED BED COMBUSTION BOILER
Scream
no . Compc
1 Coal
2 Dolomite
3 Air
Flow rate,
ment kg/hr(lb/hr)
9,979
(22,000)
16,207
(35,730)
108,410
(239,000)
4 Spent dolomite 9,526
(21,000)
5 Flue gas
124,286
(274,000)
6 Fines recycled 1,315
(2,900)
7 Fines
8 Fines
9 Gas to gas
649
(1,430)
254
(560)
turbine 27,216
(60,000)
10 Gas to turbine cascade 68,584
(151.200)
Approx.
temp.
°C(°F)
Ambient
Ambient
260
(500)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
Approx.
pressure,
kPa(psia)
1216.3
(176 . 4)minimum
1216.3
(176.4)minimum
1206
(175)
1206
(175)
1137
(165)
1137
(165)
1137
(165)
1137
(165)
1137
(165)
1137
(165)
433
-------
Table M-4
COMPARISON OF DESIGN PARAMETERS BETWEEN THE BASIC DESIGN AND THE PILOT-SCALE EXPERIMENTAL UNITS
Bed Cross-Section
Bed Height
Bed Height/Diane ter Ratloa
Tube Packing, Z bed cross-section
Z bed volume
Beat Release Rate/Bed Area
Volume Beat Release Rate
Bed Area/Coal Feeding Nozzle
Basic design"
BCURA unit
Ex»np
1.52-m x 2.13m (3.25 n )
(5 ft z 7 ft [35 ft2])
3.35 m-4.27 m (11-14 ft)
2-2.5
21.5-28.5
17-22.5
1.2-2.0 x 107 J/m2-sec
(3.8-6.3 x 106 Btu/ft2-hr)
3.6-4.7 x 106 J/D3-sec
(3.5-4.5 x 105 Btu/ft3-hr)
0.81 m2
(8.75 ft2)
0.61 B x 0.91 m (0.56 m2)
(2 ft x 3 ft [6 ft2])
1.37 m (4.5 ft)
2
17
8
3.2 x 106 J/m2-sec
(1.0 x 106 Btu/ft2-hr)
2.3 x 106 J/m3-sec
(2.2 x 105 Btu/ft3-hr)
0.14 m2
(1.5 ft2)
0.3175 m I.D. (0.08 m2)
(12.5 in I.D. [0.85 ft2]}
3.05 D-4.6 m (10-15 ft)
9.5-14.5
28
9
1.46 x 107 J/D2-sec
(4.6 x 106 Btu/ft2-hr)
3.7 x 106'J/m3-sec
(3.6 x 105 Bcu/ft3-hr)
0.08 m2
(0.85 ft2)
a For beds of rectangular cross-section, the hydraulic diameters are used.
Basic design has four fluldized beds of slightly different designs per module for preevaporatlon, superheating, and reheating.
-------
• Capability of changing Che heat: transfer surface
configuration — for example, horizontal tube versus
vertical tube configurations. This alters the pattern
of solid circulation, gas bubble size, and bed-tube heat
transfer coefficient.
• Capability of changing the boiler tube material, which
allows studies of corrosion/erosion of boiler tubes at
different steam conditions.
Recirculating Fluidized Bed Boiler
Fluidized bed combustion technology can be applied with many
different configurations and can be utilized in many different steam
2 3
and power cycles, as has been discussed elsewhere. ' Two alternative
pressurized fluid bed combustion concepts can be studied:
• A recirculating bed boiler concept
• An adiabatic combustor combined-cycle plant which
represents a modification of the base power cycle.
Conceptual designs, performance, and economics for these two concepts
were projected and compared favorably with the pressurized fluid bed
3
boiler combined-cycle plant.
The concept of a deep recirculating fluidized bed boiler is
illustrated in Figure M-5. Primary air, along with coal or low-grade
liquid fuels, is fed to the boiler at the base of an open draft tube.
The draft tube may be metallic or ceramic with water cooling and/or
steam generation. The superficial velocity of the air and combustion
gases flowing up the riser is 3.1 to 18.3 m/sec (10 to 60 ft/sec) at
the operating temperature and pressure of 704 to 1093°C (1300 to 2000°F)
and 405 to 2026 kPa (4 to 20 atm). The solids and gases emerging from
the riser pass into a bed of expanded cross-section and decreased
superficial gas velocity. The heat transfer surface may be located in
the bed section and in the downcomer region for steam generation. The
435
-------
Owg.
Cyclone
Freeboard
Down comer
Air
Down comer
Heat
Transfer
Surface
Particle
Separator
Secondary Separation
Heat Transfer
Surface
Air Primary Air
Air
Fuel Injector
Figure M-5-Deep recirculating fluidized bed boiler
436
-------
gas velocity in the bed is in the range of 0.31 to 4.6 m (1 to 15 £t)/sec.
Sorbents may be added along with the fuel for sulfur removal. Secondary
air is introduced at the base of the downcomer at a rate necessary to
permit the downward free flow of the solids. Flow of the solids in
the downcomer can be reduced or halted by reducing or cutting off the
secondary air. This capability makes it possible to adjust independently
the heat removal and heat production in the boiler.
A conceptual recirculating bed boiler design was prepared on
the basis of the boiler plant design developed by Westinghouse under
3
contract to EPA. The design selected for a 320 MW recirculating bed
boiler consisted of six individual modules of 3.7 m (12 ft) inside
diameter with one module for preheater, two modules for evaporator, two
modules for superheater, and one module for reheater. This design was
selected to permit standard shop-fabrication of each vessel.
Vessel 2 can be used to study the recirculating bed concept
with the upstream and downstream facility common with that for Vessel 1.
This two-vessel facility with common upstream and downstream auxiliaries
will substantially increase the availability of the facility. Both
concepts can be evaluated in parallel. When one unit is shut down for
modification, the other unit can be operated. The material balance for
Vessel 2 operated as a preevaporator in the recirculating bed concept
is presented in Figure M-6 and Table M-5.
Adiabatic Combustor
The pressurized fluidized bed boiler concepts operate at
excess air values from 10 to 100 percent, with up to 70 percent of the
heat released in burning the fuel with air transferred to the water/steam
in the tubes surrounding and submerged in the bed. Increasing the design
point excess air with constant bed temperature will decrease the total
heat transfer surface in the fluid bed until no boiler tube surface will
be required at an excess air of approximately 360 percent. In this case,
the power system is a combined-cycle plant with the gas to the turbine
3
expanders supplied from a coal-fired, adiabatic combustor. Combined-
437
-------
LJ
CO
t
Dwg.
r
i
G.T.
~l
Turbine
Cascade
1(9
L ^j^**' ^^
I
I '
I I
I I
-^*. To Quench
Drum
Figure M-6- Material balance for the recirculating bed concept (preevaporator)
-------
Table M-5
MATERIAL BALANCE FOR THE RECIRCULATING BED CONCEPT
Stream
no.
Component
Flow rate,
kg/hrdb/hr)
Approx.
temp.
°C(°F)
Approx.
pressure,
kPa(psia)
1
2
2a
2b
3
4
5
6a
6b
7
8
9
10
Coal
Air
Air
Air
Dolomite
Spent dolomite
(30% utilization)
Flue gas
Fines recycled
Fines
Fines
Gas to gas turbine
Gas to turbine cascade
Gas to quench drum
9,163
(20,200)
100,699
(222,000)
85,277
(188,000)
15,422
(34,000)
10,478
(23,100)
6,804
(15,000)
109,771
(242,000)
1,216
(2,680)
522
(1,150)
225
(495)
68,584
(151,200)
27,307
(60,200)
41,278
(91,000)
Ambient
260
(500)
260
(500)
260
(500)
Ambient
871
(1600)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
871
(1600)
1216.3
(176.4)minimum
1207
(175)
1207
(175)
1207
(175)
1216.3
(176.4)minimum
1216.3
(176.4)
1138
(165)
1138
(165)
1138
(165)
1138
(165)
1138
(165)
1138
(165)
1138
(165)
439
-------
cycle plants of this type, which burn natural gas and/or heavy
distillates, are now being marketed to electrical utilities for
intermediate-and base-load applications. One embodiment of such a
plant is the Westinghouse Power at Combined Efficiencies (PACE) plant.
A preliminary evaluation was made of the adiabatic combustor
fuel processing system. Two design concepts — a single fluid bed
module and a stacked fluid bed module —were considered, as illustrated
in Figure M-7. A summary of the respective design features is presented
in Table M-6. Vessel 2 can be modified to simulate a single-bed operation
in the stacked-bed design. The same auxiliaries existing for the
fluidized bed boiler can also be used here. Due to the compressor
rating, the bed will operate at about 90 percent capacity. The material
balance is summarized in Figure M-8 and Table M-7. Data obtained in this
unit are projected to be applicable for any adiabatic combustor larger
than 3.7 m (12 ft) in diameter.
Design Features
Vessel is a 3.7 m (12 ft) inside diameter unlined pressure
vessel which contains the fluidized bed boiler module as described
previously. This vessel will have an equivalent rating of 30 MW and
will represent the full-size operation of individual fluidized beds in
a multifluidized bed, 320 MW station. By operating the vessel in each
of three boiler modes in turn, piecemeal simulation of the total
commercial plant will be achieved.
Vessel 2 is designed as a recirculating bed boiler, with the
capability of being subsequently modified for operation as an adiabatic
combustor. It has refractory-lined walls of 12 ft inside diameter and a
central draft tube around which circulation of the bed material takes
place. This vessel also represents one full-size fluidized bed unit in
a 320 MW multiunit plant and could be operated in any one of the four
boiler modes. Vessel 2 uses the same auxiliaries as Vessel 1, and the
vessels can initially be operated one at a time.
440
-------
\
'\.'/. v "*"" • '* .""•»"*"•" "l"»*•'"
Air Plenum
Single Bed Design
Air
Air
Air
Air
i)wg. 6211A8R
Flue Gas
Flue Gas
Flue Gas
Stacked Bed Design
Figure M-7-Adiabatic combustor designs-Westinghouse/Foster Wheeler Design
-------
Table M-6
ADIABATIC COMBUSTOR DESIGNS
Number of modules
Number of beds per module
Module diameter, m(ft)
Module height, m(ft)
Bed depth, m(ft)
2 2
Bed area, m (ft )
Fluidizing velocity, m/sec (ft/sec)
Single bed designs
I
4
1
6.4(21)
4.9(16)
2.0(6.5)
32.3(347)
1.8(6)
II
2
1
9.1(30)
4.9(16)
2.0(6.5)
64.6(695)
1.8(6)
Stacked bed designs
I
4
3
3.7(12)
15.2(50)
2.0(6.6)
10.5(113)
1.9(6.2)
II
2
6
3.7(12)
152.4(500)
2.0(6.6)
10.5(113)
1.9(6.2)
442
-------
Owg.
OJ
Adiabatic
Fluid Bed
Combustor
Turbine
Cascade
-*- To Quench
Drum
Figure M-8 - Material balance for the adiabatic fluid bed combustor
-------
Table M-7
MATERIAL BALANCE FOR THE ADIABAT1C FLUID-BED COMBUSTOR
Stream
no.
1
2
3
4
5
6
7
8a
8b
9
10
11
Flow rate,
Component kg/hr(lb/hx)
Coal 4,808
(10,600)
Air 175,997
(388,000)
Dolomite 5,489
(12,100)
Spent dolomite 3,565
(7,860)
Flue gas 180,805
(398,600)
Flue gas to cyclones 124,286
(274,000)
Flue gas to quench drum 56,519
(124,600)
Fines recycled 1,315
(2,900)
Fines 649
(1,430)
Fines 254
(560)
Gas to gas turbine 68,584
(151,200)
Gas to turbine cascade 27,307
(60,200)
Approx. Approx.
temp. pressure,
°C(°F) kPa(psia)
Ambient 1,021
(148) minimum
323 1,010
(614) (146.5)
Ambient 1,021
(148) minimum
871 1,021
(1600) (148)
871 993
(1600) (144)
871 993
(1600) (144)
871 993
(1600) (144)
871 993
(1600) (144)
871 993
(1600) (144)
871 993
(1600) (144)
871 993
(1600) (144)
871 993
(1600) (144)
444
-------
Since the air compressor is 60 percent oversized for either
vessel, the two vessels can be operated in parallel with the airflow
to each at 80 percent maximum. In this way multibed operation as in
the commercial unit can be simulated, which is particularly useful for
developing start-up, shut-down, and control techniques. Vessel 1
would operate as a fluidized bed boiler and Vessel 2 as a recirculating
bed boiler, as before. Any consecutive pair of boiler modes could be
simulated with steam flowing in series from one vessel to the other.
A refractory-lined pressure vessel is formed by removing the
draft tube and boiler tubes from Vessel 2 and modifying the air inlet
distributor plate. This may be operated as an adiabatic combustor.
In this mode the heat of combustion is removed from the vessel as
sensible heat in the off-gas,and about 300 percent excess air is required.
With coal as the fuel, the air requirement and the product gas volume
are both twice that of either the fluidized bed boiler or the recirculating
bed boiler. Operating the air compressor at full capacity will therefore
enable the adiabatic combustor to operate at around 90 percent of its
maximum rating. The 3.7 m (12 ft)inside diameter adiabatic combustor
represents one full-size bed in a 300 MW plant, which consists of 12 such
beds.
The tube bundles in the fluidized bed boiler are designed for
easy alteration, the number of tube banks can be altered, or banks of
vertical tubes can be substituted for horizontal tubes. Appraisal of
such tube configuration can be substituted for horizontal tubes.
Appraisal of such tube configuration could most effectively be carried
out in Vessel 1.
The recirculating bed boiler is designed with vertical tubes
originating from headers positioned above the level of the fluidized
bed; this arrangement facilitates complete removal of steam-raising
equipment, and since the walls are refractory lined, Vessel 2 is most
easily converted to an adiabatic combustor.
445
-------
Vessel 1 may also be operated as an atmospheric-pressure
boiler. In this case the bed depth must be kept low to minimize bed
pressure drop, and the flue gas is no longer capable of driving a gas
turbine. The same auxiliary systems would again be utilized; the
compressor would be run at reduced pressure differential and increased
throughput. All boiler tubes except those in the bed would be removed.
Vessel 1 could also be used to assess the application of
fluidized bed boilers for use with advanced power cycles. The fluidized
bed temperature is limited to about 1000°C (1832°F) by dolomite sulphur
removal properties, but higher gas outlet tmperatures may be attained,
for example, by burning fines in the freeboard area with secondary air
injection. Also.the present heat transfer tube designs will allow for
3
higher steam temperature and pressure.
Additionally, Vessel 1 could be used to investigate the concept
of the potassium vapor cycle, in which liquid potassium replaces steam/
water in the boiler tubes and the dry saturated vapor is delivered to a
potassium vapor turbine for power generation. Such tests would involve
changing the material of the boiler tubes.
In summary, the proposed fluidized bed combustion systems combine
the characteristics of flexibility and full size simulation of fluidized
bed boilers and combustors.
This flexibility is illustrated by the following summary of
test capabilities to:
• Operate Vessel 1 as a fluidized bed boiler.
• Operate Vessel 2 as a recirculating bed boiler.
• Operate Vessels 1 and 2 in parallel to simulate multibed
operation.
• Operate Vessel 2 as an adiabatic combustor.
• Test different tube bundle configurations in Vessel 1.
• Operate Vessel 1 as an atmospheric boiler.
446
-------
• Operate Vessel 1 as a fluid!zed bed boiler with
advanced power generation cycles.
It is always difficult to decide the size for a test facility.
The basis for the size in this case was the selection of a commercial-
scale unit. In reaching this decision, several criteria were considered:
• Objectives
- Scope (e.g., first or second generation, alternative
concepts)
- Environmental control
- Component development
- Commercial scale demonstration
• Scale-up — scale-up of data to commercial operation:
What is the minimum scalable size for each component?
• Operation — continuous
- Integrated/discrete
- Time required/experiment
- Modification time
• Cost — capital
- Operating
• Construction time.
A comparison of design parameters for commercial units and for
two test units is summarized in Tables M-8 and M-9. The two test units
identified in the tables can be assessed on the basis of the above
criteria. In particular,
• Objectives — Maximum process information can be obtained
with the minimum-scale plant; the larger plant, however, is
required to develop commercial-scale operating credibility.
447
-------
Table M-8
EVALUATION OF TEST FACILITY SJZE
Equipment
Design parameters
In commercial
design
Projected
minimum
scalable size
Suggested minimum size test unit
Design
parameters
Sc.ile-up
fjc-tor
CO
commercial
size
Equivalent
capacity
Recommended test unit size
Design
parameters
Scale-up
factor
to
commercial
Size
Equivalent
capacity
00
Fluidized Bed Combustor
15 MW
1:1
30 MW
Pressure vessel size, m(ft)
Bed area, m2(ft2)
Bed height /diameter
ratio*
Tube packing,! bed
cross section
Tube packing,! bed
volume
Heat release rate/bed
area, W/m2(Btu/ft 2-hr)
Heat release rate/bed
volume , W/m2 (Btu/f 1 3-hr )
Bed area/coal-fetding
nozzle, m2(ft2)
Water walls
Regenerator
Pressure vessel size, m(ft)
Bed area, m (ft )
Bed height, m(fl)
Bed height/diameter
ratio
Partlculate Removal Sys.
Primary
Secondary, ACMM(ACFM)
Tertiary
Gas Turbine
Turbine Cascade
*
3.7(12) ID 2.4(8) ID
3.3(1.5 x 2.1) 1.9(1.5 x 1.2)
[35(5 x 7 ft)] 120(5 x 4 ft))
2-2.5 (Based on bed
height/diameter
ratio consideration)
£1 • J— £O • J
17-22.5
11.98-19.86 x 106
(3.8-6.3 x 106)
11.03-16.18 x 105
(3.5-4.5 x 105)
0.8(8.75)
Yes
2.7-3.7(9-12) ID 0.9(3) ID
5.9-13.4(64-144)
1.5-3.0(5-10)
0.5-1.0
Conv. design Conv. design
566-1132
(20.000-40,000)
7
*30 MW 3 MV
7.26 leg/sec gas
(16 Ib/sec)
2.4(8) ID
1.9(1.5 x 1.2)
[20(5 x 4 ft)]
2-6.5
10-30
10-30
7.88-23.64 x 106
(2.5-7.5 x 106)
7.88-23.64 x 105
(2.5-7.5 x 105)
0.5-0.9(5-10)
Yes
1:20-36
0.6(2) ID
0.3(3)
1.5-3.0(5-10)
2.5-5.0
Conv. design
382
(13,500) M:2
382
(13,500)
3 MW •>!()
7.26 kg/sec gas
(16 Ib/sec)
3.7(12) ID
3.3(1.5 x 2.1)
[20(5 x 7 ft)]
2-2.5
10-30
10-30
7.88-23.64 x 106
(2.5-7.5 x 106)
7.88-23.64 x 105
(2.5-7.5 x 105)
0.5-0.9(5-10)
Yes
1:9-16
0.9(3) ID
0.7(7)
1.5-3.0(5-10)
1.5-3.0
Conv. design
693
(24,500) 1:1 to 1:2
693
(24,500)
3 MW >1Q
7.26 kg/sec gas
(16 Ib/sec)
Hydraulic diameter is used when die bed is in rectangular shape.
-------
Table M-9
EVALUATION OF TEST FACILITY SIZE
Equipment
Design parameters
In commercial
design
Projected
minimum
scalable size
Suggested minimum size test unit
Design
parameters
Scale-up
factor
to
commercial
size
Equivalent
capacity
Recommended test unit size
Design
parameters
Scale-up
factor
to
commercial
size
Equivalent
capacity
VD
Recirculatlng Bed
Combustor
Pressure vessel size, m(ft)
2 2
Bed area, m (ft )
Bed height, m(ft)
Draft tube ID, m(ft)
Draft tube height, m(ft)
Tube packing,Z bed
cross section
Tube packing,Z bed
volume
Coal—feeding nozzle
Water walls
Adlabatic Combustor
Pressure vessel size, m(ft)
Bed area, a2(ft2)
Bed height, m(£c)
Bed height/diameter
ratio
Bed area/coal-feeding
nozzle, o(ft2)
Hater walls
3.7(12) ID
10.5(113)
4.6-6.1(15-20)
0.85(2.8)
3.0(10)
2C-25
20-25
1
No
3.7(12) ID
10.5(113)
2.0(6.6)
0.5
0.93(10)
No
1.8(6) ID
1.2(4) ID
15-25 MU
2.4(8) ID
4.6(50)
4.6-6.1(15-20)
0.67(2.2)
3.0(10)
20-25
20-25
1
No
2.4(8) ID
4.6(50)
2.0(6.6)
0.8
0.46-0.93(5-10)
No
7.5 MW
3.7(12) ID
10.5(113)
4.6-6.1(15-20)
0.85(2.8)
3.0(10)
20-25
20-25
1
No
3.7(12) ID
10.5(113)
2.0(6.6)
0.5
0.46-0.93(5-10)
Ho
1:1
30-50 MW
1:1
17 MW
-------
• Scale-up — Both plant sizes are specified to permit
scale-up. The larger size minimizes risk since no further
scale-up in size is required for the combustor. For
example, tube support requirements to avoid excessive tube
vibration and temperature distribution in the bed cannot
be confidently projected from a small-scale test unit.
The larger size also assumes the critical problem(s) are
being investigated, since the same problem may not occur
in a smaller unit.
• Operation — Both plants provide continuous, integrated
operation. The time required for each experiment and the
time required to make modifications are projected to be the
same for each plant. The minimum size plant is not designed
to operate the rotating turbine and test cascade simultaneously.
Thus, more experiments would be required with the small unit
to obtain the same information on turbine-blade corrosion/
erosion.
• Cost — Capital cost for the large plant (base design) is
estimated to be around $20 million. Cost for the minimum-
scale plant is estimated to be $14 to 15 million. A
demonstration plant is projected to cost in the order of
$100 million. Manpower requirements to operate the plants
will essentially be the same. Fuel and limestone costs
will be lower for the minimum-size plant.
• Construction time — The time required for operating plant
is expected to be the same for either plant size. The
time will be limited by long lead-time equipment which
will be the same for either plant.
On the basis of these comments, the only advantage of the
smaller plant is the capital cost savings 0\-$5 million) and the savings
in fuel and sorbent ($1 to 2 million over three years). The advantage
of the larger plant is the minimization of risk in guing to demonstration
450
-------
plants. The projected savings represents about $6 to 7 million. A
demonstration plant may cost $100 million or more. Potential savings
in the demonstration plant capital and operating cost could be signi-
ficant if results from the test system avoids a problem, a savings which
could easily exceed the $6 to 7 million cost addition for the larger
test system program.
Based on these considerations, a commercial-size test system
is believed to offer the greatest return of technical and economic
information needed to develop and commercialize pressurized fluidized
bed combustion systems. Data from such a system will also provide the
greatest confidence for a demonstration plant program.
Sulfur Removal Systems
On the basis of the available pilot plant data, sulfur
removal to meet emission standards can be achieved. Once-through
systems with high sorbent utilization have not been demonstrated on a
large scale, however, and economically attractive regeneration processes
have not been identified for limestone or dolomite sorbents. Alternative
sorbents are known, but again their potential has not been assessed.
Schemes to utilize waste sorbents in by-product functions or to convert
them to environmentally acceptable products are being evaluated but
have not been tested on any significant scale. The test facility would
be capable of addressing these important areas. The test facility
provides for tests on commercial-scale regenerative processes, once-
through processes, and waste sorbent processing for by-product utilization
or disposal with limestone/dolomite sorbents or alternative sorbents with
new processing schemes or reaction schemes.
The process conceptual design and operation studies carried
out in support programs will identify and characterize processing schemes
and the associated process and equipment options in the design, installation,
and operation of the test facility. Basic operating variables (temperature,
pressure, concentration, and so on), temperature control options, waste
heat utilization options, process turndown, and start-up options, would
be considered for study. In addition, processing components common to
many processes can be examined in order to identify and characterize
451
-------
gas-cleaning system options and costs. These components include the
following:
• Pneumatic transport processes for sorbent circulation
• Process-gas generation systems (for example, gasification
systems for reducing gas generation, carbon dioxide
recovery processes)
• Sulfur recovery processes (for sulfur dioxide or hydrogen
sulfide processing; for sulfuric acid production, and so
forth)
• Fluid bed reactors
• Other process components, such as high-temperature valves,
particle pulverizers, waste heat boilers, and process
Particulate Control Systems
Provision is made to test equipment for primary, secondary,
i>
and tertiary particulate collection. The three-stage system would be
capable of operating with both fluidized bed combustor test modules.
It is clear that several different dust collection systems must be
thoroughly evaluated and tested for the fluid bed combustion development
program. Specific examples are cited at this point to illustrate the
approach.
Primary collectors are considered to be conventional, state-
of-the-art cyclones which will not require further investigation.
Secondary collectors will probably be cyclone collectors.
Since broad choices are available — conventional multicyclone designs,
or specialized designs which employ clean secondary gas flows — at this
point it appears that the provision of a clean secondary gas flow may
result in a more complex and less efficient gas-turbine plant. Thus,
the secondary collector is considered to be a multicyclone unit of
conventional design.
Tertiary collectors may be required to meet stringent dust
collection efficiencies, which will limit the choice to the various
452
-------
filter systems. Both porous ceramic and porous metal filters require
back flushing with clean gas flows during the cleanup cycle. Granular
bed filters may be designed to operate without back flushing or with
cleanup by short bursts of high-pressure gas acting against the main
flow. The granular beds will be simpler to incorporate into a complete
system, and thus a unit of this type is considered attractive for the
tertiary collection stage.
Turbine Test Facilities
A test facility consisting of a rotating, multistage turbine
and several stationary test passages is recommended for the plant. This
test facility, used in conjunction with analytical studies, will allow
assessment of turbine blade erosion or deposition due to particulate
matter and hot corrosion and/or deposition initiated by alkali-metal
compounds in the gas stream. Table M-10 summarizes recommended components
for the program. The multistage test turbine will provide information
on the effects on erosion and deposition resulting from patticle con-
centrations within the turbine flow path due to passage vortex develop-
ment, radial pressure gradients, local flows due to stage-to-stage
interactions, and particulate-blading interactions. One stationary
test passage, designed with long particulate acceleration nozzles to
obtain the high-particle-impact velocities that will be typical of an
electrical utility gas turbine, will be used to correlate test-turbine
erosion to full-scale turbine erosion. An additional full-scale first-
stage turbine nozzle cascade will assess damage to this highest temperature
component. A full annular cascade passage will provide information on
the effect of the inlet design in patterning the particulates over the
flow path in such a way as to prevent localized concentrations and,
hence, prevent excessive deposition and/or erosion damage.
The turbine test facility must allow operating conditions
in a large electrical utility gas turbine to be duplicated. The potential
for condensation of volatile alkali-metal compounds — sodium and potassium
chlorides and hydroxides and subsequent reaction to sulfate or molten
453
-------
Table M-10
TURBINE TEST PROGRAM COMPONENTS
Component
Function
High-pressure, High-
Temperature, Full-Scale
Turbine Vane Cascade
High-Pressure, High-
Temperature, Full Annular
Stationary Cascade
High-Pressure, High-
Temperature Particle
Accelerator Nozzle
Impactor with Wedge
Targets
Rotating Hot Gas
Multistage Turbine
Measures vane erosion/deposition
with ability to evaluate film
cooling deposition control options
Assesses particulate patterning
Measures erosion/deposition rates
at realistic impact velocities and
angles
Verifies particle concentration
analytic model
454
-------
sulfate-chloride mixtures — can be studied if similar gas temperatures
and metal temperatures at points of corresponding pressure are duplicated
and similarity of flow conditions are established over appropriate
targets. Blading must be of hot-corrosion resistant materials so hot-
corrosion attack of the test turbine will be comparable to attack of the
large turbine. Erosion damage will be comparable if the velocities of
particulate impact with turbine components are comparable in both magnitude
and impact angle.
Rotating, Multistage lest Turbine
The recuperative version of the Solar Centaur industrial
gas turbine appears attractive for the multistage rotating test vehicle.
This machine is of the right scale for pilot-plant operation. Its gas
flow requirement — 19 kg/sec (42 Ib/sec) — is less than the bed output —
34.5 kg/sec (76 Ib/sec) -but is sufficiently large to be useful as a
demonstration of the viability of the overall concept. The inlet temperature
of 871°C (1600°F) is acceptable for use on the fluidized bed combustion
system, and the turbine can accept the required 1013 kPa (10 atm) inlet
pressure.
Corrosion resistance of the first two stages of the Solar
Turbine is roughly comparable to that of a modern electric utility gas
turbine. The third-stage stator nozzles and blades should be replaced
with those made of more hot-corrosion resistant materials such as alloys
having at least the corrosion resistance of MAR 421.
The regenerative configuration of the machine provides both
compressor outlet and hot-gas fe.ed in locations that are immediately
*
accessible. The recuperative Centaur is virtually Identical in all
important characteristics to the present production model of the Centaur
(except for the compressor outlet and hot-gas inlet) and should be
reliable.
In designing this facility, the turbine inlet design must be carefully
examined to control dust concentrations presented to the first-stage vanes.
455
-------
First-stage stationary vane erosion/deposition can be
measured in a stationary cascade, provided that full-scale vanes are
used in the passage. Such a cascade would require about 5 kg/sec
(16 Ib/sec) of gas for each nozzle passage used in the cascade. The
effect of cooling airflows on deposition and erosion could be effectively
studied in this facility.
A scaled full annular stationary cascade will allow experimental
verification of the effectiveness of inlet manifolding design on controlling
particulate concentrations over the inlet of the turbine.
This combination of test facilities allows all of the significant
deposit forming/erosion mechanisms to be studied and related to the
performance of full-scale commercial turbines at a fraction of the cost
of operating a large machine.
Operation and Control
The test facility can be used to evaluate high-pressure
fluidized bed combustion processes with particular emphasis on emissions
control, component development, efficiency enhancement, and systems
development. It also provides means for testing proposed commercial plant
operation and control procedures. These tests could demonstrate that
fluidized bed combustion power plants can meet electrical utility operating
requirements such as 2:1 turndown per module, which will permit 8:1
turndown of a four-boiler plant and be capable of changing load at the
rate of 5 percent per minute, while meeting environmental standards.
Auxiliaries
The test facility would also provide opportunities to test
alternative coal and sorbent feed systems, gas-monitoring equipment
proposed for control of particulates and trace element release, and
alternative working fluids for the boiler.
456
-------
Gas velocities relative to blading approach those of a large
electrical utility gas turbine (similar to the Westinghouse 501) at the
rotor trailing edges and at the exit to the third-stage stator. At
other locations in the turbine, however, the gas velocities are con-
siderably lower, typically reaching only about 50 percent of the
Westinghouse 501 gas velocities at blade and vane leading edges.
If the particle velocities were to follow exactly the gas
velocities, the erosion rate of the Solar blading at the rotor trailing
edges and at the exit to the third-stage stator should be comparable to
that in the Westinghouse 501, while the erosion rates at the leading
edges of stator vanes and rotor blades may be as low as 6 percent of
those at comparable locations in the full-size machine. Because of the
rapid acceleration of the gas passing through the stator vanes and the
deacceleration in the rotor blading, particles as small as 2 vim in
diameter may have too much inertia to follow the gas flow. An analysis
of this effect in a full-scale electric utility turbine showed that the
variations in particle velocities lagged the variations in gas velocity.
This effect increased with increasing particle size and resulted in
particulate velocities leaving the rotors that were significantly higher
than the gas velocity. A similar analysis of the test turbine flow path
must be made to allow the erosion data to be interpreted properly.
Stationary Test Passages
The fact that impact velocities cannot be duplicated in the
rotating test facility is the reason that the stationary passage erosion
tests mentioned earlier are required. Particle acceleration nozzles of
sufficient length (about 1.2 m [4 ft] with nozzles whose flow areas
decrease linearly with length) must be used to provide sufficient time
for particles to reach gas velocity before impacting the erosion targets.
Erosion targets must be carefully designed so that the impact angle as
well as the velocity are accurately known at the locations where erosion
rates are measured.
457
-------
By monitoring particulate levels in Che gas stream before it
enters the expansion turbine, three potentially hazardous situations can
be detected and action taken to prevent turbine damage:
• Gross failure of dust collectors resulting in very high
dust loading
t Dust loadings greater than the design values .which will
result in turbine erosion if allowed to persist over an
extended time period
• Gradual increase in loadings ,which indicates that the
dust collectors are not functioning correctly and require
maintenance.
In addition, to get maximum information about the reactions
and transport of trace elements in a fluidized bed combustion system, the
test facility should be equipped with monitoring equipment. There is a
particular need for monitoring the trace elements which are known to be
capable of forming corrosive deposits on gas-turbine hardware, notably
sodium and potassium, as well as environmentally important trace elements.
ESTIMATE OF TEST FACILITY COST
A preliminary cost estimate was prepared for the test facility
installed at an existing power plant. The major subsystem costs for the
base plant are summarized in Table M-ll. The basic pressurized fluidized
bed combustion plant includes the air supply systems; fuel-handling and
feed system; sorbent-handling and feed system; two pressurized fluidized
bed combustor test units; a single-train, three-stage particulate removal
system; and a four-element gas-turbine test system. The installed cost
(December 1974) complete with all necessary labor and materials for
auxiliaries (piping, instrumentation, electrical, foundations, structures,
and so forth), is estimated to be $21 million.
Cost estimates were also prepared for a representative
regeneration system. A one-step regeneration system was selected and
458
-------
Table M-ll
INVESTMENT ESTIMATE FOR TEST FACILITY3
Cost,
$-million
Component (12/74)
Air Supply Systems 4.7
Coal/Limestone-Handling 6.0
Systems
Fluidized Bed Boiler 4.4
Adiabatic Combustor System 0.75
Recirculating Fluidized Bed 0.25
Boiler Internals
Particulate Removal System 1.9
Turboexpander Test Systems 1.6
Auxiliary Equipment 1.4
Total Installed Systems Cost 21.0
a
Installed subsystem costs complete with all
necessary labor and materials for auxiliaries:
piping, instrumentation, electrical, foundations,
structurals, and so on.
459
-------
estimated on the basis of available design projections. The
regeneration system is estimated to cost $5 million. This includes
solids circulation, a fluidized bed regenerator, a gas producer, and
sulfur recovery.
The cost estimates are based on the preliminary design and
cost estimate for a pressurized fluidized bed boiler development plant,
subsequent conceptual design modifications to the plant, and recer.t
fuel-processing plant design and cost estimates developed by
Westinghouse.3'11"13
460
-------
REFERENCES
1. Keairns, D. L. et al. Evaluation of the Fluidized Bed Combustion
Process. Vol. III. Pressurized Fluidized Bed Boiler Development
Plant Design. Environmental Protection Agency. Westinghouse
Research Laboratories. Pittsburgh, Pennsylvania. December 1973.
EPA 650/2-73-048C. NTIS PB 232 439/3.
2. Archer, D. H. et al. Evaluation of the Fluidized Bed Combustion
Process. Vols. I-III. Environmental Protection Agency.
Westinghouse Research Laboratories. Pittsburgh, Pennsylvania.
November 1971. NTIS PB 211 494; 211 960/9; 213 152/2.
3. Keairns, D. L. et al. Evaluation of the Fluidized Bed Combustion
Process —Vols. I and II: Pressurized Fluidized Bed Combustion
i
Process Development and Evaluation, and Vol. Ill: Pressurized
Fluidized Bed Boiler Development Plant Design. Environmental
Protection Agency. Westinghouse Research Laboratories. Pittsburgh,
Pennsylvania. December 1973. EPA-650/2-73-048a, b, and c.
NTIS PB 231 162/9, 231 163/7, 232 439/3.
4. Pressurized Fluidized Bed Combustion. Office of Coal Research.
National Research Development Corporation, England. 1974. R&D
Report No. 85. Interim No. 1.
5. Monthly Progress Reports. Multi-cell Fluidized-Bed Boiler. Office
of Coal Research. Pope, Evans and Robbins Inc. Contract 14-32-
0001-1237.
6. Annual Report. Reduction of Atmospheric Pollution by the Application
of Fluidized Bed Combustion and Reduction of Sulfur-Containing
Additives. Environmental Protection Agency. Argonne National
Laboratory. June 1973. Publication No. EPA-R2-73-253. Interagency
Agreement EPA-IAG-0020.
461
-------
7. Reduction of Atmospheric Pollution. Final Report (Volume 1-3).
Office of Air Programs. National Coal Board. London, England.
September 1971. PB210 673, PB210 674, PB210 675.
8. Nutkis, M. S.,and A. Skopp. Design of Fluidized Bed Miniplant.
Proceedings of the Third International Conference on Fluidized Bed
Combustion. Hueston Woods, Ohio. 1972.
9. Exxon R&D Program. (Presented at FBC Contractors Meeting, Argonne
National Laboratories, September 11-12, 1974.)
10. Fraas, A. P. Potassium-Steam Binary Vapor Cycle with Fluidized-
Bed Combustion. (Presented at annual AIChE meeting, New York,
November 1972.)
11. Initial Design Study for 50 MW Fluidized Bed Oil Gasification
Demonstration Plant. Environmental Protection Agency. Westinghouse
Research Laboratories. Pittsburgh, Pennsylvania. Contract No.
68-02-0605. March 1974.
12. Forty-Sixth Monthly Progress Report. Environmental Protection
Agency. Westinghouse Research Laboratories. Pittsburgh,
Pennsylvania. Contract No. 68-02-0605. February 1974.
13. Personnel Communications. Westinghouse Coal Gasification Plant Cost
Study.
462
-------
TECHNICAL REPORT DATA
(Please read InOnienons on the reverse before completing)
1 REPORT NO
EPA-650/2-75-027-C
3 RECIPIENTS ACCESSION NO.
4 TITLE AND SUBTITLE
Fluidized Bed Combustion Process Evaluation
Phase TI--Pressurized Fluidized Bed Coal Com-
bust ion Development
5 REPORT DATE
September 1975
6. PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
8 PERFORMING ORGANIZATION REPORT NO
D. L. Keairns et al.
9 PERFORMING OR6ANIZATION NAME AND ADDRESS
Westinghouse Research Laboratories
Beulah Road, Churchill Borough
Pittsburgh, PA 15235
10 PROGRAM ELEMENT NO.
1AB013; ROAP 21ADB-009
11 CONTRACT/GRANT NO.
68-02-0605
12 SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Phase n Final: 6/73-12/74
14. SPONSORING AGENCY CODE
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
TThe report'gives results of a program to evaluate and develop pressurized
fluidized-bed coal combustion. The historical, technical, and economic aspects of
fluidized bed combustion (FBC) systems have been reviewed, systems analyses per-
formed, commercial plant design and cost estimates prepared, and experimental
data on the sulfur removal system obtained. Two pressurized FBC power plant sys-
tems have provided the basis for the work on system design, performance, econo-
mics, and development. The basic design and performance parameters for these two
systems are presented. The present work extends the previous work to include
collection and analysis of data on critical system parameters (e.g. , sulfur removal,
spent sorbent disposition, and trace element release); development of process options
(e.g. , particulate control); and assessment of power plant cycles and component
designs (e.g. , use of low-temperature gas cleaning, alternative cycles, and gas
turbine corrosion/erosion test rig design and construction). The report includes
an extensive bibliography.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/ClOUp
Air Pollution
Coal
Combustion
Fluidized Bed
Processing
Pressurizing
Desulfurization
Sorbents
Trace Elements
Test Equipment
Bibliographies
Gas Turbines
Air Pollution Control
Stationary Sources
Particulates
13B 07D
2 ID 11G
21B
14B
13H,07A 05B
13G
8 DISTRIBUTION STATEMENT
19 SECURITY CLASS (Thu Report)
Unclassified
Unlimited
21 NO. OF PAGES
A89
20 SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
463
------- |