PB-212 560
AFTERBURNER SYSTEMS STUDY
R. W. Rolke, et al
Shell Development Company
Emeryville, California
August 1972
DISTRIBUTED BY:
National Technical Information Service
U. S. DEPARTMENT OF COMMERCE
5285 Port Royal Road, Springfield Va. 22151
-------
EPA-R2-72-062 S.H121
August 1972
PB 212 560
Afterburner Systems Study
R.W. Rolke, R.D. Hawthorne, C.R. Garbett,
E.R. Slater, T.T. Phillips, G.D. Towell
ENVIRONMENTAL PROTECTION AGENCY
Office of Air Programs
Contract EHS-D-71-3
,-n
Emeryville, Californio
-------
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED FROM THE
BEST COPY FURNISHED US BY THE SPONSORING
AGENCY. ALTHOUGH IT IS RECOGNIZED THAT CER-
TAIN PORTIONS ARE ILLEGIBLE, IT IS BEING RE-
LEASED IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE.
-------
BIBLIOGRAPHIC DATA '• Report No. 2-
SHEET EPA-R2-72-062
4. Tide and Subtitle
Afterburner Systems Study
7. Auihor(s)
R. W. Rolke, R. D. Hawthorne, C. R. Garbett, et al
9. IVrtorming Organiz.it ion Name and Address
SHELL DEVELOPMENT COMPANY
A Division of Shell Oil Company
Emeryville, California
12. Sponsoring Organization Name and Address
ENVIRONMENTAL PROTECTION AGENCY
Office of Air Programs
Research Triangle Park, North Carolina 27711
3. Recipient's Accession No.
?p>-a i2j 3~6>o
5. Report Date
August 1972
6.
&• Performing Organization Kept.
No. S1U121
10. Pro)ect/Task/Work Unit No.
11. Contract/Grant No.
EHS-D-71-3
13. Type of Report & Period
Covered
14.
15. Supplementary Notes
16. Abstracts ^g resuits are presented of a study of afterburner or fume
incinerator technology for control of gaseous combustible emissions
from stationary sources. The scope of the study included evaluation
of current engineering technology, evaluation of existing afterburner
systems, assessment of present practices and problems, determination
of major sources and potential applications, and development of research
recommendations. The main results of this study are presented as a
handbook, allowing the potential user to be able to decide if his
particular emission is amenable to afterburning and to obtain a rough
estimate of cost and size of equipment needed. The user will also be
made aware of potential problems and recommended design features.' The
user then would deal with the appropriate equipment supplier for details
of equipment selection.
17. Key Words and Document Analysis. 17o. Descriptors
Air Pollution
Air Pollution Control Equipment
Afterburners
Catalytic Converters
Design
Design Criteria
Performance
Heat Treating Furnaces
Furnaces
Heat Recovery
17b. Identifiers/Open-Ended Terms
Stationary Sources
Air Pollution Control
Thermal Afterburners
17c. COSATI Field/Group
Construction Materials
Performance Evaluation
Expenses
Combustion
Combustion Chamber Gases
Reaction Kinetics
18. A\ailability Statement
Unlimited
19. Security Class (This
Report)
UNCLASSIFIED
20. Security Class (This
Pace
UNCLASSIFIED
21. No. of Pages
ORM NT1S-3S (REV. 3-721
THIS FORM MAY BE REPRODUCED
USCOMM-DC M852->-/2
-------
AFTERBURNER SYSTEMS STUDY
Contract EHSD 71-3
Office of Air Programs
Environmental Protection Agency
R. W. Rolke, R. D. Hawthorne, C. R. Garbett,
E. R. Slater, T. T. Phillips, G. D. Towell
Shell Development Company
A Division of Shell Oil Company
Emeryville, California
S-14121
-------
FORWARD
This study of afterburner technology was conducted under contract
EHSD 71-3 for the Office of Air Programs of the Environmental Protection Agency
with Dr. D. L. Harmon serving as project officer. The first part of this report
presents a handbook of afterburner technology and the second part presents a
review of the tasks performed, an inventory of major sources and potential
afterburner applications, and research recommendations.
In addition to the authors, Messrs. C. M. Schlaudt, D. R. Hayes,
G. A. Stenmark, J. S. Son and R. A. Brown participated in the project.
We acknowledge with thanks the cooperation and time of the staffs of
the equipment manufacturers, afterburner users and the Air Pollution Control
Districts, who provided much of the information upon which this study is based.
S-1U121
-------
ii
TABLE OF CONTENTS
Chapter
1. INTRODUCTION
2. SUMMARY
2.1 Thermal Afterburners
2.2 Catalytic Afterburners
2.3 Heat Recovery
2.4 Auxiliary Equipment - Fume Generating Process
2.5 Costs
2.6 Measuring Afterburner Performance
2.7 Major Sources of Combustible Pollutants and Potential for
Afterburner Control
2.8 Proposed Research and Development Programs
AFTERBURNER HANDBOOK
THERMAL (DIRECT-FLAME) AFTERBURNERS; DESCRIPTION AND PERFORMANCE ... 15
3~General Requirements for Satisfactory Performance 15
3-1.1 Steps Involved in Dilute Fume Incineration 16
3-1.2 Temperature and Residence Time Requirements 17
3-1.2.1 Calculation of Residence Time 18
3.1.2.2 Hydrocarbon Destruction 20
3-1-2.3 Carbon Monoxide Cleanup 21
3-1-2.4 Odors, Aldehydes and Other Oxygenated Hydrocarbons 22
3*1.2.5 Liquid Smokes and Droplets 22
3-1.2.6 Soot and Combustible Particulates 23
3.1.2.7 Nitrogen Oxides (NOX) Formation 24
3*1.2.8 Wastes Containing Chlorine, Sulfur, Phosphorous, Nitrogen,
Metals, or Other Heteroatoms 26
3-1-3 The Role of Flame Contact - Dilute Fumes Must ByPass 27
3.1.3.1 Fuel Requirements and Stream ByPassing 27
3-1.3-2 Benefits of Flame Contact 28
3.1.4 Quenching Must be Avoided - Presence of Aldehydes and CO in the
Flue Gas 29
3-2 Thermal Afterburner Design Features - Systems Utilizing
Distributed Burners 30
3-2.1 Description of Available Distributed Burners 30
3*2.1.1 Line Burners 30
3-2.1.2 Multijet Burners 31
3.2.1.3 Grid Burners 32
3.2.2 Possible Problems with Distributed Burners 33
3.2.2.1 Burner Fouling 33
3-2.2.2 Low Oxygen Content 33
3.2.2.3 Excessive Temperatures - Limit on Preheat 34
3-2.3 Afterburner Configuration - Mixing of Bypassed Fume 35
3.2.3.1 Location of Distributed Burner 35
3.2.3.2 Cross Sectional Area 35
3.2.3.3 Baffles 36
S-14121
-------
ill
TABLE OF CONTENTS (CONT)
Chapter
3*3 Thermal Afterburner Design Features - Systems Utilizing
Discrete Burners 39
3*3*1 Discrete Burners Utilized in Available Afterburner Systems .... 39
3*3*1*1 Raw Gas Burners - Nozzle Mix with Lew Pressure Primary Air
Supply 39
3*3*1*2 Raw Gas Burners - All Oxygen From the Fume 40
3*3*1*3 Raw Gas Burners - High Turndown Capability 40
3.3*1*4 Preraix Gas Burners 4l
3*3*1*5 Oil Burners and Dual Fuel Burners 4l
3*3*1*6 Use of Swirl to Obtain Rapid Heat Release 42
3*3*2 Afterburner Configuration - Mixing of Bypassed Fume 43
3*3*2.1 Axial Flame-Jet Mixing 44
3*3.2.2 Radial Entry of Fume or Flame 44
3*3*2*3 Tangential Entry of Fume or Flame - Swirl 45
3*3*2.4 Baffles 45
3.3*2.3 Bends and Changes in Combustion Chamber Cross Section 47
3.3*2.6 Packed Bed Regenerative Heat Exchange - TRAPS System 47
3*4 Thermal Afterburner Design Features - System Configuration and
Construction 49
3*4.1 Plug Flow Retention Section - Straight Through Circular Chamber . 49
3*4.2 Vertical or Stack Afterburners 30
3*4.3 Induced Draft - Avoiding Leakage 30
3*4.4 All Metal Construction - Annular Fume Inlet 31
3*4.3 Afterburner Should be Rugged and Well Designed Mechanically ... 32
3*5 Users Survey on Thermal Afterburner Performance 53
3*6 Conditions Required for Adequate Performance in Some Applications . 54
3*7 References 54
CATALYTIC AFTERBURNERS; DESCRIPTION AND PERFORMANCE 55
4.1 Introduction 55
4.1.1 Basic Elements of System 55
4.1.2 Economics Relative to Thermal Afterburners 55
4.1.3 Limitations 57
4.1.3.1 Participate Containing Streams 57
4.1.3*2 Control of Waste Generating Process 58
4.1.3*3 Performance Monitoring 58
4.1.4 Applications 6l
4.2 Performance Characteristics 63
4.2.1 Rate of Oxidation 63
4.2.2 Effect of Combustible Molecular Type 65
4.2.3 Effect of Hydrocarbon Concentration 68
4.2.4 Catalyst Deactivation and Poisoning 69
4.2.4.1 Deactivation by Aging or Thermal Processes 69
4.2.4.2 Coating of Catalysts Surfaces by Particulates or Coke 70
4.2.4.3 Specific Poisoning of Catalyst Activity 71
4.2.5 Users Reports on Catalytic Afterburner Performance 72
4.2.6 NOX Formation in Catalytic Afterburners 73
S-14121
-------
iv
TABLE OF CONTENTS (CONT)
Chapter page
4.3 Equipment Description 74
4.3.1 Catalysts 714.
4.3.1.1 All-Metal Catalysts 714.
4.3.1.2 Alumina-Base Catalysts 75
4.3.2 Catalytic Afterburner Enclosures 79
4.3-2.1 Burners fo
4.3.2.2 Enclosure Configurations 80
4.3.2.3 Materials of Construction 82
4.3.2.4 Instrumentation 82
5- USE OF FIRED PROCESS HEATERS AND UTILITY FURNACES AS AFTERBURNERS ... 85
5-1 When to Use as Afterburners 85
5«2 Manner of Introducing the Fume Stream Into the Furnace 86
5.2.1Furnaces Using Premix or Diffusion-Mixed Burners (Gas or Oil Fuel)
with Natural Draft, or Induced Draft Blowers 86
5.2.2 Furnaces with Burners Using Forced Draft Blowers 87
5.3 Hazards on Shutdown 88
5-4 Verification Tests 88
6. HEAT RECOVERY 89
671General "Considerations 89
6.2 Types of Recovery Equipment 90
6.2.1Recuperative (Gas/Gas) Exchangers 90
6.2.1.1 Heat Transfer Effectiveness 90
6.2.1.2 Cross-Flow Exchanger 91
6.2.1.3 Stacked Plate (Counter Flow) Exchanger 92
6.2.2 Rotary Regenerative Heat Exchanger 92
6.2.3 Cyclically-Operated Packed-Bed Exchangers 914.
6.2.4 Recycled Flue Gas to Process Unit 95
6.2.5 Waste Heat Recovery for Process Use 95
6.2.6 Waste Heat Boiler 96
6.3 Fouling and Cleaning 96
6.4 Thermal Expansion and Thermal Stress 97
6.5 Double Wall Construction 98
7- MATERIALS OF CONSTRUCTION 99
7.1 Scope of Materials Review 99
7.2 Refractories 99
7-2.1 Introduction 99
7.2.2 Classification of Refractories 100
7.2.2.1 Brick 100
7.2.2.2 Castable Refractories 102
7.2.2.3 Plastic Refractories 102
7-2.2A Ceramic Fibres 102
7.2.3 Selection Criteria 103
7.2.3.1 Use Temperature 103
7.2.3.2 Resistance to Chemical Attack 103
-------
TABLE OF CONTENTS (CONl)
Chapter Page
7.2.3.3 Thermal Conductivity and Density 105
7.2.3.4 Thermal Shock Resistance 106
7.2.3.5 System Geometry 107
7.2.3.6 Thermal Stress Loadings 107
7.2.3.7 Flame Impingement Problems 108
7.2.3.8 Installation Methods 108
7.2.14- Costs 108
7.2. 4.1 Materials 109
7-2.4.2 Installation 109
7.2.4.3 Maintenance 109
7*2.5 References 110
7.3 Metallic Materials 110
7.3-1 Mechanical and Metallurgical Considerations at High Temperature . Ill
7-3.1.1 Carbon Steel and l/2# Mo Steel Ill
7.3.1.2 Low Alloy Steels (l Cr, l/2# Mo to 5# Cr, 1/256 Mo) Ill
7.3.1.3 Aluminized Steel Ill
7.3.1.4 Chromium Stainless Steels (Types 405 and 4lO) 112
7.3.1.5 Chromium Nickel Stainless Steels (Types 304, 321, 347 and Cast
Stainless Steels) 112
7.3.2 Oxidation Resistance of Commonly Used Alloys 113
7.3.3 Problems of Special Chemical Environments 113
7.3.3.1 Sulfur Dioxide and Hydrogen Sulfide 113
7.3.3.2 Compounds Containing Vanadium, Sodium, Potassium, Sulfur,
Molybdenum, or Lead 113
7.3.3.3 Carbon Dioxide and Carbon Monoxide 116
7.3.3.4 Chlorine and Hydrogen Chloride 116
7.3O.5 Bromine and Hydrogen Bromide 116
7.3.3.6 Phosphorous Compounds 116
7-3.4 Further Investigation of Metallic Materials of Construction . . . 119
7.3.5 Problems of Heat Recovery Equipment and Cooled Metal Surfaces . . 119
7.3'5.1 Phosphorous Compounds 119
7.3.5.2 Sulfur Trioxide 119
7.3.5.3 Other Acid Gases 119
7.3.5.4 Steel Shell of Afterburner 120
7.3.6 Thermal Distortion and Thermal Fatigue 120
8. AUXILIARIES 123
8.1 Blowers 123
8.2 Ducts 125
8.3 Supporting Structures - Afterburner Weight 126
8.4 De-Entrainment 126
9- CONTROL AND SAFETY 127
9"7l Combustion Safeguards 127
9-1.1 National Approval Bodies 127
9.1.2 Recommended Practices 128
9.1.3 Firing and Light-Off Options 130
9.1.4 Equipment - Flame Detectors, Shut Off Valves, Pressure Switches . 130
9-1.4.1 Flame Detectors 130
9.1.4.2 Shut Off Valves 131
S-14121
-------
vi
TABLE OF CONTENTS (CONT)
Chapter Page
9.1.4.3 Pressure Switches ....................... 331
9.1.5 Typical Gas Fired Single Burner Combustion Safeguard System . . . 151
9 « 2 Temperature Control ........................ 132
9.2.1 Control Concepts ......................... 132
9-2.1.1 Oxidation Zone ......................... 132
9*2.1.2 Shutdowns and/or Alarms .................... 332
9.2.1.3 Inlet Fume Temperature ..................... 332
9.2.2 Burner Control .......................... 333
9.2.2.1 Recommended Practices ..................... 333
9-2.2.2 Equipment - Thermocouples, Controllers, Control Valves,
Over temperature or Safety Devices ............... 133
9.2.2.3 Typical Burner Controls and Alarms ............... 334
9«2.3 Inlet Fume Temperature Control .................. 335
9.2.3.1 Recommended Practice ...................... 135
9.2.3.2 Equipment - Temperature Measurement and Control, Temperature
Switches ...... „ ..................... 335
9-2. 3>3 Typical Systems . . . ..................... 136
9-3 Handling of Flammable Vapors .................... 337
9.3.1 Recommended Practice ....................... 337
9.3.2 Equipment ............................ 138
9.3.2.1 V apor Concentration Indicator ................. 338
9.3.2.2 Using the Afterburner Temperature Rise for Control ....... 139
9.3.3 Typical Vapor Concentration Control System ............ 139
9«4 Handling of Toxic Materials .................... 339
9«4.1 Recommended Practice ....................... 339
10. MEASURING AFTERBURNER PERFORMANCE-ANALYTICAL METHODS ......... l4l
lOTl Introduction ........ . ........ ..........
10.2 Sampling Considerations ...................... i42
10.3 Calibration of Instruments .................... 142
10.14- Principles Involved in Analyses .................. 144
10.4.1 Hydrocarbons .......................... l^
10. 4. 2 Oxygenated Organic Compounds, Carbon Monoxide, Carbon Dioxide . . 146
10 .4. 3 Chlorinated Hydrocarbons, Phosgene, Hydrochloric Acid ...... 146
10.4.4 Oxygen and Nitrogen ....................... 1^8
10.4.5 Nitrogen Oxides (NOX) ...................... 3.49
10.4.6 Sulfur Dioxide ......................... 152
10.4.6.1 Hydrogen Peroxide Methods ................... 152
10.4.6.2 Coulometric Methods ...................... 152
10.4.6.3 Colorimetric Method ...................... 154
10.4.6.4 Electrochemical Sensor Method ................. 154
10.4.7 Sulfur Trioxide, Sulfuric Acid ................. 155
10.4.8 Particulate Matter ....................... 155
10.4.9 Source Measurement of Odor Emissions .............. 158
10.4.9.1 Measurement of Odor Intensity Using Dilution ......... 159
10.4.9.2 Sample Collection and Dilution ................ 159
10.4.9.3 The Odor Panel-Determining the Odor Threshold ......... 159
10.4.9.4 Possibilities for Instrumental Analysis ............ 159
10.5 On-Site Instrumental Analysis ................... 160
10.6 Temperature Measurement ...................... 160
S- 14 121
-------
vii
TABLE OF CONTENTS (CONT)
Chapter Page
10.6.1 Thermoelectric Thermometry 160
10.6.2 Resistance Thermometry 165
10.6.3 Radiation Pyrometers 166
10.6.4 Other Techniques 166
10.7 Flow Pattern Determination 166
10.7.1 Use of Models 167
10.7-2 Quantitative Methods 167
10.7.3 Visual Methods 168
10.7.3.1 Static Methods . . . . * 168
10.7.3.2 Kinetic Methods 168
11. COSTS 169
11.1 Scope 169
11.2 Thermal Afterburners 169
11.2.1 Purchase Costs from Manufacturers' Data 169
11.2.2 Instrumentation 169
11.2.3 Burners 170
11.2.4 Purchase, Installed and Operating Cost from Users' Data 171
11.2.5 Fuel 172
11.3 Catalytic Afterburners 175
11.3.1 Purchase Costs from Manufacturers' Data 175
11.3*2 Instrumentation 175
11.3.3 Burner 175
11.3 A Purchase, Installed and Operating Costs 175
11.3.5 Fuel 177
11A Total Annual Cost 178
12. COMBUSTION PRINCIPLES
12.1 The Oxidation Process - A Balance Between Heat Generation and
Removal 179
12.1.1 Ignition Temperature 182
12.1.2 Combustion Limits and Flame Temperatures 183
12.1.3 Dilute Fume Must Bypass Flame 184
12.2 Mechanisms of Oxidation and Combustion 185
12.2.1 Free Radical Flame Reactions 185
12.2.2 Low Temperature Oxidation 187
12.2.3 Thermal Oxidation in the Absence of Flame 188
12.3 Oxidation Kinetics 190
12.3*1 Hydrocarbon Destruction 190
12.3.2 Carbon Monoxide Oxidation 192
12.3.3 Liquid Smokes and Droplets 195
12.3 .4 Soot and Combustible Particulates 196
12.3.5 Flame Inhibition and Combustion of Substituted Hydrocarbons . . . 198
12.3.6 Elementary Reactions - Kinetic Models 200
12.4 Mixing Processes - Control of Overall Oxidation Rate 202
12.5 Flue Gas Composition 202
12.5.1 Determining Combustion Equilibrium 203
12.5.2 Fate of Heteroatoms 203
S-14121
-------
viii
TABLE OF CONTENTS (CONT)
Chapter Page
13- BURNERS 205
13-1 Introduction 205
13-2 Equipment Types and Alternatives 205
13-2.1 Function 205
13.2.2 Discrete Versus Distributed-Source 206
13.2.3 Premix Versus Diffusion Mix Burner 206
13.2.3.1 Premix 206
13.2.3.2 Diffusion-Mixed 20?
13.2A Oil Versus Gas Fuels 208
13-2.5 Oil Atomization 209
13-2.6 Outside Air Versus Fume Oxygen Supply 209
13-2.7 Flame Stability and Turndown Ratios 210
13.2.T.1 Burner Tile 210
13.2.7.2 Target Rod and Plate 210
13-2.7.3 Fuel-Rich, Low Velocity Pocket 210
13.2.7-4 Graded Air Entry 210
13-2.7-5 Deflector Plate on Fuel Gun 210
13.2.7.6 Preheated Air 211
13.2.7.7 Holding Oil Flames 211
13.2.7.8 Excess Air Burners 211
13.2.8 Turndown Ratios 212
13.2.9 Susceptibility to Fouling 212
13.2.9.1 Vapors Carried in the Fume Stream 212
13.2.9.2 Entrained Droplets 212
13.2.9.3 Entrained Particulates 213
13.2.9.4 Combinations 213
13 «3 Flame Patterns 215
13-3-1 The Premix Burner 215
13-3-2 The Discrete Diffusion Burner 215
13-3-2-1 Effect of the Type of Fuel Used 216
13-3-2.2 The Effect of Turbulence 217
13.3.2.3 The Effects of Axial and Rotational Momentum 218
13.3.3 The Distributed Burner 221
13-3-4 Discrete Burner with Precombustor 221
13-3-5 Scale Factors and Energy Release Rates 221
13-4 installation 222
13-4.1 Setting 222
13.4.2 Piping and Controls 222
13-4.3 Initial Operation 222
14. MIXING OF FLAME AND FUME 225
14.1 Introduction 225
14.2 Characterizing Mixedness 225
14.3 Axial Jet Mixing 228
14.4 Concentric Axial Jets 229
14.5 Subdivided Parallel Jets 229
14.6 Profile Plates 230
14.7 Cross Stream Jets 231
14.8 Baffling 231
14.8.1 Bridge Wall Baffle 231
14.8.2 Ring and Disc Baffles 232
S-14121
-------
ix
TABLE OF CONTENTS (CONT)
Chapter Page
14.9 I,- Turns and U- Turns ........................ 233
Ik. 10 Annular Throat: The Eductor Principle .............. 233
14.11 Side Entry of Flame or Fume ................... 233
14.12 Mixing with Swirling Flames ................... 234
14.13 Use of Cold Air and Water Models to Study Mixing Problems .... 234
15- RESIDENCE CHAMBER ........................... 235
15.1 Introduction ........................... 235
15-2 The Plug Flow Reactor ....................... 235
15.3 Plug Flow Reactor Residence Time ................. 236
15*4 Chamber Shapes .......................... 237
15*5 Comments on Stack Arrangements .................. 238
15.6 The Use of Cold Air and Water Models ............... 238
16. MECHANISMS AND RATES OF REACTIONS IN AFTERBURNER CATALYSTS ...... 239
1671 Mechanisms' ....... .............. ....... 239
16.2 Rates of Reactions ........................ 241
16.2.1 Empirical Observations
16.2.2 Rate Expressions ........................ 24l
16.2.2.1 Chemical Rate Constants .................... 242
16.2.2.2 Mass Transfer Coefficients .................. 243
17. DESIGN METHODS FOR CATALYTIC AFTERBURNERS ............... 247
17-1 Introduction ........................... 247
17-2 Performance Calculation Method .................. 248
17-3 Recommended Correlations of Rate Parameters for Afterburner
Catalyst Matrices ......................... 251
13«3-1 Transport Coefficients ..................... 251
17.3*2 Chemical Rate Constants ..................... 251
17-4 Estimation of Performance with Partially Deactivated Catalyst . . . 253
Nomenclature for Chapter 17 .......... ...... . . . . 254
18. HEAT RECOVERY - DESIGN FUNDAMENTALS .................. 257
18.1 Introduction ........................... 257
18.2 Equations and Relations for Heat Transfer ............. 257
18.2.1 Quantity of Heat to be Transferred ............... 257
18.2.2 Heat Transfer Rate Equations .................. 260
18.2.3 Heat Transfer Coefficients ................... 260
18.2.3.1 Film Coefficients, hf ..................... 260
18.2.3.2 Wall Conductance, C^ ..................... 260
18.2.3.3 Fouling Factor, f ....................... 260
18.2.3.4 Overall Coefficient ...................... 26l
18.2.4 Heat Exchange Effectiveness, E ................. 26l
18.2.5 Number of Transfer Units, N-tu .................. 262
18.2.6 Capacity Rate Ratio ....................... 263
18.2.7 Cross Flow Exchanger Performance ................ 263
18.3 Exchanger Pressure Drop Calculations ............... 263
18.3.1 Headers ............................. 263
18.3.2 Entrance and Exit Losses in Tubes ................ 264
18.3.3 Friction Through Exchanger Tubes ................ 265
S- 14121
-------
TABLE OF CONTENTS (CONT)
Chapter
18.3.4 Pressure Drop for Flow Across Tubes 266
18.3.5 Balancing Pressure Drop Against Exchanger Size 266
18.3.6 Other Miscellaneous Equations 26?
18.4 Sizing and Exchanger - Procedure . 268
18.5 Procedure for Checking Pressure Drop^ 270
18.6 Evaluating the Performance of an Existing Exchanger 271
19- ANNOTATED LIST OF EQUIPMENT SUPPLIERS 273
REVIEW OF TASKS, MAJOR SOURCES AND POTENTIAL APPLICATIONS,
RESEARCH RECOMMENDATIONS
20. REVIEW OF TASKS 283
20.1 Evaluation of Current Engineering Technology 283
20.2 Evaluation of Existing Afterburner Systems 284
20.3 Assess Present Practices and Problems 284
20.3.1 Survey of Users 284
20.3.2 Contacts with APCD's 284
20.3.3 Questionnaire Format 285
20.3-4 Selection of Users 285
20.3.5 Replies 286
20.3.6 Treatment of Information 286
20.3.7 Follow-up Visits 286
20.1* Determination of Major Sources and Potential Applications 288
20.5 Research Recommendations 288
21. MAJOR SOURCES AND POTENTIAL AFTERBURNER APPLICATIONS 289
21.1 Summary of Investigation Results 289
21.2 Applicability of Afterburners to Air Pollution Control Pooblems . . 292
21.3 Combustible Air Pollutants; The Nature and Dimensions of the
Problem 292
21.3.1 Types of Emissions Not Covered in This Study 293
21.3.2 Emission Characterization 294
21.3-3 Potential Emissions Reduction Through Extensive Afterburning . . 294
21.4 Emission Ranking Procedure 295
21.5 Pollution Generating Activity Estimates and Projections 296
21.6 Emission Estimates and Projections; Methods and Results 297
21.6.1 Solvent Emission Rates 298
21.6.2 Oil Refinery Emissions 298
21.6.3 Gasoline Marketing Emissions 298
21.6.4 Emissions from Refuse Burning 301
21.6.5 Emissions from Carbon Black Manufacture 301
21.6.6 Emissions from Petrochemical Manufacture 301
21.6.7 Emissions from Charcoal Manufacture 305
21.6.8 Emissions from Other Sources 305
S-14121
-------
xi
TABLE OF CONTENTS (CONl)
Chapter Page
22. RESEARCH RECOMMENDATIONS 315
22.1 Thermal Afterburners 313
22.1.1 Field Test Data 313
22.1.2 Kinetics of Combustion 3lU
22.1.3 Research Afterburner - Design, Construction and Operation .... 315
22.1.3-1 Experimental Study Hydrocarbon Feeds - Research Afterburner . . 316
22.1.3-2 Experimental Study Odorous Feeds - Research Afterburner .... 316
22.1.3-3 Experimental Study Particulate Feeds - Research Afterburner . . 316
22.1.3.4 Experimental Study of Oil Firing - Research Afterburner .... 317
22.1.3-5 Experimental Study Heteroatom Containing Feeds - Research
Afterburner 317
22.1.4 Research Afterburner - Gun and Control System Development .... 318
22.1.5 Mathematical Modelling of Afterburners 318
22.1.6 Field Implementation of New Designs 318
22.1.7 Field Test New Afterburners 319
22.2 Catalytic Afterburners 319
22.2.1 Field Test Data 319
22.2.2 Kinetics of Catalytic Oxidation 319
22.2.3 Catalyst Poisoning and Deactivation Rates 320
22.2.4 New Catalyst Development 320
BIBLIOGRAPHY AND REFERENCES 323
S-14121
-------
Chapter 1. INTRODUCTION
A systems study of afterburner or fume incinerator technology for
control of gaseous (or gasborne particulate) combustible emissions from stationary
sources has been carried out. The scope of the study inclMed evaluation of
current engineering technology, evaluation of existing afterburner systems,
assessment of present practices and problems, determination of major sen, ces and
potential applications, and development of research recommendations. Information
was obtained from the published literature, equipment manufacturers, equipment
users, air pollution control agencies, research institutes and universities.
Visits were made to many of these sources of information in order to hold detailed
technical discussions and observe equipment in operation. Questionnaires were
sent to plants which were known to have afterburners in use or were likely to
have them. The respondents provided information on general performance, costs
and problems encountered with their units. Actual performance data were rarely
available. Two main classes of afterburner are considered, thermal (direct flame)
and catalytic. Disposal of liquids or solids by combustion are not included.
Flares and afterburners which are part of a solids refuse disposal system are
not included.
Tne current main uses of afterburners are for the control of odor and
smoke with some applications in hydrocarbon and CO emission control. Afterburners
are typically used to control low concentration combustible emissions. This comes
about because of the safety requirement to keep the concentrations within the
fume generating equipment below the lower flammability limit. Odor nuisance
emissions are usually very low concentrations. A combustible gaseous emission
produced at high concentrations can be used as a fuel. Alternatives to after-
burning for gaseous emissions are adsorption on a solid bed or scrubbing with a
solvent. Water scrubbing is generally not effective in controlling organic
emissions. Alternatives to afterburning of organic particulate emissions are the
use of a cyclone, filter, electrostatic precipitator or wet scrubber.
The main results of this study are presented as a handbook of afterburner
technology. The potential user will be able to decide if his particular emission
is amenable to afterburning and to obtain a rough estimate of cost and size of
equipment that is needed. The potential user will also be made aware of potential
problems and recommended design features. It is expected that the potential user
would then deal with the appropriate equipment supplier for details of equipment
selection. The handbook will be useful to equipment manufacturers and also to
operators of afterburners for performance improvement. The handbook is contained
in Chapters 1 through 20. The first part (Chapters 3 to 11) covers the
description and performance of afterburner systems. The second part (Chapters 12
to 19) covers fundamentals and design principles. An annotated list of equipment
suppliers is presented in Chapter 20. The remainder of the report contains
documentation of the methods used in collecting the information, estimates of
emissions and the applicability of afterburning to them, and recommended future
R and D.
S-1M21
-------
Chapter 2. SUMMARY
Afterburners or fume incinerators can be used to control the emission of
gaseous or small particulate air pollutants which are combustible or thermally
decompose at high temperatures. Hydrocarbons are oxidized to the products carbon
dioxide and water. This oxidation may be carried out in thermal or catalytic
afterburners, with the catalyst allowing use of a significantly lower operating
temperature. Table 2-1 lists some of the processes from which emissions have
been controlled using afterburners. Most applications have been for the abatement
of smoke, odor, and (primarily in California) smog producing hydrocarbons.
Effluent must be collected and ducted to the afterburner. In many cases, this
costs more than the afterburner itself.
Table 2-1. APPLICATIONS OF AFTERBURNERS
Adhesive tape curing
Asphalt blowing
Brake lining ovens
Cat cracker regenerator off gas
Charcoal broilers
Ceil and strip coating lines
Core ovens
Cupola furnace stacks
Deep fat frying
Fat rendering
Fiber glass curing
Herbicide and insecticide
manufacturing off gas
Lithographing ovens
Meat smokehouses
Metal coating ovens
Metal reclaiming
Pulp and paper
Packing house effluents
Paint baking ovens
Paint removal facilities
Phthalic anhydride manufacturing off
gas
Plastic curing ovens
Printing presses
Quench bath oil fumes
Resin and paint cooking
Roofing paper machine hoods
Rubber curing
Solvent degreasing
Solvent manufacturing off gas
Sulfur plant tail gas
Textile dryers
Varnish burn-off
Varnish kettles
Vinyl sponge curing
Wire enameling
It is important that a potential user work closely with an experienced
manufacturer in order to obtain a satisfactory unit. The potential problems
described in this report should be carefully considered in the design stage before
equipment selection is made. The potential user can contribute significantly to the
success of an afterburner installation by defining the concentration and flow
rate ranges carefully. A low initial cost.minimal afterburner system is a poor
investment if in a short time the unit falls apart or is incapable of meeting
the target air pollution regulations. The information in the handbook should
allow the prospective user to assess whether the system designs provided by
various manufacturers avoid the potential problem areas.
Preceding page blank
-------
2.1 Thermal Afterburners
Thermal (direct-flame) afterburners destroy combustible pollutants through
oxidation to COg and water. Temperatures of 1^00 - 1500°F are sufficient to obtain
nearly complete conversion of most substances in .1 -> -3 seconds residence time.
Destruction of most hydrocarbons occurs rapidly at 1100 - 1200°F but oxidation of
CO ->• C02 requires the higher temperatures and residence time. Dense carbon
smokes may require temperatures of up to 2000°F and/or longer residence ti es
(~1 second), but liquid smokes are destroyed nearly as rapidly as gaseous hydro-
carbons. (Gross liquid carryover into the afterburner can lead to problems, and
should be prevented through use of knockout vessels.)
This handbook is mostly concerned with dilute fume incineration.
Supplemental fuel must be burned to supply heat since the fume itself is not
combustible and often contains only a few hundred parts per million of organics.
Fuel to attain a 1^00 - 1500°F afterburner temperature cannot be premixed with
the entire fume stream since such a mixture is below the flammable limit. The fuel
must be burned separately using ~50$ of the fume for combustion air (if the fume
has a high oxygen content and is non-fouling), and the hot combustion products
(~2500°F) must be mixed with the remaining fume to achieve the ~1400°F temperature.
Mixing of bypassed fume and hot combustion gases is the most crucial
step in attaining good afterburner performance. Typically, afterburners are
designed with ~.5 second total residence time. This time is nearly all required
for this mixing step and many designs fail to complete the mixing in the distance
(time) available. Some fume escapes without being raised to a sufficiently high
temperature. To meet a performance specification (if possible at all) more fuel
must be burned than would be needed if mixing were complete. Distributed burners
are placed directly in the fume stream and divide the flame into many individual
jets or lines of flame surrounded by fume. This subdivision greatly speeds the
mixing process, and these burners are well suited to use oxygen from the fume for
combustion. The use of outside air requires an additional 30-50$ of the fuel to
be burned to heat it to l400°F. Distributed burners are subject to fouling,
have somewhat limited turndown, aren't available for use with oil fuel, may be
difficult to use with outside air, and have a few other potential drawbacks.
Therefore, many afterburners employ discrete burners which give either long or
short point sources of flame. The mixing problem is much more difficult since
there is no subdivision at the burner. Internal baffles are required in the
relatively short afterburner chambers utilized in available designs.
Many designs stress "flame contact" in an attempt to mix fume and flame
as rapidly as possible. This often leads to flame quenching and an increase in
pollutants in the fume stream since as was mentioned above, complete fume/fuel
mixing gives a noncombustible mixture. Fuel should be burned as rapidly as
possible and the hot gases should be mixed with bypassed fume. Pressure drop must
be expended in achieving good mixing through baffles and/or a long chamber.
Mixing will be faster when there is initial fume/flame subdivision.
In comparison with boilers and industrial furnaces, thermal afterburners
have low NOX emission factors. The low operating temperatures and dilution of
combustion products by fume results in stack NOX effluent concentrations of
only 20-30 ppm.
S-1U121
-------
The presence of chlorine, sulfur, metals, nitrogen, etc., in the fume
along with hydrocarbons does not rule out the use of thermal afterburners. These
components may either pass through unchanged or be converted to the corresponding
acid or oxide and will require additional flue gas treatment (e.g., wet scrubbing)
if present in high enough concentration. Nitrogen compounds in the fume would be
almost quantitavely converted to NOX in the exhaust. Also, corrosion problems
may arise and higher chamber temperatures may be required for complete destruction.
Wherever unusual components will be present in the fume, sufficient temperature
flexibility should be provided and/or prior testing should be done.
The normal configuration for construction of a thermal afterburner
involves a steel outer shell, lined with a refractory material. The purpose of
the refractory is to protect the steel shell from direct exposure to the effects
of high temperatures and corrosive materials, and to improve thermal efficiency
of the unit by limiting heat losses. The refractory serves as a thermal insulator,
lowering the temperature from a maximum of more than 2000°F on the inside of the
combustion chamber to a temperature of 250 - UOO°F at the steel shell. At this
temperature the shell retains most of its room temperature strength and is still
hot enough to prevent condensation of water vapor on the surface, thereby reducing
corrosive attack. Refractory is used in the form of bricks, castables, and dense
boards made of pressed fibers.
Refractory structures are heavy, with densities running 45 Ib/cu ft for
the lightweight insulating firebrick and castable refractories up to 185 Ib/cu ft
for high alumina materials. Refractory wall thicknesses run typically from 4 to
8 inches. This weight adds considerably to the cost and difficulty of shipping,
so that many afterburner manufacturers do not install the refractory until after
the shell has been moved into place - expecially on the large, custom-designed
units. The installation of full masonry construction calls for a relatively
skilled mason. An unskilled mason can make mistakes that will lead to premature
failure. Once in place the added weight of the masonry can be an important
consideration in the structural design of the building, especially if the after-
burner is to be installed on a roof or elevated platform. Because of its light
weight, fibre-block wall construction is being used in some afterburners. However,
this material has limited strength; care must be used in attaching it to the
metal walls, and gas velocities must be kept low to avoid the possibility of
erosive damage.
In order to avoid the heavy weight and slow heatup associated with
refractory construction, some manufacturers have designed all metal units, employing
incoming fume for cooling the outside of the combustion chamber walls. These units
have been subject to many structural failures since metals have been used at the
limit of their strength and oxidation resistance. Also thermal stresses are severe
due to the large temperature gradients and the cool wall provides a path for fume
escape without pollutant destruction.
It was previously mentioned that high concentration combustible gaseous
emissions can be used as fuel to generate process heat. It may also be possible
to use process heaters, utility furnaces, or steam boilers as afterburners for
dilute fume streams and save the purchase cost and fuel costs of a separate after-
burner. If such an existing furnace can provide sufficient residence time at a
high enough temperature and the fume meets certain criteria listed in Chapter 5>
fume can be added as part of the secondary air supply to the furnace. Uniform
treatment of fume may be difficult to achieve so performance tests should be
conducted.
S-14121
-------
Tr.ere are many poorly designed afterburners and many of lightweight
construction on the market. Many units are subject to structural failure since
therral expansion stresses can be severe and many are not capable of operation at
lUOO - 1500°F. The "smoke eliminator" operating at 700 - 800°F is merely
vaporizing liquid smoke droplets and producing lots of CO and aldehydes by quenching
of the supplemental fuel flame fired across the stack.
The user survey produced 214 completed questionnaries on thermal
afterburners. 89$ of the units were regarded as satisfactory. The efficiency of
hydrocarbon removal was reported for 66 units with two-thirds being in the 95-100$
removal range and one-third in the 90-95$ range. Generally units are adjusted
during start up and a temperature found that results in adequate removal. No
further testing is made and adequate performance is assumed if the same temperature
is maintained. Maintenance on most units is not a major problem. The main
operating problems involve safety controls, refractory linings, heat exchanger
fouling or mechanical failure and bearing failure in fans.
2.2 Catalytic Afterburners
Catalytic afterburners are an alternative to thermal incinerators as a
means for oxidizing gaseous, combustible contaminants to carbon dioxide and water.
Their successful operation is limited to a more restricted range of applications
than thermal afterburners; but where applicable, catalytic units offer the
potentials of significantly lower fuel consumption and smaller, lighter-weight
units. The basic elements of the catalytic unit are a preheat/mixing section,
designed to achieve a uniformly preheated and distributed waste stream flow, and
the catalyst bed or matrix, in which the major portion of the oxidation reactions
take place.
Catalysts in use in afterburners are typically metal mesh-mats, ceramic
honeycomb or other ceramic matrix elements with surface deposits of finely
divided platinum metals. The oxidation of most hydrocarbons and CO occur rapidly
in the range of 600 - 900°F over these catalysts (in contrast with 1200 - 1500°F
required in thermal systems).
Catalytic systems are limited to applications in which the waste stream
has a negligible particulate loading, and in which the combustible contaminant
is gaseous or can be completely vaporized in the preheat section of the unit.
Oxidation of methane presents a difficult problem for catalytic units and requires
higher than normal operating temperature.
In use, the oxidation activity of the catalyst decreases. This decline
in activity is accelerated by high temperature exposure, by particulates from the
waste stream coating the catalyst surface (inhibiting contact of catalyst and
waste stream), and by actual poisoning of the active catalytic sites by specific
contaminants in the waste stream. Performance of the unit is maintained during
catalyst deactivation by, (i) initial over-design of the catalyst bed, (ii) raising
the operating temperature, (iii) periodic cleaning of catalyst, and (iv) eventual
replacement of catalyst. Prediction of the frequency at which catalyst must be
replaced is uncertain for new use. However, for some applications, catalytic
afterburners have been in use for twenty years and manufacturers are knowledgeable
concerning factors limiting catalyst life in these uses. In some cases catalyst
has never been replaced whereas in others replacement is made every year.
S-1M21
-------
The user survey only produced 2k completed questionnaires on catalytic
afterburners. 30$ of the units were regarded as satisfactory. Efficiency data
was only reported in one case. A major maintenance problem was catalyst
poisoning, reported in 17 units.
Because of lower operating temperature, enclosures for catalytic
systems need not be so ruggedly constructed as for thermal systems. However,
refractory or stainless steels are recommended for interior surfaces and parts
exposed to preheat or afterburner temperatures. Configurations in current use
are relatively simple, usually in-line placement of preheat burner and catalyst
element. Design of the chamber and mounting of the catalyst element are based on
obtaining uniform flow of preheated stream through the catalyst and avoiding
bypassing of any part of this stream around the catalyst. Where possible,
distributed burners are used to aid mixing of waste with preheat burner products.
Catalytic afterburning differs from many other industrial applications
of heterogeneous catalysis in that afterburners are designed for unusually high
rates of reactions (per unit volume of catalyst bed). Under such conditions, the
overall rate of the catalyzed process becomes dependent on the rate of the mass
transfer of combustible from the waste gas stream to the catalyst surface as well
as on the rates of "chemical" processes taking place within the porous catalyst
itself. An analysis which allows treating mass transfer and chemical contributions
to determination of the rate, has been carried out for the catalytic afterburning
process. A method for calculating oxidation performance based on this analysis is
presented in Chapter 17. Correlations for mass transfer and heat transfer
coefficients and friction factors to be used in these performance calculations have
been developed for all of the catalyst matrix configurations in current use.
Publicly available data on the rates of the "chemical" oxidation processes with
these catalysts are scanty. A few estimated "chemical" rate parameters are
included in this report.
The NOX emissions from catalytic units will be lew, typically 15 ppra,
because of the low level of preheat.
S-1U121
-------
2.3 Heat Recovery
Since the discharge of flue gases at a temperature level of 1300 -
1500°F to the stack represents a continual wastage of expensive heat energy, the
afterburner user has a strong incentive to consider the many possible ways for
heat recovery. Recovery methods include heat exchange between hot flue gases
and incoming cool fume stream, recycling a fraction of the hot flue gases to the
oven or process from which the fumes have arisen, and the use of the heat in
other processing or heating loads, such as in generating steam for plant or
process heating, for power generation, etc. The cost of the heat recovery equip-
ment can probably be recovered in fuel savings in a few years,at least for fume
flow rates greater than 5000 SCFM.
Applications of heat recovery methods to afterburners have often had
problems.
l) Recovery must be feasible. Dirty streams may make exchangers
inoperable, and plant heat balance may preclude effective use in ways other than
exchangers.
2) Safety must be assured. Preheating the fume stream to a high tempera-
ture when it contains combustible components could initiate a fire.
3) The heat recovery equipment must be dependable, so that it does not
cause inadvertent outages in the plant.
U) Payouts must be evaluated in terms of reasonable capital and
operating costs, including realistic maintenance allowances.
5) There must be space available, or structural strength to carry it,
if the heat recovery equipment is to be mounted near the afterburner or on the
roof.
S-11*121
-------
6) Deterioration of the heat recovery equipment must not jeopardize
the pollution control function of the afterburner.
The most common type of heat exchange equipment is the cross flow tubular
heat exchanger; in Chapter 20 equations, data, and procedures have been included for
an approximate sizing of this type of exchanger. Rotary regenerative and switched
packed-bed type exchangers may also be used, although they have not been applied
widely.
Particular problems with heat exchange equipment relate first to the
fouling action of contaminants present in the fume stream, and to the loss of
performance and the fire hazards this may engender; secondly, to thermal expansion
caused by unequal heating throughout the exchanger structure.
2.U Auxiliary Equipment - Fume Generating Process
The afterburner system is centered around the afterburner itself and its
heat exchangers, but consideration must also be given to the fume source, to
process control systems, and to bringing fumes to the afterburner and discharging
clean flue gas> Large fluctuations in fume flow rate and concentration must be
limited if satisfactory performance is to be attained. Also, de-entrainment
devices should be used to eliminate heavy carryover of liquid and inert particulates,
Failure of control instruments causing major or nuisance shutdowns is
one of the most common complaints from the operators of fume incineration equip-
ment. In other cases improperly installed temperature sensing elements or
inadequate control systems have resulted in excessive fuel consumption, large
temperature variations, and operation at temperatures too low to give acceptable
pollutant destruction. Also, proper use of process controls will enable the user
to operate drying ovens and similar equipment so as to obtain higher solvent
levels in the fume stream, thereby reducing fuel costs. For these and other rea-
sons, it is important that an afterburner system be chosen which includes effec-
tive and reliable process controls. Chapter 9 is intended to guide the selection
of control and safety systems for afterburners. Descriptions of the control equip-
ment generally available and typical control systems are given. Included are
sections on combustion safeguards, temperature controls, and vapor concentration
controls.
Fume collection and delivery to the afterburner is a very important part
of the total system* Careful attention should be paid to its design since both
capital and operating costs are sensitive to the manner in which it is handled.
Long duct runs can cost more than the afterburner itself and condensation of
combustibles can occur even in insulated ducts, causing a fire hazard. Proper
hooding is crucial in order to insure that all fumes are collected and that air
dilution is kept to a minimum. Once the fume is below 25% of the LEL, all extra
air represents wasted fuel in the afterburner. Existing ventilation systems
will probably require modification to reduce unnecessary dilution and to vent
clean air directly to the atmosphere rather than through the afterburner.
Unless the fume generating process is operated at high pressure, the
afterburner system should include its own blower to overcome the pressure drop
S-
-------
10
(to provide mixing energy) through the unit. Blowers are normally placed upstream
of the afterburner so they handle cool fume at the lowest possible volume. For a
given pressure increase through the blower, the power consumption is proportional
to the actual volume handled. Induced draft blowers are placed downstream of the
afterburner so as to operate on clean gases and avoid fouling problems. Also the
afterburner will then be under slight vacuum and any leaks will be inward. Fans
are normally limited to operating temperatures below 700°F so heat exchangers
and/or cooling air must be used to reduce the temperature of the hot exha. 3t gases.
Induced draft blowers must be larger and consume more power. The volume at 600°F
is twice that at room temperature, and dilution air also increases the volume.
2.5 Costs
Purchase, installed and operating costs of thermal and catalytic after-
burners are covered in Chapter 11. The information is based on extensive dis-
cussions with manufacturers and users and also on replies by users to our
questionnaires. Data have been presented graphically where possible.
Inevitably there is a wide range in costs depending on the afterburner
type, the particular design and the nature of the installation. Thermal
units without heat exchangers range from $1.30 to $U.OO per SCFM uninstalled
depending on the type and on the size of the unit (larger units being cheaper
per SCFM). Corresponding units with heat recovery are of the order of 3-1/2
to 3-1/2 times these costs. Catalytic units are generally somewhat higher in
capital cost than thermal units with equivalent capacity. The average installation
cost of a unit is of the same order as the purchase cost, but is highly dependent
on the particular situation.
Operating costs for thermal units are largely a function of fuel require-
ments. Little operating manpower and maintenance is required. An average fuel
cost for a thermal unit would be ~$U.50 per SCFM per year without heat exchange,
and ~$2.50/SCFM/year with heat exchange. (Based on 2U-hour per day operation,
afterburner at lUOO°F, no heating value in waste stream, waste stream temperature
UOO°F, gas 50^/MSCF.)
Very few questionnaire returns were received from users of catalytic
afterburners, so that the information is limited on these units. Operating costs
are largely dependent on catalyst life, and fuel costs are not so important
because of the lower operating temperature. Where the waste stream is clean, and
a catalyst life in excess of two years can be counted on, this type of unit
should be economically attractive.
2.6 Measuring Afterburner Performance
In order to assess performance of an afterburner, concentrations of
critical components at the entrance and exit of the afterburner must be determined.
Typically these are total hydrocarbons and carbon monoxide but may include specific
hydrocarbons, halogens, particulates, etc. As a bare minimum the afterburner will
have to meet the standards imposed by the local air pollution control ag_ncy,
either emission standards for specific pollutants or a percentage destruction in
S-1U121
-------
11
passing through the afterburner. Other standards may prohibit any odor or
visible plume. The operator of the fume incinerator will often be responsible
for demonstrating compliance with regulations, either by doing his own testing
or by hiring an outside laboratory. If the afterburner does not provide the
pollutant destruction required, some investigation into interior performance of
the unit must be made. Of primary concern are measurement of temperatures and
flov patterns.
Certain measurements are also required before an afterburner is
purchased. It is important that fume stream flovrate and composition be deter-
mined, both average values and the maximum variation. The amount and nature of
entrained material will also influence design of the unit for reasons previously
discussed.
In Chapter 10 an attempt is made to indicate the state of the art
pertaining to the determination of a variety of air pollutants, measurement of
temperature and determining flow patterns. Problems encountered in sampling
and calibration are discussed. The methods available for determination of
individual compounds or classes are summarized, and available instruments are
listed.
S-1M21
-------
12
2.7 Major Sources of Combustible Pollutants and Potential for Afterburner Control
The ra^e at which organic pollutants are emitted in the United States
fro^ toth mobile and stationary sources has been estimated for 1968 and 1970, and
projected through 1980, 1990, and 2000. The results are summarized in Table 21.1.
The contribution of stationary sources is projected to rise irom 50$ of the total
in 1970 to 85$ in 2000. The projections of emissions from stationary sources were
made with the assumption that the present levels of legislated emissions control
would not change during the projection period. The projected mobile source
emissions reflect the stringent automobile emission regulations which are
scheduled to take effect in 1975 •
Several of the largest sources shown in Table 21.1 are either mobile
or are not amenable to afterburner control. Examples are motor vehicles, refuse
disposal, gasoline marketing, etc. Conversely, many sources which do not emit
nationally significant tonnages of combustible emissions are, non-the-less, prime
candidates for afterburner control because they constitute severe local nuisances.
Odor and smoke sources, e.g. rendering plants, coffee roasters, brake shoe
debonding ovens, etc., frequently fall into this category. No attempt was made
to estimate the total rates of emissions of effluents which are objectionable
primarily because of malodor. Instead, a count of the number of point odor
sources was used to give a measure of the severity of odor pollution problem.
The results are shown in Table 21.2.
Emissions from sources amenable to afterburning have been characterized
by the type of compound emitted. Each set of emissions has been assigned a
ranking based upon toxicity, malodor, and photoreactivity to allow an importance
comparison of the various pollutant sources. In addition, each source has been
assigned a ranking based upon the relative ease of control by afterburning. The
source characterization and rankings are given in Table 21.3«
Criteria to be considered when choosing between afterburners and other
control devices for particular applications have been outlined. The costs
associated with extensive afterburner application to control stationary sources
emitting nationally significant amounts of combustible pollutants excluding,
hovever, control of sources objectionable because of odor, is estimated to be
$360 million for equipment purchase and installation. The fuel cost associated
with this level of control (again excluding control of odorous sources) is
estimated to be $452 million for the purchase of 0-931 trillion scf of natural
gas annually. These estimates are based upon achieving the maximum feasible
level of control which would reduce the total emissions shown in Table 21.1 for
1970 by 11.1 billion Ibs per year, or about 15$. At this level of control,
approximately half of the total emissions from the stationary sources suitable
for afterburning would be incinerated.
It should be emphasized again that these costs make no provision for
controlling sources objectionable primarily because of malodor or carbon
monoxide, and they would be increased if these costs were included.
S-
-------
£.£ Proposed Research and Development Programs
Five-year programs at the $1 MM and $3 MM levels have been developed to
provide technology where deficiencies exist, to generate the data required to
predict and improve process performance and economics and extend the application
of afterburners to additional air pollution sources.
Since little quantitative performance data was found in this study,
field testing of thermal and catalytic units is recommended. Suitable analytical
methods, preferably of an instrumental type, need to be selected and utilized.
Basic studies of thermal and catalytic oxidation kinetics at temperatures
of interest are recommended to provide a basis for better design methods. Initial
studies would be with hydrocarbons and would continue into odors and particulates.
Final plans for the basic studies would consider heteroatoms (e.g., S, Cl, etc.).
The poisoning and deactivation rates of catalyst are also needed. Alternative
lower cost base metal catalyst studies are also recommended.
A substantial program is recommended in combustion design for thermal
systems. A research combustor is proposed with provision for changing geometry
readily and measuring internal temperatures, flow patterns and concentrations.
The interaction of mixing and combustion needs careful study to advance the design
techniques. Burner studies to find ways to get a large turndown and thereby stable
operation with varying feeds is also recommended. More reliable design procedures
can then be developed on this data base. Feeds studied in this research incinera-
tor would be hydrocarbons, odors, particulates and compounds containing
heteroatoms. Field testing of new designs are proposed to implement the research
findings.
S-11H21
-------
15
Chapter 3. THERMAL (DIRECT-FLAME) AFTERBURNERS; DESCRIPTION AND PERFORMANCE
Thermal afterburners are the most widely employed method to control the
release of hydrocarbon fumes, especially solvents from coating processes and
smokes. Pollutants are heated in the presence of oxygen to a temperature sufficient
to allow complete oxidation in the residence time available. Typically the thermal
afterburner consists of a refractory lined combustion chamber with a raw gas
burner fired at one end to supply heat. Cold, oxygen-containing fume enters through
the burner and acts as the air supply for the natural gas flame.
As is discussed in this Chapter, a thermal fume incinerator can be de-
signed and operated to achieve essentially complete conversion of hydrocarbon
pollutants to C02 and water. Its major drawbacks are the relatively high operating
expense caused by the need to burn supplementary fuel (usually natural gas) to
heat cold fume to the required high temperatures lUoo° - 1500°F and the significant
initial cost for a reliable unit. (Chapter 11 is a guide to estimating capital
and operating costs for a particular application.) Heat recovery (Chapter 6) can
be employed with an increase in first cost, but often with fuel savings large
enough to rapidly return this added capital.
Catalytic afterburners are another way to reduce fuel requirements
since in most cases pollutant oxidation will occur at a significantly lower
temperature in the presence of catalyst. Successful operation requires fume
streams which won't foul the catalyst and careful monitoring of the unit to
insure that the catalyst has not lost its activity. As a result, in the majority
of applications thermal fume incinerators are chosen for their promise of more
trouble free operation, and heat recovery is employed to obtain comparable fuel
savings. Sections U.I.2 and 11.3 should be consulted for a more thorough com-
parison of thermal and catalytic afterburners.
This Chapter is designed to guide the selection of a thermal afterburner
from among the many designs available and in use today. We first set forth the
general requirements for satisfactory performance of any thermal afterburner.
Then various design features are discussed as to how well (or poorly) they meet
these requirements and how well they are expected to stand up mechanically. The
material contained later in this handbook, especially Chapters 12-15,
should be consulted for a discussion of the bases for the guidelines developed
in this Chapter.
3.1 General Requirements for Satisfactory Performance
Successful fume incineration simply requires contacting pollutant
molecules with sufficient oxygen at a high enough temperature for the oxidation
reactions to go to completion in the time available at the high temperature.
Nearly every previous discussion of incinerator design requirements indicates
that the secret to success lies in the "3T's," time, temperature, and tur-
bulence. Essenhigh, however, has commented3'8': "The chant of Time, Temperature,
and Turbulence as being a useful basis for design is something of a mystical
incantation that at best is a truism and at worst is misleading, if not actually
wrong."
Preceding page blank
S-1U121
-------
16
The 3T's embody the correct notion of what is required for successful
incineration, but fail to indicate how this can be put into practice. Especially
troublesome is "turbulence."
Most fumes treated by afterburners are dilute, i.e., they do not
contain sufficient combustible to support combustion. Therefore, supplementary
fuel must be burned to generate heat, and the resulting hot combustion prc 'ucts
must be mixed with the contaminated fume stream in order to raise the entire
stream to a sufficiently high temperature. If the fume stream is not dilute
and can support combustion without supplemental fuel, it can be treated as a
fuel quality stream, either premixed or oxygen free, and burned in a normal waste
gas burner. The resulting combustion temperature, >2200°F is sufficient to destroy
any organic pollutant. In most cases the heating value can be recovered by
using the concentrated fume stream as fuel in a boiler or process heater.
In most applications for afterburners, the fume stream is essentially
contaminated air, and has adequate oxygen content (usually 15-21$ m) both for
burning the necessary preheating fuel and for oxidizing the contaminant. In a few
potential applications, the contaminated waste stream is an inert gas (e.g. NS)
with a low hydrocarbon content. In such cases, air must be blended with the waste
stream to supply oxygen. Some equipment manufacturers recommend that sufficient
additional air be added to assure an oxygen content in the afterburner effluent
of at least ^ m. This concentration allows oxidation of the contaminant to take
place, with a satisfactory level of conversion, under temperature and residence
time conditions within the range normally used in thermal afterburners.
3.1.1 Steps Involved in Dilute Fume Incineration
Figure 3-1 schematically indicates the sequence of steps which must
occur if fume incineration is to achieve satisfactory desbruction of pollutants
in a dilute stream. In an operating afterburner these steps are not likely to be
obvious since some of the equipment components are involved in several different
steps. All fume probably enters through the burner, which fires into a single
combustion chamber. As shown in the figure, at least part of a dilute fume stream
must bypass the supplemental fuel flame. If intimate "flame contact" is attempted
in a non-premixing system, serious flame quenching may occur. These points are
often misunderstood and lead to design and performance deficiencies. Sections
3-1.3 and 3«1-^ expand on these crucial areas.
Once the supplemental fuel has been burned to yield its heating value,
the cold fume and hot combustion products must be mixed to give a nearly uniform
temperature to all fume flowing through the afterburner. This should be done as
rapidly as possible without causing flame quenching so that sufficient residence
time can be provided at the required temperature in a reasonably sized unit.
As discussed in the Section 3.1.2 temperatures of 1^00° - 1500° are normally
adequate for complete destruction of pollutants if the stream is held for .1 - .3
seconds after the required temperature has been attained. This view of thermal
afterburner operation ascribes no special role to "flame contact" which 3-
fact is a beneficial concept as will be discussed in the Section 3«1«3«
S-1^121 Figure 3-1 follows
-------
O- i/t
NJ I
00 S
Supplemental
Fuel
Outside Air
(if Used)
Fuel Combustion
Dilute
Fume
Mixing of Fume
and Hot
Combustion Gases
Fume to Supply Oxygen
for Fuel Combustion
(Outside Air Needed if
Fume Fouls Burner or
< -16% Oxygen)
Retention of Fumes
at High Temperature
for Sufficient Time
Clean
Effluent
Figure 3-1. STEPS REQUIRED FOR SUCCESSFUL INCINERATION OF DILUTE FUMES
-------
17
For this reason, however, we refer to the broad class of non-catalytic
afterburners as "thermal afterburners," whether or not a flame is utilized as the
source of heat. The terms "thermal," "flame," and "direct-flame" afterburner are
used interchangeably.
The greatest variation among different afterburner designs is in how well
they achieve the goal of raising all of the fume to the required temperature for
the required time. Most cases of poor performance are due to non-uniform temper-
atures and flows which allow some of the pollutants to escape without adequate
treatment. This will be discussed in more detail as part of the consideration
of various afterburner designs currently available. The consequences of the
escape of some pollutants will depend on the situation. With extremely odorous
components a few parts per million or less will give a smelly exhaust, so 99«9#
destruction may be needed. Most hydrocarbon emission regulations have been
written to require >90# destruction of pollutants in the afterburner. As is
discussed in Sections 3-1-2.2 and 12.2.3 destruction rates are proportional to
the pollutant concentrations. This means that the outlet concentration must be
reduced to very low levels indeed if this inlet concentration is already low.
For these reasons and reasons of fuel economy, it is important that unnecessary
dilution be avoided. (Chapter 8 contains a discussion of proper fume-hood and
ducting design to minimize dilution in excess of that required for safe operation.)
3.1.2 Temperature and Residence Time Requirements
Temperature and residence time requirements are discussed together
since they are interchangeable to some degree - a higher temperature allows use
of a shorter residence time and vice versa. Additional residence time involves
a bigger combustion chamber and therefore a higher capital cost. However,
additional volume is relatively cheap (see Chapter 11) and the residence time
could be doubled for a 20 - 30% increase in capital. Operating temperature on
the other hand bears a direct relationship to fuel usage, which is the major
operating cost. Depending on specific considerations (fuel cost, hours/year
operation, available space, etc.) it would often appear that temperature should
be reduced and residence time increased.
However, this interchangeability of temperature and time is not of great
practical significance since oxidation rates are very strongly temperature
dependent. Figure J-2aschematically indicates the general effects of temper-
ature and residence time on oxidation rates in a flow through reactor. Over a
narrow temperature range the rate increases from essentially zero to rates
measured in milliseconds or less. At high temperatures complete conversion is
controlled more by concentrations of pollutant and oxidant than by the temperature
dependent rate. Since oxygen concentrations are almost always greater than
5$ whereas pollutant concentration is reduced to a few parts per million,
apparent rates are essentially independent of oxygen concentration and directly
proportional (first order) to pollutant concentration.
S-
-------
18
The "ignition temperature" is used to characterize this marked change
in extent of pollutant oxidation. However, as discussed in Section 12.1.1,
ignition temperature is highly dependent on the conditions under which it is
measured. Thus, autoignition temperatures (the temperature above which a com-
bustible mixture of the substance and air must be raised to initiate combustion
in the absence of a spark or flame) reported for many hydrocarbons have a large
range. Among the widest ranges and highest temperatures are for toluene,
970°F - lU90°F and methane 1000°F - 1300°F. Also, as explained in Section 12.1.1,
autoignition temperatures do not apply to hydrocarbon-air mixtures below the
lower combustible limit (Gee Section 12.1.2.) For safety reasons, most after-
burner applications involve dilute fume streams ranging from traces of combustibles
up to 25$ of the LEL. In order to achieve reasonable oxidation rates, temperatures
100°F or more above commonly reported ignition temperatures are usually required.
Table 3-1 shows typical ranges of residence times and operating
temperatures for each of the major pollution abatement categories for which thermal
afterburners are applicable. The range of conditions shown for each category is
relatively wide. To some extent this range results from differences in oxidation
rates of specific pollutants, due to their physical or chemical characteristics.
However, the major factor contributing to the width of the range of conditions is
the variability among afterburner designs in the effectiveness with which they
mix fume with combustion products, and avoid by-passing or short circuiting of
any significant part of the fume stream around the hot reaction zone. Those
designs which most successfully utilize the residence time for oxidation of all
of the fume at the average effluent temperature will permit attaining the
conversion (or abatement) goal at an operating temperature and residence time near
the lower end of the ranges quoted in Table 3-1. Those which are less successful
will require higher temperature and longer residence time for the same abatement
result.
3.1.2.1 Calculation of Residence Time
Total afterburner residence time is simply the total combustion chamber
volume divided by the volumetric flow rate of fume and supplemental fuel combustion
products. The volume of hot gases is directly related to the absolute temperature
of the gas so the volumetric flow rate must be based on the actual volume at the
combustion chamber temperature rather than the volume at standard conditions.
Operation at lUOO°F involves volume expansion -3.5 times the volume at 60°F.
Total average residence time, for cases where all oxygen for combustion is obtained
from the fume can be estimated from
S-1U121
-------
to
Increasing
Residence
Time
600
Figure 3-2a.
1000
1600
1800
2000
1200 1400
Increasing Temperature
COUPLED EFFECTS OF TEMPERATURE AND TIME ON RATE OF POLLUTANT OXIDATION
-------
100
90
c 80
ID
C
JO
o
b! 70
o
u
tt>
'u
Hydrocarbons
Only
Hydrocarbon +CO,
per LAAPCD, Rule 66
60
50
1200
1300 1400
Temperature, °F
1500
Figure 3-2b. TYPICAL EFFECT OF OPERATING TEMPERATURE ON
EFFECTIVENESS OF THERMAL AFTERBURNER FOR DESTRUCTION
OF HYDROCARBONS AND CO
S-14121
67784
-------
19
Table 3-1. THERMAL-AFTERBURNERS
Conditions Required for Satisfactory Performance
in Various Abatement Applications
Abatement Category
Hydrocarbon Emissions
(90# + Destruction of HC)
Hydrocarbons + CO
(90% + Destruction of HC + CO,
as in LAAPCD Rule 66)
Odor
(50-90# Destruction)
(90-99% Destruction)
(99% + Destruction)
Smokes and Plumes
White Smoke (Liquid Mist)
(Plume Abatement)
(90$ + Destruction of HC + CO)
Black Smoke (Soot and Combustible
Particulates )
a; Temperatures of 11*00- 1500BF may
Afterburner
Residence Time
(Sec)
0.3-0.5
0.3-0.5
0.3-0.5
0,3-0.5
0.3-P-5
0.3-0.5
0.3-0.5
O.J^l.O
be required if th
Temperature
(°F)
3.100-1250*'
1250-1500
1000-1200
1100-1300
1200-1500
8QO-1000b'
1250-1500
}.lfQO-2pOp
e'ljyd.rocprbon,
has a significant content of any of the following; methane,
cellosolve, substituted aromatics (e.g* toluene, xylenes).
b) Operation for plume abatement only is not recommended* since
this merely converts a visible hydrocarbon emission ^o an
invisible one, and frequently creates a new odor problem due
to partial oxidation in the afterburner,
S-
-------
20
v«6o ,_ .
T ~ (SCFM)3-5 ^ '
where T = total average residence time, seconds
SCFM = fume flovrate to afterburner at 60°F and atmospheric pressure
(standard conditions), ft3/minute
V = combustion chamber volume (following burner and prior to het-t
exchangers, dilution air, etc.)> ft
For more precise calculations the factor 3-5 should be replaced by
(T(°F) +U60)/520 and (SCFM) should include supplemental fuel and
combustion chamber
any outside air added through the burner. When outside air is not used, methane
will contribute less than 3% to the volume. However, if the required supplemental
fuel is burned using only outside air the volume may be increased by 50% or more.
Our survey showed that afterburners are operating with total average
residence times which vary from 0.1 to 5 seconds. It should be stressed that this
is the total residence time and is not equivalent to the time during which fume is
held at the required temperature. A significant amount of time (often greater than
the total residence time) is required to bring the cold fume up to temperature.
Also, all elements of fume do not spend an equal amount of time in the combustion
chamber. Some are swept out very quickly while others are trapped for a much longer
time. This variation in residence times is a function of flow patterns attained
and can greatly affect performance of the unit.
3.1.^.2 Hydrocarbon Destruction
Destruction of most hydrocarbons occurs very rapidly at temperatures in
excess of 1100° - 1200°F. Possible exceptions are methane, cellosolve, and benzene
derivatives, like toluene, which are stable molecules and require a higher
temperature (-lUOQ°F) for oxidation to occur in a few tenths of a second.
More information on expected rates of hydrocarbon destruction is contained
in Section 12.3.1. Except for a few compounds, required conditions for destruction
are not sensitive to the hydrocarbons involved. (Methane as noted is difficult
to destroy at temperatures below lUOO°F. However, in most cases the presence
of methane in the fume indicates inefficient combustion in the burners in
the drying ovens, i.e., flame quenching. Adjustment of these burners will
normally markedly reduce inlet methane concentration to the afterburner.)
If hydrocarbons are found in the flue gas of an afterburner operating at
a nominal combustion chamber temperature above lUOO°F (or above 1200°F for all
but a few hydrocarbons), it is due to poor mixing and non-uniform treatment of
the fume stream rather than too low a temperature. (Excluding the possibility
that the indicated temperature is higher than the actual temperature, either
because of the need for instrument recalibration or because the thermocouple
"sees" the high temperature flame.) A higher nominal temperature will usually
reduce outlet hydrocarbons in this case but the benefit comes from raising the
temperature of that portion of the fume which sees a much lower temperature or
shorter residence time than the average, rather than through increased efficiency
S-1M21
-------
21
due to the increase in the average temperature. In effect, the extra fuel is
wasted.
At these same temperatures most hydrocarbons should disappear in
~.l second or less. Since most afterburners have total average residence times
many times this long, slow mixing and bypassing must be blamed for any failure
to achieve complete destruction. Increasing the residence time should prove
beneficial since more time is available for mixing of hot combustion gases and
cold fume. However, if mixing is very poor to start with, a doubling of residence
time may produce little noticeable effect. Also, if residence time is increased
by decreasing throughput, performance is likely to be poorer since the gases will
be flowing at a lower velocity and will mix more slowly.
Hydrocarbon destruction rates will increase with initial fume concentration
and better cleanup will be possible with a higher concentration given the same
mixing efficiency and afterburner temperature. This is partly due to the first
order dependence of oxidation rate on pollutant concentration and partly due to
the significant local heat of reaction which is released during oxidation. But
primarily the ability to achieve better cleanup is due to increased concentrations
of free radicals which actually carry out the oxidation. (See Section 12.2.3
for additional discussion.) This improved performance may not be observed, if the
higher heating value of the more concentrated fume, results in much poorer mixing
because there is less mixing energy in the smaller supplemental fuel flames.
However, this falloff of rate with concentration should be kept in mind when
dealing with very odorous substances or the need to achieve a high destruction
efficiency with a very dilute incoming fume.
3.1.2.3 Carbon Monoxide Cleanup
In the previous section it was mentioned that operating temperatures
of lUOO°F and above are rarely required to achieve hydrocarbon destruction
except in the case of poor mixing of cold fume and hot combustion products and
a few stable hydrocarbons. Such high temperatures do seem to be required, however,
to obtain sufficient oxidation of CO •> C02 in <-5 seconds. As is discussed in
Sections 12.2 and 12.3, the time required for the oxidation of CO is >10X that
required for all of the steps involved in the oxidation of hydrocarbons to CO.
Carbon monoxide cleanup is required since it too is a pollutant which is detri-
mental to health and is thought by some to lead to photochemical smog. Los Angeles
Rule 66 includes the net production of CO in calculating afterburner efficiency.
Afterburner Efficiency =
{Hydrocarbons}IN-{Hydrocarbons}OUT+{CO}IN-{CO}OUT
{Hydrocarbons}
where { } indicates concentration expressed as Ci.
Thus the afterburner may achieve 98$ destruction of hydrocarbons but fail to achieve
the 90% minimum efficiency because of high CO production.
Section 12.3.2 contains some rate expressions for CO oxidation flames and
also some estimates of the oxidation rate at concentrations and temperatures of
S-1U121
-------
rost in i'urr.e incineration. Afterburner experience shows that temperatures of
° - 1^50°? &re required with an actual residence time at this temperature
l'ter -nixing is nearly complete) of .2 - A seconds to achieve nearly complete
cxiCdtion of ~CO - C02 (less than ~200 PPM of CO in the flue gas) see Figure 3-2b.
Units with poor mixing patterns exhibit outlet CO concentrations >1000 PPM though
temperatures are >1^00°F. A few units have provided sufficient CO cleanup at
•uo-i ot-atures as low as 1200°F. However, this has normally been associated with a
i'ifi1 .--olver.t concentration, and as was mentioned previously, the higher th-
conuer.tration, the easier pollutant destruction will be. Effective oxidation
i'ate_; are nigher for the reasons discussed in Section 12.2.3-
3-1.2.U Odors, Aldehydes and Other Oxygenated Hydrocarbons
Afterburners are often used to destroy odorous fumes containing only
100 PPM or less of the odor producing substances. These are likely to be aldehydes,
organic acids or sulfur containing compounds. Oxygenated hydrocarbons are inter-
mediates in the oxidation of hydrocarbons, and as a result they are easier to
destroy than the original hydrocarbon. Often afterburner temperatures of 1000°F
or less are adequate for odor removal applications. Little if any CO will be
oxidized at these temperatures in the time allowed, but so little is produced
from >100 PPM of odorous material that it can be released without concern. How-
ever, care must be taken that the fuel necessary to heat the fume stream is burned
in such a way that flame quenching does not occur. (See Section 3.1.3.) If it
does, high concentrations of CO and methane from the fuel burner will be found
in the stacx gases.
It is especially important that good mixing patterns be established when
dealing with extremely odorous compounds. Temperature uniformity is required
across the combustion chamber and bypass must be avoided if the odor is to be
completely destroyed. Regulations dealing with solvent emission control can be
met even if 5 - 10$ of the original hydrocarbon escapes treatment in the after-
burner. However, odors may still be a problem if less than 1% of the original
concentration remains in the flue gas.
3.1.2.5 Liquid Smokes and Droplets
Outside of California smoke and odor control is the most common appli-
cation of afterburners and most of these are operating at 1000°F or below. As
mentioned in the previous section, CO oxidation will not occur at an appreciable
rate in this temperature range. Condensed hydrocarbon smokes are given off by
many processes such as rubber curing, meat smoking, asphalt blowing, deep fat
frying, etc. Much of the smoke given off from refuse incineration and improper
oil burning is also liquid rather than dense carbon smoke.
Since organic liquids which are likely to be found in fume streams have
boiling points below 600 - 700°F, the smoke plume can be made to disappear by
merely heating the effluent gases to 800°F or less. About all that such low
temperature afterburners do is to vaporize the visible smoke and render it in-
visible. That's all they would do if supplemental fuel to supply heat were
burned externally in an efficient manner. However, a common arrangement is to
-------
23
place a burner at the base of the stack and fire directly into the smoke containing
effluent gases. Gross flame quenching occurs and much CO, aldehydes, and unburned
fuel is found in the "smoke free" stack gases. The aldehydes create a bad odor
problem. In many cases there is less total pollution when the afterburner is
not operating.
Liquid smokes require the same ~lUOO°F afterburner temperatures that
gaseous fumes require. In Section 12.3-3 it is shown that the additional time
required to burn a 100 micron hydrocarbon droplet is -.02 seconds at lUOO°F.
Since drops bigger than 50 - 100 microns are easily removed in simple cyclones
and knockout vessels, no special problems are introduced relative to rates of
pollutant destruction. Certain practical problems such as fouling of burners
and buildup in duct work are discussed elsewhere.
3.1.2.6 Soot and Combustible Particulates
As contrasted to liquid smokes, solid carbon (soot) smokes and other
combustible particulates are expected to increase the required afterburner
temperature and/or residence time. (See Section 12.3-^.) Particles larger than
50 - 100 microns should be removed ahead of the afterburner using simple cyclones
or settling chambers. If particles are larger than 5 microns (5 x 10"1* centimeters),
residence times of a full .5 seconds may be needed at lUOO° - 1500°F to achieve
complete burnup. Since mixing and particle heating require additional time, total
average residence times of .7 - 1.0 seconds or higher temperature capabilities
should be provided for large particulates. One method of providing longer
residence time without increasing the combustion chamber size, is to use
tangentially fired burners in a vertical upflow afterburner. (See Section 3.3.2.3.)
The resulting cyclonic action and gravity will hold large particles near the
bottom of the afterburner chamber until they burn up.
Small soot or other dense carbon particles present another problem.
Soot typically has particle diameters <.l micron but may agglomerate to several
times this size. Soot will not be affected by any cyclonic forces and appears
to exhibit surface burning rates two orders of magnitude less than these for
pulverized coal or coal char. Extreme temperature and residence time requirements
are calculated (Section 12.3.U) by extrapolating soot combustion data into the
temperature regions of interest. Temperatures of up to 2000°F and residence times
of 1 second or more may be required for dense carbon smokes.
Afterburner experience with wire burnoff, smelting operations, and cloth
carbonization shows that temperatures of over 1700°F and residence times greater
than .5 seconds are often needed to clean up the black smoke plume. The required
severity of treatment seems to depend on the amount of volatile matter present and
porosity of the particles. As mentioned in the previous section, most smoke from
refuse incinerators is a liquid smoke and is easily burned up at temperatures of
lUOO° - 1500°F. However, smoke from plastics incineration sometimes requires
temperatures greater than 1700°F for destruction in a secondary combustion chamber.
Since limited information is available, some testing and/or operating flexibility
(e.g., high temperature capability) is recommended where no direct experience with
similar dense carbon smokes is available.
S-1U121
-------
2k
3.1.2.7 liitrogen Oxides (N0y) Formation
In afterburners, as in any combustion equipment where fuel is burned
with air in a flame, some reaction of nitrogen and oxygen from the air occurs,
forming some nitric oxide (NO) and nitrogen dioxide (NOg) • (For measurement and
discussion, these two oxides of nitrogen are referred to cumulatively as NOX«)
Substantially all of the NOX formed in a thermal afterburner is formed in the
high temperature region £>28000F) of the burner flame itself. At the tempeiature
of the main residence-time section of the afterburner (1200-1500°F), the overall
rate of reaction of nitrogen with oxygen is too slow for significant formation
of NO • However, in occasional afterburner applications where nitrogen-containing
compounds, e.g. NH3, amines, pyridines, are present in the fume contaminants,
additional NO may be formed during fume oxidation. This formation of NOX from
oxidation of nitrogen-containing compounds is discussed below in section 5-l«^«8.
The normal size of afterburners and the types of burners used to
facilitate mixing of fume with combustion products, result in relatively small
flame size, with rapid quenching of flame products from >3000°F to ~1500°F.
Thus, nitrogen and oxygen passing through the flame region have a very short
residence time in the high temperature region where temperatures are high enough
for NOX formation. In comparison with larger combustion equipment (e.g. boilers
and other industrial furnaces), afterburners have relatively low NOX "emission
factors", (i.e. the amount of NOX formed per unit of fuel burned).3"9) For
gas-fired afterburners, the "emission factor" has been found to be typically
0.05-0.1 Ib NO /MM Btu heating value of the fuel burned.3-9' (in emission
factors, the "Ib NOX" is customarily expressed in terms of NOa — as though all
the NOX were present as NQZ with a molecular weight of 46.) No test results are
available from oil-fired afterburners. However, from results of tests with other
combustion equipment, oil-fired afterburners would be expected to have NOX
emission factors higher by a factor,of 2-3 than gas-fired units.3"9)
Tne use of preheated combustion air (which reduces fuel consumption)
leads to higher flame temperatures and more rapid NOX formation. Similarly,
preheated or heat-exchanged fume may lead to slower flame quenching and
correspondingly higher NOX formation per unit of fuel actually burned. Reference
3-9) recommends the use of the emission factor together with the effective
heating value — the heating value of the fuel actually burned plus the heating
value (above ambient temperature) of the preheated stream -- in estimating NOX
enissions from furnaces with air preheaters. The extension of this concept to
afterburners with ntat exchanger will probably result in a slight overestimation
of NOX emissions. However, it appears better to err in this direction than to
underestimate emissions by neglecting any effect of preheating on NOX formation.
Using the assumption of effective heating value and emission factors of 0.08
for gas-fired units and O.l8 for oil-fired units, the emission rate and
concentration in afterburner effluents given in Table 3-2, have been estimated.
S-HH21
-------
25
Table 3-2. ESTIMATED N0y EMISSIONS FOR THERMAL AFTERBURNERS
Gas fired afterburners
Fume used as
air supply for burner
External combustion
Air for burner
Oil fired afterburners
External combustion
Air for burner
Operating
Temperature
(°F)
1200
1500
1200
1500
1200
1500
NOX Emission Rate
/ Ib N0v/Hr\
VMSCFM Fume/
0.13
0.16
0.18
0.22
0.40
0.50
NOX Concn
in Effluent
(ppm)
18
22
18
22
4o
50
The afterburner design concept involves dilution of primary flame
products by mixing with cold fume stream to achieve the afterburner operating
temperature. Thus, the NOX formed in the flame is emitted from the stack at
only 0.2-0.3 times the concentration at which it was formed in the flame zone.
In contrast, in efficient larger scale combustion equipment, operated with
minimuip excess air, combustion gases are cooled to stack temperature by heat
exchange (with boiler tubes or other heat transfer surfaces) rather than by
dilution and NOX is emitted at the concentration at which it formed in the flame.
Emissions from afterburners, already lower than from other combustion equipment,
on an emission factor basis, are also at even lower concentrations because of
the dilution effect. For gas fired afterburners, estimated concentrations of
NOX in the stack effluent of 18-22 ppm are shown in Table 3-2.
Experimental studies of.NOx emissions from afterburners were included
in a I960 study by the LACAPCD3"8^ of NOX emissions from stationary sources.
The report of that study characterized measurements on 24 afterburners by
emission factors of 0.05-0.08 Ib NOX/MM Btu. One burner manufacturer reported
iieasurements of NOX concentrations of < 20 ppm when gas-fired distributed
burners are operated for 1500°F or lower afterburning. Test data supplied by
two afterburner users on three well-designed thermal afterburners operating on
paint baking oven effluents indicated stack concentrations of NOX of 20-35 ppm
when these units were operated at temperatures of l400-1500°F. Data3"10^
obtained by Scott Research (Plumsteadville, Pa.) on two thermal afterburners in
tests made for the National Coil Coaters Association are summarized below.
_ Afterburner Unit
Inlet volumetric flow (SCFM)
Outlet gas temp (°F)
Increase in NO (ppm)
Increase in NOa (ppm)
Emission factor (lb/MM Btu)
(Tr.ese data indicate a distribution between the two oxides of nitrogen of 70-90$
nitric oxide, 10-30$ nitrogen dioxide.) All of these available experimental
measurements are reasonably consistent both with each other, and with the estimation
basis used to construct Table 3-2.
A
5900
1290
25
2.2
0.12
B
3000
1070
8.4
3-0
0.06
S-Ik 121
-------
26
3.1.2.8 Wastes Containing Chlorine. Sulfur, Phosphorous, Nitrogen, Metals,
or Other Heteroatoms
Afterburners are used for the control of combustible air pollutants
but components other than carbon, hydrogen, and oxygen may be present. These
can either be chemically bound with the combustible, as a chlorinated solvent,
or merely be an inorganic dust carried along in the fume stream. In the latter
case, one would not expect any change to occur at typical afterburner temperatures,
and the dusts will appear in the flue gas. Chemically bound heteroatoms will form
the corresponding oxide or acid, e.g., chlorine will form hydrochloric acid
(or free chlorine), sulfur forms sulfur dioxide, metals form their oxides commonly
called ash. Section 12.5.2 discusses the likely fate of most heteroactoms. If
concentrations of these products are high enough, e.g., >200 PPM S02 or >50 PPM HC1,
>.l gr/SCF ash, the afterburner alone will not provide sufficient pollution
control. Wet scrubbers, cyclones, filters, etc. must be considered for flue gas
treatment.
The presence of various heteroatoms does not rule out the use of thermal
afterburners as it would catalytic units because of catalyst poisoning
(Section U.2.U). In fact, many fume incinerators are used to burn poisonous
HaS to S02 and N02 to a mixture of NO and N2. However, the heteroatoms,
particularly the halogens, do inhibit the oxidation reactions (Section 12.3.5),
and even at low concentration they may cause higher temperatures and longer
times to be needed for complete destruction of pollutants. No firm guidelines
can be provided so testing is recommended prior to design of an afterburner
to handle such wastes. Corrosion problems are especially severe when acidic
components are present in the flue gas. This must be kept in mind when choosing
construction materials for the afterburner and associated duct work.
Nitrogen containing fume components (e.g. NHa, amines, pyridines, etc.)
are of particular interest, since their oxidation can contribute to the total NOX
emissions from an afterburner. No direct information is available on the extent
to which organic nitrogen in the fume is oxidized to NOX. However, chemically
bound nitrogen in fuels oxidized in flames is converted to NOX in yields ranging
from 20-100$ of stoichiometric.3'11' The remainder, not converted to NOX, is
almost entirely converted to Ng. The yield of NOX increases with decreasing
concentration in the fuel — trace quantities are converted quantitatively, while
major concentrations show lower yields. The concentrations of contaminants in
fume streams approach most closely the situation corresponding to low or trace
concentration of nitrogen compounds in fuels. It should be pointed out that the
comparison between flame burning and fume oxidation is not ideal, since the two
processes.do take place at quite different temperatures. However, from what is
known3-12' of the mechanism of formation of NOX in flames from reduced nitrogen
compounds in the fuel, it is expected that the lower temperature of fume after-
burning would increase selectivity to NOX (over Ng)- Thus, one would expect
in most situations to see nearly quantitative yields of NOX from nitrogen compounds
present in the fume stream.
S-11*121
-------
27
3-1-3 The Role of Flame Contact - Dilute Fumes Must ByPass
Many afterburner designers feel that contact between the supplemental
fuel flame and the fume is needed for satisfactory pollutant destruction. Since
this may lead to improved mixing between cold fume and hot combustion products
and excess free radicals from the flame will enhance oxidation rates, the con-
cept of "flame contact" is a good one. However, all of the fume cannot be put
through the flame since a non-combustible mixture would result when fuel rate is
set for a 1400° - 1500°F net temperature. Therefore, too much zeal in pursuing
the goal of flame contact will lead instead of flame quenching and possibly an
increase in pollutants.
3.1.3.1 Fuel Requirements and Stream ByPassing
As is discussed in Section 12.1.2, combustible mixtures are those
which release enough heat through the oxidation reactions to raise the neighboring
unturned fuel and air above their "ignition temperature" and overcome heat losses.
The lowest concentration at which this will occur is the lower combustion
(explosive) limit, the LEL. This lowest combustible mixture yields a flame
temperature ~2300° - 2400°F for most hydrocarbon substances burned in air.
(Table 12-l) (Methane burns as low as ~2100°F.) If the mixture inlet temperature
is raised, less combustible is required since the minimum flame temperature
capable of sustaining combustion is more or less fixed.
Afterburner fuel requirements are chosen to provide the minimum required
temperature for pollutant destruction in the time available at this temperature.
As discussed in Section 3.1.^ temperatures above 1^00° - 1500°F are normally not
required. If this fuel is premixed with the cold fume a noncombustible mixture
will result. The LEL for methane in air at room temperature is 5$v compared to
the less than 3% methane required to heat a cold, very dilute fume to lUOO°F.
Heat balance calculations are used to determine the heat required to raise fume
and combustion products to a specified temperature and the fuel which must be
burned to release this heat. Heat released during oxidation of the pollutant
is often not negligible and must be included. (A 25% LEL fume stream will give
a ~500° - 700° temperature increase during oxidation depending on the specific
pollutant and temperature level.) Chapter 11 contains further discussion and
useful graphs relating to the calculation of afterburner fuel requirements.
As fume temperature and/or fume concentration increases, a heat balance
calculation predicts that no fuel is needed to achieve a 1^00°F combustion chamber
temperature. This occurs at ~50# LEL and no preheat or ~25# LEL and an 800°F
preheat temperature. A certain amount of supplemental fuel is always required,
however, to insure that incoming fume will be raised above its "ignition
temperature" so that its heat content will be released. For most hydrocarbons
900° - 1000°F temperature rise is sufficient (Section 12.1.1) and fuel burners
should be designed with enough capacity to raise cold fume by a minimum of 10CO°F.
The burner can then be "turned down" to a point where stable operation is insured.
Streams more concentrated than ~50# LEL (fume in the range from 50$ LEL to the
upper combustible limit should not be encountered for safety reasons) should not
be preheated and will require combustion chamber temperatures above l400°F.
Many afterburner failures have been caused by excessive combustion chamber
temperatures resulting from a high combustibles content in the fume stream.
This is often caused by entrainment of liquid droplets which were present c
-------
23
Wherever possible, oxygen in the fume should be used to burn the
supplemental fuel so part of the fume passes through the flame and part bypasses
and is mixed with the combustion products downstream of the flame. (Use of
outside air will increase fuel requirements by 30 - 50% and requires all of the
fume to be mixed in downstream of the flame.) For a lUOO°F combustion chamber
temperature hQ% of a cold (80°F) fume stream can be mixed with methane to give
a 5-9$v mixture. (5$v is the LEL-) Thus, approximately half of the fume must
bypass the flame and be mixed with the combustion products. If fume enter" at
UOO°F instead (or equivalently has a 320°F heating value), less methane must be
burned to provide a 1^00°F chamber temperature and more fume must bypass (-59$)
to maintain the same flame temperature (2500°F) as before. Thus good mixing
downstream of the burner becomes more crucial for higher preheat temperatures and
higher inlet concentrations.
3.1.3.2 Benefits of Flame Contact
As is discussed in Section 12.2.1, oxidation reactions in flames pro-
ceed very rapidly and primarily involve reactions of OH, 0, and H free radicals
with the fuel molecules and intermediate components. Oxidation rates are pro-
portional to the concentrations of these radicals and radical concentrations
are several orders of magnitude higher in a flame than they are in the absence
of flame at lUOO°F. As we have Just noted, it is not possible to sustain a
hydrocarbon/air flame at lUOO°F, but the high radical concentrations take
-.01 - .03 seconds to decay after fuel has been consumed in the >2200°F flame.
Therefore, pollutants in the fume will be destroyed more rapidly if hot com-
bustion products and fume are rapidly mixed; i.e., "flame contact." Care is
required, however, since CO oxidation to C02 in the flame occurs on the same
time scale as the decay of excess free radicals so rapid mixing will lead to
quenching. Also, as is discussed in Sections 3.1.2 and 12.3, required oxidation
times for hydrocarbons at lUOO°F in the absence of flame are much shorter than
the -.5 second typical afterburner residence time.
A more beneficial result of attempts to provide flame contact has
been the development of distributed burners (Section 3.2). In these burners
the supplemental fuel flame is spread out across the fume inlet to the combustion
chamber and mixing of bypassed fume and hot combustion products is more easily
accomplished because some subdivision has already occurred.
S-11*121
-------
3.1-^ Quenching Must Be Avoided - Presence of Aldehydes and CO in the Flue Gas
Temperatures in the range of lUOO° - 1500°F are required for reasonable
rates of complete conversion of hydrocarbons to C02 and H20. In Section 12.2
we discuss the fact that oxidation reactions involve free radical chain carriers
and that aldehydes, organic acids, and carbon monoxide are formed as intermediates.
These intermediates and the original hydrocarbon will be found in the afterburner
flue gases if quenching occurs and the oxidation reactions are essentially
terminated.
Quenching is most likely to occur in local regions in the combustion
chamber due to cold spots or too rapid mixing of fume with the flame. A few
possible quenching situations are:
1. Fume which has been heated to a sufficiently high temperature may be
cooled due to contact with a cold wall or mixing with cold fume.
2. If cold fume is rapidly mixed with the supplemental fuel flame (to achieve
flame contact) some of the fuel and CO from the flame may be rapidly cooled causing
reaction to stop.
3. Some fume or quenched gases from the flame may pass through the afterburner
without reaching the nominal combustion chamber temperature. (Poor mixing or
bypassing can cause large variations in temperature and residence time.) If only
moderate temperatures are attained (-1000°F) aldehydes will be prevalent and will
remain when the flue gases are quenched after leaving the combustion chamber.
This is also a problem with plume eliminators designed to operate below 1000°F.
U. High surface area within the combustion chamber may serve as a third
body for radical recombination reactions, resulting in partial quenching and a
reduction in chain carrying free radical concentration and oxidation rate.
(See Section 12.2)
Severe quenching by any of these mechanisms can lead to a worse
pollution problem with a poorly designed afterburner in operation than when
it is shut off.
S-114-121
-------
30
3.2 Thermal Afterburner Design Features - Systems Utilizing
Distributed Burners
It was noted in Section 3-1 that large fuel savings (up to 50% or more)
are associated with using oxygen in the fume for combustion of supplemental fuel
rather than bringing in outside air which must also be heated to ll+00°F. It was
also pointed out that ~l/2 of the fume stream must bypass the supplemental fuel
flame and that achieving good mixing between this bypassed fume and the hoi com-
bustion products is the most important step in successful fume incineration. For
these reasons, the distributed raw gas burner is well suited for use in thermal
afterburners. These burners are placed directly in the fume stream in order to
utilize oxygen in the fume, and required mixing distances can be made short (less
time needed for bringing all fume up to the required temperature) since the flame
is distributed across the fume inlet to the combustion chamber. Figure 3-3 is a
sketch of a system using distributed burners. Drawbacks include: fouling in
streams containing tars or particulates; tendency for flame quenching (high CO
in combustion products); and inability to burn fuel oil instead of gas in currently
available burners.
As is discussed in Chapter lU, subdivision of the cold fume stream and
hot combustion products from the supplemental fuel flame is one of the best ways
to achieve good mixing in a short distance. The distributed burner divides the
flame into many individual jets or lines of flame surrounded by cold fume.
Average spacing is 1 foot or less. Thus the scale across which mixing of flame
and fume must occur is on the order of inches rather than the full combustion
chamber diameter (2-10 feet) and mixing time will be reduced. The pollutants
have more time at the operating temperature and there is less chance for bypassing.
3.2.1 Description of Available Distributed Burners
There are three basic types of distributed burners utilized in fume
incinerators and most are based on burners developed for air heating applications.
These are line burners, multijet burners, and grid burners. As will be seen,
these are rather arbitrary distinctions and many variations exist. All of these
burners are intended as raw natural gas (or propane) burners with air taken from
the fume stream, but they can be modified as discussed in Section 3.2.2.2 for use
with low oxygen content fumes.
3-2.1.1 Line Burners
Line burners are the most commonly utilized distributed burner. The
Maxon Combustifume (Figure 3-1*) is widely employed in various manufacturers'
afterburners. The Eclipse TAH-0 (Figure 3-5a) and AH-0 (Figure 3-5b) and other
specific manufacturers' variations of these burners are also found in fume
incinerators in use and on the market today. These are all intended as raw
natural gas (or propane) burners with air taken from the fume stream but
can be modified for low oxygen content fume as noted below. Gas enters through
holes in manifold pipes placed across the duct with spacing on the order
of one foot. Air (fume) for combustion enters through holes in mixing plates
attached to the manifold pipe and forming a V shaped trough. Lines of flame
result.
S-11H21 Figures 3-3 thru 3-Ta, b follow
-------
To
Stack
Fume
Natural
Gas
Figure 3-3. AFTERBURNER WITH DISTRIBUTED BURNER
BURNER
PROFILE
OPENING
1
&J
^*&*&t^^ -—**—-<-
'M
I
PROFILE
PLATE
INCINERATOR
HOUSING
COMBUSTIFUME® Burner
Figure 3-4. MAXON COMBUSTIFUME BURNER
S-14121
67784
COURTESY; Moxon Premix Burner Co.
-------
Figure 3-5A. ECLIPSE TAH-O DISTRIBUTED BURNER
Front View - Mixing Plates
Figure 3-5B. ECLIPSE AH-O DISTRIBUTED BURNER
Back View - Inlet for Fume Used as Combustion Air
COURTESY: Eclipse Fuel Engineering Co.
S-14121
67784
-------
Profile
7 Profile Plate
Figure 3-6a. CQMBUST1FUME TYPE DISTRIBUTED BURNER WITH PROFILE PLATES
Figure 3-6B. SHORT FLAME FROM COMBUSTIFUME BURNER
S-14121
67784
COURTESY: Maxon Premix Burner Co.
-------
Natural
Gas
Fume
A. Hirt Multijet Gas Burner
^^^^^^^^^^AAAAA^A^i
Natural
Gas
S-14121
67784
1
Adjustable
Gap
\
s
Exhaust
v\\\x\\\\\\\\\\\\\\\\\\\\\x\x\\\x\\\^
ft/me
B. Afterburner System Employing Multijet Burner
Figure 3-7. HIRT MULTIJET GAS BURNER
-------
31
Figure 3-6a shows the Maxon Combustifume type arrangement. Profile
plates are used to block most of the area between adjacent burners and between
outside burners and the combustion chamber wall. The gap between profile plate
and burner determines the pressure drop across the burner and is set to force
part of the fume (-505? for reasons discussed in Section 3.1.2.1) through the mixing
plate holes to provide oxygen for the flame. The remaining fume bypasses the
flame and is mixed with the hot combustion products downstream of the profile plate.
The pressure drop across the burner is -1.5" w.c. which results in fairly intense
mixing of natural gas and fume between the mixing plates. This gives a short,
stable flame extending only a short distance beyond the V trough, as shown in
Figure 3-6b.
The Eclipse type of distributed line burner is similar except that much
longer flames are produced and higher heat input per foot of burner is possible.
Flames extend up to 5' beyond the burner for the -1.5 million Btu/hr-ft maximum
firing rate. This maximum is -2X that for the Combustifume type burner and means
that burner costs are less for a given heat input requirement (fume flow rate).
However, burner spacing will be greater since less burner length is needed,
so mixing of fume and flame must occur across somewhat greater distances.
Also, the long flame length means that several feet of combustion chamber
are lost before all of the heat from the supplemental fuel has been released
and can be used to raise the fume to the required incineration temperature
(-lUOO°F). The slow mixing of fume and fuel (which gives the long flame)
offers the advantage of somewhat greater turndown (ratio of maximum to minimum
firing rate) and spacing between burner and profile plate is less critical.
With rapid mixing it is easy to fall outside the combustible limits and get
serious flame quenching (Section 3.1.3).
3.2.1.2 Multiset Burners
Multijet burners employ many discrete flames of the type discussed in
Section 3-3. However, they are viewed as part of a single burner with one control
system and a common manifold for feeding fuel gas. The Hirt multijet burner
schematically shown in Figure 3-Ta has been widely used in afterburner applications.
It is normally installed as indicated in Figure 3-7b. Part of the fume flows around
behind the burner, mixes with the fuel Jet, and supplies oxygen for combustion.
Individual flames issue from each port in the burner. The remaining fume mixes
in with the hot combustion products after all of the supplemental fuel has been
burned. The gap is adjusted to provide the correct amount of fume flow through
the burner; this can be done easily after installation from outside the unit. On
the other hand, changing profile plates on line burners is a difficult task that
requires shutdown of the afterburner. A drawback to the multijet burner is that
bypassed fume (~5Q% of the total) is brought around the outside of all of the hot
combustion products, rather than dispersed throughout as in the line burner. This
presents a more difficult mixing problem since the fume must be mixed across several
feet of combustion products rather than several inches. Thus, this type of dis-
tributed burner is able to use oxygen in the flue gas and avoids problems of flame
quenching, but it does not take advantage of the reduced mixing distances possible
with distributed burners.
S-It 121
-------
3.2.1.3 Grid Burners
The grid burner presents another method of distributing the flame across
the fume duct. The North American Flame Grid Burner is schematically shown in
Figure 3-8. Gas enters through holes in the manifold pipes and fume flows through
the -1" X 3" slots in the grid plate. Flames are stabilized by the wake formed
behind the grid plate. (See Section 13.2.6). This burner provides extremely fast
mixing between fuel and combustion products because mixing distances are very
short; manifold pipes are -6" apart. There is relatively little control of fume
and fuel mixing, however, so some quenching will occur if the fume pressure drop
across the burner is too high. Also, flame stability is primarily due to the
eddies formed behind the grid. If fume pressure drop across the burner is too
low, these eddies will be weak and the flame will not be steady. Provision can
be made for adjusting the size of the grid openings (a second grid sits over the
first and can be moved as a shutter), thereby adjusting the fume pressure drop.
However, this grid burner is most suitable for applications where fume flow rate
is relatively constant.
S-1^121 Figure 3-8 follows
-------
Gas
Fume
Stream
S-14121
67784
Figure 3-8. NORTH AMERICAN FLAME GRID BURNER
-------
33
3.2.2 Possible Problems with Distributed Burners
Possible problems with flame quenching due to rapid mixing with excess
fume have already been discussed. Other problems are primarily related to the fact
that distributed burners are placed directly in the fume stream.
3.2.2.1 Burner Fouling
If liquid or solid particulates are entrained by the fume, buildup on
the burner can occur, resulting in plugging various burner parts. Mixing plates,
burner tiles, and grid plates are generally quite hot due to heat transferred from
the flame so combustible material will usually burn off. However, buildup in
cooler areas can occur without burnoff (or noncombustible material may be in-
volved) and this can lead to plugging. More serious is the possibility that
combustible deposits will accumulate and then ignite, releasing enough heat to
damage the afterburner. Distributed burners should not be used in fouling streams
unless frequent cleaning is possible.
3.2.2.2 Low Oxygen Content
Another limitation on some distributed burners is that fume oxygen
content should be above 1655. Lower values are said to be insufficient to
guarantee complete combustion of the supplemental fuel. Oxygen content was well
above this for most processes encountered in our afterburner user survey. Anytime
air dilution is used to keep below 25% of the LEL, oxygen content is expected to
be well above 16%, even in the case where afterburner exhaust is recycled to the
process for direct heat recovery (Section 6.2.U). Batch operated processes such as
wire burnoff, coffee roasting, and refuse incineration have a tendency to produce
low oxygen content fumes during part of the cycle. Certain processes involve
dilution by N2 or CO2 and fume oxygen content is near zero.
Low oxygen content can be partially overcome in any of the distributed
burners by partially premixing air with the natural gas fuel. The burners are
not designed for completely premixed operation since flammable mixtures would be
contained in the manifold but ~25/? of the air needed for combustion can be added
with minor modifications. (One cubic foot of methane requires 9 ft3 of air for
complete combustion so -2.25 ft3 of air can be premixed. This gives a -30$
methane mixture). In addition, the multijet burner and the Eclipse type burner
can be readily modified to utilize outside air for combustion. All fume is then
forced to bypass the flame.
It would also be possible to add air directly to the fume stream to
provide sufficient oxygen for combustion. However, as is shown in Table 3-3,
this is not very efficient since much more air is needed to achieve the 1.6%
minimum oxygen content in the fume than is needed through partial premixing or
fuel combustion with outside air.
S-11H21
-------
Table 3-3
Fume Oxygen
Content
16%
1.0%
Q%
Ft3 Of Outside
Air Needed Per
Ft3 Of Fume To
Achieve 1.6% 02
0
1.2
3.2
Ft3 Of Outside Air Per Ft3 Of Fume If Fuel Burned
With Outside Air And All Fume Bypassed
(OK For Discrete
25/£ Excess Air Burner But Low
For Distributed)
.55
.55
.55
100$ Excess Air
.87
.87
.87
Thus, large amounts of outside air would be needed for direct addition to the
fume and supplemental fuel must be burned to heat all of this air. In general,
discrete burners will be able to achieve complete combustion with less excess air
and therefore will show better fuel economy when outside air must be used.
3.2.2.3 Excessive Temperatures - Limit on Preheat
Distributed burners are subject to high rates of heat transfer from the
flame so mixing plates and grid plates tend to reach high temperatures. The
incoming fume provides some cooling of the plates but metal temperatures of
1800°F or higher have been observed. This temperature will increase with increased
preheat temperature, increased combustion chamber temperature and increased pressure
drop across the burner. High temperature alloys such as Hastelloy X are generally
utilized in burners available today. Many failures have occurred with aluminized
steel and stainless steels such as type SOU. Even Hastelloy X can be damaged if
combustion chamber temperatures exceed 1500°F and sufficient cooling is not
provided by the fume. Refractory construction, such as in the multijet burner,
imposes no limits at temperatures of interest.
Excessive preheat of fume can also cause decomposition of natural gas
in the distributor manifold, producing coke which plugs the gas orifices. This
becomes a problem at preheat temperatures greater than 1000°F. Such high
temperatures can occur if fume is raised to a temperature where oxidation of the
hydrocarbons in the fume occurs at a rapid rate and much of the fume's heat
content is released upstream of the burner.
S-11*121
-------
35
3.2.3 Afterburner Configuration - Mixing of Bypassed Fume
Once supplemental fuel has been burned to generate heat for dilute fume
incineration, the bypassed fume must be mixed with the hot combustion products and
then held at the required temperature for sufficient time. Both mixing and retention
(Section 3.1.1) occur within the combustion chamber of the afterburner and are
difficult to distinguish from one another. Here we discuss combustion chamber
configuration only as it affects the mixing step since this is dependent on whether
distributed or discrete burners are used. Design features pertaining to uniform
retention of pollutants at the incineration temperature and most mechanical con-
siderations are discussed in Section 3.k since these are the same regardless of
burner type.
Afterburners should be designed to give rapid and complete mixing of fume
and hot combustion products once the fuel has been combusted. Flame quenching should
be minimized and a balance must be struck between expenditure of pressure drop
to speed mixing and the added purchase cost for additional total residence time in
the combustion chamber. Chapter 11 contains cost information to guide this decision.
With afterburners designed to give a certain total residence time (since we cannot
predict mixing processes well enough for use as an absolute design criterion),
the afterburner configuration plays a major role in deciding how efficient the
unit will be at pollutant destruction.
3.2.3.1 Location of Distributed Burner
The distributed burner is placed directly in the duct carrying the fume
stream and fires into one end of the afterburner combustion chamber. It is im-
portant that fume flows uniformly through the entire cross section of the burner
in order to achieve a nearly uniform temperature profile a short distance from the
burner. The typical 1 - 1.5" w.c. pressure drop across distributed burners will
smooth out modest variations in fume velocity across the inlet duct. However,
wherever possible a straight section of inlet duct extending at least iI - U duct
diameters should be placed between any 90° bend and the burner. This is especially
important where flow variations can cause the burner pressure drop to fall below
1/2" w.c. Figure 3-21 indicates the recirculation or stall region which is
set up downstream of a sharp bend. If the burner is placed within this region,
gross maldistribution may result. (The pressure drop over the burner has a strong
effect on the location of such a recirculating region so the sketch should be
taken only as an indication of potential problems.)
3.2.3.2 Cross Sectional Area
Combustion chamber cross sectional area should be set to insure that
flows will be strongly turbulent so that mixing will occur rapidly on the molecular
level and pollutant molecules will be heated and oxidize. Turbulent flow requires
; 120 (3_2)
-------
36
or for a circular cross section,
(3-3)
where SCFM = Total volumetric flow rate at 60°F and atmospheric pressure
(Take fume flow where outside air not used for combustion), Ft3/min
D = Diameter (or equivalent), Ft
A = Interior cross sectional area, Ft2
This is an easy criteria to meet and is equivalent to a minimum velocity of 3.5
ft/sec for a 2* diameter combustion chamber. Minimum velocities several times this
high (10 - 20 ft/sec) should be utilized to help intensify mixing on the molecular
level. For a given total afterburner residence time the best mixing will be
obtained for the smallest cross section and therefore the longest combustion cham-
ber. High length to diameter ratio means high velocities and high pressure drops
but this results in better mixing and better pollutant destruction. Typically,
afterburner design velocities are -30 ft/second but L/D ratios are only 2-6.
Cross channel mixing is very slow in straight duct flow and mixing lengths equiva-
lent to Uo duct diameters are required for typical flow rates. Therefore, some
additional means is required to accelerate cross channel mixing in order to raise
all fume to the incineration temperature in a reasonably sized afterburner.
Chapter lk discusses the various means of improving cross stream mixing
of fume and combustion products. Axial Jets can be-used and give mixing distances
proportional to the width of the streams to be mixed. With distributed burners the
bypassed fume and flame streams are normally greatly subdivided and mixing occurs
in a distance shorter than the length of the afterburner. Reasonably complete
mixing (see Section lU.2) occurs in a distance about 10 - 20 times the width of the
streams. For a one foot spacing this takes 10 - 20 feet.
3.2.3.3 Baffles
It is common practice to employ line and grid burners without any down-
stream baffles or other mixing promoters. If the flame is sufficiently subdivided
and flame and bypassed fume issue from the burner with a large velocity difference,
Jet mixing will be rapid and sufficient pollutant destruction will normally occur.
However, problems can occur if flow through the burner is non-uniform (Section
3.2.3.1) so that mixing must occur across the complete channel width. The burner
should be located so that this does not occur.
The layer of fluid flowing along the walls of the duct will have a lower
velocity and experience little mixing with the main body of fluid. Therefore, it
is likely that a certain amount of fume will flow around the outside of the burner,
reach the combustion chamber walls, and pass through the combustion chamber without
S-14121 Figures 3-9, 3-10 follow
-------
O- t/>
VI I
85
tsi
400
E
O_
Q.
^
O
"«£
(0
t_
"c
§ 200
u
4
ID
U
2
TJ
X.
\
\
\
\
\
\
\
\
\
Wall with
Baffles
Represents repeat tests with turbu-
lence inducers reversed.
Chamber Temperature 1250°F
Measured 0.7 Seconds from Burner
Inlet Xylene Concentration 3500 ppm
11
Distance in Inches
Figure 3-9. EFFECT OF WALL BAFFLES
22
Courtesy: Despatch Oven Company
-------
0.7 Seconds
Adjustable
Sample
Probe
Fume-
Wall with
Combustifume Burner Turbulence
Inducers
Profile Plate
•* •* ' * ' f •"
J
K_£^C_C.A-^^^^
22" Square ^
Chamber
Exhaust
Smooth Wall
Figure 3-10. DESPATCH OVEN CO. EXPERIMENT TO STUDY
TURBULENCE INDUCERS
S-14121
67784
-------
37
attaining a sufficiently high temperature for pollutant destruction. The amount
of such wall flow depends on flow velocity, burner configuration, chamber con-
figuration, etc., but it is usually small enough so that overall efficiency can be
attained with a small increase in combustion chamber temperature.
Ring baffles or "turbulators" placed along the wall serve to divert this
wall flow and mix it with the main stream. Figure 3-9 is based on data taken by
Despatch Oven Co. showing the effect of such wall baffles on hydrocarbon concentra-
tion .7 seconds downstream of a Combustifume burner with the nominal combustion
chamber temperature set at 1250°F. The experimental set-up is schematically
indicated in Figure 3-10. Approximately h" wide baffles were placed along one
wall of the 22" square chamber. Concentrations were measured at various points
across the combustion chamber and considerable variation was observed between
concentrations at the "rough" and smooth walls.
Checkerwork baffles (bricks layed so that a gridwork is formed) are also
utilized with some line burners to generate additional turbulence and obtain a
uniform temperature profile more quickly. As noted in Section lU.8 such grid
baffles are quite effective at achieving mixing over the scale of the grid, -6",
but are nearly worthless if larger scale maldistribution exists upstream of the
baffle. For this reason, grid baffles with <50$ free area are effective in
improving mixing when distributed burners are utilized and will shorten the required
total residence time. All baffles should be located so that flame impingement does
not occur. As is noted in Chapter 7 such flame impingement can set up severe thermal
stresses and lead to failure even if the materials' temperature limit has not been
exceeded. In addition, partial flame quenching can occur.
AS'- shown in Figure 3-7b the multiJet burner normally employs baffles to
improve mixing between bypassed fume and hot combustion products since the streams
are not subdivided. The mixing problems are then similar to those in afterburners
employing discrete burners, Section 3.3. Good cross stream mixing is provided by
forcing the bypassed fume to enter at right angles to the combustion gases and
having an expansion immediately after the first baffle. The target or disc baffle
within the chamber forces another change in flow direction and provides good mixing
over large radial distances. To be effective these baffles must block a large
fraction of the flow area (-1/2) and this will add at least 1 - 2" w.c. to the
system pressure drop. This degree of baffling is often not utilized and poor per-
formance results unless excessively high temperatures are employed. The data shown
in Table 3-^ show the importance of proper use of baffles. Inlet concentration
variations and possible burner adjustments make direct comparison difficult, but
operation with the baffle is obviously better.
S-
-------
Table 3-k. EFFECTIVEHESS OF BAFFLES
MultiJet Burner - lUOO° Combustion Chamber Temperature
.3 Seconds Residence Time
Without
Baffle
With
Baffle
Inlet
Hydrocarbons
Carbon Monoxide
Outlet
Hydrocarbons
Carbon Monoxide
Maximum-Minimum Temperature
Found in Traverse Across
Combustion Chamber
900 - 1200 ppm
10 ppm
250 - 700 ppm
20 - 100 ppm
120° - 2UO°F
1300 - lUOO ppm
30 ppm
100 - 110 ppm
-10 ppm
-30°F
S-
-------
3.3 Thermal Afterburner Design Features - Systems Utilizing Discrete Burners
In Section 3.2 the distributed burners' subdivision of fume and hot
combustion products was listed as a major advantage since all fume can be raised
to the incineration temperature in a reasonably short distance from the burner.
However, there were also disadvantages (fouling, limited turndown, can't burn oil,
difficult to use outside air, etc.) and these can usually be overcome in systems
utilizing discrete burners. In addition, discrete burner costs are usually lower
per million Btu's of heat added, but as noted in Chapter 11, burner costs are a
very small factor in total afterburner cost. Figure 3-11 is a sketch of a typical
afterburner utilizing a discrete burner.
Our distinction between discrete and distributed burners is somewhat ar-
bitrary. The discrete burner is viewed as a point source of flame with a separately
controlled fuel supply. There may be four or more such burners in a large afterburner
but they are distinct, not like the multijet burner (Section 3-2.1.2). They may
be raw gas burners which sit in the fume stream and obtain oxygen for combustion
directly from the fume. Ring burners and one burner employing mixing plates similar
to the line burners of Section 3.2.1.1 are classified as discrete burners. The
commonly used nozzle mix or "torch" burners with their long flames are discrete
burners, but intense air/fuel mixing may also be employed to give a very short flame.
Discrete burners may utilize a completely premixed stream of natural gas and air
as well as raw gas or oil fuel.
The major distinction is that the burners are discrete sources of heat
and additional mixing measures must be employed to prevent fume passing through the
combustion chamber without being raised to a sufficiently high temperature for
pollutant destruction. Single Jet mixing and turbulent cross stream mixing are not
effective in afterburners with typical length to diameter ratios of 2 - 6. Many
existing afterburner designs rely completely on mixing provided by energy in the
flame Jet. This has led to poor performance and is especially a problem when firing
rate is low because of high fume preheat or high solvent content. High pressure
drops (>U" w.c.) and combustion chamber baffles must be used to obtain sufficient
mixing and adequate performance with total residence time of .5 seconds or less.
3.3.1 Discrete Burners Utilized in Available Afterburner Systems
All of the major burner manufacturers make gas and oil burners which can
be and are used in fume incinerators currently on the market. There are many
variations among these burners. These variations affect flame shape, turndown ratio,
and completeness of combustion. Chapter 13 and burner suppliers should be con-
sulted for guidance in choosing between different burners. In general, poor
afterburner performance is not due to the burner itself, but rather to how it is
utilized in the complete design. The burner is viewed as a source of heat and if
properly adjusted it will provide relatively complete combustion, at least near
the design firing rate.
3.3.1.1 Raw Gas Burners - Nozzle Mix with Low Pressure Primary Air Supply
Since natural gas is clean burning and readily available at a low price
S-1M21
-------
in most areas of the country it is most often used for supplemental fuel in an
afterburner. The most common arrangement employs a Jet of gas with primary com-
bustion air brought in around this gas Jet to initiate combustion. This primary
air in the nozzle mix type of burner is usually fresh air and set to supply no
more than the minimum oxygen needed to complete combustion. Additional oxygen is
taken from the fume stream. In order to reduce the heat load by as much as 30%, part
of the fume stream (if it will not foul or erode burner parts) could be supplied as
primary air for combustion instead of adding outside air. However, only one or two
installations were encountered in which this was done. Combustion air is supplied
at -0.5" - **" w.c. pressure and must mix with the gas stream to form a combustible
mixture. This mixing is very slow for low supply pressures so the resulting flame
is typically several feet long and not well defined. If excess fume is mixed in
too rapidly, serious flame quenching can occur, but simple entrainment of fume into
the flame jet should give no such problem. Turndown on fuel is limited to -5:1
with no adjustment of the burner. No automatic control of air supply is employed.
(See Section 3-3.1.3.)
3.3.1.2 Raw Gas Burners - All Oxygen From the Fume
The two raw gas burners discussed here operate much like distributed
burners except that they occupy a small fraction of the combustion chamber cross
section and are classified as discrete. Fuel savings associated with use of oxy-
gen in the fume result, but the problems associated with placing distributed
burners in the fume stream (Section 3.2.2) also are possible. In many cases a simple
gas ring has been placed in a fume stream, especially in an exit stack in order to
eliminate a visible plume. Problems inherent in such an installation have been
discussed in Section 3.1.^.5- A simple raw gas ring burner will add much CO and
unturned fuel to the fume unless flame stabilization is provided and mixing with
fume is controlled. Partial flame quenching is less of a problem in a lUOO°F
afterburner with uniform treatment of fume since CO and aldehydes from the flame
will be oxidized along with pollutant. Figure 3-12 shows one afterburner design
employing a ring burner. The stabilization bar provides some control over the
rate of fume mixing to avoid quenching. In this afterburner design, a mixing
baffle downstream of the stabilizer provides fairly uniform treatment of the
entire fume stream.
UOP uses a raw gas burner with a conical mixing plate in their thermal
fume incinerators. Figure 3-13 is a sketch of this burner. Outside air is
normally not used and fume is split between combustion air and bypass according
to the pressure drop across the burner. Operation is like the Combustifume type
line burner (Section 3.2.1.1) but a single flame Jet results and extends several
feet into the combustion chamber. As in the line burners, high temperature alloys
must be used for construction of the mixing plate or cone.
3.3.1.3 Raw Gas Burners - High Turndown Capability
Where high turndown of the supplemental fuel firing rate is desired»
a special wide range burner should be used. These burners control both air and
fuel supply and can achieve a U0:l or greater variation in firing rate without
much loss of stability. Combustion air must be supplied at -1 psig. Such a large
turndown is desirable in cases where the fume flow rate or composition is highly
variable (This makes successful afterburner operation difficult, however.) or
S-11*121 Figures 3-11 thru 3-15 follow
-------
Fume
Fuel
Exhaust
Combustion
Air
(Fume)
S-14121
67784
Figure 3-11. AFTERBURNER USING A DISCRETE BURNER
-------
MUT
rUWMLCT
Figure 3-12. RING BURNER USED IN STACK AFTERBURNER
COURTESY: Surface Combustion Division - Midland Ross Corp.
& The Heating and Ventilating Engineer Journal 12-9)
S-14121
67784
-------
Plenum
S-14121
67784
Figure 3-13. UOP RAW GAS BURNER
-------
Figure 3-14. COMBINATION OIL AND GAS BURNER EMPLOYING SWIRL
S-14121
67784
COURTESY: Combustion Equipment Associates
-------
S-14121
67784
OIL
MAM AM
Figure 3-15. NORTH AMERICAN DUAL FUEL BURNER
COURTESY : North American Manufacturing Co.
-------
highly efficient heat recovery is employed and a high firing rate is required only
for start up. High turndown can also be achieved by using multiple burners so that
one or more can be shut off when not required. Fume can be used for combustion air,
but normally fresh air is used to avoid fouling of the blower and burner.
3. 3.1.1* Premix Gas Burners
The kitchen gas range is a common example of a premix flame issuing from
a ring burner. Gas and air are mixed up stream of the burner and a relatively
short, stable flame results. (See Section 13.2.3.1.) Where premix burners are
used in afterburners they have usually been a single nozzle with the air/gas mix-
ture produced in a premixing blower. Outside air is used since burner ports are
small and subject to fouling. High turndown is possible (10:1 - 15:1 or more)
since both air and fuel are easily varied. Premix burners are not normally
utilized for heat duties above one half million Btu's/hour but rapid heat release
and short flames are attainable.
3.3.1.5 Oil Burners and Dual Fuel Burners
In certain areas natural gas is difficult to obtain and/or it is very
expensive. In such cases oil fired afterburners are attractive or one may want
the capability to substitute oil fuel during periods where the natural gas supply
is interrupted. Gas shortages are a relatively recent phenomenon so many after-
burner suppliers do not regularly supply an oil fired option. However, most designs
employing discrete burners can be adapted to oil fuel. Oil firing increases flame
luminosity and therefore will transfer more heat by radiation. (However, the
fume, unless it contains participates, will not absorb much heat by radiation. The
chamber walls and any other solid objects which "see" the flame can be expected
to exceed the average chamber temperature unless cooled from behind.) Also for
similar burner mixing patterns, oil flames can be expected to be longer than gas
flames by 50$ or more since the oil must first vaporize. These points are dis-
cussed in Sections 13.2.1+ and 13.3.2.1. Maintenance requirements will be higher
with oil burners than gas burners.
Figure 3-lU shows a typical combination oil and gas burner. Raw gas enters
through the ring at the exit end of the burner. An oil gun is inserted through the
hole in the center of the burner. Combustion air or fume enters through the louvers
along the outside of the burner. The louvers are set to give some spin to the air
giving more intense mixing and a shorter flame. Spin is discussed in the next
section. The burner is mounted on the end of the combustion chamber and surrounded
by a "windbox" which directs fume or air through the louvers. The slotted plate
surrounding the oil gun serves to stabilize the flame by recirculation and spin
(Section 13.2.6). Figure 3-15 shows the commonly used North American Dual Fuel
Burner which does not use spin on the combustion air.
Oil fuels must be sprayed into the chamber as small droplets so the
vaporization is fast and combustion complete. Typically, afterburners employ #2
fuel oil, a relatively low viscosity home heating oil. It is "pressure atomized"
through spray nozzles with oil supplied at 25 - 100 psig. Alternatively, oil can
be supplied at low pressure and air Jets used for atomization. In some cases
S-1U121
-------
U2
heavy residual (#6) oil may be utilized and this requires heating to lower its
viscosity and use of high pressure steam or air jets to break up the oil into
small droplets. (Section 13.2.k discusses oil atomization in more detail.)
Residual oil is cheaper but it normally is high in sulfur and nitrogen which will
end up in the afterburner flue gas so cleaner fuels are preferable.
Oil burners including their atomizers have a limited turndown ratio of 3:1
- U:l. Pressure drop on the fume or air side ranges from 5" - 10" w.c. f. r the
type shown in Figure 3-lU. Lower pressure drop operation (-2" - V1 w.c.) is
possible with burners similar to low pressure raw gas burners but flame lengths
are much longer and the supplemental fuel heat of combustion is not available for
fume heating until far down the combustion chamber. If oil burners are used, it
is crucial that oil guns are blown out and pulled from the burner immediately
after the oil flow is stopped. Otherwise carbonization will occur and burner
plugging and poor atomization will be a problem.
3.3.1.6 Use of Swirl To Obtain Rapid Heat Release
As noted in Section 3.1.3 little benefit is to be derived from "flame
contact" whereas partial flame quenching leading to high CO, aldehydes, and
unburned fuel in the flue gas is quite likely. Therefore, supplemental fuel should
be combusted as rapidly as possible to provide maximum residence time for mixing
with bypassed fume and retention of pollutants at the incineration temperature.
A precombustion chamber with careful control of combustion air (fume) would insure
complete combustion prior to entering the afterburner combustion chamber.
Intense swirl or vortex burners can be used to speed up air/fuel mixing
and achieve supplemental fuel combustion with a very short flame. Figure 3-l6 is
a sketch of a vortex type burner. Combustion air (which can be part of the fume
stream if it contains sufficient oxygen and isn't fouling) is supplied at high
pressure (-16 - 20" w.c.) and much of this pressure is lost in passing through the
swirl vanes (very tight louvers such as in Figure 3-1*0. Air then passes around
the fuel Jet with a high rate of spin and expands into the precombustion chamber.
This spin and expansion leads to the circulation patterns indicated in Figure 3-l6.
Mixing and combustion are rapid and very little flame extends beyond the end of the
precombustion chamber which is considered part of the burner. As is discussed in
Sections 13.3.2.3 and lU.3-2.5 flame shape and mixing intensity depend on the ratio
of axial velocity to swirl velocity at the inlet to the burner.
Such a swirl burner can be used to provide a source of hot combustion gases
without danger of flame quenching provided that the air/fuel ratio stays within
combustible limits (Section 12.1.2). Mixing is nearly complete so one must be
extremely careful that too much excess air (or fume) is not put through the burner
and that oxygen content remains high enough. Typically, such burners are run with
close to the stoichiometric amount of combustion air so high temperature (~3000°F)
combustion gases result. If part of the fume stream is used for combustion air it
must be passed through a blower to provide the required high pressure (-16" - 20"
w.c.). Fouling and erosion are potential problems. Because of the sensitivity of
mixing patterns to flow rate, turndown is limited to -3:1.
S-1U121 Figures 3-16, 3-17 follow
-------
Swirl Vanes
to Give Intense
Spin to Air
Precombustion
Chamber
Natural Gas
or Oil Fuel
s/sssss/s/ss/s
_^=. Short
—— Intense
- J^=- Flame
Combustion
Air (Fume)
Figure 3-16. VORTEX TYPE BURNER-RAPID AIR FUEL MIXING
S-14121
67784
-------
S-14121
67784
me*
fTMKI
MLCT
L
1
Figure 3-17. USE OF FLAME JET TO INDUCE FUME FLOW
/URTES
,.
URTESY: Surface Combustion Division - Midland Ross Corp.
The Heating and Ventilating Engineer Journal «*>
-------
3.3*2 Afterburner Configuration - Mixing of Bypassed Fume
Afterburners should be designed to give rapid and complete mixing of fume
and hot combustion products once the fuel has been combusted. Flame quenching should
be minimized and a balance must be struck between expenditure of pressure drop to
speed mixing and the added purchase cost for additional total residence time in
the combustion chamber. Chapter 11 contains cost information to guide this decision.
With afterburners designed to give a certain total residence time (since we cannot
predict mixing processes well enough for use as an absolute design criteria), the
afterburner configuration plays a major role in deciding how efficient the unit
will be at pollutant destruction.
As noted in Section 3.2, most distributed burners can be axially fired
into an unbaffled combustion chamber and mixing will be fast enough to give a
nearly uniform temperature profile in 10' - 20'. However, with discrete burners
simple turbulent flow cross mixing and in most cases axial Jet mixing cannot be
expected to bring all fume up to the required incineration temperature in the
distance available in commercial afterburners. Burner location, baffles, and
combustion chamber configuration are all used to improve mixing and afterburner
performance. Chapter lU should be consulted for a more detailed discussion of
various mixing arrangements and a discussion of what constitutes adequate mixing
in practical afterburners.
When Maxon was first applying their Combustifume distributed burner
(Section 3.2.1.1) they compared its performance with that of a single axial fired
burner in a -U1 diameter unbaffled combustion chamber with length -2-3 times
the diameter. Fume was preheated in a recuperative exchanger. Results showed
that outlet hydrocarbon concentration was reduced to -10 ppm at 1250°F with the
distributed burner. Concentrations were >200 ppm using the discrete burner, and
the nominal combustion chamber temperature had to be raised above 1550°F to achieve
similar hydrocarbon destruction. In addition, the combustion chamber length
had been reduced in installing the distributed burner. Since hydrocarbon des-
truction is rapid at 1100° - 1200°F (Section 3.1.U.2) this demonstrates the
problems caused by poor mixing.
As discussed in Section 3.2.3.2, cross sectional area should be chosen so
that flow will be strongly turbulent so that mixing occurs rapidly on the molecular
level to heat and oxidize all pollutants. This is important for small scale
mixing. However, additional means must be provided for cross channel mixing. In
Section 3-2.3.2 it was noted that mixing is improved by high velocities and high
length to diameter ratios. This is limited by pressure drop across the after-
burner and mechanical design considerations.
S-1M21
-------
3.3.2.1 Axial Flame-Jet Mixing
A common afterburner arrangement is to utilize a single burner firing
into the end of the combustion chamber with fume brought in around the burner
as a parallel stream (see Figure 3-12). If a high excess air burner (Section
13.2.6.6) is used, the fume may all go through the burner but part of the fume
(~505&) must still flow around the flame and mix in downstream. As noted in
Section lU.U, surrounding fume will slowly be entrained into the hot combustion
gases so there is little danger of flame quenching. However, large temperat'ire
variations would be expected across the exit end of an unbaffled combustion
chamber with length to diameter ratio less than 10 - 15- Typical length to
diameter ratios found in available afterburners are ~2 - 6 so much fume can pass
around the hot combustion gases and escape without being raised to a sufficiently
high temperature. Additional mixing promotors are required.
One means of improving simple axial Jet mixing is to reduce the diameter
of the combustion chamber after fume has been entrained into the jet of hot combus-
tion gases. One means of doing this is shown in Figure 3-17* This particular fume
incinerator utilizes the flame jet to create an eductor effect which pulls fume into
the afterburner. Only that fume which can be entrained by the jet is pulled in. If
instead, the fume were to be supplied under pressure, it would not be completely
entrained and cold fume would surround the hot gases as they flow through the
throat. The situation would be similar to that discussed above. In either case
there will be considerable flame quenching unless the flame extends only a short
distance beyond the burner block. Addition discussion can be found in Section lU.3.3.
Some early afterburner designs employed a discrete burner firing upstream
into the fume flow. This may result in some increase in mixing rate but the amount
of flame quenching was dieasterous. Flame stability was also a problem when fume flow
rate was high. This arrangement should be avoided, unless fuel is burned in a
separate chamber and only hot combustion products mix with the cold fumes.
3.3.2.2 Radial Entry of Fume or Flame
Cross stream mixing is much faster when the streams to be mixed enter at
right angles (or at least some angle) rather than parallel. Burners often enter
from the side, especially in stack afterburners but this usually leads to flame
impingement and shortened refractory life (See Section 7.2). Also the improved
mixing and attempts in many designs to "fill the cross section with flame" lead to
severe flame quenching and formation of CO and aldehydes (Section 3.1 •*<•)• In a
high temperature (-1UOO°F) afterburner such flame quenching is not too serious
(if good downstream uniformity exists) since intermediates from the fuel will be
oxidized with the fume. However, in a low temperature (~1000°F) smoke or odor
abator the CO and aldehydes will remain in the flue gas and cause an increased
pollution problem. If low temperature units are to be used, fuel should be burned
in a separate chamber (burner block extension) and only hot combustion products
should mix with excess fume.
A more satisfactory arrangement employes an axially fired burner with
multiple, high velocity cross.flow jets of fume mixed in downstream of the flame.
S-
-------
More typically the fume enters through one or two radial streams at relatively low
velocity and does not penetrate the flame Jet very far. This is similar to the
multijet burner in Figure 3-7b. Baffles must "be provided to enhance downstream
mixing since the situation is similar to axial entry of both fume and flame.
3.3.2.3 Tangential Entry of Fume or Flame - Swirl
As discussed in Section 3-3.1.6 swirl can be used in burners to create
intense mixing and give very short flames. Tangential entry in an afterburner can
also be used to speed up mixing of bypassed fume and hot combustion products.
Pressure drop must be expended to generate swirl and speed up mixing. In vortex
burners this pressure drop may be -18" w.c., but typically the total pressure
drop across an afterburner is only 2" - V w.c. so relatively slow mixing is
expected.
The cold fume stream should enter axially and the hot combustion products
should enter tangentially for improved mixing performance. If instead the fume
is brought in tangentially at high velocity, the cool (and therefore heavier) gases
remain near the wall due to cyclonic action and mix more slowly with the flame than
if both were brought in axially. Tangential firing of the burners is preferred.
A vertical upflow afterburner with tangential firing is a good arrangement for
incinerating fumes containing heavy particulates (Section 3.1.2.6) since they
are held near the bottom of the combustion chamber until consumed and experience
a much longer residence time than the average.
Flame quenching will be a problem with tangential flame entry, especially
in designs which attempt to fill the combustion chamber with flame. (In most
cases the flame actually follows a helical pattern along the wall of the afterburner
since the flame jet is weak compared to the fume stream, especially at lower firing
rates.) As noted in the previous section a certain amount of flame quenching can
be tolerated in lUOO°F afterburners with uniform flow downstream since these
generated pollutants will also be oxidized. However, it would be better to complete
the flame within a burner extention and inject the combustion products for mixing
with fume. Refractory used with tangential flame entry will be subject to flame
impingement and possibly require increased maintenance.
As in the swirl vortex burners (Section 3.3.1.6), mixing will be improved
if the combustion chamber diameter expands immediately downstream of the point where
swirling flow is generated. This generates backflow and much cross chamber mixing.
Where vortex burners are used, some swirl is still present in the combustion gases
issuing from the burner and can be used to provide some energy for mixing with the
bypassed fume.
3.3.2.U Baffles
In most discrete burner fume incinerators on the market today, the
combustion chamber is not long enough to provide adequate mixing (and pollutant
destruction) without the use of baffles. Baffles can greatly speed cross chamber
mixing of fume and combustion products. However, they must be used correctly or
else they will only add pressure drop and set up recirculating dead zones in their
wake which remove active chamber volume and shorten effective residence time
S- 1M21
-------
46
without improving mixing. Section lU.8 contains a more detailed discussion of
baffles.
Baffles can be used to compensate for lack of a distributed source of
hot combustion gases by diverting streams over relatively large distances across
the combustion chamber. Also they serve to block off a major part of the chamber
cross section so flow velocities are rapidly increased and then decreased on the
downstream side. This generates a high level of turbulence and speeds local
mixing at the expense of pressure drop.
As discussed in Sections lU.8.1 and lU.8.2 ring and disc baffles and
bridge wall baffles, shown in Figure 3-l8a and 3-l8b, are effective in forcing cross
channel mixing over large distances. They should be used in pairs for best results
and sized to block 60% or more of the chamber cross section. This will add
several inches of water pressure drop across the afterburner, but will improve
performance. Fairly good mixing has been demonstrated with ring and disc baffles
at the same point in the afterburner so that an annular slot is formed with
-25% free area (See Figure 3-12). As discussed in Section lU.8.2 less than
50°F temperature variation is obtained within 2 chamber diameters downstream of
this baffle.
Baffles should be located so as to complete mixing as rapidly as possible
leaving maximum residence time after all fume is brought to temperature. Flame
impingement is to be avoided so the first baffle is usually I1 - 2* beyond the
maximum burner flame length. Subsequent baffles should be located -1 chamber
diameter downstream, similar to what is shown in Figure 3-18. The common place-
ment at 1/3 and 2/3 of the chamber length is probably not optimum.
Section 3-2.3.3 contains a discussion of the important role ring baffles
and grids can play in•afterburners utilizing distributed burners. Ring baffles
along the wall tend to prevent escape of cold fume from the afterburner by flowing
along a cool vail. Such wall flow is diverted into the main stream. These would
also play the same role in otherwise unbaffled afterburners utilizing discrete
burners. However, overall performance will be much better if gross cross channel
mixing is provided by the disc or bridge wall baffles just discussed.
The small scale of cross stream mixing provided by grids or checkerwalls
(bricks are layed so as to form a gridwork across the chamber) also limits their
usefulness with discrete burners. The gridwork can serve to eliminate velocity
gradients if a pressure drop -1" w.c. is taken over the grid but temperature and
concentration non-uniformity will only be affected over the scale of the gridwork
(-6"). In one case where three small discrete burners were equally spaced across
the entrance to a I1 by 5' cross section combustion chamber, a checker-work grid
was put across the chamber in an attempt to improve mixing and increase the <1Q%
efficiency of pollutant destruction. The overall destruction efficiency was in-
creased only -2% because the scale of maldistribution was much greater than the
size of the gridwork and fume continued to bypass without reaching incineration
temperature. (Heat exchanger leakage also contributed to poor performance.)
S-1^121 Figures 5-l8a, b thru 3-20 follow
-------
S-14121
67784
r
14D
Fume
Fume
A. Ring and Disc Baffles
Fume
Fume
///////////////////////////// ///////////////////jf
//////////////////////////// /////////////////////f
B. Bridge Wall Baffles
Figure 3-18. BAFFLE MIXING DEVICES
-------
Burner
1
/
/
/
\
1
Fume Flow Direction
1 i
'':••:•: v.;-.*;J: 5; ;•:•;:;
^.^.•;v;.;;;;;i
\ /
~^n — — ~
Flame
)i
'',
\
— •
— -•
'///////////////.
1 »
•:..:•.-.;••.::'.••/.::/.-•;;;.•-::
:'••.'.; Packed Bed ;V
\ /
-~-- - ^-- — "
, - -*^^Z
- • " ~
7
^
^
,^-
/^ ////// /////////
1 *
.-;-./ .:.•*;•;.!.:•-•/ :-.v
:::;vV;.::'^^^'
I. /
Flame
— — -
y
i
^
Bu
^
-•:.v'-;'i://.:v:- /;.:•• ••••.•'V.''': 0; -K:.'•••'•'••'i'••'*'^'•-'•"••'',
-H
•I-!
lf// /////////////////////////////////////////////
Figure 3-19. PROCTOR-SCHWARTZ TRAPS SYSTEM WITH
REGENERATIVE HEAT RECOVERY
Top View with Ducting Removed
S-14121
67784
-------
^> Contaminated
Gas from
Process
Inlet
Valves
Outlet
Valves
Gas
Outlet
Duct
Clean
Gas
Purge
Valves
Energy Transfer
Media Courtesy: Proctor and Schwartz, Inc.
Figure 3-20. VALVE AND DUCTING ARRANGEMENT IN PROCTOR SCHWARTZ TRAPS SYSTEM
-------
3.3.2.5 Bends and Changes in Combustion Chamber Cross Section
As is discussed in Section 3.1*.! flow around a 90° or 180° bend is
not a good mixing promoter unless the combustion chamber extends for -6 - 10
chamber diameters after the bend rather than the normal 2 diameters or less.
Recirculating dead zones are set up which reduce the total residence time in the
chamber so such bends should be avoided where possible.
If strong swirling flows are involved, a sudden expansion of the com-
bustion chamber cross section can set up recirculation which significantly speeds
mixing. With simple axial flow abrupt expansions and contractions of the cross
sectional area produce some increase in mixing rate, but this is more effectively
done using baffles.
3.3.2.6 Packed Bed Regenerative Heat Exchange - TRAPS System
Heat recovery systems and associated heat exchangers are discussed in
Chapter 6 since they are usually add on devices - They make little difference in the
operation of the combustion chamber except that a hotter inlet fume must be
handled and little mixing energy will be available from the small supplemental
fuel flame required. This is not the case with a new high heat recovery fume
incineration system offered by Proctor Schwartz under the name TRAPS. A sketch
of this system (top view) is shown in Figure 3-19. The unit supplied for a mini-
mum fume flow -10,000 SCFM is composed of a minimum of three modules (as shown
in Figure 3-19) with packed bed heat exchangers along the sides of the combustion
chamber and a discrete burner firing from each end. Fume flows in through a packed
bed exchanger where it is preheated to -1100°F, flows across the combustion
chamber where it is raised to 1UOO° - 1500°F, and flows out through the packed
bed exchanger on the opposite side where it is cooled to ~600°F. Ductwork and
butterfly valves are arranged to reverse the flow direction in each module every
-2 minutes so the cold incoming fume can pick up the heat transferred to the
packed bed exchanger by the hot exhaust fume on the previous half cycle. Figure
3-20 indicates the arrangement of these valves and ducting. Between each half
cycle a short air purge is required to sweep contaminated fumes out of the region
between the inlet valve and the combustion chamber. Otherwise, when the flow
direction is reversed, this fume would be swept out the exhaust without treatment.
The regenerative packed bed exchangers offer high heat recovery efficiency;
the manufacturer claims 7555 or &3% recovery depending on the depth of packed bed
used. The ceramic packing has a high heat capacity so only 25° - 50°F temperature
variations are expected at any point in the packed bed during the -2 minute half
cycle. (There is a sharp temperature change across the bed, -1300° on the com-
bustion chamber side and -500° on the inlet (outlet) side.) Pressure drop (-8" w.c.)
is expended in obtaining this high heat recovery efficiency.
In order to minimize pressure surges in the fume inlet duct (which might
upset the fume generating process) each module is operated on an independent cycle
so that only one set of valves is closed for air purging at any time. Therefore,
adjacent modules will often have flow in opposite directions, and there will be a
strong tendency for fume to enter the combustion chamber through one packed bed
-------
and flow out through the adjacent packed bed without crossing the chamber. This
could lead to very short residence times in the chamber compared to that calculated
for flow across and out the opposite side and necessitates the use of the baffles
shown in Figure 3-19 between each module. These baffles insure a minimum flow
path between modules of .3 - . U seconds residence time at 1UOO°F. Some pollutant
destruction probably also occurs within the inlet packed bed.
Problems could arise in achieving good pollutant destruction if large
temperature gradients can exist within the combustion chamber. Cool flow paths
could then be set up and stabilized by the heat exchange process. As in con-
ventional designs, the control thermocouple only records the temperature at a
single point. Mixing within the -8' x 5.5' x 11' (for three modules) combustion
chamber is provided by the flame jets at each end. Their velocity is several times
higher than the nominal cross flow velocity of the fume but relative.1 v little heat
input is needed (-2 - 3 MM Btu/hr from each burner) because of the high heat
recovery. For fume flow rates above 10,000 SCFM more modules are added but two
burners continue to be used and must provide the mixing action.
As mentioned previously these units are new and operating experience was
not available for us to assess the adequacy of chamber mixing and therefore pollu-
tant destruction. Such experience is also needed to determine the ease of operating
on fouling fume streams. Organ!cs should be able to be removed from the packed
beds by burning off, but inorganics may require repacking of the exchangers.
S-
-------
3.*4 Thermea Afterburner Design Features - System Configuration and Construction
In this section we consider thermal afterburner design features which are
independent of the type of burner used. The combustion chamber must be designed to
provide uniform treatment of all fume once it is raised to the incineration temper-
ature and hold it for sufficient time for oxidation to occur (Section 3.1.2).
Operation can also be affected by the orientation of the unit, especially when
mounted vertically to serve as its own stack. In addition to providing pollutant
destruction, the afterburner must be constructed with sufficient mechanical strength
to withstand the lUOO° - 1500°F combustion chamber temperature and rapid heating
and cooling cycles.
3.U.I Plug Flow Retention Section - Straight Through Circular Chamber
Once all of the fume has been raised to the required incineration
temperature a plug flow condition should be maintained so that all of the fume
is given equal residence time at this temperature. (Plug flow is an idealized
picture in which fluid moves as a solid and there is no backmixing or cross
mixing between different locations in the gas.) When an approximation to plug flow
is not maintained, large variations in velocity and recirculation patterns can
exist. This means that some of the fume can be swept through the
chamber in a small fraction of the calculated residence time
while others remain trapped for a long time in recirculating "dead zones." Of
course plug flow should be avoided until all fume has been mixed with the hot
combustion products. Otherwise temperature nonuniformity will be preserved in
the short flow path provided in available afterburners and fume will escape with-
out sufficient pollutant destruction.
As is discussed in Chapter 15 a straight, unbaffled chamber with cir-
cular cross section will give a close approximation to plug flow and residence
time will be the chamber volume divided by the actual volumetric flow rate. Cross
stream mixing is very slow so nearly uniform temperatures must be obtained by
mixing upstream. Afterburners commonly have a rectangular cross section because
of ease of fabrication, especially for large sizes. Some circulation occurs in
the corners and flow may be restricted in these regions. As a result residence
time is slightly less than calculated from the total chamber volume.
More serious reduction in residence time can occur with 90° or 180°
turns in the combustion chamber. The 180° or U bend construction in which the
combustion chamber is folded back on itself is quite common in available after-
burners . It achieves a slight cost reduction and a considerable reduction in
total length since less exterior wall is used. Also, if recuperative heat
recovery is desired> it is easy since inlet and outlet are close together. How-
ever, some problems have been experienced with overheating of the dividing plate
since it is heated from both sides. Materials selection, either metal or re-
fractory, is crucial (Chapter 7). We encountered several cases in which this
baffle warped, collapsed, or required extensive repair.
The 180° turn is usually put forth as a mixing promoter which will improve
afterburner performance, but instead it more likely results in less favorable
S-
-------
50'
performance. Cross stream mixing is improved downstream of such a bend but 6-8
combustion chamber diameters are required to achieve such mixing. Since after-
burners typically have L/D ratios -2-6 with only half of this following the
180° bend, not much mixing will be accomplished. Figure 3-21a shows the recir-
culation patterns which are likely to develop downstream of the U bend and in
the corners of the chamber. These are fairly stable patterns and can result in
as much as 25? of the chamber volume being unavailable for fume flow. An equiv-
alent reduction in residence time results. As discussed in Chapter 15 large varia-
tions in velocity are found within the active chamber volume so that some fi tie will
be swept through the chamber in a small fraction of the average residence time while
other parts of the fume recirculate for a long time before being released. Figure
3-21 also shows the recirculation patterns developed downstream of 90°
turns.
A checkerwork baffle or grid should be placed at the start of the reten-
tion section if mixing patterns can be determined well enough to define the point
at which fume has been sufficiently well mixed with hot combustion gases. This
grid will serve to even out velocity variations across the chamber and generate
turbulence on a scale which speeds mixing and reaction on the molecular level.
3.U.2 Vertical or Stack Afterburners
Considerations specific to each installation usually determine whether
a vertical or horizontal afterburner will be used (location, allowable roof
loading, expected wind speeds, etc.). Gravity can cause some stratification
between hot combustion products and cold fume and make mixing more difficult in
a horizontal unit, but this is a minor consideration since flow velocities are
high and buoyancy effects are usually small. Vertical units serve as at least
part of their own exhaust stack which is an advantage but also can lead to problems.
If the vertical unit has an open top and has a short length to diameter ratio, much
heat will be lost from the combustion chamber by radiation to the sky. Insufficient
treatment of much of the pollutant can result. Also strong cross winds may affect
flow patterns within the chamber and result in cold zones. Both of these problems
make it difficult to find a suitable location for the control thermocouple.
In order to avoid these problems, a vertical afterburner combustion
chamber should have a reduced cross section at the exit end to give a higher exit
velocity and reflect much of the radiation which could be lost. Also a rain hat
can serve to reduce such losses. Short, squat vertical fume incinerators (L/D
2-3) should be avoided and a stack extension would be recommended. Horizontal
afterburners require a 90° turn into the stack so these problems do not
occur.
3.^.3 Induced Draft - Avoiding Leakage
Combustion chambers can be run at pressures either above or below
atmospheric pressure by placing the fume blower either before the afterburner
(forced draft - positive pressure) or following the afterburner (induced draft -
negative pressure). If leaks develop in a chamber under slight negative pressure,
outside air will enter and produce a very small increase in fuel requirement.
S-14121 Figures 3-21, 3-22 follow
-------
A. U Bend
Shaded Areas Represent
Recirculating Flows -
Dead Zones
2D
_[
B. L Bend
Figure 3-21. REDUCED RESIDENCE TIME EFFECT OF 180° AND 90° TURNS
S-14121
67784
-------
S-14121
67784
Control
Thermocouple
Fuel
rf^4^^4^4^Ubb£^i±4»Ui£rfi&i>4«^bb^&£JUu
^- ,. r^J
' '//////////y////yy/////////7v
LMLLLX
7- ') "
,^>Pt ~\ "
Exhaust
V^
//.////.//yy.A.y.////////////////////^
,,,,.,, ^.^..^^.r^^^ ^ j^t.j^, +^IJT
Fume
Figure 3-22. ALL METAL AFTERBURNER - ANNULAR DESIGN
-------
51
Hovever, in a forced draft system hot gases will escape through the leak and vill
probably make it bigger. The gases issuing from the leak probably still contain
pollutant and may be toxic and offer danger to nearby personnel because of their high
temperature. Through proper mechanical design and use of heavy gauge metal and
careful assembly the possibility of leaks can be minimized. Induce'd draft systems
offer an extra degree of safety, but higher blower costs are incurred (see Section
8.2). Some afterburners designed for forced draft operation, are constructed with
a double metal wall around the refractory. A separate, low volume blower delivers
outside air to the space between the walls at a pressure greater than that in the
combustion chamber. This again prevents fume and hot gases from escaping and
allows use of loose expansion Joints to prevent thermal stress buildup.
Many problems have occurred with tube leaks in recuperative heat ex-
changers due to thermal stresses and improper tube fastening. With either forced
or induced draft systems the clean, hot flue gas will be at lower pressure than the
polluted, cold fume. Therefore, untreated fume leaks into the cleaned exhaust
gases and greatly reduces the efficiency of the afterburner system. The only way
the leakage direction can be reversed is to locate the blower between the preheater
and the combustion chamber, not a common practice.
S.^.U All Metal Construction - Annular Fume Inlet
Many early afterburners were essentially burners in a stack designed to
eliminate a visible plume and operate at 600° - 800°F. At these temperatures
carbon steel or low alloys could be used successfully. However, as noted in
Section 3.1.2.3 and elsewhere, such low temperature afterburners cause more
pollution than they abate and should not be used. High temperature (-lUOO°F)
afterburners should be constructed with a refractory lining to act as an insulator
and protect the outer metal shell which provides structural strength and seals
the unit. Refractory selection, cost, and installation techniques are discussed
in Section 7-2.
Refractory is heavy, adds expense to the unit, and requires relatively
slow warm up of the afterburner. (Refractory can be damaged if heated or cooled
too rapidly, but it also has a high heat capacity so a large refractory lined
afterburner will take a minimum of 5 - 15 minutes to come up to operating
temperature from a cold start.) For these reasons all metal construction is used
in some units. The most common design involves a double shell as indicated in
Figure 3-22. Cold fume comes in around the combustion chamber and keeps the metal
wall cool.
Our survey indicates that this design has led to a great amount of
trouble, primarily structural failure. This has resulted in severe leakage, the
need for extensive repairs, and in some cases abandonment of the afterburner as
inoperable. The problems seem linked to the inability to handle thermal stresses
which are developed due to the sharp temperature gradients in this design. In some
cases there has been evidence of excessive temperatures due to combustion of con-
densate which built up in the annular space.
S-1U121
-------
52
It has also been claimed that this annular design serves to preheat
the incoming fume and recover heat, but this is only marginally true if at all.
The amount of heat transfer area is very limited so even if fins are used to
increase heat transfer, only 100° - 200°F preheat will be experienced by the
fume. But in the arrangement indicated in Figure 3-22 even this amount of re-
covered heat cannot be utilized. The heat needed to preheat the fume comes from
the gases in the combustion chamber so to hold ll*00°F at the chamber exit the
average temperature at flame end must be 1500° - l600°F. The same amount of
fuel must be burned as if the fume entered without preheat and a lUOO°F exit
temperature is maintained.
Perhaps the structural failure could be overcome through proper mechanical
design and use of heavy gauge stainless steels. However, the design still suffers
from the presence of cold walls in the combustion chamber. As discussed in Sections
3.2 and 3.3 wall flow leading to fume escaping without adequate treatment is a
potential problem with refractory lined fume incinerators. If the walls are "cold,"
quenching of the oxidation reactions will occur and fume escape is nearly inevitable.
3.U.5 Afterburner Should Be Rugged and Well Designed Mechanically
Many light weight, poorly constructed, poorly controlled afterburners
have been sold, usually at a low price often by companies anxious to get into the
pollution control field. These units are often bought by companies anxious to
satisfy some new air pollution code or to quiet complaining neighbors. In many
cases structural failure results in a year or less and the money spent is lost or
much money and effort is spent trying to beef up the unit.
Mechanical design of afterburners is beyond the scope of this handbook,
but it is important that one deal with an experienced and reputable manufacturer.
A major difficulty is involved with handling the thermal expansion which occurs on
heating the unit to lUOO° - 1500°F and even hotter in the vicinity of the flame.
A long unit may expand several inches so provision must be made to allow for this
movement and minimize thermal stresses. Such stresses are a special problem when
hot and cold gases are close together (as in a heat exchanger) or a surface is
rapidly heated and cooled. Problems also arise when materials with vastly dif-
ferent thermal expansion coefficients (metal and refractory) are used together.
Chapters 6 and 7 contain additional discussion of methods of avoiding or minimizing
thermal stress buildup.
In most units mechanical strength and sealing are provided by an exterior
metal shell. As discussed in Chapter 7 most steels, even stainless steels, lose
much of their strength at temperatures of 1000°F and above. Therefore a refractory
lining is used to keep the metal temperature much lower. The advantages and draw-
backs of various metals and refractories are discussed in Chapter 7. Afterburners
are manufactured with shell thicknesses ranging from 20 gauge to 3/8" plate. In
general, the thin metal shells have not had sufficient mechanical strength to
stand up under typical afterburner service.
-------
53
3-5 Users Survey on Thermal Afterburner Performance
Questionnaires were filled in covering 21^ thermal afterburners. A
complete tabulation of the data is given in Chapter 11 (Table 11-2). Pertinent
information has been extracted and summarized in Table 3-5- In general, the
thermal units were reported to be satisfactory overall, only 11$ of the units
being regarded as unsatisfactory for any cause.
The percentage reduction in hydrocarbon was given in only a limited
number of cases, and it was difficult to determine the reliability of the figures
quoted. Most units were claimed to give greater than 90$ reduction where numbers
were reported. Generally units are adjusted during start-up to meet the local
pollution requirements, a temperature is determined at which satisfactory operation
is obtained, and no further testing is carried out. If the unit is operating at
the temperature agreed to by the local control board, then it is assumed to be
doing an adequate job.
Main operating problems involve safety controls, refractory linings,
heat exchanger fouling or mechanical failure, and bearing failure in fans.
However, most units are available 95-99$ of the required operating time, so that
maintenance is a minor problem.
Table 3-5. USERS INFORMATION ON THERMAL AFTERBURNER PERFORMANCE
Overall Performance
Number
Installed
211*
Number with
Heat Exchangers
83
Number
Not Yet Run
7
Performance
No Comment
2
Satisfactory
Number
181*
% of Those in Use
89
Reported Hydrocarbon Reduction
Reduction in HC
Number reported as meeting requirements
Number not meeting requirements
95-100
in
90-95
21
85-90
2
1
80- 85
2
1
Total
66
2
Main Operating Problems
Category
Number of afterburners having
problems in each category
Safety Controls
Flame- Rod
32
Refractory
19
Heat Exchanger
16
Fan
10
S-1M21
-------
3.6 Conditions Required for Adequate Performance in Some Applications
Table 3-6 presents operating temperatures recommended or used for
thermal afterburner applications in various industries. The first column gives
temperatures recommended by the LAAPCD (taken from Table 51 of the revised
edition of reference 3-1) for these applications when using afterburners having
0.3-0.5 seconds residence time. The second column presents the range of
temperatures reported by users (in the Users Survey; to give generally satisfactory
operation. The final column gives temperatures which have been quoted in various
literature articles describing the use of thermal afterburners for these
applications.
Conditions in the usual paint baking and drying oven are such that the
oven burner does not contribute significantly to the destruction of oven fumes.
Claims have been made by the manufacturer of one oven design that its use of air-
jet-driven recirculation and mixing of the oven atmosphere with the hot gases
from the burner results in significant oxidation of fume material in the oven.
These claims are not substantiated by available test data. It is possible that
some material may be oxidized, but only a small fraction of the oven contents
will be heated to a temperature in excess of 1300 or 1^CO°F for a tenth of a
second or so, as is required to achieve a significant conversion. The rest of
the oven gases can be expected to remain at a well-mixed temperature near the
normal 300 to 500°F oven operating temperature and remain unoxidized.
3»7 References
Many articles on afterburners have appeared in trade and technical
journals, but few are referenced here. Most of these articles are in the nature
of advertising and discuss the features of a particular manufacturer's unit.
Articles discussing necessary design criteria and types of applications are much
less complete and objective than the material contained in this handbook. In
many cases essentially the same material has been republished time and again.
A few articles are thought to be useful because they relate experience with after-
burner installations in particular industries and contain some performance data.
However, as noted previously, most performance data is not useful since the
operating conditions are not constant and the design features of the unit are
often not reported.
The "Air Pollution Engineering Manual" prepared by the Los Angeles
Co. Air Pollution Control District3-1) is recommended for its industry by
industry treatment of the air pollution problem and control techniques. Other
useful references to supplement this handbook are 3-2 to 3-7 found in the
Bibliography.
S-Ik 121 Table 3-6 follows
-------
Table 3-6. TYPICAL THERMAL AFTERBURNER OPERATING TEMPERATURES (°F)
Industry
Asphalt blowing
Biological control, fermentation
Carpet laminating
Coffee roasting
Coil coating, sheet coating, metal decorating
Core ovens, foundry
Coating, engraving
Cloth carbonization
Deep fat fryers
Gum label drying oven
Mineral wool, f ibreglass curing
Odor control (general) sludge off -gas
Hardborad tempering
Oil and grease smoke (metal chip recovery,
heat quench baths, tempering)
Paint Bake ovens
Paper manufacture - sulfite digester off-gas
Pipe wrapping
Rendering plants
Rubber products
Petroleum refining and products
Printing, lithographing
Smelting, refining, metal recovery, wire burnoff
Smokehouse operation
Solvent control
Varnish cookers, resin kettles
Vinyl plastisol curing
Wood milling
Wire enameling
Phthalic anhdyride
Textile drying oven
IAAPCD
Recommendation3 '
1200- 1400
-
i4oo
1800
1200
1300-1500
1200-1400
1200-1500
1400
1200
1200
1300-1500
1200
1200-1400
-
Survey Data
1000-2000
1100-1250
1200-1500
1200-1500
2000
1000^)- 1450
\
1000°'
1100-1250
1200
900b'-l600
1100-1500
1200-1300
1300-1400
1300-2000
1300-1500
1300-1650
800b>-1200
1000b)-1500
1200
1250-1440
1350
Literature
1050
1300
1250
1310
1300-1425
1200
1240
1350
1200
1300
1200-1400
9005'
1200
1300-3350,1400
\
aj Reference 3-1, revised version, Table 51.
b) Low temperature generally for odor, smoke control, not true fume destruction.
-------
55
Chapter U. CATALYTIC AFTERBURNERS; DESCRIPTION AND PERFORMANCE
Int roduct i on
A catalytic afterburner is an alternative to a thermal incinerator as a
means for oxidizing gaseous hydrocarbons (or oxygenated hydrocarbons) to carbon
dioxide and water. Contact of a waste stream with a catalyst bed (or catalyst
matrix structure) allows oxidation reactions to occur rapidly in the temperature
range 700 - 900°F, in contrast to the 1300 - 1500°F required for practical
oxidation rates in thermal afterburners. The catalyst conforms to definition by
accelerating the oxidation reactions without itself being changed in any way.
The oxidation reactions which occur at the surface of the heterogeneous catalyst
produce the same products : carbon dioxide and water and liberate the same heat
of combustion as does thermal oxidation.
U.I.I Basic Elements of System
The basic elements of a catalytic afterburner system are shown
schematically in Figure U-l. Heat necessary to bring the waste stream up to
the required (catalytic) oxidation temperature is normally supplied by a preheat
burner. A combust ion /mixing chamber downstream of this burner is designed to
achieve a uniformly preheated, uniformly distributed mixture of combustion
products and waste stream, which is then passed through the catalyst bed. The
catalyst bed (or matrix) in commercial units is typically a metal mesh-mat ,
ceramic honeycomb, or other ceramic matrix structure with a surface deposit
or coating of finely divided platinum or other platinum family metals. The
finely-divided metal deposit is the actual catalyst for the oxidation reactions ,
while the matrix serves to support the catalyst on a high geometric surface
area and promote good contact between waste stream and catalyst. A surprisingly
small volume of catalyst bed is required, typically 0.5-2 ft3 catalyst/M SCFM
waste stream for 85 - 95$ conversion of hydrocarbons in the waste stream. The
small volume and the low density of the catalyst contribute to overall small
size and light weight for catalytic units. Recovery of heat from the cleaned
effluent stream may be included in the system using any of the techniques used
for this purpose with thermal afterburners (e.g., recuperative exchange with
the cold waste stream, or recycle of part of the hot, cleaned stream to a drying
oven). However, because required preheat temperatures are lower and hence fuel
consumption is lower, there is often less economic incentive for inclusion of
heat recovery in a catalytic afterburner than in a thermal afterburner system.
U.I. 2 Economics Relative to Thermal Afterburners
The comparison of equipment costs for catalytic and thermal afterburners
depends somewhat on the capacity of the unit and type of system. For small
capacity (up to 10 - 12 M SCFM), pre-engineered catalytic units, the purchase
cost is substantially equal to that of a thermal afterburner of equivalent
capacity (see Chapter 11). For larger, custom designed units, purchase costs for
S-1M21
-------
56
catalytic afterburners are somewhat higher than for thermal units. The inclusion
of a single pass, recuperative heat exchanger in both systems again gives approxi-
mately equal purchase costs for catalytic and thermal systems since the heat
exchanger for the catalytic system handles lower temperatures and may be constructed
using less expensive materials. Because catalytic systems are frequently smaller
and lighter than corresponding thermal systems, the cost and convenience of
installation usually favors catalytic systems.
The fuel consumption for preheating in a catalytic system is typically
hO - 60% of that required by a corresponding thermal afterburner. This economy
in fuel consumption is partially offset by increased operating costs for maintenance
(periodic cleaning of catalyst), and for periodic replacement of the catalyst
element in the system. Although the catalyst is unchanged by the oxidation
reactions which it promotes, it is subject to more or less gradual decrease in
effectiveness due to three factors, l) Even "particulate free" waste streams carry
low concentrations of particulates (usually inorganic materials) which gradually
build up a deposit on the surface of the catalyst, restricting contact between
waste stream and active catalyst, reducing performance of the unit. Catalyst
manufacturers recommend a number of physical and chemical cleaning methods for
removal of such surface deposits. With such periodic cleaning at 3 - 12-month
intervals, it is normally possible to maintain effective catalyst activity at a
satisfactory level. 2) With continued exposure to operating conditions, an
irreversible decline in catalyst activity takes place due to microstructural
changes in the active material in the catalyst matrix. Under normally recommended
conditions, catalyst activity loss due to this factor would require replacement
(or reactivation by the manufacturer) of the catalyst element at 3 - 5-year
intervals. Continued exposure of the catalyst to higher than normal temperature
accelerates this aging-deactivation process and shortens the effective life of
the catalyst element. 3) The presence of specific catalyst poisons (e.g., zinc,
phosphates—see discussion in section U.2.U.3) in the waste stream being treated
will also shorten the effective life of the catalyst. If the combination of
factors contributing to catalyst deactivation allow a catalyst life (between
replacements or reactivation by the manufacture) of one year or longer, a
catalytic afterburner will normally be more economical than a thermal afterburner.
Manufacturers of catalytic afterburner systems (who all also manufacture thermal
afterburner systems) will not generally recommend use of a catalytic system for
an application where catalyst life of less than one year is anticipated.
A summary of comparative costs associated with thermal and catalytic
afterburners in a typical installation is presented in Table U-l.
Figure h-I follows
Table U-l follows
-------
Preheat
Burner
Catalyst
Element
Fume Stream
70-400TF
^r 600-900°F
800-1100° F
Clean Gas
to Stack
Combustion/Mixing
Chamber
Optional Heat
Recovery
(Regenerative or
Recycle System)
Figure 4-1. SCHEMATIC OF CATALYTIC AFTERBURNER SYSTEM
S-14121
67784
-------
Table 4-1.
Comparison ol Thermal and Catalytic Incinerator*
Assume:
(1) Process gas volume Is 6000 SCFM at 300°F.
(2) Standard solvents for the metal decorating Industry:
toluene, MEK, xylene, etc.
(3) Design for 90 per cent conversion of hydrocarbon.
(4) Fuel cost Is $0.75 per MMBTU.
Item
Inlet volume
Solvent at 25 per cent LEL
Operating Temp. F
Thermal
6000 SCFM
8600 ACFM
38gph
^350 -
Catalytic
6000 SCFM
8600 ACFM
38gprt
';•- BOO ..
Preheat at zero solvent' . -.7.0 ' 4,2
MMBTU/hr
Preheat at full solvent
MMBTU/hr
Inside cross sectional
area, sq ft
Refractory thickness, In.
Approximate weight, Ibs
Cost of capital equipment
Fuel saving
MMBTU/hr
$/hr
S/5000 hrs
Catalyst replacement
Payout of catalyst bed
Installation ratio
Shipping cost ratio
Maintenance factor
3.0
11.1
9"
18,600 ..
$18.400
1/1
1
refractory
stack
burner
zero
6.0
5"
12.000
$19.000
3.1
2.33
11,850.
$7750
. 3300 hours
.75/1 .
.67/1
refractory
catalyst
burner •
Table 1 Comparison of temperatures required to convert
combustibles to CO2 and H2O.
Combustible
Benzene
Toluene
Xylene
Ethanol
MIBK
MED
Methane
Carbon
Monoxide
Hydrogen
Propane
Ignition
Tharnul
1076
1026
925
738
858
960
1170
1128
1065
898
Timp. °F
uuiytic
575
575
575
575
660
660
932
500
250
500
DIKtfUF* °F
501
451
350
163
186
300
238
628
815
398
Citirist/
TtMrmal*
53.4
56.0
62.2
77.9
76.9
68.8
79.9
44.3
23.5
55.7
WITH PERMISSION FROM: POLLUTION ENGINEERING
S-14121
67784
-------
57
Limitations
The oxidation performance (conversion of hydrocarbons) of a catalytic
system depends on a complex interaction of several factors: catalyst (type,
amount, and activity), operating temperature, combustible contaminant type and
concentration, and flow velocity of waste stream through the catalyst bed. The
amount of catalyst required in a catalytic afterburner depends strongly on the
fractional reduction of pollutant concentration which the afterburner must accom-
plish, rather than on the actual outlet concentration of pollutant. This dependence
is such that conversions up to 90 - 95$ can be attained with reasonable catalyst
volumes. However, the catalyst volume required for very high conversion (e.g.,
>9&%) generally makes catalytic afterburning uneconomic when such performance is
required. This limitation can be important in specialized odor control appli-
cations where concentrations of strongly odiferous species (e.g., aldehydes or
acrylates) must be reduced by factors of 100 or more.
For carbon monoxide and most gaseous hydrocarbons encountered in waste
streams, the available catalysts are highly effective in promoting oxidation at
temperatures of 600 - 900°F. For methane, however, the catalysts are much less
active, and temperatures 300 - UOO°F higher are required for substantial conver-
sion. In applications where the waste stream contains methane, and control
regulations require significant reduction in methane content to meet overall
performance requirements, a catalytic afterburner might have to be operated with
an outlet temperature of 1100 - 1200°F to meet these requirements. Under these
conditions, the fuel economy advantage of catalytic over thermal afterburners is
seriously decreased.
U.I.3.1 Particulate Containing Streams
Oxidation catalysts used in catalytic afterburners are relatively
ineffective in treating waste streams containing combustible particulates,
unless the particulate material can be totally vaporized in the preheat/mixing
region before reaching the catalyst bed. Catalytic afterburning is not
recommended for streams containing significant loadings of particulates which
cannot be vaporized (either organic or inorganic). The particulates tend to deposit
in the catalyst matrix, plugging flow channels, and forming a surface coating on
catalytic surfaces which inhibits contact between the waste stream and the
catalyst. Catalytic systems have been used successfully for controlling parti-
culates in waste streams from phthalic anhydride plants. In this application,
the combustible material (principally phthalic and maleic anhydrides) can be
adequately vaporized at reasonable temperatures and residence times in the
preheat/mixing section of the afterburner. Although catalytic afterburners
performed adequately in this application, most phthalic anhydride manufacturers
now use thermal afterburners with heat recovery. Displacement of catalytic
systems from this application has been principally on economic grounds. In
controlling emissions from asphalt oxidizers, where part of the waste stream
hydrocarbon content is an aerosol of high boiling unsaturated oils, catalytic
afterburners have been unsuccessful. A part of this failure is due to inability
to totally vaporize this oil mist at reasonable temperature and residence time in
the preheat/mixing section. The unvaporized oil droplets resulted in a coating
of tar or coke forming on the catalyst surface. A contributing factor to failure
of catalytic units in this application, however, has been very high and
S-11H21
-------
58
poorly-controlled hydrocarbon contents in the waste stream entering the after-
burner, resulting in repetitive exposure of the catalyst to temperatures of lUOO°F
or higher. Under these conditions catalysts rapidly lost activity. This same
factor has also resulted in materials failure in thermal afterburners in this
application.
U.I.3.2 Control of Waste Generating Process
It has been indicated above that the rate of decline in catalyst
activity and hence the frequency of catalyst replacement, are strongly affected
by extreme temperature excursions due to fluctuations in heating value of the
waste stream and by particulates and specific catalyst poisons in the waste
stream. These factors impose constraints on the waste stream, which may require
some accomodation of the process generating the waste stream in the way of control,
surveillance or even modification if catalytic afterburning is to be used success-
fully. Commonly described examples of such process accommodations are: l) galvan-
ized metal should not be used in process ovens or ductwork between process and
afterburner since zinc is a catalyst poison (stainless, aluminized steel or even
carbon steel are usually satisfactory); 2) mercury in glass thermometers should
not be used in locations where breakage would contaminate the waste stream with
mercury vapor; 3) phosphorus from phosphate metal-cleaning detergents and other
sources should be excluded from the process; U) lubricants for oven conveying
equipment in solvent evaporation, paint or core baking ovens should be selected
so as not to contain silicones or inorganic metal compounds which would contribute
to catalyst poisoning on evaporation or degradation. If such constraints are
inherently met by the waste-generating process or can be met economically and
with a tolerable level of change in operating and maintenance procedures for the
process, then catalytic afterburning should be considered on an economic basis as
a technically viable control technique.
U.I.3.3 Performance Monitoring
An additional limitation has been frequently raised by the staff of the
Los Angeles County Air Pollution Control District (LACAPCD) in explanation of their
reluctance to approve installations of catalytic afterburners. This concerns the
difficulty of predicting or monitoring performance of an installed unit. With ther-
mal afterburners it is assumed that performance is strictly determined by operating
conditions (waste stream flow rate and composition and combustion chamber tempera-
ture). Thus, if satisfactory performance has once been demonstrated under a given
set of conditions, reproducing these operating conditions will assure such perform-
ance. With catalytic units, performance depends not only on operating conditions,
but also on the current activity of the catalyst. The air pollution control
enforcement agency, and presumably also the afterburner user, would like a continu-
ous or at least frequent method of determining whether performance of a catalytic
system is satisfactory. The method should not be more complex or difficult than
the measurement of the operating temperature in a thermal unit. For afterburner
systems operated to meet odor or nuisance regulations, subjective evaluation of the
stack effluent is usually satisfactory. Difficulties arise primarily in applications
where performance must meet solvent control regulations such as the LACAPCD Rule 66.
If the hydrocarbon concentration in the waste stream is relatively const^-.t, and
sufficiently high (>2% of the LEL), the temperature rise of the waste stream be-
tween inlet and outlet of the catalyst bed can be used as a measure of the amount
of oxidation taking place in the bed. At 2% of the LEL, total conversion would
correspond to a temperature rise of 50 -60°F. However, in a catalytic afterburner
S-1H21
-------
59
system, not all of the oxidation occurs in the catalyst bed. As indicated in
Figure b-2, a significant part of the hydrocarbon in the waste stream (10% to perhaps
Uo or 50$) is oxidized in passing through the preheat/mixing section. The fraction
consumed there is dependent on the preheat temperature and on the chemical nature
and thermal oxidation characteristics of the hydrocarbons in the particular waste
stream, the flow rate of waste stream, and on the design of the preheat burner
and mixing section. Thus, since only a part of the hydrocarbon reaches the catalyst
bed still unoxidized, the temperature rise across the catalyst bed can at most
correspond to oxidation of this remaining fraction, and will indicate a conversion
less than that attained by the overall system. The overall fractional conversion
of the system, XQ, may be represented as:
Xo - Xp * fc (1 - Xp> (U-1}
in terms of Xp, the fractional conversion in the preheat/mixing section, and fc,
the fraction of hydrocarbon reaching the catalyst which is converted in the
catalyst bed. The final term is directly related to the observed temperature rise
across the catalyst:
(U-2)
where: C_ is the average specific heat of the waste stream flowing through the
catalyst bed (Btu/lb-°F).
AT, the observed temperature rise (°F).
Q, the heat of combustion of the hydrocarbon in the waste stream (Btu/lb).
Co, the initial concentration of hydrocarbon in the waste stream
(ib HC/lb waste stream).
If a relation, such as the lower curve in Figure U-2, giving the fractional
conversion in the preheat section, can be supplied by the system manufacturer on
the basis of prior experience with similar installations and waste streams; or de-
rived by the user from results of a limited program of analytical tests, then the
performance factors XQ and fc can be obtained from measurements of AT across the
catalyst bed. Both factors are of importance in determining whether the system is
meeting performance requirements, and if not, whether performance can best be
improved by raising preheat temperature or by cleaning or replacing catalyst.
The relationship between performance and temperature rise across the
catalyst bed is further complicated by the fact that oxidation reactions within
the two parts of the system may not be complete (to CO2 and water as products),
but may instead yield significant amounts of carbon monoxide and other partial
oxidation products. The heat liberated by the oxidation reaction is then not
proportional to the extent of removal of the original hydrocarbon, but depends on
the products formed from it as well. The oxidation reactions which take place
S-
-------
within the catalyst section under normal operating conditions are invariably
complete (to C02 and 1^0 as products) provided the waste stream has sufficient
oxygen content. In the preheat section, the oxidation of hydrocarbon takes place
in a temperature range (600 - 1100°F) , where thermal processes are known to lead to
partial oxidation products. (This is precisely the temperature range used in stack
burners for plume abatement where solution of the plume problem frequently creates
a worse odor problem due to formation of partial oxidation products.) These often
include strongly odorous aldehydes as well as CO. The catalyst then must complete
the oxidation of these intermediates as well as oxidize the remainder of the
original hydrocarbon. Thus, though at first glance, the curves in Figure U-2
indicate that the preheat section is contributing significantly to overall system
performance in the 300 - 1+00°C temperature range, in fact the increasing amount of
partial oxidation occurring there may be creating a more difficult problem for the
catalyst. If the concentration of aldehydes reaching the catalyst is so high as
to require 99+55 conversion (for odor abatement), a catalyst bed designed for 90 - 95%
conversion will not be able to do the job. The most satisfactory solution to this
problem is to reduce the partial oxidation in the preheater, not to attempt to get
better performance from the catalyst. Since the amount of partial oxidation
occurring in the preheater depends on temperature (as well as residence time and
burner design), that section should be designed for and operated at the minimum
temperature compatible with satisfactory catalyst performance.
From the standpoint of developing a performance monitoring system, what
is needed is a curve similar to that shown for the preheater in Figure U-2 indi-
cating the fraction of the total heat of combustion of the hydrocarbon in the
waste stream which is liberated in the preheat section (rather than the fractional
destruction of the original hydrocarbon). Development of such a curve would have
to be based on experience either by the system user or the manufacturer. With such
a curve, the equations presented above could be used to characterize performance if
X0 and Xp were interpreted as fraction of total available heat of combustion rather
than fractional conversion of the original hydrocarbon in the waste stream. Such
a performance monitoring system appears complicated to initiate. Once set up it
should be fairly simple to use. However, such a system is valuable only when
hydrocarbon content of the waste stream (Co) is known and can be held relatively
constant. A much more generally applicable monitoring system would require
development of an inexpensive analytical instrument which could be used either in-
line or intermittently for determination of total hydrocarbons and CO in the waste
stream and in the afterburner effluent. Such an instrument would be of particular
value in overcoming problems of performance monitoring for catalytic systems, but
would also find application in thermal afterburners used under a wide range of
waste stream concentrations and compositions.
S-1^121 Figure 4-2 follows
-------
100
80
J60
c
«
§ 40
20
Commercial Operation; Methyl Ethyl Ketone
Catalyst: Pt/Alumina on J/8* Cell "Torvex*
Total Combustion
Preneater Section
Combustion Alone
I
I
100 200 300
Catalyst Bed Inlet Temperature, °C
Figure 4-2. CONTRIBUTION OF PREHEAT SECTION TO OVERALL
CONVERSION IN A CATALYTIC AFTERBURNER
(From Fig. 4-8)
400
S-14121
67784
-------
61
Applications
Manufacturers' brochures and trade journal articles have frequently
quoted extensive lists of applications for which catalytic afterburners are
recommended. For some broad categories of applications, considerable development
vork has been done by manufacturers and users of catalytic systems. The probability
of success in using catalytic afterburning can be reasonably predicted within these
categories—the problems and limitations inherent in the process generating the
waste stream are reasonably veil understood. For nev applications to totally
different processes, it is considerably more difficult to predict whether operation
will be successful. Most equipment manufacturers maintain at least laboratory
facilities and several have portable pilot units for on-site testing in unusual or
new applications. Such tests are useful particularly in determining temperature
and catalyst volume required with fresh catalyst, but seldom can a sufficiently
extensive program be carried out to determine whether catalyst life of more than
a few months can be expected. The demonstration of the capability of catalytic
afterburning in a new application usually requires the installation of a full-size
(though perhaps small) unit in the field. Such an installation requires assumption
of considerable risk by the user (which may be shared by the equipment manufacturer).
However, if the unit proves successful in the application, its operating costs can
be considerably lower than a corresponding thermal unit.
Until a few years ago, most applications of catalytic afterburners were
to streams with relatively high hydrocarbon contents. Metal decorating (lithog-
raphy and coating ovens) and paint drying ovens have been subjects of extensive
field trials of catalytic afterburners. Applications have often been successful,
(though less often when destruction of methane by the afterburner is required) .where
equipment manufacturers are knowledgeable concerning the process factors which
have in some instances lead to short catalyst lives in these applications. In
wire enameling ovens, also considerable field experience has been acquired. Here,
due to high temperature operation of the enameling oven, and the volatilization of
metal contaminants in the process, very short catalyst lives have been experienced
with noble metal catalysts. Applications work is continuing in this area with
emphasis on possible use of catalysts less subject to deactivation by phosphorus and
copper poisoning. Catalytic afterburners have been used fairly extensively and
successfully in controlling emissions from varnish cookers, and to some extent in
foundry core baking. It is generally agreed by both equipment manufacturers and
users that fumes from asphalt blowing operations (and probably also from asphalt
impregnation processes) are not an appropriate application for catalytic units.
More recently, engineering and development work by equipment manufacturers
has centered on new applications involving relatively low concentrations of hydro-
carbons in the waste stream to be afterburned. In these cases, the afterburner is
usually installed to control odor or other nuisance problems. Several successful
installations have been made on the effluent air streams from dryers following
web-offset lithograph presses in the printing industry. Effluents from various
food processing operations have also been sucessfully treated by catalytic after-
burning. However, the particulates associated with effluents from coffee-roasting
and some other food drying operations would be expected to make these unsuitable for
catalytic afterburning.
S-
-------
62
Some large installations in the chemical industry have been quite
successful, technically at least. Catalytic systems are used as emission controls
on off-gas from ethylene oxide plants and have been used on off-gas from phthalic
anhydride plants. Probable success in applications to waste streams from chemical
plants is usually somewhat easier to assess than in other applications. Such
streams are well characterized in terms of chemical composition, and the presence
of contaminants which might decrease catalyst life is usually known in advance of
deciding to install the unit. The extensive use of catalytic systems for abatement
of nitrogen oxides in tail gas from nitric acid plants is outside the scope jf
this study. However, such systems are similar in design and use similar catalysts,
and are designed and supplied by the same manufacturers as are oxidative catalytic
afterburners. In some of these chemical process applications, the waste stream
is available under pressure. In such cases, the system must be designed to
accommodate this condition. In the equipment descriptions which follow, the
more normal application in which the waste stream is essentially at ambient
(atmospheric) pressure is assumed.
The preceding discussion of applications is not intended to be
exhaustive, but rather to point out general areas of applications where consider-
able experience has been acquired. In considering installation of a catalytic
afterburner, the customer should discuss in detail with equipment manufacturers,
their experiences with systems in applications most similar to his own.
S-14121
-------
63
U.2 Performance Characteristics
In the preceding section, the contribution of oxidation in the preheat/
mixing section to the overall performance of a catalytic afterburner was discussed
in some detail. Recognizing that some significant fraction of the total hydro-
carbon in the waste stream may have been oxidized in that upstream section, here
the discussion will be restricted to the factors determining the extent of oxida-
tion in passing through the catalyst bed itself. The performance characteristics
of catalysts from different manufacturers are different. Some attempt will be
made to point out major differences both in this discussion and in the section on
equipment description. However, the emphasis in the present discussion is on
characteristics common to all of the catalysts available for afterburners.
Performance requirements for afterburners may be stated either: l) in
terms of a maximum outlet concentration of contaminant in the cleaned waste stream;
or 2) in terms of fractional conversion of the contaminant from the inlet waste
stream. Nuisance and odor restrictions are most easily stated in terms of maximum
outlet concentrations, while solvent control regulations are usually stated as
required fractional conversion. The two methods of stating performance are simply
related provided the inlet contaminant concentration is known. In dealing with
performance of catalytic systems, the fractional conversion expression is more nat-
urally related to the design and operating variables of the system. In most
applications, requirements demand conversions in the catalyst bed of at least 70 -
8o£ (to give overall system conversion of 85% or higher). Fractional conversions
higher than about 95$ in a catalytic bed require uneconomically large volumes of
catalyst. Thus, the range of conversions of interest is perhaps 1Q% to 95$.
U.2.1 Rate of Oxidation
For a-i.i of the catalysts in commercial use, the rate of oxidation of a
combustible component of the waste stream at any point within the catalyst bed
is well described as being first order in (proportional to) the concentration of
the component in the waste stream at that point. As a consequence, the overall
conversion of that component can be expressed as :
- in (1 - ,)
gas
where: In natural logarithm
fc fractional conversion of component
k ff average effective first order rate constant for oxidation of
the component (sec"1)
v volume of catalyst in bed (ft3)
V volumetric flow rate of waste stream (actual ft3/sec.)
gas
The rate of oxidation and conversion are essentially independent of the oxygen
concentration in the waste stream, provided this stream after preheating contains
at least 2%v oxygen in excess of that required for complete oxidation of all
S-14121
-------
combustibles present. The conversion can then be thought of as determined by the
product of two factors: l] the amount of catalyst per unit volume of gas being
treated per unit time; and 2) an effective rate constant for the oxidation process.
The amount of catalyst is selected in the design of the unit, and hence is not an
operating variable. However, its influence on performance is important in deter-
mining the operating capability of the system. Figure U-3 shows the effect of
the (relative) amount of catalyst used on conversion performance, when the compari-
son is made for constant effective rate constant (constant operating conditions
and catalyst activity). The figure shows that twice the amount of catalyst is
required for 90% conversion as for 66%. A further increase of catalyst amount by
50% will then raise the conversion over 95%•
The effective rate constant for oxidation is directly controlled by
operating conditions. Figure U-U shows the typical effect of temperature on
conversion through its effect on the rate constant for oxidation. The overall
oxidation process is comprised for a complex sequence of steps. However, for
examining the effect of operating variables (and indeed even for correlation and
prediction of performance), these may be lumped into two process which appear to
dominate the determination of the oxidation rate. The first process is a chemical
reaction taking place heterogeneously at the surface of the catalyst element, and
having the usual strong dependence on temperature characteristic of chemical
reaction rates. It is recognized that, at least with the Pt/AlaOa catalysts, the
reaction involved occurs at Pt crystallite sites distributed within the porous
alumina coating on the matrix. Thus, its rate involves diffusion of reactants and
products through this porous layer as well as the actual chemical reaction rate.
However, in the normal range of operating condition, it appears possible and
reasonable to treat this complex process as a chemical reaction rate, proportional
to the geometric surface area of the matrix, and having a constant activation
energy (temperature dependence coefficient for the chemical rate constant). The
second process is mass transfer of the species to be oxidized (hydrocarbon) to
the surface of the catalyst from the bulk gas stream; and the same process of
mass transfer of oxidation products (COs, HaO) from the surface back into the bulk
gas stream. Typical of such mass transfer processes, the rate of this process is
relatively slightly affected by temperature. However, with at least some of the
catalyst configurations used it is affected by other design and operating condi-
tions, in particular the velocity of gas flow through the matrix. These two
processes operate in series. At low temperature, where the chemical rate process
is far slower than mass transfer, the overall rate is determined by the chemical
process, as indicated by the region A-B in Figure U-U. At high temperatures in
region C-D, the chemical rate has become much higher than the mass transfer rate
and the overall process becomes limited or controlled by the mass transfer rate.
In the intermediate region, B-D, both processes contribute to determination of the
overall rate. Figure k-5 shows the rate constants for these two contributing
processes, and their combination into an overall effective rate constant for the
oxidation process. In Figures U-U and U-5 the numerical values of conversion,
rate constants and temperature were selected for a catalyst supported on a
ceramic honeycomb matrix under typical operating and design conditions. For other
catalyst types and design conditions the same qualitative behavior with operating
temperatures will be observed. However, the temperature region and the conversion
level at which transition between chemical and mass transfer control takes place
S-14121 Figures k-3, k-k, and k-5 follow
-------
0.5 1.0 1.5 2.0
Volume of Catalyst/Volumetric Flow Rate of Waste Stream ( Relative)
Figure 4-3.
2.5
-------
VI —
100
80
60
o
0)
o
40
20
300
400
500
Mass
Transfer Controlled
800
900
600 700
Temperature, °F
Figure 4-4. EFFECT OF TEMPERATURE ON CONVERSION OVER AFTERBURNER CATALYST
1000
-------
ro
c
ID
M
o
u
£
ID
^\
X
Mass T,Wer Coefficient
600 700
Temperature, °F
1000
1200
Figure 4-5. EFFECTIVE FIRST ORDER RATE CONSTANT AND CHEMICAL AND
MASS TRANSFER RATES CONTRIBUTING TO IT
-------
65
will be shifted. The design operating point (i.e., quantity of catalyst and
operating temperature) is frequently selected to meet required performance at
or near point C on a curve similar to Figure U-U. Such an operating point makes
maximum use of preheat temperature in increasing conversion and minimizes the
amount of catalyst required. At such a point, the system performance will be
only slightly sensitive to changes in temperature. In particular, a significant
improvement in performance over the original design cannot be obtained without a
rather large increase in temperature. For the design illustrated by the figure, a
200°F increase from point C (800°F) to point D (1000°F) raises conversion from
09% only to 93%.
If the catalyst is not sufficiently active for oxidation of the particu-
lar component desired Ce.g., methane), then sufficient catalyst must be used to
attain the required conversion that the mass transfer controlled region is
shifted to an impractically high temperature. In such a case the design operating
point would be nearer to or in the region of chemical rate control and hence
performance at the design point would be more sensitive to temperature. However,
significant improvement in performance (over design) by raising temperature might
still be constrained by the need to keep catalyst and other system components
below their maximum permissible temperatures. The overall conclusion concerning
the effects of operating temperature are that it provides a moderate leverage for
affecting conversion performance, but that a major improvement (over design) in
the performance of a catalytic system could not normally be obtained by raising
temperature, but would require an increase in the amount of catalyst used in the
unit.
U.2.2 Effect of Combustible Molecular Type
The catalysts used in afterburners are effective in promoting oxidation
for a wide variety of gaseous, combustible materials. However, their specific
activity does depend on the type of molecule being oxidized. Figure U-6 shows
conversion-temperature curves for Kg, CO, CH« and solvent hydrocarbons for a
typically designed system using Pt/AlaOa supported on ceramic matrix. The catalyst
is most active for H2 oxidation, with significant oxidation below 100°F and high
conversions attainable at 300 - 500°F. CO shows significant reaction at 300°F and
high conversions at 550 to 800°F. Most of the common solvent hydrocarbons lie
within a narrow band with commercial alumina-supported, noble-metal catalysts.
These show significant activity at about UOO - U50°F, and high conversions require
750 to 1000°F. The catalyst is least active for methane with little activity
below 750°F and high conversions requiring temperatures of 1100 - 1300°F. The same
relative positions of these curves probably also apply to the all-metal, Pt or
Pt-Pd supported on nichrome ribbon catalysts. However, these catalysts show a
much broader range of activity for solvent hydrocarbon molecules. They have
comparable or perhaps even higher activity than the alumina catalysts for a
few more easily oxidized species (such as toluene and xylene), but for more
difficultly oxidized hydrocarbons (such as benzene and n-heptane) temperatures as
much as UOO - 500°F higher would be needed. With these all-metal catalysts, design
and operating conditions depend strongly on the particular composition of the
waste stream being treated. In general, their design points tend to be nearer to
the region of chemical rate control than those for the alumina-base catalysts.
S-14121
-------
66-.
Table U-2 presents a collection of ignition (catalytic) temperatures
and preheat temperatures for oxidation of hydrocarbon and other species over various
commercial catalysts. The columns labelled ignition temperature represents the
temperature at which 20 - 50$ conversion would be attained under normal design
conditions. Similarly, the preheat temperatures for 9Q% conversion refer to a
normal design ratio of catalyst volume to waste stream flow rate. However, these
preheat temperature.'usually refer to oxidation of a stream containing a significant
hydrocarbon concentration (.10$ of LEL), which results in a substantial temperature
rise (about 250°F) in passing through the catalyst. Thus, the average temperature
of the catalyst bed to attain 9Q% conversion may be 150 - 200°F higher than the
t.abulated preheat temperature. These are labelled minimum temperatures because
they refer to fresh catalysts. Presumably as catalyst activity declines in use
due to aging and poisoning preheat temperature would have to be raised.
An indication of the effect of solvent molecule type on oxidation
activity of the all-metal catalysts is given in Figure U-7. The rate constants,
presented on a relative scale in this figure, correspond to the "effective first
order rate constants" shown in Figure U-5« For toluene-xylene, the rates are
essentially mass transfer rates, but for the species more difficult to oxidize, the
rates shown are more nearly the rates of chemical oxidation.
S-1^121 Figures k-6 and k-1 follow
-------
200
400 600
Temperature, °F
800
1000
1200
Figure 4-6. TYPICAL TEMPERATURE-PERFORMANCE CURVES FOR VARIOUS
MOLECULAR SPECIES BEING OXIDIZED OVER Pt/AI,O3 CATALYSTS
-------
1.0
£
ID
oc.
o
ID
I o.i
0
Toluene-Xylene
0.01
Cellosolve Ace fate
MEK
Acetone
n-Butyl Acetate,
Benzene
n-Heptane
900 800 700
Temperature, °F
Figure 4-7. CATALYTIC OXIDATION RATES FOR SOLVENTS
OVER ALL-METAL CATALYSTS
(Ref. 4-5)
600
S-14121
67784
-------
Table 4-2. TEMPERATURES FOR CATALYTIC OXIDATION
Component
Kz
CO
Propane
n-Pentane
n- Heptane
n-Decane
n-Tetradecane
Benzene
Toluene
Xylene
CH*
Ethanol
MIBK
MEK
Mesityl Oxide
Ethyl Acetate
Dimethyl Formamide
Pyridine
Thiophene
Chlorobutane
AlgOa Based Catalysts
Ignition
Temperature
°F
(Acres, '*"1))
68
300
320
355
340
390
750
320
345
345
355
525
390
765
635
800
Minimum Preheat Temperature for 90^
Conversion with Solvent Concentration 10ff> LEL
(Acres,^-!))
68
300-390
480-570
480-570
480-570
480-570
540-930
480-570
570-660
570-660
480-570
750-840
660-750
750-840
750-840
840-930
Sowards, *-*>))
480
572
536
544
544
734
(Thomiades,1*-1^))
250
500
500
575
575
575
932
575
660
660
(Romeo & Warsh,^-10))
32
600
660
570a)
500
570
800
Nichrome Ribbon
Based Catalysts
Ignition Temp
°F
(Suter^-12))
650
590
580
500
550
440
460
470
760
a)Naphtha.
ON
-j
-------
66
U.2.3 Effect of Hydrocarbon Concentration
In the preceding discussion of operating characteristics, the catalyst
operating temperature has been characterized as a single, constant value through-
out the catalyst bed. This would be the case if the waste stream contained such
a low concentration of hydrocarbon (combustible) as to have a negligible heating
value. In many applications (e.g., metal decorating, paint baking) the waste
stream may have a significant solvent concentration, up to 25$ of the lower
explosion limit (LEL). In such a case, as the combustible contaminant is oxidized,
the heat of combustion raises the waste stream temperature. The catalyst surface
temperature normally follows closely the temperature of the gas flowing over it.
The catalyst bed then operates with a temperature gradient in the direction of
flow, increasing from the preheat temperature at the inlet end to a temperature
as much as 700°F higher at the outlet. For most hydrocarbons and other solvents,
the heat of combustion (oxidation) is about 0.5 Btu/SCF for each percent of LEL
hydrocarbon concentration. This raises the waste stream temperature about 27-5°F
for each percent of the LEL on oxidation. Thus, at 10$ LEL concentration into
the catalyst bed, the fume stream would have a heating value of 5 Btu/SCF, and
the stream temperature would rise 250°F if 90% of the hydrocarbon were oxidized
in passing through the bed.
Figure U-8 shows the effect of solvent concentration on the required
preheat temperature for a given conversion level. The lowest curve, labelled
"Q% LEL" is from the band for "solvent hydrocarbons" in Figure k-6. It represents
preheat temperature required when the entire catalyst bed is isothermal. The
upper three curves show the progressive and significant lowering of preheat
temperature for a given required conversion. This is brought about by the increase
in temperature rise through the bed with increasing hydrocarbon concentration (and
heating value) of the waste stream. Thus, the catalyst surface temperature
(which determines the chemical rate of oxidation) operates at progressively
greater temperature increments above the preheat temperature as the solvent
concentration increases. It should be noted that the ultimate limit on performance
at high temperatures is still the mass transfer of reactant to the catalyst sur-
face. This process is only slightly affected by the higher temperature resulting
from high solvent concentration in the waste stream.
When performance is viewed in terms of catalyst outlet temperature,
the temperature-performance curves for various solvent concentrations are brought
much closer together. As shown in Figure U-9, their order with solvent concentra-
tion is now reversed, with the highest solvent concentration requiring the highest
outlet temperature for a given conversion. With neglible concentration or heating
value in the stream, the entire bed operates at the outlet temperature. As the
heating value is increased, the temperature at the inlet end of the bed is de-
creased below the outlet temperature, and is characterized by a lower chemical
rate constant. To raise the average temperature and rate up to the level for required
performance then means that the outlet temperature must be increased. For concen-
trations up to 10% LEL (5 Btu/SCF), the curves spread a maximum of 60°F. Outlet
temperature in that range is a fairly good indication of performance nearly
independent of solvent concentration. At higher solvent concentrations, the
spread from the 0% T.Tgr. curve becomes quite significant (up to l65°F) and solvent
concentration must be considered in setting outlet temperature for desired
performance.
S-1U121 Figures k-8 through 4-10 follow
-------
en
Nl
100
80
60
c
o
0)
c
o
40
20
I I
Solvent Concentration
in Waste Stream
300
400
500 600 700
Preheat Temperature, °F
800
900
1000
Figure 4-8. EFFECT OF SOLVENT CONCENTRATION ON REQUIRED PREHEAT TEMPERATURE
-------
(/>
to
100
80
60
o
v
c
O
U
40
20
30C
400
Solvent Concentration
in Waste Stream
500
900
600 700 800
Catalyst Bed Outlet Temperature, °F
Figure 4-9. EFFECT OF SOLVENT CONCENTRATION ON CATALYST OUTLET TEMPERATURE
1000
-------
O t/>
NJ I
ro
80
60
^N |
^1
-------
69
U.2.U Catalyst Deactivation and Poisoning
With time and exposure to the environment within the afterburner, the
activity of the catalyst for promoting oxidation will decline. Unless compen-
sating changes in operating condition or maintenance is done, the afterburner
performance will decline along with catalyst activity. Compensation for decreased
catalyst activity may be made by: l) initial overdesign in specifying the
amount of catalyst needed to attain required performance; 2) raising preheat
temperature as chemical activity decreases; 3) cleaning or reactivation of the
catalyst; U) replacement of the catalyst with a fresh charge. The appropriate
compensating action depends upon the cause and nature of the deactivation process.
U.2.U.1 Deactivation by Aging or Thermal Processes
In the absence of more specific deactivation processes, catalyst life
is limited by thermal aging—micro-structure changes in the active metal
crystallites and the porous alumina supporting them—and by loss of active
coating from the matrix structure by erosion, attrition and vaporization. With
proper operating temperature and good control of temperature, these processes
are normally slow and satisfactory performance can be maintained for three to
five years before replacement of catalyst is necessary. With both the all-metal
and alumina base catalysts, these aging processes are accelerated by increasing
temperature. The alumina-base catalysts are somewhat more sensitive in this
respect. The three to five-year life estimate is based on operation with a maximum
temperature in the catalyst bed (normally at the outlet end) of 1100°F or less for
Pt/Als03 catalysts. If this temperature is raised to 1250 - 1300°F effective life
will drop to about one year. Above this temperature, deactivation processes
accelerate rapidly, and even short-term exposure to temperatures above lUOO -
1500°F can result in drastic (nearly total) loss of catalyst activity. The
manufacturer of the all-metal catalysts specifies 1500°F as the maximum temperature
of exposure. Although both types of catalysts are also used in NOX abatement in
tail gas from nitric acid plants at temperatures up to lUOO°F, shorter catalyst
life is acceptable in that application, and the reducing conditions in units there
(as opposed to the oxidative environment in catalytic afterburners as discussed
here) are less conducive to thermal aging processes. Regular, preventative mainten-
ance of thermocouples and the control instrument in the temperature control system
of catalytic afterburners is recommended since a failure of temperature control
could result in deactivation of the catalyst due to overheating. Similarly sufficient
control over the process generating the waste stream must be exercised that sudden
surges in combustible concentration, which might generate sufficient heat to
damage the catalyst are avoided.
The normal consequence of aging or loss of active coating is a loss of
chemical activity uniformly over the matrix surface through loss of active sites
randomly located throughout the bed. Figure U-10 compares the performance of fresh
catalyst with that for catalyst which has lost half its initial chemical activity
by random loss of active sites. The effects of activity loss are quite evident
in the low temperature region of chemical rate control. However, they decrease
at higher temperatures where the rate is controlled by gas to solid mass transfer.
If the initial design point were 90% conversion with a preheat temperature of 650°F,
the loss of half the activity of the catalyst would only decrease conversion to
S-
-------
TO
85$ at this preheat temperature. Conversion could be restored to 90% by raising
preheat 750°F. This policy of raising preheat temperature to compehsate of loss
of activity is normally followed until the added fuel cost for preheating makes
catalyst replacement a more economically attractive choice, or the outlet tempera-
ture has reached a point where catalyst activity is rapidly declining.
A second type of performance^change resulting from another pattern of
catalyst deactivatiort is also illustrated in Figure U-10. In this case, catalyst
has been totally deactivated £n only one part of the bed. This pattern cc ild
result from high temperature exposure of the inlet end due to flame impingement
from the preheat burner (poor design of the preheat/mixing section or overloading
of the preheat burner), or*high temperature at the catalyst bed outlet due to
excessive heating /value (solvent concentration) in the waste stream. The
performance pattern is equivalent to operation with less than the design amount
of catalyst. Performance is decreased not only in the low temperature region,
but also in the Lias's transfer controlled region. A total loss of activity in
only 25% of the bed would decrease conversion from 90% to 82$ at 650°F preheat and
restoring performance to 90% Conversion would require a preheat temperature in
excess of 1000°F. For this pattern of deactivation, raising the preheat tempera-
ture may be an impractical policy. In this example raising preheat above 1000°F
would raise the outlet temperature (vitB a hydrocarbon concentration of 10$ of
LEL) above 1250°F and lead to rapid deactivation of the remaining catalyst.
U.2.U.2 Coating qf ga-fralyats Surf£ce9tby..Particulates or.Coke
A second major class of processes leading to reduction in catalyst
activity is the buildup of surface 'coatings on the catalyst. These may be either
condensed (and polymerized or partially charred) organic material or layers of
inorganic particulates» -Their effect. is-, to inhibit -contact between the gas phase
and the catalyst surface (increased mass transfer resistance). The effect on
performance is similar'to the curve slfeWn for geometric area reduction in Figure k-IO
at least in the high temperature region-. Raising preheat temperature is not very
effective in compensating for this tyge;of activity decline. Periodic cleaning
of the catalyst is usually effective in restoring up to 90% of the initial catalyst
activity where surface coating Is responsible for activity loss.
A number of cleatiing-prbced^reS' are recommended by catalyst manufacturers
for removal of surface coatings. Thei* recommendations should be followed,
because some procedures are applicable,only to specific catalysts. The following
brief descriptions of cleaning-procedures are included to indicate the type of
maintenance required. The frequency e^t .which such cleaning will be necessary
varies greatly between applications and even between installations in a particular
application. However., cleaning at tntee
-------
71
Inorganic particulates may originate from pigment residues in metal
decorating, residues of silica from silicone lubricants used in drying ovens, and
corrosion products in process equipment or ducting. For inorganic materials,
removal of the catalyst from the unit for cleaning is necessary. Cleaning methods
include: l) vacuum cleaning or back-bloving through the element with clean,
compressed air; 2) washing the element with water or mild detergent solution (de-
tergents containing phosphates should not be used); 3) washing and/or soaking the
catalyst element in mild organic acid (e.g., oxalic acid solution) for dissolution
of iron oxide deposits, followed by rinsing with water. Where tenacious deposits
are not removed by such methods or where the catalyst has been exposed to high
temperatures and surface coatings have fused or alloyed with the catalytic metal
either replacement with new catalyst or reconditioning by the manufacturer vill
be necessary. Reconditioning is available only for the all-metal catalysts at a
cost considerably lower than replacement with new catalyst. However, such
reconditioning is possible only for certain types of catalyst deactivation.
U.2.U.3 Specific Poisoning of Catalyst Activity
The final class of deactivation processes is poisoning by specific
contaminants in the waste stream which chemically combine with or alloy with the
active catalyst metal. With most of the metallic poisons, the deactivation is
assumed to be alloying with the active metal. The list of such metal poisons
frequently quoted includes:
phosphorus
bismuth
arsenic
antimony
mercury
lead
zinc
tin
At sufficiently high temperatures, even copper and iron are capable of alloying
with platinum. However, at normal temperatures below 1100°F oxides of iron only
form a surface coating on the catalyst element. A recent review^"1^ lists the
first five as fast-acting poisons. Such materials, including phosphate residues from
metal cleaning detergents, should be rigorously excluded from process equipment
generating the waste stream. Even trace quantities of these fast-acting poisons
in the waste stream can lead to rapid catalyst deactivation. The last three
metals (i.e., lead, zinc and tin) are referred to in the same source as slow-
acting poisons. Experience indicates that catalysts are somewhat more tolerant
of these materials, particularly at temperatures below 1000°F. They probably
affect the catalyst by alloying and the rate of alloying increases with tempera-
ture. These materials should also be excluded, and galvanized ductwork should
not be used in the process or afterburner system.
Sulfur and halogens are also regarded as catalyst poisons. However, in
most cases, their chemical interaction with the catalyst is reversible (i.e.,
catalyst activity is restored when the halogen or sulfur-containing species is
removed from the waste stream). The high ignition and preheat temperatures
listed for thiophene and chlorobutane in Table U-2 are an indication of the
S-11*121
-------
Y2
inhibiting effect of sulfur and halogens on oxidation activity. Nevertheless, in
at least one case, a catalytic afterburner is used in odor abatement of a stream
containing butyl mercaptans as the primary odorous pollutant. Where concentrations
of sulfur and halogen containing species are significant (leading to >200 ppm SOa or
>50 ppm HC1 in the afterburner effluent) scrubbing facilities downstream of the
afterburner would be required to limit emissions of these species.
The effects of specific poisons on catalyst performance is qualitatively
similar to that described and illustrated in Figure U-10 for aging. With normally
encountered trace concentrations of poisons in the waste stream, deactivation will
be primarily uniformly distributed deactivation of chemical sites. Compensation by
increasing preheat temperature will allow performance to be sustained at a satis-
factory level until a large fraction of the chemical activity has been lost.
Incidents of short-term exp6sure of the 'catalyst to higher concentrations of
poisons may result in nearly total deactivation in a local region or throughout
the bed. In such cases» catalyst replacement is the only effective way of restoring
performance.
U.2.5 Users Reports on Catalytic Afterburner Performance
The response to the questionaires covering catalytic afterburners
was small, only 2h units being reported* A complete tabulation is given in
Chapter 11 (Table 11-2). Pertinent data have been extracted and summarized
in Table U-3. Only 30J5 of the units were reported to be satisfactory.'
The major maintenance problem was catalyst poisoning, reported
on 17 units. Other maintenance problems were negligible in comparison.
Specific data on the reduction of hydrocarbons was given by users
in only one case; 90% reduction was claimed, but the unit still did not meet
local requirements.
Table 4-3. USERS' INFORMATION ON CATALYTIC
AFTERBURNER PERFORMANCE
Overall performance
Number
Installed
2k
Number
with Heat
Exchangers
7
-. .»
Number
Not Yet
Run
1 .
«£.
Performance
No
Comment
Satisfactory , .
Number
7
% of Those
In Use
30
S- 14121
-------
73
k.2.6 NOX Formation in Catalytic Afterburners
The same factors, discussed for thermal afterburners in sections 3-I-2-7
and 3.1.2.8, will contribute to detectable, but normally low, emissions of
nitrogen oxides, NOX, in the stack effluent from a catalytic afterburner. The
emission factor for NOX is expected to be the same as for a thermal afterburner,
0.05-0.1 Ib NOX emitted/MM Btu preheat fuel value required. Because of the lower
level of preheat in catalytic afterburners, the concentration of NOx in the
effluent will be lower, typically about 15 ppm. As with thermal afterburners,
chemically combined nitrogen in the fume stream is expected to be nearly
quantitatively converted to NOX in the effluent.
S-HH21
-------
U.3 Equipment Description
U.3.1 Catalysts
The various matrix supports for afterburner catalysts have been selected
or developed to provide a compact catalyst element with: l) high geometric
surface area; 2) moderately low pressure drop; 3) structural integrity and
durability; and U) a system in which uniform distribution of the flow of waste
stream through the catalyst element can be accomplished readily. More con^ ^n-
tional catalyst pellets have been used as supports for similar platinum catalysts
for other applications (NOX abatement in nitric acid plant tail gas) and are
apparently used in some afterburner applications in Europe and Canada. Typically,
spherical pellets or short cylindrical extrudates (L/D = l) with a diameter of
about 1/8 inch, are supported within a basket as the catalyst element in these
units. To attain reasonably low pressure drops through these elements, the
flow rate through the bed must be low and the element thin (l - 2 inches) in the
direction of flow. Settling of pellets within the basket and attrition of pellets
lead to problems with bypassing of part of the waste stream and poor flow
distribution. The matrix elements in use in this country overcome these difficulties.
Commercial afterburner catalysts may be divided into two categories; the all-metal
catalysts and the Pt/AlaOa catalysts supported on ceramic honeycombs or other
matrix elements.
U.3.1.1 All-Metal Catalysts
Only one manufacturer, UOP Air Correction Division, supplies the all-
metal catalysts. These have been available and used in a wide variety of after-
burner applications since 19^9 • In 1957, over 1000 installations using these
catalysts in afterburners were reported:1"11). Although the number of installations
at present appears considerably smaller than this (based on results of the user
survey made in connection with this study), a great deal of experience in use of
these catalysts has been accumulated in the last twenty years. The catalyst
support is made from nichrome ribbon (l/l6-inch x 0.005-inch) which is crimped
and formed into a mat having a void fraction of 90 - 93$v, for the manufacturer's
standard "D" series catalysts. A second series of catalysts termed "Expanded",
having a void fraction of 0.95 - 0.97 is offered by the manufacturer for applica-
tions requiring very low pressure drop. The active Pt or Pt/Pd component of the
catalyst is applied by an electrodeposition technique to form a spongy, porous
layer about 0.001 inch in thickness on the nichrome surface. The mats are
mounted in stainless steel frames with heavy-gauge l6 mesh screens supporting
front and back faces of the modular element. Figure U-ll shows the manufacturer's
standard "D2" module, 18" x 2U" x 2-1/2", with a face area of 3 ft2 and a volume
of 0.625 ft3. These modules weight about 25 pounds and are easily handled.
Other shapes and dimensions of catalyst modules are manufactured. Most installa-
tions use the "D-2" or the similar "D-3" or "D-U" which have the same face area
but are 3-3/1*" and 5" thick, respectively. The catalyst modules are arranged in
banks as shown in Figure U-12. The catalyst element mounting framework is line-
welded into the afterburner system, and the modular units are champed to this
framework against asbestos tadpole gasketing to prevent by-passing of unreacted
waste stream.
S-11H21 Figures k-Ii and k-12 follow
-------
f
Figure 4-11. STANDARD MODULAR ELEMENT FOR ALL-METAL CATALYST
S-14121
67784
COURTESY: UOP Air Correction Div.
-------
Figure 4-12. MODULAR CATALYST ELEMENTS MOUNTED IN FRAMEWORK
IN AFTERBURNER SYSTEM
COURTESY: HOP Air Correction Div.
S-14121
67784
-------
75
Geometric characteristics of the UOP metal catalyst and typical design
conditions for its use in an afterburner are given in Table U-U, along with
similar characteristics for the other commercial afterburner catalysts. Figure
U-13 shovs pressure drop characteristics (at a typical operating temperature of
800°F) for the same group of materials. In both table and figure, 1/8-inch
spherical catalyst pellets have been included for comparison. Design methods for
calculation of required catalyst bed dimensions and pressure drop are discussed in
Section 17.
The all-metal catalysts are rugged; the nichrome support resists
oxidation and chemical attack from materials likely to be present in waste streams
at temperatures up to 1500°F (the maximum exposure temperature recommended by the
manufacturer). The manufacturer has developed methods for reconditioning and
reactivation of these catalysts, when the deactivation or poisoning process has
not damaged the support. In such cases, catalyst replacement with reconditioned
elements can be considerably less expensive than replacement with new elements (see
Chapter 11).
The metal base catalysts are inherently less active than the alumina base
catalysts (except for a few more easily oxidized hydrocarbons) because of
microstructural differences in the active metal crystallites. Consequently,
designs with these catalysts usually require greater catalyst volume and/or higher
preheat temperatures than do those using ceramic honeycomb substrates. However,
costs per unit volume are lower for the metal catalyst, and comparisons should be
made in terms of overall economics. The typical design conditions indicate low
superficial velocity (face velocity) and exceptionally low pressure drop through
the catalyst bed (typically less than 0.5 in HsO through the entire bed), rela-
tive to the other catalyst configurations. The lower superficial velocity requires
a larger cross section of the afterburner unit for a given throughput, as well as
lower velocity in the upstream preheat/mixing section. Under similar flow
conditions, the crimped-ribbon array would be expected to have a much higher
efficiency for collection of particulates from the waste stream (leading to
suppression of activity by surface coating) than would the honeycomb matrices. The
lower operating velocity used with the metal ribbon mats partially compensates for
the fact that their geometry favors particle collection.
U.3.1.2 Alumina-Base Catalysts
Alumina-base catalysts are available in three geometric types from four
different manufacturers. A fourth geometric structure has also been used
experimentally (including commercial scale tests). All are similar in having
the basic geometric shape formed from a dense, non-porous ceramic structure
(ot-aTumina, or one of the silica-aluminas, e.g., mullite or cordierite). On the
surface of this geometric matrix, a thin, wash coat (approximately 0.005 inch in
thickness) of porous a-alumina is applied. The active noble metal is deposited
or dispersed within this thin layer of porous alumina as microcrystallites.
Methods of application of the wash coat and dispersion of active metal within it
vary among manufacturers and techniques for these processes are either patented or
considered proprietary. Activities of the finished catalysts vary somewhat between
manufacturers but all exhibit similar chemical oxidation characteristics. For
S-1U121
-------
Table k_ U. PROPERTIES AND TYPICAL DESIGN PARAMETERS FOR AFTERBURNER CATALYSTS
Catalyst Type
Manuf acturer( s )
Geometric Surface Area
(ft2'ft3)
Density (lb/ft3)
Cost ($/ft3)
Metal
Ribbon
UOP
330
UO
1*00 - 500
Oxycat
Oxy catalyst
36
50
U80
1/8 inch Hex-Cell
Honeycomb
Oxycatalyst "Oxycomb"
Mat they-Bi shop "THT"
duPont
270
UO
1300 - 1800
Design Conditions for 90% Conversion of Typical Solvent
Hydrocarbon at 10% LEL
Catalyst volume
(ft3/M SCFM waste)
Bed length (inch)
Superficial gas velocity
(SCFM/ ft z)
(ft/sec at preheat temp.)
Temperature (°F)
Preheat
Outlet
AP (in HjO)
2
2.5 - 5
150 - 200
6-8
800
1050
0.3 - 0.5
5 - 9
9-18
200 - UOO
8-15
650 - 700
900 - 950
0.5 - 1.5
1 - 1.2
8-10
600 - 1000
25 - 35
650 - 700
900 - 950
2 - U
8C/inch
Corrugated
Honeycomb
Engelhard
700
UO
2200 - 2500
0.6 - 1
2-6
200 - UOO
8-15
650 - 700
900 - 950
1.5 - 3
S-1M21
Figures U-13 and 4-14 follow
-------
10 102 103
SUPERFICIAL FLOW VELOCITY OF WASTE STREAM, scfm/ft2
Figure 4-13. PRESSURE GRADIENTS FOR FLOW THROUGH CATALYST
MATRICES AND BEDS AT 800°F/ ATMOSPHERIC PRESSURE
S-14121
67784
Oxycats Torvex J/8 in *B*
UOP *D* Nichrome Ribbon Catalysts
EngeI hard-American Lava 8C/in
J/8in Spherical Catalyst Pellets
-------
'
Figure 4-14. QXYCAT
5'/2" Long x S3/,^ V.'ide x 3" Deep
CN
•*}• CO
•— rx
i rv
IS) vQ
m
u
h
tr.
3
o
u
-------
77
many materials they are significantly more active in promoting oxidation than are
the all-metal catalysts. Each manufacturer may have several formulations for noble
metal content and type, and methods for preparing catalysts with special activity
for particular waste stream components.
The ceramic matrix structure longest in use is the "Oxycat" (Oxycatalyst,
Inc.). The basic element is a grouping of ceramic rods, arranged with end plates
into a small, brick-like structure 5-1/2 inches long x 3-3/16 inches wide x 3 inches
deep (in the direction of flow) shown in Figure U-lU. The individual rods are of
neutral air foil cross section with blunt ends facing the oncoming gas flow.
These have an effective diametral dimension of about 0.15 -0.2 inch and have the
porous alumina/noble metal dispersion on their external surface. These modular
bricks are arranged into a wall or bed usually 3 to 6 bricks deep in the direction
of flow (which may be either vertical or horizontal). Channel connectors may be
used to hold adjacent rows of "Oxycats" together in this type of bed. The flow
of preheated waste stream through this array is then similar to cross flow through
a staggered tube-bundle.
Properties and normal design range for Oxycats are given in Table U-U
and estimated pressure drop characteristics in Figure U-13. Design methods are
discussed in Chapter 17. Costs are given in Chapter 11.
As shown in Table U-U, the surface area per unit volume of this
structure is much lower than for the other matrix supports. Consequently the
catalyst bed would need to be 5 - 15 times the volume of Oxycats as of the other
types of catalyst. This results in an afterburner unit much larger and heavier
when Oxycats are used. The manufacturer, Oxycatalyst, Inc., now uses "Oxycomb"
catalysts (based on 1/8-inch hex-cell honeycomb ceramic) for most new installations.
Oxycats are still manufactured for replacement in existing units and have been used
in some new installations where very low pressure drop was required.
Catalysts prepared on two types of ceramic honeycombs are commercially
available and in use.
A 1/8-inch hex-cell honeycomb, "TORVEX-B" (manufactured by E. I. duPont
Company) is used as a base for "Oxycomb" (Oxycatalyst, Inc.), "THT" (Matthey-Bishop)
and catalysts manufactured by duPont. Figure U-15 shows the structure of this
honeycomb ceramic (in several hole sizes). The honeycomb support is available in
12" x 12" x 1" thick sheets. Catalysts prepared on these sheets are used to
assemble modular catalyst elements or total catalyst beds for incorporation into
afterburners. Typical modules may be U - 13 inches (sheets) thick. Manufacturers
differ in their design of catalyst module or catalyst bed enclosures, but all are
designed to facilitate removal of the catalyst for cleaning or replacement.
Figure U-l6 shows an exploded arrangement of catalyst, catalyst module frame, and
a section of mounting framework. The module shown is 12" x 12" x 3". The photo
shows the module made up with an experimental "cross-flow" honeycomb rather than the
S-1U121
-------
78
hex-cell type. The overlayed sheets of honeycomb catalyst are first framed with
a coating of compressible, castable ceramic. This inner element then fits snugly
into an outer stainless steel frame which can be mounted in the framework in the
afterburner system.
Other manufacturers include retaining screens and structural supporting
members at front and back sides of the metal frame of the catalyst element.
Figure U-17 shows such an element where the entire catalyst charge for an afterburner
unit has been made up as a single element. The 12" x 12" modules may weighi 20 -
30 pounds each, while the total bed element shown might have a total weight of
500 pounds for a moderate size unit.
The significant design and construction features for honeycomb ceramic
catalyst modules and mountings are related to providing uniform flow of waste
stream through the catalyst and avoiding by-passing of the bed by any part of
this stream. As with the all-metal catalysts, the mounting framework should be
line welded or otherwise tightly sealed to the afterburner enclosure walls.
Catalyst elements should be tightly seated and sealed in mounting frameworks.
With normal pressure drops of 3 - 5 inches across the catalyst element, even small
openings around the catalyst could lead to significant by-passing. Mounting
frames are designed to seal around the ceramic honeycomb without placing it
under strain during thermal cycling (e.g., startup or shutdown of the afterburner).
Some cracks do develop in these catalysts in use. However, they are prevented
from opening to significant size by the element mounting frames.
A similar but geometrically and dimensionally different honeycomb
structure is used in catalysts manufactured by Engelhard Industries. The ceramic
matrix, "Thermacomb" (Anerican Lava Division of 3M Company) is of corrugated con-
struction with passageways for gas flow more nearly triangular than circular in
cross section (see Figure U-18). In current afterburner catalysts the eight
corrugation/inch size of the honeycomb material is used. Gas flow passages have
a dimension of 0.05 - 0.06 inches. This material has the highest surface to volume
ratio of any of the matrix materials used. Modules made up from this material are
shown in Figure U-19.
Properties and typical design conditions for ceramic honeycomb catalysts
are given in Table U-Us and pressure drop characteristics in Figure U-13. Design
methods are discussed in Chapter 17- Costs are given in Chapter 11.
An additional honeycomb structure, "TORVEX-C", has been used on an
experimental basis as a support for afterburner catalysts. This material, shown
in Figure U-20 has flow passages directing gas flow at ^5° to the main flow
direction. It has a significantly higher pressure drop than "TORVEX-B" honeycombs,
and its performance as a catalyst support does not differ greatly from "TORVEX-B".
Data available on performance are inadequate to evaluate it for this use at present.
It has been used in connection with the "TORVEX-B" honeycombs as an inlet
distribution layer, and a commercial trial has been made in a paint baking
application.
S-1^121 Figures ^-15 through k-20 follow
-------
n
o
c
x
H
n
Ul
~
Figure 4-15. "1ORVM B* HONEYCOMB ALUMINA CATALYST SUPPORT
79 £
-------
I
EXPLODED VIEW OF MnMFYrOMB CATALYST MODULE
ai
u
0
u
-------
•ill!!!
1
•
ill!
s
Figure 4-17. CERAMIC HONEYCOMB CATALYST ELEMENT
Single Element Contains Entire Catalyst Charge for Afterburner
I
Figure 4-18. *THERAM COA/B* HONEYCOMB CATALYST SUPPORT
S-14121
67784
COURTESY: Oxy-Gttalyst, Inc.
-------
"V I
— r»v
rv.
>o
,.
3
i
I
4-19. CATALYST MODULES MADE FROM THERMA COMB 8C/1NCH
a:
O
u
-------
Figure 4-20. "TORVEX C" CROSSFLOW CERAMIC HONEYCOMB
USED EXPERIMENTALLY AS AFTERBURNER CATALYST SUPPORT
S-14121
67784
COURTESY: duPont, Industrial and Biochemicals Dept.
-------
79
U.3.2 Catalytic Afterburner Enclosures
The enclosure for a catalytic afterburner unit provides a duct in which
the catalyst bed or matrix is mounted. Upstream of the bed, the enclosure pro-
vides means of preheating the waste stream to be treated, mixing the waste stream
with preheat burner combustion products, and distributing uniformly the flow of
preheated waste stream through the catalyst. The system also includes appropriate
instrumentation for measuring and controlling temperature of the catalyst bed
and for operation of the preheat burner. It may include a fan to overcome
pressure drop through the afterburner if the process exhaust fan does not have
sufficient head to overcome this pressure drop. Downstream of the catalyst the
system may include provision for heat recovery, or exhaust directly to a (usually
short) stack.
U.3.2.1 Burners
All of the known installed catalytic afterburner systems use natural
gas as fuel for preheat burners. Some designs using oil-fired burners have been
proposed by manufacturers for specific installations. In such units oil burners
have the usual disadvantages vis-a-vis gas burners, e.g., lower turndown ratio,
primary air required (normally from external source) increases the total flow
through the catalyst, and the single discrete oil burner contributes far less to
mixing of the fume with combustion products than does the distributed gas burner.
In addition, the sulfur in the fuel (oil) may exert a significant depressing or
poisoning effect on catalyst activity, requiring the use of more catalyst or higher
temperatures in an oil fired unit. Even No. 2 fuel oil normally has a sulfur
content of 0.1 - Q.5%v and may exceed 1.0/?w. Thus, where possible, gas-fired
preheat burners are used.
Nearly all recent designs and installations of catalytic afterburners
utilize distributed raw-gas burners (e.g., Maxon "Combustifume", Eclipse "AH",
North American "Flame Grid",Pyronics or UOP burners). With such a burner,
mixing of the waste stream and preheat-burner combustion-products is accomplished
rapidly downstream of the burner due to design features of the burner (see
Section 3.2). The use of a raw-gas burner (using the waste stream as combustion
air) also has the advantage of requiring no additional combustion air and hence a
lower total flow through the catalyst bed. (With external primary air to the
burner, the flow through the catalyst would be increased perhaps 15 - 25$, depending
on the percent excess air used). As explained in Section 3.2, such use of a raw
gas burner requires a minimum oxygen content in the waste stream of about 15 - l6j5v.
However, in most applications the waste stream either has an oxygen content greater
than this minimum, of has been air diluted (for safety reasons) to such an oxygen
content prior to reaching the afterburner.
In catalytic afterburner installations, distributed burners are usually
installed with blanking or "profiling" around the burner shape (see Section 3.2.1.1)
such that the pressure drop across the burner is 0.5 - 2.0 in HsO. While this
practice may contribute significantly to the total pressure drop through the after-
burner, it is necessary to obtain good mixing and flow distribution downstream of
the burner.
S-1M21
-------
30
U.3.2.2 Enclosure Configurations
Figure U-21 shows a configuration much described in the literature for a
catalytic afterburner enclosure. Although this design is not currently manufac-
tured, many units of this type are still successfully in use in various applications.
This configuration was developed before the availability of distributed burners,
and hence relies on design features of the enclosure to accomplish mixing between
preheat burner combustion products and waste stream. The burner used is a discrete
"torch" type with its own primary air supply. The waste gas stream is brough;
in above and upstream of the burner. Two means are provided to promote mixing.
The combined stream is passed through a fan located between burner and catalyst.
Fluid mechanical mixing within the duct was promoted by raising duct velocity
to the maximum possible and by arranging flow around a U-bend in the chamber.
However, as noted in Chapters 3 and lU neither of these techniques is very effective
at providing good mixing. The configuration requires a fan rated for high
temperature use (900°F). To reduce unit size and allow higher velocity through the
auct, than was permissible through the catalyst element, the latter was mounted at
an angle to the duct flow. In some applications, these mixing provisions were
inadequate and stratified flow through the fan occurred, or non-uniform distribu-
tion of flow through the catalyst element resulted. Thus, the enclosure or duct
design features contributing to gas mixing and flow distribution in this design
cannot be said to have been successful in all applications. In some more recent
designs, perforated metal plates have been placed between the discrete preheat
burner and the catalyst, to promote mixing. A significant pressure drop (-1 in
HgO) must be taken across such a plate in order for it to contribute effectively
to mixing.
With the development and availability of distributed raw-gas burners,
simpler configurations for catalytic afterburner enclosures have been used. In
these, mixing of waste stream and preheat-burner combustion products is accomplished
rapidly downstream of the burner due to design features of the burner (see Section
3.2). The enclosure then becomes a straight duct with a uniform cross section
leading from burner to catalyst, as shown in Figure 4-22. Figure U-23 shows an
external view of a completed installation of this type of catalytic unit.
The duct cross section may be either circular or rectangular (as shown
in the figure) and is sized to provide an appropriate velocity of the preheated
stream through the catalyst bed which is mounted normal to the flow and fills the
duct. The cross-sectional area of the enclosure will be directly proportional to
che throughput (including preheat combustion products) and inversely proportional
to the flow velocity. Typical design velocities for gas flow through the enclosure
duct are given in Table 4-4. These vary considerably with the type of catalyst matrix
being used in the unit. Thus, cross-sectional areas for enclosures for normal
designs with 1/8-inch Hex-cell honeycomb catalysts will normally be smaller (by
a factor of 2 - 5) than for designs with the other catalyst matrix types.
The length of the enclosure between preheat burner and catalyst is
normally in the range of 5 - 10 feet, independent of the enclosure cross section
(which is determined by unit throughput). This length is selected to avoid over-
heating of the upstream face of the catalyst element due to flame impingement or
S-14121 Figures 4-21 through 4-23 follow
-------
Clean, Hot
Gases
Catalyst
Elements.
Oven
Fumes
Preheater
Figure 4-21. SCHEMATIC DIAGRAM OF CATALYTIC AFTERBURNER
USING TORCH-TYPE PREHEAT BURNER WITH FLOW OF PREHEATED
WASTE STREAM THROUGH FAN TO PROMOTE MIXING
~. /'
S-14121
67784
-------
Stack
(Optional)
Davit
(Optional)
Catalyst
Access Door
Rain Enclosure
(Optional)
Removable Burner
(Gas Fired)
Blower
(Optional)
A. Vertical Mounting
Catalyst in Removable Container
Removable Burner
(Gas Fired)
Process Blower
(Optional)
B. Horizontal Mounting
Figure 4-22. IN-LINE ARRANGEMENT OF PREHEAT BURNER AND CATALYST ELEMENT
S-14121
67784
-------
IPs
1.3-
EXTERIOR INSTALLATION OF CATALYTIC
Figure 4-23.
AFTERBURNER UNIT
S-14121
67784
COURTESY: Oxy-Catalyst, Inc.
-------
81
excessive radiation from the preheat burner. The distance required to avoid these
factors varies with the particular burner being used. New designs and installations
have been satisfactory from this standpoint. However, in at least one installa-
tion where an older unit has been modernized by installation of a distributed
burner, overheating of the catalyst by location of the burner too close to it is
believed responsible for relatively short catalyst life experienced after the
modernization.
In horizontally mounted units, some extension of the enclosure downstream
of the catalyst element, before turning into a vertical stack is desirable, a
minimum distance of 2 - 3 effective duct diameters is usually necessary to avoid
maldistribution of gas flow through the catalyst due to effects of the flow around
the bend into the vertical discharge. In vertically mounted units, a short
discharge stack is usually provided to direct hot gas discharge and protect the
downstream face of the catalyst from cooling by excessive radiation.
:
Extension of the enclosure upstream of the preheat burner is essentially
inlet ducting and plenum which is subjected to relatively mild temperature condi-
tions (UOO - 600°F even when a recuperative heat exchanger is used). Except in
those older designs where the fan was placed between preheat burner and catalyst,
the fan is placed upstream of the afterburner unit and the afterburner operates
under (slight) positive pressure. The overall pressure drop through catalytic
units (including preheat burner and catalyst) is normally in the range of 2 - 5
in HgO; or when a tubular, recuperative heat exchanger is included in the system
6 - 10 in HaO.
At least two equipment manufacturers are currently offering afterburner
systems which can be used either as catalytic or thermal systems. Presumably
the system would be installed and operated initially with catalyst in place. If
operation in this manner were satisfactory and adequate catalyst activity obtained,
catalytic operation would be continued. If such operation did not prove successful,
or if local performance regulations were made more stringent, such that they could
not be met by catalytic operation, the catalyst could be removed and the tempera-
ture of operation increased to operate the unit thermally. Such a dual-purpose
design requires that the enclosure meet all the normal requirements (e.g., tempera-
ture capability and residence time) for thermal operation in addition to providing
for catalyst mounting and even distribution of flow through the catalyst matrix.
Such units would have purchase costs somewhat greater than, and installation costs
about equal to, thermal units. However, once installed they would provide the
guarantee of performance of a thermal unit, and the potentiality of the fuel
savings attainable with a catalytic unit. With use of 1/8-inch Hex-cell ceramic
base catalyst, the normal design superficial velocity through the catalyst
(25 - 35 ft/sec) is in the range satisfactory for thermal operation. Thus design
of the enclosure to meet normal thermal design requirements would only involve
making the length adequate for about 0.5 second residence time at lUOO - 1500°F
and the use of materials capable of withstanding the higher temperatures of thermal
operation. With the other catalyst supports, the design of a dual purpose
enclosure requires mounting of catalyst other than normal to the flow to obtain
sufficiently low velocity through the catalyst for normal catalytic operation.
S-1M21
-------
82
U.3.2.3 Materials of Construction
Catalytic afterburner enclosures are subject to considerably less
severe conditions than are thermal units. Consequently, somewhat less expensive
and lighter weight materials and techniques can be used in their construction.
Two general techniques of enclosure wall construction are used by manufacturers;
l) an outer steel structural shell lined with castable refractory (similar to
construction of some thermal afterburners; 2) double wall construction (similar
to the walls of paint baking or food drying ovens) in which a metal liner foims
the interior wall of the enclosure. This is backed by lightweight block or
mineral wool insulation, covered on the outside with a second metal shell.
Units of the first type normally have a shell of 3/16-inch thick
carbon steel plate, with U-l/2 - 5 inches of lightweight castable refractory
lining. These units weight 1*200 to 8500 pounds for capacities of 2500 - 12,000
SCFM waste stream. Units in this size range are shipped pre-piped and pre-wired
to minimize field installation costs. Even afterburners of this construction
up to 20,000 SCFM are shipped as single, pre-assembled units.
The second (double-wall) construction method is used by a number of
manufacturers, with considerable variation among them on metal types and thick-
nesses used. The metal inner liner is generally 1/16 inch or No. 16 gauge metal
either of stainless steel or aluminized steel. The range of temperatures
encountered in catalytic afterburners makes the use of aluminized steel marginal
^or this application (see Section 7.3.1.3). In the preheat mixing section
average temperatures are normally 600 - 950°F. However, local regions may run
higher, and temperatures downstream of the catalyst may be as high as 1100 - 1150°F
with preheat raised to obtain satisfactory performance with a partially deactivated
catalyst. Furthermore, aluminized steel would have little margin of safety in the
event of a control failure which allowed operating temperature to rise above the
normal range. Instances of failure of aluminized steel liners have been reported
in the survey of users made for this study. Insulation between walls of the
enclosure is generally 3 - U inches thickness with type and temperature capability
selected to withstand 1200 - 1500°F next to the inner liner, and to provide
sufficiently low temperatures for personnel protection at the outer wall. In
adequately designed units, the outer wall provides a substantial structure for the
enclosure, with 1/8 - 3/l6-inch carbon steel plate used. One manufacturer's
designs, have, however, specified construction with No. l6 gauge inner shell and
No. 20 gauge outer shell. This construction is not considered adequate for a
durable installation.
Internal support frameworks for catalyst modules are usually constructed
of a stainless steel or other alloy capable of withstanding temperatures of
lUOO°F. Construction of these parts with lesser temperature capabilities is not
recommended.
U.3.2.U Instrumentation
Primary operating control of a catalytic afterburner normally uses the
outlet temperature (downstream) from the catalyst element to adjust fuel to the
-------
83
preheat burner. Most manufacturers provide a thermocouple to sense this tempera-
ture and an indicating (or recording) proportioning controller to actuate the
fuel valve. A second thermocouple in this downstream location and thermocouple
upstream of the catalyst bed (which senses the temperature of the preheated
stream) are connected to a temperature indicator (or recorder) and high tempera-
ture limit instrument. This functions to prevent overheating the catalyst or
afterburner in the event of fuel controller failure or sudden surges of high
concentration of hydrocarbons in the waste stream. The temperature difference
across the catalyst bed gives an indication of the activity of the catalyst
and provides a basis for (manually) resetting the outlet (control) temperature
or for making decisions regarding cleaning or replacement of catalyst (see
Section U.I.3.3). The desirability of development of a more direct means of
measuring catalyst performance was discussed in that section. However, at present
no inexpensive instrument exists for that purpose.
Safety instrumentation corresponding to FIA, FM, or other insurance code
specified by the customer is normally included in a catalytic afterburner
installation (see Chapter 9).
Pressure taps (and in some cases draft gauges) are provided for measure-
ment of pressure drop across the catalyst bed and across the preheat burner.
Excessive pressure drop across the catalyst bed is usually caused by plugging of
the matrix element by buildup of particulate matter, and indicates the need to
remove and clean the catalyst.
Frequent preventative maintenance of electromechanical controls is
recommended for catalytic as well as thermal systems. Although thermocouples are
not subjected to as high temperatures in catalytic systems, they should nevertheless
be periodically tested and/or replaced. The exposure of the catalyst to excessive
temperature (deactivating it) in the event of thermocouple or control instrument
failure can be a costly reminder of the need for routine maintenance on these
control elements.
S-1U121
-------
Chapter 5. USE OF FIRED PROCESS HEATERS AND
UTILITY FURNACES AS AFTERBURNERS
3.1 When To Use As Afterburners
When waste fume streams occur in an operation which is part of a plant
complex containing process heaters, utility furnaces, or steam generating boilers,
it may be possible to use these units as afterburners. Most such heaters operate
with firebox temperatures in excess of 1800°F and with residence times in the fire-
box of 0.5 to 3 seconds, so that they might well serve as efficient fume incinerators.
If practical, the advantages may be:
a! Avoid purchase of new or separate capital equipment.
b) Avoid fuel and operating costs of separate afterburner.
c) Recover heating value of fuel portion of fume.
Whether such a system is practical will depend on satisfactory resolution of
the following problems:
a) It may not be satisfactory for the fume source unit to be made depen-
dent on the continuous operating ef the heater or furnace. Possibly
the fume stream mi^ht "be vented temporarily if the furnace should be
shut down. (This wiy not "be a problem if two or more furnace or
boiler units are in operation, so that the fume stream can be shifted
from one to the other if required.)
b) Fuel value or oxygen addition rates brought in with the fume should repre-
sent a small fraction of that regularly consumed in the furnace; otherwise,
rather elaborate interlocking and regulating controls may be required.
Such controls might include continuous oxygen and unburned hydrocarbon
analyzers, which can be very expensive and have questionable reliability,
in order to maintain the fuel/air ratio within safe and operable limits.
(The problem here arises from the fact that ordinarily the air supply is
made slave to the fuel supply, which is controlled by the load demand on
the furnace or boiler. The addition of the fume may add an appreciable
"secondary air" supply without being subject to the control action of the
fuel regulator. The added fuel that may be present in the fume is not a
major problem, since the temperature controller will cut back on the normal
fuel supply to compensate.)
c) Fume constituents should not be such as to substantially alter flame
patterns and flame radiation in the furnace. Sodium salts and highly
unsaturated aromatics can make a flame more luminous, which might lead to
excessive local heat transfer rates and burnouts.
d) Furnace brickwork can tolerate no fluxing elements, such as from suspended
salts of sodium, potassium, vanadium, etc., if the brickwork is held at a
temperature above the melting point of the salt materials. The molten
flux dissolves away the brickwork; it also attacks the oxide film that
protects metal against oxidation.
Preceding page blank
-------
e) Furnace structures, boiler tubes, etc., will be vulnerable to attack
from halogenated hydrocarbons and sulfur-containing fumes. Metal parts
are usually of low-carbon steel, may be expensive, and require considerable
furnace outage time to replace.
f) Plume-forming pollutants should be avoided, such as streams containing
phosphorous, or high in sodium and potassium.
g) The furnace must provide good fume destruction. By-passing would likely
be a problem with a large unbaffled furnace. Also, the temperature may
drop below 1500°F in some process heaters. The size of the furnace must
be sufficient that the addition of the fume stream, when treated as a
secondary air supply, does not cause the temperature of the mixed gases
in the combustion chamber to fall below the required fume destruction
temperature, even when the furnace is being operated under its minimum
load condition.
h) The presence of fouling materials that could deposit on the tube walls
should be considered. Fouling of these surfaces would reduce the capa-
city of the furnace or boiler by interfering with the heat transfer,
and frequent maintenance would be necessary.
A full resolution of all of these problems should be apparent before use
of this approach to incineration should be considered as a workable alternative.
Reference 5-1 states that boilers have been used as afterburners successfully to
control visible emissions from meat smokehouses and odors from rendering cookers;
they have been used in refinery processes involving cresylic acids, hydrogen
sulfide, mercaptans, sour water strippers, ammonia compounds, regeneration air from
treating plants, oil mists and vapors from process columns. It seems unlikely that
a large number of occasions will arise where the right combination of fume sources
and suitable furnaces or boilers can be found.
5.2 Manner of Introducing the Fume Stream Into the Furnace
The variety of fume streams, and the furnaces, boilers and associated
burner combinations is so large that no detailed recommendations can be made. Pre-
sented here are a number of suggestions and comments. They apply to handling fume
streams which contain less than 25$ of the LEL fuel value; streams containing more
than this amount of fuel should be treated as fuels and handled in separate
"waste heat boilers," or flared, with full regard for the special problems of
handling fuel and air mixtures that might be within the combustible or explosive
range.
5.2.1 Furnaces Using Premix or Diffusion-Mixed Burners (Gas or Oil Fuel) With
Natural Draft, or Induced Draft Blowers
Commonly used in oil refineries and chemical plants, this type furnace
offers the best expectation for incinerating a fume stream. If the combustion air
is induced through the burner, however, the substitution of a fume stream for the
air supply is possible only if it is non-corrosive, non-fouling, and if its oxygen
S-1U121
-------
87
content is very nearly the same as ordinary air. If the furnace uses a secondary
air register, separate from the burner, it should be introduced through a tee into
an extension of the register, as shown in Figure 5-1 (Reference 5-1), thereby taking
advantage of the existing arrangements for insuring that the fume will be well mixed
into the combustion gases. If there is no secondary air register, then one must in
effect be created. It should be located in the immediate vicinity of the burners;
Figure 5-2 suggests good and poor locations. The bottom entry immediately below the
burner shown is suggested for use where fouling conditions are encountered (Refer-
ence 5-1)> however, this would not be so if the furnace floor were located at
ground level, since the ductwork would not be accessible for cleaning; also it would
not serve well if condensible material is present, since a sub-floor channel would
serve as a liquid trap.
Care must be taken if fume is introduced as additional secondary air to
insure that adequate draft capacity is available to handle the increased flow of
flue gases. Fumes containing about the same amount of oxygen as fresh air can
merely displace the normal secondary air, with adjustments to the secondary air
dampers; in this case no added draft capacity is needed. However, if the fume
s"cream is deficient in air, the same amount of oxygen will be required in the
furnace (in fact, a little more, since the furnace efficiency will be reduced by
the extra gases that must be heated to the flue temperature), and the extra blower
or draft capacity must be supplied. If large amounts of inert gas (C02, N2, H20
(vapor)) are added via the fume, burner stability may be reduced.
A reformer furnace, although commonly using premix gas burners, would
be a poor candidate for incinerating fumes. This type of furnace uses a large
number of burners which are individually adjusted in order to achieve a close
control of the heating distribution to the tubes in the furnace. Possibly the
fume might be introduced into a plenum which supplied air to all of the burners,
if such a plenum could be applied, but this would interfere with access for
lighting-off and adjusting the individual burners, as is common practice.
5.2.2 Furnaces With Burners Using Forced Draft Blowers
The problem here is that most fumes are available only at or near
ambient pressure, and they would have to have their own blower or else pass through
the furnace blower in order to enter the furnace. Blower units are particularly
subject to difficulty in fouling services, and this arrangement is not recommended
if either fouling or corrosive conditions are present. If the fume is non-fouling,
an arrangement such as in Figure 5-3 may be considered, but such an arrangement is
not easily applied when the burner unit comes as an integral package that includes
its own blower.
A steam-Jet ejector may be considered as an alternate to a rotating
blower. Properly selected and applied, it is less susceptible to fouling.
In all cases involving extensive changes to the burners, the control
system, or the provisions for supplying draft, the manufacturer of the equipment
should be consulted. He may have had previous experience with a similar problem,
and be able to supply equipment modifications and recommend alternative arrangements.
S-1M21
-------
88-'
He should be made a responsible party for any changes that effect the safety or
functioning of the equipment he has supplied.
5-3 Hazards on Shutdown
"Contaminants in most exhaust gas streams are normally well below explosive
concentrations. In a few processes, however, combustible gas concentrations can
accumulate during shutdowns with resultant explosion hazards on lightoff of the
boiler. For instance, a batch of raw or partially cooked animal matter mitn't be
left overnight in a rendering cooker ducted to a boiler-incinerator. This could
generate enough methane, hydrogen sulfide, and other organics to produce an ex-
plosive mixture in the ductwork leading to the boiler. If, subsequently, the burner
were ignited without first purging the line, an explosion could occur. To avoid a
rare possibility such as this, both the boiler firebox and the ductwork should be
purged before igniting the burner.
"Some fire hazard is created by the accumulation of organic material in
ductwork. Lines such as these must usually be washed periodically. The degree of
organic accumulation can sometimes be reduced by frequent steam purging or by
heating the ductwork to prevent condensation "5-1/.
5.U Verification Tests
When a process heater or boiler has been adapted to serve also as a fume
incinerator, its optimum adjustment and satisfactory performance should be verified
by tests both before and after the addition of the fume. The furnace may have been
performing poorly before the fume addition; the fume stream will certainly not
improve matters. Even if the furnace performs well for its ordinary function it may
do poorly as an afterburner, as it is by no means easy to ensure the furnace will
give the right combination of mixing, heating, and oxidation that will bring about
the destruction of the fume. An improperly introduced stream may remain cool and
bypass the main combustion chamber process , or it may be introduced in such a way
that it partially quenches the flame before combustion is complete, forming CO and
aldehydes. (See Chapter lU for a discussion of the problems of mixing, and Chapter
10 for test procedures.)
S-Ik 121 Figures 5-1 through 5-3 follow
-------
Contaminated
Air Di
Boiler
Firebox
Variable
Louvers
Custom Made Air Register for
Multi-Jet Burner
A. Good method of introducing contaminated air to boiler firebox through a
custom made air register. There is good flame contact. Contaminated air
enters firebox through burner. Note: Type of burner is critical; contam-
inated air is portion of combustion air; not applicable where contaminated
gases are corrosive.
Contaminated
Air Duct
Boiler
Firebox
Diffuser
Variable Louvers for
Multi-Jet Burner
B. Poor method of introducing contaminated air from diffuser to boiler firebox
through the burner air register. Diffuser restricts combustion air to burner.
Moreover, louver may partially close restricting flow of contaminated air into
boiler firebox.
Reference: LAAPCD Engineering Manual
Figure 5-1. INTRODUCING FUME STREAM VIA AIR REGISTER
S-14121
67784
-------
Burner (Sealed or
(Open Type)
Boiler Firebox
Contaminated
Air Duct
Contaminated
Air Diffuser
Burner
A. Boiler firebox showing entry of contaminated air
through a diffuser in the floor near the burner.
The possibility for avoiding bypassing is good.
Note: Type of burner is not critical; contaminated
air is secondary air for boiler; applicable where
contaminated gases are corrosive.
Contaminated
Air Duct
Boiler
Firebox
Better
Entry
Point,
Boiler
Firebox
Contaminated
Air Duct
Entrance
Burner
B. Poor method showing entry of contaminated air
near boiler firebox rear firewall. Mixing will be
poor.
C. Boiler firebox showing entry of contaminated
air through a duct at front of boiler.
Reference: LAAPCD Engineering Manual
Figure 5-2. DUCT ENTRY OF FUME STREAM INTO FIREBOX
S-14121
67784
-------
Air
Vent—-
Combustion Air
to Furnace
Damper
Control
Non-Fouling
Fume
Source
Emergency Shutdown
(Furnace Flameout)
Control
Figure 5-3. SAFETY PROVISIONS FOR ADDING A FUME STREAM
TO AN EXISTING FURNACE
S-14121
67784
-------
89
Chapter 6. HEAT RECOVERY
6.1 General Considerations
This section describes methods and equipment designed to make use of the
heat energy contained in the flue gases discharged from the afterburner, which
would otherwise be wasted. A number of recovery methods are in use or have been
proposed:
a) Heat exchange may take place between the hot flue gases and the cool fume
stream, using a shell-and-tube (recuperative) or a rotary or cyclically regenerative
heat exchanger, and thus preheating the fume stream before it enters the afterburner.
b) A portion of the hot stack gases may be returned, or "recycled"
to the oven or process unit which is the source of the fumes.
c) The hot flue gas heat may be removed in some other manufacturing
process unit, either directly (by exposure to the flue gases), by a heat exchanger
to a circulating heat medium, such as oil, water, or molten salt, or to a steam
generator for process steam, space heating, or power generation.
Since heating a large fume stream to a temperature as high as 1500°F
represents a major consumption of fuel energy, and one that will continue
indefinitely so long as the afterburner is operated, the economics and practica-
bility of heat recovery should be weighed very carefully. Many factors must be
considered:
a) Is it feasible? Dirty or foul fume streams may render heat
recovery equipment inoperative in a very short time. Plant heat balance may not
allow for an effective use of recovered heat.
b) Is it safe? On fume streams containing combustible material,
excessive preheat by heat exchange with the flue gas might lead to flammable, or
even explosive conditions. Recycle of hot incinerator stack gases could intro-
duce contamination hazards in the process or oven equipment.
c) Is it dependable? If heat recovery equipment forces frequent
shutdowns of the afterburner for cleaning and this in turn shuts down the plant
unit, the economic value of any energy saving could be quickly nullified.
d) Is it economically Justified? A proper economic evaluation
should account for installed cost, operating costs (including fuel, operator and
supervision, maintenance, etc.) interest charges, etc. Some guides for making
this evaluation are given in Chapter 11. One manufacturer gives as rule of thumb
that heat recovery equipment is economically Justified only on afterburner units
of more than 10,000 SCFM capacity.
S-1U121
-------
901
e) Can the extra space and weight be accommodated? Heat exchangers
are generally bulky, and so is the ductwork. If mounted on the roof, major
structural changes might be required to carry the added weight.
f) Can deterioration of the heat recovery apparatus result in
loss of control over the pollution source? Leaking exchangers can bypass fume
material into the stack.
These questions, and others, are often complicated, and no simple rules
can be made. All the conditions noted above must be answered affirmatively, one
way or another, before a heat recovery system can be considered Justifiable. The
record of heat recovery systems is not very good; many have given trouble. The
potential user should review proposed heat recovery systems with considerable
skepticism. This is not to say he should reject the idea out of hand—rather
that he should look for proven designs and applications which show successful
use in a situation similar to his own.
A valuable aid for deciding whether or not to use waste heat recovery is
to make a balance sheet of heat requirements and potential heat sources of the
plant. Only if places can be found to use effectively the waste heat—in terms
of load cycles, heating load quantities, and temperature levels—does it make
sense to consider heat recovery schemes.
Table 6-1 summarizes the alternatives, the advantages, and limita-
tions of these methods. The text material discusses the points of the table
more fully.
6.2 Types of Recovery Equipment
6.2.1 Recuperative (Gas/Gas) Exchangers
6.2.1.1 Heat Transfer Effectiveness
The fume and the flue gas streams have about the same mass flow rates
and specific heats, so they can be made to exchange heat with each other on
nearly a degree-for-degree basis. Commonly this is done in a shell-and-tube
(recuperative) exchanger, with one stream passing through the tubes, and the other
stream passing over the outside. Usually the fume stream is put through the tubes,
although manufacturers differ in their practice.
The performance of such an exchanger is measured by its "heat
transfer effectiveness". The entering hot flue gas and the cold fume stream
give the maximum temperature difference experienced in the exchanger, and this
measures the maximum amount of heat that might be transferred. The temperature
rise of the cold stream in the exchanger can never exceed this maximum difference,
and the extent of its approach measures its heat transfer effectiveness, which
is calculated thus:
c (T — T
E _ Q _ p ' fume stream leaving fume stream entering exchanger)
Q ~ C (T T "" \
inax p flue gas entering - fume stream entering exchanger)
(C = the specific heat at constant pressure of the fume stream, Btu/(lb - °F))
Table 6-1 follows
S-1U121 Figure 6-1 follows
-------
Table 6-1. FORMS OF HEAT RECOVERY
Type
T-bdlar exchanger
Gas/gas
Regenerative
(rotary) exchanger
Flue gas recycle tc
oven
Steaa, generators.
boilers, water
heaters
Process *ieat--vla
circulating heat
transfer salt
(Hytecl. oil. air
Effectiveness
Ratio. E
1 sta-50? oox
2 sta-629( cax
3 it a- 85* cax
up to 85*
-30S
to 75*
to 75*
Additional Auxiliary
Equipment
•».
Safety controls
Eitra burners and
controls, safety
controls
Extra ducting.
blowers, controls
Exchange -• , piping ,
reservoir, pusp.
controls
Limitations, Problems
1) May be easily fouled.
frequent cleaning and
caintenance
2) Failures , differential
thereal expansion
3) Hot surf aces may crack
or polymerize fuae
components, ley eoa-
bustlble deposit,
initiate a fire.
M Bulky, heavy, added
roof load and/or floor
space .
5) Corrosion if cools be-
lev dev point of flue
gaa.
1) Easily fouled Use
only on relatively
clean streams
2) Burnout If failure on
rotary drive motor
3) Requires attention to
pressure balance to
control leakage at
seals.
t) Avoid cooling flue gas
to dev point, but
otherwise is relatively
Insensitive to corro-
sion.
5) Ignition if overheat
fuel-rich stream
l) Process must be compa-
tible vlth flue gas
(condensation? Sulfur
in fuel? CO or CC^T
Reduced oxygen? Un-
burned fuel?)
2) Usefulness depends on
tenperature and heat
requirements of fuae
generating process.
1) Ties steam generation
to fume process and
vice versa
2) Match steam heating
load to afterburning
heat release
3) Dev point and condensa
tlon on cold-vater
colls
1) Ties fume generation
to process
2) Hatches process heat
load to afterburner
heat release (but can
use supplementary
firing for added pro-
cess heatX
Commonly Used For
tesln curing ovens
>aint drying ovens
Chemical plants
0
Odor control , water
treat sent unit*.
Resin, lacquer curing
ovens, (if lev solvent
release)
Lltho ovens
Plant steam supply
CO burner and boiler
(fluid bed catalytic
cracking unit in oil
refinery Uses
supplementary fuel
firing)
Plant heating and air
conditioning units
Asphalt bloving, pre-
heat ing
CoiBonly Hot Used For
Rendering plants
Any fumes containing oils,
dusts, resins
Varnish cookers (com-
bustibles hazard) ovens
requiring human access.
Smoke ovens (low
temperature)
Bo Cases
Reported In
Survey
uestionnaire
Returned
81 thermal
7 catalytic
2
18
(not tabulated)
(not tabulated)
a) Ore unur>cv_-«r claim to have used he»t ejzhtrgeri nceeiifully for rendering planU
S-14121
67784
-------
Fume
Stream
Flue
Gas
Shell and Tube Type Exchanger
Cross-Flow Type
Temperature
Temperature
100°F
Entering
Cold
Fume
Stream
Leaving
Heated
Fume
Stream
Leaving
Cooled
Flue Gas
14008F
Entering
Hot Flue
Gas
Three Dimensional Plot of
Fume Stream Temperature
Three Dimensional Plot of
Flue Gas Temperature
Figure 6
• 1. GAS TEMPERATURE PROFILES IN A CROSS FLOW HEAT EXCHANGER
S-14121
67784
-------
91
Clearly, the greatest heat economy will be achieved if E can be made to approach 1.
However, this would be impossible without having an infinitely large exchanger.
The effectiveness is also limited by the practical arrangements needed to confine
and direct the gas flow.
6.2.1.2 Cross-Flow Exchanger
The need for a simple structure usually dictates using what is called a
"cross-flow" arrangement, as shown in Figure 6-1. Although the exchanger can
theoretically be made more compact (and heat transfer coefficients improved) by
using small diameter tubes, the need for a low pressure drop and for easy cleaning
usually dictates that all passages be an inch or so across, and tubes an inch,
inch-and-a-half, or even two inches in diameter are common. In services that are
subject to external fouling, tubes are arranged in vertical rows (not staggered)
for easy cleaning.
Sometimes tubes with a rectangular cross-section and with fluted walls are
used; the corrugations improve the heat transfer coefficient, and make a somewhat
smaller exchanger. For a given overall exchanger effectiveness, there is little
preference, one tube shape for the other, and the choice of surface is primarily
a matter of cost. Tubes are anchored in a "tube sheet", and the tube bundle is
enclosed in a metal "shell", which also confines one of the process streams.
Hence, the term "shell-and-tube" exchanger.
From the standpoint of maximum heat recovery, the ideal arrangement
for an afterburner would give counter-flow between hot and cold streams.
Physically a true counter flow arrangement is difficult to obtain in a shell-
and-tube construction (although a stacked-plate design approaches counter flow
performance and is discussed later). The simplest and cheapest is the cross flow,
and stacked or "multiple pass" cross flow units approach counterflow performance.
There are limitations to the effectiveness obtainable with a cross flow
exchanger. Figure 6-1 shows that although the fume and flue gas streams each
enter at uniform temperature, portions of the stream see different temperatures,
and the leaving streams are not uniform. On one side the fume stream temperature
may closely approach the flue gas temperature. On the other it sees an already
cooled flue gas stream, and cannot reach as high a temperature. The average of
all the streams may gain only a fraction of the maximum possible temperature rise
(the effectiveness is less than unity). An effectiveness over 65% is possible
on a single pass cross flow exchanger, as pictured on Figure 6-1, but only with
much more surface area, and at the cost of a higher pressure drop; UO to 50%
is a commonly accepted compromise.
The exchanger can be arranged for multiple passes of either or both
streams. A higher effectiveness is obtained if one of the streams is doubled
back into a second pass, as in Figure 6-2, or even a third pass, with effective-
ness values commonly running about 65% and &5%\ the cost of the exchanger is
roughly doubled or tripled to do this, and the blower pressure drop increased too.
S-1M21
-------
92<
Chapter 18 gives an approximate method for calculating the tube surface
area required for an exchanger of specified effectiveness of the cross flow type.
The size so determined should be used for preliminary evaluation only. Final
judgment and equipment selection should be based on firm proposals or bids.
Design proposals may be based on different assumptions as to fouling, pressure
drop, heat transfer coefficient, etc., and on different materials from those
assumed, and the cost quotation may differ appreciably from that estimated from
the procedure given. The basis for assumptions on which bid proposals are to be
made should be clearly defined if alternative proposals are to be Judged fairly.
6.2.1.3 Stacked Plate (Counter Flow) Exchanger
A stacked plate exchanger may be encountered (Figure 6-3). Plates
separate alternate channels for fume stream and flue gas, and a compact, near-
counter-flow arrangement can be achieved. Heat transfer coefficients may be
improved by embossing corrugations into the plates (thereby increasing turbulence
and breaking up the insulating effect of the boundary layer at the wall surface).
However, this arrangement suffers from severe local thermal stresses at the
end sections, where much complicated welding is necessary to keep flue and
fume streams separate. It is also difficult to clean. Its use should be
limited to small temperature changes on clean service.
6.2.2 Rotary Regenerative Heat Exchanger
This type exchanger is diagrammed in Figure 6-U. A heat storage element
made of corrugated sheet metal plates is mounted on a rotating cage structure and
is passed alternately through the hot and cold streams. The plates are formed or
stacked into a matrix structure, facing edgeways into the gas flow so that the
gas may flow between them. The corrugations improve the heat transfer coefficient
from gas to metal, and also serve as spacers to hold the plates apart. The matrix
structure puts a large heat transfer area into a small volume, and an exchanger
for a high heat transfer effectiveness for a large gas flow rate can be built
into a very compact unit. Corrosion, so long as it does not impair the structural
integrity or plug the flow channels, does not impair the performance.
This type of exchanger, when applied and operated correctly, can func-
tion very well. It is almost universally used for recovering heat from the
boiler stack gases in central-station power generation units. However, there
are important limitations and precautions to its use as an adjunct to the
afterburner:
a) The small passageways between the plates are easily plugged.
The fume stream must be free of condensible (and polymer!zable) vapor, heavy
smokes, or sticky particulates. Nevertheless, for some services, the passages
are self cleaning (where non-sticky, dry particulates are encountered).
b) Isolating seals must be used to control leakage between hot
and cold sections. These are subject to wear and warpage with time, and the
leakage may become excessive. Effects can be minimized by arranging blowers so
that pressure differences cause any leakage to flow into the fume stream where it
will be recycled through the afterburner; this arrangement may be difficult to
provide for some applications.
S-1^121 Figures 6-2 through 6-k follow
-------
Cold
Fume
Stream
Hot
Flue
Gas
Heated
Fume
Stream
A. 2-Pass, Crossflow Exchangers (Arranged to Place Units Counter-Flow)
Heated
Fume
Stream
Cold
Fume
Stream
B. 3-Pass, Crossflow Exchangers (Arranged to Place Units Counter-Flow)
Figure 6-2. MULTI-PASS COUNTER FLOW ARRANGEMENTS
S-14121
67784
-------
Cool
Fume
Stream
Cooled
Flue
Gas
Heated
Fume
Stream
Figure 6-3. COUNTER-FLOW PLATE-TYPE EXCHANGER
S-14121
67784
-------
Rotating
Matrix
Cool
Fume
Stream
Figure 6-4. ROTARY REGENERATIVE HEAT EXCHANGER
S-14121
67784
-------
93
c. As the matrix revolves, gas trapped within is carried from
one flow channel to the other. The effect is minimized "by blank areas in the
cover plates. Ordinarily this trapped gas would bypass the afterburner
and reduce its efficiency. As noted in (2) above, the flue gas pressure should
be the higher, and the size and position of the blank area adjusted to give
a positive sweep-leakage of flue gas into the fume gas side.
d. Since the seals must be close-fitting to avoid excessive
leakage, they are sensitive to the strains caused by thermal expansion differ-
ences between hot and cold streams. Especially for exchangers taking flue gas
directly from the afterburner (at around 1500°F), care must be exercised in
designing to minimize effects of warpage within the exchanger structure, and of
strains introduced by expansion of the attached ducting. Some early installa-
tions using the rotary regenerator have experienced mechanical difficulties on
this score. Some manufacturers restrict use of this type exchanger to a second
stage of heat recovery, with a recuperative type for the first stage; this
exposes the structure to a smaller temperature difference and results in lower
stresses and less warpage.
e. An inadvertent shutdown on the rotation of the matrix will
allow it to be heated to stack gas temperatures. This could be damaging. Also,
if any coating of combustible material has accumulated from the fumes, a fire
might be started. A burner shutdown interlock system should be provided to
protect against this type of failure, or else the rotor should be designed to
withstand the maximum temperature of the flue gas.
f. Although corrosion effects are not as immediately serious
as in a turbular exchanger, they should be guarded against. The metal of the
matrix should never be allowed to operate below the dew point temperature at any
point; preferably a 50°F margin above dew point temperature should be allowed
when corrosive materials are present.
g. Because of the high effectiveness ratios to which this exchanger
is best suited, the fume stream exit temperature becomes high, and combustible
vapor streams that are well below the combustible limits at room temperature may
ignite at the exit temperature. This is important for fume streams having a
high fuel content.
With these limitations in mind, the rotary exchanger should be considered
for installations on clean service, where high heat transfer effectiveness
ratios are Justified. Examples would be for odor control, or for bacteriological
sterilization. A manufacturer who makes both the rotary regenerator and the
shell-and-tube type exchangers reports that, as a rule-of-thumb, the rotary type
is more economical to build and install when the desired exchanger effectiveness
is over 60%.
The design of the rotary-regenerator type exchanger is complex, and
should not be undertaken by the ordinary user. References on design and
performance are:
Basic data on heat transfer and pressure drop of matrix: V. K. Migay,
Heat Transfer - Soviet Research. Volume 2 No. U July, 1970 pages 151-156.
S-1U121
-------
9k
Rotary heat exchanger: W. M. Kays, Compact Heat Exchangers. McGraw-Hill,
196U. See pages 27-30, UU-U9, 58-62, 70-71, and 259-260.
6.2.3 Cyclically-Operated Packed-Bed Exchangers
Closely related in principal to the rotary regenerative exchanger,
although very different in physical arrangement, is the use of a pair of packed-
beds, through which the cold and hot streams alternately pass in a switched
cycle. The packed bed consists of screen-supported ceramic balls or saddles
(see Perry's Handbook of Chemical Engineering for a more complete description of
the packed bed and common packing materials)7 Normally used for catalytically con-
trolled chemical reactions, the system may also be used as a heat exchanger with
high effectiveness (how high depends on the thickness of the bed and the pressure
drop available). One manufacturer offers the packed-bed exchanger for heat
recovery on afterburners, and quotes design effectiveness ratios of 75$ to 83$.
Advantages of the arrangement are:
a) Packing material is relatively inexpensive
b) Packing material is relatively inert, non-corroding
c) Exchanger bed is compact. There is a lot of exchange surface area
and thermal storage capacity (mass x specific heat) in a small volume.
Disadvantages are hypothesized, since there are an insufficient number
of installations to give a meaningful record in afterburner service;
a) Switching valves, some of which must work in a high temperature
environment, are potential trouble sources.
b) During switching, the gas trapped in the bed from one stream is
carried into the other. This would give a bypassing around the afterburner
and a reduced system efficiency for controlling emissions. The manufacturer
proposes to avoid this trouble by purging with clean air back into the fume
stream. The arrangement has not been adequately demonstrated.
c) The packing, although not corroded, may be coated and insulated
by fouling material carried in the fume stream, and the interstices blocked.
Unless a pilot test can be arranged to prove otherwise, using the proposed foul
stream, this exchanger should not be considered for use in a fouling type of
service. Under some circumstances fouled material can be cleaned from the bed
by burning.
Thermal performance of the cyclically operated packed bed exchanger is
given in Compact Heat Exchangers (op. cit.) pages 27-30, 35-37, UU-U9, 131.
Pressure drop and heat transfer data on saddle packing should be obtained from
the manufacturer, or from Perry's Handbook (op. cit.).
S-1U121
-------
95
6.2.U Recycled Flue Gas to Process Unit
This method of heat recovery has been used successfully on baking,
curing, and drying ovens. It should be considered wherever large volumes of hot
air or hot gas can be put to use in the process. Usually the oven temperature
operates well below that of the afterburner stack, so that only a fraction of the
stack gases can be recycled to the oven, being tempered on the way by mixing with
fresh air to match the required oven temperature. The oxygen in-the stack gas is
partly depleted, hence the fume stream coming from the oven will also be reduced
in oxygen content, and the burner flame stability should be verified. If a high
recycl* ratio is used (i.e., a high fraction of the flue gas is returned to the
oven), external air will probably be required by the burner to satisfy combus-
tion n«eda—with adverse effects on the fuel economy, especially when using a
raw gas burner.
Figure 6-5 shows a possible arrangement with recycling for heat
recovery where no appreciable fuel contribution comes to the afterburner from
the process unit.
Use of this recycle arrangement would not be desirable if liquid fuels
containing sulfur, or other corrosive or contaminating materials were to be
encountered in the system. Recycling produces a concentration effect on con-
taminants as well as a depletion of the oxygen, and these effects may not be
tolerable to processes going on in the oven (some lacquers may be very sensitive
to contaminants, for example).
6.2.5 Waste Heat Recovery for Process Use
Where process heat is required, and neither recycling flue gas nor heat
exchange with the fumes entering the afterburner is feasible, a number of
possibilities remain.
a) The hot flue gases may be used directly in some other process
unit. For example, they might be used as a heat source for evaporating and
concentrating a wet slurry, or for a reboiler in a distillation column.
b) Hot clean air may be obtained for the oven to which the afterburner
is attached using an air/flue gas exchanger. (Figure 6-6)
c) Process heat loads may require steam, which can be generated
from heat in the afterburner exhaust (see Section 6.2.6).
d) Process heat loads may be best satisfied through an indirect
transfer medium, such as hot oil, hot water, or a molten salt.
Whatever the process use, alternate bypass controls should be included
so as to permit operating the afterburner as a heat source even though the fume
source is not running, or to permit the afterburner to function on the fume stream
when the other process unit is down. Stacks should be designed so as to withstand
the temperature of the flue gases coming directly from the afterburner.
S-1M21
-------
96-•
The amount of heat recoverable from the flue gas can be estimated with
the aid of Figure 6-7» which is based on the assumption of normal flue gas
composition derived from natural gas fuel. Some deviation from this will come
about from different fuels and fumes containing combustible components.
6.2.6 Waste Heat Boiler
The exit gases from an afterburner are normally at a temperature of
1200°F to 1500°F, and considerable heat recovery can be obtained by passing tl.ese
gases through a steam generating unit, or waste heat boiler. Note that this
arrangement does not substitute the boiler furnace for the afterburner (such an
alternative is discussed in Chapter 5); the boiler is considered only as a
separate heat recovery device.
If desired, the waste heat boiler can also be equipped with separate fuel
burners and controls; in this way, both afterburner and steam generator may be
operated independently or in conjunction, without impairing the flexibility of
either. Such an arrangement is shown in Figure 6-8- The choice of whether
to fire the boiler, and how much firing to do, will depend upon the relative
capacity demands on afterburner and on plant steam load. Generally the added
complexity of supplementary firing adds appreciably to the cost, since the
unfired version can be a relatively simple arrangement.
The variety of boiler designs useable for waste heat recovery is large.
Some versions are of the locomotive type, with straight tubes fitted into tubesheets
(up to 250 psig, 20 M Ib/hr). Still larger boilers (perhaps >10 MM Btu, or >10 M
Ib steam/hr) will have the water contained in the tubes, with a steam drum for
disengaging. Superheating coils may be included with a boiler of this type.
Boilers of this size are available as package units, shipped complete with
burners (if used) and auxiliaries, such as blowers, firing controls, feed pumps,
etc.
The amount of steam that can be obtained from the flue gases of an
afterburner without supplementary firing is given approximately in Figure 6-9.
This plot, like that of Figure 6-7, is based on natural gas fuel, without
supplementary fuel value from the fume stream; it also assumes that the flue
gases will not be cooled below 350°F, because of the possibility of condensing and
corrosion on cold surfaces.
With the addition of supplementary firing, there is essentially no
top limit on the size of the boiler. With whatever version is used, safety and
controllability dictate using bypass around the boiler with blinds, so that
either afterburner or boiler may be opened up for service. There should be a
failsafe interlock control on firing that will provide emergency shutdown pro-
tection if there is a malfunction. Materials in the bypass/vent system should
be able to withstand the afterburner exit temperature.
6.3 Fouling and Cleaning
A study of the equations of Chapter 18 make it apparent that heat
exchanger surfaces must be clean for maximum effectiveness. A coating of
S-1^121 Figures 6-5 through 6-9 follow
-------
(5) P
*—•*• **—•«•—*—«*«-J
Stack
Combustibles
in
Fume
Fuel ''
Afterburner
Afterburner
Blower
Notes:
A Two-Way Gate to Have Maximum Position Stop and
Cut-Out Switch Actuated on Failure of Blower Pres-
sure, to Avoid Overheating Oven.
B Emergency Override Cuts Fuel to Afterburner if
Dangerous Over-Temperature in Oven.
A, B Should be Failsafe on Loss of Electrical Power or
Instrument Air.
Figure 6-5. RECYCLED FLUE GAS TO PROCESS UNIT
S-14121
67784
-------
Tempering
Air
Oven or
Process Unit
Clean Air
Fuel
Gas :
Temp. Override Switch
Shuts Down Afterburner
on Excessive Temp.
Afterburner
Cold Air
Heat Exchanqer
Exchanger Must be Able to
Withstand Stack Temp. With-
out Damage Upon Loss of
Cold Air Stream.
Figure 6-6. HEAT RECYCLE TO OVEN VIA HEAT EXCHANGER
S-14121
67784
-------
20
16
m 12
ID
0)
10
W
8 <>
V
Flue Gas Temperature / 1500°F
1400°F
1300°F
1200°F
Note:
Final Temperature of Flue Gas = 350°F
Fume Stream Used for Ox/gen Supply
(No Outside Air Used in Combustion)
I
4,000 8,000
Afterburner Capacity, scfm
12,000
16,000
Figure 6-?. PROCESS HEAT RECOVERABLE FROM AFTERBURNER
S-14121
67784
-------
Fuel
Blower
(May be Located
Elsewhere)
Water
Preheater
Notes:
A Automatic Diversion to Bypass
1) When Steam Pressure/Temp. Excessive and Boiler
Fuel Cut to Minimum (Using Dual-Range Controller
on Boiler Fuel).
2) Upon Loss of Boiler Feed Water.
3) Upon Loss of Boiler Flame.
4) (Closure of Both) Upon Shutdown of Afterburner.
C Blind When Boiler Not in Service.
B. C Blinds When Afterburner Not in Service
Figure 6-8. WASTE HEAT BOILER
S-14121
67784
-------
20
16
ID
Of.
ID
0
8
Flue Gas Temperature /1500°F
1400°F
1300°F
200°F
Note:
I
Steam at 100 psig, Saturated
Water at 60°F
Stack Temperature 350*F
No Supplementary Firing
I
4,000
12,000
8,000
Afterburner Capacity, scfm
Figure 6-9. APPRftXlMATE STEAM GENERATION RATE
WASTE MEAT BOILER UNIT
16,000
S-14121
67784
-------
97
polymer, dust, carbon, on the exchange surface provides resistance to heat
transfer, and less heat will be recovered. This is commonly accounted for in
the equations by an empirical "fouling factor". An eighth-inch thick layer
of polymeric, carbonaceous deposit on the fume side of a fume/flue-gas exchanger
would give a fouling factor of about .12; a cross flow exchanger having an
effectiveness of 50$ when clean might have an effectiveness of only U2J8 when so
fouled. The reduction in effectiveness, AE, represents a loss of fuel heat
value, AE , which can be a significant fraction for a high-effectiveness
1-E
exchanger.
To minimize this loss, frequent cleaning may be needed. Most exchanger
manufacturers prefer to put the fouling stream inside the tubes. Tough deposits
may then be reamed out by a tool called a "turbine". If fouling deposits are less
adherent, the fouling stream can be run over the outside of the tubes, and
cleaning may be accomplished in place by water or solvent washing, steam lances,
air Jets, or by "rodding" between the tubes with hand tools. Rectangular tube
arrays are preferred over triangular ones for this operation. The cleanliness
after such an operation frequently leaves much to be desired, and placing the
foul stream inside circular tubes is preferred. Cleaning by chemical or solvent
treatment may not work in the afterburner, since the deposits are likely to be
hard-baked to the point that they are not easily soluble.
The need for frequent cleaning also dictates that the exchanger and its
ductwork be designed for easy access. This can be done with cover plates and
access doors, or can be done by removing sections of ducting. There are no simple
answers to accessibility for cleaning. The desire for low first cost and minimum
weight and floor space favor closely coupling afterburner and exchangers; this
often hampers accessibility for cleaning. It is expensive to bring in a hydraulic
crane to lift ducting to one side. Flange Joints are heavy and (especially at
high temperature) develop leaks. The same is true of access doors. Using ample
space for access complicates thermal expansion problems in the ductwork.
A user contemplating heat recovery for his afterburner when fouling
conditions are expected would be wise to seek experience on similar installations
on a closely related process in his industry. In the absence of such favorable
experience, it is recommended that the prospective user view the use of a heat
exchanger with a somewhat Jaundiced eye, and include substantial allowances for
frequent servicing in his economic evaluations.
6.U Thermal Expansion and Thermal Stress
The metal temperature in a cross-flow exchanger varies from point
to point (as viewed in plan-form in Figure 6-1)> consequently, each tube
row will have a different increase in length as it heats up to operating tempera-
ture. A rigid structure so heated will warp or tear itself apart (or, upon
several cycles of heating, develop cracks from a "thermal fatigue" effect.
Flexibility to accommodate this thermal expansion must be built into the tube
fastenings in the tube plate (failure to do this in early versions of heat
S-
-------
98
recovery equipment resulted in many examples of buckled tubes and ruptured tube
fastenings. Sometimes tubes are attached to the tube plate with convolutions or
"expansion bellows" to give the needed flexibility.
Attached ductwork should also be built to accommodate thermal expan-
sion. This is especially true if any of the structure is to be built of stainless
steel, which has an expansion coefficient nearly twice that of carbon steel. On
large installations both afterburner and heat exchanger can be heavy, and duct
connections must be provided with flexible expansion Joints, or else one of the
units (the heavier) should be rigidly tied at one point, and the rest of the
equipment should ride on tracks that will allow free movement.
Where long hot ducts are used, as in the run from an oven to its
afterburner, or in connecting the afterburner flue to a heat exchanger, the thermal
expansion can.be large. For example, a hundred foot long fume duct at U50°F,
made of carbon steel, will expand about 3-1/2 inches over its cold length. A
25-foot long duct of stainless steel at 1500°F will expand about 5 inches.
Provision must be made for this movement; otherwise an attempt to tie such a
duct to an afterburner (which usually is built with a light gauge sheet metal
shell) will surely cause a structural failure.
A fair degree of sophistication is required in the design of heat
exchangers for use with large temperature differences. It requires a detailed
knowledge of temperatures' at each point in the system, and imagination in fore-
seeing the warping and distortion effects they may produce. Transient conditions
during startup may impose severe thermal distortion effects (one part may heat up
to operating temperature quickly, another part slowly). A major source of
difficulty is that flow and heating distribution is not uniform in ducts; one
fdde may run hot (from flame impingement?), the other side cool (from a recircu-
lating eddy that sweeps a cold wall?). Flow through a heat exchanger bank may
not be uniformly distributed, so that one group of tubes received more heat than
anticipated, another less. Where possible a user would be wise to select a
heat exchanger design that has proven itself under comparable service
conditions.
Chapter 7 gives some data on thermal expansion effects and coeffi-
cients on some commonly used materials.
6.5 Double Wall Construction
This refers to the use of the afterburner shell as a heat exchanger ele-
ment. The fume stream passes along and around the burner and holding chamber,
between the outer shell and the inner ceramic or metal liner. Only a marginal re-
duction in fuel burned can be gained this way, although the cool fume stream helps
make the outer wall less of a personnel hazard, and the insulation thickness needed
is reduced. Some vapor condensation can be avoided, but there is a heavy penalty
if the stream is a fouling one—the arrangement is very difficult (if at all
possible) to clean. Also, it is a hazard, in that fouling deposits may accummulate,
become ignited, and start a fire, releasing sufficient heat to damage or destroy
the afterburner. Further, as discussed elsewhere, a 255& LEL fuel concentration in
the cool fume stream may become a combustible mixture at elevated temperature; this
mixture might be ignited by hot spots on the liner surface, and cause burning inside
the wall of the afterburner. The prospective user should consider this design
arrangement with skepticism.
S-11H21
-------
99
Chapter 7. MATERIALS OF CONTRUCTIOH
The purpose of this chapter is to summarize the state of the art
in selection of materials for afterburner systems, and to give data which vill
be useful in equipment selection and design.
7.1 Scope of Materials Review
Presented here is a brief review of materials commonly used in the
construction of afterburners. No attempt has been made to review all the published
literature on this subject. Only technical information readily available was used
in this survey.
Some emphasis has been given to problems of corrosion, although most
afterburner applications will give rise to no special problems of corrosive attack.
Only a small minority of the afterburner installations involves the handling of
corrosive fumes. The corrosivity can arise from the fuel used in the afterburner -
that is, from burning oil fuels containing sulfur, or residual fuels containing
ash-forming organo-metallics - or from the fume stream itself, which may contain
sulfur, halides (chlorine, flourine, bromine, etc.). or very fine mists or parti-
culate clouds of an endless variety of materials. Note that any appreciable
quantity of some of these materials may require that the afterburner be followed
by some form of scrubber; the problems of applying scrubbers or other collectors
are outside the scope of this handbook. Because of the increasing scarcity of gas
fuels and the need for using oil fuels and the increasing emphasis on the cleanup
of chemical waste streams, some data on corrosion effects on metallic materials
have been included.
Two classes of materials are discussed - ceramic and metallic. They
are covered in two separate sections.
7.2 Refractories
7.2.1 Introduction
The normal configuration for construction of a thermal afterburner in-
volves a steel outer shell, lined with a refractory material. The purpose of the
refractory is to protect the steel shell from direct exposure to the effects of
high temperatures and corrosive materials, and to improve thermal efficiency of the
unit by limiting heat losses. Normal designs are based on the principle that the
refractory serves as a thermal insulator, lowering the temperature from a maximum
of more than 2000°F on the inside of the combustion chamber to a temperature of
250 - UOO°F at the steel shell. At this temperature the shell retains most of
its room temperature strength and is still hot enough to prevent condensation of
water vapor on the surface, thereby lessening corrosive attack.
The refractory serves as a physical barrier to the high temperature
combustion gases, allowing only minor amounts to percolate through to contact the
steel shell. The gases are also cooled as they migrate through the refractory,
S-1M21
-------
100
so that corrosion by gaseous attack on the shell is greatly reduced.
Finally, the refractory serves to minimize thermal losses from the
system, so that the burners may be run with relatively economical amounts of
fuel.
The refractory may have any one of a number of chemical compositions
and physical forms. Most refractories used in afterburners are made up of mix-
tures of aluminum oxide (usually called "alumina") and silicon dioxide ("silica").
For very special applications, bodies made up of magnesium oxide, chromium oxide,
zirconium dioxide, or silicon carbide might also be used. The physical forms
normally encountered are bricks, castables (refractory concrete), and dense boards
made of pressed fibres.
Among the many factors to be considered in choosing materials for
lining an afterburner are operating temperature, resistance to attack by com-
bustion gases, thermal conductivity, material cost, installation cost, density,
resistance to thermal shock, and vessel geometry. Additional operating factors,
such as amount of operating supervision available for the unit, accessibility for
repair, and frequency of shutdowns, also may affect materials selection. Each of
these factors will be discussed here in order to demonstrate its importance in
the materials selection process.
Refractory structures are heavy, with densities running 1*5 Ib/cu ft for
the light weight insulating firebrick and castable refractories up to 185 Ib/cu ft
for high alumina materials, and with wall thicknesses running typically from U to
8 inches (see Table 7-3). This weight adds considerably to the cost and difficulty
of shipping, so that many afterburner manufacturers do not install the refractory
until after the shell has been moved into place - especially on the large, custom-
designed units. The installation of full masonry construction calls for a rela-
tively skilled mason; an unskilled mason can make mistakes that will lead to
premature failure. Once in place the added weight of the masonry can be an important
consideration in the structural design of the building, especially if the afterburner
is to be installed on a roof or elevated platform. Because of its light weight,
fibre-block wall construction is being used in some afterburners (see Section 7-2.2.U),
However, this material has limited strength, care must be used in attaching it to
the metal walls, and gas velocities should be kept below 50 ft/sec to avoid the
possibility of erosive damage.
7.2.2 Classification of Refractories
Refractories composed of mixtures of alumina and silica are by far the
most common lining materials for thermal afterburners. They are generally classified
in four groups, depending on their form during installation: brick, refractory
concrete (castables), plastic refractories, and ceramic fibre materials.
7.2.2.1 Brick
Refractory brick consists of a mixture of clays, carefully selected by
composition and particle size, which has been fired to a high temperature to drive
-------
101
off volatiles and to form strong ceramic bonds between the component particles.
As a general rule, the maximum use temperature, thermal conductivity, and density
all increase with increasing alumina content.
The maximum use temperature corresponds to the temperature at which a
glassy silica phase within the brick softens, and the brick begins to flow. This
temperature is indicated by reference to pyrometric cones. These cones are
ceramic bodies of known softening point, which are placed in a furnace with the
brick. The cone which slumps at the same temperature that the brick flows is
then used as a reference indicator of maximum use temperature. Thus "cone 29"
or "P.C.E. (pyrometric cone equivalent) 29" refers to a brick which softens in
the same range as standard cone number 29, approximately 3050°F. Table 7-1
shows the classes of fireclay and high-alumina bricks, with corresponding com-
positions, cone ratings, and maximum use temperatures.
Table 7-1. CLASSIFICATION OF FIRECLAY AND HIGH-ALUMINA BRICK
Type
Low-Duty
Medium-Duty
High-Duty
Super-Duty
50% A1203
60* A1203
10% A1203
80$ A1203
90% A1203
Mullite
Composition
(Wt % A1203)
31 - 33
36
38 - k2
1*1 - 1*6
50
60
70
80
90
72
Pyrometric
Cone Equivalent
15 - 27
29
31-1/2
33
3U
35
36
-
-
-
Maximum Rated
Temperature, °F
2600 - 3000
3050
3100
3200
3250
3300
3350
3^00
3500
3350
A separate class of relatively low density brick, called "insulating
fire brick", is used for its insulating properties, usually as a back-up for a
denser brick lining. Brick is applied by the same hand lay-up techniques used
in the building trades. The mortars used are special versions of the castable
or plastic refractories discussed below.
S-14121
-------
102
7.2.2.2 Castable Refractories
Castable refractories are composed of mixtures of both raw and calcined
clays of carefully chosen composition and particle size distribution. They also
t'outain compounds, primarily calcium aluminate, which hydrate on addition of
water. The castable behaves in much the same way as ordinary concrete, except
that it forms very strong ceramic bonds after heating and dehydration. It may be
applied by casting in forms, by gunning, or by troweling.
The classification of castables is usually on the basis of density and
maximum use temperature; they may be obtained with ratings as high as 3200°F
(even higher if exotic materials are to be used). See Table 7-3. There are
light weight insulating castables, analogous to insulating fire brick, and denser
castables used for more general applications. Special extra-dense abrasion-
resistant castables are also available. These are particularly useful in systems
where considerable particulate matter is entrained in flue gases. The temperature
limit of a castable is usually stated in degrees, so that typical nomenclature
involves terms like "lightweight 2700° castable."
7.2.2.3 Plastic Refractories
Plastic mixtures of ground fireclays with chemical binding agents are
called plastic refractories. They are shipped wet from the manufacturer in
sealed bags, and are usually applied by ramming with pneumatic hammers. They
may be formulated to harden upon exposure to air ("air-set" varieties) or to
harden when heated, by formation of a ceramic bond ("heat-set" varieties).
The "air-set" types lose their strength on heating, and finally form ceramic
bonds after exposure to high temperature. Plastic refractory temperature ratings
are obtainable to 3200°F. Both castables and plastic refractories require
special anchoring to the afterburner shell. Anchors may be either metal pins
of various shapes welded to the shell, or refractory blocks attached to the
shell by metal clips. Every refractory manufacturer and contractor has a
particular preferred anchoring system, and as the choice is so dependent on the
specific application, a contractor or manufacturer should choose the system to
be used.
After firing at high temperatures, properties of bricks, castables,
and plastics of similar composition are almost identical, except that the
monolithic materials tend to be slightly less permeable than their brick analogs.
7.2.2.U Ceramic Fibres
There has been a growing trend in recent years to use ceramic fibres for
refractory linings. These fibres are usually of the kaolin composition (53#w SiO ,
U7J5 A12C>3) and are limited to applications below 2300°F. The fibres may be
packed together to form^a loose blanket, like mineral wool, or may be pressed with
a sodium silicate binder to form a rigid board. The pressed materials have a
somewhat lower (-1800°) temperature limit.
These materials have very limited strength, and must be installed with
S-1M21
-------
103
special clip fittings. They should not be used where maintenance calls for foot
traffic. The gas velocities should be kept low (not over 50 ft/sec) in order to
minimize the possibility of erosive damage.
7.2.3 Selection Criteria
7.2.3.1 Use Temperature
The single most important consideration in most refractory selection
procedures is the ability of the material to withstand the design operating
temperature. As shown in Table 7.1, most alumina-silica brick refractories have
use temperatures far above the normal afterburner operating range. However,
refractories near the burner may be exposed to temperatures in excess of 3000°F,
so that special attention must be given to the refractories of material in this
zone.
There is a common tendency to specify materials with a maximum use
temperature far above the predicted operating conditions, to allow a considerable
safety factor. This is not sound procedure, especially in the case of monolithic
(castable and plastic) materials. The preparation of refractory brick usually
involves firing to temperatures Just below the maximum use temperature in order
to develop ceramic bonds between individual grains. The process is not complete,
and brick will continue to increase in bonding strength during use. The bond-
formation process, known as sintering, is very sensitive to temperature, and
use of a brick at temperatures well below its rated temperature will minimize
any additional bond formation, leaving the brick relatively weak. Therefore,
selection procedures should aim for materials with use limits close to design
temperatures for the process.
In the case of monolithic lining materials, no ceramic bonding is pre-
sent until the lining is thoroughly cured near a maximum use temperature. Because
the cool face of the lining never reaches temperatures necessary for formation of
ceramic bonds, strength is usually provided by the hydraulic bonds formed at low
temperature. The center of a section of monolithic material usually has been
heated sufficiently to destroy the hydraulic bonding but not enough to form ceramic
bonds, and is relatively weak. If the hot face of the lining is never thoroughly
cured by soaking at near maximum temperatures, the entire structure will be quite
weak, and likely to spall.
The presence of a relatively weak center in a monolithic wall makes
the structure a little more flexible than rigid brick construction, and often
allows it to perform a little better in resisting thermal shock.
7.2.3.2 Resistance to Chemical Attack
Alumina-silica refractories are relatively inert, resisting chemical
attack by a great many materials. As a general rule, problems of chemical attack
are not a major factor in selection of afterburner materials. However, some com-
ponents of flue gases can severely attack refractories, and any selection process
must take these into account.
S-
-------
Most chemical attack causes refractory failure of three kinds. In
one case, the presence of fluxing agents will cause melting of the refractory
surface. If the melted material is removed by gravity flow or some other mechanism,
then fresh surfaces are repeatedly presented for attack, and reaction rates can
become so fast as to cause catastrophic failure. If the reaction product is not
readily removed from the refractory face, further reaction may be essentially
stopped. The reaction product will almost invariably have a sufficiently differ-
ent coefficient of thermal expansion from the host refractory that severe stresses
will build up during cool-down or subsequent start-up. The result is usually
cracking and spalling (sloughing off of layers of refractory parallel to the hot
face), with exposure of fresh surfaces to further attack.
The second primary mode of failure involves the formation of solid
reaction products. In the case of refractories for incinerators and afterburners,
operating temperatures are normally above the dew point for most corrosive species.
Condensation can occur, however, in the pores of the refractory away from the hot
face. The result of this condensation-reaction process is usually the formation
of a solid reaction product with thermal expansion coefficient markedly different
from that of the builk refractory. Depending on whether thermal expansion of
the reaction product is greater or smaller, spalling will occur on start-up or
shutdown, respectively.
A third mode of attack occurs only during shutdowns. If the reaction
products of high temperature attack are hygroscopic, that is, attract water at
room temperature, there may be swelling and spalling. Also, rapid heating on
start up may cause evolution of steam from these compounds. When steam is evolved
more rapidly than it can migrate out of the refractory, pressure builds up, and
the refractory actually explodes. This same phenomenon, of course, occurs during
initial curing of monolithic materials, and can be prevented only by very cautious
initial heating rates.
The elements which are most destructive in attacking alumina-silica
refractories are sodium, potassium, and vanadium. They react with aluminum oxide
and silicon dioxide to form very low melting compounds. Sodium may also attack
the alumina portion separately, forming the compound Na20*ll Al203, which tends
to promote bloating and spalling.
Calcium, zinc, phosphorus, iron, and cobalt are also often corrosive.
These elements attack refractories mostly by formation of solid state (at normal
afterburner temperatures) reaction products which later cause spalling during
shutdown.
Specific operating experience with afterburners damaged by some of these
elements includes:
In an automobile industry application burning used oil filters, the
refractory was attacked by zinc volatilized from the galvanized cases. The
attack was in the form of condensation behind the hot face, reaction to form zinc
silicates, and spalling.
5-1^121
-------
105
Sodium silicate binders used on wet-pack cardboard boxes were the
source of sodium which severely fluxed afterburner refractories.
Crematoria and pathological incinerators, as well as some smokehouse
afterburners, are often attacked by calcium and phosphorus compounds; the common
reactions are bloating and spelling.
Phosphorus can be particularly troublesome in afterburner linings con-
taining a significant fraction of iron oxide impurities. The common mode of
attack is solid state reaction, followed by bloating.
In fibre wall constructions using relatively low density pressed fibre
board, the presence of sodium in flue gases leads to embrittlement of the fibres
at temperatures above about l600°F.
7.2.3.5 Thermal Conductivity and Density
The thermal conductivity of refractories useful in afterburner linings
varies over a wide range. The primary controlling factor is the density of the
material. Table 7-2 lists some representative values for thermal conductivity
of refractories.
Table 7-2. APPROXIMATE THERMAL CONDUCTIVITIES OF REFRACTORIES
Refractory Type
Dense 90% Al2®3
Super-Duty Fireclay
High- Duty Fireclay
Insulating Firebrick
(50 Ib/cu ft)
Dense 60% A1203 Castable
(3000°F rating)
Light Weight Insulating
Castable (2000°F rating)
Conductivity at T, °F
Btu/hr -ft-°F
200
25.7
8.7
8.U
1.8
U.6
1.5
1000
19.3
9.3
9-0
2.U
5.1
1.9
2000
17.7
10.0
9.8
3.2
5.7
2.1
S-1U121
-------
106
Insulating firebrick, which is often used as a back-up to a dense fire-
clay brick liner, is almost four times lower in thermal conductivity than super-
duty brick. Obviously, if no factors other than lining thickness and fuel economy
were considered, dense brick would rarely be used. However, insulating firebrick
offer much poorer resistance to chemical attack than do denser brick of the same
composition. The low density of the insulating brick allows much greater pene-
tration of corrosive materials, and the higher surface area increases reactivity.
Density, in addition to influencing thermal conductivity, and reactivity,
is also of importance in economic considerations, for structural design, and in
operating factors. The use of lightweight lining materials, which are usually
cheaper on an area basis for equivalent thermal performance, also reduces shipping
costs in those units in which refractory linings are shop-installed. The lighter
linings permit the use of lighter steel construction, with attendant savings. In
many afterburner applications roof mounting is desirable or even unavoidable;
and the savings in lining weight permitted by lightweight materials, particularly
fibre board construction, can make these applications economically feasible.
Typical densities of various refractory materials are shown in Table 7-3.
Table T-3. DENSITIES OF REPRESENTATIVE AFTERBURNER REFRACTORIES
Material
Super-Duty Firebrick
High-Duty Firebrick
Dense 90% A1203 Brick
2300° Insulating Firebrick
2000° Lightweight Insulating Castable
3000° Dense Castable
General Purpose 2200° Castable
3000° Plastic
Density, Ib/cu ft
1U5
135
180 - 185
U5
50 - 55
lUO
120
lUO
7.2.3.1* Thermal Shock Resistance
Because of the brittle nature of refractory materials, they are subject
to cracking when subjected to sharp temperature fluctuations. Because refractories
are poor thermal conductors, the maintenance of strong temperature gradients is
favored; and the stresses caused by differential thermal expansion from hot face
S-Ik121
-------
107
to cold face can be considerable. Most manufacturers of refractory-lined
equipment specify the maximum heating and cooling rates which the units can
tolerate; the deciding factor in setting these limits is the resistance of
the lining to thermal shock.
Thermal shock resistance increases with increasing thermal conductivity
because stresses due to thermal gradients are lessened. Increase in density
decreases the resistance to thermal shock, particularly among very dense materials.
The presence of pores prevents spreading of microcracks; reduction of pore volume
with increased density allows greater crack propagation.
The mineral mullite (3 A1203'2 Si02) is an important material because
of its excellent resistance to thermal shock. The combination of low thermal
expansion coefficient, moderate conductivity, and high strength make this the
best alumina-silica material for applications requiring good thermal shock
resistance.
7.2.3.5 System Geometry
The shape of a reactor often strongly influences the choice of which
class of refractory will be used for a lining. The use of monolithic materials
permits design and installation of much more complicated linings than would be
possible with brick construction. A large number of shapes are available in
refractory brick, but often many of these are not shelf items, and long order
times are needed.
As a first approximation, brick construction is limited to simple
patterns - straight walls, cylinders, cones, and arches. Castable refractories
may be used for any shape where installation of forms is possible. This is
especially useful for providing baffles for combustion chamber mixing. Castables
are particularly suited to small, relatively complex parts which may be fabricated
in the shop. Plastic refractories can be used in any configuration where there
is sufficient access to permit ramming.
The geometry of the reactor strongly affects the economics of the
lining installation, as is discussed in Section 7.2.U on costs.
7.2.3.6 Thermal Stress Loadings
The combination of ceramic and steel construction can give serious
problems if proper allowances are not made for the relative thermal expansion
of ceramic and metal parts exposed to temperature variations. This is discussed
further in Section 5.3.6. Figure 7.1 shows the linear thermal expansion of
common types of brick. Castable materials tend to exhibit less thermal expansion
than brick but it isn't actually negative as often claimed. When castable
ceramic cures it shrinks by an amount greater than subsequent thermal expansion
when it is raised to lUOO° - 1500°F. Therefore in afterburner applications
castable refractory occupies the most volume when it is first installed and no
thermal expansion need be allowed for.
S-1U121
-------
108
7.2.3.7 Flame Impingement Problems
Ordinarily one would expect that flame impingement would not be a
problem in afterburners provided the ceramic is selected so as to have a rating
capable of withstanding the flame temperature. However, trouble could certainly
be anticipated for situations where a 3000°F flame impinges a refractory that is
rated at only 1500°F (as might easily be the case with some of the castable
refractory materials). Moreover, there is a high probability for difficulty if
the flame impingement gives an alternating exposure to the hot flame and the
unmixed cold fume, brought about by the random and turbulent motion of the flame.
The alternate heating and cooling of the surface will succeed in bringing about
a slow spoiling attritition. The problem is especially acute with oil flames
bfcause of their highly radiant character.
7.2.3.8 Installation Methods
The different types of linings call for vastly different installation
methods; the choice of a lining is largely determined by the relative costs of
the possible alternative installations.
Brick linings require experienced, skilled workmen for installation.
Because of their relatively high weight, brick linings are almost never shipped
installed, but are installed by refractory contractors at the field installation
site.
Castable linings are much simpler to install, and withstand damage
during shipment rather well. For this reason, plus their applicability to a
wide range of geometries, and the economy and chemical resistance of lightweight
castable linings, they are probably the most widely used materials in afterburners,
Unskilled plant personnel can install (and repair) these linings, which are
normally dried to about 200°F before shipping. This drying removes excess water
and strengthens the hydraulic set of the lining.
Most plastic refractories do not have adequate strength to withstand
shipping damage; and skilled installation is necessary. They offer few advan-
tages over other lining types, and are relatively less common in afterburners.
Fibre linings offer the optimum properties consistent with plant
installation and shipment: excellent mechanical shock resistance ease of
installation, and light weight. They are nearly always shop installed.
7.2.U Costs
It is almost impossible to place real values on costs of linings
because of temporal and regional fluctuation in materials prices, freight
rates, and labor costs. The values presented in the following sections are
intended only as guides to indicate the range of costs that might be expected.
S-1^121 Figure 1-1 follows
-------
2.0
1.6
70%AI2O3
—^—^ High Duty or Super Duty
Fireclay Brick
2600°F Insulating Fireclay Brick
1.2
S
ID
Q.
X
ID
0)
0.8
0.4
I
800
2,400
1,600
Temperature, °F
Figure 7-1. LINEAR THERMAL EXPANSION OF COMMON BRICK TYPES
3,200
S-14121
67784
-------
109
7.2.U.I Materials
Monolithic materials may cost more than the equivalent brick on a
thermal efficiency basis. That is, the amount of castable or plastic necessary
to provide a given shell temperature may cost somewhat more than the amount of
firebrick needed for the same Job. However, for brick shapes other than very
simple configurations, this advantage may vanish, and will also be offset by
labor savings for monolithics. In late 1971, the following price ranges were
realistic (all prices at point of manufacture): superduty firebrick, about
$250/1000 brick (brick prices for "9" straights," a brick 9" x U-l/2" x 2-1/2");
2500° castable, about $150 per ton; and a comparable plastic about $125/ton (wet).
On a volume basis, these figures are equivalent to about $U.50/cu ft for brick,
$3.00 to $12.00/cu ft for castable (depending on density), and $7-00 to $10.OO/
cu ft for plastic.
7.2.U.2 Installation
The speed of installation of various linings is extremely dependent on
geometry. For very simple configurations (walls, floors, etc.) manufacturers
estimates range from 2l» - 28 hours of labor/1000 bricks, and 10 - 2U hours/ton
of castable or plastic. This would make installed costs of a simple superduty
firebrick structure $UO - 60/cu ft, and castables $15 - 50/cu ft, depending
on density; these prices do not include freight charges for the materials.
Fibre wall construction consisting of fibre blanket and mineral
wool block back-up has been estimated in the trade Journal at approximately
$15/cu ft, installed.
It should be stressed that labor for a complex installation of brick
could force brick prices very much higher, and that the use of low-priced labor
for castable installation could lower that estimate. Also, these figures do
not take into account the differing thermal efficiencies of the materials. The
real figure for comparison would be cost per square foot for a wall thick enough
to provide the necessary shell temperature.
7.2.U.3 Maintenance
The maintenance costs of various lining types depend strongly on the
mode of operation, amount of supervision, and accessibility for repair. In
simple applications where the unit is rarely shut down, no corrosives are present,
and reasonable care is exercised in heating and cooling rates, many years of
trouble-free service can be expected.
If the unit is subjected to frequent over-temperature excursions,
brick construction may have an advantage. When thermal shocking is unavoidable,
the fibre approach may be very attractive.
It is probably somewhat more economical to make small repairs on
monolithic linings than on brick; but so many factors, such as availablility
S-
-------
110
of labor, accessibility, etc., enter into this situation that firm guidelines
are dangerous.
1.2.3 References
There is very little published literature which bears specifically on
the problems of refractory linings in afterburners. However, there is an
enormous amount of information in background areas, such as calculation of wall
thicknesses, or proper curing methods for monolithic linings. The reference;
listed 7-1 - 7-6 are intended to provide convenient sources for detailed dis-
cussions of subjects only touched on here.
7.3 Metallic Materials
Ordinary low carbon steels are used exclusively for the structural
components of afterburners . Many manufacturers use a gas-tight structural con-
tainer to enclose a combustion chamber that is built of metal, but which is lined
with ceramic material for heat insulation. For most afterburner services this
container vessel is also of low carbon steel, as it experiences a temperature
of no more than UOO°F. Other units are designed with metal parts exposed to the
inside of the combustion chamber; these units make extensive use of alloy
steel construction. Units of this type do not expose the metal to the hot part
of the flame; at most they require that the metal withstand the 1200 to. lUOO°F
temperature of the residence chamber, and they have often been designed so that
one side of the metal parts is cooled by the incoming cool fume stream, so that it
experiences a temperature below that of the gas in the chamber.
The low temperature, non-corrosive applications that permit of carbon
steel construction will not be discussed here, since they lie within the realm
of common engineering experience.
The creep resistance and the creep rupture life of metal parts exposed
to the high temperature levels becomes important, and data is given on some
commonly used materials. Data is also provided on the oxidation resistance of
metals at high temperature. Special problems of chemical environments likely to
give corrosive attack are reviewed; also a related one of cool surfaces where
corrosive conditions can arise because of condensation of components in the flue
gas. Some thought is directed to the problems of dealing with thermal fatigue.
Reactions of metals and nonmetallic materials with corrosives is a
highly complex subject. In this summary it is possible to give only preliminary
consideration to the many combinations of corrosives and materials of construction
that may be encountered. The purpose here is to indicate the factors that in-
fluence selection of construction materials . Any specific environment of interest
would need to be considered in greater detail. Thus, the information presented
is not to be considered as complete for final selection of materials for a speci-
fic afterburner application.
The afterburner may be called upon to handle vapors of varying composi-
tion. The combustion gases may contain a number of acidic components in addition
to the gases commonly found in boiler or furnace operation. Thus, selection of
S-1M21
-------
Ill
materials of construction cannot be based solely on previous experience in other
types of furnaces burning conventional fuels. The limitations on materials of
construction that relate to the fire side of steam boiler environments will apply;
but in addition, the environment may be more severe producing additional limita-
tions on materials of construction.
The additional environmental components to be considered here include
low concentrations of chlorine, hydrogen chloride, oxides of nitrogen, phosphorous
pentoxide, sulfur compounds plus various metal oxides and salts such as A1203,
Na20, NaCl, Si02 and oxides of potassium and cobalt.
7.3.1 Mechanical and Metallurgical Considerations at High Temperature
Metal parts within the heated enclosure must be designed to withstand
the operating temperature. The effect of temperature on creep strength and stress
rupture strength of a number of metallic materials of construction is illustrated
by the data in Table 7-1* and Figure 7-2 and 7-3. It will be seen that creep is
a significant factor in use of many metals at temperatures as low as lUOO°F.
In addition to changes in creep strength and other mechanical properties,
many alloys are subject to changes in metallurgical properties at elevated tempera-
ture. Specific comments on use of various common materials of construction at
elevated temperature are presented below.
7-3.1.1 Carbon Steel and 1/2% Mo Steel
After prolonged exposure at temperatures in the 800° - 1300°F range,
carbon steel and 1/2% Mo steel in general experience spheroidization of carbides
and graphitization. Significant reductions in strength accompany both spheroidi-
zation and graphitization. Although spheroidization will affect plate steel uni-
formly, graphitization occurs preferentially in weld heat affected zones. Following
elevated temperature service, e.g., delayed coking units operating in the 900°F
range, 1/2% Mo steel has been found to be more susceptible than carbon steel to
cracking. The minimum commercial alloy necessary to prevent graphitization is
1% Cr - 1/2% Mo steel.
7.3.1.2 Low Alloy Steels (l% Cr. 1/2% Mo to 5% Cr. 1/2% Mo)
These steels exhibit the normal reduction in tensile properties with
increasing service temperature. However, if heated above the transformation
temperature (on the order of 1300°F) and subsequently cooled rapidly, they have
reduced ductility at lower temperatures. To prevent this decrease in ductility
when overheated, the material must be cooled slowly (50 to 100°F per hour) down
to 1100°F.
7-3.1.3 Aluminized Steel
Experiences with the use of aluminized steel in afterburner service
have not been good. The aluminizing process is a coating process by which aluminum
is applied to and diffuses into carbon steel or alloy steel as a surface layer.
S-1U121
-------
112
The coating has good resistance to high temperature oxidation and to high temperature
sulfur corrosion. Damage to the coating is likely to occur and in such cases metal
areas that are not aluminized are also exposed to the corrosive environment.
Aluminizing is not recommended as a substitute for the proper alloy to provide
oorrosion resistance. Its use may be considered if the coating is applied to
increase the life of a component that can perform satisfactorily in the uncoated
state. However, the aluminizing process may produce adverse effects on the
mechanical properties of the base metal.
7.3.1.U Chromium Stainless Steels (Types U05 and UlO)
The 12% Cr stainless steels (Types U05 and UlO) are generally satisfactory
materials for use above 800°F. However, higher chromium steels - 17$ Cr or more -
are not acceptable because they are susceptible to "885°P embrittlement." This
aging phenomenon produces extensive decreases in toughness and ductility in the
high chromium alloys after long term exposure in the 700 - 900°F range. Because
the minimum chromium concentration necessary to cause 885°F embrittlement has never
been conclusively established, Type 1*10 stainless steel (13.5$ Cr maximum) is to
be preferred over Type U05 stainless steel (lk.5% Cr maximum).
Close attention must be directed toward welding of the 12% Cr steels
because variations in weld deposit composition can exert marked influence in
mechanical properties, especially toughness. We are aware of no experience with
large vessel fabrication involving solid 12% Cr materials. However, these steels
have been used extensively as vessel linings.
7.3.1.5 Chromium Nickel Stainless Steels (Types 30U, 321, 3^7 and Cast
Stainless SteelsT
Metallurgical factors associated with long term exposure at high tempera-
ture can result in greatly reduced corrosion resistance on subsequent exposure in
liquid corrosives at lower temperatures. Due to carbide precipitation (sensitiza-
tion), all of these stainless steels will be less corrosion resistant after pro-
longed exposure above 800°F. Thus, it is desirable to ensure that they are not
exposed to liquid acid corrosives of the environment, even at ambient temperature
such as during shutdowns.
Exposure of these steels in the temperature range of 1000 to l600°F
can result in formation of "sigma phase." This can markedly reduce subsequent
room temperature ductility. However, above 1000°F, presence of sigma phase may
increase tensile strength with ductility being only slightly affected.
Type 3^7 stainless steel is ^subject to pronounced weld heat affected zone
cracking and is not recommended for elevated temperature service.
There are a number of cast stainless steel compositions for high temper-
ature service. Any specific alloy would have to be considered in detail before it
is selected for high temperature service. The HT alloy, for example, is susceptible
to hot shortness, an embrittlement phenomena producing cracking at elevated
temperature. The preferred composition HK-UO (carbon content 0.37 to
Table 7-b follows
S-lU-121 Figures 7-2 and 7-3 follow
-------
Table T-k. CREEP STRENGTH
Metals
Iron Base Alloys
Steel
1/2 Mo
1 Cr-1/2 Mo
1-1/1* Cr-1/2 Mo
2 Cr-1/2 Mo
2-1/lt Cr-1 Mo
5 Cr-1/2 Mo
7 Cr-1/2 Mo
9 Cr-1 Ho
12 Cr - Type U10
Stainless - }OU
Stainless - }16
25 Cr-12 Nl Steel
25 Cr-20 Nl Steel
Incoloy 600
Copper Base All»s
Copper
Tin Bronze
Aluminum Bronze
Arapco Ho. IE
Red Brass
Yellow Brass
Mun-.z Metal
Admiralty
Silicon Bronze
70-JO Cu-Nl
Nickel Base Alloys
Mor.el !>00
Monel K-500
Nickel ZOO
Ineor.el £00
Hastelloy 3
Hastelloj C
Hastelloy F
Hastelloy X
Super Alloys
Nlronic 80
Incanel X-750
M-252
Refraetaloj 26
A -266
Discalco
N155
S-590
S-816
Haynes Alloy 31
Haynes Alia) 25
Ha\nes Alloj 21
16 Cr-25 Nl-6 Mo
19-9 DL
17— PH
Other Metals
Aluramia
Lead
Titanium
Zirconium
Tantalum
Stress for Creep Rate of 1* In 10,000 Hours, kpsi
Temperature, 'F
'•,00
-
.
_
-
-
-
-
-
_
-
-
.
-
6.0
_
_
_
9.5
5.0
-
19
15
-
_
.
_
.
-
_
-
-
_
-
_
-
-
_
-
.
.
-
_
-
.
-
_
_
_
_
-
600
_
.
_
.
-
-
-
-
_
.
.
.
-
_
_
_
.
-
_
.
2
_
-
_
-
ItO
.
-
_
_
-
_
.
_
-
-
_
.
_
-
-
_
.
_
135
_
_
_
_
-
SCO
1?
29
29
29
25
-
33
33
.
_
-
_
_
_
_
-
_
.
0.2
_
-
26
_
6
hO
-
_
-
-
_
_
_
.
-
_
-
_
.
-
_
.
_
£
_
_
_
_
-
1000
5
11
13
11
11
13
10
6
12
11
18
25
16
16
"•5
_
_
_
_
-
_
_
_
_
-
3.0
_
12.5
_
_
-
_
_
_
_
80
_
_
s
Nickel Base Allo.^^
S. per Allots
ASM Handboox, ASTM Bulletins 100, 12U, 157, 180, 228, Bulletins Tlmken Roller Bearing Co. National
Tube Co., and Babcock and Wilcox Tube Co.
ASM Handbook, Bulletins Ampco Metals, Inc., Bridgeport Brass Co., Scovill Mfg. Co., and ASTM
Bulletin 161.
ASM Handbook, Bulletins International Nickel Co. and Haynes Stellite Co.
ASM Handbook, ASTM Bulletin 160, Engineering Alloys Digest, International Nickel Co., Hejnes
Stellite Co., Allegheny Ludlum Steel Co., Cyclops Sieel Co., and Crucible Steel Co.
S-14121
67784
-------
Long-Time Elevated-Temperature Properties • 1400 F •leooF DHIlisooF
4 8 12 16
Stress to Rupture (looopsi)
20
Limiting Creep Stress (looopsi)
•Not recommended for service above isoo P.
Figure 7-2. STRESS RUPTURE AND LIMITING CREEP STRESS OF
CAST STAINLESS STEELS AT ELEVATED TEMPERATURES
Source: Schoefer, E.A., Machine Design, April 2, 1959
Letters at Left Margins of Graphs Refer to Alloy Designations;
See ASTMStd. A 297
S-14121
67356
-------
Superolloys
nickel-bose
Superolloys
Cr,Ni,FeaCf,Ni,
Co,Fe(heot-
treoted)
Superolloys
cobolt-bose
Superolloys
os worked^
Cortndge
bross
Molybdenum
alloys
Aluminum-
base alloy
Cermet
K-161 8
APMP
257 -
^vi Costmognesium-bose
x| olloy EM-62
Cast HT stainless
300
50O
1000
Temperature, °F
1500
2000
Figure 7-3. STRESS TO RUPTURE IN 1000 HR vs TEMPERATURE
H.C. Cross and W.F. Simmons, ASM Preprint, 1954. See also
'Super Alloys", Universal-Cyclops Steel Corp., Bridgeville, Pa.
5.-14121,
67356
-------
113
is commonly used for applications below 1900°F.
7-3.2 Oxidation Resistance of Commonly Used Alloys
The relative oxidation resistance of a number of steels, stainless steels
and nickel base alloys is indicated by the data in Table 7-5 and Figure 7-*+. It
will be seen that high nickel and chromium contents are associated with increased
oxidation resistance.
Low alloy steel boiler tubes can be used frequently in heat exchangers
because the metal temperature is much lower than the furnace temperature. However,
if there is a fouling deposit on the cold side of the tubes, it reduces heat
transfer and increases the metal temperature. If this occurs, the rate of oxidation
increases and tensile properties are reduced, both contributing to possible tube
failure.
7.3.3 Problems of Special Chemical Environments
7.3.3.1 Sulfur Dioxide and Hydrogen Sulfide
The relative effect of S02 in air or in air plus water vapor on corro-
sion of steel at l650°F is seen from the data in Table 7-6. Sulfur dioxide signi-
ficantly accelerates corrosion under these conditions.
The relative resistance of various metals and alloys in sulfur containing
atmospheres is indicated by the tabulation in Table 7-7.
We have no information on corrosion of metals by 803 in flue gas under
comparable conditions.
7»3. 3.2 Compounds Containing Vanadium. Sodium, Potassium, Sulfur. Molybdenum.
or Lead
Compounds containing the above materials are capable of accelerating
corrosion of steels, stainless steels and other metallic materials of construction
at elevated temperature. Some of these constituents are found in residual fuels.
While the concentration in the fuel may be quite low, deposits on metal surfaces
in boilers may contain these components in high concentration. Vanadium and
sodium react to form products melting as low as 980°F. This liquid and others
from the elements mentioned above promote severe corrosion. The mechanism of
corrosion appears to be transfer of oxygen to the metal via the concentrated
liquid to form a porous metal oxide as the corrosion product. Thus, oxygen is
essential to the reaction, and the liquid on the metal surface is not used up in
the corrosion reaction. Corrosion by vanadium in fuel oil is illustrated by the
corrosion data in Figure 7-5.
Calcium oxide, strontium oxide, and other compounds combine with vana-
dium oxides and have been added to fuel oils to minimize corrosion.
S-
-------
Ill,
Table 7-5. MAXIMUM TEMPERATURE WITHOUT
EXCESSIVE SCALING IN AIR*)
Type of Steel
Maximum Temperature
Without Excessive Scaling,
°F
Carbon steel
% Mo steel
1 Cr | Mo steel
2fr Cr 1 Mo steel
5 Or £ Mo steel
12 Cr AISI UlO
18 Cr 8 Ni AISI
18 Cr 8 Ni Nb AISI 3^7
18 Cr 8 Ni Mo AISI 3l6
25 Cr 20 Ni AISI 310
1050
1050
1100
1175
1200
1250
1600
1600
1600
2100
a; Data from "Corrosion Engineer's Reference Book",
2nd edition, BIPM, 1965, p. 133.
Table 7-6. EFFECT OF INDUSTRIAL GASES OF DIFFERENT
COMPOSITIONS ON GASEOUS CORROSION OF
CARBON STEEL AT 1650*1^
Atmosphere
Pure air
Air + 2% S02
Air + 5# HsO
Air + 5# S02 + 5% IfeO
Relative Corrosion
100
118
135
276
a; Source: Tomashov, N. D., "Theory of Corrosion
and Protection of Metals", The MacMillan Company,
New York, 1966, p. llU.
Figures 7-1* and 7-5 follow
-------
Figure 7-4. AIR CORROSION OF Fe-Ni-Cr ALLOYS
This nomographic chart is based on the 100-hr test data of Brasunas, Gow,
and Harder, ( Proc. Am. Soc. Testing Mater., 46:870, 1946) and presents
air-corrosion data in concise and convenient form which would otherwise
require pages of tables or numerous curves. Loss in accuracy is to be
expected in-condensing data and, should more precise data be required,
the original article may be consulted.
S-14121
67356
-------
100
9O
80
70
r60
8 50
9
b c
f 40
3O
2O
F\J Clomp
I22 Support
Bunker C oil
(200 ppm V, 50 ppnh No, 2.5% S)
at 1500 F
<7-450Ohr
0-645O hr
C-9429 hr
Figure 7-5. OXIDATION OF FURNACE TUBE SUPPORTS EXPOSED TO
HIGH-VANADIUM FUEL OIL COMBUSTION PRODUCTS
Nickel Topics, 18:4, 1955, International Nickel Company, New York
S-1412'
67356
-------
115
Table 7-7. RESISTANCE OF VARIOUS METALS
AND ALLOYS TO SULFUR-CONTAINING GASESa)
Materials listed in order of increasing corrosion resistance
HaS
Ni, Co
Mild steel
Fe
Fe-Mn
Inconel
Cu
Fe-15 Cr
Fe-25 Cr
Cr
Fe-l8 Cr-8 Ni
Fe-22 Cr-10 Al
Cu-10 Mg
Fe-12 Al, Ni-15 Al
Ta, Mo, W
Al, Mg
S
Ni, Cu
Mild steel
Fe
Fe-lU Cr
Cu-Mn
80 Ni-13 Cr-6.5 Fe
Mn
Cr
Fe-17 Cr
Fe-18 Cr-8 Ni
Hastelloy
Al, Mg
a; Source: 0. Kubaschewski and B. E. Hop
SQa
Ni
Fe
Cu-10 Mg
Fe-15 Cr
Ta
Cu, brass
Al alloys
Mo, W
Fe-30 Cr
Fe-l8 Cr-8 Ni
Inconel
Cu-12 Al
Zr
kins, "Oxidation
of Metals and Alloys", p. 277, Butterworth and Co.
(Publishers), Ltd., London, 1962.
-------
116
7-3.3.3 Carbon Dioxide and Carbon Monoxide
The presence of carbon dioxide and carbon monoxide in flue gas probably
does not accelerate corrosion in most cases with iron base alloys. However, in
high C02 concentrations or at sufficiently high temperature a CO-CO? atmosphere
is capable of decarburizing or carburizing steels (and other alloys) by the forward
and reverse directions of the reaction
C(Fe) + C°2 *=? 2 C°'
In addition, a C0-C02 atmosphere is capable of directly oxidizing iron or reducing
iron oxide by reactions such as
Fe + C02 l — ; FeO + CO.
An equilibrium diagram for Fe-C0-C02 is given in Figure 7-6 which shows the regions
of stability of the various species.
7-3.3.U Chlorine and Hydrogen Chloride
Carbon steel and low alloy steels are not suitable for exposure at
furnace temperatures in gaseous C12 or HC1. In Tables 7-8 and 7-9 are listed the
suggested maximum temperatures for exposure of various alloys in dry chlorine
vapor and dry HC1 vapor at elevated temperature. It will be seen that the
temperature listed for steel is UOO°F in C12 and 500°F in HC1.
Corrosion tests (Figure 7-7) at 900°F with a few percent Cl and steam
present indicate severe corrosion of carbon steel and Type U05 stainless steel.
Hastelloy C was the most resistant of the alloys tested.
7'3.3.3 Bromine and Hydrogen Bromide
We have no specific information on corrosion by Br2 or HBr at elevated
temperature. Temperature limits may be similar to those for C12 and HC1.
7.3.3.6 Phosphorous Compounds
When phosphorous compounds are burned in the presence of water vapor,
the combustion product will condense as a liquid except at very high temperatures.
The boiling point of the azeotropic mixture of phosphorous pentoxide and water
(92% phosphorous pentoxide) is l600°F.
Corrosion of carbon steel and Type 3l6 stainless steel is greater than
3 inches/year in concentrated phosphoric acid (97 to 127$) at U50°F. It appears
that a liquid phase of P20g or other phosphorous compound would probably be very
corrosive to most metallic materials of construction at elevated temperature.
S-1^121 Figures 7-6 and 7-7 follow
-------
400
600
800
1000
1200 1400
Temperature, °C
Figure 7-6. EQUILIBRIUM DIAGRAM OF THE Fe-CO-CO, SYSTEM
W.D. Jones, "Fundamental Principles of Powder Metallurgy", p. 537,
Edward Arnold (Publishers) Ltd., London, 1960
S-14121
67356
-------
320
160
0.9 psi Chlorine
11 psi Steam
800
_L
405.
320
-------
117
Table 7-8. HIGH TEMPERATURE CORROSION
BY DRY HYDROGEN CHLORIDE GAS
Based on data from short term laboratory
tests by Brown, DeLong, and Auld (ind. Eng.
Chem., 39, P- 8J9-M, July 19^7).
Material
Platinum
Gold
Nickel
Inconel
Haste Hoy B
Haste Hoy C
Type 316 SS
Type JOU SS
Carbon steel
Monel
Copper
Aluminum
Suggested Upper Temperature
Limit, °F, for
Continuous Service
2200
1600
950
900
850
850
800
800
500
U50
200 a)
Not suitable at ambient temp
a; Uhlig, H. H.f "Corrosion Handbook", John
Wiley and Sons, page 767.
S-
-------
118
Table 7-9. HIGH TEMPERATURE CORROSION
BY DRY CHLORINE GAS
Based on data from short term laboratory
tests by Brown, DeLong, and Auld (ind. Eng.
Chem., 39, p. 8J9-UU, July 19^7).
Material
Nickel
Inconel
Hastelloy B
Hastelloy C
Monel
Type 316 SS
Type 30U S3
Platinum
Carbon steel
Copper
Aluminum
Tantalum
Gold
Suggested Upper Temperature
Limit, °F, for
Continuous Service
1000
1000
1000
950
800
650
600
500
Uooa'
Uoo
250^'
300** '
<250
a) Ignites at ^50 to 500°F.
b) Ignites at 270 to l*50°F; literature
references vary widely.
c) Metals Handbook, A»1, vol. 1, p. 1222.
S-1U121
Figure 7-8 follows
-------
400
350
300
S 250
o»
•I 200
a
a>
° 150
100
Curve is correlation
of field measurements
using clean electrodes.
0.001 0.002 0.003 0.004 0.005 0.006
Sulfer trioxide in flue gas, volume % on the dry basis
Figure 7-8. WATER DEW POINT WITH SO3 IN FLUE GASES
Chemical Engineering, Vol. 75, No. 16, p. 168, July 29, 1968
S-14121
67356
-------
119
7-3-1* Further Investigation of Metallic Materials of Construction
The information presented above on metals in furnace environments has
served to illustrate the many problems that can be encountered. For any further
detailed consideration of this subject it would be desirable to survey published
literature, manufacturing practices, and other sources of information on the
current state of the art. However, it is likely that laboratory investigation
would be required to supplement information available from others. For afterburners,
corrosion of metals in atmospheres containing low concentrations of acid gases
at furnace temperatures appears to be such an area for possible investigation.
7.3.$ Problems of Heat Recovery Equipment and Cooled Metal Surfaces
Exchangers or steam generators following the afterburner for heat recovery
usually require installation of banks of metal tubes that would be exposed to the
furnace atmosphere. The temperature of the surface of the tubes would determine
the type and extent of corrosion that could occur. High temperature corrosion by
various components of the furnace atmosphere has been discussed above.
Another type of corrosion can occur from condensation of a corrosive
liquid phase on metal surfaces. The temperature at which such a liquid phase
could condense would be determined by the specific components present in the
furnace atmosphere.
In any case, prior to shutdown of the furnace it would be necessary
to purge all acid gases from the furnace atmosphere in order to prevent corrosion
by condensed liquid acidic components on metal surfaces when the furnace is not in
operation. Comments on specific acidic components in furnace atmospheres are
presented below.
7.3.5.1 Phosphorous Compounds
As indicated above, it appears that concentrated phosphoric acid could
condense at temperatures as high as l600°F.
7.3.3.2 Sulfur Trioxide
Minute amounts of S03 normally present in flue gas are strongly absorbed
in water droplets so that the corresponding acid condensates are quite concentrated.
The sulfuric acid concentration of initial condensate droplets typically falls in
the 20 to 80$ H-SO^ range. Such acid compositions are severely corrosive to
steels, stainless steels and many nickel base alloys. The dewpoint of water with
very small amounts of S03 in flue gas is illustrated by the data in Figure 7-8.
These data indicate the dewpoint may be as much as 3Uo°F with S03 present.
However, other published information indicates condensation temperatures as high
as 1*65°F.
7.3.5.3 Other Acid Gases
The dewpoint of HC1, N02> etc., is lower than the acid forming components
S-1U121
-------
120
mentioned above. However, it is necessary to maintain metal surfaces well above
such dewpoints to prevent condensation corrosion. For example, in HC1 containing
a small amount of water vapor it is recommended that the temperature be at least
50°F above the dewpoint to prevent corrosion. This is due to the hygroscopic nature
of metal chloride corrosion products.
7.3.5.U Steel Shell of Afterburner
The afterburner often consists of a steel shell lined with firebr* ck or
castable refractory. As the lining is not totally impervious, furnace gases may be
in contact with the steel shell. The temperature of the steel shell is a significant
factor in the amount of corrosion that can occur. One way of preventing corrosion
by most acidic components is to maintain the temperature of the shell above the
condensation temperature of acid gases in the furnace atmosphere (normally above
U50°F).
This approach cannot prevent condensation by phosphorous compounds. At
one .installation, an incinerator was constructed recently that periodically burns
phosphourous compounds. Based on a survey of similar furnaces, it appeared that
phosphorous compounds would form slags with the refractory. This would tend to
seal the surface, preventing exposure of the underlying steel. In addition,
acidic phosphorous compounds that penetrated the brick would have been essentially
neutralized by reaction with components of the refractory. Thus, it appeared that
a steel shell could be used. Similar reactions would not be expected to occur with
most other acidic components in the furnace atmosphere.
Another method of preventing corrosion is to have the metal surface at a
low enough temperature so that an organic membrane can be used to cover the steel
inner surface and thus protect it from the acid gases. For most membranes, this
means a temperature of less than 200°F.
Another approach that has been used is to construct a pressure tight
steel box from tubes (for heat recovery) and tube fins inside the furnace to
prevent flue gas from contacting the exterior.
7.3.6 Thermal Distortion and Thermal Fatigue
Metals, especially the high-nickel stainless steel alloys, have high
thermal coefficients of expansion compared with ceramic materials. On the other
hand, ceramic parts are often exposed to high temperatures, but supported and res-
trained by the metal structure. It is of the utmost importance that careful
allowance be made for the relative movement between cold and hot parts in such a
way that the structure is not damaged. The problem is especially severe during
startup conditions, when some parts of the system directly exposed to the flame will
heat up quickly, while other parts will heat only after an extended period of oper-
ation. Figures 7-9 and 7-1 give data on the thermal expansion coefficients
of metallic and ceramic materials at room and elevated temperatures.
Many instances of failures have been encountered where proper allowances
for this differential movement have not been provided for. It should be kept in
S-1^121 Figure 7-9 follows
-------
1600
1400
1200
Expansion
1000 -
e
*
0
15 600
v
Q.
Room Temperature
Coefficient of Expansion
per °C x 10-6
Steel
Cast Iron
Copper
18-8 Stainless
Monel
Nickel
InconeI
Aluminum
Lead
-200
-400
200
S-14121
67784
40 80 120 160
Percent of Room Temperature Coefficient
of Expansion Retained
Figure 7-9. THERMAL EXPANSION DATA ON METALS
-------
121
mind that restraint on the hot ceramic with a cold structure will often load the
ceramic portion of the system in compression; most ceramics are very strong in
compression, and will force a containing structure to yield, and the structure will
be loose upon cooling. Or if the loading places the ceramic in tension, the low
strength and brittle behaviour of the ceramic will lead to the formation of cracks,
and possibly to its collapse.
Still further problems arise in the design of metal parts exposed to
large variations in temperature. If the thermal expansion is restrained, the
structure will warp, twist, and deform. If subjected to repeated cycles of
temperature loading in this fashion, the structure will fail from "thermal fatigue."
Many instances of warped and cracked metal members in the near vicinity of the flame,
even though they may be made of heat -resisting alloys designed for the correct
temperature range, attest the severity of the problem.
S-1M21
-------
123
Chapter 8. AUXILIARIES
The afterburner system is centered around the afterburner itself and its
heat exchangers, but consideration must also be given to bringing fumes to the
afterburner and discharging the clean flue gas. Also, refractory-lined fume
incinerators are heavy and special supporting structures or concrete pads will be
required as part of the installation.
Fume collection and delivery to the afterburner is a very important
part of the total system and much attention should be paid to its design since
both capital and operating costs are sensitive to the manner in which it is
handled. Long duct runs can cost more than the afterburner Itself and condensa-
tion of combustibles can occur even in insulated ducts, causing a fire hazard.
Proper hooding is crucial in order to insure that all fumes are collected and that
air dilution is kept to a minimum. Once the fume is below 25% of the LEL, all
extra air represents wasted fuel in the afterburner. Existing ventilation systems
will probably require modification to reduce unnecessary dilution and to vent clean
air directly to the atmosphere rather than through the afterburner.
This chapter is not intended to guide the design of the fume collection
and delivery system. The reader is directed to "Industrial Ventilation"8"1^ and
Chapter 3 of the "Air Pollution Engineering Manual"8~2^ for comprehensive treatments
of hood and duct design and blower selection. Forced versus induced draft blowers
are discussed in this chapter since afterburner installation will nearly always
involve a separate blower. Cost and power requirements are estimated. Cost
estimating information is also supplied for ducting, and curves are given for
estimating the afterburner weight which must be supported. A brief treatment is
given of particle de-entrainment devices installed before the afterburner to
reduce the pollutant load entering the afterburner. Some of these same devices
could be used on a cooled flue gas stream to remove particulates or acidic compo-
nents formed within the afterburner (Section 3*1.^.7).
8.1 Blowers
Afterburners are usually supplied with their own blowers which may be
either forced draft (waste gas pushed through the afterburner) or induced draft
(waste gas pulled through). With induced draft systems, the high temperature of
the afterburner discharge usually makes it necessary to blend in cool air
immediately before the blower to reduce the effluent temperature to around 600°F.
High efficiency heat exchangers (>50# recovery) will achieve this temperature
reduction without air dilution. Induced draft fans could be designed to operate
at the full afterburner exhaust temperature of lUOO°F but this would require
special high temperature alloys which would greatly increase fan costs. Also,
bearing failures are a problem at high temperature, and even at 600°F, external
bearings are utilized.
For a given pressure increase through the blower, the power consumption
is proportional to the actual volume of gases handled. The volume at 600°F is
twice that at room temperature, and dilution air also increases the volume. Both
Preceding page blank
S-1U121
-------
124
these factors will increase power requirements. Therefore induced draft
blowers must be larger (increased capital) and involve higher operating costs
than forced draft blowers.
Induced draft systems do offer some advantages, however. The fan sees
only a clean, treated stream and there will be no tendency for buildup in the fan
blades. Such buildup will occur with forced draft fans in systems where tarry
materials are likely to condense or impact on the blades, e.g., smoke house or
asphalt blowing offgas. This buildup can cause fires due to friction genei ited
heat or more commonly it may unbalance the fan. Also, leakage of hot and
possibly toxic gases can occur in a forced draft system but all leaks will be
inward with an induced draft system. This is important since hot gas leakage
through a small hole can quickly enlarge the hole and require major repairs.
Normal system pressure drop ranges from 2 in. to 10 in. water column
depending on whether the unit is thermal or catalytic, and whether it is fitted
with a heat exchanger, and the number of passes in the heat exchanger. The
blowers may be generally classified as industrial fans. They are usually fitted
with heavy duty radial blades which are simple, robust, and least affected by
particulate matter. A number of manufacturers supply them in standard sizes to
30,000 cfm capacity with static pressures from 1 in. to 20 in. water column.
Normal maximum operating temperature is 350°F, but fans may be modified for
operation up to 900°F by fitting cooling wheels and loose fitted water cooled
bearings. Blowers are available with higher pressure ratings but are in a
different classification.
Power requirements for blowers are proportional to the actual volume
moved and to the total static pressure increase. Plots of brake horsepower against
cubic feet per minute at pressures from 2 in. to 28 in. water column are shown in
Figure 8-1. An efficiency of 60% has been assumed. When sizing the motor for a
blower, care must be taken to insure that the worst conditions have been allowed
for. A fixed rpm blower connected to a system of ducts at a uniform temperature
throughout, will deliver an actual volume which will be approximately constant
regardless of the temperature level. The pressure developed by the blower and the
pressure drop through the system are both proportional to gas density and both will
decrease as the temperature rises (for constant ACFM). Thus power requirements,
which are proportional to the volume handled and the pressure developed will
decrease as the temperature rises, and maximum power will be needed on a cold
startup. An afterburner installation is somewhat more complicated since the
temperature of the waste gas from the process may vary as well as the temperature
of the afterburner, and the blower may be either before or after the afterburner.
Pressure drop calculations must be made for each part of the equipment (ducting,
afterburner, heat exchanger, etc.) for startup conditions to determine maximum
motor horsepower needed. Electrical power usage during normal operation is calcu-
lated from the operating conditions. In some installations power is conserved
during startup by fitting vane control inlets to the blower which reduce the volume
of air handled1. This same system can be used to reduce power consumption whenever
the process is operating at reduced flow rates.
S-14121 Figures 8-1 and 8-2 follow
-------
S-14121
67784
103
102
X
u
.2!
"o
Q.
X
m
3
JS
J
(0
U
10
10
-i
103
'/
= 28" Water
10*
6"
2"
104
Blower Actual cfm
Figure 8-1.
10s
-------
10
, 1
28Water Column
Operating Temp.: 350°F
Material: Carbon Steel
10-'
I I I
I I
103
104
Capacity, cfm
Figure 8-2. BLOWER COSTS
Costs Include Motor, Belt Drive and Guard
10s
S-14121
67784
-------
125
A comprehensive treatment of blowers is given in References 8-1 and
8-2, and much information is available from manufacturers' brochures.
Typical blower costs complete with motor and drive are shown in
Figure 8-2, for a system operating at 350°F. In addition to normally encountered
pressures of 2 in., 6 in., and 10 in. water column, a plot for 28 in. pressure
drop has been included. This indicates the cost penalty incurred when mixing
is improved at the expense of pressure drop through the afterburner. Electricity
costs will also increase in proportion to the pressure drop. For example, at
10,000 CFM and 2 in. pressure, the annual electrical cost would be $2UO per year,
whereas if the pressure were increased to 28 in., the cost would be $3300 (based
on 6000 hrs/hr and electricity at Ic/KWH).
8.2 Ducts
Ductwork requirements depend largely on the type of installation, the
number of units served, and on the location of the afterburner in relation to the
process equipment. The ducts are often insulated to prevent condensation of
volatile material on the inside. This type of condensation can be a serious
fire hazard, and even where the ducts are insulated, it is a wise precaution to
install access panels for inspection and clean-out if it should be required.
Most systems operate below the 25% LEL, but for those situations where
this concentration is exceeded, provision must be made to prevent flashback. This
can be done by providing high velocity sections in which the waste gas velocity
is higher than the flame propagation velocity. If concentrations are above the
HEL, fumes may be ducted without the need for costly air dilution. (Fuel must be
burned to heat up the dilution air to lUOO°F.) However, it is essential that air
be excluded at «*"n points between the fume generating process and the afterburner.
An alarm system and/or careful maintenance are required for rapid detection of
leaks.
The cost of ductwork is highly variable, but the following figures may
be used as a guide for galvanized carbon steel in relatively short runs:
Size Diameter,
in.
6
9
12
15
18
2U
Ga.
2U
22
22
22
20
18
Duct,
$/ft
U.OO
T.Uo
9.80
12.20
18.00
29.60
Insulation ,
$/ft
2.30
3.50
U.70
5-90
7.00
9.UO
Stainless steel ducts would cost approximately 2.5x galvanized ducts.
S-1M21
-------
126
Methods of calculating pressure drop in ducts, the design of hoods for
providing ventilation for different types of equipment and general information on
various hydrocarbons, including flammability limits, are covered in some detail in
References 8-1 and 8-2.
8.3 Supporting Structures - Afterburner Weight
Installation costs are extremely dependant on the supporting structure
required for the afterburner. Figure 8-3 shows the range of equipment we.'ghts
obtained from manufacturers' data. If the afterburner can be mounted at ground
level on a concrete pad, the weight will have little influence on the installation
cost. However, afterburners are frequently mounted on the roof of a building.
Roof mounting often allows a minimum of ducting since the afterburner can be located
directly above the fume generating process. Also roof mounting saves valuable
(or unavailable) space within the building and eliminates the need for a tall
stack to discharge flue gases above the building. If the roof itself is not
strong enough to take the additional load (very common in old buildings), a
special supporting structure with expensive steel work will be needed. On a one-
story building, it may be fairly easy to run columns to the ground and the additional
expense for supports will not be high. However, on a multistory building,
costs may be prohibitive and special lightweight designs or ground level mounting
will be required. Thus, it is important to consider the weight factor and the
cost of supporting structures at the time of the design.
8.U De-Entrainment
Carryover of particulate matter, both liquid and solid, may add sub-
stantially to the load on an afterburner. Solid particles may not burn or may
not have sufficient residence time at temperature in the afterburner to be destroyed
(Section 3.1.4.6). Excessive liquid carryover of a combustible material may
result in excessive heat release in the afterburner with combustion chamber tempera-
ture exceeding the design value. This will result in shutting down the unit and/or
damaging the installation. It is therefore necessary to provide equipment for
removing these particles before the waste gases reach the afterburner. Typical
examples are the removal of chaff from coffee roasting off-gases and removal of
entrained hydrocarbon mists from asphalt blowing.
There is a large number of different types of equipment for removing
particles and mists and there is extensive literature on the subject. References
8-U through 8-13 cover a representative selection which give general information
of the equipment and design criteria. Chapter U of Reference 8-2 also covers this
subject.
S-14121 Figure 8-3 follows
-------
S-14121
67784
70
60
50
. 40
30
20
10
10
75% Heat Recovery
20
M scfm
30
40
Figure 8-3. AFTERBURNER WEIGHTS - THERMAL
Based on Manufacturers' Information
-------
127
Chapter 9. CONTROL AND SAFETY
Failure of control instruments causing major or nuisance shutdowns is
one of the most common complaints from the operators of fume incineration
equipment. In other cases improperly installed temperature sensing elements or
inadequate control systems have resulted in excessive fuel consumption, large
temperature variations, and operation at temperatures too low to give accept-
able pollutant destruction. Also, proper use of process controls will enable
the user to operate drying ovens and similar equipment so as to obtain higher
solvent levels in the fume stream, thereby reducing fuel costs. For these
and other reasons, it is important that an afterburner system be chosen which
includes effective and reliable process controls.
This chapter is intended to guide the selection of control and
safety systems for afterburners. Descriptions of the control equipment
generally available and typical control systems are given. Recommended
practices in regard to control and safeguards are discussed in detail. In-
cluded are sections on combustion safeguards, temperature controls, and vapor
concentration controls. This chapter is intended to supplement, not supplant,
the recommendations of national approval bodies and current Federal, state and
local codes and regulations. All process controls for afterburner systems
should be approved by local authorities having jurisdiction and the
insurance underwriter.
9.1 Combustion Safeguards
9.1.1 National Approval Bodies
There are several bodies in the United States that make recommendations
and/or approve equipment and systems relating to safety and fire protection.
These bodies include both non-profit organizations and groups sponsored by
insurance underwriters. A brief description of the major approval bodies
follows.
1. Underwriters Laboratory, Inc.
207 East Ohio Street
Chicago. Illinois 60611
Underwriters' Laboratories, Inc. (UL) is a non-profit laboratory for
the testing of equipment and systems. They are primarily concerned with the
manufacture of equipment in accordance with standards of safety. Once an item
or system is approved, they are concerned only that the item continues to be
made to the same specifications as the original that was tested.
2. Factory Mutual System
1151 Boston-Providence Turnpike
Norwood, Mass. 02062
Factory Mutual System (FM) is the testing laboratory of nine fire
insurance companies which specialize in insuring manufacturing plants against
fire, explosions, and other losses.
S-11H21
-------
128
FM publishes many of their findings in manuals and bulletins. They
also publish a list of approved devices for industrial fire protection. They
consider each system or installation on its own merits before approval is
given.
3. Factory Insurance Association
85 Woodland Street
Hartford, Conn. 06102
The Factory Insurance Association (FIA) is composed of member stock
insurance companies. FIA is concerned with all phases of fire protection and
other perils that are insured by its member companies.
FIA generally, accepts devices that are approved by Underwriters
Laboratories, Inc. However, they approve or disapprove each system or
installation individually.
U. National Fire Protection Association
60 Battery March Street
Boston, Mass. 02110
The National Fire Protection Association is a non-profit voluntary
membership organization with over 200 national trade and professional organi-
zations as Organization Members and more than 2U,000 Associate Members in
industry, fire departments, government, etc.
All NFPA Official Codes, Standards, and Recommended Practices are
published as the National Fire Codes. These standards are used as the basis
of local codes and ordinances on building and fire regulations and are widely
adopted by industry as a basis of good practice.
It should be noted that for a system to be approved by one of the
national agencies, it should be constructed of approved devices where they are
available. However, the use of devices which bear an approval label does not
necessarily mean that the system is approved. Approval of the system is made
on an individual basis.
9.1.2 Recommended Practices
The recommended practices of the national agencies for the installations
of combustion safeguards for afterburners are basically the same as those set
forth for single burner boilers and furnaces. Although there are some minor
differences in definitions and recommendations of the various agencies, on most
points there is agreement. All of the approval bodies require a pre-ignition
purge to clear any combustible gases from the burner. However, the recommended
purging rate varies from 50 to TO per cent of the maximum air flow and the minimum
recommended purge volume ranges from four to eight air changes. A purge timer
is required with an interlock that will not allow the fuel valves to open until
the purge period has been completed. In the case of afterburners, facilities
should allow the pre-ignition purge to be made with fresh air rather than the
S-
-------
129
vent stream to "be burned, especially if material may have condensed in the
duct during the period the equipment was idle.
An interlock to insure the presence of combustion air is recommended.
This is normally a pressure switch plus an electrical interlock to the fan motor
starter. Loss of either power to the fan motor or loss of air supply pressure
(this could be either positive or negative pressure depending on whether the
unit has a forced or induced draft fan) will cause the fuel valves to close.
If the afterburner uses the vent stream for the combustion air supply, then
the interlocks would be on this stream.
The recommendations for fuel gas valves vary somewhat as shown in
Figure 9-1- FM recommends a manual opening safety shutoff valve for manual and
semi-automatic burners. NFPA allows automatic opening, spring closing safety
shutoff valves but recommends that two safety shutoff valves be installed in series
with an automatic closing, spring opening vent valve between them to give a
double block and bleed arrangement. They further recommend a position indicator
and interlock on the valve nearest the burner. This interlock prevents initiation
of the firing sequence until this valve is closed. FIA recommends a double
block and bleed arrangement similar to NFPA except that the upstream valve shall
be a manual opening safety shutoff valve and the position interlock is not re-
quired. For gas to the pilot NFPA and FIA recommend a double block and bleed
arrangement with automatic safety shutoff valves. FM allows a single safety
shutoff valve.
Any of the above systems should provide safe operation if the equipment
is inspected regularly and kept in good condition. However, the FIA and NFPA
system provide redundance for added protection. The final choice should, of
course, comply with local fire regulations.
For oil fired burners a single automatic opening, spring closing,
safety shutoff valve in the oil supply line is recommended by all agencies.
Provisions should be made for heating and recirculating heavy oils.
Fuel pressure supervisory switches are recommended. Gas burner
installations should have both high and low pressure interlocks with the safety
shutoff valves. Oil burners normally require only the low pressure interlock,
but other interlocks may be needed on the oil temperature and atomizing steam
or air.
All agencies recommend the installation of an approved non-recycling
flame safety shutdown system which will stop the fuel flow within 2-U seconds
if the flame is extinguished. The system may detect the presence of a flame
by means of a flame rod, or other device sensitive to visible, infra-red or
ultra-violet flame energy. The merits of each method will be discussed later.
In addition to stopping the fuel on flame failure, the flame safety shutdown
system should include timers which will close and lockout the fuel valve if
the pilot and/or main flame is not established in a reasonable time.
S-1M21
-------
130
9.1.3 Firing and Light-Off Options
Many options are covered by recommendations of the FM, FIA, and NFPA.
Gas only, oil only, or both gas and oil can be used for fuel to the main burner.
The fuel selected will probably be based on the economics but the use of a single
fuel, preferably gas, will result in a simpler control system that will be easier
to maintain.
The degree of automation available for lighting off the burners ranges
from a permissive system where the operations are performed manually to one
where all the operations are automatic after the start button is pressed. The
definition of these systems as supervised manual, semi-automatic, or automatic
varies with the recommending agencies. The user should select the system that
best fits his requirements keeping in mind the frequency of startups and the
increased maintenance that the more automated systems will require. Any system
selected should have a continuous pilot and should not attempt to relight on
a flame failure (be non-recycling).
9.1.U Equipment - Flame Detectors. Shut Off Valves. Pressure Switches
9.I.U.I Flame Detectors
A flame detector is an essential item in the combustion safeguard
system. Most practical flame detectors use one of two basic principles. One
type uses a flame rod in actual contact with the flame and employs the rectification
of an electric current by the ions in the flame to determine the existence of
the flame. The other type depends upon the detection of radiation from the
flame. The radiation wave length detected can be in the infra-red, visible
or ultra-violet region depending upon the design of the unit.
The flame rod has a fast response. However, its direct contact with
the flame results in oil and carbon deposits on the rod which limit its usefulness.
They must be cleaned and replaced quite often and, in some types of fuels which
have a low resistance flame, the low output of the flame rods results in nuisance
shutdowns.
The flame detectors based on radiation do not come in contact with
the flame but they must have a clear viewing port so they can see the flame.
The detectors are also limited to an ambient temperature of approximately 200°F
and may require cooling in some applications. Photocells which sense radiation
in the visible region are limited in their ability to reject radiation from
the refractory. Infra-red detectors overcome the problem of hot refractory
but they can be fooled by seeing shimmering hot air. Ultra violet detectors
are reliable in applications to both gas or oil flames. An additional safety
feature is available on some ultra-violet detectors that interrupts the view of
the flame about 20 times per second. The circuit will re-establish only if ultra-
violet radiation still exists. The circuit, in effect, proves to itself 20
times per second that a flame exists.
S-114-121 Figure 9-1 follows
-------
Vent
to
Fuel Atmosphere
Gas |
Control
Valve
0 .
Supply
$ <
— 1W-* tljjt fc.
'^J — i 1
y^-1^- * I
(T
^
^
•iV 1 1
.!/ . 1— 1
to! i ^
y [^^^~ "
Gas Safety
Pressure Shut-Off
Regulators Valves Gas ^
\ X \ Pilot /^
-LA r
i —
\*y i
^^ I
•«v Burner (pSL/*l
>* v " y ii
• • X.O ^
T
J Spark
1 Plug
r\ Flame |
^y^->. Detector | i
®- — 1 i i ,
i
•
fc .. t i u
Vent •* •"•-*• — i Combustion _J
Combustion
Air
[A]« L Safeguard
PSL
PSH
Press. Switch- Low
Press. Switch-High
Notes:
1. Valves Shown in Solid Lines are Required by FM.
2. Additional Valves Shown in Dotted Lines Required by FIA and NFPA.
3. NFPA Required Interlock Switches but Not Manual Reset.
Figure 9-1. TYPICAL COMBUSTION SAFEGUARD ARRANGEMENT FOR A SINGLE GAS BURNER
-------
9.1.1*.2 Shut Off Valves
There are several types of shut-off valves used in combustion safeguard
systems depending upon the requirements of the approving authority and the degree
of automation of the system. The safety shut-off valves required by FIA, FM,
and NFPA are electrically operated valves which shut-off the fuel automatically
when de-energized. They must be constructed so that they cannot be blocked
open and they must be closed by a spring that maintains a minimum force of 5
pounds on the seat. On the valves nearest a gas burner, NFPA requires position
switches on the stem for use in the safety interlock system. The manual reset
safety shut-off valve is designed so that it can be opened and reset manually
only after it has been energized. It closes automatically when the circuit
is deenergized regardless of the position of the reset handle. All safety shut-
off valves should be approved by the appropriate authority.
9.1.U.3 Pressure Switches
Various switches which sense air flow, temperature, fuel pressure,
etc. are necessary to complete the combustion safeguard system. The switches
which meet the standards of the UL and/or FM are approved by these bodies.
9.1.5 Typical Gas Fired Single Burner Combustion Safeguard System
The system shown in Figure 9-1 is typical. The recommendation of
the appropriate authority should be considered for a specific system.
S-14121
-------
132
9.2 Temperature Control
9.2.1 Control Concepts
9.2.1.1 Oxidation Zone
The obvious need for temperature control in the combustion area occurs
because a minimum temperature must be maintained in order to properly oxidize
the fumes. Maintaining temperatures much in excess of this minimum results
in higher fuel costs and also possibly higher maintenance costs. There are
also other requirements which must be considered when a temperature control
system is selected. In afterburners which operate at high temperatures the
refractory may not be able to stand the shock of a rapid change in temperature.
In this case a temperature programmed startup system may be required with the
temperature control point raised slowly to the final operating point. This
can be done either automatically or manually if the temperature span of the
controller is broad.
Some burners may have a limited turndown and at low fume loads may
result in higher than desired fire box temperature. This problem can usually
be handled by split range control of the burner valve and another valve which
admits dilution air to the afterburner. As the load on the afterburner is reduced,
the controller reduces the fuel to the burner until the minimum allowable setting
is reached. The dilution air valve then is opened and additional cool air mixed
with the incoming fumes to maintain the firebox temperature at the control point.
9.2.1.2 Shutdowns and/or Alarms
In addition to the normal temperature controls there should be a separate
thermocouple and high temperature shutdown that will shut-off the fuel valve
and vent the fumes in case the firebox temperature exceeds the safe limit. A
low temperature alarm should also be considered if the fumes cannot be properly
oxidized below certain temperatures.
Whenever the burner is shutdown either manually or automatically,
the fumes should be shut-off and fresh air admitted to the burner. The fan
should continue to run long enough to purge the fumes from the combustion chamber,
but not long enough to chill the refractory if this can cause damage. A timer
could be considered to provide the purge automatically.
9.2.1.3 Inlet Fume Temperature
In some afterburner installations the combustion products pass through
a heat exchanger to preheat the incoming vapor stream. This reduces operating
costs but unless precautions are taken a reduced flow of fumes to the afterburner
can cause overheating, possible preignition, and damage to the heat exchangers
and/or ductwork. If condensation has previously taken place in the ductwork
the increased temperature at reduced flows may cause re-evaporation and result
is an explosive mixture in the ductwork. Control of the temperature of the
S-1M21
-------
133
incoming fumes will usually prevent trouble. In certain installations, it may
be necessary to provide a controlled bypass around the exchanger in order to
adequately control the preheat temperature (see Chapter 6).
9.2.2 Burner Control
9.2.2.1 Recommended Practices
The control of the temperature of the combustion area of the afterburner
will normally be accomplished by modulating the flow of fuel to the burner.
It is of the utmost importance that measurement of the temperature to be controlled
is properly made. The temperature measuring point must be located where the
temperature is representative of the mixture of fumes and the combustion products
of the burner. This will require that the probe be in a location where the
gases are well mixed and are moving at a velocity high enough to insure good
heat transfer. The probe should not be mounted where it is subject to flame
radiation or impingement or in a dead spot where the gases are essentially stagnant.
9.2.2.2 Equipment - Thermocouples, Controllers, Control Valves, Overtemperature
or Safety Devices
Thermocouples
Most afterburners will operate in a temperature range of 850-1500°F.
In this range, base metal thermocouples are the most inexpensive and trouble
free means of measuring temperature. Iron-Constantan (Type j) couples are inexpen-
sive and are good to lUOO°F. At temperatures up to 2300°F, Chromel-Alumel (Type K)
couples should be used. The thermocouples should be installed in wells to protect
them from oxidation and/or corrosion.
Controllers
The selection of the temperature controller will depend upon economic
factors and the required precision of the control. Regardless of the type selected,
the controller, or the transmitter if one is used, should have upscale thermocouple
burnout protection. This means that in case of a faulty thermocouple circuit,
an abnormally high signal will be seen by the controller which will shut-off
the fuel to the burner.
The most simple form of controller is the electric temperature controller
with relay output. This controller accepts an input directly from the thermocouple
and operates relay contacts which cause an electric motor to open or close the
fuel valve. Since the thermocouple connects directly to the controller, the signal
level is very low and the lines must be kept short to avoid losses and interference.
The form of control is usually simple three position control. If the temperature
is too low, a relay closes which opens the fuel valve. Another relay operates
and closes the fuel valve if the temperature is too high. The "dead zone" between
the two relay operating points is adjustable as well as the proportional band
or the speed that the valve operates. This controller provides a reliable
system at a minimum cost and requires only electric power for operation. However,
-------
because of the "dead zone" necessary to prevent continuous cycling of the valve,
the temperature will float between the high and the low setting. This is not
a problem unless very precise temperature control is required.
If the afterburner is installed in a plant which has conventional
pneumatic or electronic process control instrumentation, it may be desirable to
use the same type of equipment for controlling the afterburner temperature. The
temperature can be controlled more precisely and spare equipment for maintenance
may be reduced. However, more equipment is usually required and the cost w'll
be greater. Oil-free air at 20 psi will be needed to operate the valve. A
transducer is normally required to convert the thermocouple signal to a standard
pneumatic or current signal that the process controller will accept. Since
the pneumatic or current signal is a high level signal, it can be transmitted
a considerable distance to a remote controller if desired. The output signal
from the controller is a standard pneumatic or current signal which ultimately
operates a pneumatic control valve in the fuel line.
Control Valves
The type of control valve will be dictated by the controller selected.
The electric temperature controller with the relay output requires a reversible
proportioning motor acturator on the control valve. The use of the proportioning
motor actuator is limited to valves 3-1* inches and under with low pressure drops.
The stroking speed is relatively slow compared to the pneumatic operated valves.
However, these are not serious disadvantages in installations of one or two
burners.
Both the pneumatic and electronic process controllers require a pneumatic
valve operator. An electronic valve positioner or transducer is also required
for the valve operated by the electronic controller.
Overtemperature or Safety Devices
The sensing points of any combustion area temperature alarms should
be located Just as carefully as the points for the temperature control. Again,
thermocouples are probably the best choice for sensors for low cost and reliability
at elevated temperatures. Reliable and inexpensive solid state switches are
readily available which will accept a thermocouple input. These switches can
operate high or low temperature alarms or be connected to the combustion safeguard
system to shutdown the burner.
9.2.2.3 Typical Burner Controls and Alarms
The system shown in Figure 9-2 is typical of a system that will provide
reliable temperature control at a minimum cost (approximately $750).
A more precise and complex system is shown in Figure 9-3* This system
might be installed as part of a large plant where instrument air is available
and remote*monitoring may be desirable. The cost for a pneumatic system would
be approximately $1200 and for an electronic system about $lUOO.
S-1^121 Figures 9-2 and 9-3 follow
-------
Electric
Motor
Operated
Val
Gas
Supply
Relay Contact Output
Part of Combustion
j Safeguard System
PSL Pressure Switch-Low
PSH Pressure Switch-High
TIC Temperature Indicating
Controller
TSH Temperature Switch-High
Figure 9-2. MINIMUM COST TEMPERATURE CONTROL SYSTEM
S-14121
67784
-------
To
Combustion
Safeguard
System
Pneumatic
Or Electronic
Process Controller
Temperature
Transmitter
Gas
Supply
Cooling
Air ~-
r
"^
i
1
»
\ jj
1 Pneu
1 (JSH)
-.^
\ I
" ^ 5^ ^ rf
1 1
1 AH
^ VQ -jl Afterburner |
matic
* Control Valve
To
Stack
\
Split Range Option:
Valve Positioners are Installed
on Gas Control Valve and Cooling
Air Damper and Arranged to Oper
ate Sequentially.
Inlet
Stream
Figure 9-3. TEMPERATURE CONTROL SYSTEM USING
PROCESS CONTROL INSTRUMENTATION
S-14121
67784
-------
X55
Alarm points or shutdown system cost approximately $150 to $200 for
any of the systems.
9.2.3 Inlet Fume Temperature Control
9.2.3.1 Recommended Practice
When a heat exchanger is installed to preheat a widely varying flow
of inlet fxomes with the combustion products, temperature control on the inlet
fumes should be considered. The temperature can be controlled by either by-
passing the fumes around the exchanger or by adding cool air to the fume stream
ahead of the exchanger. Bypassing the inlet stream results in the most economical
use of fuel but the heat exchangers must be constructed to withstand the temperature
of the exhaust gases with reduced cooling from the inlet stream. Diluting the
fumes with cool air provides continuous cooling for the heat exchanger but requires
fuel to heat the added air up to the required oxidation zone temperature. Bypassing
the hot gases around the heat exchanger is another approach but the bypass valve
must be able to handle the hot exhaust gases. The selection of the temperature
control method will depend upon an economic balance of increased fuel costs
versus the increased construction costs.
If prolonged operation at reduced rates is planned, a control system
should be provided that will continuously measure the inlet temperature and
either add air to the inlet stream or bypass gases around the heat exchanger.
If only occasional operation at reduced flow rates is expected, a simpler inter-
mittent control system can be used. This system opens a damper to admit a given
amount of air to the inlet stream whenever the temperature exceeds the safe
value.
9*2.3.2 Equipment - Temperature Measurement and Control, Temperature Switches
Temperature Measurement and Control
The same type of equipment can be used here as was used for measuring
and controlling the oxidation zone. However, since the temperature of the inlet
stream is considerably lower than the oxidation zone, a gas or liquid filled
transmitter can be used with the pneumatic or electronic process controllers
if either of these systems are selected. The filled transmitters cost about
$200 less than the thermocouple transmitters but a pneumatic or electronic control
system will still be more expensive than the electric temperature controller
operating directly from a thermocouple.
Temperature Switches
Thermocouple switches can be used here as well as in the combustion
zone. Temperature switches based on the expansion of liquid in a bulb system
are also satisfactory at a somewhat lower cost.
S-
-------
136
9-2.3.3 Typical Systems
Figure 9-U shows a typical system where continuous control is maintained.
The output from the controller goes to either the air dilution valve or the by-
pass valve but not both. The system shown in Figure 9-5 is typical for an inter-
mittent control system.
S-11*121 Figures 9-k and 9-5 follow
-------
To
Stack
Temperature
Alarm-
High
Incoming
Stream
A
/\
Exhaust Gases
}
) i JL JL JL
Afterburner
-^TIC
Dilution
Air
To Combustion
Safeguard System
Figure 9-4. TYPICAL SYSTEM FOR CONTINUOUS CONTROL
OF INLET GAS STREAM
S-14121
67784
-------
Incoming
Stream
Dilution
Air
Temperature Switch
Very High
To Combustion
Safeguard System
Figure 9-5. TYPICAL SYSTEM FOR INTERMITTENT CONTROL
OF INLET GAS STREAM
S-14121
67784
-------
157
9.3 Handling of Flammable Vapors
9.3.1 Recommended Practice
Detailed discussions of ventilation systems which handle flammable
vapors are contained in the National Fire Codes published by the National Fire
Protection Association and in the Handbook of Industrial Loss Prevention published
by the Factory Mutual System. These agencies recommend that the flammable vapors
be diluted with fresh air so that the concentration does not exceed 25% of the
lover explosive limit at the maximum evolution of flammable vapors. This is
especially important when a source of ignition such as an afterburner is located
at the end of the ventilation system.
Interlocks must be provided that will shutdown the afterburner and
vent the fumes in the event of failure of the ventilation fans. An air switch
alone is not an acceptable method of interlocking the fans in ventilation systems
where flammable \olatiles are handled because condensation or deposits can cause
the switch to become inoperative. Acceptable interlocking methods include rotational
switches on the fan shaft, extra contacts on the fan starter, relay contacts
operated in parallel with the fan motor, and powering the combustion safety
systems from the load side of the fan starter. This insures that the fan is
turning when the afterburner is operating.
There should be an automatically timed purge of the duct system with
fresh air prior to lighting the afterburner to remove any flammable vapors or
to evaporate any condensed material from the duct.
A vapor concentration indicator on streams with flammable vapors is
highly recommended regardless of the concentration of fumes. This instrument,
if properly installed and maintained, will give a continuous reading of the
concentration in terms of the per cent of the lower explosive limit. In addition
to indicating or recording the concentration to insure that safe conditions
are maintained, by adding alarm or control features, the instrument can be made
to control the vapor concentration by opening or closing dampers or to alarm
and/or shutdown the afterburner if the vapor concentration exceeds a safe value.
When a continuous vapor concentration indicator or controller is installed
and properly maintained, the concentration of flammable vapor in the inlet stream
can be increased up to 505? of the lower explosive limit. Any increase in the
concentration above this point should sound an alarm and shutdown the afterburner.
The use of vapor concentration controls on the dilution air can result
in considerable saving in fuel cost when the afterburner is used in service
where the vapor load varies. However, the measuring instrument and the controller
must be properly maintained and calibrated or a dangerous condition can develop.
This may require daily servicing in certain applications. If there is any doubt
that adequate maintenance can be provided, sufficient dilution air should be
provided to keep the vapor concentration below 25/fc of the lower explosive limit
at all times.
S-1M21
-------
138
In some special cases, flammable vapors may be generated with little
or no air present. In this case, the mixture would be above the upper explosive
limit and would not burn. Because of the chance of air leaking into a duct
and creating a hazard, the usual practice is to dilute the vapors below the
lower explosive limit at the source and to transport the diluted vapors to the
afterburner. However, if it can be assured that the vapors will always be above
the upper explosive limit and the local authorities and/or fire underwriters
agree, the vapors could be treated as a fuel source, mixed with air at the after-
burner and burned. A system such as this would require additional safeguards
such as flame arresters and blow out discs and possibly storage facilities to
even out the vapor flow. Any system which handles vapors that are or can be
made explosive or combustible by diluting with air should be approved by local
authorities and the fire underwriters.
9.3.2 Equipment
9.3.2.1 Vapor Concentration Indicator
Most currently used vapor concentration detectors operate on the catalytic
combustion principle. When a specially treated platinum filament is heated
and contacted with a mixture of air and a flammable gas or vapor, combustion
takes place at a temperature considerably lower than the normal ignition temperature.
The heat generated by the combustion is sensed either by a thermocouple attached
to the filament or by changes in the resistance of the filament itself.
The cell life of the detector measuring the resistance of the filament
is shorter and more subject to zero drift than the model using thermocouples
for temperature measurement. Therefore, the thermocouple type detector should
be preferred to the type measuring the resistance changes of the filament itself.
The sample system for the detector is a very important item. The
system must be constructed so that vapors do not condense before they reach
the detector and the sample flow must be great enough to give a rapid response.
The location of the sampling point must be selected to obtain a representative
sample. A well mixed location is necessary since most flammable vapors are
heavier than air and may stratify if the duct velocity is low.
A diffusion head analyzer is slightly different in the method of sample
handling. The diffusion head is located at the sample point and does not require
a sample pump or flow control system. Although large amounts of particulate
matter, moisture and dust can cause plugging, the diffusion head should be used
in clean, dry streams because of its simplicity and reliability.
To insure the safe operation of vapor concentration detectors, alarms
should be considered to warn in case of:
Filament failure
Power failure
Low sample flow (not available for diffusion heads)
S-14121 Figure 9-6 follows
-------
en
K>
Venf
Afterburner
A
Sut
—Hi
—4
ir
jply
1TT-
P-r
Ve
ft \
i
•• PSi
From
Combustion
Safeguard
System
1
i. t
•nt
t
\
Vapor
Concenti
Switch-
High
-------
X39
The cost of a vapor concentration monitoring system with sample system
is between $1500 and $2000. The addition of control features can add another
$1000.
9.3.2.2 Using the Afterburner Temperature Rise for Control
Split range temperature control is used to first turn down fuel to the
burner as the fume concentration rises to some limit, e.g.,. 25#-50$ of the LEL.
At this point the fume stream will account for a certain fraction of the AT across
the afterburner (~600°F). As fume concentration rises further, the controller
operates a damper to provide more air flow through the oven or other source of
fumes and prevents further concentration increase.
9.3.3 Typical Vapor Concentration Control System
The usual method of controlling the concentration of flammable vapors
is by means of a damper to admit air to the duct. This can be a continuously
adjustable damper if the fume load varies widely or a two position damper if
the load swings between two relatively constant flow rates such as in a batch
process. Figure 9-6 shows a typical arrangement. In addition to the control
features there should also be alarms and shutdowns to protect the system from
fan failure or high vapor concentration. In the event of a shutdown, the fumes
should be bypassed around the afterburner.
9.U Handling of Toxic Materials
9.^.1 Recommended Practice
All of the previous recommendations for handling flammable vapors
should be followed where applicable. However, when the material is toxic and
cannot be vented temporarily, additional precautions are required. Among the
precautions to be considered are additional and/or redundant alarms to warn
of possible trouble, a stand-by afterburner with automatic startup and switching,
and temporary storage to hold the toxic material until it can be burned. Emergency
shutdown of the fume generating process should also be provided. Particularly
in special cases such as this, the local authorities having Jurisdiction and
the insurance underwriters should be consulted.
S-1U121
-------
llu
Chapter 10. MEASURING AFTERBURNER PERFORMANCE-ANALYTICAL METHODS
10.1 Introduction
In order to assess performance of an afterburner, concentrations of
critical components at the entrance and exit of the afterburner must be determined.
Typically these are total hydrocarbons and carbon monoxide but may include
specific hydrocarbons, halogens, particulates, etc. As a bare minimum the
afterburner will have to meet certain standards imposed by the local air pollution
control agency, either emission standards for specific pollutants or a percentage
destruction in passing through the afterburner. Other standards may prohibit
any odor or visible plume. The operator of the fume incinerator will often be
responsible for demonstrating compliance with regulations, either by doing his
own testing or by hiring an outside laboratory.
If the afterburner does not provide the pollutant destruction required,
some investigation into interior performance of the unit must be made. Of
primary concern, for reasons discussed in Chapter 3, is measurement of two
parameters: 1) Temperature - Is the temperature experienced by the fume stream
high enough for the time the waste stream is in the afterburner? and 2) Flow
Patterns - Has good mixing been achieved? Is there bypassing or severe non-
uniformity of flows in the afterburner?
Certain measurements are also required before an afterburner is purchased.
It is important that fume stream flowrate and composition be determined, both
average values and the maximum variation. The amount and nature of entrained
material will also influence design of the unit for reasons previously discussed.
In this Chapter an attempt is made to indicate the state of the art
pertaining to the determination of a variety of air pollutants, measurement of
temperature and determining flow patterns. Problems encountered in sampling
and calibration are discussed. The methods available for determination of
individual compounds or classes are summarized. Table 10-1 shows the components
and concentration ranges which will be considered in this handbook. The high
limits on these ranges are estimates of entering fume concentrations experienced
in afterburner applications. The low values are based on estimates of stringent
emission standards which may have to be met. Considerably more material is given
for NOX and S02 analysis than for other pollutants. This is an indication of the
emphasis which has been placed on these components during the past decade.
In order to keep this Chapter to a reasonable size, no detailed pro-
cedures are included. References are given to sources of methods and appear in
the Bibliography. The following references are particularly recommended to give
an overall picture of the subject:
Jacobs10~25) and Leithe1^'26' have given good reviews covering the
subject of air pollutant analysis in general. Cooper and Rossano10"19' present
a useful reference book covering sampling and analysis for particulate matter.
Hanson et a!10~12' give the most complete and detailed procedures covering wet
Preceding page blank
S-14121
-------
chemical methods for determination of organic compounds in air. Ruch10~27) reviews
more compounds than Hanson, but the complete procedures are not given. Reference
10-U5 is a good basic reference on temperature and flow measurement.
Odorous emissions are generally complex mixtures at very low concentrations
which makes evaluation by chemical or instrumental methods difficult. Because of
this, the usual approach to the measurement of odor emissions is to determine
the magnitude of dilution required to obtain a concentration at which human Judges
can just detect the odor (the odor threshold). The strength of an odor is described
in "odor units" and one odor unit is defined as the amount of odor necessary to
contaminate 1 cubic foot of odor-free air to the threshold level. Odor measure-
ments are at best semi-quantitative, but they have been useful in Judging the
effects of odor control measures (principally afterburners). It has been esti-
mated10'30' that an odor strength of less than 150 odor units per scf in stack
gases will not cause community odor problems. It also appears that an emission
of up to one million odor units per minute from a plant will not cause an odor
nuisance.
10.2 Sampling Considerations
The collection of representative and meaningful samples is the most
basic aspect of all practical analysis. Samples must accurately represent the
conditions at the point of sampling and they must be preserved unchanged until
the analysis is made. Sample size and sampling frequency are important
considerations. Wherever possible, a continuous in-place analysis is preferable.
The flow rate, temperature, and pressure of the gas stream at the
point of sampling must also be measured for calculation of mass flow of the
components measured.
Sampling probes must be positioned with the above considerations in
mind. Steam-tracing or heating-tape-tracing of sample lines may be necessary to
prevent condensation of components at ambient temperature. Detailed recommen-
dations for sampling atmospheres for analysis of gases and vapors is given in
ASTM D l605-6010~1|. A method for sampling stacks for particulate matter is
ASTM D 2928-7110-1'. In the sampling of streams containing liquid or solid
particulate matter, sampling for these particles is particularly critical owing
to segregation according to particle size due to acceleration effects. Sampling
in this case should approach iso-kinetic conditions, i.e., the sample stream
at the sampling point and the main stream should be the same in linear flow
rate and direction. An excellent discussion of the errors associated with an-
isokinetic sampling is given by Lundgren and Calvert10"2). Sampling is discussed
further under 10.U.8 Particulate Matter.
10.3 Calibration of Instruments
Accurate calibration of monitors, especially in field use, presents
difficulties as with any gas measuring devices. The preparation of meaningful
standards has not been easy; a real advantage has become available through the
S-
-------
Table 10-1. COMPONENTS OF INTEREST FOR
MRASTTRTMr.
Components
Hydrocarbons
Total
Non-Methane
Paraffins
Olefins
Aromatics
Oxygenated Organics
C02
CO
Acids
Aldehydes
Particulates
Combustible
Non-combustible
Visible Emissions
Concentration
Range, ppm
10 -
10 -
10 -
10 -
10 -
0.5
10 -
1 -
1 -
•
20,000
20,000
20,000
20,000
20,000
- 10*a)
20,000
500
1000
0.01-10 grains /cu ft
Other Components
so2
SO 3
H2S, Mercaptans
Halogens
NOX
Water
Oxygen
5 -
5 -
1 -
5 -
5 -
0.1
0.1
20,000
1,000
10,000
10,000
5,000
- 10*a)
- 20*a>
a) Not a pollutant, but determination may be needed for control.
S-
-------
use of permeation tubes10-17»23'. These are Teflon tubes which contain the
liquefied substance for which the calibration is to "be made. Tubes are available
for sulfur dioxide, nitrogen dioxide, hydrogen sulfide, chlorine, hydrogen
fluoride, ammonia, carbon disulfide, carbon tetrachloride, acetaldehyde, several
mercaptans and hydrocarbons. Tubes can be made to order for many compounds
boiling above -70°F10~18 . Each will generate a low concentration of the
substance as a function of temperature.
A calibrated permeation tube containing the desired liquefied gas i->
maintained at a constant temperature while air flows over the tube at a fixed
rate. From the permeation rate at the fixed temperature and the flow rate of
the diluent gas the concentration of the test gas can be calculated. This
mixture of known concentration is used to calibrate the analyzer. A typical
calibration procedure as applied to sulfur dioxide is described in ASTM D
Several manufacturers (see Table 10-8) sell calibration equipment
for use with permeation tubes.
10.k Principles Involved in Analyses
The components of interest are discussed separately below. Under
each component, a brief discussion of the chemistry involved in the existing
analytical methods is given. Commercial instruments which have been developed
are listed. Wherever a choice of several approaches is available, an attempt
is made to recommend the optimum one for a given concentration range.
10.U.1 Hydrocarbons
Hydrocarbons in low concentration are measured by a flame ionization
detector (FID). The principle of the FID is based on the linear relationship
of ion formation to the concentration of a given organic compound in the flame.
The gaseous mixture containing hydrogen is burned at a small Jet with air or
cxygen and the change in electrical conductivity is measured. Response is a
function of the number and type of oxidizable carbon atoms in a molecule. The
relative response of a carbon atom follows the decreasing order -CH2-, -CHOH-,
-CO-. Relative response factors of some typical compounds are shown in Table 10-2.
In general, there is a small variation in the response of widely differing hydro-
carbon types.
Oxygenated organic compounds have a lower relative response, decreasing
with increasing oxygen content. C02 and CO, and inorganic gases show essentially
no response. Chlorinated hydrocarbons also are measured; response is dependent
on the number of CH bonds in the compound.
Federal regulations on ambient air quality standards are concerned with
"non-methane" hydrocarbons rather than total hydrocarbons. This is due to the
fact that in some areas appreciable concentrations of methane exist from natural
sources and furthermore, methane does not become involved in photochemical smog
reactions so it is not a problem.
S-14121
-------
Table 10-2.
RELATIVE REPONSE FACTORS FOR THE FLAME
IONIZATION DETECTOR1 °~3
Compound or Type
Relative Response Factor per
gram (n-heptane = 1.00)
Paraffinic hydrocarbons
methane
propane
n-pentane
n-octane
2,3> U-trimethylpentane
Cyclic hydrocarbons
eyelopentane
1,1 dlmethylcyclopentane
ethylcyclohexane
Olefinic hydrocarbons
acetylene
ethylene
hexene-1
Aromatic hydrocarbons
benzene
p-xylene
tert-butylbenzene
Oxygenated compounds
methanol
n-propyl alcohol
octyl alcohol
formaldehyde
acetaldehyde
acetone
di ethyl ether
n-propyl acetate
formic acid
acetic acid
hexanoic acid
Chlorinated compounds
chloroform
carbon tetrachloride
1,3 dichloropropane
tri chloroethy lene
t etrachloroethylene
Inorganic Gases
argon
nitrogen
oxygen
carbon monoxide
carbon dioxide
water
0.97
0.98
l.OU
0.9T
0.99
,OU
,03
1.01
1,
1.
07
02
0.99
1.12
1.00
1.02
0.23
0.61
0.85
.1
.3
O.U9
0.58
0.53
0.01
0.2U
0.
0.
0.63
0.082
O.OU5
0.37
0.23
0.19
essentially no response
S-1U121
-------
Instruments previously available measured total hydrocarbons
including methane. A new group of instruments are now becoming available
which will measure both total hydrocarbons and methane separately, thus
giving the desired non-methane hydrocarbon value by difference. The technique
can also determine carbon monoxide separately following its catalytic hydro-
genation to methane. A measured volume of air is sampled and the total hydrocarbon
content is measured with a flame ionization detector. An aliquot of the same
aii- sample is passed through a stripper column which retards all hydrocarbons
other than methane. The eluted methane is then measured by the flame ionizat-'on
detector. The stripper column is then automatically backflushed to prepare it for
subsequent samples on a semi-continuous basis.
The Los Angeles Air Pollution Control District describes analytical
methods for organic solvents and vaporslo~24). Total organic materials are
measured by a total combustion analyzer (TCA). In this equipment, a large sample
is passed through a dry-ice trap. The trap is then warmed through 2 temperature
levels so that low-boiling and later, high-boiling components are evaporated
into a carrier gas stream consisting of 5$ oxygen in nitrogen. The gas mixture
then passes through a catalytic oxidizer where organic compounds are converted
to COa, which is measured in an infrared analyzer. Two peaks for organics
result - one for low-boiling and the other for high-boiling materials.
Individual hydrocarbons are determined by gas chromatography using a
flame ionization detector. A typical method for determination of Cj to C5 hydro-
carbons is described in ASTM D 282010~1'. The lower limit of measurement of the
individual components is in the parts per billion range.
10.U.2 Oxygenated Organic Compounds, Carbon Monoxide, Carbon Dioxide
Since oxygenated organic compounds respond to the FID and are measured
as hydrocarbons, they may be included in the category of hydrocarbons. However,
in most instances, the oxygenates can be determined specifically by virtue of
their functional groups. These may be determined by absorption in suitable re-
agents followed by a titration, colorimetric, or spectroscopic measurement. For
example, carboxylic acids are readily absorbed in basic solution and determined
by titration. Aldehydes and ketones react with 2,U-dinitrophenylhydrazine to form
a hi-ghly colored product which can be measured spectrophotometrically. Table 10-3
lists a number of possible approaches which may be considered for several oxygenated
species which may be encountered.
10.U.3 Chlorinated Hydrocarbons, Phosgene, Hydrochloric Acid
The use of the flame ionization detector for measurement of chlorinated
hydrocarbons has been cited under "Hydrocarbons". In addition, other procedures
are available for determination of carbon tetrachloride, chlorobenzene, chloroform,
ethylene dichloride, tetrachloroethane, trichloroethylene, and ethylene chlorohydrin.
These procedures involve absorption in suitable solvents, color development, and
measurement of the color by spectrophotometric means.
S-11H21
-------
Table 10-3- METHODS FOR DETERMINATION OF OXYGENATES
AND OTHER COMPONENTS
Compound Determined
Aldehydes
Aldehydes and
Ke tones
Acrolein
Formaldehyde
Carboxylic Acids
Esters
Carbon Monoxide
Carbon Dioxide
Method
Methylbenzothiazolone
Hydrazone
Spec trophotome trie 650 nm)
Dinitrophenylhydrazine
Spec trophotome trie
4-Hexylresorc inol
Spectrophotometric £05 nnO
Chromo tropic Acid
Spectrophotometric (5 70 nm)
Absorption- titration
Hydroxamic Acid
Spectrophotometric 630 nm)
Non- dispersive Infra- Red
(NDIR) (see Table 10-10
NDIR (see Table 10-U)
Orsat.
Lower
Limit,
ppb
20
ko
20
20
100
1000
2000
0-5*
Absorptivity,
L/mol cm
50,000
27, 000
17,000
19,000
1,100
Reference
to, 5, 8
4e, 6
7, 8
ifd, 8
9
ke, 10
11
S-1U121
-------
148
Carbon tetrachloride, chloroform, ethylene dichloride, trichloroethylene,
tetrachloroethylene and ethylene chlorohydrin are all determined by a method in-
volving absorption in pyridine, reaction with aqueous caustic, and measurement of
the resulting color10"12*'. The intensity of the color varies with the compound
and the method should be calibrated against known quantities of the particular
compound being measured. In general the method is sensitive in the range
10 - 1000 ng of the compound.
Chlorobenzene is measured by a method which involves absorption ana
reaction with a solution of formaldehyde in concentrated sulfuric acid. The
yellow color produced is measured. The reaction is not specific for chlorobenzene,
as aromatic hydrocarbons in general are measured10"12^).
Upon passing through an afterburner, chlorine containing compounds
should produce hydrochloric acid as a major product and traces of phosgene. The
method most generally applied for hydrochloric acid involves absorption in aqueous
caustic, acidification and titration with silver nitrate11-120). Phosgene is
determined by a number of colorimetric procedures10-27'. In a typical pro-
cedure10"25', the air sample is drawn through a filter impregnated with ethyl-
hydroxyethylaniline and dimethylaminobenzaldehyde. Phosgene produces a blue color
whose intensity is proportional to concentration.
10.U.U Oxygen and Nitrogen
Oxygen and nitrogen content may be of interest for the purpose of
calculating excess air in the process. An Orsat analysis can be applied to
measure C02 and oxygen. Nitrogen is determined as the difference between 100
and the sum of the C02 and oxygen. The concentrations are in the 2 - 20% range,
where the Orsat equipment is of adequate accuracy, (see EPA method 2)10-11'.
Process analyzers are also available for oxygen and CO,.
Table 10-U.
ANALYZERS FOR 02 AND C02
Component
Instrument
Principle
CO.
Beckman C2
Eeckman Model
Teledyne 320P
Beckman IR-215
variety of concentra-
tion ranges is
available.
Paramagnetic
Electrochemical sensor
Non-dispersive Infra Red (NDIR)
S-14121
-------
10.U.5 Nitrogen Oxides (NOX)
A number of different oxides of nitrogen exist but only three are of
importance in the atmosphere. These are
N20 Nitrous Oxide
NO Nitric Oxide
N02 Nitrogen Dioxide
Since nitrous oxide is not toxic and it does not appear to take a
part in photochemical smog reactions, its determination is not of general concern
in air pollution problems.
NO is produced in appreciable amounts in combustion processes and
whenever air is heated to a high temperature. The reaction
N, + 0, < 2 NO
* 2 >
has an equilibrium constant which increases markedly at elevated temperatures.
The nitrogen oxides from high-temperature sources are predominantly (98 - 99%}
in the form of NO. In the presence of excess oxygen, NO reacts at a moderate rate
to produce N02, which is in equilibrium with the tetroxide (NgO^). At the low
concentrations at which the oxides of nitrogen exist in the normal atmosphere
the equilibrium
N20^ <"-> 2 N02
is almost entirely on the side of the dissociated species. Of the oxides of
nitrogen, N20 and NO are colorless, while the monomeric N02 is deep brown.
The concern regarding NO and N02 in air pollution involves three
areas:
i These oxides of nitrogen are chain carriers in the generally
accepted mechanism for photochemical smog formation,
ii The brown tinge of the Los Angeles atmosphere may be due to
the NO2 present, and
iii Limited tests indicate that N02 may be toxic at parts per
million levels.
The applicability and limitations of the principal methods for
the determination of NO are outlined in Table 10-5. The two chemical methods
have been used for some years and, in general give good agreement in the range
of 10 to UOOO ppm. That involving the reaction of phenol disulfonic acid to a
nitrated product (ASTM D 1608)10"1) is of moderate sensitivity and can be employed
for the determination of oxides of nitrogen, with the exception of nitrous oxide,
at levels above 5 ppm. This method is generally applied to samples taken in
S-
-------
Table 10-$. SUMMARY OF APPLICABLE METHODS FOR NO,
METHOD
PRINCIPLE OF PROCEDURE
INSTRUMENTATION
APPLICABILITY
LIMITATIONS
Phenol disulfonic
acid method
ASTM 0 1608
(Method 4 EPA)
NO
Sulfanilic Acid
method
(Saltzman)
*STM D 2012
SAE J177
Phenol disulfonic acid
nitrated to produce
yellow color.
N02 converts sulfanilic
acid to diazonium salt.
Salt couples with amine
to produce deep violet
color
Ultraviolet ab-
sorption
method
N02 has maximum
at 400 rap with
absorptivity « 170
liters/mole-cm.
NO is transparent above
230 mw.
Infrared absorp- NO has band at 5.4 u
tion method with absorptivity
« 2 liters/mole-cm.
Electrochemical
Sensor
NO*N02 permeate
membrane on sensor
and are electrochem-
ically oxidized. Re-
sulting current is
proportional to NO
concentration.
Chemiluminescence The light resulting from
the reaction of NO with
ozone is measured with
a photomultiplier.
must be converted
to NO to be measured.
Laboratory spect ro-
photometer at
400 mu
Laboratory spect ro-
photometers at
550 my.
Continuous analyzers:
Beckman Acralyzer.
Technicon Autc-
analyzer.
Continuous Analyzers:
Beckman NDUV K'odel
255 plus oxida-
tion system.
Continuous Analyzers:
Beckman NDIR Model
315A.
Mine Safety Appli-
ance Model LIRA 200.
Continuous Analyzers:
Dynasciences NX-110
and NX-130
EnviroMetrics Model
N-122
Theta Sensors Model
LS-800-ANX
Continuous Analysers:
Thermo-Electron Corp.
Bendix, Environnental
Science Division
REM, Inc. Model f.42
Range: 5 to 1000ppm
for all nitrogen oxides
except N20.
Range: 0.01 to 4000 ppm.
Specific for N02
NO determined by
prior oxidation.
Useful for air
and exhaust analysis.
Faster than disulfonic
acid method
Range: 10 to 6000 ppm.
Determines N02 directly.
NO determined by prior
oxidation.
Range: 10 to 4000 ppm.
Determines NO directly.
Range: 2 to 10000 ppm
Models available for NO
or N02
Not sensitive below
5 ppm.
Equipment somewhat
more complex than
for phenol disul-
fonic acid method
NO is a reactive
gas and can be
partially lost un-
less precautions
are taken.
Water vapor inter-
feres and must
either be con-
stant, or pre-
ferably removed.
S02 interferes but
can be eliminated
or compensated for.
Range: 0.01 to 10000 ppm No known interference*
-------
151
flasks which contain an oxidizing absorbent. The phenol disulfonic acid method is
used by a number of local control districts including the Los Angeles County Air
Pollution Control District1°-13>l"»).
The sulfanilic acid method, otherwise known as the Saltzman or
Griess-Saltzman method, is of considerably higher sensitivity than the phenol
disulfonic acid methods and is specific for nitrogen dioxide. It must be stressed
that NO, the predominant species present in incinerator effluent, is not determined
directly by the sulfanilic acid reaction and it must first be oxidized to N02 for
its determination. An SAE method for NOX in diesel exhaust10-15) involves taking
a sample into a syringe along with air and sulfanilic acid reagent. The NO is
oxidized by the air during a reaction period and the color formed is a measure
of the N02 plus NO originally present. ASTM D 201210-1), a method designed for
ambient concentrations, measures NO after oxidation to N02 by permanganate.
Spectrophotometric instrumentation is particularly useful for the
monitoring or analysis of the nitrogen oxides in combustion exhausts. The
infrared region is suitable for the determination of NO and this can be done
effectively with infrared instrumentation using a narrow band-pass non-dispersive
filter to provide discrimination against interfering components. Either the
Beckman NDIR 315A or the Mine Safety Appliance LIRA 200 offer feasible direct
measurement of NO in the range of 10 to UOOO ppm with minor interference except
for water vapor which must be kept constant or preferably eliminated.
In the ultraviolet and visible region, NO is transparent above 230 my
but NO2 has a strongly absorbing band with a maximum in the vicinity of UOO my.
The Beckman Model 255 Nitrogen Dioxide Analyzer is a nondispersive ultraviolet
(NDUV) spectrophotometer and is effective for the direct determination of N02
in effluents. For the application of the ultraviolet technique to the deter-
mination of NO or total nitrogen oxides, it is necessary to convert the NO into
N02 by oxidation before the gas sample is introduced into the sample cell. The
Beckman Model 255 (NDUV) Nitrogen Dioxide Analyzer, when used in conjunction
with the Beckman Model 315A (NDIR) Nitric Oxide Analyzer offers a direct, con-
tinuous system for the simultaneous measurement of NO and NO.. For the best
results both water vapor and hydrocarbons should be removed and this can be done
with cold traps and Drierite dehydrating agent without excessive loss of the
nitrogen oxides.
Instrumentation for measuring nitrogen oxides by means of electro-
chemical sensors has been available since 1969- The sensor consists of a
pair of electrodes immersed in an electrolyte. A permeable membrane covers
one electrode and encloses the cell. The NO (N0+N02) present in the sample
gas can permeate the membrane and be electrochemically oxidized at the applied
potential. The resulting current, which is proportional to the NO and N02
present, is amplified and read out on a meter or recorder. The NO sensor
also responds to S02 on an approximately equivalent basis. The S02 interference
can be removed by using a selective scrubber solution provided by Dynasciences,
or by electrical compensation provided by a second sensor for S02 with the
EnviroMetrics and Theta Sensors instruments. Carbon monoxide can interfere to
a slight extent if present in high amounts (about 1% or more).
S-1M21
-------
152
The chemiluminescence method for NO is quite new but appears to have
great promise. It is based on the reaction between NO and ozone which results
in the emission of light. The light is measured with a photomultiplier tube.
The method is very sensitive and can be used to determine low ambient levels of
nitrogen oxides. However, N02 must first be converted to NO by passing the
sample gas through a furnace unit. At present there are no known interferences
to the method.
10.U.6 Sulfur Dioxide
Since regulations limit the maximum concentrations of S02 in air to
the range of tenths of a part per million, sensitive and reliable analytical
methods and monitoring instruments have been developed for this determination.
Some of the important features of the principal methods in current use are
summarized in Table 6. Although there are a number of variations of these
methods, they are essentially based on four types of chemical reactions.
10.U.6.1 Hydrogen Peroxide Methods
S02 in intimate contact with dilute hydrogen peroxide solution is
dissolved and oxidized rapidly according to the reaction:
S02 + H202 > HgSO^
The resultant sulfuric acid can be titrated with base or the sulfate determined
by titration with barium perchlorate. These titration techniques find most
common use in the determination of the S02 concentration at higher levels than
is normally encountered in the atmosphere, such as in stack gases at levels of
200 to 3000 ppm. This approach is probably most useful for incinerator effluent.
For the determination of low concentrations of S02 in air, advantage
can be taken of the increase in conductivity of the hydrogen peroxide solution
due to the formation of sulfuric acid which is highly dissociated. Automatic
analyzers based on this principle have found use since they can provide continuous
results in the range of 0.01 to 5 ppm with little need for operator attendance.
The scrubber in the conductivity instrument is usually filled with a dilute
hydrogen peroxide solution containing sulfuric acid in order to avoid interference
from carbon dioxide. The conductivity instruments must be calibrated by procedures
based on primary standards. It is important to realize that the conductivity
method is not specific for S02 since other electrolyte-producing materials can
interfere. In making measurements near the ocean, salt aerosols have been found
to interfere; acidic and basic gases or aerosols which are likely contaminants in
some industrial areas, are also common interferences. The useful range of this
equipment, 0.01 to 5 ppm, may be too low for analysis of incinerator effluent.
10.>t.6.2 Coulometric Methods
The reactions involved in the coulometric methods are essentially
the following:
-------
Cn
Table 10-6. METHODS FOR S0s>
Method
Hydrogen Peroxide
Titration
Method
(EPA method 3)
Hydrogen Peroxide
Conduct ometric
Method
Electrolytic
(Coulometric)
Method
1IEST-GAEKE
Colorinetric
Method
ASTM D 2914
Electrochemical
Sensor
Principle
S02 • H202 -» H2S04
Titration with base or Ba
S02 * H202 -> H2S04
Measure Conductivity
S02 * Br2 * H20 -* 2HBr * H2S04
S02 * I2 * H20-> 2HI « H2S04
Br2 and I2 generated electro-
lytically.
Formation of dyestuff by
reaction with bleached
pararosani line.
Oxidation in a membrane-
covered eel 1
Instrumentation
Laboratory Equipment:
Absorber and titration units
Monitors:
Leeds J Northrup, AEROSCAN
Wosthoff U3S ULTRAGAS ANALYZER
Instruments Development, IDC 902-1
Scientific Instruments, SI -67
Monitors:
Consolidated, TITRILOG
Beckman Instruments, Model 906
Barton 286 SULFUR TITRATOR
Phillips Instruments, Model PW 700
Atlas Electric Devices, Model 1200
Laboratory Equipment:
Spectrophotometer
Monitors:
Atlas Electric Devices, Model 1500
Technicon Corp., AUTO ANALYZER
Monitors;
Dynasciences, SS-330
Envirometrics NS-200
Theta Sensors LS-800-AS
Applicability
Range: 0.01 to 100 ppm
Range: 0.01 to Sppn
Range: 0.01 to 5 ppm
Monitors simple to
operate and reliable
for unattended
service.
Range: 0.01 to 5 ppm
Most nearly specific
method for S02.
Range: 0 to 5000 ppm
Simple to operate
Limitations
Requires reagent
additions.
Interference by salt
aerosols and acidic
and basic gases
which may be eli-
minated by filters.
Interference by oxi-
dizing materials,
aldehydes, olefins,
and hydrogen sul-
fide. (Some inter-
ference can be
eliminated by
filters).
Procedure cumbersome.
Continuous analyzer
needs close
attention.
NO and N02 interfere
slightly.
IN
-------
15*
S02 + I2 (or Br2) —> 2HI (or 2HBr) + H2SOU.
Instruments are commercially available which operate on the principle of coulometric
generation of free halogen to react with the S02 removed by the cell from the gas
stream. The current generated to produce sufficient halogen to satisfy the
reaction above is a measure of the S02 content of the inlet gas stream. These
coulometric instruments can be used in the range of 0.01 to 5 ppm of S02 in the
atmosphere. Oxidizing substances such as chlorine, NO, and ozone will give low
results for S02, but the oxidants may be removed in some cases by a filter con-
taining ferrous sulfate crystals. Oxidizable compounds (other than SO,) such as
hydrogen sulfide, mereaptans or aldehydes, will give a positive error unless removed
by a suitable filter system. Some of the commercial instruments based on the coulo-
metric technique are finding extensive use as monitors since they are quite simple
to operate and require only minor attention.
10.U.6.3 Colorimetric Method
The method generally considered most specific for the determination of
S02 is that of West and Gaeke^0"16) which involves the initial fixation of S02
as a stable complex with sodium tetrachloromercurate. Addition of a mixture of
acid-bleached pararosaniline and formaldehyde to the complex produces the deep red
pararosaniline methyl sulfonic acid.
The detailed procedure published as ASTM D 291U10-1) for conducting the
West-Gaeke method is recommended for the determination of concentrations of S02
in the range of 0.003 to 5 ppm. Acidic or basic gases or solids (sulfur trioxide,
sulfuric acid, ammonia, or calcium oxide) cause no interference with the method,
but the gaseous oxides (ozone and N02) interfere if present in concentrations
greater than that of the S02. The interference of N02 can be eliminated by adding
a small quantity of sulfamic acid during the analysis. The interference due to
ozone can be reduced by allowing the sample solution to stand, and the ozone to
decompose, prior to analysis.
Although the West-Gaeke method, when employed with due precautions
concerning possible interfering materials, is considered the most specific for
the determination of low concentrations of S02 in the air, the procedure is
elaborate and the continuous analyzers available are cumbersome. Since the
method requires close attention, it is not as widely used as the conductivity
or the coulometric techniques.
10.U.6.U Electrochemical Sensor Method
The same instruments of this type used for determination of nitrogen
oxides can be used for the measurement of S02 by interchanging sensors or by
using a dual unit with sensors for both NOX and S02. The same type of cell is
used but the selectivity is altered by changing the applied potential, and perhaps
the electrolyte and membrane. The current resulting from the electrochemical oxi-
dation of the S02 is proportional to the S02 concentration. NO may give a slight
positive interference while N02 may give a negative response. The extent of the
interferences will vary some depending on manufacturer.
S-1U121
-------
155
10.U.7 Sulfur Trioxide. Sulfuric Acid
Any combustion process in which a sulfur containing fuel is used will
usually produce some sulfur trioxide along with the sulfur dioxide. The SOa
may range from 0.555 or less up to about 1055 of the total sulfur oxides emitted
and the concentration may range from about 1 ppm up to 50 ppm or more.
Many methods have been published for the determination of 803 in the
presence of S02- These usually consist of separating the SO, and the S02
and then determining the resulting acidity or sulfate ion from each. The two
most widely used techniques for separating the S03 and S02 are based on
absorption 10-l>*,«t3) and condensation1O-"*1*). In the one case, the sample gas
is passed through a set of absorbers in which the 803 is first removed by an
isopropyl alcohol-water solution and the S02 then removed by an aqueous hydrogen
peroxide solution. Some S02 dissolves in the first IPA-HzO scrubber solution
and must be subsequently purged out with nitrogen to transfer it all into
the peroxide scrubber. The IPA inhibits the oxidation of S02 to S03 before the
separation is completed. In the second case, advantage is taken of the fact
that as the hot combustion gases are cooled down, the SO, combines with water
vapor and becomes a sulfuric acid mist. This sulfuric acid mist is then condensed
out at a temperature below its dew point, but above the dew point of water vapor.
The SO3 is thereby separated from the S02 which passes through the condenser to
a peroxide scrubber.
The solutions obtained following the separation of SO3 and S02 may be
titrated for acidity with standard base if no other acidic interferences are
present. More frequently the resulting sulfate ion is determined by titration
with barium ion using thorin indicator, or colorimetrically following reaction with
barium chloranilate.
At present there are no commercially available instrumental methods for
determining SO and it must be determined by the manual wet chemical methods
outlined above.
10.U.8 Particulate Matter
Particulate matter can be emitted from incinerators for at least
three reasons: combustion is incomplete and soot is formed, some of the feed
to the incinerator escapes combustion and appears as a condensible vapor in
the stack gas, or the fume contains inorganics which appear as oxides. In
any event, the effect of a significant amount of particulate matter in the
incinerator stack gas is the formation of a visible plume. In most air pollution
control districts, plumes above a certain opacity are illegal and steps must be
taken to prevent their formation. In addition to compliance with control district
regulations, there are sometimes other reasons for controlling particulate matter
emissions. These include the use of soot formation as a guide to controlling
the combustion process, and the prevention of the loss of economically valuable
ash products. The first step in such control is the estimation of the amount of
particulate matter in the stack gas. Cooper and Rossano10-19) review the methods
for this purpose and cite 15 references to methods used by regulatory agencies.
-------
156
The most easily applied test for the amount of particulate matter in
a stack gas emission is the visual evaluation of the plume opacity. It has been
found to be possible to train observers to estimate the plume opacity to within
10 per cent light transmission. The unit commonly used in this field is the
Ringelmann Number. The appearance of the plume is compared with a series of
standards prepared by ruling a white surface with sets of lines of increasing
width. In the Ringelmann standard, 20 per cent of the surface is covered by
black lines, in Ringelmann 2, Uo per cent, and so on up to Ringelmann 5 which is
completely black.
In the basic method for the determination of Ringelmann Number, large
cards ruled in the manner described are placed about 50 feet away from the viewer
and in the same field of view as the plume. The lines then fuse into fields of
grey of different density. The number assigned to the plume is the number of the
standard which most closely matches the plume in appearance. The basic method
is often impractical due to restricted access and the need for viewing the plume
from several vantage points. Consequently, control districts often train their
observers to evaluate the plume opacity from memory. Miniature cards are
available as guides for occasional observers but are no substitute for training
by the evaluation of standardized plumes. Such training becomes especially
important in the evaluation of white or colored plumes.
The accurate sampling of incinerator stack gases for particulate matter
content is a difficult Job. The difficulties arise from the nature of the medium
which is sampled. It is usually hot (over 1000°F), wet, and moving at a velocity
of one to several thousand feet per minute. Furthermore, the composition of the
gas sometimes varies from point to point in the duct and with time. The con-
centrations of interest are often low (.01 to .1 grains/cubic foot) and sometimes
the gas contains condensible vapors other than water which form particles on
cooling in the ambient air. For these reasons, there are no simple techniques
which give reliable particulate matter concentrations in all circumstances. The
methods used by air control districts generally require a day or so for a skilled
two man team to test an installation thoroughly. The method outlined in the next
paragraph is typical.
S-14121
-------
157
The apparatus used in the test is shown in Figure 10-1 and is the one
described by Smith10"20'. During the test, sample gas is withdrawn through the
train at a rate such that the velocity of the gas through the nozzle matches that
of the gas in the duct. In order to achieve this rate, the duct velocity is
measured with the aid of the Pitot tube. Calculations made with the aid of a
nomograph indicate the pressure drop to be established across the orifice meter.
Particles existing as such at the stack conditions are collected in the cyclone
and filter. Water and other condensible vapors are collected in the impingers.
During a test, the nozzle is moved to several preselected locations in the
duct to ensure the collection of a representative sample. At the end of the
test, the particles deposited in the probe are combined with those in the cyclone
and filter and weighed. The impinger contents are evaporated to dryness at about
220°F and weighed also. The particulate matter concentration is calculated
from the sum of the weights of material, both in the dry collectors and probe
and in the impingers.
Accurate measurement of particulate matter content is probably impractical
for any purpose except compliance testing. However, less rigorous techniques can
give information which is useful for control purposes. These techniques use
monitors to provide automatic records of some property of the stack gas related
to the particulate matter burden. Cooper and Rossano10~19' discuss the state of
the art here and point out that there.are two main approaches to particle monitoring.
The first employs the attenuation of the intensity of a light beam through the
stack and the second employs tape samplers.
The elements of an in-stack smoke photometer are shown in Figure 10-210"21).
They consist of a light source and photocell mounted on opposite sides of the stack.
The windows in the light path are kept clean by streams of fresh air directed over
them. Output from the photocell is directed to a readout device which can be an
indicator, recorder or alarm. This kind of unit is widely used in large power
generating stations and other combustion installations. Because the light
absorption depends upon particle size as well as concentration, the smoke photo-
meter record is related more to Ringelmann number than to weight concentration
of particulate matter. Moreover, vapors which would be counted as particulate
matter in a wet collection method such as that described above, would not be seen
by these photometers. However, these units do not require special sampling methods
nor is the temperature or water content of the stack gas a problem.
Another monitoring technique uses the tape sampler. Gas is drawn through
a filter paper tape at a fixed rate for a given time. Then the tape is automatically
advanced to make the spot visible and another sample taken. Models of sampler are
available which measure the light transmittance of the spot right after the tape
advance and print out the value on paper tape. Tape sampling requires more attention
to the means used to convey the stack gas to analysis point than do the photometers.
A sample is withdrawn through a probe which must be located properly and at a rate
which corresponds to the velocity of the stack gas if a representative sample is to
be taken. When the gas is hot or wet, a known amount of dilution gas must be added
to bring the mixture to ambient conditions without precipitating the water. Tape
samplers with optical readers are subject to changes in output signal introduced by
changes in the color of the particle or anything else which affects the optical
S-1U121
-------
158
density of the spot on the tape. The complications introduced by the sampling
problem and the intermittent nature of the output make the tape sampler less
desirable than the photometer for most installations.
A portable opacity meter is manufactured by Shell Development Company
for the measurement of stack plumes. The method is based on using the sun as a
light source and the measurement is expressed as Ringelmann number or as percent
transmissions. The instrument is adjusted to read 100$ transmission in un-obstructed
sunlight. Then the observer places himself so that the plume is between him and
the sun and a reading of the sun obscured by the stack plume is made. The light
path of the instrument contains optical filters which produce an instrumental spectral
response closely approximating that of the human eye. A silicon photodiode con-
verts the light to an electrical signal which is amplified and displayed on an
internal meter. Since the legal constraints placed on plume opacity are based on
visual effects, any instrumental technique should measure the same quantities.
The instrument can be used under most weather conditions, except when the sun is
obscured or when moving clouds change the available sunlight during calibration or
measurement. Under normal conditions the instrument's precision is ±0.25
Ringelmann number, markedly better than the trained human eye which usually can-
not discriminate beyond ±0.50 Ringelmann number.
There are several good references available in the field of particulate
matter sampling. The most complete and the most recently published of them is the
work by Cooper and Rossano1°~19). They provide a comprehensive guide to the
literature, together with a good summary of the corpus of knowledge on this subject.
Particularly valuable to the man in the field is the collection of numerical data
and computation methods. Somewhat older are Chapter 28 and 29 in the work edited
by Stern10~9'. These cover much of the same ground as do Cooper and Rossano but
with less depth and less practical information. The methods used by the Los Angeles
Air Pollution Control District are described by Devorkin et al10~22). A recently
published ASTM method, D2928-7110~1', describes methods for measuring particulate
matter concentration.
10.U.9 Source Measurement of Odor Emissions
Odor is a physiological response of individuals, and it is because of
this response that odor complaints are made. There is no instrument that parallels
the nose in the observation of odors and because of this all measurements of odor
emission ultimately rely on observations made by the human nose.
Odorous emissions are generally complex mixtures at very low concentrations
which makes evaluation by chemical or instrumental methods very difficult. Because
of this the usual approach to the measurement of odor emissions is to measure odor
intensity by determining the magnitude of dilutions required to obtain a concentration
at which human Judges can Just detect the odor (the odor threshold). The techniques
to accomplish this will make up the main part of the discussion which follows. Some
note will also be made of progress toward physico-chemical techniques for the
S-1^121 Figures 10-1 and 10-2 follow
-------
iV
Heated
Equipment
Probe
IRgl'
Pitot
Tube
Stack
Wall
L -J
Cyclone
Differential
Manometer
Orifice
Meter
"^-SrSr1'
y
Differential
Manometer
Dry Test
Meter
Greenberg-Smith Impingers
Vacuum Gage
Pump
Spray
Droplet
Filter
Figure 10-1. PARTICULATE MATTER SAMPLING TRAIN
S-U121
67784
-------
Receiver
Optical Extensions
to Reduce Smoke Path Length
Chimney or Duct
Projector
Alarm Bell
or Light
71 Alarm Supply
Connection
Control Unit
and
Photocell Amplifier
Mains
°' Connection
Recorder
Figure 10-2. 1N-STACK SMOKE PHOTOMETER
S-14121
67784
-------
159
determination and measurement of odorants. No consideration will be given to the
determination of odor quality or objectionability since this in itself is a very
complex subject, and usually the goal in industrial odor control is the removal of
the odor whether it is Judged pleasant or unpleasant.
10.U.9.1 Measurement of Odor Intensity Using Dilution
There are two main concerns in the measurement of odor emissions: how
serious is a given odor problem, and how effective have control measures been.
To date the relative strength of an odor is generally determined by dilutions. The
greater number of dilutions with odor-free air to bring an odorous gas sample to
the threshold, or barely detectable, level, the stronger the odor. The strength of
the odor is described in "odor units" and an odor unit is defined as the amount of
odor necessary to contaminate 1 cubic foot of odor-free air to the threshold level.
The number of odor units of a given sample represent the dilution with odor-free
air necessary to bring the sample to the threshold level.
10.If.9.2 Sample Collection and Dilution
Samples of odorous gases are collected from the source, usually
through a glass probe into a glass, plastic, or metal container. Where conden-
sation is a problem the samples are diluted with dry, odor-free air during
collection.
Once the sample is collected it is diluted to threshold level as
observed by an odor panel. Many authors have described the equipment and
techniques used in these operations.10'28"40'
10.^.9.3 The Odor Panel-Determining the Odor Threshold
The number of people required on an odor panel has varied from 2 to 8.
The training and use of an odor panel is at best cumbersome and expensive and
there is always a tendency to use as few people as possible. The Air and
Industrial Hygiene Laboratory of the State of California Department of Public
Health describes an apparatus which may allow good measurements with only
2 observers.10'34) According to ASTM Standard Dpl391-57lo~29) results that
are reproducible within ^0$ may be achieved with a single observer.
While these techniques are at best semi-quantitative, and more study
is necessary to determine losses during dilutions, they have proved of value in
judging the effects of various odor control measures (principally after-
burners)10-30'33/. Benforado10-30) estimates that an odor strength of less than
150 odor units per scf in stack gases will not cause community odor problems. It
also appears that an emission of up to about 1,000,000 odor units per minute
from a plant will not cause an odor nuisance.
10.1*.9«** Possibilities for Instrumental Analysis
While the above dilution to threshold techniques can give meaningful
measurements, they are not amenable to automatic monitoring. Techniques that could
identify which of the pollutants are key odorants might supply information that
S-
-------
i6o
would be useful in the design of monitors that would indicate effectiveness of
odor control. One approach to this problem is to combine gas chromatography of
the pollutants with sensory evaluation of the emerging resolved materials. In
the study of diesel fuel exhausts, Dravnieks10"1*2' has made use of a GC equipped
with a sniffing port at the outlet to identify key odorants. This technique is
only in the development stage and is not ready for routine application, but it
does offer promise for the future.
10.5 On-Site Instrumental Analysis
The present trend in analysis of gas streams is toward instrumental
methods. Instruments are presently available or can be adapted for analysis of
most of the components of interest. Table 10-7 summarizes the types of commercial
equipment currently available.
10.6 Temperature Measurement
Measurement of temperature can be divided into four basic techniques:
l) Thermoelectric thermometry (thermocouple technics)
2) Resistance thermometry
3) Radiation and Optical Pyrometry
U) Other technics used in flame research
The first two are primarily used to study an afterburner and the last two will
only briefly be described. A good basic reference on temperature and pressure
measurement is Fundamentals of Temperature. Pressure and Flow Measurement by
Robert P. Benedict. *»-•»»)
10.6.1 Thermoelectric Thermometry
Thermocouples rely on the principle that Junctions of two dissimilar
metals generate an electrical potential which is a function of temperature
(Seebeck Effect). When a closed circuit is formed with one Junction at a
reference temperature (usually ice water at room temperature) and the other at
the temperature of interest the resulting potential difference can be measured
and used to indicate temperature. 10~1*5^ Some of the most commonly used thermo-
couple materials are shown in Table 10-8.
S-
-------
CO
Table 10-7. SUMMARY OF ANALYZERS AMD MONITORS
Purpose
Manufacturer
Model
Pricea
Address
SO,
•Consolidated Electrodynamics
•Beckman Instruments
*Barton
•Philips Electronic Instruments
*Atlas Electric Devices
*Atlas Electric Devices
•Leeds and Northrup
•Calibrated Instruments
•Instruments Development Co.
•Scientific Instruments
•Technicon Corporation
Dynasciences Corporation
Dynasciences Corporation
EnviroMetrics
Theta Sensors
Titrilog
Model 906
286
PW9700
1200
1500
Aeros can
Wosthoff U3S
IDC 902-1
SI-67
Auto-AnalyzerlV
SS-130
SS-330
S-6U
LS-800-AS
$2900
$5^30
$2000
$1900
$1150
Pasadena, California
Fullerton, California 92631*
Monterey Park, California
91751*
Mount Vernon, New York 10550
Chicago, Illinois 60613
Chicago, Illinois 60613
Philadelphia, Pennsylvania
New York, New York 10023
Reston, Virginia
Hampstead, New York
Tarreytown, New York 10025
Chatsworth, California 91311
Chatsworth, California 91311
Marina Del Rey, California
90291
Orange, California
NO.
•Beckman Instruments
•Technicon Corporation
Mine Safety Appliance
Dynasciences Corporation
EnviroMetrics
Theta Sensors
Thermo-Electron Corporation
Bendix Environmental Science
Division
REM. Incorporated
IR-315A
Acralyzer
NDUV-25S
Auto-AnalyzerlV
LIRA-200
NX-110, NX-130
N-76
LS-800-ANX
6U2
$5^30
$2500
$2000
$1900
$1150
$6500
Fullerton, California 9263k
Tarrytown, New York 10025
Pittsburgh, Pennsylvania 15208
Chatsworth, California 91311
Marina Del Rey, California
90291
Orange, California
Waltham, Massachusetts
Baltimore, Maryland
Santa Monica, California
a) Approximate, 1970.
CA
-------
CO
Table 10-7. SUMMARY OF ANALYZERS AND MONITORS (Contd)
rv>
Purpose
Manufacturer
Model
Price
Address
Oxygenates
CO
CO 2
Hydrocarbons,
Methane, and
CO
Particulates
Calibration
Devices
Mine Safety Appliances Co.
Bendix
Beckman Instrument Co.
Mine Safety Appliances Co.
Beckman Instrument Co.
Beckman Instrument Co.
Teledyne
Beckman Instrument Co.
Bendix Corp.
Mine Safety Appliances Co.
Tracer, Inc.
Union Carbide
Monsanto Enviro-Chem
Byron Instrument Co.
Bendix Corp. Environmental
Science Division
Shell Development Co.
Analytical Instrument
Development, Inc.
Kin-Tek Laboratories
Tracer Analytical Instruments
Lira Model 200
Sensor Model 902
IR-215
Lira Model 200
CZ
715
320P
6800
Environmental
Chromatograph
350F
3020
Tape sampler-
Recorder 23605
Opacity Meter
305
670
$2500
$ 575
$2UOO
$2500
$1000
$ U70
$ 835
$7000
$1212
$ 995
$ 725
$ 865
$2000
Pittsburgh, Pennsylvania 15208
Fall River, Massachusetts
02722
Fuller-ton, California 92631*
Pittsburgh, Pennsylvania 15208
Fullerton, California 92631*
Fullerton, California 92631*
San Gabriel, California 91776
Fullerton, California 92631*
Ronceverte, W. Virginia 21*970
Pittsburgh, Pennsylvania 15208
Austin, Texas 78721
White Plains, New York
Dayton, Ohio
Raleigh, North Carolina
Baltimore, Maryland 2120U
Oakland, California 9U623
West Chester, Pennsylvania
19380
Texas City, Texas 77590
Austin, Texas 78721
* These instruments are primarily for ambient air concentrations
-------
163
Table 10-8
Thermocouple Material
Iron-Constantan
Chromel-Alumel
Platinum-Platinum/Rhodium
Tungsten 5* Rhenium-
Tungsten 26# Rhenium
Temp. Range
-300°F - 1390°F
0°F - 2l*90°F
32°F - 2990°F
To UOOO°F
* Requires inert gas atmosphere
All these thermocouples are available commercially in a variety of sizes,
from microprobes to industrial sizes and in many sheath materials as shown in
Table 10-9.
Table 10-9
Sheath Material
Notes
Melting
Temperature
Stainless Steel UU6
Inconel* X
Inconel* 702
Aluminum (type 606l)
Copper
Hastelloyt C
Hastelloyt X
Haynest 25
Monel"
Nickel
Platinum
Platinum-lOJJ Rhodium
Platinum-20$S Rhodium
Tantalum
Often used in preference to Chromium-Nickel
alloy in sulfur bearing atmospheres
Nickel-Chromium alloy, age hardenable
Addition of Aluminum increase oxidation
resistance
Alloy, principally Nickel, Chromium and
Molybdenum
Alloy, principally Nickel, Chromium Iron
and Molybdenum
Cobalt case Alloy
Nickel-Copper Alloy
Addition of Rhodium increases service
temperatures, strength & corrosion
resistance
Useful to UOOO°F in inert gas environ-
ments or in a vacuum
2600 to 2750°F
25UO to 2600°F
2UOO°F to 2555°F
1080 to 1200°F
198l°F
2310 to 2U50°F
2350°F
2U25 to 2570°F
2370 to 2U60°F
2615 to 2635°F
3217°F
3370°F approx.
3l*50°F approx.
T.M. International Nickel Co., Inc. t. Union Carbide Corp.
S-
-------
It should be remembered that a thermocouple inserted into a moving
gas stream does not measure the static temperature of the gas nor the total
or stagnation temperature but somewhere in between. However for most afterburner
systems where the velocities are low (less than 100 ft/sec) the difference
between the total and static will be less than 1°F. Of far greater importance
is the effect of radiation primarily and secondarily stem conduction. Basically
the temperature that the probe assumes will be due to a heat balance between
convection and radiation to the probe and conduction and radiation away from the
probe. The influence and importance of each of these will depend on the following
factors:
a) Probe size and shape
b) Gas velocity
c) Probe thermal conductivity
d) Emissivity of the probe and surrounding
e) Fluid properties including gas emissivity and absorption
The trends of these various effects are presented in the literature
as for example, in Reference 10-U5 and 10-U6. Typically for an unshielded thermo-
couple in a furnace with a gas temperature of around 2000°F and cold walls at 80°F
the temperature of the probe could be as much as UO°F cooler than the gas. If the
probe "sees" the flame the probe temperature could be as much as 200°F higher than
the gas temperature. Thus either the probe should be designed to significantly
reduce the effect of radiation or the probe should be corrected for radiation. It
should be remembered too that any probe either for velocity or temperature measure-
ment is going to disturb the system. Thus probes should be made as small as is
economically possible. A properly designed probe can yield temperature measurements
as accurate as ±1°F
A great variety of probe types have been developed of which several are
shown in Figure 10-3-
a) Half shielded - this probe of the three is least sensitive to radiation,
Mach and Reynolds Number effects.
b) Pencil type - most sensitive to radiation, Mach and Reynolds Number effects.
c) Bare Wire - intermediate between the above two.
These probes can also be built in the form of rakes8' for temperature traverses.
In addition some special probes have been developed for use in highly
radiative regions. These are known as suction pyrometers and pneumatic pyrometers.
a) A rake is assembly of probes spaced at regular intervals.
S-Ik 121 Figure 10-3 follows
-------
A. Half-Shielded
C. Bare
THE LAND SUCTION PYROMETER TYPE 4
fVAT. PLAT IJ%RHOO
THERMOCOUPLE
1 J- M.P.
SUCTION
CONNECTION
STANDARD MEL BOX
CONTAINING SPARE
THERMOCOUPLE WIRE
D. Suction Pyrometer
COURTESY •. Land Instruments
RESISTANCE
THERMOMETER
(Tc)
OPERATING
LENGTH
'HOT'
VENTURI
EXPANSION
BELLOWS
-
WATER
CONNECTIONS
PROTECTION
TUBE
'COLD-
VENTURI
(APC)
COURTESY: Land Instruments
E. Section of Land Venturi Pneumatic Pyrometer, Type VPP15
Figure 10-3. THERMOELECTRIC PROBES //
S-14121
67784
-------
165
A suction pyrometer (see Fig. 10-3d) is a device that completely shields the
thermocouple junction from radiation and draws the hot gases over the junction
at velocities in excess of 150 ft/sec. Accuracies of 1-1/2% to 2% are typical.
Some of these instruments are water cooled and typically operate to 2900°F. They
are available commercially.
To make measurements in the flame region to UOOO°F pneumatic pyrometers
have been developed. Pneumatic probes are generally of two types, venturi pneumatic
pyrometers feee Fig. 10-3d) and sonic orifice pneumatic probes. Both of these probes
rely on the thermodynamic relationships of gas flow across a series of sonic
orifices or venturies. In the venturi probe (see Fig. 10-3e) aspirated hot gas
density is measured by a front venturi restriction; after being cooled by passage
down the water cooled probe, the density is measured by a second venturi and the
cool gas temperature is measured by a thermocouple or resistance thermometer.
Generally these devices are used where high temperature, dirty gas conditions or
faster response requirements would prohibit the use of suction pyrometers. How-
ever, care must be exercised if a double sonic orifice is used in a dirty environ-
ment since corrosion or deposition can take place. It has also been reported
that it is difficult to interpret the results when there are rapid fluctuations
in pressure, temperature and velocity.10~U7) Again these probes are available
commercially and can be made to perform traverses either radially or axially.
In summation the following advantages and disadvantages may be listed
for thermoelectric thermometry:
Advantages Disadvantages
a) High precision characteristics a) Radiation error correction needed
of electrical measurement (Can use null technique to eliminate
radiation)
b) Extremely small thermocouple can
be made b) Spatial resolution limited to wire
size
c) Can be made of materials to with-
stand high temperatures. c) Probe can generate disturbances due
to vibration or catalysis (The latter
may be eliminated by coating the T/C
with silica10-1*6/
10.6.2 Resistance Thermometry
Resistance thermometry relies on the principle that a wire's resistance
changes in a known manner with temperature. Basically there are three characteristics
of these devices: 1) Simplicity of the circuit (no reference Junction is needed)
2) Sensitivity of measurement and 3) Stability of the sensors. There are two
basic types of sensors, resistance temperature detectors (RTD's) and thermistors.
RTD's are solid wire probes, platinum, nickel or copper typically and have a
positive coefficient of resistivity. They have a temperature limit of 1000°F.
Thermistors which have a negative coefficient of resistivity are electrical circuit
elements formed from solid semi-conducting materials. They are quite accurate,
S-14121
-------
166
±0.1°F but are limited to a temperature range of 32°F - 600°F. These temperature
ranges are rather low for direct application in an afterburner but the elements
are used as secondary elements in the pneumatic pyrometers (Section 10.6.1) for
greater accuracy.
10.6.3 Radiation Pyrometers
There are two classes of radiation pyrometers, total radiation and
optical pyrometers. Total radiation pyrometers measure the total radiation
emitted by a hot body. Optical pyrometers compare the brightness of the hot
body in a narrow frequency or color band with the brightness of a standarized
radiation source. Typically these could be used to measure the refractory
surface temperatures. Many instruments in most any price range and accurracy are
available commercially. Table 10-10 shows a typical relation of color to tempera-
ture of iron or steel.
Table 10-10. RELATION OF COLOR TO TEMPERATURE OF IRON OR STEEL
Dark blood red, black red. ••
Dark cherry red
Orange , free scaling heat
Yellow
White
Deg F
990
1050
1175
1250
1375
1550
1650
1725
1825
1975
2200
10.6.1* Other Techniques
Many other sophisticated techniques have been used for detailed flame
studies such as the sodium line reversal method, spectral line intensity methods,
and temperatures derived from measurement of gas density gradients. The reader
is referred to the book by Fristrom and Westenberg (Reference 10-U6) for the
details of these techniques.
10.7 Flow Pattern Determination
Two approaches may be taken in attempting to determine flow patterns
in an afterburner. One can study the full scale system either hot or cold or
one can study a geometrically similar model.
S-1U121
-------
167
10.7.1 Use of Models
Typically, a plastic model would be employed with water or air as the
fluid in order to simulate afterburner flow patterns. Care must be exercised in
modeling to insure that the important parameters are held constant. That is, the
experimenter must decide whether velocity ratio, Reynolds number, density ratio
and/or some other dimensionless groups are the most important characterizing para-
meters. A good discussion of this is given in Reference 10-U9. By using a
model, considerable time and expense can be saved and a good qualitative feel for
the flow patterns can quickly be ascertained through visual techniques. Quantita-
tive methods for determining flow patterns are also available and can be used with
either cold models or full scale systems.
Water models have been very popular for simulating furnaces and combustors
because of the ease of pumping water and the many qualitative and quantitative
visual techniques which can be used to determine flow patterns (Section 10.7*3).
It should be remembered though that with a water model, an incompressible fluid is
being used to simulate compressible fluid flow. Mach number effects and expansion
due to heating cannot be adequately simulated. However, in afterburners velocities
are generally quite low and compressibility effects will not be too important.
Also, good correlation has been obtained between water model studies and observa-
tions in operating gas turbine combustors10"50', industrial furnaces and domestic
heating equipment.
10.7.2 Quantitative Methods
Velocity measurements are usually made using pitot tubes or hot wire
anemometry. The latter method is only suitable for cold studies. For measurements
in highly turbulent flows, a straight pitot tube is difficult to use successfully
since the orientation must be within a few degrees of the flow direction. Thus,
a variety of multihole probes have been developed to measure velocity and flow
direction in either two or three dimensional systems. Examples of 2-D probes are
claw probes, three hole yaw probes, cobra probes and wedge probes (See Fig. 10-Ua-dX
The procedure is simply to balance the two outer pressure taps to orient the probe
and then make the velocity measurement through the center tap. Suitable methods
of measuring the angle to some reference line and traversing micrometer must be
employed in order to map the flow chamber. Three directional probes are the
Four or Five Hole Yaw Probe. Spherical probe, five hole directional probe and
3-D cobra (See Fig. 10-5a-c). All of these are described in detail in Reference 10-51.
Many of these probes can be cooled for use in hot systems and are available
commercially.
Hot wire anemometry can also be used for velocity measurement, flow
direction and turbulence intensity and is particularly good in measuring unsteady
flows. It is basically an indicator of rate of heat transfer from a hot wire to
the flowing system. The reader is referred to Chapter VI of Aerodynamic
Measurements, Reference 10-51 for a very complete discussion of the fundamentals
of hot wire anemometry. Hot wire anemometry has primarily been used to study
boundary layers and turbulence. See also Reference 10-52. A few tracer materials
have been used successfully such as helium or hot air injection where the concentration
8-11*121
-------
166
levels can be mapped with a thermal conductivity cell. These concentration levels
then give an indication of the amount of mixedness and indicate flow patterns.
These techniques are somewhat limited if the system is in an unsteady state as
the thermal conductivity cells will indicate time averaged values. In addition
if the system is highly turbulent with a great deal of back mixing it may be diffi-
cult to obtain a good concentration map. Radiosotopes have been used in a similar
manner.
Many of the visual techniques described in Section 10.7•3 can be used in
conjunction with photography to yield quantitative information.
10.7.3 Visual Methods
Visual methods in conjunction with air or water models are probably the
least expensive and fastest way to obtain a great deal of qualitative information.
(As is discussed below, it is also possible to get some quantitative information.)
Design changes can be easily made on a model and visual observations made to
assess its effect on flow patterns. Visual methods may also be used in the full
scale apparatus, either cold or at operating temperatures. Visual techniques may
be divided into two categories, static methods and kinetic methods.
10.7.3.1 Static Methods
Static techniques are most often used for boundary layer flow visualization
and may involve a surface coating (oil paint, dye, or fluorescent oil film10"53')
or tufts attached to the wall for directional measurements (regions of reverse flow
may be detected)10'51*). This gives a general picture of flows in the region near
the surface under observation but gives little if any information about disturbance
in the main flow. For this reason, static methods would not be used for afterburner
flow pattern studies.
10.7.3.2 Kinetic Methods
The most direct method for flow visualization is with suspended micro-
scopic dust particles. Flow is made visible in a cross section by a high-intensity,
narrow plane beam of light cutting a slice across the flow path and can be viewed
or photographed at right angle to the plane beam. Table 10-11 lists some of the
common tracers for air or water systems. For good photographic records the particles
should be as large as possible and for hydrodynamic reasons as small as possible.
In addition the tracers should be as near as to a neutral density as possible so as
not to be influenced by buoyancy, inertia or gravitational forces. Great care must
be exercised in preparing the background, selecting a strong light source,and if
quantitative information is desired a suitable technique for pulsing either the
tracer or the illumination for a photographic record. These aspects are discussed
in great detail in References 10-U6, 5U, 55, 57. Probably the easiest to use of all
these techniques (limited to water models) is the hydrogen bubble technique or air
bubbles. It has the advantage particularly in a closed system that particles do
not obstruct any pitot tubes that happen to be in the system.
Figures I0-k and 10-5 follow
S-1^121 Table 10-11 follows
-------
S-14121
67784
TAW TUBE*
rTAGNATION - PRESSURE TUBE
"MEAaURINO FOOT"
AXB Of DOTATION
A. Claw Probe
DIRECTION TAPS
STAGNATION -
PRESSURE -TAP
-HOLE AXIS
B. 3-Hole Yaw Probe
STAGNATION-PRESSURE TUBE
C. Cobra Probes
X \
D. Wedge Probe I(# Q
Figure 10-4. TWO DIMENSIONAL PROBES
-------
ALTERNATE STEM "•"
A. Spherical Yaw Probe
B. 4-Hole Yaw Probe
C. 5-Hole Directional Probe
Figure 10-5. THREE DIMENSIONAL PROBES
S-14121
67784
-------
Table 10-11
References
10-55, 5l» ,"»6,U9
10-56
10-56
10-55, 5"»,1»9
10-U9
10-56, 1»9
10-58
10-55, 5>»
10-55,57
10-55, 57, 5*
10-59, 55, 57, 5"»
10-56
10-56
10-56 »
10-56
10-56
10-56
Material
Al
NH|,C1
TiCll,
Polystyrene
Balsa Oust
Smoke
Metaldehy de-
Flocks
Air Bubbles
°2
C
H2
Dye Streaks
A1203
Pyrex
MgO
Quartz
Ti02
Zr02
Dispersed In
Air = A
Water = W
A, W
A
A
W
A
A
A
W
W
W
W
A
A
A
A
A
A
Melting
Point
°F
1220
Sublines
>3272
1U8-200
.
3632
752
5072
3092
3866
5"»32
Boiling
Point
OF
3735
968
1*010
6512
1»OU6
Decomposes
7772
Spec. Gr.
g 60°F
2.7
1.53
5.»»7
.98 - 1.1
.1 - .2
.0012
.0013
i».o
2. 2*4
3.65
3.8U1».17
5.73
Particle
Size
U-lU mills
2 mm
1-5 v
Suitable In
Hot = H
Cold = C
C
C
H.C
C
C
C
C
C
C
C
H
H
H
H
H
Comments
Forms T102 in flames.
Sublimes and is re-
duced by flame gases.
Typically boiling oil
vapors . Not partic-
ularly good in turbulent
flow regimes.
Good light scatter.
v Difference in densities
most serious , drawback ,
electrochemical technique,,
> See Reference 55 & 57-
Used mostly for 2-D
studies. Easy to get
' pulsed traces.
Primary laminar flow;
difficult to use if a
closed system. See
reference for actual
dye materials.
Very reflective
Generally considered best.
-------
169
Chapter 11. COSTS
11.1 Scope
This chapter covers the purchase, installed, and operating costs of
thermal and catalytic afterburners with and without heat exchangers. The informa-
tion is based on extensive discussions with manufacturers and users, and also on
user replies to our questionnaires. The data have been presented in graphical
form so that they may be used easily.
Fuel requirements are discussed in detail including the effects of heat
exchange and of hydrocarbons in the waste gas stream. Comparisons have been made
between the total annual costs, including depreciation, for typical thermal and
catalytic units with varying levels of heat exchange.
11.2 Thermal Afterburners
11.2.1 Purchase Costs from Manufacturers' Data
Purchase costs obtained from sixteen manufacturers of thermal afterburners
are plotted as broad bands in Figure 11-1. They cover units without heat exchangers,
with 35 to 1+5J5 heat recovery, and with 55 to 75$ heat recovery. The range in costs
for a particular type of unit may be accounted for by different methods of con-
struction, type of materials used, and instrument package. For instance, a simple
burner in an unlined stack is sometimes used for odor control. It has an absolute
minimum of instrumentation, short residence time, low operating temperature, low
cost, and although it may reduce odor and plume, its performance expressed as
percentage reduction in hydrocarbons would not meet stringent requirements
such as Rule 66. The more expensive units are designed with 2000°F refractory
lining, are made of heavy gage metal with adequate expansion Joints, have baffles
for good mixing and an expensive instrument package. Heat exchangers are generally
shell and tube design, and the cost depends on materials of construction (e.g.,
aluminized tubes, carbon steel shell, versus stainless steel), the number of
passes and provisions for expansion. Rotary regenerative exchangers and packed
bed regenerative exchangers are sometimes used for high heat recoveries; both tend
to be expensive. It must be emphasized that a unit high in first cost does not
guarantee that it will have superior performance. Bids from manufacturers must be
first evaluated on the ability of the afterburner to do the Job, (i.e., residence
time, operating temperature, mixing, etc.), and then compared as to construction
details and costs.
11.2.2 Instrumentation
Instrumentation requirements to meet safety codes have been discussed
in Chapter 9. For a gas fired unit a minimum cost for temperature control would
be $750 with an additional $1,1*00 for combustion safeguards to meet FIA safety
requirements (total $2,150). For more accurate and reliable pneumatic control
instruments, also using a UV detector rather than a flame rod for shutdown
protection, the cost would be increased to $3,600. These costs can be much
S-
-------
170
higher on large systems with built-in control features and alarms to meet special
process requirements with complicated start-up and shutdown procedures. Oil-fired
systems require additional controls which are discussed later under burners.
Instrumentation costs can run from as low as 5$ of the purchase cost to
as high as 30$. On small units, in particular, they can be a substantial portion
of the price. Thus, it is extremely important to compare instrument packages on
different bids on an equivalent specification basis.
11.2.3 Burners
The bare cost of oil or gas burners without auxiliaries or instrumentation
is relatively small. Distributed raw gas burners give better mixing of the waste
gas but are more expensive than single nozzle-mix burners. Dual fuel gas/oil
burners are only slightly more expensive than single fuel burners. Costs, of
course, depend on the particular manufacturer, design, and materials of construc-
tion. Some afterburner suppliers use standard burners from burner manufacturers,
whereas others design and build their own. The following figures were quoted to
us by burner suppliers and are reproduced here to indicate the approximate burner
cost in relation to the overall cost of the afterburner. They should not be
considered as firm prices:
Type of Burner
5- inch nozzle mix gas
Distributed raw gas
U-inch dual fuel gas/oil
MM Btu/hr
3.25 - 5
3.5
3.U
Bare Cost of Burner
$318
$525
$260
Thus, the difference in cost between oil fired, gas fired, or dual
fuel fired systems is primarily due to additional instrumentation and auxiliaries
for fuel atomization including air blowers, pumps and heat exchangers. Manu-
facturers have indicated that a unit burning Number 2 or Number U fuel oil would
cost $2 M - $U M more than a corresponding gas fired installation. For burning
Number 6 fuel oil, assuming that steam is available for heating the oil to
reduce its viscosity, an incremental cost of $3 M - $6 M could be incurred over
a gas fired unit.
It must be emphasized that information on instrumentation and
auxiliaries is extremely dependent on the details of the installation. Firm
prices must be obtained on a strictly comparable basis in order to be able to
evaluate the relative merits of proposed burner installations by different
manufacturers.
S-1U121
Figure 11-1 follows
Table 11-1 follows
-------
140
120
100
80
$M
60
40
20
Thermal 55-75%
Heat Recovery
Thermal 35-45%
Heat Recovery
CC Catalytic Custom
Unit 45% Recovery
BB Catalytic Custom
Unit No Exchanger
Thermal
No Exchanger
AA Catalytic Pre-Engineered
No Exchanger
I
S-14121
67784
10 20 30 40
M scfm
Figure 11-1. AFTERBURNER PURCHASE COSTS FROM
MANUFACTURERS' DATA
50
-------
171
11.2.U Purchase, Installed and Operating Cost from Users' Data
A comprehensive summary of the data obtained from users of afterburners
has been tabulated in computer format and is given in Table 11-2. Items
which have been included, and are of direct interest with respect to costs,
are (a) capacity of the unit in thousands of standard cubic feet/minute (MSCFM),
(b) purchase cost, (c) installed cost, (d) year of installation, (e) operating
cost, and (f) hours of operation per day.
A computer program was written to sort the data into categories, apply
corrections, and machine plot the treated data. It had the following features:
Purchase and installed costs were adjusted to a 1970 cost base by
applying factors from the Chemical Engineering Fabricated Equipment Index.
Operating costs were corrected to a 2U-hr/day basis. (A correction
based on the number of hours worked per year would have been better, but was not
available in most cases).
The data were separated into units with and without exchangers.
Six separate plots were generated covering purchase cost, installed cost,
and operating cost versus capacity, for units with and without heat exchangers.
They are shown in Figure 11-2 through 11-7.
Although there is a large amount of scatter in the data, this is not
unexpected because of the varied conditions under which the units operate, and
the wide differences in design. Most of the purchase costs fall within the broad
bands of prices obtained from the manufacturers.
Installed costs for units without heat exchangers range from 1.2 - 10.0
times the purchase cost with an average of 2.2 times (obviously the ratio of 10 is
exceptional; 85$ of the units are covered by a ratio of 3 or less). For units
with exchangers the range is from 1.2 times to 3-7 times with an average of 1.9.
Thus for an average plant the installation cost is of the same order as the
purchase cost. The lower installation costs are associated with smaller after-
burners shipped as package units requiring little more than a concrete pad and
gas supply to be furnished on site. The more expensive units are associated with
large equipment, often installed on special steel supporting structures, with
long service lines and interconnecting ductwork, and an expensive instrumentation
package interconnected with the process.
Operating costs are largely fuel costs since maintenance is a relatively
small item except in a few cases where the unit was unsuitable for the application.
There have been instances where it has been necessary to do extensive repairs on
an afterburner after only one year in service due to cracks in the afterburner
shell and heat exchanger. This was usually due to inadequate design, insufficient
allowance for expansion, and in some cases, over-temperature operation. Fortunately,
such incidents represent a small number of afterburners.
S- 11H21
-------
172
Most companies ao not keep separate detailed records of maintenance
costs on afterburners so that the information reported to us has been meager.
Eight different units without heat exchangers had an average annual maintenance
cost of 1.9% of installed cost, whereas 11 units with heat exchangers had a
corresponding figure of 3-5$. The higher maintenance on heat exchangers is
largely due to the requirement to replace burned out tubes from time to time in
some applications. It should be noted that the 3.5$ factor is applied to a larger
installed cost for the same throughput since units with exchangers are considerably
more expensive than units without. Choice of materials for heat exchangers can
make a large difference to the frequency of repair of these units. The substitu-
tion of SS tubes for aluminized tubes may be paid off in reduced maintenance cost
over a relatively short period in some cases. In view of the fact that many
companies reported negligible maintenance costs, it is felt that the 3.5$ factor
may be on the high side for a well designed system, and the correct figure may be
closer to the 2% for units without exchangers. This would cover minor repairs to
the refractory lining, maintenance of control instruments and cleaning of flame rods,
if used.
Afterburners are often brought on stream without using special slow
heat-up cycles, although some manufacturers do recommend them. Operator attention
is minimal and consists of starting up the unit, noting by inspection of the flame
that it is operating satisfactorily, and then relying on the temperature controller
to take over. Usually the operator in overall charge of the process equipment is
responsible for the afterburner, and additional operators have not been required
after installing an afterburner on an existing process unit.
Fuel costs are discussed in some detail below, but for comparison
purposes lines have been drawn on Figures 11-U and 11-7 showing typical fuel costs.
They are based on 6,000 hours per year operation, heating from UOO°F to lUOO°F,
with gas at $0.50 per MSCF. The waste gas has been assumed to have no heating
value, and for the case with a heat exchanger, h5% recovery has been assumed.
Fuel costs lower than these curves may be explained as being due to the heating
value of the contaminants in the waste gas; higher costs could be due to the local
cost of fuel.
11.2.5 Fuel
Most afterburners use natural gas as a fuel since it is easy to handle
and control, it burns cleanly, and is free from sulfur. It is presently low in
cost because of price regulation, but unfortunately it is in short supply. It
is expected that the price will be allowed to rise in the future to meet the
supply-demand situation, and also to encourage exploration for new gas fields.
Alternative fuels are numbers 2, U, and 6 fuel oil with low sulfur content, which
are also in short supply. Prices of both natural gas and fuel oil vary depending
on the geographic location. Fuel oil prices also fluctuate seasonally reflecting
demand, and the local supply situation. A typical natural gas price would be of
the order of 50c/1000 SCF, equivalent to 56c/MM Btu (low heating value 900 Btu/SCF).
Fuel oil could be of the order of $5.00 per barrel, equivalent to about 89-5c/MM
Btu, (low heating value 133,000 Btu/gallon, 1 barrel - !*2 gallons). It must be
emphasized that both these costs can vary widely and local prices must be ootained
when the economics of an installation are being considered.
S-1U121 Figures 11-2 through 11-7 follow
-------
NJ
o
o
o
CM.
O
O
o
o.
» s
c o
CD •
•g °
x «
o
i
U
o.
(O
UJ
< °
I °
I 5
0.
0
o
o.
CM
o
o
°0.00
4 .00
8. 00
12.00
16.00
MSCFM
20-00
24.00
28-00
32.00
Figure 11-2. PURCHASE COST THERMAL AFTERBURNERS WITH NO EXCHANGER
( Data T-'">« frmn Users Survey)
-------
VJ I
>i -r
o
o
o
o
CO
o
o
•
o
in
CM
1 ?
£ o
5 °-l
O (SI
X
UI
2 o
o
5 o
8
o
o
UJ
o
o
_l O
< 2-
(/)
o
o
o.
X " XX
x.
X
Xx
*x
xx x
x xx
X
x x x
2*
y
°0.00 4.00 8.00 12.00 16.00 20.00 24.00 28-00
MSCFM
Figure 11-3. INSTALLED COST THERMAL AFTERBURNERS - NO EXCHANGER
(Data Taken From Users Survey)
32.00
-------
O- t/>
vj i
K>
O
o
o
o
I
u
X
pj
Z
£ o
ID C3
O.
o
o
Line representing Fuel Cost Only based on heating
from 40CPF to l40tfF. no heating value in waste
gas, natural gas SOclmscf. 6000 hour operation per
year combustion air from waste stream.
•fo.OO 4.00 8-00 12.00 16-00 20.00 24.00 28.00 32-00
MSCFM
Figure 11-4. ANNUAL OPERATING COST THERMAL AFTERBURNERS - NO EXCHANGER
( Data Taken From Users Survey)
-------
O
o
•
o
^.
O*
O
o
o
CM.
i
o
o
•
o.
o
o
Q
Q Q
IX) Dm Q Q
UD
Q
D _
Q Q
on D° o
o m
Q
o
Q
Q
Q
Q
8.00 16.00 24.00 32.00
MSCFM
40.00 48.00 56.00 64.00
Figure 11-5. PURCHASE COST THERMAL AFTERBURNERS WITH EXCHANGER
(Data Taken From Users Survey)
-------
VJ I
VJ —
o
o
o
00
o
o
o
o.
5 o
Ol O
I o
« rJ-l
X f
Ul
3.
H ~
(O
O
U
_J O
< CO
o
a
o.
CO
o
o
O
O or O.
o
-------
VI I
25
CO
O
o
O
CM
Line representing Fuel Cost Only based on heating
from 400°F to I40(fF using 45% Exchanger, no
heating value in waste gas, natural gas SOc/mscf
6000 hnlyr combustion air from waste stream.
°0.00
8.00
16.00
24.00
32.00
MSCFM
40.00
48-00
56.00
64.00
Figure 11-7. ANNUAL OPERATING COST THERMAL AFTERBURNERS WITH EXCHANGER
( Data Taken From Users Survey)
-------
173
In order to obtain a 90% reduction in hydrocarbons, thermal afterburners
generally need to operate at lUOO°F. Thus, if advantage is taken of the heat in
the discharge from the afterburner to preheat the incoming waste gas, a considerable
fuel saving can be made. Figure 11-8 shows a diagram of such a system in which the
waste gas is preheated from TA to TB through the exchanger, the afterburner tempera-
ture is controlled at lUOO°F, and the hot gas from the afterburner gives up its
heat through an indirect heat exchanger to the incoming gases. A commonly used
expression for the heat exchanger recovery is:
T - T
B A
Simplified exchanger recovery = TTQQ m~
A
(A more accurate criterion of heat exchanger performance is the heat exchanger
effectivness ratio discussed in Chapter 6).
Plots have been made in Figure 11-8 of the temperature Tg against
exchanger recovery for incoming waste gas temperatures (TA) of 100°F to UOO°F.
If temperature TB is found in this way, the heat to be supplied by the afterburner
may be calculated, and fuel requirements estimated, as is discussed later. The
economic incentive for using a heat exchanger with a particular recovery may thus
be obtained.
Fuel requirements are determined by the temperature rise through the
afterburner, the heating value in the waste gases, the specific heat of the waste
gases and whether combustion air for the fuel is brought from outside. After-
burners are often used on waste gases containing only a dilute concentration of
hydrocarbons and in these cases, the oxygen in the waste stream can be used to
supply all that is needed for combustion. This can amount to a 30% reduction
in fuel requirements compared with bringing in the stoichiometric amount of air
required for combustion from outside.
In calculating the heat requirement in the afterburner, the fact that
the combustion products have quite different specific heats from the gases entering
the afterburner is often neglected. Figure 11-9 shows the average specific heats
of a number of gases and how they change with temperature. Items of particular
interest are the values for air and carbon dioxide.
The concentration of hydrocarbons in the waste stream has a large
effect on fuel requirements. For example, hexane at the 25$ LEL has a heat con-
tent of about 13 Btu/SCF which would be sufficient to raise the temperature of
the waste gas by about 650°F.
In order to allow for these various factors, equilibrium calculations
were made using a computer program to determine the fuel requirements to reach
llfOO°F in the afterburner for four main cases. These included natural gas or fuel
oil with either stoichiometric air from outside or combustion air supplied by the
waste stream. In each case the calculations were done to cover a range of inlet
temperatures to the afterburner from 100°F to 900°F with a range of hexane in the
waste gas from 0 ppm to 3000 ppm. (The amount of hydrocarbon in the waste stream
is commonly reported either as parts per million carbon (ppm Cj.) or parts per
million hexane. Since hexane contains 6 carbon atoms, ppm Ci = ppm hexane x 6.)
The results are plotted in Figures 11-10 through 11-13- (For the purpose of these
calculations the natural gas was assumed to have a low heating value of 900 Btu/SCF,
and the fuel oil, 136,000 Btu per gallon.)
S-1M21
-------
In order to calculate the fuel requirements for a particular application,
it is first necessary to assume a heat exchanger recovery. Knowing the waste gas
temperature, then the temperature into the afterburner may be determined from
Figure 11-8 (TB)- Entering the appropriate graph 11-10 through 11-13 at tempera-
ture TQ, and knowing the hydrocarbon concentration in the waste stream, the fuel
requirement can be read directly. This does not make any allowance for heat
losses which may add of the order of 5% to the heating load.
S-1U121 Figures 11-8 through 11-13 follow
-------
1000
"~^r - .
i ' i >•
Of T400T from Afterburner
ed Excha nge r Re covery
500
400 -
300
S-14121
67764
0.2 0.4 0.6 0.8
Exchanger Heat Recovery ( Simplified Expression)
Figure 11-8. EFFECT OF EXCHANGER RECOVERY AND WASTE GAS
TEMPERATURE ON INLET TEMPERATURE
Thermal Afterburners
-------
6
800
1600 2400
Temperature, °F
3200
4000
Figure 11.9. MEAN MOLAL HEAT CAPACITIES OF GASES AT CONSTANT PRESSURE
Mean Values from 77° to t°F (From References 11-1)
S-14121
67461
COURTESY: John Wiley & Sons, Inc.
-------
o
£
-------
2.8
Operating Tf
StOiqhiometr
- r—r-r^r-i^-f
ppm C, * 6 x ( pfp Hwwne)
I ]
0 200 400 600 800
Afterburner Inlet Temperature, °F
Fipure 11-11. FUEL CONSUMPTION; THERMAL AFTERBURNER
S-14121
67784
-------
• M
0
0
£
V)
8
0
1
•s.
5
200 400 600
Afterburner Inlet Temperature, °F
Figure 11-12. FUEL CONSUMPTION: THERMAL AFTERBURNER
1000
S- 14121
67784
-------
Operati ng T
Fuel; Oil
en iperatur ;:
!"•• I
Stoichiometric
Outside
t =s 6x(Helxane)
Figure
200 400 600 800
Afterburner Inlet Temperature, °F
II- 13. FUEL CONSUMPTION: THERMAL AFTERBURNER
1000
S-14121
67784
-------
175
11.3 Catalytic Afterburners
11.3.1 Purchase Costs from Manufacturers' Data
Purchase costs obtained from the three major manufacturers of catalytic
units have been averaged and plotted in Figure 11-lU. The data have been broken
into three main categories, (l) pre-engineered units, generally in smaller sizes,
(2) custom designed units without heat exchangers, and (3) custom designed units
with UO - 50$ heat recovery. The same information has been plotted on Figure 11-1
to show the comparison with the cost of thermal units. Pre-engineered catalytic
units without heat exchangers fall in the middle range of costs for thermal units
without exchangers, but the custom designed units are considerably more expensive.
Catalytic units with 1*5$ heat recovery are in the higher part of the cost range
for similar thermal units. Increased capital cost of the catalytic units should
be offset by fuel savings in an application for which they are suitable.
Catalysts are based on noble metals (platinum or palladium), and
represent a substantial initial investment. Even though different manufacturers
use different types of support for the catalyst, the average cost of the charge is
approximately $1,500 for 1000 SCFM throughput. This cost has been shown on
Figure 11-lU to demonstrate the large proportion of the initial cost which is tied
up in catalyst alone, especially for the small pre-engineered units.
11.3.2 Instrumentation
Instrumentation requirements are generally the same as for thermal units.
The waste gas is preheated prior to entering the catalyst bed with the final
oxidation taking place in the bed. Fuel is controlled to maintain a constant
temperature of the order of 900°F out of the catalyst bed.
11.3.3 Burner
Natural gas is used generally for preheating the waste gases although
manufacturers do have designs for oil fired units. Both nozzle mix burners and
distributed gas burners have been used. Again requirements are the same as
for thermal units except for the lower heat input required.
11.3.U Purchase. Installed and Operating Costs
Only Sk catalytic units were reported in replies to our questionnaires
and these included two units for treating nitrogen oxides. The units were mostly
old and had a low throughput. It was not felt to be worthwhile tabulating the
data separately, but they are included in Table 11-2. Purchase and in-
stalled costs are not meaningful because they were installed so long ago, and
in some cases because of the extensive modifications which have been made to the
original equipment.
Extensive discussions were held with users of catalytic afterburners either
in person or by telephone, in order to obtain as much relevant information as
possible to supplement the small amount of data obtained, from returned questionnaires.
-------
176
Maintenance costs are largely associated with catalyst washing and
replacement. In the Los Angeles area washing three times per year and annual
catalyst replacement is required to maintain the level of performance demanded
by the local APCD. At the other end of this scale, (not in the Los Angeles area),
catalyst cleaning with an air blast every two years, and catalyst replacement every
five years has been adequate. If the waste stream is free from catalyst poisons
and does not contain particulate matter, over two years catalyst life would be
expected with washing one to three times per year.
Washing is a relatively simple operation, but does involve removing the
catalyst from the unit. In most designs the catalyst modules are small enough
that they can be removed by hand. One manufacturer uses a system where the whole
bed must be removed in one piece. This could be an advantage where there is
adequate space available for handling, but would be a problem if the afterburner
were installed in a confined space which often seems to be the case.
Catalyst replacement costs depend on the type of catalyst support being
used. Where the catalyst is deposited on nichrome ribbon, it is often possible to
use the same module, clean off the old catalyst, and put on new. This type of
replacement cost would be of the order of $500 per MSCFM throughput. If the old
module could not be re-used, replacement cost would be close to the price of a new
charge at $1500 per MSCFM. Oxycat brick-type catalyst can usually be replaced for
about $700/MSCFM since some of the supports, can be reclaimed. This is compared to
a new charge price of $1000/MSCFM in this case.
At this time manufacturers have been unable to develop a satisfactory
method of reclaiming honeycomb-type supports, although they are working on the
problem. Currently some manufacturers give a 10$ discount for the old catalyst
which represents recovery of the noble metals. Thus, replacement may be as low as
$1350/MSCFM throughput. It is expected that this cost will be reduced further
when techniques for reclamation are developed.
The figures quoted above are only average prices based on discussions
with manufacturers. They will vary depending on the exact catalyst formulation
and the manufacturer involved. It has been assumed that 1 cubic foot of catalyst
is required for each 1 MSCFM throughput. (See Chapter k for further discussion of
this.) Firm prices for catalyst replacement should be obtained directly from the
manufacturer at the time the installation is being designed.
Maintenance costs, apart from catalyst replacement, would be expected to
be slightly less than for the corresponding thermal units because of the lower
operating temperature.
Operating labor is again a negligible factor.
Fuel costs are considerably lower than for thermal units because of the
lower operating temperature.
S-1^121 Figures 11-14 and 11-15 follow
-------
Figure 11-14 and Figure 11-15 (pages numbered 174-b and
174-c) are in correct text order.
-------
XI —
° 5 100
80
? 60
8
U
j
U
| 40
20
0
0
^\
r-«^
i
L
;
•
-;
-—
;:•':
•
m
.
-
- •
. .
n .
H
|. .= ..
:
• -
- . .
. . . t
i
I
-
: "
I
•
.
g
.
-
•
. _.
i
•
1
:-
t
" • • T
I-'
•
^""
**\-
~~^
•
4
Figure 11-14.
•
— **•
.
•
•
•
#£.
.
/"
--**
- •!!>. 1
^>
-^
•'.
'
•
- —
1
i
•
1
'
i
/
S^
*** .
•
^
' '
•
/
^
j
X
r—
^
Pre
•
irvg
j
•
^Em
Ur
- - -
+**1*
71
I _
•
/
f
H
: ' :
in*
!*•
,
' :
ii
8 12
CATALYTIC AFTERBURNER
sre<
• ..
• :
i
.: . :'
'**
' ' ~. '
M
.
/
-
<^"^
•
.•
•fi|
{$
.:
: h
j
• •
'
X
^Y
\>
•
\
\
. I -. .
t
-fe
i tal yst C
1500 pe
t . -•
K^
*-*•
ost
M<
"
:!:'
r-**
I/'
^jj
...,
CF^
.
(:'
Js'
I '
- -
'
.1 . •
; ^
^=fr
..i- -
. j. ..
. : ':
, -r^
Etit
:
i
fe
l
-}-7-r
g
. '
• " T
;:
rvix
.H.'v
:•
!
Custt
^ w? th
f4
-------
na
o
•6
OJ
O
^0
z
Tempe nature Outrof Bed:i 90
Fuel;; Natural
Combustion Air Norn Waste 0
•.. .4 ;_.:nT: :.]TJ „ t-rT.i
• ——r —••"
•. m
• 4
i — h- - • - +''' *^ ' "I *
... , .
mm
7—J ----- '
0.2 —
200 400 600
Afterburner Inlet Temperature, °F
Figure 11-15. FUEL CONSUMPTION: CATALYTIC AFTERBURNER
1000
S-14121
67784
-------
177
11.3.5 Fuel
Catalytic afterburners should give satisfactory hydrocarbon reduction at
an outlet temperature of 900°F from the catalyst bed (see Chapter U). Part of the
temperature rise is provided by fuel in the preheat section, and part by oxidation
in the catalyst bed.
Fuel requirements have been calculated and presented graphically in
Figure 11-15 through 11-18 for the. 900°F discharge temperature. They include both
natural gas and fuel oil with combustion air either from the waste gas or from
the outside with different levels of hydrocarbon concentration in the waste gas.
Figure 11-19 shows the effect of heat exchanger recovery on the temperature rise
through the exchanger.
Determination of fuel requirements follows the same procedure as for
thermal afterburners. An exchanger heat recovery is assumed, then from the
waste gas temperature T^, the temperature into the afterburner TB, may be read
from Figure 11-19- Entering the appropriate graph 11-15 through 11-18 at tempera-
ture TQ, and knowing the hydrocarbon concentration in the waste stream, the fuel
requirement can be read directly. These numbers are average values, and somewhat
lower fuel usage would be expected with a new catalyst bed, and somewhat higher
as the catalyst ages. Generally, catalyst on a ceramic base is more active than
catalyst on a nichrome ribbon base, so the former would give slightly lower fuel
consumption. An allowance for heat losses of the order of 5% should be added.
S-1U121
-------
178
11.U Total Annual Cost
The total annual cost of operating an afterburner depends on the installed
cost, depreciation, interest charges, as well as the maintenance, fuel and power
costs. All these factors must be considered when an installation is being planned,
and must be based on firm quotations from equipment suppliers coupled with the
local costs for fuel, pover, maintenance labor and construction.
Illustrative examples of the cost of catalytic units and thermal uiits
with and without heat exchangers have been calculated for comparison, and the
results are given in Table 11-2. It must be emphasized that these do not represent
real costs since factors on capital have been used to estimate installation and
maintenance costs which will vary depending on the unit and the plant location.
The following assumptions were made:
1) Waste stream 100°F, 10,000 SCFM (= 10,800 ACFM at 100°F), containing
1000 ppm hexane;
2) Combustion air provided by the waste gas;
3) Installed cost 2X purchase cost;
b) Ten-year life straight-line depreciation;
5) Interest charge on capital, 9%;
6) Catalyst replacement every two years;
7) Maintenance 2% of installed cost;
8) Fuel natural gas 50c/1000 SCF. Requirements calculated using
Figures 11-8, 11-10, 11-15, and 11-19;
9) Blower power calculated assuming forced draft, 6 in. and 10 in.
pressure drop for catalytic units with and without exchangers; 2 in., 6 in., and
10 in. for thermal units without exchanger, with kO% heat recovery and with 60$ heat
recovery, respectively.
10) Electricity Ic/KW hr.;
11) UOOO hours/year operation;
12) Individual costs have been rounded off;
13) One ft3 of catalyst/1000 SCFM waste gas.
Using this particular set of assumptions , it is seen that the pre-
engineered catalytic unit is somewhat cheaper than all the others, followed by
the catalytic unit with heat exchanger, and then by the thermal unit with a kO%
exchanger. A thermal unit with a 60% exchanger is the most expensive, but this
cost is heavily influenced by using straight percentage factors for installation
and maintenance.
Figures 11-16 through 11-19 follow
S-1^121 Tables 11-1 and 11-2 follow
-------
1.4
Temperature Out of Bed: 900°F
Fuel: Natural Gaa
Stoichiometric Combustion Air from Oi
•
ppm C jj = o x (ppm Hekane i
200 400 600 800
Afterburner Inlet Temperature, °F
Figure 11-16. FUEL CONSUMPTION; CATALYTIC AFTERBURNER
1000
S-14121
67784
-------
10
8
(0
O
_
g.
•2
-S.
O)
2 h
Figure
400 600
Afterburner Inlet Temperature, °F
11-17. FUEL CONSUMPTION: CATALYTIC AFTERBURNER
1000
/ 7?-A
S-14121
67784
-------
200 400 600
Afterburner Inlet Temperature/ °F
Figure 11-18. FUEL CONSUMPTION; CATALYTIC AFTERBURNER
1000
S-14121
67784
-------
Constant Discharge Temperature of 90C
Catalyst Bed
:
/
ftttrtthtffi
' : - : | -
Exchanger
m
Tf- •
Exchanger Recovery =
.
100
S-14121
67784
0.2 0.4 0.6 0.8
Exchanger Heat Recovery (Simplified Expression)
Figure 11-19. EFFECT OF EXCHANGER RECOVERY AND WASTE GAS
TEMPERATURE ON INLET TEMPERATURE
Catalytic Afterburners
-------
Table 11-1.
c
p •
• 1
0 C
I 0 n u
fl 'ft! S 1
fl ffll < 1
fl fftl S 1
II IfOO 1 1
II ItOO S C
tf IfOO S 0
77 IfOO 9 1
10 I7lf 1 1
IT l|f| 9 C
It ll'l I D
10 Itll I I
11* llll S 0
ll> llll » 0
1 KM 9 D
II KM N |
111 I'M 9 0
1IC KM 9 1
110 KM 1 I
III KM 9 I
IIP Klf 9 I
ll« I'M 1 I
11
• L
C Cl U
T tt 1
T ft C
T fl 1
T ff G
I fO I
t to i
T H I
T I
T ft 1
T tt 1
T If •
T ff I
1 ff G
T C
1 G
T ft G
T G
T I
T G
T G
T G
0
i in
?
IT.O 7|
lO.O IS
It.O Tt
It.O |10 II
l.t TO
IBOO
lO.O ISO t
•' T| to 10000
•1 100 110
I.O f'O
too
too
1.' 110 II 10
I.O 100 II t IIOO
i.o mo it t iioo
I.O 100 I* S |IOO
I.O 100 It § llOO
I.O 110 10 ISO
too
ll.O ISO 11 It fll
t.I ISO 11 It ItO
.1 110 1 TTO
I.T tit lOOO
1.0 100
tit 110 >100
l.f 110 IIOO
i.i no iioo
T.I 110 7 100
I.I ISO IIOO
PROCI99 C*TCGOB| rfn* US GOvCRNHCMT ITtNnARO
runt ••ooof I.NC fCouCTiow c«roiic purtpi c
Fun G.NtTutii G>9 0*011 -.c in.cn a Of i
H
0
u
flu S
Pit 0
t
9CF 1
C IS • N
C IS k N
C It k T
C It II T
C H • N
C 1 • N
C It • •
tll.O I 1 N
H N
C If f N
C It II N
I f t N
C 7 • N
• INN
C 7 t N
C If N T
t.n c II » N
I.O CHUN
t.O CHUN
C It H N
C It N T
1
N
0
F
0
•
C II NIP
C
0 IN
F IIOO
It
F 100
1 70
F TO
1 IIOO
P III
F lit
1
F Tt
F ttO
too
i too
r no
F no
F ItO
1 100
1 100
1 100
1 100
1 100
1 110
F
C I • II | ItO
C • • N | ItO
C I • II 1 ItO
C 1 II H | ISO
C 10 • N | 110
C I • N | IIS
C N N | tIS
C It I II 1 100
C II • N F 110
C I • N F
C II N N F
C I « N F
C k N F
• t 0 N F
MOU9TIIU CItStlFICII
. SNO>C/P«IICULITC9
I'OTNC* PPN . PtllS
IINP
OUT
IIOO
IIOO
IttO
IttO
I'lO
IIOO
IIOO
lltO
IIOO
1110
1110
IfOO
1000
IIOO
ItOO
ItOO
1171
IIOO
1110
1110
ItOO
MOO
1110
ItlO
I'lO
I'lO
ItOO
ItOO
lltO
lltO
lltO
ItSO
IIOO
IIOO
IIOO
ItOO
IIOO
ItOO
IfOO
ON
E'OTHI
•CP. I
CO '
Oil
t.t; i.o
S.S I.O
l.t S.t
l.t t.t
I.O It.O
T.t ll.O
I.O 10. 0
l.t I.I
l.t It.O
I.O 1.0
l.U t.O
t.O I.O
.
l.t 10.0
•
I.I ll.O
l.f ll.O
l.f fl.O
l.f tl.O
I.I t.I
I.J '.1
S.S t.o
t.t I.O
l.t t.S
1.0 I.I
1.1 ll.O
t.t I.O
I.O 1.0
l.t It.O
t.t It.O
1.1 It.O
l.t ll.O
• »••••
IlllOK Cl
ORG
out
N
H
M
N
N
• tt N
N
1 N
N
N
N
N
ITO N
111 T
lit T
111 T
T
T
100 H
100 N
N
100 i
•II II
N
ITO T
N
N
N
N
N
N
N
N
fuel
UN
ITU
10.0
l.t
t.O
III.O
.t
t.T
1.1
l.f
l.t
l.t
l.t
I.I
I.I
I.S
l.t
It.f
t.I
I.T
l.t
It.O
.1
H'A.O
••••C09I *••••••
OP
n»
I.O
ll.O
f.O
ll.O
t.t
t.o
s.o
ll.O
t.o
f.O
IS.O
ll.O
I.O
I.O
II. 0
ll.O
ll.O
l.t.
110. 0 10.0
ll.O t.O
It.O '0.0
I.T
If.O I.O
7.1
I.O
I.O
It.O
ll.O I.O
ll.O 7.0
ll.O T.O
to.o 1.1
10.0 I.I
ll.O 1.0
ll.O 1.0
to.o
•0.0 It.O
to.o i.i
T.O
10. 0
1.0
If.O
I.I
f.l
10. 0
1.1
t.t
tt
tf
11
II
11
II
It
• IPNICTOX IT
If
TO
70
11
N01 70
71
TO
TO
StFITT OOICCS tS
9>r(TT OCVICI9 II
9>FI1T Ol'ICIt II
9IFCTT DOICCI |7
9IFITT OOICC9 II
9IFCTT OIHICCS II
•IFFlC FtlllD II
IIFFlC FtlllO II
••IC>9 LOOSC
•CFMCTo'T tl
•CFKCTO'T II
FIN CONTHl ICFR
II
II
It
tf
II
II
II
FLINI 100 POSH. IT
IK
PC^FOfMtNCC !•!tTltriCTORV N.MOT SMISptCfORT ••NOT TCT ffuH
DIxmSIONf, in PIC! IC'PC»lF|
rMTfc't «.«o«i C-CKLONC O'onuiioii s*sc>u»c> •••oiocLOni C.PICCIPIKTOI p«oiLt"i OS-OUT OF noicc
-------
Table 11-1. (Contd-1.)
c
1
c
0
< U 1
i
tl
•I
(• u
CONrlHfHNT
0
1 LBJ llu
1
H MM
0 P 1 0
• C T F
1 1 0
C
c
T
UCL
NN
STU
CO T HI
OP
UN
JM Kit • T ft
«*• »«7f S B 1 «0 ft
«*C HM » B T «fl C
711 lift I B 1 G
I.O
io.o mo 10
i o.o mo 10
••* 1*0
us iooo ift. a c
IIS iQOO IIS.O C
*o s.r e
1
c
10
ID
10
M
N
N
N
H
H
1
v r
T p
1
T P
T P
aoo
•00
»io
•so
1200
MOO
MOO
IIOU
1100
MOO
MOO
0.0
*.Q
).?
i.O
11.0
11.0
14.0
JO.O
N
300 N
300 N
N
1.0
1.0
1.0
I.I
IU.O
17.0 71
45*0 *0.0 ll.O PIN Ml iCMftNfiKR*!
*)S*D *0.0 lft*0 PAN HT iCNANCCIUI
JJ.O U.O HOT OPIR*T« TtTTI
OT.ITST IIP!
PtOCCIS C*TC60*V F"0« US 60VFtii«riiT Sr*Nit*to |NuufTII|*L CLASSIFICATION
FUHC ••000* B.NC •COi'CVIOH c*inilc FUHrS 0. VNOtC/P*IITICULI'CS C*0|HCII Fal*o «*0*D HaA.O I«A*0*0
FUCI c>HiTutii sis g.oii >.riTHCt o o. G MOTNII PPH . PUTS Pt« HILLION ci
OIIAHICS OUT ciiNtt unit rci< BIH.ION 01 ito.tn pltcinT tcoucTiON IHNU IFIINIUIMN
PC*PONNI.Cf S«Stll«r«CIBNl N-H01 ItTlfrlCTONT I.HOT fit NUN
DI-IHtlMt tl Hit TI.Pt»TUt[S flrlHtCl F IHO/PO«CPO*INDUCIO OH POUCIO ON»FI H.H»TU»L 0«»PT
PNIINtll H.HOHI C'OCLOHC otfllluTIOH S'SCNUIOEN ••NOTOcLOHl t.PttClPI'»TOI PNOIllHS OJ-OUt OP SCIvUC
-------
Table 11-1. (Contd-2.)
c
c r •
• 0 T*V
T I tl
C B »l
0 * U C (B U
*• J«7* • H T 6
100' J«M NIC G
1001 >47* NBC a
IQOC I«M h 1 C G
1000 !"'« NIC fi
IBOC Jiff NIC a
100' sift NIC «
10QM »»7» * 1 C «
1001 Hff lie 6
100J J-iTf i fl C 6
IDOL 1*17* SIT «
II 11*1 SO* 6
«« 1J«I N 0 1 G
11 107* » 1 T G
IS 10** S 1 T *f 0
PlOCtSS CMCaOVr r«0-
0
1 LBS
f
fi ft*
t
7.0 2't
II. 0 lit
|1.« ISO
1.7 Z&O
JO.O ISO
t.l ISO
t.l ISO
t.l 110
t.t 110
t.t ISO
t.t ISO
t.t JSO
t.t ISO
t.s isn
«.S ISO
t.O ISO
B.O ISO
lO.O HO
S.O ISO ISO
44.0 111
too
1.4 100 II 171
1.0 ISO
••0 ]>S
|I.O ISO
1.0 100
US GOvC«NiF*T ST
• ••
4SO
410
11
11
• 1
«J
• 1
11
• 1
• 1
HI
11
HI
I]SOO
I OSS
B
t
1
C
Bill /
C
PC*. 0
N
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
B
B
B
1
c
B
B
ft
H
0
U
I
0
*
11
11
11
1,
2*
I
1
B
B
B
B
B
B
B
•
B
B
B
B
B
1*1
11
11
2*1
11
Ii
2«
I*i
21
2*1
2*1
B
2*1
C 11
C 11
C 11
I«DUST»UL
1
H N
C 0
0
c •
I C
c c
in T r
» i r
N N r
M N r
N N r
N N r
N T r
N T r
N T r
N T r
N T r
N T r
N f F
N T r
N T r
N t F
N T F
H f F
N T F
N T F
N N F
N
N
N
H
N
N
N
N
N
r
i
N
N N F
N M
N T F
N T F
* T |
* f 1
Cl*SSI'
TCNP
•to
Bfcll
ITS
271
17S
III
too
too
400
400
4(0
400
400
400
400
400
4(0
400
400
400
ISO
BSD
BSD
(SO
(SO
BIO
BIO
BSD
BSO
BSD
400
(00
(00
too
too
L
1
N
UUP (0 1
1
OUT OU X
1100 t.U 10.0
1100 4.0 10.0
1100 111 II. 0
1100 1.1 II. 0
1100 t.l 11.0
1100 1.1 II. 0
1100 t.l II. 0
1100
1100
1100 t.t 11.0
1100 t.l II. 0
1100 t.t 11. 0
1100 t.t 12.0
1100 t.t II. 0
1100 l.h IB.O
1100 l.t1 II. 0
1100 1.1 17.0
1100 1.1 17.0
1100 1.1 17.0
I10O 1.1 17.0
1100 l.t IB.O
lltO
1110
1110
lltO
1110
1110
1110
1110
1110
1100
1110
1110
1110
1410 1.0 11.0
1100 I.I
4.t t.O
ISO lilt 1.« 14.0
too 1100 t.o 10.0
410 1100
IC'TION
FUCL ••••COST «(•••••
nit OP
0«t ITU PC*
10.0
II. 0 41
1.1 II. 0 (.0 000* HC IN Oil.
1.1 II. 0 1.0 000* HC IN Oil.
1.7 II. 0 1.0 0001 HC IN Oil.
1.7 II. 0 1.0 000* HC IN oil.
1.7 II. 0 B.O 000* HC IN Oil.
1.7 II. 0 1.0 000* HC IN Oil.
It II 41.0 1.1 41
II N (1.0 10 ((
1 N (.7 (7
1 N 10.7 41
1 N 11. 1 41
1 • Hit 41
1 N 11. 1 41
1 N 4it (7
1 N t.O 47
1 H l.t 47
1 II 1.1 41
1 N l.S 41
1 H 4.S (1
N S.I II. t ClTALTST IIP!
II t.l II. 1 C4T4LMT lire
N t.l 17.1 C4T41T1T LIFl
II t.l II. 1 CITtLTIT LIPI
• t.l 11. t C4T4LTII LIFC
N t.l 17.1 C«T«LTST IIPI
N 1.1 17.1 C«T«LTIT LIFl
40.0
40.0
10.0
N 1.0 10.0 10.0 II. 0 COHTRLI 71
• 14.0 (0.0 JUST ON ]T*C>N II
N .1 1.0 I.C «>! »1IL 4t
N 10.0 40.0 It.O 10
N It.O 100. 0 1(0.0 41
B* N 1.2 11.0 *S.Q 12*0 FAN BURNER 4S
107 N l.S Jft. 0 40.0 IB-0 **>
oi>[»tio«s IN FCCT ic-pc<'!u*ts
F INO/FOICEO>INOUCCO o* FOKCCO DP.IPI
FIC1IE>I n.NONC c-CtCLOHC D-OIIUTION S-SCIUBKC" K.HOIOrl 0«F
H-NITUXL O»FT
• *CC IPH'IOH P>OBLl»I OJ-OUT OF (CNVlCC
-------
Table 11-1. (Contd-3.)
t
c - •
• 1 C 1
0 1 i>L
t o • u ce u
s • : M NL c
1 1 Ml | S 0 T fS 6
73 Ml 1 S 1 I f t G
«P Mil * C G
«! Ml 1 S | r ft 0
S7 Mil 51 1 M i
*B Ml l S 6 1 ft I
102 2*1 | 5 • T ft l
I03C Mil S C I fS I
IP* Ml | S C 1 t> G
2S MOO S » 1 ft «
Ii iBSl S B 1 ft G
7* MSI S 1 t ft G
34 1BS| S B C tS G
37M f|Sl B T G
*l >4t| S • T *• 6
«» ieS| S * I ft G
t* 1«S| S D 1 ft G
SB* 1834 5 i 1 fS G
t4 Za3« S » T G
3 fais s r i 1 7 i
*S> 7alS S B I fO «
in moo (* o i to G
»i 7*00 * c T *7 G
IB* 770O N H T G
ia» MOO *• H t «
13 M7» SOT 0
IT" 2QtS »• G I ft G
>O lOtS S O t *' G
70B ID'S S 1 1 ft G
10
• l LBS
CT
CO «*
2.0 120 s -ona
s.o >no
1*0
5
t.O 120 13
1.0 «00 IS 100
3.0 "SO
70.0 fib B?*00
70.0 tio B?aoo
IS.O sno l *7000
27.0 700 7
.1 |to 1 11)000
7.0
1 .S
•* "0 ZO 1 ISO
.1 7« 1*
.1 «BO SS
7*.0 |«0 It 40
7S.O *S 2Q 10
l.O 7SO 1* l*
]3.0 MO 10 IISO ISJOO
*}.0 |SO -BOO
70.0 77Q
1.0 400
1.0 40O
4.0 100
•.0 «no 14 BOO
s.o ino 17
1.0 US 11
• 1
• H M H
C V I > /
91ti /IT 0
C /ft*
" IICC
C 4 « « 110
C 10 N • |
e 21 s » i 140
i 10 s T r 100
Pi 4 N r F
• 40.0 i F ISO
ii. n c 24 c r F no
11.0 C 24 C 1 F 420
C 14 N « 100
C 24 N N 100
B l» S r 1 70
1 SOO
S S T F 700
II 1
• 14 0 • 1 10
• 14 S N | IS
C 14 N II F 410
c 14 s r i 100
C 14 S 1 F ISO
7.» C 21 c I F 1010
c 21 N i r 1000
C 24 s N F 270
24 N V r 140
• 24 II II F 100
C 24 s 1 F ISO
B 20 M N F 400
C 1 • H F ISO
C N T 1 1200
•1C B C • F IIS
L
C
N
T
1000 1.5 11.0
• 00
1100 |I.O 10.0
1100 S.O 1.0
IIIS l.S 14.1
IBOO 4.0 >.5
1000 10.0 11. II
ISIS 10.4 24.0
4.2 12.0
B.U IO.O
IBSO |4.0 10.0
IBSO |4.0 10.0
IISO |4.U 10.0
1400 1.0 20.0
1400 0.0 20.0
ISOO 2.d ft.O
II DO
1200
1100 2.0 1.0
ISSO 2.0 1.4
1100 1.0 14.0
IISO S.S 20.0
I2SO 4.1 II. II
1100 ll.l 40.5
• SO 4.O 14.0
1100 1.0 1.0
ISOO 4.L 10. U
1100
1100 5.5 15.0
IISO 4.0 S.O
• DO 4.5 1.0
1400 If.O
1100 1.1 I.t
1200 1.4 7.5
FUEL ••••COS? !!••••••
mm OF-
OR* HIU PC*
II. 0 20. O 4.0
C05I FIILI 01
N l.S 24.0
ii 2.1 ii. o 11.0 HINT noos ncauo
N 1.4 II. 5 SO.O HINT HECH FtllS
• .« 11.7 .5
R*0 N 12.0 )«•• 70.0 FLAHI OuT.TINP
N 1IS.O UMUIIIICO '<*!'
N 225.0
N 21.0 28.0 47. O
N
II
ISO H 110.0 tlCHCI FOVLIIIt
ISO II 1)0.0 flCMSK rOULIKt
ISO N 110.0 e»CMt« FOULIH
COIIIIIOLI
N 4.5 ISO.O 10.0
N l.S 150.0 10.0
n 1.0 10.0 100.0 1.5
II ELCCIHIOL
II 15.0
N I.I 1.0 10.0 2.4
N .4 IB.O 25.0 i.o msiminciiTt
II 10.0 110.0 141.0 70.0 DOItlll (I. men.
N ll.l 144.0 104.0
•to ii LI 6i.o nim 111 oucii
«1» II 10.1 114.0 100.0 «7.0 PDOCISS UFSII
• 10 ii ll.l ito.o 100.0 15.0 men ocn.
« .1 is.o »7.o conosioii
IH 4.0 COLOII HOI
II (.5 10.0 •V»C« FLU* OCC
N II. S 7.5
2 1 4.f S4*0 2V5.0 IS.Q
II 1.1 10.0 tc FtlLUIE
T 1.0 21.0 1.0 HC tCOUCIIOK
15.0 ISO.O
II .1 1.1 4.0 .1
N .7 1.1 4.0 >•
44
44
45
64
4f
64
61
55
70
10
51
64
44
44
54
51
4»
41
41
70
7U
64
60
71
6V
4>
66
71
71
61
61
48
4*
44
44
71
42
42
40
67
71
70
10
10
0» US G
i««iir$
• ILllON OH K10.10 pcICtnT lEOUCIIOH tx«u «fTE*»U«« FORCCD DBcFI ••
F»0>LlHS OSiOul Of SCIvlCC
-------
Table 11-1. (Contd-4) .
CONTI
C *
TCI eo
E *L 41 LBS
0 N U C EB U E« «•
BIC 20*ft S fi T G 1.4 220 1
BIE 2n«ft I « T 6 1.4 220 1
101* 2ti*ft s r i t* « .4 *fto
IOIC 2Q«& i f 1 •« G .4 *ftO
2* ro*« sat G io.o mo
I»C 2011 S D T f* G 2.1 200
|*r 201) S D T «f G 2.1 20(1
B 1
•<|NNT * H
1 0
C U
C /
•PM SCr T T IN
"°
i««oo e • N N r
1*100 B • N N r
B 8 C N r *JftD
e • c ri r o&o
l.ft B 20 N T F 540
B 18 M N F
H IB N N r
B * N N r
our
1200
1200
I2SO
1150
1200
1200
1100
on
1.1
1.1
1.4
1.4
1.1
1 . J
2. ft
L
E
N
H OUT
ft. 8 N
ft.B N
7.0 N
r.o N
4.0 N
4.0 N
(IN Of
i.i
I.i 1.1 .1 71
J.ft 7.2 .ft 71
.7 |0.0 14.0 S.) 4ft
.7 |0. 0 14.0 5.1 4ft
Ift.O fl.O 143.0 40iO 4*
.ft ft.D 1.0
.ft ft.O 8.0
47
47
/7f-
-------
Table 11-2 TOTAL ANNUAL COST
Purchase, $
Installed, $
Total Capital, $
Depreciation, $
Interest on Capital, $
Catalyst replacement, $
Maintenance, $
Fuel, $
Blower pover, $
i
Total Annual Cost, $
Catalytic
Pre-engineered
No Exchanger
22,000
22,000
111* ,000
i*,i*oo
• It, 000
7,500
900
16,000
500
33,300
Custom Unit
No Exchanger
38,000
38,000
76 ,000
7,600
6,800
7,500
1,500
16,000
500
39,900
Custom Unit
hot
Exchanger
148,000
148,000
96 ,000
9,600
8,600
7,500
1,900
7,100
800
35,500
Thermal
No Exchanger
22,000
22,000
UU.OOO
U,l400
14,000
900
32,300
200
Ul,800
Uo?
Exchanger
J46.000
146,000
92 ,000
9,200
8,300
1,800
17,900
• 500
37,700
60%
Exchanger
7*4,000
7U,000
lUS.OOO
1*4, 800
13,300
3,000
10,100
800
142,000
-------
179
Chapter 12. COMBUSTION PRINCIPLES
In order to logically design an afterburner, one should understand the
principles which govern the oxidation process, especially as it is utilized in
dilute fume afterburners. Therefore, in the sections which follow we discuss the
oxidation process with its associated heat generation and loss ; the meaning of
ignition temperature and flammability limits; the reasons why some of the waste gas
must bypass the flame; the mechanisms involved in the oxidation process; estimated
oxidation rates for various pollutants; the role of mixing processes; and the
fate of various heteroatoms. It is not our intention to present a complete
treatise on combustion or combustion chamber and burner design. Good general
references are books by Spaulding,18'1' Bradley, 18~8' and Minkoff and Tipper. 18~3>
12.1 The Oxidation Process - A Balance Between Heat Generation and Removal
Steady state combustion or thermal oxidation processes are defined by
a balance between heat generation by the exothermic reaction and heat removal
from the local region where reaction is occurring. Actually, combustion is far
from a true steady state operation, as any observation of flickering flames will
indicate, but one can view flames and thermal oxidation as overall steady state
processes. We shall use this approach to define ignition temperatures, flame
temperatures and combustible limits in order to provide an understanding of the
basic requirements of afterburner design. The treatment is brief, but an entire
book by Vulis18~4' is available on the subject.
Oxidation involves bimolecular reactions between the combustible com-
pound and the oxidizer. As shown in Section 12.2, the oxidizer is usually not
molecular oxygen but this is not important for the discussion here. The rate
at which oxidation proceeds can be approximately represented by the overall
expression
r • «*." "combustible1 (12-1)
where r is the rate at which the combustible compound is oxidized, moles/cm3sec;
k is the rate constant, moles/cm3sec; and x is the mole fraction. The rate
constant, k, is an exponential function of temperature
-E/RT (12-2)
k =, A e
with activation energy, E, typically in the range of 10-60 kcal/gmole°K for this
overall rate constant. This means that oxidation rates are extremely temperature
sensitive. For an activation energy of 30 kcal/mole°K, the reaction rate will increase
by 103 times for an increase of 508F at 80°F and by 1.52 times for the same 50°F
increase at 131*0°F. Equation 12-1 shows that the oxidation rate is also dependent
on the concentration of combustible. Since combustible is consumed in the
S-1U121
-------
l8o
reaction, the rate decreases an oxidation proceeds. This leads to the familiar
S curve (Figure 12-1) of extent of reaction (i.e. fraction of the hydrocarbon
contaminant, which has "been oxidized to C(>2 and H20) versus temperature for a
constant reactor (combustor) residence time. The curves in Figure 12-1 were
computed using the CO oxidation kinetics reported by Yuster, et al,12~5^anci
given as Equation 12-37- (The CO oxidation kinetics should be considered
suspect at these low temperatures, however.) Oxygen is assumed to be in
large excess, so its concentration is constant. (Section 12.3.2 discusses
the various rates measured for CO combustion.) As can be seen, for a given
residence time and temperature, the plug flow (zero backmixing) reactor gives
more complete destruction of combustible than the stirred tank reactor (com-
plete backmixing) in which the concentration is everywhere taken equal to
the exit concentration.
The extent of reaction can also be expressed as a dimensionless heat
generation (by the exothermic oxidation reaction), if the heat of combustion
is taken as constant. The upper section of Figure 12-2 shows the S curve of
this dimensionless heat generation versus temperature for a single well stirred
reactor with 0.1 seconds residence time. Steady state operation requires
that this heat generation be balanced by heat losses and convection. If heat
losses can be ignored, the dimensionless convective heat removal rate (as a
fraction of the maximum heat generation rate) is
T _ T
= inlet (12_3)
convective heat removal ~ x
where T = temperature within the stirred tank reactor
(combustor)
Y = mole fraction of combustible in the incoming
inlet stream
AH = heat of combustion, Btu/lb mole (for complete
conversion to COs and HgO)
C = average of combustible heat capacity of
p flowing stream, Btu/lb mole°F
The straight line given by this equation is shown as the dotted line in the
same coordinates as those used for S-shaped heat generation curve. The line
shown was calculated for a waste stream inlet (preheat) temperature of 800°F
and for a stream hydrocarbon concentration equal to one-fourth the lower
explosion limit (lA LEL). At steady state, heat generation and removal must
be equal. Thus, points A, B and C, the intersections of the heat generation
curve with the heat removal line represent possible steady state operating
conditions for the system.
S-14121 Figures 12-1 and 12-2 follow
-------
o
<0
0)
c
V
CO Oxidation Kinetics
as per Eqn. 12-37
0.01 Second Residence Time
1200 1600
Temperature, °F
2000
2400
Figure 12-1. EFFECT OF TEMPERATURE AND BACKMIXING
ON EXTENT OF REACTION
S-14121
67784
-------
X
o
I
0.4
0.2
Heat Generation for
0. 1 sec Residence Time
Xonvective Heat Removal by Combustion
Products for Waste Stream with HC Con-
centration of ]/4 LEL, Preheated to 800°F
Heat Generation
Heat Removal
CO-Oxidation Kinetics as
per Eqn. 12-37
I
I
I
400 800 1200 1600
Temperature, °F
2000
2400
2800
S-1412J
67784
Figure 12-2. HEAT GENERATION AND HEAT REMOVAL
Effects of Temperature and Residence Time on Generation.
Effects of Concentration and Preheat on Removal.
-------
181
The lower portion of Figure 12-2 shows heat generation curves for three
residence times, and five heat removal lines, representing four values of hydro-
carbon concentration in the waste stream and two preheat temperatures. The LEL
for CO is taken as 12.55&V, AH is taken as 1.217 x 105 Btu/lb mole°F. As can
be seen, the slope of the heat removal line increases as the concentration
decreases. For very low concentrations a nearly vertical line is obtained.
If heat losses occur and these losses are proportional to the combustor
temperature, the slope of the heat removal line will be greater for a given
concentration.
Each intersection between a line and curve represents a possible
steady state condition for the residence time specified on the heat generation
curve, and the hydrocarbon concentration and preheat temperature specified on the
heat removal line involved in the intersection. With a low combustible concen-
tration and low inlet temperature, the oxidation rate and extent of reaction
are near zero. As the inlet temperature is raised, the heat removal line shifts
to the right and will eventually intersect the generation line at an arbitrarily
high extent of reaction. As discussed below, this defines the ignition
temperature.
S-1U121
-------
182
High concentration heat removal lines have three intersections with
the heat generation curve. The central one is unstable and defines the ignition
temperature. If the temperature of the mixture is raised above this, the high
temperature steady state will be attained.
Figure 12-2 is only presented to give a qualitative picture of the
coupling between heat transfer and reaction. The temperatures shown, however,
are much lower than found to be necessary in practice (see Section 3.1.2.3).
A major reason for this is that heat losses have not been included in Figure 12-2.
These are very important, especially with a dilute fume stream incapable of
generating much heat itself (near vertical heat removal line). Heat losses
(conduction and radiation) limit the extent to which reaction time can be in-
creased to compensate for a low reaction rate. Reactions must occur rapidly
or the heat generated will be lost to the surroundings instead of raising the
adjoining fuel and air mixture to its ignition temperature. Another problem
may be that the reaction rates used to calculate Figure 12-2 were too high.
In Section 12.3.2 we compare various rate expressions for CO oxidation.
12.1.1 Ignition Temperature
The term ignition temperature can be understood in light of the discus-
sion in Section 12.1 and Figure 12-2. For high concentrations of combustibles
there are three intersections of a heat generation curve with a given heat loss
line, and the middle (unstable) operating point will define the ignition tempera-
ture. If the mixture is raised above this, the combustor will Jump to the upper
operating point characterized by nearly complete conversion and a high temperature.
Low combustible concentrations (typical of afterburner applications)
give a nearly vertical heat removal line since very little heat is generated by
the oxidation reaction. In this case there is only one intersection of the
generation and loss lines and ignition temperature is more difficult to define.
One way is to associate it with the temperature required for 9Q% (or some other
value) conversion. Another is to take the point of inflection (maximum slope)
on the heat generation curve (~50# conversion). For high activation energy
reactions these are not far different.
But it should be apparent that these "ignition temperatures" will be
extremely sensitive to experimental conditions. The measured value would be
affected by:
Changes in heat removal line
1. The concentration of combustible component (and oxidant)
2. The inlet temperature
3. The rate of heat loss from the apparatus
Changes in the heat generation (extent of reaction) line
1. The combustor residence time and flow pattern (deviations
from complete backmixing)
-------
183
2. Inhibition or catalysis by vessel walls so results depend
on vessel geometry and material of construction
Therefore, published values of autoignition temperatures should be used
only for general guidance as indicated by the wide range of temperatures
for many compounds.
12.1.2 Combustion Limits and Flame Temperatures
The upper and lover flammability or explosive limits (maximum and minimum
concentrations of fuel in air which will support combustion) are also determined
by a balance between heat generation and loss. If combustion is to be self-sus-
taining, heat generation by reaction in an ignited region must exceed heat losses
by enough to raise neighboring unburned fuel and air mixture above their "ignition
temperature." Empirically, it has been observed18'6) that sustained combustion
requires the release of at least 10 kcal per mole of combustion products above
that needed to raise these products to the ignition temperature. This provides
the driving force for transport of heat to unburned regions.
The rate of combustion and flame temperature are greatest for a stoichio-
metric mixture of oxygen and fuel since no diluent (excess oxygen or fuel) is
present to absorb heat. Also, reactant concentrations for the bimolecular reac-
tions are optimal. For pure hydrocarbons, this requires (x + 1/Uy) moles of
oxygen per mole of hydrocarbon CxHy. When air is used, the 79/21 moles of
nitrogen per mole of oxygen serve to lower both rate and flame temperature.
For most hydrocarbons, the adiabatic flame temperature for a stoichiometric
mixture of air and hydrocarbon is ~UOOO°F. As heat sink (excess air or fuel)
is added, both the flame temperature and oxidation rate drop. For most hydro-
carbons in air, combustible mixtures at room temperature range from 1/2 •* 3
times the stoichiometric concentration as measured in mole percent. Table 12-1
gives a few representative values for the reported lower explosive limit LEL,
heat content, and temperature rise for the minimum explosive mixture. As can be
seen, the LEL for these compounds vary, but the corresponding heat content and
AT are not very different. This would be expected from the discussion above.
Table 12-1. HEAT CONTEMT OF VARIOUS FUME STREAMS
AT THE LOWER EXPLOSIVE LIMIT
Bases: 379 SCF/lbmole
C^ = 8.5 Btu/lbmole°F
Combustible
Methane
Propane
Hexane
Toluene
Methyl ethyl ketone
LEL at
Room Temperature
5.03&V
2.1$v
1.2$
1.2%
1.955
Heat Content
U5.7 Btu/SCF
U8.7 Btu/SCF
52.9 Btu/SCF
51.3 Btu/SCF
U9.U Btu/SCF
AT Corresponding
to Combustion
20UO°F
2190°F
2380°F
2310°F
2220°F
-------
Table 12-1 also shows that stable flame temperatures for these repre-
sentative combustibles cannot be less than ~2100°F •* 2UOO°F. Imperfect mixing
and heat losses result in slightly more fuel than the LEL being required to
sustain combustion. If air dilution is increased to give lower flame temperatures
(mixtures below the LEL), the flame will be extinguished.
This then is the reason why unburned fuel cannot be completely mixed
with a dilute fume stream in order to achieve a lUOO°F-1500°F operating tempera-
ture. The heat content of the fuel will not be released since the maximum heat
generation is not sufficient to insure ignition of the cold mixture. For the
same reason, a stream containing ~50-60$ of the LEL cannot be destroyed without
burning supplemental fuel despite the fact that the heat content of the fume is
sufficient to raise it to lUOO-1500°F. The incoming fume must be raised to its
"ignition temperature" in order to release the available heat and oxidize the
pollutants. Preheating the fume will minimize the required supplemental fuel.
12.1.3 Dilute Fume Must Bypass Flame
The discussion in the above sections shows why dilute fumes must be
mixed with the combustion products from burning the supplemental fuel rather
than with the unburned fuel itself. Where possible, oxygen present in the fume
should be used for combustion of the fuel in order to avoid introducing outside
air which requires additional fuel to raise it to the operating temperature (see
Section 3.1.3.1). However, the remaining fume must be mixed in downstream of the
combustion zone. This amounts to i50> of the total fume in the case of 80°F fume
with negligible heating value, 100$ excess air in the burner (near the LEL), and
lUOO°F combustion chamber temperature. More fume must bypass if the burner is to
operate with less excess air or if less fuel is needed because the fume is pre-
heated or has a significant heating value. If the fume enters at UOO°F rather than
80°F, ^65% of the fume must bypass the burner, other conditions remaining the
same. Schematically, the process can be viewed (Figure 12-3) as three distinct steps.
1. Supplemental fuel combustion (generates heat)
2. Mixing of fume and combustion products
3. Oxidation of the combustible pollutants in the fume (hold
fume at temperature for required time)
The way these three steps are accomplished in available afterburners varies
considerably and leads to a large range of effectiveness from complete conversion
to COa and H20 to increasing rather than decreasing the amount of pollutants in
the fume stream. This is discussed in more detail in Chapters 13, 1^, and 15.
If the fume stream contains a high enough concentration of combustibles
to sustain combustion at the temperature it is supplied, it can be treated as
a premixed (or raw) fuel stream. In this case only a single combustion step
is required since the fume itself supplies all the required heat. For safety
reasons, such high concentration streams are likely to have little, if any,
oxygen so combustion air must be provided.
S-14121 Figure 12-3 follows
-------
xj i
00 £
Supplemental
Fuel
Outside Air
(if Used)
Dilute
Fume
Fuel Combustion
T
Mixing of Fume
and Hot
Combustion Gases
Fume to Supply Oxygen
for Fuel Combustion
(Outside Air Needed if
Fume Fouls Burner or
< ~16% Oxygen)
Retention of Fumes
at High Temperature
for Sufficient Time
Clean
Effluent
Figure 12-3. STEPS REQUIRED FOR SUCCESSFUL INCINERATION OF DILUTE FUMES
-------
185
12.2 Mechanisms of Oxidation and Combustion
Dilute fume incineration involves oxidation reactions at combustible
concentrations and temperatures much lower than are normally studied in combustion
research. As discussed above, flames cannot be supported and kinetic rates and
mechanisms determined using flames may not be applicable. In addition, the normal
operating temperatures 1200-1500°F are of the same order as the transition between
"cool flame" kinetics (where peroxides are important chain carriers) to hot flame
kinetics (involving excess concentrations of 0, OH, and H). High temperature
mechanisms are discussed first.
12.2.1 Free Radical Flame Reactions
Within flames, oxidation does not take place by reaction of molecular
oxygen with fuel but rather through free radical mechanisms. This reflects the
near zero activation energy of free radical oxidation reactions. For instance,
the activation energy for the oxidation of carbon monoxide by OH radicals
CO + OH -»• C08 + H (12-4)
is only 1 kcal/mole, whereas that for the direct oxidation by molecular oxygen
CO + 08 -»• COa + 0 (12-5)
is U8 kcal/mole. In addition, in flames. one finds OH, 0, and H radical concen-
trations orders of magnitude higher than would be predicted by the dissociation
equilibrium for HaO, Os, and Ha. This can be explained by the high rates for
the chain branching reactions (two radicals produced for every one involved).
H + 08 •* OH + 0 (12-6)
0 + H80 -»• OH + OH (12-7)
0 + Ha -»• OH + H (12-8)
H + HaO -»• Ha + OH (12-9)
A partial equilibrium is quickly established between radicals via these reactions
and the reverse of these. However, the much slower three-body recombination
reactions
OH + 0 + M -*• H80 + M (12-10)
H+H+M+Hg+M (12-11)
S-It121
-------
0 + 0 + M-»-Oa + M (12-12)
vhere M represents the third (energy absorbing) body
do not return the system to complete equilibrium with molecular species until
sometime after all fuel has been consumed. Typically, peak OH concentrations
in flames (~3000°F) are in the range .01 •* .05 mole fraction and may be as high
as .10. The corresponding equilibrium mole fractions are one or two orders jf
magnitude lower, ~.001 •* .002. At lower temperature, the equilibrium values
are even lower, e.g., OH mole fraction ~10~7 SlUOO°F. Combustion reactions can
be quenched by a drop in temperature which drastically reduces rates or by the
presence of a high concentration of third bodies, e.g., high surface area or many
inert molecules.
Radicals can be generated by initiation reactions involving slow reac-
tions between fume and fuel. However, in continuous flow combustors these
initiation steps are not very important since radicals are carried from burning
regions to unturned regions by mixing and diffusion. (Flame stabilizing bars
or discs are used to form stable vortices in their wake which provide backmixing
of heat and radicals. This stabilization can also be accomplished aerodynamically.
Chapter 13 contains additional discussion of flame holding.)
Initial attack of the vaporized fuel molecule in an oxidizing flame is
probably by an OH radical. For methane, this is
CH4 + OH -*• HaO + CHa (12-13)
Then the chain continues
CH3 + Oa * CH80 + OH (12-lU)
or
0 + CH3 •* CH80 + H (12-15)
forming formaldehyde. This reacts further to give CO
CHaO + OH -»• HaO + CHO (12-16)
and
CHO + OH •* HaO + CO (12-17)
which is oxidized to COa via reaction 12-U
CO * OH * COa + H (12-U)
S-1U121
-------
187
All evidence points to CO oxidation as the slow step in oxidation of hydrocarbons.
As discussed in Section 3.1.2.3, this rate often controls the required residence
.time in afterburners.
Hydrocarbons composed of more than one carbon will undergo a similar
sequence, with organic acids and aldehydes appearing as intermediates. With hydro-
carbons of more than 3 or U carbons, the radical formed in the equivalent of
reaction 12-13 is unstable and is likely to undergo pyrolytic decomposition
(endothermic) and/or dehydrogenation.
RCHaCHa' * R'CH = CHa + R" • (12-18)
or
RCHaCHa' * RCH = CHS + H (12-19)
As a result, aldehydes containing more than two carbon atoms are not likely to
be found in flames. Aldehyde concentration in flames is generally very low since
all steps except the oxidation of CO to C0a are very fast at flame temperatures.
If concentrations of 0 and OH are low (poor mixing or fuel rich flames), dehydro-
genation may be accompanied by polymerization to yield material of a high C/H
ratio—smoke and soot. Steam is often injected to reduce soot formation. This
helps to increase OH concentration via reactions 12-7 and 12-9-
12.2.2 Low Temperature Oxidation
At very low temperatures oxidation rates can be taken as zero. However,
for temperatures between ~500°F and the "ignition temperature" (see Section 12.1.1)
relatively slow chain reactions occur which yield "cool flames" for combustible
mixtures. As in high temperature flames, OH radicals are a major oxidizing
species but in this case their concentration is much lower and peroxide radicals
(H0a) play a major role. Combustion times at 500-700°F are on the order of 10-30
minutes rather than milliseconds as in normal flames.
To illustrate cool flame mechanisms we shall again use methane as an
example. Initiation occurs by
CH4 + Oa •*• CH3 + HOa (12-20)
which is very slow. A chain reaction is set up with initial attack on CH* via
CH4 + OH -»• HaO + CH3 (12-21)
and
CH* + H0a •* H80a + CH3 (12-22)
Then methyl radicals react with oxygen to give formaldehyde.
S-1IH21
-------
188
CH3 + Oa * CHS0 + OH (12-23)
which leads to branching by
CHaO + Os * CHO + H0a (12-2U)
Formaldehyde can also react with OH and HOa radicals
CHaO + OH -»• HaO + CHO (12-25)
CH80 + H08 •* H80 + CHO (12-26)
The oxidation continues to CO via
CHO + Oa -»• CO + HOs . (12-2?)
CO oxidation is very slow at low temperatures but probably occurs via an ozone
intermediate or other excited species.
CO + 0 + M •* COs + M (12-28)
As can be seen, some of the reaction steps are the same at low and high
temperatures. However, the rates are much slower and concentrations of OH and 0
are much lower. As a result, the less active peroxide radicals play an important
role. Also, aldehydes are much more prevalent and provide for chain branching
through reaction 12-2U.
Higher hydrocarbons follow a similar sequence with corresponding
aldehydes, acids, alcohols, and peroxides formed as intermediates. As with
higher temperature processes, large hydrocarbon molecules tend to decompose
following the initial hydrogen abstraction step.
12.2.3 Thermal Oxidation in the Absence of Flame
The above discussion shows the important role free radicals play in
the oxidation of hydrocarbons to COa and HaO. It has also been pointed out that
radical concentrations in flames are one or two orders of magnitude above equilib-
rium values with correspondingly high reaction rates.
Dilute fume incineration, however, involves thermal oxidation in the
absence of flame. The reasons for this were discussed in Section 12.1.3. In
afterburners radical concentrations are expected to be much closer to equilibrium
values once excess levels from the supplemental fuel burner have decayed. Chain
branching reactions dominate only while fuel is being consumed. Within the
combustion chamber combustible concentration in the fume will determine the amount
by which radical concentrations can exceed the overall equilibrium values set by
the temperature and concentrations of water and oxygen. Radicals generated oy
S-14121
-------
189
reactions such as 12-lU or remaining from the fuel burner will tend to recombine
by reactions 12-10 •* 12-12 rather than lead to branching. The lower the concen-
tration of combustibles, the nearer the radical concentrations will be to overall
equilibrium values since chain terminating collisions of radicals with third
bodies are more likely than reacting collisions with hydrocarbon molecules. In
Sections 12.3.1 and 12.3.2 CO and methane oxidation rates axe compared for flame
conditions (excess free radicals) and for radical concentrations at equilibrium
values. Experience with afterburner operation indicates that actual oxidation
rates are somewhere between these extremes as would be expected from the above
discussion. Also, very dilute fumes are more difficult to clean up than more
concentrated fumes.
The concept of "flame contact" would appear to have much value since
the excess free radicals formed in the supplemental fuel flame could be utilized
to oxidize the combustible pollutants in the fume. Typical decay times for these
excess radicals are on the order of .01 -*• .03 seconds. However, this is of the
same order as CO oxidation in the flame and attempts to rapidly mix fume and flame
on this time scale will lead to flame quenching. As discussed in Chapter lU,
normal mixing times are an order of magnitude longer.
S-1U121
-------
150
12.3 Oxidation Kinetics
i
Once fume and combustion gases have been completely mixed and brought
to a temperature above the ignition temperature, the oxidation reactions must be
given sufficient time to occur before the gases are quenched in a heat exchanger,
by air dilution, or by exiting through a stack. As is discussed in Section 12. U,
fuel combustion in flames is essentially controlled by mixing rates, and oxidation
kinetics usually are of secondary importance. In afterburners, however, low
temperatures and reduced radical concentrations make oxidation rates comparatle
to mixing rates. In this section we give some estimates for these required times.
Data is extremely sparce, especially at temperatures and concentrations of interest
in fume incineration. Therefore, a certain amount of design flexibility and test-
ing is required to allow even the most stringent effluent criteria to be met.
12.3.1 Hydrocarbon Destruction
The most common use of afterburners is to destroy solvents or other
hydrocarbon fumes and odors. In the combustion of fuels it has been found that
the oxidation of CO -»• COa (reaction 12-U in Section 12.2.1) is the slowest step
and the kinetics of the steps leading from the original hydrocarbon to CO can
be ignored as a first approximation.18"7' Temperature levels in afterburners are
much lover than in flames so all rates are lower, but carbon monoxide oxidation
again seems to control. A possible exception is methane oxidation which has a
very high activation energy (~60 kcal/mole), and except at flame temperatures
(>3000°F), the oxidation of methane to CO is considered rate limiting.la~8'
The curves shown in Figures 12-U and 12-5 represent18'9' experimental
data taken by Surface Combustion Division of Midland Ross Corporation. Measure-
ments were made in the absence of a flame and at temperatures and concentrations
of interest. Prevaporized solvent was added to air heated to the desired operating
temperature. Mixing was designed to be nearly instantaneous, and plug flow was
maintained in the reactor. As can be seen, the rate of hydrocarbon disappearance
was taken to be first order in solvent concentration and somewhat dependent on
solvent type. Toluene, which is often reported to have a very high ignition
temperature, disappeared more slowly than cyclohexane or hexane.
Figure 12-5 shows the rate at which toluene is oxidized as a function
of afterburner temperature. Rates are slow below lUOO°F but increase rapidly
above this temperature. (The asymptote shown for concentrations <10 ppm may be
real, but more likely is a sampling and analysis problem.)
As can be seen from Figures 12-U and 12-5, the rate of toluene destruc-
tion is very fast at temperatures greater than lUOO°F. Conversions greater than
95% can be obtained with residence times of .1 second or less without any need
for flame contact. However, several problems remain.
1. These data were taken in a plug flow reactor so all solvent was
given equal treatment. There were no mixing delays and no cold spots.
2. The measurements were based on flame ionization detector readings
(see Section 10.3) and only represent rates of hydrocarbon disappearance.
S-1U121 Figures 12-U and 12-5 follow
-------
102
10
<5
u
£
u
10
I 1
10
1-1
Incineration Temperature 1410°F
0.05 0.1 0.15
Time, seconds
0.2
0.25
Figure 12-4. HYDROCARBON OXIDATION RATES
IN ABSENCE OF FLAME
S-14121
67784
COURTESY: Surface Combustion Division - Midland Ross Corp.
& The Heating and Ventilating Engineer Journal 12*)
-------
104
Toluene Added To Heated Air
1390°F
103
0
to
CL
0.
q
g 10*
§
o
o
0)
-------
191
Aldehydes and CO do not show up. At these temperatures aldehydes should be
rapidly oxidized but CO conversion is rather slow. Figure 12-6 is an estimate
of the equivalent first order oxidation of CO compared to toluene from Figure 12-U.
Methane is one of the most difficult hydrocarbons to oxidize either
thermally or catalytically. Its rate of combustion has been measured many times
but always at high concentrations and temperatures. An estimate for this rate
.12-107
•d[CH4] _ i. 7
jj_ ~ H* [
at
ie
-56600/RT
i/8
R'T
gmoles/cm sec (12-29)
where P [=] atm
R1 = 82.06 cm3atm/gmole°K
R = 1.986 cal/gmole°K
T [=] °K
Xi = mole fraction of i
Using this equation, the extent of conversion versus temperature has been plotted
in Figure 12-7 for lQ% Os and 6% H20. It should be stressed that this equation
has been extrapolated far below its intended range of methane concentration >5$
and temperatures >2000°F. Again, there is only CH4 oxidation to CO. Conversion
of CO -»• C0a is not included.
Methane oxidation rate can also be calculated from the rate of the
reaction
CH4 + OH -»• CH3 + HaO
(12-lU)
This rate expression is
dJ[CH
dt
ki4[OH][CH4]
(12-30)
As discussed in Section 12.2.3, radical concentrations in afterburners are close
to equilibrium values. Therefore, the concentration of hydroxyl radicals [OH]
can be taken at its equilibrium value for
and
This gives
0 + HS0 7 OH + OH
(12-13)
(12-8)
8-14121
-------
192
K 1/anH n
- _S .H8° Og
nOH - Kiai/4(R.T)i/4
where n^ = gmoles of component i
R1 = 82.06 cm3atm/gmole°K
T [=] °K
KJ_ = equilibrium constant for reaction 12-i
Taking the rate of reaction 12-lU as18'11)
ki4 = 50 x io-13e-9900/ET cm3/gmole sec (12-32)
we get
alSSjl = 2.82 x 1016e-UT200/RT[CH4][Oa]1/4[H80]1/a (12-33)
CLo
where [ ] = concentration in gmoles/cm3
Figure 12-7 also shows the extent of conversion versus temperature using this
equation.
12.3.2 Carbon Monoxide Oxidation
As previously noted, the oxidation of CO -»• COa is usually the rate
limiting step in premixed flames and is expected to be the slowest step in the
destruction of gaseous pollutants in afterburners as well. There are a great
variety of rate expressions given for this reaction, and these are discussed
below. Afterburner experience shows that temperatures in excess of lUOO°F are
required for satisfactory CO cleanup in .5 seconds total residence time (including
mixing time). Figure 12-6 compares the estimated rate of CO oxidation with that
shown in Figure 12-U for hydrocarbons. It is obvious that longer times and/or higher
temperatures are needed when CO cleanup is desired rather than merely hydrocarbon
destruction.
In the absence of water, CO is extremely difficult to burn. This is
because no OH radicals can form to oxidize CO via reaction 12-U. Therefore, the
much slower reaction of CO with Og is required. Afterburners are generally
employed to clean up fume streams containing water vapor and solvents composed
of hydrogen as well as carbon. In addition, the combustion of methane generates
much water. Therefore, adequate water is always expected to be present.
As with methane, carbon monoxide oxidation kinetics have been determined
by many investigators, but concentration and temperatures were much higher than
those of interest in dilute fume incineration. The rate expressions developed
S-14121 Figures 12-6 and 12-7 follow
-------
\
*
\
\
\ Estimated CO Destruction
\
\
\
\
\
1410°F
Afterburner
Temperature
\
\
Hydrocarbon Destruction
from Figure 12-4 12'9)
0.2
0.4
0.6
Time, seconds
Figure 12-6. COMPARISON OF HYDROCARBON AND CARBON
MONOXIDE DESTRUCTION RATES
S-14121
67784
-------
O CO
XJ I
NJ
200
temperature, °F
1600
2000
2400
Conversion Based on Completely
Mixed Combustion Chamber with
T = 0.1 Second.
( Does Not Allow for CO — CO2
Conversion)
750
1000 1250 1500
Temperature, °K
Figure 12-7. METHANE OXIDATION
Comparison of Rate in Flame with Rate Calculated for Equilibrium (OH)
2000
-------
193
ir -eY«:rs.l of these flame studies are given below and these have been used to
calculate the extent of CO oxidation at various temperatures for 6% water, 18J5
oxygen and .1 second residence time. These are plotted in Figure 12-8. Again,
~he rate expressions have been extrapolated far below their intended temperature
and concentration range and should be suspect. As can be seen, there is some
disagreement between the various expressions.
Field, et al., 1J3~is' surveyed several of the findings of investigations
and give the expression
-d[CO] = - 1Qlo .3 .B /_P_\1<8 -16000/RP ,,p -M
dt J * -LV XcoX0a XHa0 VR«T C0a + H (12-5)
S-1U121
-------
is given by '
k6 = .71 x 10'13 e-T700/RT cm3/gmole sec (12-38)
Using the equilibrium concentration of OH radicals given by Equation 12-31, we get
"dj£°l = l» x 1013[CO][0S]1/4[H20]1/S3 (12-39)
dt
where concentrations are given as gmole/cm3
It can be seen that this equation gives very much lower rates of CO
oxidation than the flame determined data. Actually observed rates in afterburners
tend to lie in between these two extremes. The CO oxidation rate shown in
Figure 12-6 is representative of expected rates, and this is about halfway between
the extent of reaction predicted by Equations 12-36 and 12-39-
Equation 12-39 has also been used to calculate the extent of reaction
vs time for a constant temperature of 1250°K. The results for a plug flow and
completely mixed combustion chamber are shown in Figure 12-9- This shows the
importance of obtaining a high enough temperature since if backmixing is important,
it is difficult to achieve high levels of conversion even for long residence times.
Of course, the rate given by Equation 12-39 is unrealistically low, but a similar
plot in practice would correspond to temperatures of 1300-lUOO°F.
All of the expressions reviewed above for the rates of oxidation of
hydrocarbon and CO indicate a significant dependence of the oxidation rate on
the oxygen content of the stream. In normal afterburner application, the waste
stream is essentially contaminated air, with an oxygen content of 15-20$m even
after passing through the afterburner. However, in some potential applications,
the waste stream comprises combustible contaminants in an inert carrier
(e.g. N or C02); and air (as a source of oxygen) must be blended with the
waste stream before passing through the afterburner. The amount of air added
must be sufficient to bring the oxygen content of the stream being treated up to
a reasonable level, if the oxidation rate is to be high enough for adequate con-
version under normal afterburner temperature and residence time conditions. Some
equipment manufacturers have specified 1* - 5$m QZ as a minimum in the after-
burner effluent. The kinetic expressions indicate that lower concentrations could
be used, but at the expense of increasing temperature or residence time to
achieve the necessary conversion.
S-1^121 Figures 12-8 and 12-9 follow
-------
tsj
o
"o
(0
0)
on
-s
"c
LLJ
Temperature, °F
1200 1600
2000
2400
Conversion Based on Completely
Mixed Combustion Chamber with
T = 0.1 Second.
750
1000 1250
Temperature, °K
Figure 12-8. CO OXIDATION
Comparison of Flame Measured Rates with Rate
Calculated for Equilibrium OH Concentration
1500
1750
-------
0.15 0.2
Reaction Time, seconds
0.25
0.3
0.35
Figure 12-9. CO OXIDATION RATE
Plug Flow vs Completely Mixed Combustion Chamber.
Rate Calculated for Equilibrium (OH) .
-------
12.3.3 Liquid Smokes and Droplets 195
Combustion of liquid smokes and droplets requires an additional step
compared to gaseous pollutants since vaporization must precede oxidation. All
organic "liquids" which are present in fume streams are expected to have boiling
points below 600-700°F. Therefore, vaporization should take place rapidly at
typical afterburner temperatures of ~1UOO°F. In fact, when smoke emissions are
eliminated using afterburner temperatures of 800-900°F, vaporization is the major
phenomenon occurring. The effluent is invisible but analysis would find high
aldehyde and CO concentrations since oxidation is very slow at these temperatures.
Liquid fuel combustion involves sequential vaporization and burning in
the gas phase so the problem has been extensively studied. An excellent review
was recently published.ls~13) The rate at which liquid droplets vaporize and
burn is dependent on droplet size and the rate at which heat can be transferred
to the surface. Rapid mixing and radiation of heat from the combustion chamber
walls or flame give a high rate of heat transfer.
Time required for vaporization and destruction of original hydrocarbon
can be estimated byls~14)
S-11H21
-------
196
29800
(12-UO)
where pft = oxygen partial pressure, atm
Ug
d = original drop diameter, cm
T [=] °K
M = molecular weight
For a 200 molecular weight hydrocarbon, this formula predicts burning times at
1000°K (13UO°F) of .OOU6 seconds for a 50 micron drop and .0185 seconds for
a 100 micron drop. If temperature is raised to 2000°K (3lUO°F), which is a typical
flame temperature, lOOy drops burn in .0055 seconds. These times are of the same
order as are commonly reported for single droplet combustion. Afterburners
should not have to handle droplets larger than 50 -»• lOOy since these are easily
removed in simple cyclones. Therefore, the additional time required to burn up
droplets rather than gases (<.02 seconds) is not an important consideration in
sizing the afterburner.
12.3.1+ Soot and Combustible Particulates
As contrasted to liquid smokes, solid carbon, soot and other combustible
particulates are expected to increase the required afterburner temperature and
residence time. For low temperatures (<2000°F) and small particulates (<100y)
the rate of burnup is controlled by the rate of reaction rather than diffusion.
This is expressed as a surface reaction rate q [=] g/cmssec. If one assumes
spherical particles, this results in a linear burning rate variation with particle
size
•OW
D(p TT d3)
6 Dt
where D indicates differentiation
p = density of particle
d = particle diameter
Integration gives the time for complete burnout as
Field12"la) has reviewed data relating to the combustion of pulverized
coal and coal char and found that most results can be correlated by
q = 8710 e-35700/RT P (12-U2)
-------
197
where q [=] g/cmasec
T [=] °K
= partial pressure of oxygen,
R = 1.987 cal/gmole°K
atm
This has been used by Niessen, et al. , S-1B) to predict burnup of smoke and char
particles in incinerators. For oxygen partial pressures of .1 atmospheres, this
equation predicts the burning times given below.
BURNING TIME FOR COAL CHAR
Temperature
lU6o°F
2000°F
25UO°F
Particle Size
100 y( micron)
U.3 seconds
1.3 seconds
. lU seconds
10y
. U3 seconds
.13 seconds
.OlU seconds
iy
.01*3 seconds
.013 seconds
.OOlU seconds
The numbers in this table have been generated by extrapolating Equation 12-U2 far
below the temperatures and particle sizes used in experiments which were the basis
for the equation. In general, particle sizes were in the range used in pulverized
coal boilers (30-60y) or larger (up to 2000y). Temperatures were typical of flame
combustion, 2000°F and higher. Therefore, the burning times should be taken as
estimates only.
These char burnup rates do not appear to accurately reflect the rates
of soot combustion. Lee, Thring, and Beer, ~16' Fenimore and Jones,la~y ' and
Magnussen18-18) have measured the oxidation rates of soot formed in fuel-rich
hydrocarbon flames. (Once formed, the soot is burned with an excess of air.)
They find that the surface reaction rate,q, for soot is approximately two orders
of magnitude below that given by Equation 12-U2. The reasons for this may be
the small size of soot particles,
-------
SOOT BURNING RATE
(pQ = .1 atra)
Temperature
1520°F
1TOO°F
2060°F
2600°F
Particle Size
lOy
1750 seconds
UOO seconds
U3 seconds
h seconds
IP
175 seconds
Uo seconds
U.3 seconds
. U seconds
• iy
17.5 seconds
U.O seconds
. U3 seconds
.Oh seconds
Fenimore and Jonesla~17 ' correlate their results in terms of collision
frequency between OH radicals and soot particles. If OH concentration is taken
as the equilibrium value for p0g = .1 atm and pR Q = .08 atm, one calculates soot
burnup times roughly 1/2 of those given in the previous table. Magnussen's
correlation13'18^ agrees fairly well at temperatures above 2000°F, but predicts
much lower rates than shown in the table at low temperatures. Oxygen concentra-
tions of 21$ will double the predicted rates.
This has serious consequences for afterburners designed to eliminate
dense carbon smokes like soot. It may be necessary to use very high temperatures
and long residence times to achieve burnup. Experience with wire burnoff, cloth
carbonization, and smelting operations shows that temperatures of over 1700°F
and residence times greater than .5 seconds are often needed to clean up the smoke
plume. However, refuse incinerator smoke can often be eliminated at temperatures
below 1500°F. The amount of volatiles present, particle size, and surface
properties all help determine the burning rate.
12.3.5 Flame Inhibition and Combustion of Substituted Hydrocarbons
As previously mentioned (section 12.1), any diluent will reduce the
rate of combustion by absorbing heat and lowering the concentration of the reacting
species, thereby making it more difficult to maintain a flame. In this way,
nitrogen, carbon dioxide, water (except when present in small amounts during
combustion of hydrogen free fuels), and other inerts inhibit combustion. Since
both reactant concentration and heat balance are affected, the effect on combus-
tion rate is fairly strong, e.g. , maximum flame speed for propane is decreased
by a factor of 3 as the oxygen concentration in the enriched combustion air is
halved from hO% to 20$.
Certain substances, however, exercise much stronger effects even when
present at a very low concentration. Halogen and phosphorous compounds act in
this way, and as a result, they are used as flame retardants and fire exti^guish-
ants. In fume incineration, these same substances sometimes must be burned.
S-11*121
-------
199
The mechanism involved in this later form of flame inhibition is still
not fully understood but it is thought to be primarily through interference in
the free radical, chain branching mechanism. Recently, a literature review was
published by Creitzla~ao' giving complete coverage of the experimental findings
and mechanistic interpretations of inhibition by halogen compounds. One of the
more detailed analyses of the mechanism of flame inhibition was presented by
Rosser, Wise, and Miller.ls~81' They claim that most inhibitors first decompose
to yield a halogen atom which forms the halogen acid by reacting with the fuel.
The halogen acid then reacts(with )one of the chain branching species (H, OH, or 0)
to replace it with a halogen atom and prevent branching, e.g.,
CH3Br •* CH3 + Br (12-UU)
CH4 + Br •*• CH3 + HBr (12-U5)
OH + HBr •* HaO + Br (12-U6)
This mechanism has not been proven but it does form the basis for most discussion.
Fenimore and Jonesia~aa) consider the mechanism to be primarily through a reduc-
tion in concentration of H atoms.
H + CH3Br -»• CH3 + HBr (12-U?)
with no further effect assigned to the halogen acid. Inhibitors are more effective
in rich flames than lean flames since the concentration of H atoms is high and
10— a j
hydrogen plays a more active role in the combustion process. In fuel lean
flames, the concentration of H is low but reactions 12-6 -*• 12-9 are in local
equilibrium so a reduction in H results in reduced OH and 0 as well.
Phosphorous compounds and various metal salts can also serve as flame
inhibitors, probably again by interferring with chain branching mechanisms and
serving as third bodies for recombination reactions. They are widely used as
fire retardants in fabrics and paints. Also, high surface area in the combustion
chamber will inhibit oxidation by speeding recombination reactions. If the
inhibiting surface is also a heat sink, quenching is likely to occur.
Many studies have been conducted to determine the extent of inhibition
for low concentrations of various halogens. This is reported as the flame speed
for premixed flames at various air to fuel ratios with various amounts of inhibitor
in the mixture. Flame speed or propagation velocity is a measure of the kinetics
of the reaction but.cannot be directly interpreted in terms of rates of reaction.
It is measured18'83' by optically determining the surface area of the flame and
dividing this into the volumetric flow rate of reactants. Typically, halogens
show a 2-20% maximum reduction in methane flame speed per 0.1 mole percent of
inhibitor added. Bromine is one of the most effective of the halogens. There
S-
-------
200
is much scatter among the various investigators, at least in part because flame
speed is dependent on the experimental set up, initial temperature, and gas flow
rate.
Substituted hydrocarbons generally have a lower heat of combustion and
are more difficult to burn. Methane has a heat of combustion of 23,875 Btu/lb,
while that for CH3C1 is 5850 Btu/lb and that for CH3Br is only 3^92. Methyl
bromide is barely combustible with range of only ±1/2$ of stoichiometric mixture.
Certain wastes are not sufficiently combustible to support combustion alone and
supplementary fuel must be added. CHCla with a heat of combustion of 13^5 Btu/lb
will burn alone only in pure oxygen and even then, only under carefully controlled
conditions.
A few studies have been conducted in which the flame speeds for pure
halogen compounds were measured. In order to sustain a flame, oxygen concentra-
tions greater than 21$ must often be used. In pure oxygen, the flame speeds of
the various chlorinated methane compounds compared to pure methane were
~l/3 for CH3C1, -1/8 for CHaCla, and -1/UU for CHC13. Combustion is inhibited
both by interference in the mechanism and a lower rate of heat generation per
mole of reaction.
From the above discussion one can get a feel for the fact that in general
substituted hydrocarbons are more difficult to oxidize than unsubstituted or
oxygenated species. An afterburner handling these species must allow for opera-
tion at a higher temperature and/or a longer residence time. Some preliminary
test work should be conducted to determine design specifications.
12.3.6 Elementary Reactions - Kinetic Models
As noted in the previous sections kinetic data applicable to fume
incineration are relatively limited. The overall rate expressions allow order of
magnitude calculations but there is disagreement between various experiments.
Also, temperatures and concentrations are far different than those of interest.
Prediction and extrapolation can be done with more confidence by using a model
based on the actual mechanisms involved in the combustion. This involves using
the elementary reaction steps (free radical reactions) discussed in Section 12.2.
Computer programs for integrating large systems of non-linear differential equa-
tions have been adapted to handle reaction rate expressions and many kinetic
models of flames have been developed. These have been especially important in
understanding nitrogen oxide formation in flames. Heat generation and simple
mixing processes can be accommodated. This type of model cannot be used (given
current knowledge) for a priori prediction of reaction rates but it holds much
promise as a semi-empirical tool. Difficulties involved with a completely
theoretical approach lie mainly in determining the reaction mechanism (to decide
which elementary reactions should be included in the model), and in determining
the reaction rate constants. Methods are available for estimating these rate
constants for simple reactions (e.g., hydrogen abstraction) and many rates have
been determined through mass spectrometer studies of flames or in non-flame
experiments. Some data are given in References 12-25 and 12-26. (Reference
12-26 presents the results of using a kinetic model to calculate CO oxidation
5-1^121
-------
201
rates in an automobile exhaust reactor.) It is most likely, however, that this
type of model will be utilized for interpreting experimental results and making
them useful in design. In this way, a minimum amount of experimentation will be
required and interpolation and extrapolation can be done with confidence.
S-1M21
-------
202
12.U Mixing Processes - Control of Overall Oxidation Rate
The oxidation rates discussed in Section 12.3 assume that the combustible
component and oxidant are well mixed and in molecular contact at time zero. In
most cases of fuel combustion, fuel and oxidant are not premixed and combustion
rates are controlled by mixing rates. (At flame temperatures, the initial hydro-
carbon will disappear in <1 millisecond and even CO oxidation only takes 5-10
milliseconds once mixing has occurred.) As discussed in Chapter 13 flame shapes
can be predicted quite well for simple geometries by calculating the zone where
mixing between fuel and air is Just sufficient to produce a (stoichiometric) mixture,
The chemical kinetics can be ignored. As is discussed in Chapter lU, mixing time
can also be an overriding consideration in afterburner design. Poor mixing is
often the reason for poor performance.
Mixing processes are discussed in detail in Chapters 13 and lU. However,
it is of interest to estimate the effect of mixing time on the rate of combustion
from observed versus theoretical combustion intensities as measured in Btu/ft3-hr.
It has been determinedls-a7 / based on bimolecular reaction theory that the maximum
theoretical combustion rate for a stoichiometric mixture of butane and air with
instantaneous mixing is ~T30 x 106 Btu/ft3-hr. This is for optimal addition of
premixed reactants to the combustion zone which gives a temperature ^80% of the
adiabatic flame temperature. The maximum combustion intensity which has been
observed experimentally s-1'is -100 x 106 Btu/ft3-hr using an intensely mixed
(uniform temperature and concentration) spherical combustor. The highest rates
attained in practice, ~10 x 106 Btu/ft3-hr, are in highly efficient gas turbine
combustors employing strong recirculation patterns (to transport heat and radicals
to incoming reactants) and rapid mixing. Normal burners, even with some recircu-
lation, usually exhibit combustion intensities of 0.5 x 106 Btu/ft3-hr or less and
most combustion chambers are designed for heat release rates an order of
magnitude lower than this. Thus, we can see that mixing times appear to be more
than two or three orders of magnitude slower than chemical kinetic times, at
least at flame temperatures. Within an afterburner, times for mixing and complete
oxidation are closer to the same order.
12.5 Flue Gas Composition
Given sufficient reaction time, equilibrium considerations will determine
flue gas composition and may effect the choice of design variables and operating
conditions such as temperature, amount of excess air, use of supplemental fuel.
When dealing with pure hydrocarbons in the presence of slightly more than the
stoichiometric amount of oxygen, equilibrium is essentially completely toward
CO2 and HaO. Any presence of intermediates in the flue gas is due to quenching
of the oxidation reactions and is due to kinetic rather than equilibrium
considerations. Equilibrium considerations are much more important when consider-
ing the fate of halogens, phosphorous, metals, nitrogen, sulfur, or other non-
hydrocarbon components introduced into the fume incinerator.
S-1U121
-------
203
12.5.1 Determining Combustion Equilibrium
Equilibrium constants can be generated for most any desired reaction
by using the thermochemical data contained in the JANAF Tables.18 a8' Among
other data, one can obtain free energy data as a function of temperature for
compounds with respect to their elements. This can be used to calculate equilib-
rium constants since
= e
(12-U8)
where AF = Ev^F^ - E.v^Fi = change in free energy
i products *• reactants due to reaction
F. = free energy of component i
v. = stoichiometric coefficient of component i
K = equilibrium constant at constant pressure
Much of the information in the JANAF Tables is also available on a magnetic tape
for use with computer programs requiring thermochemical data. Computer programs
are available which utilize this data to calculate equilibrium compositions for
large numbers of components by minimizing total free energy for the system.
12.3.2 Fate of Heteroatoms
When the fume stream contains contaminants other than plain hydrocarbons
or oxygenated species, the flue gas may require treatment besides incineration.
This may entail wet scrubbers, cyclones, filters, or other particulates removal
devices. In deciding whether or not such cleanup is required, one must determine
the concentration and form in which heteroatoms and inert particulates will be
found in the flue gas. Some guidance as to maximum acceptable concentrations
was presented in Section 3.1.2.7- The presence of acids in the flue gas and
combustion chamber will introduce special corrosion problems unless planned for
in the design (see Sections 7.2 and 7.3).
When halogens are present in the waste to be burned, they can appear in
the flue gas as either the corresponding acid or as free halogen. In most cases,
it is desirable to maximize conversion to the acid, since it is more easily
removed from the flue gas and results in fewer corrosion problems.
The equilibrium distribution between the halogen and its acid is deter-
mined by the following equation, shown for Clg/HCl.
2 HC1 + 1/2 Oa i Cla + H30 (12-U9)
Hydrochloric acid is favored by low levels of excess air, high levels of water
vapor (a high H/C1 ratio in the total feed to the incinerator, including
-------
'_m'. nt»iL lu',1 ••jri'J .f.r.'jrn j -, '.^uivalont ) , and high temperature since the
204
'.quilibrium constant, L,hift', toward IIC1 as temperature increases. (K-j for reaction
12-1*9 is -2.0 at 1000°F, ~.l at 1700°F, and -.02 at 2UOO°F.) At 2000°F and 255?
excess air with a 2:1 H/C1 mole ratio in the feed, the equilibrium ratio of Cla/HCl
is .016. If the temperature is dropped to 1700°F this ratio increases to ~.04. If
the H/C1 ratio is increased to 2.5:1, the Cla/HCl ratio drops to .011 at 2000°F.
Incineration of dilute fume containing some halogenated solvents involves high
levels of excess air and low temperatures and probably results in the formation
of considerable free halogen. Equilibrium ratios of Brg/HBr are two orders of
magnitude greater than Clg/H Cl under similar conditions. Efficient combustion
is essential when handling halogen substituted organics to avoid formation of
poisonous components such as phosgene, COCls. This is especially a danger when
burning at low temperatures with insufficient oxygen in the burning region.
Organophosphous compounds break down to give PgOe in the flue gas .
This has a very low vapor pressure below ~700°F and forms submicron hygroscopic
solid particles below this temperature. These exit as a phosphoric acid mist when
water is present and result in a highly visible plume, unless concentrations are
below 1 mg/SCF. Corrosion of metallic components and refractory can be severe
unless proper materials of construction have been utilized.
Metals which are present in the waste stream are most likely to be
converted to the corresponding oxide if the incineration temperature is high
enough. (inert particulate matter is most likely already present as an oxide
and will not be changed. ) Some of the metal oxides will react with water to
form hydroxides and will be found in the flue gas as a fine mist. Sodium and other
alkali metals usually appear as hydroxides. Whatever form the metals take in
the flue gas, they can be expected to be highly visible, submicron drops or
particles.
Sulfur and nitrogen are other components likely to be found in wastes.
The fate of sulfur in flames has been extensively studied since it is present in
fuel oil and is the cause of our sulfur oxides pollution problems. In an oxidiz-
ing flame, the sulfur is primarily converted to SOa with 10/2 or less appearing
as SOa. Nitrogen oxide formation is a problem whenever combustion occurs since
some atmospheric nitrogen is oxidized along with much of the nitrogen in the fuel.
Afterburners do not seem to be a significant source of nitrogen oxides due to
the relatively low temperatures at which they operate. Below 2000°F very little
NO (indicating various oxides of nitrogen) is formed due to low oxidation rates.
Only in the flame is there much of a problem and rapid mixing with cold fume
limits the amount which can be formed. Most of the chemically-bound nitrogen
in the waste can be expected to form NO along with that formed by nitrogen fixa-
tion in the burner flame. When burning natural gas in the preheat burner,
less than 100 ppm NOX would be expected to be formed due to nitrogen fixation
in the primary flame products from the burner. Subsequent dilution of these
flame products with the fume stream being treated would reduce this NOX
concentration to 25-30 ppm or less in the afterburner effluent.
S-14121
-------
205
Chapter 13. BURNERS
13.1 Introduction
Chapter 3 has given general description, performance, and selection
information on flame-type afterburner systems. Chapter 12 immediately proceeding
has described the general principles of combustion, mixing, and retention at
temperature which apply to the flame-type afterburner. The following section
describes the physical equipment in which the operating principles are carried
out. It is intended to give more detail and a broader basis for understanding
that lies behind the description of Chapter 3.
In this chapter the convenient breakdown into separate steps, as shown
in Figure 12-3, will be adhered to. That is, the burner, the subsequent mixing of
burner combustion products and fume stream, and the retention at an elevated
temperature for a specified time, will be discussed as separate phenomena. In
varying degrees, depending upon the actual afterburner design, these functions are
intermixed and it becomes very difficult to say where combustion is complete,
where the mixing function begins and ends, or where the beginning and ending of
the reaction that consumes the fume pollutants takes place.
13.2 Equipment Types and Alternatives
Although this chapter is titled "Design Fundamentals", no attempt
will be made to instruct on how to design a burner. There are a great variety of
commercial burners on the market. The objective here will be to give some under-
standing of how the different burners work, and what their capabilities and limita-
tions are. When selected and installed so as to match their capabilities and
accommodate their limitations, these burners will give good service.
13.2.1 Function
The function of a burner is primarily to provide a source of heat. This
function is initiated by bringing together a fuel source and an oxygen source at
proportionate rates which will support combustion reaction in the form of a flame.
Usually this ratio is proportioned to give a small excess of oxygen over that
required for a chemically complete conversion of the fuel to combustion products.
The chemically correct ratio is called the "stoichiometric" ratio. The mixture
containing more air than the stoichiometric ratio requires is sometimes called
"fuel-lean"; that containing less than the required air is "fuel-rich".
The burner must be designed so that the flame is retained in a fixed
position with respect to the burner, with a predetermined flame pattern. It
should have an ability to retain the flame in position over a range of firing
rates.
A secondary function of the burner is that its flame serves as a source
of free radicals, which are highly reactive but transitory molecular species that
S-1U121
-------
206
can attack the organic components of the fume stream and initiate their being
broken down and oxidized. The "half life" of these free radicals is generally very
short compared to the time required to intimately mix combustion products from the
burner with fume introduced external to the burner; nevertheless, they may play
an important role in decomposing the fume.
13.2.2 Discrete Versus Distributed-Source
In a discrete source burner, as shown in Figure 13-2, a single large
flame plume is formed into which the fume stream must ultimately be blended. A
distributed source burner, on the other hand, Figure 13-3, gives an array of small
flames. Fume and flame are already subdivided, thus substantially reducing the
space required for mixing in the fume stream. If one can assign a "characteristic
diameter" to the flame plume, as shown in the sketch, and describe a length, L,
which is required in order to get a completion of the combustion process and a
mixing-in of the fume stream, then the ratio L/D will be a characteristic number
for describing the process, whatever the scale. This L/D applies of course only
if the flame patterns are similar, and ignores the presence of nearby walls, which
can effect the shape of the flame appreciably. One can further assume that the flow
velocities are about equivalent, since they are governed externally by the pressure
drop available from the blower and the fuel supply. Clearly the distributed source
burner completes the combustion reaction and subsequent mixing-in of fume streams
in a considerably shorter distance and time scale than the single source burner
can do. The distributed source design is particularly applicable where gas fuels
are available and where the fume can serve as the oxygen supply for the burner.
13.2.3 Premix Versus Diffusion Mix Burner
13.2.3.1 Premix
When gaseous fuels are available at elevated pressure, usually above
2 psi, premix burners may be used. Combustion air and fuel are mixed in a Jet
nozzle-mixing device before passing into the combustion space. The porting which
admits the mixture to the combustion chamber are designed to give a fairly high
velocity through a number of small orifices to avoid the possibility of having
the flame "flash back" through the entering pipe to the mixing chamber. Figure
13-U is an example of a premix burner. The arrangement shown would also give a
distributed flame, but fume admitted around the burner must have a low velocity to
avoid blowing the flame out. Porting may be distributed over the length of a
pipe channel, giving a "line" or a "ribbon" burner, or it may be laid out in grid
or wheel-and-spoke patterns. The kitchen gas range is a common example of the
premix burner. The premix burner is usually found on very small afterburners
with heat releases below 200,000 Btu/hour, although the size is not limited. Un-
less completely free from the possibility of fouling, the premix burner should
not use the fume stream as an oxygen supply, since the deposit of any material
would be sure to plug the small orifice openings that serve the burner as a
flame arrester. This burner may also be operated with only part of the air
introduced in the premix section, requiring that additional oxygen be supplied
by the fume stream above the burner. It will give about a ten-to-one turndown
ratio; the Jet induced air supply remains in nearly constant ratio to the fuel
S-14121 Figures 13-1 through 15-5 follow
-------
S-14121
67784
Fume
Fuel
////////////////////////y
Combustion
Air
Figure 13-2. DISCRETE-SOURCE BURNER
-------
Fume
Fuel
Combustion
Air
Figure 13-3. DISTRIBUTED-SOURCE BURNER
S-14J21
67784
-------
VJ I
£ **
Movable
Air Register
Fuel
Nozzle
Mixing Throat
Fume
Stream
Figure 13-4. PREMIX BURNER
-------
O> t/>
VI I
Burned-Out Pockets
of Combustion
Products
Burner Tile
Swirl Vanes
Oxygen-Rich Eddy
Being Engulfed Within
Flame Envelope
Combustion Air
Commonly Introduced
With Swirl Motion
J -"?
Fuel-Rich Eddies
Mixing into
Fuel-Lean Section
Detached Pockets'
of Burning Material
Figure 13-5. SCHEMATIC REPRESENTATION OF A DIFFUSION-FLAME GAS BURNER
-------
207
supply. Using external combustion air in this way, the burner is not bothered
by having foul material in the fume stream, but this is done at the expense of
poorer fuel economy.
13.2.3.2 Diffusion-Mixed
Alternative to the premix burner is the diffusion-mixed type, also
called the nozzle-mixed type. The majority of burners found in afterburner
service are of this type. Raw gas or atomized oil droplets are mixed with the
air supply by a turbulent diffusion process occurring within the plume of flame.
Figure 13-5 is a schematic representation of a diffusion flame gas burner with
this mixing process taking place. Air (or fume) is introduced through a register
that can control the amount passing through, depending upon the pressure available;
often the register is in the form of vanes which will give the combustion air a
swirling motion. Gas Jets are introduced by a nozzle or a distributor ring
placed in the air entry channel, with the Jets disposed for mixing into the air.
However, it is only some distance downstream from the gas Jets where the velocity
has slowed and where enough mixture within the combustible limits has been formed
to support a flame. It is only within restricted limits of an air to fuel ratio
that the burning will persist. The actual burning process is an extremely
complicated one, and there are many chemical and physical processes occurring
simultaneously. In the diffusion flame, the rate controlling process is that of
the mixing between air and fuel, which involves a transport of both mass and
momentum between regions of high velocity and regions of low velocity. Only in a
relatively thin interface is there a combustible mixture, and this governs the
formation of a sheet of flame. The sheet is very much warped and wrinkled as
eddies of fuel are swept out into the air section, or as eddies of oxygen-
containing air move into the fuel. Pockets of flame will be carried into unignited
regions, which will burst into new flames; or they will be projected into zones that
have already been burned out, whereupon the flame is diluted and quenched. Since
the flame front cannot move at very high velocity, it remains always detached from
the gas nozzle. A flame front will sweep through a stagnant mixture of cold gases at
relatively low velocity, of the order of 2 to 30 feet/second. At too high a
velocity the flame will actually blow out and away, and the flame will become
extinguished; special provisions must therefore be made to hold the flame in place,
as discussed below under "flame stability".
Some nozzle-mix burners distribute the gas supply through a series of
jets or from a pipe ring with many openings into the throat section of the air
inlet. In this way the air and fuel mixing are well under way before entering the
combustion space. Such a burner is not clearly a premix or a discrete diffusion-
mixed type.
The size and shape of the flame is largely determined by the means used
for mixing gas and fuel, and for preheating it to ignition temperature. The
resulting flame patterns differ widely from one burner to the next. This will be
a topic of a later section.
S-1U121
-------
208
13.2.U Oil Versus Gas Fuels
The burning of oil further complicates the combustion process. Liquids
lo not burn (unless they are self-reactive materials), and it is only after a
droplet has been formed, an envelope of evaporated combustible vapor has developed
around the droplets, and some of this vapor has mixed with the air to form a
combustible mixture, that a flame can take place.
Figure 13-6 gives a schematic representation of liquid flame phenomena.
Depending upon the type of atomizer used the spray pattern can take a number of
forms. Shown is a mechanically atomized spray, formed by feeding the oil under
pressure through a nozzle. The liquid leaves the nozzle in the form of a thin
liquid sheet, conical in shape, which then proceeds to break up into filaments,
which further subdivide into tiny droplets. As a droplet passes out into the
flame zone, it is heated by both radiation from the flame and by convection from
the hot gases that surround it, and an evelope of vapor forms around it. The vapor
in turn mixes with the surrounding air and becomes ignited. Depending upon the
fuel used, a droplet may completely evaporate or it may reach a point where it
chemically cracks to form burnable vapors and residual carbon char. The charred
remains of droplets take the form of tiny hollow spheres, called "cenospheres".
If the combustion chamber is hot enough and there is oxygen available, the
cenospheres will be consumed; otherwise, they may leave the combustion chamber
and be carried out the stack as unburned material. When burning heavy residual
oils, droplets tend to be large (because of the high viscosity) and the flame
pattern approximates a cloud of tiny projectiles trailing flames. With more
volatile liquids, and with a fine degree of atomization, droplets are easily
vaporized, and the flame becomes essentially the same as that of the diffusion
mixed gas burner.
As with the gas flame, there are many patterns of flame possible with
the oil burner. These depend partly upon the spray pattern of the atomizing
system, and also upon the manner of introducing the combustion air, its axial and
tangential or swirl momentum. Because of the problems of oil atomization (dis-
cussed in the next section) oil burners for afterburner service are discrete
burners, used singly or at most there may be two or three burner units.
Oil and gas may be handled in the same burner, with provisions for
withdrawing the oil burner gun with the atomizing nozzle when it is not in use
(otherwise heating the nozzle tip when no fuel is flowing through it causes the
fuel contained inside it to crack to carbon char and plug the small openings).
Gas fuel is expensive in some parts of the country and available only
on an interruptable basis in other parts. For this reason many afterburners are
designed with provisions for both gas and oil burning.
Oil burners often generate substantially higher levels of nitrogen
oxides than do gas burner units, although the amount depends on the size and
operating conditions of the burner, and the presence of bound nitrogen in the oil.
In areas where nitrogen oxides are an important source of air pollution, the use
of oil burners may be restricted.
S-1U121 Figure 13-6 follows
-------
Filament Breakup
Into Droplets
Filaments
Liquid Sheet
Droplets Distributed by
Turbulent Movement of
Entraining Gas
-»..-^^
^ ^- v'
Pressure Atomizing
Nozzle
> ' Mostly Vaporized
>*„ ' Droplets; Behavior
I' I Essentially Same as
N ^ I With Gaseous
' i") Diffusion-Mixed Flame
Heat Transfer by Flame
Radiation and Convection
Fuel Droplet
(Greatly Enlarged)
Droplet Motion Relative
to Surrounding Gas
Envelope of Vapor
Around Droplet
Trailing Flame
Behavior of Individual Droplet
Figure 13-6. LIQUID FUEL FLAMES
S-14121
67784
-------
209
13-2.5 Oil Atomization
Pressure atomization is the most commonly used method, particularly for
the lighter oils and the smaller sizes. In the usual pressure atomizing nozzle,
the oil enters a small cup through tangential orifices, which cause the oil to
swirl at high velocity in the cup. The outlet forms a dam around the open end of
the cup, and the oil spills out over the dam in the form of a thin, conical sheet,
which subsequently breaks up into filaments and droplets. The Jet orifices are
small, and dirt or solids in the fuel will plug them. The pressure drop through
the jet orifices varies as the square of the oil velocity, and at reduced flow rates
the atomization is poor. The turndown with pressure atomization is commonly
about U to 1. Operating pressures run from about 25 to several hundred lb/in2.
Oil viscosity must be maintained within prescribed limits, usually under 600 SSU,
and oil heaters may be necessary to do this.
Steam and compressed air are alternative drive media to give the
necessary energy for fuel atomization, Steam under pressure is a preferred medium.
Under the right conditions, steam reacts with soot and carbon monoxide formed in
the flame, and helps to give clean combustion without smoke. Depending upon the
viscosity of the oil fuel and on the design of the atomizing nozzle, the amount
of steam used may run from as little as 5$ to as much as 30% of the mass rate of
fuel. Since about 1100 Btu/pound are required to vaporize the steam, the latter
figure represents a large loss of heat energy. Compressed air is also preferred
method for atoraization if a dependable source of compressed air is available;
however, one can rarely justify air compression Just for atomization. Other
burners use a rotating cup; oil is slung as a sheet from the lip of the cup across
the air channel. This arrangement is not commonly used for afterburners because
of difficulty in retracting the rotary cup mechanism when it is desired to convert
the unit to gas fired operation. Other burner units use high pressure steam or
air to generate trains of shock waves through which the oil Jet is passed; the
shock waves break the oil jet into an atomized spray. It is claimed that these
sonic atomizers successfully atomize heavy pitch materials.
In general, oil burners are less desirable than gas burners because
distributed burner arrangements are not practicable and because the oil burner
tips require frequent maintenance and cleaning. Improved serviceability of oil
burner gun equipment can be obtained by using automatic retractors, which will
retract the guns from the hot part of the burner as soon as the oil flow is cut
off; these devices are expensive.
13.2.6 Outside Air Versus Fume Oxygen Supply
There is a considerable economic incentive for using the fume as a supply
of oxygen for the combustion process, if at all possible. Outside air brought in
for combustion must be heated along with the fume stream to the afterburner exit
temperature of lUOO or 1500°F. Figures 11-10 and 11-11 permit comparison of the
gas fuel required with fume air and with outside air for combustion. The saving
in fuel by using the oxygen in the fume stream without additional air is about 30$
for a 1UCO°F afterburner. (This saving in fuel is translated into costs in Chap-
ter 11.) The principal limitation to doing this comes from the possibility of
fouling the burner passageways with deposited material. Where mildly fouling
conditions are anticipated, it may be possible to use the fume in a discrete
S-14121
-------
210
burner since the burner passageways are relatively large. Where fouling con-
ditions are very heavy, it becomes necessary to use outside air rather than the
fume stream, since the fume stream can then be brought into the afterburner
through a single register opening of relatively large size, and made accessible
for cleaning without disturbing the burner.
13-2.7 Flame Stability and Turndown Ratios
Holding a flame under a variety of operating conditions and firing rates
depends upon maintaining a localized zone within or adjacent to the normal flame
envelope which will always lie within the combustion limits for the fuel and
oxygen mixture, and in holding the zone where the fluid velocities are not
sufficiently high to sweep the flame away. The following are examples of ways
that are used to bring about improved stability.
13.2.7.1 Burner Tile (Figure 1.5-7)
A conically shaped ceramic tile piece surrounds the burner opening. The
diverging flow area slows the air velocity. The tile is heated by exposure to the
flame to a glowing red temperature, and it re-radiates and conducts to the incoming
mixture. In addition, its conical or cup shape, combined with a tangential swirl
to the incoming air or fume stream may cause a recirculation zone in the flame
center directly opposite the fuel nozzle which brings hot combustion products and
radicals back into contact with the fresh fuel and air mixture for ignition.
13.2.7.2 Target Rod and Plate (Figures 13-Tb and f)
The flame is made to impinge on a cylindrical rod or a flat plate (made
of ceramic or temperature-resistant alloy). Eddies form on the downstream side of
this target, and these contain pockets of fuel-rich and air-rich mixtures.
Portions of these are swept off into the stream, forming multiple ignition
sources; portions are retained and regenerated on the back of the target. The
retained flame serves as a continuous pilot.
13.2.7.3 Fuel-Rich, Low Velocity Pocket (Figure 13-7c)
The pocket is located downstream from the main fuel addition, and has its
own fuel supply. The low velocity in the pocket allows a flame to be retained,
and to act as a continuous pilot for the main stream. John Zink Company uses a
ledge in the wall of the throat of one of its gas burners in this manner.
13.2.7.1* Graded Air Entry (Figures 13-7d and e)
The gas jet enters by a channel having a diverging cross section; air is
introduced by holes that produce cross-stream jets of air. The diverging channel
slows the mixture; meanwhile, some of the air Jets are too fuel-rich, some too
fuel lean; at some point in the channel they are near optimum. Wake eddies are
formed that retain the flame. The Maxon Combustifume burner is of this type.
13.2.7.5 Deflector Plate on Fuel Gun (Figure 13-8)
A vaned conical plate, attached to the end of the fuel gun serves as a
baffle and swirl-inducing device, holding a low velocity turbulent wake zone im-
mediately opposite the burner tip. Although the mixture in this region is
S-1^121 Figures 13-7 and 13-8 follow
-------
Recirculating
Flame
Pattern
Hot Burner Tile (Heated
by Radiation from Flame)
A. Recirculating Flame
and Hot Refractory Tile
Raw Fuel Jet
Ceramic Rod
Flame Holder
'Local Recirculating Eddies
in Wake of Flame Holder Rod
B. Target Rod
Air
S-14121
67784
SIP
xLJf
Fuel-Rich
Ledge
C. Flame Holder Ledge
(After John Zink)
Gas
E. Graded Air Entry
(Grid Plate Version)
Air or
Fume
X
Fuel-Rich
Zone
D. Graded Air Entry
Maxon Combustifume
Gas
K
Fuel
F. Target Plate
Figure 13-7. EXAMPLES OF FLAME HOLDING ARRANGEMENTS
-------
Combination Baffle
and Swirl Vanes
Gas or Oil
(Atomizer)
Low Velocity Turbulent
Wake Zone
Figure 13-8. FLOW PATTERN, AXIAL JET BURNER
CONICAL OIL FUEL SPRAY NOZZLE
S-14121
67784
-------
211
generally fuel-rich, pockets of burnable mixture recirculate in the zone to hold
the flame in place. Additional air sweeps past the plate undisturbed, and mixes
with the central wake zone further out from the burner. This arrangement is
especially common with packaged pressure-atomized oil burner units.
13.2. 7.6 Preheated Air
A preheated air or fume stream used for combustion will improve the
stability of the flame and increases the turndown range; the higher the temperature,
the greater the effect. Preheat of the fume stream will result from the discharge
from an oven, or from the use of heat recovery. When using heat exchangers a
reduced flowrate through the exchanger will result in an increase in the temperature
of the fume to the afterburner, provided the afterburner exhaust temperature is held
constant; this may partly compensate for a limited turndown ratio rating of a
burner that is based on ambient air ratings.
13.2.7.7 Holding Oil Flames
Flame holders of the types described are mostly applicable to gas burners.
Retention of a flame in an oil burner is more difficult, since the complications
of atomizing and vaporizing the fuel prevent introducing target plates, or gener-
ating small pockets of slow-moving mixtures. Flame retention is partly accomplished
through using the hot burner tile and the recirculating eddy opposite the nozzle
(Figures 13-7a) and aided by adding an air deflector plate near the nozzle tip
(Figure 13-8). Sometimes the air supply is introduced in two sections, annularly
disposed about the fuel nozzle. By inserting turning vanes, the air can be made
to enter with varying degrees of swirl, which can promote mixing and burner
stability (mostly this approach is used to control the shape of the flame as
discussed below).
13.2. 7.8 Excess Air Burners
Stability of the flame also depends on the achievement of near-
stoichiometric mixture ratios of fuel and air (or fume) , and most burners are
designed to take from 5 to 1*0 percent excess air through the burner. The after-
burner, as mentioned earlier, operates as an excess air device, in that sufficient
fuel is added to heat the fume stream to about lUOO°F (which is well below the
stoichiometric flame temperature of about 3300°F). Thus, the afterburner (without
preheat or heat recovery) runs about 250$ excess air. Most burners will not tolerate
putting the entire air or fume stream through the air register, since it will
extinguish the flame. However, some burners have been specially designed as
"excess air burners" for air-heating, oven-heating, and afterburner use, and the
additional air can be tolerated. Generally, these burners are of the diffusion-
mixed type, arranged so that the excess air is introduced into the outer part
of the flame envelope, or downstream of the flame-proper. There is a danger that
such a burner will quench the flame before the burning is complete, forming
aldehydes, carbon monoxide, and other pollutants, which must then subsequently
be destroyed in the afterburner holding section. Provided satisfactory per-
formance in this respect is demonstrated, this type of burner avoids the separate
step of blending in the bypassed fume, which is accomplished only with difficulty.
(See the following section.)
Another stability phenomenon that may be encountered is the generation
of some form of acoustic resonance, coupled with a cyclic rate of energy release
by the flame. A number of different conditions can give rise to this "feedback"
S- 11H21
-------
212
effect. Often the effect, which manifests itself as a loud musical note, can be
controlled by adjustments to the burner. Other cases prove much more difficult
to control and require changes to the burner or its control system.
Burners that operate at high velocities have an additional stability
problem. An instantaneous interruption in the fuel supply (as with a slug of
water in the oil supply) is more likely to extinguish the flame than in a slow,
low-intensity flame, since the residence time in the flame is shorter, and the
recirculating eddies are removed more quickly and are not available to re-ignite
the flame when the fuel supply is resumed. The flame-out detector will then shut
the burner down, since a continued flow of fuel into a hot combustion chamber
might ignite explosively and destroy the afterburner.
13.2. 8 Turndown Ratios
The forgoing discussion shows that special provisions are needed to
obtain a burner which will operate stably over a wide range of flow rates and with
a broad range of air-to-fuel ratios. The simple diffusion-flame gas or oil burner
usuallv has a turndown ratio (maximum to minimum fuel rate) of about U to 1. With
the flame-holder provisions described above, turndown ratios of 20 to 1 or more are
possible. Where this is important, the requirements should be specified at the
time of purchase of the burner. Some afterburner units receive a nearly constant
fume flow, and the only variation in burner heat release is that required to
maintain operating temperature in the holding section. Other units may be called
upon to accept a variable fume feed rate or fume composition (that is, with
combustible components), and an adequate turndown ratio in the burner is impera-
tive. Sometimes this can be accomplished by using multiple burners, which may be
cut in or out as the load demands.
13.2.9 Susceptibility to Fouling
Fouling conditions can arise from a number of sources:
13.2.9.1 Vapors Carried in the Fume Stream
Vapor components will tend to condense on cold surfaces. Cold ductwork
walls, fuel distributor manifolds, and burner passages in which the fume stream
is accelerated, are likely points. For water-based condensibles, these can also
be points of corrosion. Water and some oily materials can be drained away and
disposed of separately. Other oily materials, most resinous materials, and
most greases, will tend to form polymerizing layers that can plug passageways in
the burner.
13.2.9.2 Entrained Droplets
These tend to collect on narrow passageways, baffles, distributor
plates, turning vanes, grid plates, etc.—wherever a flowing stream changes
direction, or a stagnant gas pocket can form. As with condensed vapor materials,
the liquid may be drained away harmlessly, or there may be a buildup of layers
of solids on the surfaces. Since the mechanism of deposit is by impingement on
S- Ik 121
-------
215
the surface, deposits may accumulate in small openings in the burner, and char to
form a hard coke. They may also constitute a fire hazard, and initiate a fire
in the ductwork leading to the afterburner.
13.2.9.3 Entrained Particulates
Solid particles will drop out of the stream when the velocity is low, or
impinge on surfaces where the flowing stream changes direction. They will collect
in the wake of brackets, flanges, baffles, etc. If combustible they can be a fire
and possibly an explosion hazard. If sticky, they are likely to coat all passage-
ways , and block burner passageways.
13.2.9.U Combinations
Most fouling materials are likely to contain «-TLl, the fouling constituents:
Condensible vapors, resinous (and polimerizable) droplets, and solid particulates.
The afterburner purchaser should have a clear idea of what kind of fouling condi-
tions he may experience when he approaches an afterburner manufacturer, and how
frequently he is willing to shut the unit down for cleanout.
For gas-fired burners, where fouling is not anticipated, distributed
burners using the oxygen available in the fume stream are recommended. For
oil burners in clean service, likewise, the fume stream should be used for
combustion air.
Applications of this class would be:
Most odor destruction units
Paint bake ovens
Resin curing ovens (depends on resin)
When light fouling is anticipated, use of distributed burners and the
fume oxygen may be possible, but should be reviewed carefully. Burner passages
for the fume stream should avoid openings less than 3/^-inch, and should be
accessible for cleaning. Operating schedules should permit time for regular
cleaning. It may be necessary to lag cold gas manifolds to avoid condensing
vapors, and to insulate walls and ducts exposed to cold ambient conditions.
Applications of this class would be:
Resin curing ovens (depends on resin)
Under conditions of heavy fouling discrete burner units using outside
combustion air should be used.
Applications of this class would be:
Rendering plants
Varnish cookers
Asphalt blowing units
S-
-------
21k
Some manufacturers have made special burners intended to operate under
severe fouling conditions. Fume passages are made without abrupt changes in
direction, and with smooth channel walls. Fuel nozzles project into the stream
with a minimum of frontal area, and with a streamlined shape. Such a burner
would be used only under special conditions, and would be designed and built
specially for the application.
S-14121
-------
215
13-3 Flame Patterns
The size and shape of the flame, or flame pattern, will vary greatly vith
the type of burner and the way it is operated. Not so immediately apparent, but
fully as important to the afterburner, are the velocities, the axial and rotational
momentum of the gas stream, the intensity of the eddy mixing process within and at
the boundary of the flame, the concentration profiles and temperature profiles in
the cross-section of the flame and around it. Moreover, the flame interacts
strongly with the combustion chamber that surrounds it; the gross flow patterns
affect the mixing-in of furnace gases and fume with the flame and the formation of
recirculating eddies; the hot refractory radiates to the flame and affects its
temperature and energy release rate. Thus, the pattern of the burner is not
unique to the burner, but depends also on the conditions of its operation and
installation.
Predicting the resulting shape of the flame is largely the result of
art and empiricism. A start has been made on representing the process mathe-
matically, but only the simplest of geometries have been handled, and with rather
arbitrary assumptions on many of the intermediate processes involved. Figure 13-9
gives an example of a mathematical solution of the temperature and fuel concentration
profiles in a single-Jet diffusion burner in a cylindrical furnace, as calculated
on a digital computer. This approach may prove very useful in the future, but is
not an easily accessible design tool.
What follows, then, is a pragmatic and approximate guide to the flame
patterns that can be expected, which must be established as a basis for mixing of
flame and fume. This mixing should be completed before specifying the retention time
under "plug flow reactor" conditions, during which destruction of the fume is
assumed to take place. Published data on actual flame dimensions are almost non-
existent; what there are apply only to specific burners. In the absence of either
a general analytical approach or of extensive data, we can only suggest that the
afterburner designer apply to the burner manufacturer for recommendations on flame
dimensions for his burner for the specific application.
13.3.1 The Premix Burner
The flame pattern for the premix gas burner can be treated as a simple
Jet of length which is ten times the diameter of the port opening. This
approximation is crude, since'the turbulence, the composition of the fuel, the
air-to-fuel ratio, the degree of preheating, all effect this Jet; mostly they
shorten it. Since the premix burner is used in multiple port arrangements with
small port openings, the flame will in any case be completed within a short dis-
tance of the burner. It is blue in color, non-radiant.
13.3.2 The Discrete Diffusion Burner
The range of flame patterns obtainable with this class of burner is very
large. For a given rate of heat release, the flame can be a compact, high-intensity
brush, or a large, billowing, "lazy" flame; it can be a pale bluish color, nearly
non-luminous, a brilliant radiant yellow-white, or a smoky orange; it can flare out
in a shallow, saucer-like cone, form a bushy ellipsoid, or be concentrated in a
long, narrow pencil-like flame. The principal variables controlling this variety
of flames are the type of fuel used, the amount of turbulence generated in the
mixing between fuel and air, and the amount of axial and rotational momentum
S-1M21
-------
216
different parts of the fuel and air are given as they leave the burner. The
most certain guide for any given burner is to obtain the prediction and
recommendation of the manufacturer. Not all of the possible flame patterns will
be useful for the afterburner, but some idea of the choices is desirable. The
controlling variables will be discussed here in a qualitative way, with some
approximate guides on the size of flame.
13.3.2.1 Effect of the Type of Fuel Used
Natural gas flames tend to be of low luminosity, especially when there is
good mixing between the gas and the fuel, and the air supplied is only slightly
in excess of that required for complete combustion. An increase in the air supplied
makes the flame more compact, turn a hard, bluish white, and give a noticeable
roaring noise. A decrease in the air supplied makes the flame longer, paler, and
"lazier". A luminous flame can be gained by separating the air supply into two
parts, with about a third of the air well mixed into the gas, and the remainder
introduced at low velocity in the outer annulus. The incomplete combustion of
the first stage forms soot, which gives the luminosity.
Unsaturated hydrocarbon materials present in the gas increase the amount
of soot formed in the flame and make the flame luminous. Ethylene, propylene,
butenes, butadienes, etc., are typical components involved.
Oil flames are generally luminous (exception: see paragraph below, des-
cribing the effect of using a pre-combustion chamber). Luminosity increases with
the "heavy" oils, which usually have less hydrogen, and tend to form more soot
particles, which become heated to incandescence. Because of the added time
required to evaporate the fuel droplets and to mix the vapor with air, and
because of the relatively slower mass transfer between the air and the soot
particles, the flame is longer than with gas. (Also the gas flame is usually
partly premixed.) If poor atomization results in forming large droplets, the
flame may be very much longer, noticeably streaky, with the formation of particles
of char that may not be fully burned.
When the flame of either a gas or an oil fuel is discharged into a hot
refractory-lined chamber, it burns more quickly (i.e., making a shorter flame),
and with somewhat less luminosity than when discharged into a cooled metal chamber
(which tends to cool the flame and form more soot). When fuel, fume, or combustion
air is supplied hot, the effect is to shorten the flame and usually to decrease its
luminosity.
S-14121 Figure 13-9 follows
-------
Fuel-Air Reaction is Assumed Instantaneous at
Point Where Stoichiometric Ratio is Reached as a
Result of Turbulent-Diffusion Mixing.
Velocity
Profiles
Stoichiometric Concentration
Defines Limit of 'Flame'
From Printout of D.B. Spaulding Program
Figure 13-9. CURVES SHOWING CONTOURS-FUEL
CONCENTRATION AND VELOCITY
S-14121
67784
-------
217
13.3.2.2 The Effect of Turbulence
Turbulence gives rise to the eddies vhich are the means for mixing the
fuel gas or vapor in with the air or fume stream. Some turbulence is generated
within the flame itself, but this is probably secondary to the effects of shear
forces generated by the gross velocity differences across the flame. The source
of this is the kinetic energy available from the gas and air streams. It is
directly reflected from the available pressure drop supplied the burner. Since the
mass rate of the air is some 15 times that of the fuel, unless a high gas pressure
is available, most of this energy comes from the air stream.
In a gas burner much depends on the manner in which the gas is distributed
into the air (or fume) stream in the burner. Some burners provide much better
mixing inside the burner, leaving less to be done in the combustion chamber. This
can be done by arranging multiple ports for releasing the gas into the air, with
parts directing Jets of gas across the air stream. Units with integral blowers
will sometimes introduce the gas ahead of the blower, so that the blower blades
aid in premixing the gas into the air. (The mixture must then pass into the
combustion chamber with a velocity higher than that of flame travel, so that the
flame will not flash back through the burner.) One design, the "Fanmix", intro-
duces gas through channels in the rotating fan blades, obtaining very good distri-
bution with relatively low pressure loss to the air stream. The better this
initial distribution, the more closely the flame pattern becomes that of the
"premix" burner; however, on this scale, the 10:1 L/D ratio between flame length
and port diameter mentioned earlier would be excessive, and a perhaps 3:1 is
more realistic.
In an oil burner this premixing is not possible, and the turbulence of
the air stream and the passage of the fuel spray must be relied upon, taking their
effects in the open combustion chamber.
The turbulence so far discussed, and its effects, are important, but
subservient to the gross movements of the flame interacting with the rest of the
gases in the combustion chamber; these gross movements result in important patterns
of recirculation, and depend on the axial and rotational momentum of the gases
as they enter the chamber.
S-1M21
-------
13.3.2.3 The Effects of Axial and Rotational Momentum
The mass flow rate of the fuel and the air (or fume) entering the
chamber with a finite velocity give the flame a substantial momentum. (Momentum =
mass x velocity; it is a vector quantity, since it must be specified both as to
magnitude and direction). If the stream enters by way of a straight tube, the
momentum is axial, and the flame has the form of a straight Jet. If the stream
enters the chamber through a set of turning vanes that give it a whirling motion
(i.e., like a cyclone), the flame has angular momentum, or swirl. Some burners
provide only axial momentum; others provide swirl in varying degrees. Swirl can
be obtained by delivering air into the annular channel around the fuel delivery
tube through a tangential scroll; it can be obtained with hinged vanes in the
entering air register (sometimes the vanes are adjustable); or the annular passage
may have helical or screw-shaped inserts. Some burners have two concentric
annular passages, one of which will be arranged with no swirl, the other with;
there will then usually be a means of adjusting the fraction of air passing to
each.
Axial Jet Flame
The axial jet flame produces a long slowly expanding, conical flame
brush. It's envelope is bounded by the ability to mix the fuel and air to the
required near-stoichiometric ratio, which is further discussed in Section 1^.3
and Ik-.k of the next chapter. For an unconfined flame, with a straight gas
nozzle and a surrounding annular air tube having no swirl, with an excess air
ratio of 30$, a flame length of about 5-1/2 feet will result when firing at
a heat release rate of 1 MM Btu/hr. The width at the far end of the flame is
about 1/3 that of its length. The flame dimensions can be scaled as the cube
root of the heat release rate. Very approximately, the flame length,
3.
L ~ 5 • 5 x ^/Heat Release Tin Btu/hr x ICT0
The flame is irregular and of ill-defined shape.
The flame length depends on the available oxygen supply — the flame
becomes longer as the excess air ratio is reduced to stoichiometric. However,
the afterburner usually uses an excess of air, and only if an outside-air-burner
is fired with low excess air will this effect be noticed.
When the free expansion of this jet is limited by the walls of the
furnace enclosure, i.e., when the flame tends to impringe on the walls, the length
of the flame is increased as compared with a flame firing into a large open chamber
(see Reference Ik-3).
The use of oil fuel can alter the flame pattern somewhat. With Number 2
fuel oil and fine atomization, the pattern will not differ greatly from that of
the gas fuel burner. Most spray nozzles produce a conical spray pattern, which
imparts a radial velocity to the jet; the result will be to spread the flame
wider. Moreover, the conical spray tends to produce a low pressure zone
immediately in front of it; with low velocity air and high pressure atomization,
there will probably be a zone of recirculation at this point, that will tend to
widen the flame spread even further. (See Figure 13-8.) This recirculatory
pattern improves the stability of the flame, as mentioned earlier.
S-1U121
-------
219
Reradiation by enclosing and confining the straight jet flame in a
cylindrical box of hot refractory will increase the burning rate, so that less
flame volume will be required; the required flame length in a box such that the
flame just touches the walls would be somewhat reduced. Cold walls will decrease
the burning rate and lengthen the flame and can result in a high residual concen-
tration of CO, aldehydes, and unburned fuel in the combustion products, especially
if the flame impinges on the cold wall.
Jet With Swirl
The long flame of the axial jet makes it awkward to use in the
afterburner, and burners with swirl are commonly employed to reduce the flame
length and the size of the chamber into which it fires. The addition of swirl
can make striking changes to the flame shape and burning rates, especially if
the annular entrance opening is divided into concentric zones (i.e., "primary"
and "secondary" air) with separate control on the degree of swirl in each.
Figures 15-8, 15-10a, b, and c show several markedly different patterns obtainable
with swirl. Of these patterns only those of Figures 13-8 and 13-10 a are likely
to be of use to the afterburner. Neither the Coanda nor the pencil flame burner
is likely to contribute to good mixing conditions or the desired plug flow for
the retention zone.
Swirl effects can be characterized by a swirl number, S, which compares
the anguler momentum of the entering stream with the axial mementum (see
Reference 18-5). For low to moderate values of swirl (.1 < S < .5) the flame
develops a small internal recirculating pocket opposite the burner, as in
Figure 15-8. Externally, the swirling flame narrows down beyond this pocket with
a more crisply defined boundary. Its length decreases somewhat, especially with
increasing swirl (S ~-5), and its shape, becomes more like that of a cylinder,
with diameter about 1/4 that of the length. (This flame has high stability,
since the recirculating zone opposite the mouth of the burner returns hot flaming
products containing highly reactive free radicals back to the starting point of the
flame, thus helping to ignite the entering mixture.) As the amount of swirl
increases still further (S >l), the recirculating pocket enlarges, and the visible
flame opens out into more of a cup shape, as in Figure 15-10a. The length will
become less than half that of the axial jet, and the diameter will be roughly
equal to the length. (This flame is relatively less stable since the central
recirculation now returns relatively cool furnace gases rather than hot parts of
the downstream flame containing the reactive free radicals.)
The transition in flame shapes becomes more complicated when the air
(or fume) supply is subdivided into primary and secondary streams, each with its
own swirl-generating register. When primary and secondary air have opposite
rotation, the momentum of the combined stream becomes the algebraic sum of a +
and - rotation. Greater turbulence can be generated, shortening the flame
without forming the open cup pattern. Complex double-vortex ring recirculation
patterns can be formed with different register settings. The pattern of the flame
and its recirculation can also be affected by the shape of the surrounding
ceramic burner tile or "quarl".
S- Ik 121
-------
220
Interaction vith Surrounding Enclosure
The forgoing discussion was premised on having an enclosure that is
considerably larger than the envelope of the flame. An adjacent wall, or a.
closely confining chamber may simply constrict the flame, which impinges on
the wall, with effects noted below. For flames that depend on gross recirculation
effects to determine their shape, the wall effect can be more drastic, either
reducing or completely suppressing the recirculation effect, and thus materially
changing the shape and character of the flame.
In the event of flame impingement on the wall of the enclosure, if
the wall surface is a hot refractory, and if the gross circulatory patterns of
the firebox are not upset, the effect is not large. With a highly radiant flame
the burning intensity may be increased, at the expense of decreased refractory
life (unless the refractory material is specially selected to have the necessary
temperature resistance; because of the turbulent nature of such a flame, the
surface is alternately swept by hot licks of flame and cold eddies of unburned
gas; this can cause a thermal fatigue effect, causing the refractory surface to
spall away). Castable refractory materials can also be locally overheated to the
point of de-bonding the cement, causing spalling.
In the event of flame impingement on a cold surface, as in a metal
enclosure that is cooled by sweeping the incoming fume stream over the back side of
the wall, the hot spots can cause severe warping of the metal, and possibly cracking
it from thermal fatigue. Additionally, the cooling effect of the wall quenches the
flame before the combustion reaction is complete, leading to the formation of
carbon monoxide, aldehydes, etc., which must then subsequently be destroyed in the
holding section of the afterburner if the unit is to satisfy its anti-pollution
requirements. The cold surface also acts as a trap, or sink, for the deactivation
of the free radicals, which serve an important function in the formation of the
flame, and may also play an important role in the destruction of the fume.
A large afterburner equipped for oil firing may require more than one
burner; as the number increases, the array becomes a distributed-source arrangement.
An array of two, three, or four, still behave as adjoining discrete burners. Placed
side-by-side, their plumes will interfere—especially if the flame pattern is one
that involves a considerable degree of recirculation. With flames having no
swirl, and with little or no recirculation, the individual plumes are merely
squeezed and narrowed. Swirling flames, when all flames have swirl in the same
direction, tend to Join to form a single large vorticular pattern; hence (assuming
this is undesirable) , the adjacent burners should be arranged for swirl in alternate
rotational directions. The effect of gross rotating flow patterns in the combustion
chamber and retention space can have important consequences in the mixing-in of
fume material into the flame, and in the degree to which one can obtain a plug-flow
condition in the retention space; this is discussed in later sections.
5-1^121 Figure 13-10 follows
-------
Fuel
Air
SWIRL
A. Air Entry with Swirl
(Conical Flame)
Fuel
Axial Air
Flame Hugs
the Wall
Surface
777//T
B. Air with Swirl on Outer Boundary
(Coanda Burner)
Axial
Air
\\\\\
Fuel-
Swirl Air — \\\\\
C. Air with Axial Jet, Swirl in Center
Figure 13-10. EFFECTS OF SWIRL
S-14121
67784
-------
221
Burners are sometimes arranged to enter the combustion chamber from the
side; this is done in vertical "stack" units; the fume passes straight through,
while the flame passes across it. This presents problems especially from the
point of view of the mixing the flame and the fume streams. The flame possesses
considerable momentum in the direction of its axis, and the wall constraint
effect is substantial, as shown in Figure 13-11.
13.3.3 The Distributed Burner
The distributed burner is simply a multiple array of discrete burners.
The flame pattern in each burner of the array will reflect the pattern of its
discrete element, except that it will be altered by being placed close to its
neighbors. Distributed burners are designed only for gas fuels. Some individual
versions were described earlier in the section on flame stability; the Maxon
"Combustifume", the Hirt "Multi-Jet" burner, the North American "Flamegrid",
and the Hydro-Co"ibustion Corporation gas burners are sketched in Figure 13-12.
For all practical purposes the flame can be considered as extending no more
than a few inches from the open end of the burner structure. At low firing
rates, the flames are largely confined within the burner structure unless the
fume flow is reduced to the point where slow mixing occurs. Then a long "lazy"
flame may result.
13.3.^ Discrete Burner With Precombustor
With increasing limitations on burning gas fuels, oil fuels may become
essential. On very large units there may be occasion to consider the use of
residual-oil fuels. Residual oils are much harder to burn without generating
smoke and forming high NO and SO2. There is limited data on the NO formed in
afterburners. Satisfactory results in avoiding the smoke have been obtained
using oil burners in conjunction with a small refractory-lined "precombustion"
chamber. The burner head delivers the air and atomized fuel into this chamber
with a fairly high velocity (as much as 25 inches of water draft may be required)
and with a carefully pre-set degree of swirl. Combustion occurs largely within
the chamber, under conditions of high turbulence (in some designs generated by
counter-rotating vortices), and with high temperature re-radiation from the
very hot refractory walls. The flame does project from the chamber opening,
but it extends only about one or two opening diameters. The energy release
rate within the chamber is very high, of the order of 106 to 108 Btu/(hr-ft3),
and the high temperature of the refractory requires special materials. Burners
of this general type are supplied by Thermal Research, and John Zink.
13.3.5 Scale Factors and Energy Release Rates
Burner manufacturers specify the fuel rates that can be handled in their
units. Burners are usually supplied in standard "frame size" units, and one
can adapt these to a range of conditions by changing the size of the gas flow
orifices, nozzle tips, and air register settings, etc. In the case of the dis-
tributed burner, ratings are given for a given burner box element, which can be
incorporated in multiple arrays as needed for increased capacity. To specify
the capacity, the burner manufacturer must know the fuel to be used, the air or
S-14121
-------
222
fume stream (if used in the burner) temperature, the nature of the fuel components
and their heating value that may be carried in the fume stream. He must also know
the pressure drop or draft available on both air/fume and fuel streams, or he
may specify requirements for his burner, which must be provided in the installa-
tion. The variety of burner designs and their pressure requirements is so large
that no general rules can be given here.
Typical heat release rates for discrete burners range from a few
hundred thousand to several million Btu/hr, both in oil and gas fired units. For
the distributed burner, Maxon lists 500,000 Btu/hr per lineal foot of the burner
assembly; other makes are equivalent. Although the burner manufacturers often
work on the basis of a heat release rate per unit volume of the combustion chamber,
such numbers are not relevant here, since the chamber volume is governed by the
space required for mixing of flame and fume components, and for holding them at
elevated temperature for the length of time that is necessary for the destruction
of the pollutants. For scaling purposes, it is sufficient to recognize that the
capacities of the burner and the afterburner unit will vary with the volume, or
the cube of the linear dimension, for units which have dimensional similarity.
13. U Installation
13.U.I Setting
A great variety of arrangements are used by the different manufacturers
for mounting and supporting their burners and for introducing the flame into the
combustion chamber. The manufacturer's installation drawings should be referred
to.
13. U.2 Piping and Controls
Refer to Chapter 9 for a review of requirements that must be met for
control and safety. Fuel and combustion air, control valves, shutoff values,
pilot, and flameout detector are included in some burner assemblies as package
units. Alternatively, the afterburner manufacturer may assemble his own version
of these components. Whatever the arrangement, the purchaser carries the
responsibility of installing a safe operating unit.
13. U.^ Initial Operation
Afterburner manufacturers regularly include supervision by an experienced
operator for the initial startup for a stated time interval as part of their con-
tract. Additional services are available for a fee. On larger and more complex
units, the purchaser should also obtain a manual of standard operating procedure.
Maintenance instructions, and service parts lists are also necessary items.
The initial operation of the burner on some units will be a simple
operation of pushing a button, and making trim adjustments on the air/fume damper
settings. Other units may be more complicated, involving a schedule of slow
initial heating to cure the cement used in refractory linings. All emergency
alarms and shutdowns should be proven by test.
S-1M21 Figures 13-11 and X5-12 follow
-------
Recirculating
(Non-Flame)
Flame
Impingement
A. Radial
Recirculating
(non-Flame)
Flame
Impingement
B. Tangential
Figure 13-11. FLAME PATTERNS - SIDE ENTRY
S-14121
67784
-------
is r^ -/
Profile Plate
^ "* -N
V.M ?
A. Maxon "Combustifume"
B. Hirt "Multijef
Gordon & Piatt "Turbofire"
Flame Holder Rod
C. North American "Flame Grid'
D. Hydro Combustion
Figure 13-12. DISTRIBUTED FLAME GAS BURNERS
S-14121
67784
-------
223
Once initial operation has "been completed, verification tests on emission
control by the afterburner are desirable, and may be required by the pollution
control agencies. Some adjustments of the burner controls may be necessary to
achieve the required emission control at this time. It is very desirable to have
this breaking-in and adjustment completed before the control officer is asked to
approve the installation.
Chapter 20 of this manual, gives test procedures for determining the
emission control performance of the afterburner.
S-14121
-------
225
Chapter lU. mXIHG OF FLAME AND FUME
lU. 1 Introduction
In the preceding chapter it was noted that the afterburner process, for
convenience of analysis, was broken into flame, mixing, and retention processes.
This section discusses the second, or mixing process. It bears to be repeated that
in the actual afterburner these three processes are not separate and distinct.
Particularly, as regards the mixing operation, the formation of a flame and the
mixing-in of the fume may be closely tied together, especially when the fume stream
is used as the source of oxygen for the combustion. Also to be repeated is the
role the flame may play in initiating the chemical breakdown of the fume constitu-
ents because of the presence of the free radicals in the flame; this role is mini-
mized in the present discussion (see Section 12.2.3). It is assumed that uniform
mixing of combustion products and of the remaining bypassed fume stream is essential
for proper performance of the afterburner. Anything less than this results in a
possible failure of the unit to control the pollutant emissions, and a wastage of
fuel since the temperature setting of the afterburner stack must be set excessively
high.
Also repeated: The entire fume stream, even if used as the oxygen source,
cannot be passed through the flame, since the mixture ratio of fuel and fume lie
outside the flammable limits for temperatures in the range used in the afterburner
(see Section 12.1.3). Hence, at least a portion of the fume stream must be passed
around the flame and blended in subsequently, (in some "excess air" burner designs,
the excess air or fume is bypassed through a channel within the burner unit,
however.)
1*4.2 Characterizing Mixedness
An important and difficult question faces the designer who is attempting
to insure good mixing in his afterburner. What constitutes good mixing? In
terms of the concentration of the flame products to be mixed into the fume, does
he want to achieve a uniform blend with a maximum deviation of 10%, 5%, l£, or 0.01??
The answer must depend upon the objective of the mixing. If the mixture leaves the
mixing section of the combustion chamber with a non-uniformity in temperature
amounting to 100°F, the chances are that this disparity will remain and carry part
way through the holding section; and if lUOO°F is required for the time available
to destroy components of the fume stream, then the temperature setting on the
controller would have to be set at 1500°F to attain the necessary destruction.
If, on the other hand, the fume stream contains a highly toxic vapor that must be
controlled to the parts-per-billion level, mixing deviations at the 0.01$ level
may be needed.
By way of definition, concentration here refers to the fraction of the
material in one stream that is being blended with a second, and which remains in
a unit volume (or mass) of the mixture at any point in time (or position) after
the mixing has begun. If one arbitrarily assigns a value C = 1 to the first
stream, C = 0 to the second, then a number between 1 and 0 will characterize the
Preceding page blank
-------
226
degree of mixing achieved at any downstream point. The local mixing effectiveness
can be defined as 100 (l-C)/(l-C . ,) in percentage points, for mixing the first
stream into the second. Clearlym£n'is definition will not apply upstream from the
rdxing. Referring to the question posed above, it is necessary to determine if
the mixing effectiveness required is 90%, 95%, 99-99%- Unfortunately, the
mathematical models and experimental correlations cannot be trusted with
anything like the predictions for 99-99%- The following discussion will attempt
to make it clear why this is so.
Mixing between two fluids is actually a process of subdividing and
recombining on a random basis. As the subdivisions become smaller and smaller the
fluids will finally become mixed on a molecular scale. For chemical reaction
between fume stream components and oxygen, complete molecular-scale mixing will
ultimately be required if pollutant components are to be fully destroyed. We
may start this by mechanically subdividing the streams (i.e., the distributed
burner—we can only go a little way with the distributed burner; however, it is
an important way in terms of saving time and space). We then depend on natural
hydraulic blending through the mechanism of eddies, and finally depend on molecular
diffusion.
Eddies are relied upon for the predominant part of the mixing process.
Eddies are generated between parallel moving streams when the streams have differ-
ent velocities. The velocity gradient between the streams generates shear forces,
which in turn cause a circulating motion that becomes an eddy. The eddy tends to
carry material with it across the line of flow. Naturally, high velocity fluid
from one stream has momentum which thereby becomes transferred over into the
adjacent low velocity stream. Similarly, low velocity fluid acquires momentum as
it is carried over into the high velocity stream. Both material and momentum are
thereby exchanged between the two streams, with what are initially rather gross
eddying motions. Continued motion of these eddies result in their breakup or
decay into a series of smaller and smaller eddies, with finer and finer distribution
of material. The process as a whole is a form of turbulence, and the scale of the
turbulence is measured by the size of the eddies that predominate at a point. If
one can pick out a streamline vhich describes the smoothed, average flow path
of a portion of the stream, unaffected by the random motions imposed by the eddies,
one can imagine a ball riding and floating in the fluid, and following only the
streamline. The ball would then see a localized movement of fluid past it, whose
velocity and direction would depend on the eddying motion in its neighborhood.
The time averaged value of the square of these unsteady velocities would be a
measure of the turbulence intensity (it will be seen that this intensity is a
measure of the kinetic energy associated with the eddying motion). For a more
detailed review of this complex subject, see Reference lU-l).
The mobility of an individual molecule with respect to its neighbors,
under the impetus of its own velocity, is relatively low; it keeps colliding with
the many other molecules nearby. That is, the rate of molecular diffusion is low,
especially compared with the gross eddying motion we have described above.
Diffusion becomes important primarily after the eddies have broken up to a very fine
scale, and the local turbulence intensity has diminished greatly. The chemical
reaction that is desired from the afterburner can be completed only after the
final diffusion mixing process is largely completed for it is only by diffusion that
the oxygen and pollutant molecules can find each other.
S-14121
-------
227
The analytical models and experimental correlations describing the mixing
process quantitatively give only the time-averaged value of the velocity, concen-
tration, and temperature in rather specialized cases. Experimental measurements
made on tests of afterburners are usually obtained with thermometers, pressure
probes, and sampling devices that cannot follow the rapid fluctuations of the
eddying motion. Highly sophisticated laboratory instruments have been used to
follow the turbulent velocity and temperature fluctuations for frequencies up to
several thousand cycles a second, but these have not been used to study afterburner
mixing. No instrument can tell the effectiveness of mixing on the molecular
scale, which is what ultimately counts. Thus, small eddy-pockets can be formed
that will persist in composition and temperature for some distance beyond where
the correlation has predicted uniformity. And mixing on a molecular scale (and
uniformity of temperature on a micro-scale), may not be achieved for some real
distance beyond this. Also, real combustion chamber shapes deviate from the
idealized model, and corners and bends can introduce gross changes in the flow
pattern; on a smaller scale, wall roughness and boundary-layer effects are not
accounted for.
In a paragraph above we gave a criterion for measuring the degree of
local mixing (C a number between 0 and l). This does not apply for the mixing as
a whole. There is no simple, rigorous, easily calculable criterion. Using the
deviations of concentration or temperature from the long-term mixed values , or
from experimental values found in a traverse, can one find an average (if so,
how weighted?), or a mean deviation? Refer to Figure lU-1 for three different
temperature profiles that might be obtained across an afterburner chamber.
Profile (a) might be found in a fairly well mixed afterburner, showing a temperature
peak opposite a burner. Profile (b) shows a nearly uniform, but slightly skewed
pattern. Profile (c) shows cool, unmixed zones along the walls. Using a weighted
average temperature (weighted according to cross section area — that is using
_ E T ( area element)
avg ~ whole area
where the cross section has been divided up into equal area elements ) , one can
find a maximum deviation from the average, d^ . We suggest using a ratio,
for determining the mixing effectiveness of the chamber as a whole (or a similar
ratio based on concentration). By this criterion, curve (b) would be the best
choice for the afterburner; however, common sense would dictate that (a) would be
the choice (the small peak will do no harm). Clearly profile (c) is the least
satisfactory.
To initiate a high rate of mixing, one should generate the large scale
random eddy patterns in the stream which generate the turbulence. This is brought
about by large velocity differences in the fluids to be mixed. The high velocity
means high kinetic energy — which makes another point. Mixing costs energy, which
S-Ik 121
-------
228
is supplied by the fan or the pressure of the flowing gas. More rapid and more
effective mixing will require more energy. A casual examination of the equations of
Figure lU-2 suggest that the mixing effects are only a function of the distance from
the nozzle; further thought recognizes that for a given mass flow rate the higher
velocity uses a smaller Jet diameter, thereby reducing the size of the entire
operation. Since the jet velocity, Vo, varies inversely with the nozzle area,
i.e., as 1/D2, the distance, L, require to achieve a given degree of mixing will
vary as I/ Vo> and inversely with the rate at which energy is supplied.
As a general observation, it can be said that the process of mixing flame
and fume streams to the point of having a uniform temperature for decomposing the
pollutants is the most difficult part in the design of the afterburner. It is the
least well understood and the least effectively handled in many working units. Its
accomplishment is rarely demonstrated by adequate tests in afterburner units; its
inadequacies show up as residual effluents in the stack and wastefully high
temperature settings.
lU.3 Axial Jet Mixing
In order to understand mixing behavior of various afterburner designs it
is helpful to work with idealized models. Although these models may be unrealistic
in describing in detail the complex behavior of the many burner configurations,
they give an insight as to what to expect, and the effects of variations in
arrangement can be examined for the ways in which they alter the general pattern
of behavior exhibited by the idealized models.
The first model is shown in Figure lU-2. It consists of a single "free"
axial jet stream without swirl discharging from a cylindrical pipe into an open space
filled with fluid at rest. The velocity distribution at various distances away
from the nozzle are shown on the sketch (see Reference lU-2). Immediately
downstream from the nozzle exit there is a "potential core" of moving fluid in
which the velocity remains undiminished. The core is eroded by the eddy mixing
process, but does not fully disappear fcr a distance of about 6 nozzle diameters
downstream. Thereafter, the centerline velocity decays logarithmically. Meanwhile
the eddies spread the Jetted material laterally, and the zone of mixed-in material
increases in width. The concentration of mixed-in material is greatest along
the centerline of the Jet, diminishing to the vanishing point at the edge of the
Jet stream.
The model describes with good accuracy the distribution of velocity,
temperature, and concentration (temperature and concentration spread a little
faster than the momentum, as measured by the velocity). The distribution shown
represents a time averaged one and the corresponding time-averaged streamline is
a straight line. Because of the eddy motion of the mixing process, the actual
trajectory of an individual particle only approximately follows the streamline;
it wanders and weaves in an irregular path as shown in the figure.
This model is helpful because it shows what happens when a discrete
burner discharges its flame into an open combustion chamber without swirl. The
jet here represents the flame plume, which does not necessarily have the same
diameter as the opening of the burner (the effect is close, however, if the Jet
S-14121 Figures 14-1 and lh-2 follow
-------
avg
avg
Jmax
'avg
max
Duct Width
Figure 14-1.
S-14121
67784
-------
O» in
VI I
ro
Time-Average
Velocity Distribution
Fluid at Rest
Trajectory of Particle
Locus Where Velocity
(Concentration, Temperature)
Has Fallen to % its
Centerline Value
Potential Core
Beyond the Potential Core Range, at Section A:
rm = L/8
Lateral Profile Cross Sect/on
Lateral Spreading Rate
4D
Decay of Centerline Velocity
Figure 14-2. MIXING OF A FREE JET IN AN OPEN CHAMBER
-------
229
velocity and size is corrected for the high temperature of the gas (Reference lU-3) .
Clearly apparent is the very large distance required to dissipate the Jet velocity
and mix the flame into the contents of the firebox. Roughly 20 times the flame
diameter is required of this mixing to reduce the core concentration to 20% of its
initial value.
lU.U Concentric Axial Jets
The second idealized model is that of two concentric Jets, as given in
Figure lU-3. There is a central core Jet, entering with a higher velocity, an
inactive gap, and an annular jet surrounding it, entering at a lower velocity.
The Jets have symmetry about the centerline of the core Jet. There is no swirl,
and the Jets discharge into a large, stagnant chamber.
This model shows that a relatively stagnant recirculating zone develops
in the gap between the two Jets (Reference lU-U) . This gives a low pressure area
that diverts the two jet streams, so that beyond b they join to form a single Jet.
Thus, the second model approximates the behavior of the single Jet model when suffi-
ciently far away from the nozzle exit .
The concentric jet model represents the situation of a flame (i.e., the
inner jet) mixing with a secondary air or fume stream (the outer, annular Jet).
The velocity profiles of Figure lU- 3 are insufficient to describe the mixing
between the two jet streams; Figure lU-1* shows that the velocity profile at section
a-a is actually made up of two profiles, a* -a* and a"-a", from which an idea can
be gained of the degree of interpenetration that has occurred. A temperature
profile of a hot core stream mixing with a cold fume stream would look like the
profile a'-^a1 . As with the single Jet model, a distance of at least 20 diameters
of the core Jet (the flame) is required to get an effective blending of the two
streams .
To account for the effects of putting the burner in an enclosure , con-
sider a third model (refer to Figure 13-8). Jets are introduced into a combustion
chamber of finite dimensions. No longer is there an infinite sink into which to
dissipate the momentum of the Jets. Material from the surroundings entrained into
the jet stream by the eddy mixing process must be replaced; it comes from developing
a toroidal recirculating pattern in the corner of the chamber, around the jet
openings, so that there is a recycling of material taken off the outer envelope
of the Jet stream. Far down the chamber the concentration and temperature
approach a homogeneous mixture of the two streams .
The presence of the enclosure makes it hard to predict the effects on
mixing. Data is available for the mixing of two concentric streams mixing in a
pipe (Reference l^*-5) • Decay of velocity and concentration of the center stream
(downstream from the nozzle) can be calculated, but the procedure is involved.
lh.5 Subdivided Parallel Jets
A multiple burner installation may be simulated with a model consisting
of an array of confined coaxial Jets of the type discussed in the preceding section.
S-
-------
The simulation can only be qualitative, since no array can give a true radial
symmetry for each element of the array. It also suffers in that with the single
coaxial Jet model, the presence of a wall contributes considerable turbulence
from its wall boundary layer, which is not present when placing the Jets side by
side. The single unconfined jet has a large reservoir of dead air surrounding it
with which to exchange momentum that will slow it down; adjacent moving streams can
provide no such braking action. Hence, parallel arrays of Jets are not as effective
in promoting mixing as one would predict if scaling from a single large Jet to
determine the performance of an individual small Jet in the array. In one case
(Reference lb-6), the use of 6k small nozzles mixing gas into a secondary stream
in a pipe was found to decrease the distance required to obtain a mixed condition
by a factor of 5 as compared with a single nozzle mixer; scaled from the free Jet
model, one would have anticipated an 8 fold reduction. This particular test was
based on very low Jet velocities, however, so that pipe turbulence rather than Jet
turbulence controlled the mixing; the same numbers would not necessarily apply to
the multiple burner case. For an accurate determination of mixing, the designer
should make model tests of his own configuration, and verify the final design
with carefully made traverses on the actual afterburner.
Despite the difficulties in predicting the degree of mixing obtainable
with an array of parallel jets, the principle is one of the most useful for
obtaining good mixing where space is limited, i.e., the multiple burner approach
should be used wherever possible.
A precautionary note: Some afterburner units are designed with multiple
Jet type gas burners that form an array of the general type discussed. If the
entire fume stream is passed through the individual burners or between them, the
principles discussed above apply. If, on the other hand, the fume stream passes
in part around the outside of the burner array assembly, then the mixing must
be considered from the standpoint of a single large burner source.
lU.6 Profile Plates
Profile plates are used to control the bypassing of fume stream around
distributed gas burners when the whole stream cannot be passed through the burner
(see Figure lU-5). Burner exit stream and fume stream entering by way of the slit
between the burner box and the profile plate can be considered as parallel
rectilinear Jets. The problem here is the distribution of the bypassed material
into that issuing from the burner opening. Approximately 20 times the width
of the slit opening should be allowed for nearly complete uniformity. The effective
opening width is less than the actual one because of the sharp-edged-orifice effect;
also the turbulent condition of the flame leaving the box, tends to reduce the
number of L/D's required for this mixing process, but these effects should not be
counted on. Test data obtained by Maxon on their "Combustifume" unit show a
nearly uniform temperature distribution on a section taken U2 inches away from the
profile plate (temperature variation 990°F to 1007°F); at a distance of 18 inches,
a temperature varation of 998°F to llUO°F was measured across the channel.
Assuming a slot width of 1-1/2 inches and an effective Jet width of 0.9", their
results would be consistent with the recommendation Just given.
Figures lk-3 through 14-5 follow
-------
VI I
VJ —.
ro
Stream B
Stream A
Stream B
Potential
Core Boundary
Velocity Distribution
Profiles
Streamline
(Outer. Annular Jet)
Streamline
(Core Jet)
Figure 14-3. MIXING OF CONCENTR1X AXIAL JETS (REF. 14-4)
-------
Velocity Distribution
at Section a-a
(See Fig. 14-3)
The Velocity Distribution is the
Sum of the Two Distributions from
Inner and Outer Jets, A & B;
Also a'-a' Represents Concentration
Profile of A. and a"-a" that of B.
Figure 14-4. VELOCITY COMPONENTS IN CONCENTRIC AXIAL JETS
S-14121
67784
-------
Profile Plate
Fume
Profile Plate
Figure 14-5. DISTRIBUTED BURNER WITH PROFILE PLATES
S-14121
67784
-------
231
lU.7 Cross Stream Jets
Another approach to improve mixing is to introduce cross-stream momentum
into the system, to carry the one stream into the other by a mechanism more powerful
than the eddies generated by velocity gradients between parallel streams. This
can be done by directing the streams at an angle to each other, or it can be induced
by baffles.
Figure lU-6 shows results of an experiment (Reference lU-7) measuring the
rapidity with which a Jet stream entering from a side wall will mix with a main
channel stream. The Jet stream is deflected into a curved path, but penetrates
across the channel to a considerable depth. The degree of turbulence is greater
than with the parallel Jet, and the Jet velocity decays much faster than for the
parallel Jet case. Also, downstream from the side-entering Jet a wake pattern
forms which includes a double line of counter-rotating vortices which aid iii the
cross stream mixing. Where the cross stream Jet differs appreciably in density
from that of the main stream, the dimensionless correlations of Reference lU-8
should be used. See also References 1^-19 and 1^-20.
The model tests indicate that when using a discrete burner unit, the
fume stream would be better blended into the flame through a series of cross-
stream Jets, perhaps arranged from alternate sides as in Figure lU-7 so as to avoid
the impinging effects shown in Figure 13-12). In rectangular array, this might
be applied as in Figure lU-8, particularly as applied to a "line burner".
lU.8 Baffling
Baffling can be used to induce cross-channel mixing; however, its use
must be understood, otherwise it is costly in added pressure drop and in wasted
chamber volume used up in recirculatory "dead zones". Its application should be
made with full regard for the possibilities of impinging the flame on the baffle,
leading to excessive maintenance problems.
Baffling must be designed to divert the stream laterally an amount
equivalent to the lateral scale of unmixedness. That is, if the variation of
concentration or temperature to be evened-out spreads from the center of the
chamber to the wall, then the stream needs to be diverted by at least this amount.
Checkerworks or grids are sometimes used with the object of improving ths mixing;
they are not very effective for this purpose, since they promote mixing on the
scale of the grid of the checker or screen; they will not help much in exchanging
material across the whole channel. On the other hand, grids and checkers are very
effective in the small-scale mixing, and in improving the flow distribution and
velocity distribution in the channel (see References lU-9, .10).
The following baffles are commonly used.
lU.8.1 Bridge Wall Baffle
A dam or bridge wall placed across the chamber is a fairly effective
mixer. For best results, bridge walls should be used in pairs, as in Sketch A
S-
-------
232
of Figure 1^-9- Used singly, the bridge wall tends to acquire a recirculating
zone on its back side, and the main stream may not mix well across the stream lines,
as in Sketch b. With a single bridge wall partition having a blockage of 60 - 75$,
good mixing was achieved in a distance of 10 times the duct height, according to
Reference lU.9. The presence of a high velocity core flow ahead of the bridge wall,
as from a burner Jet, helps somewhat to break up this streamline flow pattern
downstream from the dam, where a pair of counter-rotating vortices develop, as
in Sketch c. The bridge wall should extend at least half way up the chamber and
the second wall (if used), should be placed at least one chamber diameter (or width)
downstream. For very high effectiveness mixing (99-9#?) the bridge walls should
reach almost to the opposite sides of the chamber, giving a high velocity to the
gases, and the distance between the two walls should be increased. The chamber
formed between the two bridges then becomes a "mixing chamber", with highly
turbulent flow generated by the dissipation of the kinetic energy of the flow
through the narrow openings. Such an arrangement can be costly in terms of
pressure drop and space required, but this may be Justified if high effectiveness
of mixing is required.
U.S.2 Ring and Disc Baffles
A series of rings and discs inserted in a pipe, as in Figure lU-10, has
been in common use for blending and mixing in pipeline reactors. One or two sets
of ring-and-disc should give effective mixing and act similar to pairs of bridge
walls. It is suggested that the ring and disc each act as orifices and each
should have a free-flow area about half that of the chamber for good mixing.
Spacing should be no closer than half the minimum width of the chamber. As with
flow through an orifice the pressure drop across such a mixer can be estimated
by finding the velocity of the hot gases through an area which is 0.6 times the
free-flow area (vena contracta); then, for a mixed gas density w, It^/ft3, the
pressure drop for each ring and disc pair should be about wV2/gc. Space required
for mixing is about 10 chamber diameters for a ring and disc pair, which makes it
costly in space.
A modified version of the ring and disc is used by one afterburner
manufacturer, Figure lU—11. Ring and disc are aligned to form an annular orifice,
which is placed above a ring burner, as in Sketch a. The arrangement is very
similar to the idealized free Jet model of Figure lU-2, except that an annular
opening is used; furthermore, the full annular opening width cannot be used for
the Jet width dimension because the opening acts like a sharp-edged orifice.
d should be taken at 0.6 times the annular gap vrofcfr. The manufacturer's
data on velocity, concentration, and temperature profiles (Reference 1U-12) is
given in Sketches b, c, and d of Figure lU-11. The values of L/d on the basis
Just described would be,
Distance L/fl.
5 inches 2.lU
12 inches 5.1
2k inches 10
60 inches 26
d is the effective gap dimension, allowing for vena contracta.
S-1^121 Figures lk-6 through 1^-12 follow
-------
S-14121
67784
Line of
Maximum
Penetration
D
Downstream
Distance,
inches
Pj.Vj
C0 = 100%
p = 0.152
d0 = 0.394 inches
6
For the Lines of "Maximum Penetration of Concentration", OD,
V -DBS / X \O.34 .. /..
•T— = p^"" I-,—1 WHERE P = Vs/Vj
* ' V is Velocity
p is Density
REFERENCE:
M.A. Patrick; "Experimental Inv. of Mixing and Flow....." ,
Journal Inst, Fuel, v. 40, Sept. 1967, pp. 425-432
Figure 14-6. SIDE ENTRY OF A JET INTO A MAIN CHANNEL
-------
Burner Unit
Fume
Fume
Cross Stream
Fume Entry
(Staggered)
(Section of Box - Looking Toward Flame)
Figure 14-7. SUGGESTED ARRANGEMENTS FOR CROSS-STREAM
MIXING OF FLAME AND FUME IN RECTANGULAR BOX
S-14121
67784
-------
Cross-Stream Jets
Profile
Plate
Burner'
Figure 14-8. SUGGESTED APPLICATION OF CROSS STREAM
JET MIXING TO PROFILE PLATES
S-14121
67784
-------
Fume-
Flame •
A.
B.
Figure 14-9. BRIDGE WALL BAFFLES
S-14121
67784
-------
f— VzD .1 • V2D— H
• •• • •''••'*••'* i ''••'•
f ^^y J \ _^^^^ T ^^^^ * * '^^^^^
, i^l/^\ 0 B^/xTTv
N \
j
Figure 14-10. RING AND DISC BAFFLES
S-14121
67784
-------
EXHAUST
STACK
Q
tlillTIItli-^
-BURNER
GAS
INLET
FUME
INLET
A. Burner Configuration
4COO
3000
o
c
z
oc
111
2000
1000
C. Tracer Concentrations Indicating
Mixed Performance
RADIUS
B. Velocity Development Downstream
of Baffle
1600
1200
°F 800
400
60" D.S.
18 12 « 0 6 12 16
RADIUS
D. Temperature Uniformity
Figure 14-11. PERFORMANCE OF AFTERBURNER WITH RING
BURNER AND ANNULAR BAFFLE
(See Fig. 3-1."4
-f
S-14121
67784
COURTESY: Surface Combustion Division - Midland Ross Corp. (Ref. 14-10)
-------
PLANE OF MEASUREMENTS
S-14121
67784
in
Figure 14-12. JET INCINERATOR
COURTESY: Surface Combustion Div. Midland Ross Corp.
-------
233
One observes that the undisturbed potential core flow has disappeared
faster than would be predicted from the Jet model; this is probably attributable to
the higher level of turbulence generated by the flame beneath the opening. The Jet
model is based on the assumption of a smooth nozzle flow generating the Jet. The
baffle demonstrates good mixing as measured by temperature uniformity at the
traverse taken 60 inches downstream from the opening, which w>uld be about 26 L/
with respect to the jet, but is less than two diameters with r»_3pect to the chamber.
lU.9 L-Turns and U-Turns
A right angle or reverse turn in the channel of the combustion chambe1*
has mixing properties. There is considerable data on velocity distribution
downstream from turns, but very little data on their effectiveness as mixing
devices. Data of Reference lU-11 shows that from 6 to 10 diameters (or cht'inel
widths) distance downstream from the turn is necessary to re-establish a uni/orm
velocity distribution; it can be assumed that this distance would also be neeiLfx1
to achieve a uniform condition of mixing. Thus, it would appear to be of limited
value for the afterburner designer. More will be said about the use of L-and
U-Turns in the next chapter on the design of the holding section.
lU.jp Annular Throat: The Eductor Principle
Using the flame jet as the motive power, both mixing and pumping can be
accomplished in an arrangement that makes the burner a jet eductor. Figure lU-12
shows an example of a burner supplied by one manufacturer, who claims that the
arrangement is especially suited to handling dirty or liquid-droplet-containing
streams, since the fume passages are not passed through the burner, and the use
of an exhauster fan can be avoided. The burner uses outside air (which places
an added heating burden on the system) , and the blower used on this air supply
provides the necessary motive power for the eductor. Figure lU-13 gives per-
formance data on this incinerator (from Reference lU-5). References lU-1 and
lU-l"5 give bases for designing the Jet eductor. The Jet must fill the throat, and
the throat should have parallel walls for at least 2 diameters; for really
effective mixing 8 or 10 diameters should be used. Care should be taken to bi sure
the flame is complete before it is used to entrain the fume; too rapid a quench of
the flame will generate CO and aldehydes. Following the throat a diverging.
diffuser section may be used to recover the kinetic energy of the stream leaving
the throat. Since such a diffuser must have an angle of divergence of no more
than 15 degrees to maintain a completely filled duct, the length of the section
is likely to make it uneconomical.
lU. 11 Side Entry of Flame or Fume
Mixing of flame and fume can be promoted by introdv.ing the flame Jet in
the side of the afterburner, with the fume stream entering on the end and flowing
along the axis. The desired pattern in the combustion chamber is a general random
pattern of eddies. However, a purely radial or tangential entry is likely to
develop one of the stable recirculating vortex patterns shown in Figure 13-12;
neither of these patterns is desirable; the latter especially is undesirable, as
it may set up a stable circulating flow that is likely to persist over the length
S-1U121
-------
of the whole afterburner, unless controlled by baffles. Model experiments with air
in rectangular boxes (Reference lU-13) suggest that the burner or burners should be
directed about 30° from the radial as they enter the chamber. (The model tests
used 8 Jets in two planes, with U directed for one rotation, the other U in counter-
rotation; such an arrangement would be impractical for any but very large after-
burners.) The model tests also showed that the axial entry of the fume stream
should be at low velocity to allow good flame Jet penetration.
Burners arranged for side entry may encounter flame impingement p oblems
that could result in shortened refractory life and prematurely quenched flames.
Flames designed to "fill the chamber cross section with a bath of flame" also
run the risk of quenching the flame with the fume, thus leading to the formation
of CO and aldehydes. It would be better on this score to complete the flame before
it spreads across the chamber, thus using the flame products for mixing with the
fume.
There is the alternate possibility of introducing the flame in the end
of the chamber, and bring the fume in from the side. If this is done, the fume
should be brought in radially and with a low velocity. If brought in tangentially
at high velocity, the cool (and therefore denser) fume gases remain in a stable
cyclonic flow adjacent the wall with little tendency to mix with the flame.
Using the low velocity radial-entry fume arrangement the flame length can be kept
short by using a burner having swirl in its air supply (see Chapter 13).
lU.32 Mixing with Swirling Flames
Air model studies with the central Jet (or burner) stream having swirl
and the secondary air (or fume) stream having no swirl, mixing in open chambers,
(Reference lU-3.6), in cylindrical (Reference lU-17) and in rectangular (Reference
lU-l9) chambers, show that the swirl significantly reduces the distance required
for mixing as compared with a straight Jet. In effect, the flame is made shorter
and larger in diameter. Moreover the central potential core can be diminished
or even eliminated. If the spreading Jet impinges the wall to any appreciable
degree, a central recirculating zone develops, as in Figure 13-7&. Results of
the tests on a 1-inch diameter Jet for no swirl and with a U5° swirl vane angle
are shown on Figure lU-lU. The effective length of the jet with the swirl has
been cut in half.
lU.13 Use of Cold Air and Water Models to Study Mixing Problems
Where there is doubt that mixing conditions are adequate in a proposed
afterburner design, the use of small cold air or water models is strongly
recommended. Much information can be gained from relatively inexpensive equip-
ment. Such a program is far less likely to be expensive than a campaign of cut-and-
try on a full scale afterburner. The subject is covered more fully in Chapters 10
and 15.
Figures 14-13 and lU-1^ follow
-------
10
8
6
4
2
8
16
24
32
Maximum Entrainment Ratio vs
Dimensionless Jet Length
0.8
0.6
0.4
0.2
0
;* 'p-—.
1
—
-
—
i
1 i 1 i i
6420246
Radius
Measured Mixing Efficiency
10
8
6
4
2
8
AP Theor.
12
16
Maximum Observed Pressure vs
Theoretical Pressure
1.0
0.8
0.6
0.4
0.2
0 0.4 0.8
AP/AP Max.
Jet Incinerator Characteristic
Figure 14-13. PERFORMANCE OF AN EDUCTOR TYPE MIXER
USED AS AN INCINERATOR
S-14121
67784
COURTESY Surface Combustion Division - Midland Ross Corp
-------
O» to
vj l
g 5
NJ
o>
Q
w
*-
0)
U
C
o
u
Straight Jet - No Swirl
0.86
Length: 26"
I
I
I
8 12 16
Inches from Jet Entry
20
Decay of Concentration Along Jet Axis
Central Jet Diameter = 1
8 16 24
Axial Distance, inches
0.57 0.07 0.014
0.14 0.029 0.0
/////////I
/ f / / r //r r /]
Secondary Stream - No Swirl
Primary Stream - Swirl Vane Angle = 45*
Distance to Reference Concentration on Axis
with Varying Degrees of Swirl
20 r
0)
•5 10
8 8
c
(0
« 6
Q
_ 10% of Initial Concentration
I i i i i
i i i i i i
10-'
1
Tan ( Swirl ,'ane Angle)
10
Figure 14-14. CONCENTRATION PROFILES - JET MIXING WITH SWIRL
-------
235
Chapter 1$. RESIDENCE CHAMBER
15.1 Introduction
Previously noted has been the division of internal processes of the
afterburner into three parts: The burner, the mixing section, and the residence
chamber. This is for convenience in analysis; in reality these processes overlap
each other, and the burning of fuel and oxygen, the blending in of the dilute fume
stream (only part of which can go through the flame without extinguishing it),
and the holding at elevated temperature for a finite time to decompose the fuae
components cannot clearly be distinguished (Figure 12-3 ).
In this chapter it will be shown that a plug-flow condition is the
ideal for promoting complete destruction of the noxious fume elements. Possibilities
for achieving this ideal in practical designs will be reviewed. The effects of
using common geometric shapes and chamber shapes will be noted, including the
folded chamber, round vs. rectangular chambers, and the effects of baffles.
Comments will be made on some exhaust stack arrangements, insofar as they affect
the afterburner operation (i.e., not from the standpoint of dispersing effluents).
Finally, there will be a discussion and recommendations on the use of cold air and
water models as means of testing out and exploring different burner, mixing, and
residence chamber alternatives prior to the construction of full scale units.
1$.2 The Plug Flow Reactor
At some distance along the length of the afterburner the designer will
conclude that mixing is essentially complete, that henceforward the space will be
assigned to holding fume components and oxygen at elevated temperature long enough
(-.1 - .3 seconds at lUOO° - 1500°F for most pollutants) to ensure that the reaction
will be completed within the required limits. The last chapter took considerable
pains to show that mixing of flame and fume is in reality very difficult, and that
a truly uniform distribution of materials and that even profiles of temperature
and velocity are unlikely. Let us assume that there is a known distribution of
temperature and flow velocity at the entrance to the residence chamber, as doj._-r-
mined by experimental measurements with thermocouples and pitot tube, or as
estimated with the help of Chapters 13 and Ik. Let us explore the consequences
of this distribution.
Figure 15-1 represents a hypothetical traverse of velocity and
temperature across an afterburner at the start of a long, straight residence
chamber, of length A to B. Having entered this section in a nearly-well-mixed
state, there is very little velocity gradient in the central p-.it of the chamber
to generate the eddies that would promote further mixing. There may be some
residual turbulence from the previous eddy motion; otherwise, any eddy mixing or
further turbulence must come from wall drag, and the resulting velocity gradients
next to the wall. That is, the flow will approximate turbulent flow in a pipe.
In general, mixing under conditions of turbulent flow in a straight pipe are not
very good; commonly UO diameters is allowed to even out any initial mal-distribution.
(This depends somewhat on Reynolds Number, of course.) Since the usual afterburner
S-1M21
-------
236
residence chamber has a length which is typically from 3 to 5 times the diameter
or width, very little mixing can "be expected while traversing the length of the
chamber. As seen in Figure 15-1, gas entering section A at temperature T and
Velocity V will pass along a stream tube and leave at B with very little change.
The situation shown in the sketch is a favorable one - the higher velocity element
of gas also has the higher temperature, which promotes a faster conversion to off-
set the shorter residence time. A much less favorable situation would have P
high velocity portion of the stream passing through at low temperature; this will
happen if a slipstream of fume should bypass around the burner without subse-
quently mixing with material passed through the flame. Clearly, the most favor-
able condition exists when both velocity and temperature profiles are flat.
This condition gives what is called a "plug flow" reactor; that is, the gas
moves along as though it were a solid plug.
15.3 Plug Flow Reactor Residence Time
Many pollution control agencies require that an afterburner have a
design that will hold the fume stream at a specified temperature for a specified
"residence time," with temperature and time values fixed according to the com-
position of the fume stream or the type of service. Strictly speaking, this time
should apply to the plug flow reactor section. Conservatively, the time and space
used by the burner and the mixing of burner combustion products with the fume
should precede and be independent of this residence time. In estimating the
residence time in the plug-flow reactor section, the average velocity of the
mixed gases should be computed, based on the chamber pressure and temperature
(see equations given in Chapter 13 for calculating the gas density), and finding:
Length of Plug Flow Reactor
Average Velocity
For the computer-minded, an approximate estimate of the conversion of
the pollutant components should be possible, using the plug flow reactor concept
as applied to each flow element across the channel cross section. Such a cal-
culation can be only approximate, since
a) Despite the arguments given above, there is indeed some cross-stream
mixing, hence all the gas coming out the opposite end of the stream
tube will not have the same temperature or will have experienced the
same residence time.
b) Reaction must be describable as first-order for the pollutant, which
also requires that there be an excess of oxygen at all times.
By estimation or by physical measurement one can establish a profile of initial
concentration C of each of the pollutant elements, such as CO, methane, hexane,
etc., over the section A-A. The area may be subdivided into equal elements, dY.
Chapter 12 has given reaction-rate data for some pollutant components, and many
other materials can often be treated as though they were one or another of
components.
5-1*4-121 Figure 15-1 follows
-------
\
Temperature
Velocity
B
Figure 15-1. THE PLUG FLOW REACTOR
-------
237
The time-of-flight of the gas entering the chamber at A-A and leaving at B-B
can be found for each area element from the velocity distribution,
*•-*-
For each element, the residual concentration at B-B can be founa,
C = C0 e~kt
where
k = A
E is the activation energy, R the gas constant, T the absolute temperature.
Values of A and E are given in Chapter 12. The total conversion, and the
remaining residuals may be found by summing over the area elements.
1$.U Chamber Shapes
The simplest shape, and functionally the best for most closely achieving
the plug flow reactor condition, is the straight cylinder. A rectangular cross
section with a straight axis would be a second, "but generally acceptable choice -
there is some tendency for the corners to have a lower flow velocity. The
rectangular shape is easier to construct, especially for large sizes of afterburners
Some afterburner manufacturers have endeavoured to save space needed for
the afterburner by bending the residence chamber back on itself, or by using an
ell-shaped space. This is not good practice. On the inner face of the wall,
downstream of the bend, and in the corners, flow separations occur. The stream
does not follow the wall, and large, recirculating eddies are formed. These eddies
remain stably in place, become stagnant pockets of burned-out gas, and do not
contribute to the pollutant oxidation reaction. Figure 15-2 shows the volume
occupied by these recirculating eddies. Their volume should be subtracted from the
effective volume of the residence chamber and may reduce total residence tirv by as
much as 30%- The bends are also unsatisfactory in that they tend to give F highly
non-uniform velocity distribution, even outside the recirculating pockets, so that
some of the material is bypassed with less than its proper residence time in the
chamber, some is held for more. The arrangement is far from that of a plug-flow
reactor (Reference 15-1). Figure 15-3 shows some estimates of the effect of an
L-bend on the residence time and velocity distribution of material traveling along
a streamline with a velocity differing from the average.
Other manufacturers have used the bends, or baffler as a means of
improving the mixing or contacting in the residence chamber. Such mixing or
contacting should be carried out before the residence section.
Chapter lU mentioned the checker wall or screen in connection with
devices for improving mixing. Its use at the entrance to the residence section
can be beneficial, especially if it follows a U-bend or L-bend, or if there is
a poor velocity distribution at the entrance. The checker generates a pressure
drop that redistributes the flow and gives a more even velocity distribution
S-14121
-------
238-
downstream. It does generate additional turbulence, but this is a relatively
small scale turbulence which dies away a short distance downstream (Reference 15-2).
15o Comments on Stack Arrangements
•
The residence section should terminate in a channel of reduced cross
section, especially if the chamber axis is vertical. The flow velocity in the
residence section is usually low, and a large vertical stack under low flow
conditions may develop recirculating flow patterns, aided by the bouyancy ol
the hot gases and the cooling on the walls. Under no circumstances should the
interior of the chamber "see" the open sky, since the sky is a very effective
"sink" for radiative cooling of the chamber walls. Walls of the chamber should
be insulated, as otherwise material passing along the cooled surfaces will escape
without being oxidized.
15.6 The Use of Cold Air and Water Models
As mentioned in Chapter lU, and discussed in Chapter 10, where there is
doubt about the effectiveness of the mixing between flame and fume streams, and
also where there is uncertainty over the success in providing conditions at the
entrance to the residence section which will approximate the desired plug-flow
conditions, cold air or water models should be considered. Such a model can be
set up in the laboratory for a small fraction of the cost of constructing a complete
afterburner. It can be built of sheet metal or wood, although "Plexiglass" is
much to be preferred, since it allows visual probing and observation with smoke
tracers and dyes. Mixing effects can be studied while blending warm air or water
into room-temperature air or water. Air is to be preferred here, as the density is
much closer to that of the hot gas in the afterburner, and the Reynolds Number will
likewise be more representative. However, water models can be especially useful
where blending or mixing with colored dyes is desired. Sensitive laboratory-type
thermocouples and pitot tubes must be used with the air models to obtain satisfactory
traverse readings of temperature and velocity distribution. Sections 10.6 and 10.7
should be consulted for more details.
In preserving the scale validity between air model and full scale after-
burner, corrections should be made to account for the differences in density of
the model and the prototype. Thring (Reference 15-3) gives a method for this
correction, based on the relative momentum of the stream being considered.
S-1^121 Figures 15-2 and 15-3 follow
-------
r-
r
%D
^^^^j*^^^^
2333Z3
1
%D
J
1
%Q
J
. U Bend
L—
7
1
S-14121
67784
H
J_
C. Wall Baffle
Figure 15-2. BLOCKAGE EFFECTS OF TURNS AND BAFFLES
ON RESIDENCE CHAMBER
-------
! - .
Distance Downstream
of Bend
Number of D
2
4
6
N
\
V 1 UJL V ^ Ul «- * * *• '"^ v
Time Taken by
Streamline A*
Time Based on
Average Velocity
0.83
0.87
0.89
* Estimates Based on Unpublished Data.
Ratio is Probably Higher Than Esti-
mated Value Due to Cross Mixing from
Secondary Flows.
A i
w^iy
A
I
"^J^
— oj-
Velocity
Figure 15-3. EFFECT OF L-TURN ON VELOCITY DISTRIBUTION
AND RESIDENCE TIME DISTRIBUTION IN CHAMBER
S-14121
67784
-------
239
Chapter 16. MECHANISMS AND RATES OF REACTIONS IN AFTERBURNER CATALYSTS
16.1 Mechan i sms
In an afterburner catalyst, react ants flow through the- bed (or matrix)
and reaction products are carried away by the flow of a waste (gas) stream. (See
descriptions of the structures of such catalysts in Section U.3.1) The reac-
tions which convert reactants to product molecules, however, take place at
discrete catalytic sites distributed throughout a thin, porous si'bstrtte layer
coated on the surface of the matrix. The overall process of converting react°r.ts
to products in the flowing stream involves a number of sequential steps including
transport processes between gas stream and catalytic sites as well as the cotual
steps in the reaction at the catalyst sites. These steps can be listed:
l) Mass transfer of combustibles and oxygen to the external surface 01
the catalyst,
2) Diffusion of combustibles and oxygen into the pores of the substrate
layer,
3) Adsorption of combustibles and/or oxygen on the active catalytic
sites,
U) Reaction (oxidation) at the active site,
5) Desorption of combustion product molecules from the catalyst sites,
6) Diffusion of combustion products through the porous substrate
to the external surface,
7) Mass transfer of combustion products from the catalyst surface into
the flowing waste stream.
The oxidation reactions taking place in Step (U) are highly exothermic,
liberating considerable heat at the active sites. This in turn must be transferred
to the waste stream by thermal conduction through the porous layer and by
convective or conductive transport from the catalyst surface to the flowing gas
stream.
The qualitative effects of these sequential processes are that concen-
trations of combustibles are lower within the catalyst, at the active sites,
than in the gas stream flowing over its surface (and that comb- _,tion product
concentrations and the temperature are higher within the cat.J.yst). The con-
centration and temperature differences which develop, provide driving forces
for the various diffusion and heat and mass transfer steps in the overall sequence.
Their quantitative importance depends on the relative magnitudes of rate constants
describing each of the steps in the sequence.
A complete description of the rate of the overall process would require
a knowledge and description of each of the seven steps in the sequence. The
present state of knowledge (and available data) on afterburner catalysts does
not permit such a detailed breakdown. However, sufficient data are available
S-Ik 121
-------
to allow independent description of the rates of convective mass transfer processes
(Steps 1 and 7) taking place external to the catalyst, and of the net process
taking nlace within the porous catalyst structure (Steps 2 through 6). The
theory1^~5'7^ and some basis also exists for treatment of the interaction of
diffusion and reaction within the porous catalyst substrate. This interaction
has not been treated explicitly in the data correlations of this report. However,
sufficient calculations using the theory have been made to make it evident that
under normal design rates for these catalysts, the effective reaction rate is
impeded by slow diffusion of reactants through the porous substrate—to sue! an
extent that only that fraction of the catalyst near the surface of the thin
porous substrate layer ("0.005 inch in thickness) is being used effectively to
„catalyze the oxidation reactions. Under these conditions the "chemical rate"
(i.e., the net rate of Steps 2 through 6) is proportional to the external surface
area of the catalyst matrix rather than to the volume of the porous catalyst.
A second consequence of the inhibition by diffusion is the reduction in the
temperature dependence of this rate to approximately half what it would be with-
out diffusional interaction. However, the entire region of practical rates
appears to lie within the region in which this lower (constant) temperature
dependence prevails. Thus, a pragmatic description of the "chemical rate" is
possible without the need for separate calculations of "effectiveness factors"
to account for the interaction of intracatalyst diffusion and chemical reaction.
Heat transfer between catalyst and gas stream is important because it
determines the internal temperature of the catalyst, the temperature at which
chemical reactions take place at the active sites. To be most useful, any cor-
relation of rates of chemical reactions should be in terms of this internal temp-
erature rather than the temperature of the gas stream flowing over the catalyst
surface. Calculations for typical design conditions indicate negligible dif-
ference in temperature between catalyst surface and internal points within the
catalyst (the solid thermal conductivity is high and the layer of porous substrate
is thin). However, near the inlet end of the catalyst bed, when the hydrocarbon
content of the waste stream is high, the temperature difference between gas stream
and catalyst surface can be significant (25 - 100°F).
The same steps are involved in the use of heterogeneous catalysis in
the chemical or petrochemical industries, where the objective is to manufacture a
valuable product; as in catalytic afterburning where the objective is to destroy
an unwanted reactant. However, the relative importance of the steps may be
different since conditions and objectives are different. In industrial hetero-
geneous catalysis, typically 10 - 100 cu ft of catalyst bed are used for each
1000 SCFM gas stream to be reacted, whereas 1 - 2 cu ft/1000 SCFM is typical in
afterburners. If the comparison is made on the basis of actual porous catalyst
in the bed; the rates per unit volume of active catalyst are 2-3 orders of
magnitude higher in afterburners than in conventional, industrial applications
of porous catalysts. To a large extent the higher rates are possible in after-
burners because the desired products are drastic (total) oxidation products and
selectivity to these products is high and is not decreased by conditions leading
to high rates. In industrial catalysis, operation is normally limited to a
relatively narrow region of moderate reaction rates in which the catalytic reaction
to form the desired product (of partial oxidation, partial hydrogenation, -tc.)
S-14121
-------
is strongly favored over other possible reactions of the same re act ants . As a
consequence of the extremely high reaction rates of afterburner catalysts, after-
burners represent one of the few applications of porous heterogenous catalysts in
which the rate of heat and mass transfer from the flowing reactant stream to the
catalyst surface have a strong effect in determining the overall reaction rate.
16.2 Rates of Reactions
l6.2.1 Empirical Observations
The oxidation of most hydrocarbon (and oxygenated hydrocarbon^ species
over supported noble metal catalysts under typical afterburner design conditions
is complete when sufficient oxygen is present. The reaction stoichiometry may
be written:
n C02
with COa and HgO as the products of the catalytic reaction.
The rate of oxidation (reaction) of a given hydrocarbon species over the
catalyst is proportional to its concentration in the gas stream and independent of
the concentration of oxygen in the gas stream (provided there is at least 2$v
oxygen in excess of that required for complete oxidation of gaseous combustibles).
For oxidation of waste streams containing several oxidizable species it is
usually assumed that each species is oxidized independently with its own rate.
The rates of oxidation (under the same conditions) for similar catalysts
prepared on supports of different geometric surface areas are proportional to
the geometric surface area of the matrix (rather than the volume of catalyst).
16.2.2 Rate Expressions
The above empirical observations are consistent with the expression for
the local rate of reaction at a point within an afterburner catalyst bed:
- \ ' *, • • CHC <^-2>
where C.,- is the concentration of a particular hydrocarbon species (lb -mole /ft3)
nu
Vg the superficial velocity of gas through the bed (ft/sec)
Z the length dimension in the direction of flow (f+?
— the ratio of geometrical (external) surface to volume of the
V catalyst matrix (ft2/ft3)
IT is the effective overall rate constant for oxidation (ft/sec).
S-14121
-------
The effects of temperature, flow conditions through the catalyst and catalyst
geometry are consistent with J"~ representing two rate-limiting processes in
series: s
l) Mass transfer of reactant hydrocarbon from the waste stream to the
catalyst surface,
2) An effective chemical reaction within the catalyst, localized near
its external surface.
The rate of each process is_ first-order in reactant hydrocarbon concen-
tration. The overall rate constant ks can then be represented in terms of the
rates of these two processes:
(16-3)
16.2.2.1 Chemical Rate Constants
The chemical rate constant, k , , which represents the rate of the net
processes taking place within the porous substrate layer,is determined by the chemi-
cal composition and microstructure of the active metal crystallites distributed in
that layer and by the microstructure of the porous substrate. These factors are
in turn determined by the method used in preparing the catalyst. The chemical
rate constant depends also upon the species being oxidized, although for some
catalysts, hydrocarbon species covering a relatively wide range of molecular
structures have quite similar rate constants (see Section k.2.2). It has a
relatively strong dependence upon temperature, which can be represented over a
reasonable temperature range by the Arrhenius form:
k. =Aexp(-E/RT) (l6-U)
chem * a g s
where A is the pre-exponential factor depending upon catalyst type and
species being oxidized
E& is the activation energy
R the gas constant
D
T the absolute temperature of the catalyst.
S
At lower temperatures the behavior of the overall rate constant, k , is limited
by k . and its -""alue approaches that of k , . It is from data obtained at
* chem chem
lower temperatures, where the overall rate shows strong temperature dependence,
that one is best able to derive correlations for k , . Data concerning the
chem
chemical activity or rate constants of afterburner catalysts are usually considered
proprietary by the equipment manufacturer. Consequently, few such data are
generally available for development of correlations for k . Some data (see
S-1M21
-------
Figure U-5) have been published on relative overall rate constants for all-metal
catalysts. These data provide some basis for estimation of an average value
of Ea for this catalyst and for estimating relative values (ratios) of A factors
for the hydrocarbons included in the figure. However, absolute values of A
factors cannot be derived from the data presented. Data, such as that in Table U-2
on ignition and operating temperatures give some indication of values of kchem>
A few additional data have also been published in the form of plots such as
that reproduced in Figure U-2, giving conversion versus temperature for a specific
hydrocarbon/catalyst combination under defined operating conditions. These data
are the most useful in developing correlations—but are very limited in number.
16.2.2.2 Mass Transfer Coefficients
The other rate limiting factor, k^, is determined by the geometry
(matrix structure) of the catalyst bed, and by the hydrodynamics of flow of the
waste stream through the catalyst bed. It has a relatively slight dependence vptn
temperature and upon the nature (molecular size and shape) of the hydrocarbon
being oxidized. The overall rate constant, ks, approaches the value of k^ as a
limit at high temperature when the value of kcnem has become large relative to km^
(see Figure U-5). Data on performance of afterburner catalysts at high tempera-
tures provides one source of information for developing correlations of k^.
However, a much broader base of information is available from the heat transfer
literature, if the analogy between heat transfer and mass transfer under the same
hydrodynamic conditions is used. Theoretical calculations or experimental studies
have been made to develop heat transfer correlations for many matrix structures
either identical with or similar in geometry to those used as supports for
afterburner catalysts16"2^ 3'**'6'. These correlations provide directly a basis for
obtaining heat transfer coefficients to allow calculation of the temperature
difference between gas stream and catalyst surface. Indirectly, they allow
independent development of correlations for gas-to-solid surface mass transfer
coefficients by analogy.
The results of heat transfer studies on a particular matrix geometry
can usually be expressed in a correlation of the form:
"NU ' T • f
-------
p is the viscosity of the gas (ib/ft-hr)
e the fraction void of the matrix.
N is the dimensionless Prandtl number, Cp p/k
Cp is the heat capacity of the gas (Btu/lb-°F)
f is the correlating function, the form of which depends on the matrix
geometry .
The analogy between heat and mass transfer, as used here, utilizes such a
heat transfer correlation to develop the mass transfer correlation, which is
written:
"Sh • T- • f <"Be, V
in which for a given matrix, f is the same function of Reynolds' number as in
Equation (16-5), but with the Schmidt number, Ng , replacing the Prandtl number
where it appeared in the function in Equation (lo-5).
Nc, is the dimensionless Sherwood number, k , L/D
on mt
k . is the gas to solid mass transfer coefficient (ft/hr)
mt
D is the gas diffusivity (ft2/hr)
Ng is the dimensionless Schmidt number, y/pD
p is the gas density (lb/ft3).
This analogy has been used to develop correlations for k . for all of the matrix
structures in use as afterburners. These are presented in Chapter 17- Where
possible these have been compared with mass transfer data obtained from high
temperature performance of afterburner catalysts.
The matrices used as catalyst supports fall into two general classes
with respect to their hydrodynamic interaction with the fluid flow. The honeycomb
types are essentially bundles of parallel tubes with developing (or in the limit
fully developed) laminar flow through these tubes. The all-metal and Oxycat con-
figurations approach more nearly the case of cross flow through a tube bundle ,
or flow through a screen or mesh array.
Kays and London16'2^ present analytical solutions for the Nusselt (or
Sherwood) numbers for fully developed laminar flow through tubes of various
cross-sectional geometries and the results of numerical solutions of the exact
governing equations for development of laminar flow in smooth tubes of circular
cross section. The numerical solution was curve-fit and combined with the
analytical solutions for fully developed flow to give a general semi-theoretical
correlation for mass (or heat) transfer coefficients in honeycomb matrices. This
is:
Ngh = B {I + 0.078 L/X NRe Ngc] ' (IC-7)
-------
2U5
where L the length dimension used in Ngn and N^e and also appearing
explicitly is four times the hydraulic radius of the individual
tube in the matrix.
X is the length of the uninterrupted tube (the thickness of a layer of
catalyst matrix material in the catalyst element).
B is a constant depending upon the shape of the tube cross section —
B = 3.66 for circular tubes, 2.35 for triangular tubes.
The mass transfer coefficient used in Ng^ is the length-mean coefficient over the
length, X. The form of the equation, with respect to dependence of 1:^ on L,
X, and G was supported by data obtained by D. M. Sowards of the Indust.-ial and
Biochemical Department at E. J. duPont. On the basis of comparison of the equation
with those data, the equation was modified to contain a somewhat higher dependence
on the tube length between interruptions in the developing boundary layer.
0.«*5
Ngh = B [1 + 0.095 L/X NRe NSc] (16-8)
(This empirical adjustment is not without theoretical basis , since the original
equation was developed for flow through tubes with smooth walls . Surface rough-
ness in the catalyst matrix material may well delay boundary layer development).
It should be noted that as tube length is increased or diameter decreased the
Sherwood number approaches its asymptotic value, B, for fully developed laminar
flow. With decreasing length to diameter ratio of the tube (X/L) and increasing
Reynold's number the second term within the brackets can make a significant contri-
bution to the value of the Sherwood number. In a typical design with 1/8-inch
honeycomb used in 1-inch thick layers , the Sherwood number may be 13 compared with
3.66 for fully developed flow in the tubes. Thus, for this catalyst, frequent
interruption of the boundary layer and high mass velocity contribute to higher
values of kmt and improved conversion over the catalyst in the region of mass
transfer control. On the other hand, under typical design conditions for catalyst
supported on 8 C/inch Thermacomb , the Sherwood number may be only 2.8 compared
with the value of 2.35 for fully developed laminar flow. With the smaller tube
size of this latter catalyst, the mass transfer coefficient becomes almost
independent of gas flow rate through the catalyst.
For the Oxycat and metal-ribbon supports, which approximate cro£,s-flow
configurations, correlations of the Grimison form1^1*^ are recommended for
Sherwood and Nusselt numbers.
where a and b are empirical constants, with values between Q.'s and 0.6 for banks
of staggered tubes. The length dimension used in Reynolds', Sherwood and Nusselt
numbers is an equivalent diameter, De, equal to the perimeter (of the tube or
ribbon) divided by IT. The void fraction used in the Reynolds1 number is that for
the minimum flow cross section (rather than the average e. for the entire matrix) .
S-
-------
Correlations of the type given in Equation 16-9 have a history of
successful use in correlating and predicting heat transfer coefficients for banks
of circular tubes . The extension here to arrays of elements of other than
circular cross section is an extrapolation , but using accepted methods . Most
of the data on which the Grimison correlations are based were obtained at relatively
high Reynolds' numbers (>5 x 103). In selecting the coefficients for Equation
l6-9, data for heat and mass transfer coefficients to arrays of wire screens at
low Reynolds' numbers were also taken into account16-2^ 6^ . The correlation given
in Chapter 17 for the metal ribbon catalyst gives reasonable agreement with -Miese
data. Unfortunately, no direct confirmatory data from catalyst performance are
available for the correlations proposed for either the Oxycat or the metal ribbon
matrix. (Although presumably the manufacturers of these catalysts do have such
data, they consider them to be proprietary.)
A third general type of form for heat and mass transfer correlations
for catalysts is that for conventional beds of catalyst particles (generally
spherical or short cylindrical pellets packed randomly). Beek16"1^ reviewed
available data and recommended the form:
NQV, = 2.U2 JL X/3 N 1/3 + 0.129 N 0.8 „ 0.4 + 1-U . 0.2 (i6_lO)
Sh Re Sc Re Sc Re
This equation, like Equation l6-9j is a strictly empirical correlation, but does
take into account with the three separate terms, the shifting dependence of the
transfer coefficient with Reynolds' number as Reynolds' number increases. In
Equation 16-10, the length dimension used is the effective particle diameter,
and the Reynolds ' number is defined as :
d G
NRe= ^T
using only the superficial mass velocity, G, without including the fraction voids
in the bed as a factor.
S-1M21
-------
Chapter 17. DESIGN METHODS POR CATALYTIC AFTERBURNERS
17-1 Introduction
The objective of this chapter is to outline and present c method for
calculating the oxidation performance of a given volume and geometry (length and
cross-sectional area) of catalyst bed or matrix. These calculations of performance
are based on a minimum of experimental data obtained from measurements of
catalyst oxidation performance (in correlated form) augmented by theoi°tical
and semi-theoretical predictions of heat and mass transfer coefficient? and
friction factors. The model used in formulating the method is sufficiently
realistic in representing the processes actually taking place within the ctxt-alyst
bed that effects of gas velocity, temperature, concentration of hydrocarbon ii-
the waste stream, and catalyst bed length are well represented in their combined
effects on performance.
The calculation of catalyst performance is not the only problem in
design of a catalytic afterburner. The mechanical and geometrical design of the
preheat-mixing section to provide a uniform flow of preheated waste stream through
the catalyst bed is also important. The considerations and design methods involved
for the preheat-mixing section are similar to those discussed in Chapter lU for
thermal afterburners. Although, in general, the ratio of combustion products to
fume stream to be mixed in the preheat section is lower for catalytic than for
thermal afterburners, it lies within the range discussed in that chapter.
The performance calculation described here is aimed explicitly at the
performance of fresh catalyst (where activity has not been reduced by poisoning or
surface deposits). Its extension to calculations for aged, poisoned, or
suppressed catalyst would be straightforward, but would require significantly
more information than is currently available concerning both the nature and
quantitative extent of catalyst fouling. It was pointed out in Section U.2.H
that over-deisng of the catalyst bed to allow for decline of catylst activity
is an important aspect of overall afterburner design.
To a large extent, in current practive the catalyst volume and geometry
are specified on the basis of subjective evaluation and extrapolation from
the results of pilot experiments. At least one catalyst manufacturer has
developed a computer program for performance calculations similar to that described
here. His method does not separate mass transfer from chemical rate effects and
hence does not describe the influence of gas velocity on oxidation performance.
All of the equipment and catalyst manufacturers have avai" able correla-
tions (at least graphical) of the effects on pressure drop of temperature, gas
velocity and bed length for the catalyst(s) they use. The method described here
incorporates such pressure drop calculations into the overall performance
calculation.
S-
-------
21*8
17.2 Performance Calculation Method
The method requires that the composition of the waste stream entering
the catalyst bed be characterized in terms of the concentrations (ib moles/lb
waste gas stream) of n combustible species (or groupings of species). The heat of
combustion (Btu/lb mole) , "chemical" oxidation rate constant as a function of
temperature as defined in Chapter l6, and diffusivity must be supplied
for each of the n species. Performance is calculated as the fraction of each of
the n species remaining unoxidized after passing through a specified bed lenL th
with specified inlet temperature (preheat) and mass velocity.
Differential equations describing chemical and physical processes in the
catalyst bed are formulated using the models of these processes described in
Chapter l6. These equations are formulated using normal methods used in developing
reactor design equations for industrial catalytic reactors. The only major
assumptions made are that the catalyst bed operates adiabatically and at steady-
state , and that flow rate of the preheated waste stream is uniform over the bed
cross section.
The governing differential equations for the axial changes of concen-
trations, temperature and pressure of the flowing gas stream can be written (see
Nomenclature at end of this chapter):
dC _
- ^^s.i vCi pg (i-1.2- ' • n> <"-!>
dT
_
C G -r=£ = E k . - C.p Q. (17-2)
p dZ s,i v iKg ^i
An auxiliary relationship of gas temperature and catalyst temperature is needed
to allow calculation of the rate constants, k~s ^ at the solid (catalyst)
temperature :
The other parameters appearing in these equations may be calculated from
correlations:
1 = 1 + 1 (17-5)
k . k . . k .
s,i mt,i chem, i
i, _ A I I •*•"» 1 I •*• i II \ J- I —" I
chem.i
S-
-------
214-9
L NNu (17'7)
k = Di N
mt.i — WSh (17-8)
For the simple case in which the heating value of the fume stream is
negligible, all of the Qi's are taken as zero. In this case, Ts = T_ = constant,
p is constant, and Equations 17-1 and 17-3 become integrable (analytically).
(Equations 17-2 and 17-U are trivial for this case.)
For the more general case, when numerical methods must be used to solve
the equation set, it is convenient to rewrite Equations 17-1 in terms of the
fractional concentration (rather than actual concentration), y. of the hydrocarbon
species.
where C°^ is the initial concentration upstream of the catalyst bed.
or
d !« yi _ s P» ,
IT- =-ks,iv- G* (17-10)
The equation for gas temperature change, ATg, over any interval of bed length may
be written in terms of the changes in the dimensionless concentrations ,
over the same length:
For a sufficiently short interval of bed length, Equation 17-1* for the solid
temperature can also be written in terms of the Ay± over the length AZ:
°. Q, (17-12)
where the temperatures , Ts and T are average values over the increment of
length. The Equations 17-3, 17-5 to 17-8, 17-10, 17-11 and 17-12 form a convenient
set for numerical solution for the general case when the Qi are not zero. This
set is reproduced in Table 17-1.
S-11H21
-------
250
Table 17-1. RECOMMENDED EQUATION SET FOR NUMERICAL SOLUTION
= -k . -7T- (i = 1,2. • • n) (17-10)
P o . i- t Ay. C°. Q. (17-11)
g Cp i
T = T - —— ? Ay. C°. Q. (17-12)
Sff R T T T T
B i o * r» A J- A *
dP G
dZ ~ 2pg gcez L
ks,i kmt,i kchem,i
(17-3)
(17-5)
(17-6)
^ NNu (l?-7)
k = ^N_. (17-8)
nrt ,1 L Sh
Tables 17-2 and 17-3 follow
-------
TABLE 17-2
c LOGICAL SECTION FROM FORTRAN PROGRJAM FOR! SOLUTION OF DESIG.N |
J£ .EQUATIONS Fn«S TATAI YTTT AJFTFRBURNFRS ' J . ...
-i • • -1
C DIVIDE TOTAL BED L
DELZ = ZTOIAL/.FLOA
C ;SET INITIAL :CCNDIT
1C_ 'Ltffi£Xi|ll LDENJI.FUE
i |C INDEX(IN) IDENTIFIE
! P ( 1 ) =i P I N
|C ;N=1 IS INLET TO BE
... C ;N=N IS INLET TO NT
,T(1) = TIN i
:i C _I._LS_IHE_fiASLl£ME£
• ! DO 1 Il=l,NHC
. 1 Y(I.l) = l.Q
C Y( I) IIS THE IFRACTI
!C CINI Ii)» OF THE I •
C CALCULATE CI:N,Q PR
i 1 CQU ) - CLNJ I )*QU
ENGTH :FOR CA
T(NSEG)
IONS i
S HYDRlQCARBO
S LENGTH SEG
D, N=N]SEG-H
H SEGMENT* N
RATURE!
LCULATlONt ZJTOTAL,: INTO
N SPEdlES i i
MENT I|N BED !
IS OUTLET !
*N+.l IjS OUTLET FRdM NTH.
; J
i i i
N "SEGMENTS
j
SEGMENT
njn.i-wjvu,.. — ,., - - ,. , . | . .
i ; ; . i
: ; ! ! i
j | i ; I \
ON REMAINING] OF THE ORIGIINAL C]ONCENTJRATIOH,
T H SPEC I.ES 1 ! i | I
ODUCT. Q IS THE HEAT OF) COMBUSTION |OF SPECIES
) . i : i i i
': ;C START CALCULATION [FOR LENGTH SEGMENTS :
'"'. TBAR =! TUN) ' ' 1 ; !
JC TBAR IS ARITH, AVG. TEMP. OF GAS ACROSS ..SEJGMENT ;
: TS = TJCN) i ! ! j ! !
1C TS IS MEAN. TEMP OR CATALYST SURFACE_JA.AEGHEJlL_L _ a
i 4 TSTOR' - fs ; j I ; i
C TSTOR IS MOST RECENT ASSUMED CATALYSiT TEMpj FOR SEGMENT
CALL RiATE i i i :
c RATE SUBROUTINE USJES CORRELATIONS TO c.Al-cyiLAT.E HIHT co
C F(FRICTION FACT)^ KSBAR(I) (OVERALL 1ST CJRDER RATE CO
(C RHQ(GAS DENjSlTY) j i _[_ ! _], L . „
: 1 DELT = o.o : ! i j \
1C DELT IS THE GAS TEiMP RISE ACROSS INCREMENTj
DO 3 "i|=i"»NHC 1 i i : !
! Y(N+1,|I) = Y(N.l)/;EXP(KSBAR(I)i*SONV*DELZ*RHO/G)
•C PSE'JDdANALYTlCAL SOL ' N USED FOJR EQN ;17-10 J
C SONV IS SURFACE TQ VOLUME RAllM^F _MA.TRI X^ G LS^GASJ^?
3 DELT = DELT ;+ CQ( li) *( Y( N, I >-Y< N+l , I )!) i
•T ( N+ i.) .. « TIN.). . .+. . PELT /CP i J > ; _
; TBAR = (T(N-H) + T(N) J/2.0 , ! ;
i TS .= .TJBAR ± DELT*G:/..(H*SaiNy*DELJZj ] ,
C REPEAT] CALCULATION: UNTIL CATALIYST TEMP, TS, CONVERGES
; IJFjLABS
-------
Table 17-3. GEOMETRIC PROPERTIES OF AFTERBURNER CATALYSTS AND
RECOMMENDED CORRELATIONS"FOR TRANSPORT PROPERTIES
Catalyst Type
Torvex 2B
(1/8 In. hexcell honeycomb)
Thermocomb 8 c/in.
honeycomb
Oxycats
Metal ribbon "D" series
fc
Spherical catalyst pellets
S/V
(n-1)
268
695
36.5
•
336
'
6c
dp
c
0.6l
0.6
0.515
• .
0.93
normally
0.35 - 0.1.
L
(ft)
0.0091
0.0031.5
0.0133
O.OOU
dp
X
(ft)
0.0333
nominally
O.lfr' - 0.333
-
-
-
NNu, NSh. f
NNu = 3'66 (1 * °'°95 NRe NPr -^'^
N ' ? 0 -5
NSh = 3<66 (1 i °'°95 MRe NSc X} "
N = 2.35 (1 + 0.095 N M — ) '
Ng. = 2.35 (1 + 0.095 H Ng -)
f = 53.3/NRe (1 + 0.01.1.5 NRe |) '
NNu=0.51NRe0.5eNpr0.333
NSh=0-51NHe°-56HSc°-333
f = 0.5/NRe°-15
MNu"0-«NRe°-5Npr0-"3
f =(l-el/e(5.2 + 3600 (l-e)/HBe)
-AP = r G2/2e2p g f /L dZ f 2.1 C2/2p gj.
Jo « ° g
N =2.1.211 J/3 H J/2 °'8 °-1* k °'2
f = U-c)/ e(3.5 * 300 (l - e)/H_ )
"e
-------
251
A short section presenting the logic for numerical solution of the
set is presented as a partial FORTRAN computer program in Table 17-2. A complete
program using this logic has been used in preparing the figures presented in
Chapter U illustrating the effects of operating variables on catalyst performance.
No problems with convergence or stability of the integration procedure were
encountered when using ten axial increments through the bed in the numerical
integrations, even under conditions where rates were high. Oth^r numerical methods
could be used. However, the pseudo-analytical solution of Equation 17-10 is
recommended since it offers a strong stabilizing influence on the numerical solution
when source terms are large.
17.3 Recommended Correlations of Rate Parameters for Afterburner Catalyst Matrices
17.3.1 Transport Coefficients
Table 17-3 presents geometric properties and recommended correlationj "or
Nusselt and Sherwood numbers and friction factors for the four matrix types in
commercial use, and for spherical catalyst pellets. Geometrical properties are
based on data from catalyst manufacturers. The basis of Nusselt and Sherwood
number correlations was discussed in Chapter l6.
Friction factor correlations for the two honeycomb supports are based
on curve fitting numerical solutions for "f " presented in Figure 6-22
of Reference l6-2. The tube cross section is assumed to be circular for the
"TORVEX" and triangular for the "THERMACOMB" material. These correlations for
friction factor have also been confirmed by comparison with experimental data
covering a wide range of flow rates and temperatures for both "TORVEX" and
"THERMACOMB" catalysts. Data were made available for this comparison by the
Industrial and Biochemicals Department, E. I. duPont.
The "OXYCAT" configuration is a close approach to a conventional heat
exchanger tube bundle. The friction factor correlation recommended is taken
directly from Reference l6-U for cross flow through tube bundles. No confirmatory
data for the correlation are available.
The friction factor expression for the metal ribbon catalyst is b*>.sed on
data presented in Reference U-5 and other data from the manufacturer. The
correlation is in the conventional Ergun form, but coefficients are unusually
high. Because this catalyst is normally used in thin beds and with low overall
pressure drop, the inlet and outlet retaining screens contribute significantly
to the overall pressure drop through the element. An expression for AP incorporating
an additional term which accounts for this contribution is included in Table 17-3.
All of the transport property correlations for sphe.ical catalyst
pellets are those recommended by Beek1^"1^ .
17.3.2 Chemical Rate Constants
The publicly available data on afterburner catalyst performance provide
a meager basis for "chemical" rate constant correlations. A few estimates of
-------
252
rate parameters are presented in Table IT-k. These are based on normal design
specifications, published "ignition" and operating temperatures and the limited
performance data available in the literature. They provide some guide as to the
values to be expected for these parameters. It is to be hoped that catalyst and
afterburner manufacturers can supply more and better data by analysis of their
performance experience.
Table 17-1*. ESTIMATED "CHEMICAL RATE" PARAMETERS FOR
AFTERBURNER CATALYSTS
k . . = A.
chem,i i
Component
k at T a'
chem.i" Ref
(ft/sec)
Alumina- Based Catalysts
H2
CO
Solvent HC's
CHi,
8.0
O.U5
0.1 - 0.2
0.005
Metal-Based Catalysts
Toluene/xylene
MEK
n-heptane
0.2 - 0.3
0.03
0.005
Ea
(Kcal/gm mole)
Ea/Rg
(°R)
8
11
11
15
7,250
10 ,000
10,000
13,600
15
15
15
13,600
13,600
13,600
a) For all of the cases presented in the Table, TD . = 1010°R (550°F).
S-14121
-------
253
Estimation of Performance with Partially Deactivated Catalyst
An estimate of the consequences of partial deactivation of the catalyst
canjbe obtained if the calculation made for fresh catalyst is repeated using values
of k . modified to account for deactivation effects.
For deactivation by aging or poisoning in_which active sites throughout
the substrate are reduced in number, the values of k . could be calculated using
a value of k . . defined. Sjl
chem,i
k ~
chem,i
where f]_ is a multiplier between 0 and 1. The quantity, 1 -f can be regard1 as the
fractional deactivation of the catalyst sites.
If activity decline is due to surface coating by particulates, the
chemical activity would be unchanged. However, mass transfer resistance would
be increased since reactants would have to diffuse through the inert layer to
reach the active catalyst. For this case, the mass transfer coefficient, k , .
might be defined: sl
mt ,i
_
DiNSh kf
where l/k_ is the effective diffusive resistance of the surface coating layer.
For more complex patterns of deactivation, adaptations of the calculation
method could be developed.
S-1M21
-------
25 k
riQMLMCLA'i.'URE
(Chapter 1?)
Pre-extonential fac'-or, value of k , . at reference temperature
(f-/sec). CheD1'1
C Heat capacity of gas stream (Btu/lb-°F)
C. Concentration of ith species in waste stream (ib mole/lb waste stream)
C. ° Concentration of ith species at catalyst bed inlet end (ib mole/lb)
d Particle diameter of spherical catalyst pellets (ft)
De Equivalent diameter of catalyst support (ft)
(Ea/Rg)^ Exponential factor for temperature dependence of chemical rate (°R~1)
f Friction factor
C- Superficial mass velocity of gas through matrix (lb/sec-ft2)
gc Gravitational conversion factor (lbf-ft/lbm-sec2)
h Gas-to-solid heat transfer coefficient (Btu/sec-ft -°F)
k Thermal conductivity of gas (Btu/sec-ft-°F)
k . . Rate constant for oxidation of ith species (ft/sec)
k . Rate constant for mass transfer of ith species (ft/sec)
Hi 0 ) 1
k . Overall rate constant for oxidation of ith species (ft/sec)
L Matrix length dimension used in dimensionless numbers and in pressure
drop equation (ft)
Dimensionless Nusselt number, hL/kg
N.
Np __ Dimensionless Prandtl number, C
N_, Dimensionless Reynolds number, LG/^e
^t, Dimensionless Schmidt number, p/pgD^
"•„. Dimensionless Sherwood number, k . .
bh ' mt,i
P, ^P Pressure, pressure drop through matrix (lbf/ft2)
S-
-------
255
NOMENCLATURE (CONT)
(Chapter 17)
Q. Heat of oxidation of ith species (Btu/lb mole)
r Hydraulic radius of honeycomb matrix (ft)
h
s/v Surface to volume ratio of matrix (ft"1)
Tg Gas temperature (°F or °R)
T Reference temperature for chemical rate constant (°F or °R)
T Temperature of catalyst surface (°F or °R)
X Length between boundary layer interruptions in honeycomb matrices
(thickness of individual layers of catalyst in the bed) (ft)
y. Dimensionless (fractional) concentration of ith species, C./C.°
Z Length dimension in direction of flow through the catalyst bed (ft)
e Fraction voids of matrix
pg Gas density (lb/ft3)
V Gas viscosity (ib/ft-sec)
S-11H21
-------
257
Chapter 18. HEAT BECOVERY - DESIGN FUNDAMENTALS
18.1 Introduction
Chapter 6 described various means for achieving heat recovery in a system
to be equipped with an afterburner. This section gives a presentation of working
relations and data intended for planning the installation, for checking design
proposals, and for evaluating performance of the heat recovery equipment, particu-
larly of cross-flow tubular heat exchangers. The section is not meant to be
complete enough to give a working basis for design of equipment. Hovever, an
engineer having a general familiarity with heat transfer apparatus, >rith fluid
flow, and with elementary structural analysis, could undoubtedly prepare a design.
References will be included, where appropriate, for more complete data or' more
specialized calculations.
The heat transfer relations included here are:
Quantity of Heat - which may be recovered from cooling a hot stream, or
or which is required to heat a cool stream (Equations 18-1 through 18-U,
Section 18.2.1)
Heat Transfer Rate Equations - (Equation 18- 5, Section 18.2.2)
Heat Transfer Coefficients - ( Figuresl8-2,-3, Section 18.2.3)
Heat Exchanger Effectiveness - (Equation 18- Q, Section 18.2.U)
Number of Transfer Units - (Equation 18-9, Section 18.2.5)
Capacity Rate Ratio - (Section 18.2.6)
Cross-Flow Exchanger Performance - (Section 18.2.7)
Exchanger Pressure Drop - (Section 18.3)
Miscellaneous Equations and Relations - (Section 18.3.6)
18.2 Equations and Relations for Heat Transfer
18.2.1 Quantity of Heat to be Transferred
Q = 60w Cp (Tf.nal - T.nitial) (18-1)
Preceding page blank
S-14121
-------
253
where Q is the amount of heat, Btu/hr, for producing the change of
temperature between the limits, T. ... , and T „. ,. It is
assumed there are no condensing or evaporating liquids (that is,
there are no phase changes).
w is the mass flow rate of gases involved, lb/min.
Cp is the average value of specific heat over the temperature r^nge,
(T - T) at constant pressure, in Btu/(lb-°F).
final
The specific heat of a gas varies in a non-linear way with
temperature, but may be approximated for the purposes here by using
the average "instantaneous" values between the two limiting
temperatures .
(Cp) (Cp)
(Cp). - initial _ i final (l8-2)
Average ~ -
T is temperature in degrees F.
Values of specific heat for air and different gases may be taken
from Figure 18-1.
The specific heat of a mixture of gases is found from the specific heats of the
components, all taken over the same temperature range,
Cp2 . .
Cp • WiCpi + w2Cp2 • • • wCp (18-3)
where wif w2 • • • wn are the weight fractions of the n components. If the
composition is given in terms of volume fractions, VA, v2, . . . vn,
then the corresponding weight fractions can be found by
VJ.MI
where MI, Mg . . . are the molecular weights of the gases, as given in
Table 18-1. (Note: mole-fraction is the same as volume fraction.)
In this simplified approach, all gases are treated as "perfect gases",
and all vapor components are assumed to have a low partial pressure so that perfect
gas behavior is approximated.
S-14121 Figure 18-1 follows
-------
0 psia
0.30
0.28
0.26
m
Q.
u
0.24
H2O (Vapor)
0.22
S-14121
67784
CO,
0.56
i
E
_D
0.52 >
£
o
Q.
to
0.48
O
CM
X
o.
(J
0.44
1967 Steam Tables, St. Martins rress, N.Y.
Janaf Thermochemical Data, The Dow Chemical
Company, Midland, Mich.
500
1,500
1,000
Temperature, °F
Figure 18-1. SPECIFIC HEATS OF GASES
2,000
-------
259
Table 18-1. MOLECULAR WEIGHTS OF GASES
.£.
Dry Air
Flue Gas
02
N2
CO
C02
IfeO
CRt Methane
C2Ha Ethane
C2H4 Ethylene
CaHa Propane
CaHe Propylene
^Composition 02
Ha, A
29.0
(Nat. Gas, 10$ Excess Air) 29.3
(Nat. Gas, 400# » » ) 29.0
32.00
28.02
28.01
U.01
18.02
16.04
30.07
28.05
44.09
42.08
20. 91# vol 23.15#wt
79.09 76.85
100.00 100.00 0
S-14121
-------
260
18.2.2 Heat Transfer Rate Equations
The heat transfer rate between two fluids separated by a wall, or
between a fluid and a wall, is calculated by the equation
q = uA (Th - Tc) (18-5)
where q = heat transfer rate, Btu/hr
A = heat transfer area, ft2
T^ = hot fluid (or surface) temperature, °F
T • cold fluid (or surface) temperature, °F
u = overall transfer coefficient Btu/(hr-ft2 - deg F)
18.2.3 Heat Transfer Coefficients
The overall transfer coefficient is made up of a number of components,
depending on the nature of the heat transfer system.
18.2.3.1 Film Coefficients, hf
For a gas-to-gas exhcnager there will be two film coefficients; one on
the inside, one on the outside of the tube, hf accounts for the insulating effect
caused by a thin layer of slowly moving gas, adjacent to the metal surface of the
tube, hf is a complex function of surface geometry, flow channel geometry, fluid
properties (density, viscosity) and average fluid velocity over the surface.
Values of hf have been plotted for a useful range of gas velocities, for flow of
air or flue gas inside tubes (Figure 18-2) and across tubes (Figure 18-3). The
choice of the correct velocity to be used in these graphs will be discussed later.
For other geometries, and for a more complete treatment of a complex subject,
consult W. M. Kays, A. L. London, Compact Heat Exchangers, Reference 18-1.
3.8.2.5.2 Wall Conductance, Cy
For thin-walled tubing this may be approximated by
Cw = (l8-6)
where t is the thickness of the tube wall in inches, k is the thermal conductivity
of the conducting metal in Btu/(hr-ft2-deg F/ft). Values of k for common metals
are given in Figure 18-4. k is evaluated at approximately the median temperature
between flue gas and fume stream. For most afterburner applications this term may
be neglected.
18.2.3.3 Fouling Factor, f
Thick fouling layers on the tube walls may predominate in resisting
heat transfer from one gas to the other. However, predicting f in the absence of
experience on a particular application is largely arbitrary. Conditions may
range from clean (f = 0) to extremely dirty. Figure 18-5 gives values of f that
would correspond to a thickness of asphalt and to petroleum coke. At low
S- llj-121 Figures 18-2 through 18-5 follow
-------
10Z
o
a.
i*.
4,0
00
G = 3600 pV, where p is in lbm/ft3 and V is in ft/sec
Heat Transfer Film Coefficient for Flow of Air Inside
0.868 in. dia. Tubing.
The same curve may be used with Flue Gas with very
little error.
For more complete data, see W.H. McAdams,
"Heat Transmission", Ref. 18-2, Chapter
on Heating and Cooling Inside Tubes.
I I I I I
1
I
I I I I I
103 104
G = Mass Velocity, Ib/hr-ft2
Figure 18-2. HEAT TRANSFER FILM COEFFICIENT FOR FLOW OF
AIR INSIDE TUBES
105
S-14121
67784
-------
102
'"
CQ
Gmax — Mass Velocity, based on minimum free-flow area between
tubes (see Figure 18-10); Gmax = 3600p Vmax/ where p
is in lbm/ft3 and Vmax is in ft/sec.
Approximate Heat Transfer Film Coefficient for Flow of Air Across
Banks of 1.00 in. dia. Tubes.
The same curve may be used with flue gas with very little error.
For more complete data, see W. M. Kays and A. L. London, "Compact
Heat Exchangers'7, Ref. 18-1, pp. 127, 128
I I I I
1 1
103 104
Gmax, Ib/hr-ft2
Figure 18-3. HEAT TRANSFER FILM COEFFICIENT FOR FLOW OF
AIR ACROSS BANKS OF TUBES
105
S-14121
67784
-------
400 800 1200
Temperature, °F
Figure 18-4. THERMAL CONDUCTIVITY OF METALS
1600
S-14121
67784
-------
Btu/(hrft2 °F/ft)
Ref. 18-2
X12
I, in.
1/f = £
f ~I
Asphalt
.43
5.16
1/4
20.4
.049
y*
82.6
.012
}A
41.3
.024
3/
/I6
27.5
.036
Pet. Coke
3.4
40.8
]A
167
.006
V*
653
.0015
%
326
.003
3/
/I6
218
.0046
0.05
0.04 —
0.03 —
0.02 —
0.01 —
Thickness of Deposit
Figure 18-5. FOULING FACTOR
S-14121
67784
-------
261
temperatures, greasy materials, polymerizing resins, etc., will be more or less
like asphalt. At elevated temperature (say 500°F or more), any such materials
are likely to turn into coke-like layers; hence, coke is probably more typical
of what would be encountered in an exchanger exposed to 1500°F flue gas on the
other side of the tube wall. Fouling material may not deposit uniformly over the
exchanger surface, and the conversion to coke will be affected by the local tube
wall temperature in the exchanger. Where possible, the nature of fouling in a
given service should be established by experience.
Table 18-2 gives suggested fouling factors for specific services.
Table 18-2. SUGGESTED FOULING FACTORS FOR
RECUPERATIVE HEAT EXCHANGERS
Service
Deodorizers, odor control
Paint drying ovens
Lacquer drying ovens
Print drying ovens
Steel drum reconditioning
Meat packing and rendering
Pattern Curing Ovens
(Foundry)
Biological Cleanings of
Contaminated Air
Factor
.005
.01
.005
.005
.02
.05
.01
.005
Note: These factors are highly tentative and
should be modified as experience warrants.
18.2.3.1* Overall Coefficient
The overall transfer coefficient is calculated,
1 - 1 * r * n * If "
d.
Chf7
inside tube
w
hf
(18-7)
outside tube
This is evaluated on the basis of the inside area of the tube (d^ = inside
diameter, do = outside diameter), with the further supposition that the cool
fume stream will pass inside the tubes.
18.2.4 Heat Exchange Effectiveness, E
The effectiveness is defined (Ref. 18-i):
S-1U121
-------
262
U (T - T I
r, _ ^ _ fumex fume leaving fume entering/ /,o o.\
c ~ n ~ — /T ^~f \ VIO-OA;
* C in \ flue gas enterinS fume enteringj
C
(T - T ^
\ flue gas entering flue gas leaving/
Also = flue 6as \ flue gas entering flue gas leaving/ / « g \
C . ( flue gas entering ~ fume entering)
~C = wC for the appropriate stream; it is the heating capacity rate
fca.1 the stream
C . = Refers to that stream having the lowest heating capacity rate.
Since the fume stream has less mass flow than the flue gas (which
includes the fuel and added combustion air), and Cp fume K Cp flue
gas
^min = °fume = ^Wfume CPfume'
Cp = The average specific heat of ttie gas at constant pressure over
the temperature range Btu/(lb-deg F). (See Eq. 18-3)
w = Mass flow rate of stream, Ib/hr.
Equation 18-8A thus reduced to:
I
_ \
(
T - T i
fume leaving fume entering/ Qg g x
flue gas entering fume entering
Equation 18-8B becomes :
"C I T - T I
flue gas \ flue gas entering flue gas leaving'
- - - - - - -
_
E -
T; \
Cf ( flue gas entering fume entering)
The effectiveness is a measure of the performance of the exchanger,
comparing the actual recovery of heat with that which might be obtained using an
ideal exchanger of infinite size.
18.2.5 Number of Transfer Units, N
The number of transfer units is a non-dimensional grouping of terms
that characterizes the size of the exchanger.
N = Au . • A = v CP Ntu dS-o)
Ntu (w Cp) ' A - S -
where A is the heat transfer area of the exchanger inside the tubes, ft2.
S-
-------
263
A = (n/10 (1/lWO NL d.;2
1 (16-10)
= .005^5 NL dj2
taking A = number of tubes
L = length of tubes between tube plates, ft.
di = tube inside diameter, in.
The exposed area of the tube sheets is ignored, u is the overall heat transfer
coefficient, Btu/(hr-ft2-deg F) based on inside area, w is the fume gas flow rate,
Ib/hr, and C is the average specific heat of the fume stream over the tempera tui °
interval.
18.2.6 Capacity Rate Ratio
fume _ wfume stream Pfume stream _
_.
~
C w xC
flue gas flue gas stream Pflue gas stream
Again, w is in Ib/hr, and Cp is in Btu/(lb-deg F) and is taken at the average
value over the temperature interval of fume or flue gas stream. Since the flue
gas stream is enlarged by the mass rate of the fuel, and the value of Cp for the
flue gas is increased somewhat by the average temperature, R for this application
is usually a number slightly less than 1.0.
18.2.7 Cross Flow Exchanger Performance
The heat transfer effectiveness can be expressed as a function of Ntu> R
and flow arrangement. Solution of the mathematical equations for heat transfer
effectiveness with different arrangements has been carried out by computer and is
given in Figure 18-6 for a single pass cross-flow exchanger, with fume stream
passing through the tubes, and flue gases across the outside of the tubes (unmixed),
Figure 18-7 gives a similar plot for a 2-or-more passes, with R = 1. For more
than two passes, the performance closely approaches that of a true counterflcw
exchanger. For a more complete exposition of the different exchanger arrangements
and their effectiveness, see Kays and London, Ref. 18-1, Chapter 2.
The effectiveness-N-tu approach used here is preferred over the somewhat
more familiar relation based on log-mean-temperature-difference, since the latter
requires a solution by successive approximations for many practical problems,
especially in predicting the performance of an existing exchanger, for which the
geometry is fixed. It is also apparent from Figures 18.6 and 18.7 that the
achievement of high effectiveness is gained only at the expense of relatively
large size exchangers (high N-tu); also that the maximum atta-\iable effectiveness,
even for a very large size exchanger, is limited to about 6Cff> for a single-pass
cross- flow arrangement.
18.3 Exchanger Pressure Drop Calculations
The calculation of pressure drop through the heat recovery system
requires some knowledge of the detailed geometry of both the exchanger and its
ducting. Since the latter is subject to the great variety of the many possible
S-1U21
-------
26k
installation!, the notes given here relate primarily to the heat exchanger, which
is assumed to be a shell-and-tube cross flow unit.
For method and data on calculating pressure losses in ducting, including
elbows, bends, tees, changes in cross section, headers, etc., refer to the American
Society of Heating, Refrigeration, and Air Conditioning Engineers, Handbook of
Fundamentals (commonly known as the ASHRACE Guide), Reference l8-3«
The pressure drop through the exchanger is made up of several par s:
Headers (Section l8.3«l)
Entrance losses into exchanger tubes, cold side (Section 18.3-2)
Fluid acceleration from changes in temperature. Since velocities are
generally low, this component of pressure drop will be ignored here.
Tube wall friction (or channel flow friction) (Section 18.3-3)
Exit pressure gain on leaving exchanger tubes, cold side (Section
18.3.2)
These pressure drops must be repeated for cross flow over tubes, hot
side (Section 18.3-4)
All these items contribute to the overall pressure loss and affect the
part of blower power which is chargeable to the heat recovery system.
18.3.1 Headers
The distribution of fume and flue gas streams into the exchanger is
important; improper header design will lead to non-uniform gas velocities through
the exchanger, possibly to impingement of fouling materials, and consequently to
hot spots, warpage (or failure), and reduced effectiveness. Abrupt expansion in
area from the supply duct, immediately following an elbow, for example, can give a
very poor flow distribution into the exchanger. On the other hand, turning vanes
in the elbow intended to improve distribution might become badly fouled and thus
affect distribution—probably systems subject to such severe fouling should not
be candidates for a heat exchanger. The ASHRACE Handbook (fief. 18-3) can give
some guidance on distribution and header design. Figure 18-8 gives examples.
"A" shows an inclined duct arrangement that gives good distribution without
turning or splitter vanes. "B" shows the use of vanes on the inlet sides of the
exchanger. A relatively abrupt transition may be used on the outlet side without
appreciable effect on distribution in the exchanger.
Header pressure losses, if accounted for, should be included in the
ducting calculations.
18.3.2 Entrance and Exit Losses in Tubes
The entrance pressure loss and exit pressure gain for flow through the
exchanger tubes are ordinarily not large, and may be neglected for most cases.
S-1U121 Figures 18-6 through 18-9 follow
-------
O> to
COURTESY W.M. Kays and A.L London. "Compact ffeat Exchangers'
2nd Ed.. McGraw Hill. 1964. with the Permission of rhi'
Publisher
Mixed
Fluid
Unmixed Fluid
Number c1" Transfer Units/ Nj.y = AU/C
mn
.x^^v
-t-M-
-H-i-
4-H-
.
*M.'
90
80
3?
« 70
i/i
o>
u
'! °0
v
50
40
Counterflow
0123
Number of Transfer Units/
45
= AU/Cm;n
Figure 18-6. HEAT TRANSFER EFFECTIVENESS AS Figure 18-7. HEAT TRANSFER EFFECTIVENESS AS
A FUNCTION OF NUMBER OF TRANSFER UNITS
AND CAPACITY RATE RATIO; CROSSFLOW
EXCHANGER WITH ONF FLUID MIXED
A FUNCTION OF NUMBER OF TRANSFER UNITS
AND NUMBER OF PASSES, MULTIPASS CROSS-
COUNTERFLOW EXCHANGER Cmin/Cmax = 1;
UNMIXED FLOW WITHiN PASES/ ONE FLUID
UNMIXED BETWEEN PASSES
-------
To Stack
Fume Stream
A. Good Practice
To Stack
To
Afterburner
From
Afterburner
To
Afterburner
Flue Gas
B. Good Practice
S-14121
67784
C. Poor Practice
Figure 18-8. HEAT EXCHANGER HEADERS
-------
Pressure Loss at Entrance
— — Pressure Rise at Exit
For turbulent flow conditions, with Reynolds Number
between 3000 and 10000, with fluid density of 0.075
Ib/ft3 (i.e., dry air at ambient pressure and tempera-
ture.) For different gas density, adjust in simple
proportion.
Reference: W.M. Kays and A. L. London, "Compact
Heat Exchangers", Ref. 18-1.
a is the ratio of tube free-flow area to the
frontal area of the bundle.
= 0.3:
C 5
14-
o
0
C
I
10 20
Velocity in Tubes, ft/sec
Figure 18-9. ENTRANCE AND EXIT LOSSES FOR
TURBULENT FLOW IN TUBES
30
S-14121
67784
-------
265
For example, for air at atmospheric pressure with an approach velocity of
10 ft/sec and an area ratio, a of 0.25 (defined below), the entrance pressure
loss is .03-inch water.
The entrance pressure drop can be calculated for flow through the tubes
by (Ref. 1, Chapter 5J:
Ap =0.003 pVi2 (1.5 - O.lf a - a2) (l8-12)
The exit pressure rise regained after flow through the tubes can be
calculated from:
Ap =0.003 pV22 (2 + 1.9cr - 1.602) (18-13)
Here Ap is the pressure drop in Inches of water; Vi and Va are the approach and
exit velocities in the header, ft/sec; p is the gas density (an average value
through the exchanger can be used), lb/ft3; and a is the area contraction ratio
between the header cross section and the tube openings
n nd-j3 /, Q ., x
"thru = £ "I"- (l8-ll)-)
where n = number of tubes
d^ = inside diameter of tubes, in.
A = the area of the tube plate exposed to the header, in.2
Figure 18-9 gives a plot of entrance and exit Ap for a range of approach
velocities and a.
l8.3'3 Friction Through Exchanger Tubes
This pressure drop can be calculated as follows:
Ap = 0.0011 -up pV2avg (18-15)
where Ap = pressure drop from tube wall friction, in. water.
L - the length of the tubes, in.
d^ • the inside diameter of tubes, in.
p • average gas density between inlet and outlet exchanger ft/lb
V « average gas velocity inside the tubes, based on mean of
entrance and exit densities, ft/sec.
(The equation is based on a Fanning friction factor of .008, which is in turn
based on the assumption of Reynolds Number between 3,000 and 10,000, and on smooth
tube walls; the friction factor becomes smaller at higher Reynolds Numbers, so Ap
calculated is conservative. A more accurate calculation will account for
variations in tube diameter and physical properties of the stream; as in Ref.
18-2.)
S-1U121
-------
266
18.3.4 Pressure Drop for Flow Across Tubes
For flow across the tubes, entrance and exit effects are not accounted
for separately. The entire flow channel through the exchanger is treated as a
series of abrupt contractions and expansions.
For smooth round tubes with in-line arrangement:
Ap = a0010 NpV2 (18-16)
ulcuC
If the tubes are staggered in a triangular pattern:
Ap =0.0014 NpV2 ov (18-17)
max
Here Ap is the pressure drop through the tube banH inches of water, N is the
number of tube rows deep through which the gases must pass, p is the average gas
density in passing through the exchanger pass, and Vmax is the velocity based on
the smallest area between the tubes in the flow direction. Figure 18-10
illustrates the basis for determining this velocity. (The equations are based
on equations given by McAdams Heat Transmission, Ref. 18-2, McGraw-Hill, for tube
spacing of two diameters and Reynolds Number of 10,000. A more accurate calcula-
tion will allow for the effect of different tube layouts, and variations in fluid
properties. See Ref. 18-4.)
18.3.3 Balancing Pressure Drop Against Exchanger Size
In selecting an exchanger, one often starts by specifying a desired
effectiveness. Usually there is also a limited total pressure drop available for
the complete system, which is determined by the flow resistance of afterburner,
ductwork, heat recovery system, stack, and perhaps the oven or process unit as
well, if all are driven by a single blower. The pressure drop assigned to the
heat recovery unit, assumed here to be a cross flow heat exchanger, is thus a
rather abritrary matter. It is a matter of economic balance to trade off size and
cost of the exchanger that will achieve the desired effectiveness against the
pressure drop that can be afforded. The key variable is the velocity used in th<-
heat exchanger passes. The higher the velocity, the larger the value of the V"^at
transfer coefficient, the smaller the passage flow areas, and the smaller the
exchanger; the reduced exchanger size must be paid for by increased blcwer power.
The pressure drop is very sensitive to the velocity, since it varies as the
square of the velocity.
The process of design is a trail-and-error one, starting from a range
of velocities, from which are calculated the exchanger sizes, effectiveness ratios,
and pressure drops, so that one can establish which design meett, requirements at
the lowest cost. Ideally one could go to great lengths adjusting the many design
variables to achieve a balanced and optimum design. Practically, the design is
limited by the ready availability of standard designs, based on standard tubing
sizes, tube layouts, frame sizes, etc., and a familiarity or predilection a given
manufacturer may have for his own special design of flattened tubes, rectangular
tubes with corrugations, etc., which are his stock in trade. Further, the
manufacturer accumulates his own data and operating experience, which simplify
his application.
S-14121 Figure 18-10 follows
-------
n = No.
of Rows Deep
(4 in This Diagram)
Flow
Direction
o o o o
o o o o
o co o
<5P~cb o o
Tubes In-line
Flow
Direction
m = No. of
Flow Passages
Across Exchanger
n = No. of
Rows Deep
(4 in This Diagram)
(Equilateral Spacing)
2~O O O
O/'O O
o o ^o o
l-p-1
0066
Tubes Staggered
m = No. of
Passages Across
Exchanger
P
V
Q
m, n
L
p
o
= Tube to Tube Centerline Distance, in.
= ft/sec
= ft3/min at Gas Pressure and Temperature
= Numbers of Tubes in Row
= Tube Length, ft
= Minimum Tube Clearance, ft
= (See Figure 18-11)
Figure 18-10. TUBE ARRANGEMENTS FOR CROSS FLOW
S-14121
67784
-------
267
18.3.6 Other Miscellaneous Equations
Average Gas Density,
Average Mole Weight
Mavg ' ^-VT "-8-1"
or M = vi MI + v2 M2 . . . (18-20)
avg
where wi,w2 . . . = weight fraction (percent by weight/100) of components
1,2, . . . whose molecules weights are MI, M2 . . .
vl> V2 • • • = volume fraction (percent by volume/100 of components
1,2. . . , etc.
p = pressure of gas, lb_/in^ gauge
T
avg = average temperature of gas, °F.
p = average density, Ib /ft3
-------
268
l8.4 Sizing an Exchanger - Procedure
Tne assumptions used in the following section are to be taken as
suggestions, not prescriptions. It is suggested that Chapter 6 be reviewed before
attempting to size an exchanger, since the general design considerations discussed
there may be of overriding importance in choosing design assumptions.
Despite the direct relationship between E, R, and N^ (given here in
graphical form) the design problem is a trail-and-error one, since a number of
constraints must be applied (available pressure drop, practical tube layout, etc.)
which cannot be included in a direct relationship. Steps for sizing to meet
heat transfer requirements are:
l) Select a desired effectiveness ratio.
2) Determine inlet, outlet, and average temperatures for both fume and
flue gas streams (Equation 18-8).
3) Find inlet, outlet, and average gas densities for both streams
(Equations 18-18, 18-19, or 18-20).
4) Select tube size, spacing, and arrangement.
Suggested Assumptions (Based on Typical Installations)
a) Fume stream passes through tubes
b) Flue gas passes across tubes
c) Tube diameter 1" OD, 0.868" ID stainless steel
d) Tubes staggered, on 60° centers, spaced 1-1/2" apart.
5) Using Figures 18-10 and 18-11, determine the free flow area ratios,
a cross and CTthru*
6) Assume a design velocity, (vmax)inside* ^or ^•"•ow through the tubes.
The pressure drop is a strong function of this velocity; it is suggested that two
or three values be assumed, and that corresponding calculations be brought along.
Suggested values: 40, 50, 60 ft/sec.
T) Based on this velocity (or velocities), using Figure 18-2, determine
the inside heat transfer film coefficient, hf^. Note that Figure 18-2 gives only
an approximate value of h^.. based on an assumed tube diameter and gas composition;
for more accurate values of h^^, standard correlations based on Prandtl and
Reynolds Number should be used, as given in Reference 18-2.
8) Assume fouling factor (see Figure 18-5). Suggested allowances are
given in Table 18-2 for different services.
9) Determine wall conductance (Equation 18-6).
S-1M21 Figure 18-11 follows
-------
Tubing: BSG No. 14 Gauge
1" O.D.
0.4 -
0.2 -
1.25 1.375 1.5
1 Spacing, in.
1.625
Flu?
Fume
_ Free Flow Area (Thru or Across Tubes)
Frontal Area (Approaching Exchanger)
Figure 18-11. FREE FLOW AREA FOR FLOW ACROSS A TUBE BUNDLE
S-14121
67784
-------
269
10) Assume a design velocity, (VmaxJoutside* for flow across the tubes.
In order to make the pressure drop roughly equal in both through- and cross-flov/
passes, it is suggested that the velocity be taken:
max inside
where C = 1.1 for in-line tubes
= 0-9 for staggered tubes
1(1 )
.1 avg outside
The term l/Tr)~.~allows for the different gas densities in the parses:
V avg inside f '
it is based on the absolute temperature, °Rankine ("Fahrenheit + 1*60). For a
cross-flow exchanger having an effectiveness of k^*f> and a flue gas temperature of
l400°F, the term has a value of about l.U.
11) Determine the outside heat transfer film coefficient for flow across
the tubes using Figure 18-3, pavg from step 3, and (Vmax)outside from steP 10'
As for step 7, this value of hfo is approximate, and more accurate values can be
found by referring to the generalized correlations of Kays and London (Ref. l8-l),
McAdams (Ref. 18-2), or Bell Ref. 18-O. In particular, with some tube
arrangements, the value of hfo will vary with the number of tube rows deep (n in
Figure l8-ll).
12) Determine overall heat transfer coefficient (Equation 18-7).
13) Find average specific heat (Figure 18-1, taken at average fume
temperature through exchanger).
14) Determine mass flow rate of fume through tubes (Ib/hr).
15) Determine mass flow rate of flue gas = fume + fuel + combustion air
(if any is added), in Ib/hr.
16) Determine N-^ (Figure 18-6) using selected effectiveness from
step 1 above.
17) Find R - C- /Cr, (Equation l8-ll). Flow through tubes is
- P,
1\LU16 X .1.116
"unmixed"; flow across tubes is taken as "mixed".
18) Calculate the required inside tube area (Equat* .ui 18-9 and Steps
16, 12, and 13).
19) Calculate the number of tubes required, using tube inside diameter,
Vmax from step 5, mass flow rate, and average gas density (Step 3)«
(ljA/6Q)w _ , ,
" *
S-1^121
-------
270
20) Calculate the required length of tubes, L, using Equation 18-10.
Adjust if necessary to use standard tube length.
2l) Determine the required area for cross flow over the tube bank to
carry the required flue gas, using (Vmax)ou-tside from steP 10> Pavg from Step 3,
and the mass flowrate of the flue gas. This is the net free flow area between
the tubes.
w
net cross flow 60 Pavg^rnaxToutside
22) Calculate the number of tube rows across the exchanger (perpendicular
to the flow direction), m, using Figure 18-10 (which defines the free space width
between the tubes, p), the area from Step 21, the length L from Step 20.
m = •k cross flow _ ^
P L
23) Determine the number of tubes deep, n, in the direction of flow,
using the total number of tubes from Step 19 •
m
The number of tubes, N, their length, L, and the number of tube rows across and
thru, n and m, should be adjusted to achieve integer values for m and n.
2U) The corresponding header frontal areas, A header, of the exchanger
may be found using the values of cr from Step 5- The proportions of the headers
are fixed by the number of tube rows across, m, and deep, n, and the tube length,
L.
18.3 Procedure for Checking Pressure Drop
The steps above have sized the exchanger — tube diameter, tube spacing,
number of tube rows deep and wide, and tube length. It remains, now, to determine
pressure drops. This is done as follows:
l) Determine gas density at entrance and exit conditions in the
headers (Equation 18-18).
2) Determine approach and leaving velocities Vavg = Q/Aheader*
3) Determine tube entrance loss and exit gain (Equation 18-12, 18-13).
4) Using average velocities and densities through both passes,
determine friction pressure drop;
a) Through tubes, (Equation 18-15).
b) Across cubes, (Equations 18-16, 18-1?).
Note that the values for f given in sections l8.3«3 and 18.3.^ are approximate,
and reference iQ-k should be used for more accurate calculations.
S-14121
-------
271
5) Sum the pressure drops (with due regard to pressure rise at tube
exit). Pressure drop through the two passes should be roughly equal. If not,
re-adjustment of the number of tube rows—width and depth—may be needed, keeping
the total number of tubes and area constant.
6) The total pressure drop should be compatible with the combined
system and the blower supply pressure. If not satisfactory, tht. design should be
repeated with higher or lower velocity assumptions (Setps 6 and 10 in the heat
transfer design).
18.6 Evaluating the Performance of an Existing Exchanger
This is a much more direct problem than that of sizing the exchanger,
which involves successive trial calculations to arrive at an acceptable desi;rn.
It is assumed that the important design and operating variables are fixed:
l) The dimensions of the exchanger and its ducting,
2) The entering fume stream and flue gas temperatures,
3) The corresponding flow rates and compositions.
One may proceed to calculate the effectiveness, the exit gas temperatures, and
the pressure drops. Steps are not outlined in detail; the first step is to
assume approximate exit gas temperatures and determine average gas densities
and specific heats (subsequent recalculation will not be necessary in many cases,
since the changes in density and specific heat are not large). Then average
velocities and heat transfer coefficients can be found. The number of transfer
units and the capacity rate ratio may then be calculated, and the effectiveness
determined. The effectiveness allows recheck of the assumed temperatures.
Calculation of pressure drops is straightforward.
S-14121
-------
273
Chapter 19. ANNOTATED LIST OF EQUIPMENT SUPPLIERS
The following listing of afterburner manufacturers and associated equip-
ment suppliers is only partly complete. Annotations cannot give a complete
resume or a fair presentation of a given manufacturer's capabilities. The field
is a rapidly changing one, and equipment supplied, and the manufacturers engaged
in business change continually. However, many of the companies listed have been
designing and manufacturing afterburners and other incineration equipment for
periods of 5 - 20 years or more. The listing is included primarily to aid a
potential afterburner user in starting to look for the equipment he *ants.
The authors of this report were unable to visit and discuss details
with all of the equipment suppliers. Also, some were more willing than oLhers
to discuss details of design and design procedures. As a result more space his
been spent describing the products of some suppliers then of others. This should
not be considered as a reflection on the relative merits of one supplier over
another, or of their respective products.
Afterburners - Thermal
AER Corporation, Ramsey, New Jersey, 07^6. This company makes custom
designs based on the Maxon "Combustifume" burner and box construction lined with
fibre block insulation. They will supply tubular crossflow heat exchangers.
Applications have been mostly in paper and fabric industries.
Air Preheater Company, Wellsville, New York, 1U895- They manufacture
pre-engineered units in the 2000 to 30,000 scfm range as well as custom designed
units. These are mainly straight-through horizontal afterburners, light in weight
(insulated with soft block and fiberfrax); they are available fired either with
gas (using the Maxon "Combustifume" burner) or with oil (No. 2, No. U, or No. 6
fuel oil, using a proprietary Air Preheater mixing region together with a North
American dual-fuel burner). Heat exchange units are emphasized, using the pro-
prietary "Cor-Pak" exchanger (corrugated tube surfaces), with recuperative heat
recovery of U5 to 5058 and 65 to 10% available as secondary recovery using a
rotary regenerative exchanger. Air Preheater has carried out applications j-
odor elimination in rendering plants, fume and odor elimination in rubber juring,
fume and odor problems in fiber glass impregnation, particle board curing, and a
number of more conventional solvent evaporation processes. They maintain test
facilities including a 2000 scfm prototype at their plant, and have available
a 100 scfm portable unit for on-site testing.
Alkar Engineering Corporation. 105 Spring St., Lodi, Wisconsin 53555-
Their basic design is a horizontal venturi-shaped chamber, ax-'^ily fired with a
Maxon "Ovenpack" burner, and with radial fume entrance at the burner end of the
chamber. The chamber is lined with medium, low density castable refractory. They
offer a recuperative heat exchanger in connection with this design, but only one has
been installed. They have extensive experience with the special application prob-
lems of afterburners for smokehouses (e.g., fouling of fans and ducts with fume
condensates), and a large fraction of their business is supplying smokehouse equip-
ment to meat processors.
Preceding page blank
S-Ik 121
-------
21k
Bayco Industries of California, UTO Hester St., San Leandro, California.
Eayco builds small package units and custom designed units of larger size. Their
package units are rectangular boxes with U-bend and stack, lined with castable re-
fractory, and a small straight stack unit. They use a variety of burners. Small
units are used for smokehouses, electrical insulation burnoff, sewage odor control.
Custom units are used for burning oil vapor off filter recovery units, and for
metallurgical furnaces. They have no experience with heat recovery.
Bigelow-Liptak Corp., Northwestern Highway and 10-1/2 Mile Road, Sc ith-
field (Detroit), MichiganU8075• Their equipment, custom-designed for each
application, characteristically involves heavy-duty, brick-lined incineration
units. They are aiming primarily for the large-scale and special conditions
applications rather than pre-engineered package units. They have designed units
as large as 150 M scfm, utilizing waste heat boilers (rather than recuperative
heat exchangers) for heat recovery. They have their own design for distributed
raw gas burners. Bigelow-Liptak have experience supplying equipment for incin-
eration of liquid, solid, and gaseous wastes. This experience includes such
difficult materials as streams containing halogens, other inorganics, and partic-
ulates. For such applications, they have designed systems including pre- and
post-treatment of the stream as well as incineration.
Blaw-Knox Company, 1575 Fillmore Ave., Buffalo, New York 1U211. Blaw-
Knox offer an afterburner designed and manufactured in conjunction with Thermal
Research and Engineering for moderately-large, continuous coffee roasting equip-
ment , and one of their own design for use on smaller, batch equipment. Neither
unit involves heat exchange, both are relatively short stack incinerators. Their
experience is in manufacturing afterburners in connection with food processing
machinery, primarily for coffee roasters.
Clean-Air-Rator Co., 19^0 Linwood Blvd., Oklahoma City, Oklahoma 73106.
This company builds stack-type units with metal combustion chamber lining, bottom-
fired, with package burners using gas fuel. They have been used in food process-
ing and refuse incineration applications.
Coats Warner Co., 15130 Ventura Blvd., Sherman Oaks, California. They
build gas-fired afterburners, custom designed, primarily in conjunction with their
paint handling equipment, 750 to 1500 scfm, to LACAPCD standards. Recycle flue
gas to ovens has been used for heat recovery.
Combustion Equipment Associates, Todd Division, 120 Park Ave., New York
City 10017. Their thermal afterburner units are custom designed for relatively
large applications, and include combination gas/oil fired burners. Their tech-
nical expertise is focussed on design of an effective, light-weight combustion
chamber. Heat recovery is not stressed and standard equipment of other manufactur-
ers is used for this purpose. They have limited experience with fume incineration,
but broad experience in combustion and thermal incineration of liquid and solid
wastes.
Despatch Oven Company, P.O. Box 1320, Minneapolis, Minnesota 55hkQ.
Their basic design is a horizontal-folded chamber with a Maxon "Combustifurn?"
burner (gas firing) at the inlet. Units are insulated with low density castable
refractory and hence are relatively light weight - they are also relatively low
cost. Despatch maintains laboratory and proto-type test facilities at their plant,
S-14121
-------
275
including a 3000 scfm unit, to develop designs and investigate application problems.
Their business is primarily the manufacture of industrial ovens and they have entered
the fume incineration business in order to treat effluents from their ovens. They
offer a set of nine sizes of a standarized design with or without heat exchanger
ranging in capacity from UOO to lU ,000 scfm. Over 90% of their afterburners are
sold in conjunction with new oven installations, mostly for paint spraying or baking
applications.
FECO, 5855 Grant Ave. , Cleveland, Ohio l*Ul05. Afterburners available
from FECO include conventional (brick or castable lined) thermal afterburners,
with or without tubular cross-flow heat exchangers. Units are custom designed
with either rectangular or circular cross sections with brick checkerwork baffles
to promote mixing, and are either gas or oil fired. They also make econonc'. light-
weight thermal units , using silica fiber insulation , and supplied with or wi ihout
rotary wheel regenerative heat exchangers. These units are of rectangular cross
section with a folded (U-bend) flow path. They are designed for 0.5 to 0.9 second
residence time and flow velocities of 25 to 30 ft/sec at 1500°F. They are designed
for natural gas firing only, using a distributed raw gas burner. FECO also makes
catalytic units (see section below). FECO's principal business is the manufacture
of ovens for metal decorating and foundry core baking. They have added afterburners
to their product line primarily in order to supply complete installations to
purchasers of their ovens .
Frimberger Corp.. 15301 12 Mile Rd, Rosevillem Michigan U8066 (formerly
Mahon Industrial Div. ) . This company builds fume incinerators, dust collectors,
mist eliminators, and precipitators.
Gas Processors. UoU No. Berry St., Brea, California 92621. Standard
units, 250 - 10,000 scfm are supplied with gas burner and have a horizontal cylin-
drical holding chamber, lined with castable refractory. They will supply units
with steam boiler for waste heat recovery. Their experience has been chiefly
on paint baking ovens and spray systems, and on odor elimination.
Gran co Equipment . Inc . . 1958 Wilson Ave. S.W. , Grand Rapids, Michigan
U950U. This company uses a metal U-tube combustion chamber; incoming fume st-eam
surrounds the U-tube and exchanges heat; metal fins increase heat transfer Area.
Design uses "Flame Grid" burner. Their first installation is undergoing test.
Hayes Albion. 8550 W. Michigan Ave., Parma, Michigan U9269. This design
is a vertical metal stack unit with distributed gas burner using outside air and
a second manifold for introducing fume. It has been applied to pulp mill digester
vapors for odor control.
Hirt Combustion Engineers. 931 So. Maple Ave., Montebello, California
9061+0. Hirt provides both standarized and custom designs. The combustion chamber
uses a castable refractory lining inside a cylindrical metal shell. Blowers operate
on exit gases to minimize the possibility of fouling; air dilution is used at the
blower entry to reduce the temperature. Hirt's gas burner uses a multiple array
of ceramic cup-shaped burner heads, each with its own gas Jet, and operated at
near-stoichiometric conditions. It gives a turndown of about 10:1, and groups are
sectionalized if more turndown is required. Provided it is not a fouling stream,
or deficient in oxygen, the fume air is used in part for combustion; the remainder
is introduced after the burner assembly for dilution and cooling to the required
holding section temperature. A throat ring and ring-and-donut baffles are used
S-
-------
276
to insure good mixing in the holding section. Hirt's oil burner (their own design)
is steam or air atomized. Because of its lower turndown, several burners may be
used. Hirt use standard designs, but give an engineering review of the process.
They have supplied thermal afterburners to a wide variety of applications, many of
which have included heat recovery, using either tubular exchangers or rotary
recuperators.
Koch Engineering Co. . Inc. . Ul E. tend St., New York, New York 10017-
Koch supply custom-designed afterburners, based on commerically availabl • burn-
ers, using refractory-lined combustion chambers.
T. Melsheimer Co., Inc., 22610 So. Western Ave., Torrance, California
90501. This company has thermal incinerators (8000 and 15,000 scfm) for sewage
plant odor control. Vertical fire-tube heater units are used for exchange-pre-
heating the fume stream.
Michigan Oven Co., P.O. Box 336, Romulus, Michigan U817U. Afterburners
are built for use in conjunction with charring ovens (wire insulation burnoff).
Mid-South Manufacturing Co., 1221 Bank of the Southwest Bldg., Houston,
Texas 77002. They use a small concentric-tube metal chamber with internal baf-
fling, 150 - 1300 scfm, temperature 1100°F; stack gas exchanges heat with fume
stream. Units have been used for sewage lift station odor control.
B. Offen and Company, 29 East Madison St., Chicago, Illinois 60602.
They build custom units using the gas-fired Maxon "Combustifume" burner in rectan-
gular or cylindrical chambers (depending on size and where it is to be located).
They have used tubular exchangers for heat recovery. Offen's experience have been
in supplying equipment primarily to the graphic arts industry for use with print-
ing presses.
Pacific Coast Incinerators, 710 Delaware St., Berkeley, California. They
use a vertical cylinder with castable refractory lining, and a package gas burner.
Units have been supplied for restaurants, paint-bake ovens.
Proctor and Schwartz. Incorporated, Pollution Control Equipment Division,
7th St. and Tabor Rd. , Philadelphia, Pennsylvania 19120. They have developed a
novel regenerative heat exchange technique coupled with thermal fume incineration,
which is the basis of their "TRAPS" system. They have one such unit in operation,
and have designed the system for simple operation and modular construction. The
first unit of this nev design is now being installed. The system allows a high
degree of heat economy, and has been developed through use of a pilot scale unit.
This novel approach does require a cyclically operated valve arrangement, and uses
ceramic structural members operating at high temperatures.
Ross Engineering Division, Midland-Ross, P.O.Box 1^7, New Brunswick,
New Jersey 08903. Ross Engineering design and supply medium to large thermal
afterburners with emphasis on heat recovery through both preheating of the con-
taminated air stream and heating a secondary fluid system. They have experience
with normal solvent systems and maintain a test facility at New Brunswic!. Their
normal design is a downfired rectangular unit using a Maxon "Combustifume" burner
fired under a bridge wall and through a heat exchanger. They have also designed
end-fired cylindrical units using both gas and oil fired burners. Their units are
designed for low overall pressure drop, and use a metal structure with insulating
masonry lining. Applications have been for resin and paint curing and drying ovens.
S-11H21
-------
277
J. B. Seage Cc., 2U19 E. 56th St., Los Angeles, California 90058.
Custom-designed units use a vertical stack, are refractory lined; fume enters at
the bottom, gas burner (package unit) enters at side. Units are in service in
curing ovens, crematories, metallurgical furnaces, burnoff ovens.
Services and Equipment Co., khOI S. Western Blvd., Chicago, Illinois
60609. They build small, inexpensive stack-type units, with side ?ntry of gas
flame, short residence time. Have been used for coffee roasters, restaurant boiler
vents, wire burnoff ovens.
Smith Engineering Co., 1903 Doreen Ave., South El Monte, Calitcmia
91733. Their designs include horizontal, vertical, induced and f orced-dj. af t
types; units are custom designed. Induced draft designs are preferred. ThPir
designs are generally brick lined, use a distributed raw gas burner of Smith
Engineering design, and stainless steel construction in burner and heat exchangcu
(when used). Larger units use masonry construction, smaller units use castable
refractory. Experience includes paint baking ovens, coffee roasters, air
sterilizers, resin curing ovens.
Spencer Boiler and Engineering Company, 12106 So. Center St., Southgate,
California 90280.Units are of standard design, but are supplied as custom appli-
cations. They use a multijet gas burner of their own design and a cylindrical
holding chamber, lined with castable refractory. They have used waste heat boilers,
exchangers, water heaters, and recycle gas streams for heat recovery, and have
applied units to smokehouses, paint and solvent areas, and food dehydrating and
roasting.
Surface Combustion Division, Midland-Ross Corporation, P.O. Box 907>
Toledo, Ohiok3601.Their afterburners are standardized, pre-engineered units
which cover low and middle-range capacities (100 to 10,000 scfm), and among the
four types cover rich fume streams as well as more common lean streams. All units
are gas-fired. Primary emphasis is on adequate design of the combustion chamber
and low pressure drop through the units; no standard heat exchanger design is
offered. Larger units, including heat exchange would be designed on a custom
basis. The test facilities used in the research program, including a 3000 scfm
prototype, are maintained to study unusual applications problems. They have oarried
out an extensive research program in developing designs for the four types of after-
burners which they offer. Published results of this program represent an important
contribution to defining the fundamental requirements for thermal afterburner design.
Tailor and Company, Inc., Bettendorf, Iowa. This company have built
chemical incinerators, but have no experience in fume afterburners. They use a
unique design of cyclone burner.
Thermal Research and Engineering. Conshohocken, Pennsylvania 191*28.
Their units are custom designed, tend to be for medium to large applications;
many feature a very high degree of recuperative heat exchange, and the use of a
combination fuel oil/natural gas fired burner. Several different burner/combustion
chamber configurations are used, depending upon the application. Burners are of
Thermal Research design. They have a broad range of experience, and maintain a
flexible test facility used in developing designs for new applications. Applica-
tions have included chemical and organic wastes, smelter off-gases, sulfur plant
tail gases, chlorinated wastes, tars. They have tended to concentrate on unusual
wastes where experience and engineering skills count for more than equipment cost.
S-1U121
-------
2?8
UOP Air Correction Division, Tokoneke Rd. , Darien, Connecticut 06820.
UOP build a range of pollution control equipment (e.g., vent scrubbers, cyclones,
precipitators, adsorption and absorption equipment) as well as both thermal and
catalytic afterburners. Both thermal and catalytic unit designs have employed
lightweight, all-metal construction, using stainless steels without refractory
lining. Current designs use refractory board linings. They have pre-engineered
units in sizes to 6,500 scfm. Larger units may incorporate heat exchangers. They
normally use a distributed raw gas burner with a perforated conical baffle for fume
inlet, of UOP design. UOP experiences have included industrial chemicals and oil
refining applications, paint and varnish cookers, sewage vent treating.
John Zink Company, hkOl So. Peoria St., Tulsa, Oklahoma 7^105. Zink Co.
design and build custom engineered installations, called "Thermal Oxidizers". They
use varied arrangements depending upon the job. One arrangement is a squat verti-
cal cylinder, joined by a conical section to a vertical stack. Opposed burners
fire radially into the cylindrical part, and are without internal baffles. On
waste streams not previously encountered, they run a pilot test. Halogenated
streams are sometimes given as much as 3 seconds residence time. Zink makes their
own burners, which can be supplied for gas or liquid fuels, or both, and follow
practices on process heaters. They commonly use heavy masonry construction.
Applications have largely been in industrial chemicals and refining plants. Units
tend to be of large size, may be used as fume oxidizers or as gaseous or liquid-
fuel incinerators; they may be part of a system that includes scrubbers, knockouts,
and precipitators when waste streams are halogenated (or otherwise corrosive or
objectionable). They have not had experience with installations using heat
exchangers.
Vari-Systems, Environmental Systems Division, 1295 W. 78th St.,
Cleveland, Ohio kkW2. Standard models, oil or gas fired, use fume or outside
air for combustion. Refractory-lined horizontal cylindrical chamber are end fired
or have double-wall construction. Custom units will be designed for large sizes
or special duties.
5-1*4-121
-------
279
Afterburners - Catalytic
Englehard Minerals and Chemicals, Gas Equipment Section, 205 Grant Ave.,
East Newark, New Jersey 07029- They will provide catalytic afterburner equipment
for oxidation of waste fume streams. They have considerable parallel experience in
catalytic units for reduction of nitrogen oxides, but limited experience with oxida-
tion units. Their basic design is a downfired preheat burner, wiLh heated gases
flowing under a bridge wall and up through the catalyst matrix. Englehard uses
noble metal catalysts supported on a honeycomb ceramic grid "Thermocomb" supplied
by American Lava.
E. I. du Pont de Nemours and Co., Industrial and Biochemicals Department,
Wilmington, Delaware19898. L»u Pont supplies the honeycomb ceramic matrix material,
"Torvex", to other catalytic afterburner manufacturers, and supplies noble rrfcral
catalysts based on this material both to users and manufacturers of catalytic c.' ier-
burners. Their development group has done extensive work on catalyst development.
and characterization, and is available as consultants in design and application of
afterburners.
FECO, 5855 Grant Ave., Cleveland, Ohio MH05. Catalytic units are custom
designed, using ceramic honeycomb-based catalysts supplied by du Pont or Matthey-
Bishop. Normal designs use a rugged, double-wall construction with stainless steel
inner wall and heavy gauge carbon steel outer wall with ceramic fiber block or wool
between for the enclosures and distributed raw-gas burner for preheat. As v/ith
their thermal units, most applications have been in supplying air pollution control
systems to FECO's oven customers.
Matthey-Bishop, Malvern, Pennsylvania 19355- Their catalysts are noble
metals on du Pont "Torvex" honeycomb supports. A modular approach to mounting the
catalyst element allows effective utilization in afterburner units with either
rectangular or circular cross-sections. They also have design concepts and capabil-
ity for design and construction of thermal afterburners, but have not yet supplied
any. They have supplied catalytic units for reduction of NOX in nitric acid plant
tail gas. They have designs for catalytic systems for hydrocarbon and odor abate-
ment, and have supplied replacement catalysts for existing catalytic (oxidatio^)
afterburners. To date, their emphasis has been on relatively large installations
in the wire-enamelling, metal decorating, printing and food processing industries,
and for odor problems.
Oxycatalyst, Inc., E. Biddle St., West Chester, Pennsylvania 19380.
Both catalytic and thermal afterburning equipment for mobile and stationary sources.
Their early units were based on using "Oxycat" (brick) catalyst elements. These
were arrays of ceramic rods with noble metal deposited on their -arfaces, and fumes
were passed in cross-flow over these rods in the array. "Oxyr.d.ts" are still manu-
factured for replacement and a few new applications. Newer Oxycatalyst units use
a noble metal catalyst on a du Pont "Torvex" honeycomb base. They have pre-
engineered afterburner designs for flow rates between 2.5 and 12 M scfm as well as
custom designed units for larger capacities. These units generally use distributed
raw-gas burner to supply preheat and have a castable-refractory-lined enclosure of
rectangular cross-section for the preheat burner, mixing and catalyst sectiors.
They include heat exchangers when required. Their design configurations uce single
unit construction and are shipped as pre-packaged units. They will furnish units
intended for use as catalytic afterburners, but which have the capacity (sufficient
-------
280
residence time, velocity and temperature resistance) of being converted to thermal
afterburners. Such convertible units have a higher initial cost than straight
catalytic units, but allow for the possibility of catalytic operation being inade-
quate either because of unforeseen problems in the application or because of changes
in the nature of the user's fume stream composition or conversion requirements.
Oxycatalyst has experience in a wide range of applications of catalytic afterburners.
UOP, Air Correction Division, Tokoneke Rd., Darien, Connecticut 06820.
The catalyst used in catalytic units is the basic Suter-Ruff type with spongy noble
metal deposited on nichrome ribbon, which is crimped and assembled in the unit as a
mat in a module similar to a furnace air filter. Catalytic units are now generally
of an inline arrangement of preheat burner and catalyst element rather than the
folded configuration which has been much described in the literature. Enclosures
are of light-gauge, double metal wall construction, with aluminized steel normally
used as the inner liner. UOP's distributed conical burner is normally used for
preheat. UOP has extensive experience in a number of afterburner applications, and
has available rate constants for both thermal and catalytic oxidation of a number
of typical fume constituents. These are used in proprietary computer programs which
calculate temperature and time requirements for designs.
-------
281
Burner Manufacturers
All of the major burner manufacturers make gas and oil burners which can
be used in afterburners currently on the market. The four manufacturers in
this list have had extensive experience with fume incineration and «n have
developed distributed burners.
Eclipse Fuel Engineering Company, 1100 Buchanan St., Rockford, Illinois
61101. They supply gas and distillate oil burner packages. They have a distributed
source "line type" gas burner with provision for using the fume for th-» burner
oxygen supply if the fume is non-fouling.
Hydro Combustion Corp. , 22805 So. Avalon Blvd., Carson, California
9071*1*. Although they have supplied thermal fume incinerators, their principal
specialty is burners—they make a distributed gas burner end a heavy oil burnei .
Maxom Corporation. Muncie, Indiana U7302. Maxon manufactures gas
burners and auxiliary combustion equipment (combustion air blowers, controls, etc.).
Their raw gas, "Combustifume" burner is widely used in fume afterburners. The
"Ovenpak" premix gas burner is a package unit with blower and controls. The
"wide range" type provides for a turndown of kO to 1 on firing rate, using a pro-
portioning air/fuel controller. They have carried out extensive research and
development work on the time-temperature requirements for afterburning fume
streams containing a variety of contaminants. They maintain a development test
facility at Muncie and are continuing work on specialized application problems.
North American Manufacturing Company, UU55 E. 71st St., Cleveland,
Ohio UU105. They are manufacturers of burners and auxiliary combustion equip-
ment for both gas and oil firing. Their "dual-fuel" burner line is supplied
where oil firing or optional oil/gas firing is required. They make a distributed
raw gas burner, termed "Flame Grid", developed specifically for fume afterburner
application, which can make use of the fume as an oxygen source for combustion.
They supply a "high excess air" burner which will tolerate several hundred percent
of stoichiometric air (using internal bypassing).
S-14121
-------
283
Chapter 20. REVIEW OF TASKS
The purpose of tnis section is to briefly describe the methods used in
carrying out the afterburner systems study. The study covered afterburner or
fume incinerator technology for control of gaseous (or gasborne -oarticulate)
combustible emissions from stationary sources. Disposal of liquiae or solids by
combustion was not included. Two main classes of afterburner were considered,
thermal (direct flame) and catalytic. Flares and afterburners which were part of
a solids refuse disposal system were not included. The objectives of ^he system
study were to l) evaluate present status of technology, 2) evaluate existing efter-
burner systems, 3) determine current practices and problems in applications,,
k) identify potential future applications, and 5) develop R and D plans to c/1vance
technology. Information was obtained from the published literature, equipment
manufacturers, equipment users, air pollution control agencies, research instil.^t?s
and universities. Visits v/ere made to many of these sources of information in
order to hold detailed technical discussions and observe equipment in operation.
Questionnaires were sent to plants which were known to have afterburners in use or
were likely to have them. The respondents provided information on general
performance, costs and problems encountered with their units. Detailed performance
data were rarely available.
20.1 Evaluation of Current Engineering Technology
The current status of theoretical and applied technology applicable to
catalytic and flame afterburners was reviewed. The first source of information
examined was the published literature. While the combustion literature is very
large, only a small amount could be found on afterburners. The researchers at
universities and research institutes have been concentrating on combustion phenomena
of fuels rather than waste destruction. The theoretical principles are, however,
the same and this literature is utilized in this report. Even for fuel combustion,
the theoretical aspects are not advanced far enough to be able to use them in
design calculations for equipment. The bulk of the published literature has been
written by the equipment manufacturers and is often only descriptive material
and occasionally contains empirical calculation procedures. There is a little
literature written by the users of afterburners but again this is quite general
and empirical in nature. Part of the difficulty in developing a technological base
is the lack of good test data on performance. Many early applications of after-
burners were for odor and smoke abatement where no quantitative performance is
defined. The second source of information was the equipment manufacturers, a number
of whom were visited by members of our staff. A visit typically was arranged after
a letter contact outlining the objectives of our study and listing the information
we were looking for. The actual visit was made by two staff members and usually
lasted about one day. Reports of these discussions were then • ritten.subsequent
to the visit. These visits enabled us to obtain a clear picture of the technology
in use for equipment design. While principles were always freely discussed, in
some cases specific parameter values used in design were considered proprietary and
were not revealed to us. By and large, designs have evolved in an empirical way
and mistakes are still made through lack of understanding. Nevertheless, we were
able to learn the principles of the approach taken by the manufacturers. The
proprietary material is available to the potential user since it will be brought
to bear on his problem when he approaches the manufacturer for equipment selection
and purchase. The technology is reported in various parts of this handbook.
We have brought together the theoretical background in fuel
combustion with the empirical methods used in afterburner design. This provides
insight for many of the problems encountered and points the way towards improvements,
S- Ik-121
Preceding page blank
-------
20.2 Evaluation of Existing Afterburner Systems
A review of the features of existing afterburner systems is reported.
The main sources of information here were the equipment manufacturers. Their
brochures and published literature contains this material in a readily available
form. This information was supplemented by our staff visits to the manufacturers.
Additional information on specific features of existing systems was obtained from
users (the user survey is described in the next section). The advantages and dis-
advantages of the various features are presented above.
Some guideline information on operating and capital costs was obtained and have
been summarized. This economic data can be used by a prospective user for orienta-
tion, preliminary selection of equipment and an approximate capital and operating
cost. It is envisioned that more accurate costs would be developed in the specific
dealings a user would have with an equipment manufacturer. A listing by manufac-
turer of equipment types available is contained in Chapter 19•
20.3 Assess Present Practices and Problems
20.3.1 Survey of Users
A survey of users of afterburners was made by sending out questionnaires
to plants which were known to have afterburners, or it was thought likely that
they would have them. Follow-up visits were made in some cases, and telephone
calls in others, to obtain further information.
Response to the questionnaires was disappointing. Out of 559 question-
naires sent out, completed forms were received from 103 plant locations covering
238 afterburners. Although some companies responded promptly, in many cases
it was necessary to make repeated telephone calls and send out duplicate
questionnaires before replies were received. In a number of cases, information
promised some time ago has not been received.
Data from the replies have been tabulated in summary form. A
computer format was used in order to facilitate manipulation and machine
plotting of the information.
20.3.2 Contacts with APCD's
Local air pollution control districts were contacted as part of our
survey to:
l) Identify present afterburner users, especially those encountering
difficulty with their units.
2) Obtain test data on the effectiveness of afterburner installations.
3) Obtain pollutant inventory data and learn of present and projected
regulations governing hydrocarbon and odorous emissions.
Regional offices of APCO were contacted to determine the most Active
local districts which we then contacted, mostly through telephone calls. In most
areas of the country outside of California, there has been little concern with
pollutants other than smoke and S02• Therefore, most of the local APCD's had
little to contribute to our study.
S-14121
-------
285
Where the air pollution control district indicated that they had
afterburners in operation, indicated that they had experienced some problems,
and where we were planning a trip to the area to visit an afterburner manufacturer,
we conducted discussions with district personnel. Especially useful discussions
were held with districts in the San Francisco Bay Area, St. Louis, and New Jersey.
Eight other local and state agencies were visited. The Los Angeles County
APCD, which has the most extensive test data and experience with afterburners
discussed philosophy of afterburner design,testing, and regulations but names of
users and test data were considered proprietary. This concern over release of
user names was widespread. Test data is almost non-existent, but we were supplied
much of that obtained by the Bay Area APCD after user names had been removed.
The St. Louis district took us to visit several afterburner installations, but
felt that all local user contact should be handled through them.
20.3.3 Questionnaire Format
The objective of the questionnaire was to obtain information
on the various afterburners being used in industry so that performance and cost
could be evaluated. This presented a problem as to what questions should be
included and how long the questionnaire should be.
A draft questionnaire was submitted for approval in early October,
1970. This was later amended, and the final version was given Bureau of Budget
approval on December 29, 1970. A copy of the final version, together with the
cover letter is attached. In final form the questions cover six single side
pages, and the length may have contributed to the poor response.
20.3.1* Selection of Users
Names of users were first of all obtained from lists supplied by
afterburner manufacturers and by Air Pollution Control Districts. Most of
these addresses were incomplete (and sometimes wrong), so a great deal of
checking had to be done using our standard library reference sources and
telephone directories. A number of afterburner operations have been reported
in the literature, and these were added to our list as they were located. In
addition, a number of questionnaires were sent to companies where we did not
have specific information that they had afterburners in operation, but their
plants were such that we would expect them to have such units (i.e., oil
companies, chemical companies, coffee suppliers, etc., especially located in
California).
A total of 559 addresses were obtained in this way, and most of them
were known to have had afterburners in operation at one time.
S-1M21
-------
20.3.5 Replies
Replies vere received from 156 plant locations, of which 53 replied
that they did not have afterburners. The remaining 105 had a total of 238
'ifterburners installed. Many of these replies were only obtained after
repeated solicitations both by letter and by telephone calls.
The quality of the replies varied considerably depending on the
sophistication of the plant engineering staff. In most cases test data had
not been taken, and in others the companies were reluctant to divulge it.
Purchase and installed costs were available in a large percentage of the
cases. Many plants do not keep records of fuel costs for afterburners
separately, so in these cases, operating costs were not available.
20.3.6 Treatment of Information
The information obtained from the questionnaires was tabulated in
computer format (see Chapter 11, Table 11-2). This allowed internal manipulation
of the data including sorting by category, adjusting purchase and installed costs
to 1970 values, adjusting operating costs to a 2h hour basis, computation of
average residence time and machine plotting selected data.
Process categories were selected using the Standard Industrial
Classification Manual. The classification used was by the immediate process
to which the afterburner was fitted, rather than the overall description of
the manufacturing facility. For example, painting of automobiles and of
metal cans were both included under 3^79 (Coating), rather than using 3712
(Motor Vehicle Bodies) and 3^11 (Metal Cans), respectively.
A listing of the process categories used is given in Table 20-1.
20.3.7 Follow-up Visits
Follow-up visits were made to users of afterburners who had specific
information and wide experience and had no objections to a visit by our staff.
Most companies were willing to cooperate but few had data or experience enough
to justify a visit trip. Telephone conversations were sufficient in most cases.
Visits were made to:
1. Automobile body painting operations in the Los Angeles area.
2. Can companies in Chicago and San Francisco with wide experience of both
thermal and catalytic units.
3. A large chemical company in Philadelphia with experience on phthalic
anhydride.
h. A coffee company which has recently replaced two old afterburners with a
Proctor and Schwartz Traps Unit.
5. A number of smaller companies in the Bay Area on widely differing industries
from asphalt blowing to drum manufacture.
-------
28?
Table 20-1. CLASSIFICATION OF USERS BY STANDARD INDUSTRIAL CLASSIFICATION MANUAL
8061 Hospitals
3900 Miscellaneous manufacturing industries
371** Motor parts and accessories
36Ul Electric lamps
3621 Electric motors and generators (rebuilding)
3^79 Coating and engraving (enameling and painting)
33^1 Secondary smelting and refining non-ferrous metals
3323 Steel foundries (investment casting)
3296 Mineral wool (glass, wool, fiberglass)
3292 Asbestos products (brake shoes vinyl tiles)
3079 Miscellaneous plastics products (e.g., phenolic laminates, film)
3069 Fabricated rubber products (foam rubber)
3000 Rubber and miscellaneous plastic products
2911 Petroleum refining (including asphalt)
2900 Petroleum refining and related industries
2851 Paints, varnishes, lacquers, enamels and allied products
2831* Pharmaceuticals
2815 Cyclic intermediates (phthalic anhydride, toluene, xylene, etc.)
2800 Chemicals and allied products
2700 Printing, publishing and allied industries
2k26 Hardwood dimension and flooring mills
2095 Roasted coffee
2091* Animal and marine fats and oils (rendering)
2013 Sausages and other prepared meat products (smokehouses)
2000 Food and kindred products
^952 Sewage systems
S-14121
-------
20. U Determination of Major Sources and Potential Applications
Emission inventory information was found to be quite variable in scope,
availability and existence. Data is especially sparce relative to hydrocarbon
emissions since previous concern has been with particulates and S02- Some states
have developed complete information whereas others have none. The emission
estimates that we used were primarily based on emission factors and total production
of materials that normally find their way into the atmosphere (e.g. , solvents).
Projection of emissions in the future was then tied to the output growth of
particular industry or to the population growth. A tentative ranking of the
emissions was made by combining quantity, toxicity and nuisance factors.
20.5 Research Recommendations
The research recommendations are based on the deficiencies identified
in the review of technology. The deficiencies are discussed in the various
sections in this handbook describing the technology. The research and
development recommendations are presented in Chapter 22.
3-1^121 User's Survey Follows
-------
SHELL DEVELOPMENT COMPANY
A DIVISION OF SHELL OIL COMPANY
1400-53rd STREET
EMERYVILLE, CALIFORNIA 94608
__2/?/y "
April 8, 1971
User Survey - APCO Afterburner Systems Study
Contract BHS-D-71-3
We are currently engaged in an Air Pollution Control Office
(APCO) sponsored engineering study of afterburner systems for
control of gaseous or gas borne emissions from stationary sources.
APCO was formerly the National Air Pollution Administration and is a
part of the new Environmental Protection Agency. The objectives of this
study are to increase the effectiveness and utilization of flame and catalytic
afterburners. Upon completion of our work, we will publish an ''afterburner
handbook", to be available publicly, containing information on the
following subjects:
l) What emissions are suitable for treatment by afterburning.
Which components might cause difficulties.
2) What design features are desirable with regard to effective-
ness, trouble-free operation, and low capital and operating costs.
3) What types of units are available. Who are the active
manufacturers.
U) Unbiased data on initial costs and operating expenses
including maintenance.
This handbook will be most useful in guiding the selection of required
control equipment.
We believe that industrial users of afterburner equipment
comprise an important source of information for this study, particularly
concerning specific applications of afterburners as pollution abatement
devices. Of special interest are the ability of these systems to perform
satisfactorily in these applications; and costs associated with acquiring
-------
-2-
and operating afterburner systems. For this reason, we are surveying a
representative sample of afterburner-users.
We ask that you participate in this survey by supplying the
information requested on the enclosed questionnaire. Your participation
will be valuable even if you can only supply part of the information
requested. It is especially important to the success of the study that
you point out any present problem areas so that possible improvements
can be identified. The information you supply will be held confident!. 1.
Some of it may appear in tabular or statistical form in our reports and
handbook but without identifying your company.
We would appreciate your filling out a separate copy of pages
2-6 for each afterburner application or each type of unit. We can supply
additional copies if needed. A self-addressed, postage-paid, envelope
is enclosed for the return of the completed questionnaire. If you have
any questions, please call us collect at: (Ul5) 653-2100 and ask for
E. R. Slater, R. D. Hawthorn, R. W. Rolke, C. R. Garbett, or G. D. Towell.
Your cooperation in participating in this survey will be greatly
appreciated.
Sincerely yours,
G. D. Towell,
Project Manager
GDT:psb
Enclosures: Survey Questionnaire
Return Envelope
-------
OMB-85-S700J6
Expires 9/JO/71
NAPCA CONTRACT EHS-D-71-5
Survey of Users of Afterburner Equipment
Date:
Section I. COMPANY IDENTIFICATION
1. Name and location of company:
a. Name:
b. No., Street:
c. City:__ State: Zip Code:.
2. Location of plant if different from above:
a. Plant/Division:
b. No., Street:
c. City: State: Zip Code:.
3. Person to contact regarding information contained in this report:
a. Name:
b. Department/Division:
c. Telephone: (Area Code)
Principal products manufactured at this plant:.
5. How many afterburner systems do you have at this location? (if two or more,
please complete pages 2 to 6 for each system.)
6. Person completing questionnaire (if different from above):
a. Name:_
b. Department/Division:
c. Telephone: (Area Code)
-------
- 2 -
Section II. GEMERAL INFORMATION
1. a. Name of company completing this form:_
b. Afterburner identification (if more than one at location):
c. Date of installation:
d. Name of manufacturer of afterburner:
Did this same company design and install the system?_
If not, name of company(s) which did:
2. Name of process generating waste gas stream (varnish cooker, paint spray booth,
etc.):
3. Capacity of process (ibAir, items/day, etc.):
k. Is the afterburner operated principally to control (check one):
a. Odor nuisance
b. Emission of hydrocarbons (ib/hr or ppm level at afterburner inlet)
c. Toxic nuisance
d. Smoke or other particulate emission
e. Other (such as a condensing plume) - Please specify
5. a. Does the afterburner provide sufficient emission control?_
If not, please specify (residual odor, residual hydrocarbon emission, etc. )
b. What problems have you had with Iceeping the system operable?
Approximate $ of time afterburner is inoperable:
Section III. WASTE GAS STREAM FED TO AFTERBURNER
1. Description of uaste stream leaving process:
a. Average flou rate: _ SCFM
b. Average temperature: _ °F
c. Average composition
Oxygen: _ % Volume
Combustible , _ (ppm as hexane or other basis )
contaminant: ( _ lb/hr of contaminant
-------
- 3 -
d. Heating value of waste stream BTH/SCF
e. Are the flow rate, composition, and temperature highly variable? (process
operated batchwise, etc.)
If so, what are the ranges In flow rate? to scfm
composition? to ppm
temperature? to °T
f. Does the waste stream consist of anything other than gaseous hydrocarbons,
air, water vapor, and COa? (chlorinated hydrocarbons, inorganic dust,
liquid mist, etc.)
If yes, please specify_
g. How many hours per day is process generating waste stream?_
2. Is the waste stream pretreated (other than heat exchange) prior to after-
burning? (scrubbing, filtration, air dilution, etc.)
If yes, please specify_
3. Is a heat exchanger used to preheat the waste stream before afterburning?_
If yes, are the hot gases leaving the afterburner used as the heating
medium?
Section IV. THERMAL AFTERBURNER UNITS (if unit is catalytic type, skip to Sec Aon V)
1. Operating temperature
Preheater (if used) inlet °F outlet °F
Afterburner inlet °F outlet °F
2. Operating pressure (incv>,s of water or psig)
3. Fuel type (check one):
a. Natural gas
b. Fuel oil
c. Other (please specify)
k. Rate of fuel consumption
-------
5. Combustion chamber dimensions
Diameter Length
6. Is air supplied to the burner? (other than in waste stream)_
Section V. CATALYTIC AFTERBURNER UNITS (if unit is "thermal type, skip to Section VI)
1. Operating temperature
Heat exchanger preheater (if used) inlet °F outlet °F
Afterburner inlet °F
After direct fired preheater °F
Outlet of catalyst bed °F
2. Operating pressure (inches of water or psig)
3. Is a direct fired preheater used?
If yes, a. What is fuel type? (check one)
l) Natural gas
2) Fuel oil
3) Other (please specify)
b. Rate of fuel consumption .
c. Is air supplied to the burner? (other than in waste stream)_
k. Catalyst
a. What is catalyst configuration? (check one)
l) Mesh mats
2) Parallel cylinders
3) Pellets in baskets
k) Other (please specify)
b. What is active catalyst constituent? (copper oxide, platinum, etc.)
What is the nature of the catalyst carrier? (metal, ceramic, etc. )
-------
- 5 -
c. Typical life of catalyst "before replacement xtanths
d. Is periodic cleaning of catalyst unit employed?
If yes, l) How frequently? .
2) What method is used? (air blasting, water wash, etc.)
Section VI. TYPICAL PERFORMANCE DATA
a. Inlet pollutant concentration ppm
(If analyze for individual _^ _F'7m
compounds, please specify each!
b. Outlet pollutant concentration ppm
ppm
c. Do you have specific test data?
Would you mate it available to us for this study?
d. Do you monitor performance of the unit? How? (check one)
1) Continuously 2) Periodically 5) Occasionally
Section VII. GASES LEAVING AFTERBURNER
1. Is a heat exchanger used to recover heat from gases leaving afterburner?
2. Is part of the hot afterburner effluent recycled to the process?
3. Is posttreatment (other than heat exchange) employed? (scrubbing, filtration,
etc.)
If yes, please specify
Section VIII. OTHER SYSTEM COMPONENTS
1. What safety features are included in the design? (flame arrester, high or low
temperature shutdown, shut down if burner flame is lost, etc.)
-------
- 6 -
2. Is a separate fan used in the afterburner system? ___»_
If so, is it located (check one)
a. Before afterburner
b. Following afterburner
3. Are unusual materials of construction (other than carbon steel and normal
refractory) used?
If yes, please specify
Section IX. COSTS
1. Approximate installed cost of afterburner system $_
2. Purchase cost of afterburner unit itself $
3. Approximate operating cost of afterburner system $/year
^. Do you have a more detailed breakdov/n of operating and capital costs which you
would discuss with us?
Section X. POLLOWUP
Would you be willing to discuss in more detail system performance, cost
data, and design features through a telephone call or visit to your plant by one
of our representatives?
If you have any questions, please call one of the people listed in the
cover letter at (Ul5) 653-2100. An envelope has been enclosed for return of this
questionnaire to Shell Development Company, P. 0. Box 21*225, Oakland, California
9^23, ATTN: G. D. Towell. Thank you again for your cooperation.
RWRrpm
Enc.
-------
289
Cnapter 21. ilAJOft SOURCES AI1D POTENTIAL AFTERBURNER APPLICATIONS
21.1 Summary of Investigation Results
The rate at which organic pollutants are emitted in xhe United States
from both mobile and stationary sources has been estimated for 1^68 and 1970,
and projected through 1980, 1990, and 2000. The results are summarized in
Table 21.1. The contribution of stationary sources is projected to rise from 50$
of the total in 1970 to 85$ in 2000. The projections of emissions from stationary
sources were made with the assumption that the present levels of legislated
emissions control would not change during the projection period. The projected
mobile source emissions refleec the stringent automobile emission regulations
which are scheduled to take effect in 1975•
Several of the largest sources shown in Table 21.1 are either mobile
or are not amenable to afterburner control. Examples are motor vehicles, refuse
disposal, gasoline marketing, etc. Conversely, many sources which do not emit
nationally significant tonnages of combustible emissions are, non-the-less, prime
candidates for afterburner control because they constitute severe local nuisances.
Odor and smoke sources, e.g. rendering plants, coffee roasters, brake shoe
debonding ovens, etc., frequently fall into this category. No attempt was made
to estimate the total rates of emissions of effluents which are objectionable
primarily because of malodor. Instead, a count of the number of point odor
sources was used to give a measure of the severity of the odor pollution problem.
The- results are shown in Table 21.2.
Emissions from sources amenable to afterburning have been characterized
by the type of compound emitted. Each set of emissions has been assigned a
ranking based upon toxicity, malodor, and photoreactivity to allow an importance
comparison of the various pollutant sources. In addition, each source has been
assigned a ranking based upon the relative ease of control by afterburning. The
source characterization and rankings are given in Table 21.3.
Criteria to be considered when choosing between afterburners and oth'r
control devices for particular applications have been outlined. The costs
associated with extensive afterburner application to control stationary sources
emitting nationally significant amounts of combustible pollutants excluding,
however, control of sources objectionable because of odor, is estimated to be
$360 million for equipment purchase and installation. The fuel cost associated
with this level of control (again excluding control of odorous sources) is
estimated to be $466 million for the purchase of 0.931 trillion scf of natural
gas annually. These estimates are based upon achieving the max-l.ium feasible
level of control which would reduce the total emissions shown in Table 21.1 for
1970 by 11.1 billion Ibs per year, or about 15$. At this level of control,
approximately half of the total emissions from the stationary sources suitable
for afterburning would be incinerated.
It should be emphasized again that these costs make no provision for
controlling sources objectionable primarily because of malodor or carbon
monoxide, and they would be increased if these costs were included. The
assumptions underlying these estimates are given in section 21.3.3-
S-14121
-------
Table 21.1. ESTIMATES OF NATIONWIDE EMISSIONS OF ORGANIC
CHEMICAIS FROM MOBILE AND STATIONARY SOURCES
Mobile Sources
Motor Vehicles
Other Transportation
Total Mobile
Incidental Sources
Forest Fires
Coal Banks
Building Fires
Total Incidental
Stationary Sources
Solvent Evaporation
Refuse Disposal
Agricultural Burning
Refineries
Gasoline Marketing
Carbon Black Manufacture
Wood Burning
Charcoal Manufacture
Metallurgical Coke Manufacture
Fuel Oil Burning
Petrochemical Manufacture
Automobile Body Incineration
Smoke Houses
Coffee Roasting
Steel Drum Reclaimation
Brake Shoe Deb ending
Total Stationary
Total All Sources
Emissions, 10° Ibs
1965
36300
2000
38300
4400
400
200
5000
13400
3430
34oo
2940
1710
1570
800
417
389
324
168
6
1
1
1
< l
28600
71900
1970
33200
2030
35200
4400
400
200
5000
14500
3510
3400
3340
1850
1450
800
426
404
364
194
6
1
1
1
< 1
30200
70400
1980
8810
3090
11900
4400
400
200
5000
22100
880
3400
4300
2470
1390
800
649
404
441
484
9
1
2
1
< 1
37300
54200
1990
4320
4700
9020
4400
400
200
5000
38100
1330
3400
5620
3440
1630
800
989
4o4
567
1200
12
1
2
1
< 1
57500
71500
2000
4900
7160
12060
4400
400
200
5000
75000
2010
3400
7370
4760
2150
800
1510
404
924
3020
X3
2
2
1
< 1
101400
118500
S-14121
-------
291
Table 21.2. THE TOTAL NUMBER OF POINT ODOR SOURCES
Odor Source
No. of Plants
Plastics Fabrication
Meat Packing
Pharmaceutical Manufacturing
Poultry Dressing
Soap Manufacture
Fish Canning
Rendering
Breweries
5000
2700
875
668
320
268
185
a)Census ofManufactures, 1967-
S-1JH21
-------
2^2
^1.2 Apolioability of Afterburners to Air Pollution Control Problems
Ir.iineration can control the emissions of any gases, vapors, or aerosols
.:.'.icr. are comcus^ibls or thermally decompose at high temperatures. Most appli-
cations are for the control of effluent streams objectionable because of malodor
or smog potential, e.g., rendering or paint baking effluents, but incineration
nas also been applied successfully to effluents containing combustible
particulates, e.g., phthalic anhydride manufacture and smokehouse effluents.
Hov/ever, when an effluent stream is objectionable because of particulates con ,ent
only, mechanical collection devices such as fabric filters or wet collectors give
r.ighly efficient control at a considerable cost saving relative to
afterburners.21'19)
Before incineration is chosen as a control technique, it should be
determined that the combustion products are not offensive pollutants themselves.
For example, incineration of organic chlorides produces hydrogen chloride, a
serious pollutant. Other effluent components producing undesirable combustion
products are the other halide compounds, nitrogenous compounds or sulfurous
compounds. When incineration is the only feasible control method for a particular
effluent it is possible, of course, to cleanse the afterburner effluent using a
secondary control device, e.g., a wet scrubber. In this way the range of
incineration applicability is considerably extended.
Finally, incineration requires that the effluent be collected and ducted
to the afterburner with minimal addition of diluent air. The fuel costs for
afterburning are directly proportional to the total volume of effluent to be
treated. Thus unless the effluent can be efficiently collected the cost of
afterburning may be prohibitive.
Malodorous emissions are particularly well suited to afterburner control.
Alternative methods such as scrubbing or adsorption frequently cannot achieve the
high degree of removal of malodorous gases and vapors required for satisfactory
odor control. This is not a problem for well designed afterburners, however.
In addition, the afterburner can control combustible particulates which sometimes
accompany malodorous gases.
21.j Combustible Air Pollutants; The Nature and Dimensions of the Problem
We have estimated the rate at which combustible compounds are emitted
in the United States for 1968 and 1970 from processes which are and are not
amenable to afterburner control. The emission rates have been projected thru
the years 1980, 1990 and 2000, thus giving an indication of how the pollution •
picture will change in the next 30 years if further controls are not implemented.
In most cases it was assumed that either there were no controls at present, or
that only those portions of an industry located in California were presently
controlled. Solvent emission sources are the incinerable sources most significant
on a national basis. As shown in Tables 21.1 and 21.3 the total estimated solvent
emissions for 1970 are 1^.5 billion Ibs; only the gasoline powered automobile,
emitting 32 billion Ibs, accounts for a greater tonnage. Other significant
sources of incinerable emissions are refuse incineration, oil refineries, gasoline
marketing, carbon black manufacture, charcoal manufacture, metallurgical coke
manufacture, and chemical plants (as represented by the petrochemical portion of
the industry).
S-14121 Table 21.3 follows
-------
o» l/l
NJ I
CO
Table 21-3- CHARACTERIZATION OF COMBUSTIBLE EMISSION SOURCES
^^-^^^^
SOURCE ^"""^"---^^^
SOLVENT L*'nHC
adheilves
degrees Ing octal parts
dry cleaning
fiberglass curing
graphic arts
herbicides
rubber products fabrication
surface coating nanufacturc
surfice oo* tinge ( Industrial)
surface coatlnps (trade)
surfice coatings recovers
POOD AND BYPRODUCTS
breweries
charbrollLng rectaurant
coffee roasting
deep fat frying
fish canning
neat and bone meal drying
neat packing plant"
poultry 'rfsltif plants
rcrJerlr.- plantg
snbker-ni-cj
vegetable oil ei tract Ion
COMBUSTIO'I FRXESSES
autc body incineration
brake 'hoe reclaim lion
copper fire Insulation burnoff
fuel oil burning
refuss incineration
secondary nonferroua aetal recovery
steel drun recondition ing
GENERAL INDUSTRIAL HIOCESSES
carbon black Manufacture
charcoal manufacture
iheBlral plants
hardboard aanufacturt
grey iron foundries
natallurgical cote mmmuwfactura
oil refineries
paper Manufacture (pulp Bills)
pharsaceutical nanufacture
plastics fabrication
soap nanufacture
MISCELLANEOUS SOURCE
autODObilra
gasoline Barks tlrj*
oil tesverlnf of n-tala
•average systems
a) InfoBBBtlon not available.
No Sources ur
• J,prul 1970
Enl.alcn, m Ifry
ol Oruanlcs
NbA m Ibs
185 plants
a)
i 3 m ibs
a)
320 plants
a)
2700 plants
8W> plants
268 plants
1 2 m Ibs
338 plants
3 m Ibs
0 2 m Ibc
<2 m Ibs
36k m Ibs
3505 m Ibs
b03 plants
0.8 m ibs
1U9 Ml Ibs
>2£ m ibs
19>> m Icr
a)
10{< plants
288 Ib.
2788 h. Ibs
279 Pla s
875 plan <
5000 plsnu
668 plants
32060 m Ibs
1852 m Ibs
.)
a)
Ease of
A! terburner
control '
2
1
2
5
5
1
3
3
1
1
3
3
3
3
2
J
2
2
2
5
3
3
3
3
b)
3
3
3
3
J
1
J
2
2
1
3
1
3
b)
b)
KNOWN OIIU.ION:,
Hydn .BrtMinj
Aromatic
,X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
non-
ercamtlc
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Organic
Halldeo
<
X
X
X
X
X
X
Organic
Ailds
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Alcohols
X
X
X
X
X
X
X
X
X
Aldehydes
X
X
X
X
X
Jl
X
X
X
X
X
X
X
X
X
X
Nitrogenous
Organics
X
X
X
X
X
X
X
Esters
X
X
X
I
X
X
X
,
X
X
Krtnnrj
X
X
X
X
X
X
X
X
X
X
Lthrr
X
X
X
X
X
X
X
Sulfurouo
Organic
or lljS
X
X
X
X
X
X
X
Odor
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
1
1
X
X
•
i
X
X
X
X
Oil. r
smoke
•inokf
caokc
ajaunle
saoke
co»ke
SOj
smoke
smote
sxoke
combustible
partlculates
CO
ammonia
co, no,
smoke
.ill. i >il
Kill,
^ *>
3 •>
1.6
b 1
3.6
7 8
-
-
k 6
S 7
85
-
-
•
8.1)
1.5
2 7
k 7
'•.r
k>7
k 5
U.7
-
it I
J.7
b.O
"> 9
2 7
-
V^
a.b
8 3
•
5 5
"
It. 6
i 5
2.7
~
b) adsWioni have been Included even though the PTOCMS is not asenable to afterburrvr control |
d) Trai^d3%d"WL^ii.p^to*JM.5 Sur?a^iorttn8CtoIStot.»r it was easilj cinirollab-e, .e3 controllaole irtth considerable difficulty, or control ..at 'avoeslb.o or p. wiible enl/ rith extrw difficult*, resp-cilwly
e) The Index <-L. range frcm sera for etcUiiow Jlth no offensive qualities to trn for si eKicslon vlth saxiiu. offons.vei.es8
tf
-------
293
Not all of these sources are controllable by afterburners, e.g.,
automobiles, open burning of refuse, some of the industrial process emissions,
and some of the solvent emissions, but they have been included to give a clearer
picture of the total problem.
Of the large tonnage emission sources shown in Table -21-1, afterburning
is commonly applied only to solvent sources. In the Los Angeles acd the
San Francisco Bay areas afterburners have played an important role in the efforts
to control photochemically reactive emissions. As photochemical smog becomes a
problem in other metropolitan areas, afterburners will be applied -to tMs problem
on a much wider scale.
Many smoke and odor sources, while not significant in terms of national
tonnage, do represent severe local nuisances. These sources are the most comtu~n
applications of afterburner control. Typical examples are emission control foi
smokehouses, coffee roasters, rubber curing ovens, waste incinerators, and
rendering processes. The offensiveness of odorous emissions is only weakly
reflected in the total tonnage emissions, because odor intensity is proportional
to the logarithm of the odorant concentration, as described by the Weber-Fechner
law,
I = k In C,
where I is the intensity of odor sensation,
k is the constant peculiar to the odorant in question, and
C is the odorant concentration.
For any concentration above the perception threshold, the apparent intensity
increases slowly with increasing concentration. This implies l) that a large
scale odor source may not be a much greater nuisance than a small source, and
2) that an odor control procedure must remove a very high percentage of the
odorant to be effective.
21.3.1 Types of Emissions Not Covered in This Study
No attempt was made to estimate the rate of emissions from sources which
are offensive because of severe odor problem. For these cases, it was felt that
the total number of point sources would give a better estimate of the magnitude
of the pollution problem. The number of sources was also given for emission
sources for which there was no basis available to compute the rate of emissions.
These counts are shown in Tables 21.2 and 21.3*
Combustible particulate emissions were excluded from -Mie emission rate
estimates. There were two primary reasons for this exclusion. First, published
data on particulate emissions rarely differentiate between combustible organic
particulates emitted by a process and noncombustible inorganic participates which
may be generated by a process. Second, as mentioned above, incineration is an
uneconomic method for particulates removal.
Also, carbon monoxide (CO) was excluded from the estimates because
industrial stationary sources contribute less than ten percent of the nationwide
emissions.21'31' Thus maximum possible contribution to lowering ambient CO levels
by afterburning is only about 10 percent.
S-1M21
-------
21.3.2 Emission Characterization
The chemical compounds known to be present in emissions from various
sources are identified in Table 21.3- Each emission has been assigned an arbitrary
!ndex to characterize the relative offensiveness of the emission* The index was
based upon the toxic ity, odor potential and photochemical reactivity of the
emission's components. This index is one of many possible indices, and another
investigator would likely derive a different index from the same facts. The
rankings should be used with this in mind. The index has a range of from ze_o
for no objectionable qualities to 10 for the maximum offensiveness. Assigning an
overall ranking to each source was considered, but rejected because an equitable
basis for such a ranking was not found.
It is hoped that the information presented will be sufficient to allow
interested users to utilize it in devising source ranking schemes which reflect
their particular requirements. This is, in fact, the primary reason for conducting
this phase of the study. A detailed presentation of the emission ranking system
is presented in 21. if. The methods used in projecting levels of pollutant
generating activity and the methods used in estimating emissions are given in
detail in 21.5 and 21.6 respectively.
Each source was also considered from the viewpoint of the suitability of
applying afterburners to control their combustible emissions. A rank of 3, 2, or 1
was assigned to each source according to whether it was easily controllable, was
controllable with considerable difficulty, or control was impossible or possible
only with extreme difficulty, respectively. For example, degreasing was- ranked
one because the trichloroethylene and perchloroethylene used as solvents are very
difficult to burn, and yield noxious combustion products, while fiberglass curing
was ranked three because the fumes are easily collected, burn well, and yield
carbon dioxide and water as combustion products.
21.3*3 Potential Emissions Reduction Through Extensive Afterburning.
If nationwide pollution control laws required the maximum feasible
application of afterburning to stationary sources, it is possible that in 1970
nationwide emissions could have been reduced by about 11.1 billion Ibs. This
figure is a rough estimate which was obtained by assuming that it is possible to
control fifty percent of the organic solvent emissions and organic emissions from
metallurgical coke manufacture, oil refineries, and chemical plants, and that
the emissions from all sources in Table 21.3 with ease of control rank 3 could
be controlled at the 95 percent level. It is instructive to consider a rough .
estimate of the costs associated with this level of control. We assume that the
organic emissions are in a 400°F fume at a concentration of 0.1$ v as hexane, and
that they will be incinerated at 1^00 °F using natural gas as fuel with no heat
recovery. At 0.1$ v hexane, there are 1000 scf fume/scf hexane,
1010 lb hexane\ /379 scf hexanex / 1000 scf fum
scf fume = f}'10 1010 lb hexane\ /379 scf hexanex / 1000 scf fumex .
\86 Ib/lb mole hexane / \lb mole hexanej V, scf hexane )
4.85 x 1013 scf fume/year
S-
-------
295
From Figure 11-10, operation of these afterburners at ll*OO°F will require the
burning of 0.0192 scf natural gas/scf fume.
required gas = (4.85 x 1013 scf fume)(0.0192 scf gas/scf fume) »
0.931 trillion scf gas/year.
gas cost at $.50/103 scf a $l*.66 million/year
The fume flow of ^.85 x 1013 scf/year corresponds to a flow rate of about 9 x 107
scfm, assuming 2k hour per day, 7 days per week operation. If we assun.^ a
purchase cost of $2/scfm (a low estimate as per figure 11.1 or 11.2) the purch^sa
cost of the required afterburners would be $180 million, installed costs vould
be about twice this amount, or $360 million. This cost would be quadrupled Vy
estimating on the five day - kO hour week basis which is common to many of the
industries emitting combustible pollutants. Heat recovery could reduce the fuej
usage by about half, but would increase the purchase costs by a factor of three
or four. For comparison, the marketed production of natural gas In the United
States in 1970 was 21.9 trillion scf. It should also be noted that the
calculations make no provision for controlling malodorous fumes whose emission
rates were not estimated in this study, or for controlling CO emissions. The
very high cost and natural gas consumption associated with extensive afterburning
should be considered when pollution control policies are formulated by regulatory
agencies.
21.k Emission RariVjng Procedure
The objectionable factors which were used to characterise the degree of
offensiveness of a particular compound were toxicity, odor, and photochemical
reactivity. Toxicity factors were based on the threshold limit -values (ILV) for
airborne contaminants adopted by the American Conference of Governmental, Industrial
Hygienists (ACGIH-.21"7) The TLVs are given in the second column of Table;21-4.
A toxicity factor (TF) was then assigned according to *the scheme (Shown in .
Table 21-5. For compounds which have not been assigned•TLVs by .the ACQtH, a TF
was estimated by interpolating between TLVs for compounds of similar structure.
For each class of compounds the TFs were then averaged to obtain a suitable -ST.
for ranking the entire class. The TFs are shown in colijtmn 3 of. Table 21-^ „
The odorant rankings were adapted from rankings proposed in a NAKJA
sponsored study of the social and economic impact of odors. ~a) The original
rankings were extended by assigning an odor factor (0F) of two to aromatic, hydro-
carbons and to organic halides, two classes of compounds which are included in the
present study but which were not treated separately in the odor Ptudy. tie OFs
are shown in column k of Table 21-4.
Photochemical reactivities for the compounds were ranked usizlg an
adaption of the system proposed by Altshuller.21-8/ Rankings for caapcsjBMls not
included in Altshuller1 s original work were obtained by comparing hie xsbkingp
with the reactivities reported by other workers in the field.2*--16*11*12*20*^)
The comparisons were rather arbitrary since different invest igatojc used 'different
measures of photochemical reactivity, and also because data could not be found
for all compounds of interest. Again the rankings for individual compounds were
combined to form an average ranking for each compound class. The rankings*
designated as smog factors (SF), are shown in column 5 of Table
-------
296 •
Table 21-3. TOXICITY FACTORS AS
FUNCTIONS OF TLVS
TLV mg/m3
0-10
11 - 50
51 - 200
201 - 500
501 - 2000
2001 +
TF
k
3
2
1
0
In order to obtain an overall index for the offensiveness of each
effluent, these rankings were applied in the following manner. The TF and OF
rankings were scaled to a 0 - 10 range by multiplying by 2 and 2.5 respectively.
The overall index was taken as a weighted sum of the three factors. Weights of
3, 2, and 1 were assigned to the toxicity, odor, and smog factors respectively.
Thus for an emission containing classes of compounds GI, Cg, ... Cn, the overall
index was computed by
n
Index -
(2)(2-5 )(OFi)
1=1
Consider a rendering plant, for example. The effluent contains nr$rogeneous and
sulfurous organics (Table 21.3), thus n = 2, and the offensiveness~index is:
Index =
(67127
[6(5) + 500 + 1 + 6(5) + 500 + 0] • Q.k
This formula was arbitrarily chosen by the author, and it is recognized that
other combinations are possible. Note that this system is intended to provide a
combined offensiveness index for all the compounds in an effluent, not an index
for the various classes of compounds as entities. Some pollutants exhibit
offensive qualities which were not included in the present rankings, ' e.g*,
particulate formation, allergenic potential, or phytotoxicity. Particulfctes were
not included in the rankings because particulates were not included in the emission
estimates for the reasons noted in 21.3-1.
Allergenic potential was not included because, apparently, nearly all
aeroallergens are of natural origin, and thus not suitable to afterburner-
control.21"13) Although some industrial products, e.g., insecticides, linseed
oil, and some organic solvents, are allergens, their contribution to th# 'overall
problem appears minimal. Phytotoxicity was not included in the rankings because
apparently, among the common combustible pollutants, only ethylene"is a signifi-
cant phytotoxicant.21"14' (Products produced in smog reactions do show phyto-
toxicity, however.)
21-5 Pollution Generating Activity Estimates and Projections
Three methods were used to project the levels of air pollutant generating
activities. The first method tied future activity levels to the rate of increase
of the population as projected by the U.S. Bureau of the Census series D
projection.21-3) This method was applied to projecting levels of coffee roasting
and meat smoking, using base data from the Census of Manufactures, 196?.g1-15)
S-2.1H21
Tables 21-4 and 21-6 follow
-------
Table 21-4. TOXICITY, SMOG, AND ODOR FACTORS FOR COMBUSTIBLE POLLUTANTS
.rw it '•vie s
a;- i.ie-^ie
i:^i:^'
acrvlo-It-Iie
aniline
•se'.* i aniie
tolje*ieil lascyaate
rltroben:e-e
Eatera
acrylates
aa. 1 acetate
st-er b-t/1 acetates
c*- ' ace. ate
e.r.,1 uetyl acetate
iiopropyl acetate
aefyl acetate
Ethers
etwyl et-er
gl/ccl ethers
Isoprocy- e'-har
Ketanei
ase'c-e
c/clohexara-e
dlacetone alcohol
dllBobutyl ketone
1 aop^orone
cethyl ethyl ketone
oethy. iaobutyl ketone
Ron-Aroratlc Hydrocarbons
aeetylere
butadiene
eye lone xane
ethylere
Mg-er oleflna
napthaa
prop/»eie
turpeitlno
Aromatic Pjdrocarbons
ber=ere
ethjlbeiaere
styeie
toluene
xjle-ies
Organic iSildes
a.*/! chloride
beizy- cblorlde
o-dlch lore benzene
;-zic*lcr3benzena
ethyl chloride
-Blhyl cv.orid>
atfyl chlorofom
eethylena eflorlde
percwlorocthyl*D<
tr'cbl'->retrylene
vinyl chloride
Organic Aclda ard Anhydride!
acetic aciJ
cresol
eoleic on^vdrlde
Jier^l
xylene based aclda (eg,
phthallc anhydride)
Alcohol a
vffl alcohol
1 Bi-bui.yl alcohol
•oc-butyl alcohol
tert-tutyl alcohol
decyl nlcohola
ethyl alcohol
glycerine (oiflt)
Iso-octyl ilcohol
iioprcpyl alcohol
cethyl alcohol
aethyl Itobutyl carblnol
oxo-alcoholi
E1 Ifurous ."'Cbiii^a and hs*)
JI.-G. ij- sul1 a^
fl'hjl -crcapnar
hidro.en sulflde
"elW ' 'ie-tflP«n
-^'
)6C
!!'5
w
.9
12
S
35-100
7'0
lliCO
950
610
1200
30-TbO
2100
2UOO
200
210
290
lliO
590
HO
A
2200
10SC
A
-1000
LOO
/,
560
90
'?
750
3
5
300
<*50
2(00
210
1900
mo
670
535
1300
«
??
1
19
12
300
1*50
300
1900
330
?(>0
10J
,
A
l'i
'
""^»rjCl
3
2
«
5
li
li
L
5
5
2
L
1
1
1
1
1
1
1
1
1
2
0
2
0
3
2
2
3
1
2
1
0
0
1
0
1
2
0
1
2
3
2
2
1
2
Z
5
5
2
2
0
2
1
1
1
1
1
li
L
!,
5
L
'a
1
2
2
2
?
?
j
0
3
1
2
3
s
5
5
t.
5
££'
3
li
1
1
2
1
2
2
3
1
1.
Fictor
5
1
.1 5
•
1
j
:
;
C
3
-
2
0
0
3
3
3
1
k
5
0
6
1
li
7
5
1
5
I
0
3
8
3
«
2
;
3
0
0
0
0
0
0
0
5
I
0
1
I
3
2
0
.
0
a) Threshold Llv.it Values o' Airborne Coi.tanlnants adopted by Aaarlcan Conference of
GoverrL-untal Induatrlal i<>g!enlata, 1970 'T>.rcs*-3ld Unit vnluai (TLV'a) refer to
airborne concentration! of aubatarcei anJ rep-cse-^ conditions under vhlch It IB believed
that nearly nil workers cay bo rcpvatoll) ) Blank sjiocea in the tabli. indicate thnt inforaatlo-. was not a-allable P"
S-14121
67784
c) Toxic it/ pBclo-n range fron r«ro to f » for n-isun _o roxlmiD toxlclty rcBpectlvely
d) Odor facto-t innge '-en _eio to '» for on- to -&A n-." odor offenalveneaa
*) s-aog foctprn range *rop -e-o to 10 for none to -hx'-u- s-»g pracuroor activity
-------
Table 21-6. SOLVENTS PRODUCTION AND EMISSIONS
£on2*::e
et:;ylben:ene
,?r-e:ial ::a:r:has
*oV:wne
f-ii'Der.tine3^
vvlunes
o-dichlorobenzeue
p-dichlorobenzene
methyl chloroform
methylene chloride
p^rcjiloroethylene
tric-.lorsethyl-Mie
ar^v 1 alcohol
n- butyl alcohol
otlier butyl alcohols
,q.e:yi alcohols.
str.yl alsohola''
glycerine
isirropyl alcohol
:r_e*,r.yl alcohol
~ethyl iscbutyi cardinal
:xc-alcohcls
eresc'Is
--^3
arc' I acetate
r.-cutyl acetate
other butyl acetates
ethyl acetate
J-ethyl-hexyl acetate
iscbutyl isobutyrate
i = c-prcpyl acetate
ne t hy 1 ac e t at e
,._v_,
B^ = ^"'-a
cy c 1 ohex anone
iiisotutyl ketone
isophcrar.e
nethyi ethyl ketone
methyl isobutyl Ketone
rTTPEH=
ethyl ether
glycol others
isopropyl ether
MISCELLANEOUS
nitrobenzene
'*• 01'
"iears
11
5
5
17
16
9
25
20
2
20
20
20
ON?
19
17
4
15
7. .
_b)
20
a
6
20
GBP
13
8
20
GSP
GUP
5
4
20
6
01.T
ON?
19
7
19
3
6
6
% Growth
Projected
9
6
1.2
9
-6
5
7
0:5
9
10
9
4.5
4.3
5
5
10
5
2
0-4
7
2.5
3
5
4
1.5
8
1.
1*
4
li
0
6.5
9
- r
It
4
7
U
2.5
8
10
10
Production Estimate
1968
7,320
4,500
7,080
5,050
2.3
3,867
61
7C
300
302
637
519
38
433
500
153
732
100
1,890
3,817
43
620
81
10
63
60
179
3
5
48
9
1,516
131
86
26
451
182
100
1(41
6
396
1970
8,530
5,600
7,730
5,030
1.8
3,150
65
67
355
1(16
702
610
111.
500
560
160
760
105
1,990
4,944
46
690
93
11
72
64
159
3
6
52
9
1,598
500
95
95
31
500
210
103
520
7
500
I960
22,700
10,000
8,650
13,300
1.0
5,100
123
70
870
1,180
1,500
920
63
900
1,000
530
1,260
126
2,300
9,500
58
880
155
16
85 .
136
240
4
9
79
9
3,100
1,300
lit 5
145
46
1,000
300
135
1,200
ai
1,500
1990
55,600
18,500
9,700
36,300
0.6
8,200
2000
135 ,000
36,000
11,000
91,500
0.3
13,500
% Used
Solvent
1.5
0.25
100
15
100
15
Hydrocarbon Totals
245
77
2,100
3,500
li,600
1,500
470
82
5,200
10,900
12,000
2,400
30
90
84
60
80
98
Organic Chloride Totals
104
1,600
1,750
1,600
2,120
156
3,400
20,000
74
1,200
260
24
100
290
370
6
13
120
9
6,000
3,300
240
70
1,950
430
180
2,800
62
4,300
163
3,000
3,200
4,600
3,440
190
i(,900
44 ,0 00
92
1,620
450
47
20
10
15
100
100
11
10
83
3
10
Alcohol Totals
35
115
620
570
9
19
180
9
85
83
10
90
80
100
75
80
Ester Totals
12,000
8,300
330
330
108
3,900
620
31
• \
70
63
100
74
100
Ketone Totals
240
6,600
180
31
40
60
Ether Totals
12,000
3
All Solvents Total
Solvent Use, MM IDS.
1968
110
11
7,080
758
2.3
580
8,541
18
63
252
181
510
498
1,522
IB
87
50
23
732
100
208
382
36
19
8
1,663
9
52
6
161
2
5
36
7
278
471
19
61
28
334
182
1,149
' 31
176
U
211
12
13.376
1970
126
14
7,730
775
1.8
473
9,101
20
60
898
851
562
598
1,789
21
100
56
27
760
105
• 219
494
38
21
9
1,550
9
60
6
143
2
6
39
7
272
495
20
67
60
31
370
210
1,253
32
208
4
244
15
14,524
1980
341
25
8,650
2,070
1.0
765
11,85?
37
63
731
708
1,440
902
3,881
30
180
100
80
1,260
128
253
950
48
26
16
3,071
14
71
14
216
3
9
59
7
393
961
52
102
91
46
740
300
2,292
42.
480
13
535
45
22,069
1990
834
46
9,700
5,445
0.6
1,230
17,256
74
69
1 764
2," 100
3,680
1,470
9,157
49
320
175
240
2,120
156
374
2,000
61
36
26
5,557
20
83
29
330
5
13
90
7
577
1,860
132
168
151
70
1,443
430
4,254
56
1,120
37
1,213
129
38,143
2000
2,025
90.
11,000
14,175
0.3
2,025
29,315.
141
74
4,368
6,540
9,600
2,352
23,075
77
600
320
690
3,440'
190
539
4,400
76
49
45
10,426
30
95
62
513
7
19
135
7
868
3,720
332
231
208
108
2,886
620
8,105
74
2,640
108
2,822
360
74.9T1
a) Ajr.ount used as solvent
b) IPA growth expected to go from zero to -4? in 1975
c) Froductioa histories vere obtained primarily from Chemical Economics Handbook by Stanford Research Institute and from Synthetic Organic
Chemicals published by the U.S. Tariff Commission. Some data vere obtained from the petrochemical aeries by F. 0. Sullivan in
Hydrocarbon Processing and from Solvent Emission Control Laws and the Coatings and Solvents Industry by the Environmental Sciences
Services Corp., Stanford, Connecticut. Solvent usage fractions vere obtained from the above named publications and aleo from
Oil. Paint, and Drug Reporter Chemical Profile.
S-14121
67784
-------
297
Ir.e £rc.-wth tri production of most industrial products was assumed to
follow an exponential growtn law. Examination of production data for commodities
plotted against tL~e on a semi-logarithmic plot reveals that exponential growth
often persists over very long periods before a decay develops in the rate of
erowth. Because a decay inevitably develops in the growth at some time, however,
a conservative approach was taken in fitting a straight line to t-he production
data curves. In particular, when the latest data indicated a downward trend, heavy
emphasis was given to the latest data in projecting future production rates. The
additional constraint was imposed that growth rates were limited to a maximum of
10$/year. While this method can lead to considerable error over a pericd of
thirty years, better estimates would have to be based upon a careful analysis of
all product end use markets. Such an analysis was clearly beyond the scope of
the present study.
The chemical industry was treated in a special manner. Because of the
general complexity of the industry with an extremeley diverse mix of products and
feedstocks, v/e confined attention to only a limited portion of the industry.
Specifically the fraction of the industry processing the major petrochemical
feedstocks, acetylene, benzene, propylene, xylenes, toluene, ethylene, butylenes,
and methanol to their primary first generation derivatives was considered. It
was felt that a treatment of this portion of the industry would give an estimate
for the order of magnitude of the chemical industry's air pollution problems.
For the larger volume feedstocks, the growth rate of the products were each
evaluated separately, and the total consumption of feedstock was obtained by
summation. For the butylenes, methanol, toluene, xylenes, and acetylene, the
feedstock growth was estimated from past history, and the most recent product mix
information available was assumed to apply for the period of estimation.
The third method used for predicting production rates was to assume
that growth proceeded at a rate equal to that projected for the GNP. This method
was applied only to products where insufficient historical production data was
available. The second method was preferred wherever production data for three
or more years could be found.
Solvent usage projections were made by multiplying the production
estimate for each particular solvent by the fraction of total production which is
used as solvent. Data on solvent usage was taken from Chemical Profiles, the
Chemical Economics Handbook, and a study of the impact of solvent emission control
laws .2*~16-17-18) The estimates of solvent production and usage are shown in
Table 21-6. Also shown are the number of years for which past production data
could be obtained and the projected annual production growth rate.
21.6 Emission Estimates and Projections; Methods and Results
We have estimated the emission rates of combustible pollutants for most
of the large volume emissions sources. The methods used and detailed projections
of the emission estimates are presented below for each source of effluent. As
noted above, the effluent rates of sources objectionable primarily because of
malodorous emissions have not been estimated. The data for the number of process
point sources reported in Table 21-2 were taken from the 1967 Census of
Manufactures.
S-14121
-------
21.6.1 Solvgri.il ^.TJEoion Raxes
Is-i.-ates of solvent emissions have been made on the assumption that all
e.--e~iicEl ana petroleum produces sold as solvents will eventually find their way
into the atrr.ospr.ere. The fractions of production of solvent chemicals actually
used as solvents were obtained from References 21-16, IT* and 18. It was also
assumed that the fractions used as solvent remained constant through the 30 year
projection period. The solvent emission estimates are shown in Table 21-^.
Solvent end use markets are extremely diverse. Probably a large fraction of the
solvents used commercially could be collected and incinerated if necessary, while
those sold on the retail markets, e.g., in paints, polishes, etc., could not.
21.6.2 Oil Refinery Emissions
Oil refining is a very large industry in the United States, and
refineries are a significant source of combustible emissions. As a result, an
unusually large amount of information is available about 1iie emission character-
istics of refineries. This allows a detailed estimate of refinery emissions by
source. The detailed emission estimates and projections are shown in Table 21-7
and summarized in Table 21-8. The emission factors from Reference 21-1 were used.
It was generally assumed that only the 13$ of the refining capacity located in
California was controlled.21"24/ This assumption was made because controlling
nydrocarbon emissions is generally not profitable and at the present time hydro-
carbon emissions are specifically controlled only in California. Catalytic
cracker emissions are an exception to this rule because the installation of a CO
(carbon monoxide) boiler to utilize the heat content of catalyst regenerator off
gases is an economically attractive investment. Consequently, it was assumed
that 50$ of present catalytic cracker capacity and all new capacity additions
are controlled. The fractions of crude input which are fed to catalytic crackers
and to vacuum distillation are assumed constant at 1971 values, 45$ and 37$
respectively.21"24) The ratio of effluent water to crude capacity was taken as
1.11 bbl waste water/bbl crude capacity.21"23' The cooling water circulation rate
used was 41.3 bbl water/bbl crude, and the heat requirements were 1.94 x 10s
3tu/bbl crude from refinery gas and 4.64 x 105 Btu/bbl crude for fuel oil.21"23)
Storage tank emissions were conservatively estimated by using the ratio of storage
tank emissions to crude capacity reported for Los Angeles County, California.21"23)
Possibly from 35 to 65$ of the 1970 combustible pollutants from
refineries could be controlled by extensive afterburning if required. The primary
obstacle to control of many of these sources is the expense of collecting the
emissions for incineration. Sources such as cooling towers, valve and flange
leaks, and pump seals are too widely distributed for efficient control.
21.6.3 Gasoline Marketing Emissions
An average emission factor, 0.882 Ib HC emitted/bbl gasoline, was used
to estimate emissions from filling service GLaLLon'c and automobile tanks.21"1)
The projected gasoline production and marketing emissions are shown in Table 21-9.
Incineration is not an efficient method for controlling emissions from t' is
source; vapor recovery should be employed.
S-14121
-------
Table_ 21 ..I_..REFINERY EMISSION BY SOURCE. MM LBS
299
Production est., MM bbl/yr
a)
cat crackers HC
blowdown systems
process drains
vacuum jets
cooling towers
pipeline valves and flanges
vessel relief valves
pump seals
compressor seals
other
storage tanks
furnaces HC
HC, Total
cat crackers - aldehydes
furnaces , aldehydes
Aldehyde , total
1968
377U
172.1
988.9
769.9
157.8
39.2
105.7
Ul.5
6U. 2
18.9
37.7
U62.8
60.1+
2919.0
15.3
9.2
2U.5
1970
1*281*
195.7
1122.1*
873.9
179.1
1*1*. 6
120.0
1*7.1
72.8
21.U
1*2.8
525. u
68.5
3313.0
17.5
10.5
28.0
1980
5600
195.7
1U67.2
lll*2 . 1*
231*. 1
58.2
156.8
61.6
95.2
28.0
56.0
686.7
89.6
1*272.0
17-5
13.7
31.2
1990
7UOO
195. f
1938.8
1509.6
309.3
77.0
207.2
81. I*
125.8
37.0
71*. o
907.1*
118.1*
5581.0
17-5
18.1
35.6
2000
9800
195.7
2567.6
1999-2
1*09-6
101.9
277. U
107.8
166.6
1*9.0
98.0
1201.8
156.8
7332.0
17-5
23.9
1*1.1*
a) Assume 50$ of 1970 capacity and 100$ of post 1970 capacity increases are
controlled by CO boilers, with zero emissions.
S-11H21
-------
Table 21.8 REFINERY EMISSION. MM LBS
Year
1968
70
80
90
2000
Crude
Oil
Demand,
MM bbl
3,77^
1*,28U
5,600
7,1*00
9,800
HC
2,919
3,313
4,272
5,581
7,332
Aldehydes
2U. 1*
28.0
31.2
35-6
1*1. U
Table 21.9 GASOLINE MARKETING EMISSION, MM LBS
Year
1968
70
80
90
2000
Gasoline
Production,
MM bbl
1.93U
2,100
2,800
3,900
5,1*00
HC
1,705
1,852
2,1*70
3,M*0
U.763
a)Emissions from filling sertice station
and automobile tanks.
S-14121
-------
301
21.6.^ ETissicr.s fror. Refuse Burning
Data on the volume of solid wastes and methods of disposal were taken
from Technical-Economic Study of Solid Waste Disposal Needs and Practices.21"26)
The assumptions used were:
l) that municipal waste incineration will increase at> projected in
the solid waste study;
2) that industrial waste incineration would increase proportionately
with the GNP, i.e., at k.3/year;
3) that open dump burning will be eliminated by 1980;
k) that sawmill waste generation will level off after 1975-
The projected emissions are shown in Table 21-10. Note the sharp decrease in
emissions forecast for 1980 which results from stopping open burning. Of the
total emissions from refuse disposal only approximately 20 percent of the 1970
emissions which are from municipal and industrial incinerators (excluding "TeePee"
burners) could be afterburned. This fraction could probably be controlled more
efficiently using modern multiple chamber incinerators.
21.6.5 Emissions from Carbon Black Manufacture
The manufacture of carbon black is a large source of combustible
emissions, as shown in Table 21-11. The decrease in emissions in 1980 results
from the trend to replace natural gas fueled channel and furnace production with
production from the more economical and more efficient oil fueled furnace process.
Emission factors were taken from Reference 21-1. Possibly from ^0 to 70$ of the
organic emissions from carbon black manufacture could be incinerated.
21.6.6 Emissions from Petrochemical Manufacture
There has been very little emissions data reported for chemical
processes. However, because chemical plants are generally considered to be fairly
large scale sources, it was decided to make a rough estimate of their contribution
to combustible air pollution. It was assumed that chemical plants would emit at
the same rate as an uncontrolled refinery, about 0.005 lb effluent/lb
feedstock.21'23' Further, half of the effluent was assumed to be emitted as the
feedstock, and half was assumed to be converted to the reac tion product. For
example, assuming theoretical yield, 1 lb of ethylene produces 2.3 lb ethyl
chloride, so the emissions were assumed to be 0.0025 lb ethyle^/lb feed and
0.00576 lb ethyl chloride/lb feed. When the reaction product was a solid, the
emissions were counted as being entirely composed of the feedstock. The estimated
emissions are shown in Table 21-12. The magnitude of emission estimates found by
this method are probably within an order of magnitude. However, the identities
of the emissions are probably less accurate since a considerable fraction of
chemical emissions would probably be reaction by-products, process solvents,
catalysts, or other compounds. The fraction of chemical plant effluent which
could be burned is hard to estimate, but perhaps from 30 to 60$ of the total
emissions could be controlled.
S-14121
-------
Table 21.10 EMISSIONS FROM REFUSE DISPOSAL. MM
Year
1968
1970
I960
1990
2000
EC
1912
1966
628b)
954
1UUU
Aldehydes
50
5U
1*6
68
Organic
Acids
lUTO
1U86
206
308
100 U6U
aj Totals for municipal and industrial
incineration.
b) Assume pollution control legislation
will eliminate open burning by 1980.
Table 21.11 EMISSIONS FROM CARBON BLACK MANUFACTURE. MM LBS
Year
1968
1970
1980
1990
2000
Production ,
MM Ibs
2,812
2,935
UJOO
6,822
10 ,170
HC
1,574
1,4U9
1,392
1,627
2,153
S-14121
-------
Table 21.12 SCISSIONS FROM PETROCHEMICAL MANUFACTURE. MM_LB§
303
Non- aromatic Hydrocarbons
acetylene
butylenes
cy-hexane
ethylene
higher olefins
propylene
Total
Aromatic Hydrocarbons
benzene
styrene
toluene
xylenes
Total
Organic Halides
allyl chloride
benzylchloride
di chlorobenzenes
ethyl chlorides
methyl halides
vinyl chloride
Total
Organic Acids and Anhydrides
acetic acid
maleic anhydride
phenol
xylene base acids
Total
Alcohols
butyl alcohols
ethyl alcohol
isooctyl alcohol
isopropyl alcohol
methyl alcohol
Total
1968
2.1
1U.U
3.0
33.5
6.9
ll*.0
73-9
16.3
6.5
6.0
3.2
32.0
0.6
0.1
0.2
13.7
0.7
1.3
16.6
0.1
0.5
3.1
3.2
6.9
0.8
2.6
0.6
3-8
8.0
15.8
1970
1.9
17.0
3.1*
Ul.6
7.9
17.3
89.1
19-1
6.5
6.U
2.6
3U. 6
0.7
O.l
0.2
13.8
0.9
1.2
16.9
0.2
0.5
3.8
2.6
7.1
1.0
2.8
0.7
u.o
10. U
18.9
I960
1.2
31.9
8.6
139-9
7.9
1*3.3
232.8
1*2.6
19-9
9-5
I*. 2
76.2
0.9
0.1
0.3
67.6
1.8
0.7
71. U
0.3
1.3
9.U
U.2
15.2
1.8
3.8
l.U
6.0
21.1
31*. 1
1990
0.7
59-7
21. U
376.5
7.9
108.7
571*. 9
95-6
51.1*
iu.o
6.8
167.8
1.3
0.2
0.5
193.7
U.3
0.5
200.5
0.8
3.3
22.7
6.8
33.6
3.1*
5.3
2.6
8.9
50.6
70.8
2000
0.1*
78.0
53.6
1,050.2
7-5
270.3
l.Ut'n.l*
i
222.1
126.1
19.9
11.2
379-6
1.9
0.3
0.7
560.2
10.1
0.3
573.5
l.J
6.7
58.7
11.2
78.1*
1*.5
7.3
l • -*
3.1*
13.6
118.0
1U6.8
S-11H21
-------
Table 21.12 (CONTINUED)
Aldehydes
acetaldehyde
acrolein
butyraldehyde
formaldehyde
Total
Nitrogen-organics
acrylonitrile
aniline
methyl amines
toluene diisocyanate
Total
Esters, Ethers, Oxides
acrylates
ethylene oxide
glycol ethers
propylene oxide
Total
Ketones
acetone
Total Petrochemical
1968
1.2
0.1
1.5
2.6
5.U
3.7
0.5
0.3
0.5
5.0
1.3
7.1
0.3
2.1
10.8
1.2
167.6
1970
l.lt
0.1
1.6
3.3
6.U
U.6
0.5
0.3
0.5
5-9
1.6
8.8
O.U
2.8
13.6
1.5
19U.O
1980
3.3
0.3
3.5
6.7
13-8
9.7
0.9
0-7
0.8
12.1
2.U
20.0
0.9
10.6
23.9
U.5
U8U.O
1990
7-7
0.7
7-7
16.1
32.2
21.8
1.6
1.6
1.1
26.1
5-3
U5.U
2.1
28.2
81.0
12.7
1,199.6
2000
18.1
1-5
16.9
37-5
7U.O
67.1
2.5
3.7
1.6
7*+.9
12.0
101.5
U.8
79-8
198.1
35.2
3,020.9
S-14121
-------
305
Emissions fro-. Charcoal Xar.ufacture
Production data for hardwood charcoal T.anufacture is reported in the
1967 Census of Xanuf actures.21"15 It was assured that hardwood made up 75$ of
•:he total charcoal production. Charcoal manufacture emissions are sharply reduced
••hen chemical recovery of the effluents is practiced.21"1 Sines about 10 million
Its/year of -sethanol are produced from natural sources, while the potential yield
from charcoal manufacture is about 80 million Ibs/year, it was assumed that
chenical recovery is practiced with 12$ of charcoal production.21~16) The
estimated emissions from charcoal manufacture are shown in Table 1.2-15 The
emission factors were taken from Reference 21-1. Probably almost all combustible
emissions from charcoal manufacture could be burned, but it would probably prove
cheaper to recover the chemicals as byproducts.
21.6.8 Emissions from Other Sources
Emissions from metallurgical coke manufacture are shown in Table 21-14.
The emission factor is taken from Reference 21-1. The potential for afterburner
control here is limited, because about 55$ of the emissions are associated with
loading and unloading the coke ovens. This would present a problem in collecting
emissions. Perhaps 20-40$ of the emissions could be incinerated.
Emissions from coffee roasting are shown in Table 21-15- Coffee roasters
are very common users of afterburners for controlling the malodorous, smoky
emissions. The emission factor was taken from Reference 21-1. Coffee roasting
emissions are readily incinerated and probably 95$ could be treated.
Estimated smokehouse emissions are shown in Table 21-16. Smokehouse
emissions are also frequently controlled with afterburners. These emission factors
were taken from Reference 21-1. The 1967 Census of Manufactures reports smoked
meat production data only for those firms whose primary products are smoked
meats.21"15) For this study it was assumed that an additional 25$ was produced
by wholesalers and retailers. Smokehouse emissions could probably be controlled
at the 95$ level.
Afterburners have been applied to several metal recovery operations
which involve burning off nonmetallic components from the original products.
Examples are auto body incineration, brake shoes reclaimation, steel drum recondi-
tioning, and copper wire insulation burnoff. Of these processes, only auto body
incineration is a significant emission source. The emissions shown in Table 21-17
for auto body incineration assume that all caru scrapped in a given year ;m-
burned and that an average of 7-1$ of the car population iu scrapped uuuh
year.21"27/ The estimated automobile population wys Lakeri f'rorr iteforoTioo 2J.-'i,
and the emission factors from Reference 21-1.
Emissions from the burn-off step of steel drum reconditioning were
estimated by assuming that the emission factors for auto body incineration could
be applied to steel drums. This is a very rough approximation, but it should
give a result correct within an order of magnitude. Thus it was assumed that
incineration of 60 48 Ib steel drums produced emissions equal to those from one
2800 Ib auto body. Presently, about 50 million drums are being reconditioned
each year.21"28) It was assumed that this rate would increase at about
l$/year.21~28) The estimated emissions are shown in Table 21-18. Although the
S-14121
-------
Table 21.15 EMISSIONS FROM CHARCOAL MANUFACTURE,
Year
1968
70
80
90
2000
a)
Production
M tons
536
5^8
835
1273
1939
Combustible
Particulates
189
193
291*
UU8
683
HC
^7
U8
73
112
171
Methanol
72
73
112
170
259
Acetic
Acid
109
112
170
259
396
Total
Ul7
U26
6U9
989
1509
a) Assuming 75$ derived from hardwood and 25$ from softwood.
b) Assuming 12$ of production associated with chemical recovery since
natural methanol production is about 10 MM Ibs/yr.
Table 21.Ik EMISSIONS FROM METALLURGICAL COKE MANUFACTURE. MM LBS
Year
1968
70
80
90
2000
Production
MM tons
62.9
65
65
65
65
HC
389
kok
kok
IK*
IK*
S-14121
-------
307
Table 21.15 EMISSIONS FROM COFFEE ROASTING. MM LBS
Year
1968
1970
1980
1990
2000
Coffee
Production,
MM Ib
2002
20U3
2265
2535
279U
Organic
Acid
Emissions
1.1
1.1
1.2
l.U
1.5
Aldehyde
Emissions
0.2
0.2
0.3
0.3
0.3
Table 21. 16 EMISSIONS FROM SMOKE HOUSES. MM LBS
Year
1968
1970
1980
1990
2000
Total
Meat
Smoked ,
MM Ibs
12,525
12,800
lU,l69
15,863
17, ^82
HC
O.U
O.U
0.5
0.5
0.6
Aldehydes
0.5
0.5
0.6
0.6
0.7
Organic
Acids
1.2
1.2
1.3
1.5
1.7
S-
-------
Table 21.17 EMISSIONS FROM AUTO BODY INCINERATION. MM LBS
Year
1968
1970
I960
1990
2000
PQV-Q
Scraped,
MM
5.68
5-91
8.U6
11.6
13.3
HC
2.8
3.0
U.2
5.8
6.7
Aldehydes
1.1
1.2
1.7
2.3
2.7
Organic
Acids
1.7
1.8
2.5
3.5
U.O
Total*)
5.7
5.9
8.5
11.6
13.3
a) Rows jay not add "because of rounding.
Table 21.18 EMISSIONS FROM STEEL DRUM RECONDITIONING. MM POUNDS
Year
1968
1970
I960
1990
2000
No. of
bbls, MM
50
51
56
62
69
HC
O.U2
0.1+2
O.U6
0.52
O.U8
Aldehydes
0.16
0.16
0.18
0.20
0.22
Organic
Acids
0.26
0.26
0.28
0.32
0.3U
Total
0.8U
0.8U
0.92
l.OU
l.lU
S-
-------
309
emission levels are quite low, this process can create a severe local smoke ana
odor nuisance which can be controlled by afterburning.21"29'
Brake shoe debonding emissions are estimated using emission rate data
from Reference 21-29- It was assumed that a set of four brake shoes were reclaimed
for every 30,000 vehicle miles. Estimated annual mileages were obtained from
EPA thru 1990, and these mileages were extrapolated to year 2000.21~2) The
projected emissions are shown in Table 21-19.
Data on the annual rate at which copper wire insulation is burned off
were not found. However, an upper limit for 1970 emissions v/as obtained by
assuming that all the wire and light and heavy copper reclaimed by secondary
smelters was burned.21'30' Using emission rate data reported in Reference 2J-29
for automobile starter field coils and armatures (assuming 50$w copper), it wa*
estimated that the total combustible emissions are less than 2 million Ibs per
year. Probably 95$+ of the emissions from these reclamation sources could be
controlled with properly designed ovens and afterburners.
Emissions from two additional sources were estimated to provide a useful
perspective on the national significance of the combustible emissions problem.
Automobile emissions are shown in Table 21-20. The data for hydrocarbons were
obtained from EPA, except for the year 2000 estimates, which were extrapolated
from the EPA data.21'2/ The aldehyde and organic acid emissions were estimated
using the EPA mileage estimates and emission factors from Reference 21-1. These
emission factors do not project any decrease in aldehyde or organic acid emission
rates which might result from control devices installed on future automobiles
to enable them to meet federally specified carbon monoxide and hydrocarbon
emission levels. The federal regulation of automobile emissions should result in
a sharp decrease in the relative importance of the automobile as a pollutant
source.
Projected emission levels from fuel oil combustion are shown in
Table 21-21. They were estimated by using a rough average of the emission
factors for various types of fuel oil furnaces given in Reference 21-1. Historical
data indicate an approximate 3^/year growth of fuel oil consumption.
This completes the list of processes for which emission rates wore
projected. The processes studied do not represent an exhaustive list of all
sources amenable to afterburning. However, they were chosen to include the
important sources in terms of national tonnage, and thus to give a picture of the
potential for future air quality improvements through more extensive afterburning
of combustible emissions.
S-14121
-------
310
Table 21. 19 BRAKE SHOE DSBOITDING EMISSIONS. MM LBS
Year
1968
1970
1980
1990
2000
No. of
Shoes ,
Millions
132
lU2
192
235
267
HC
0.02
0.03
0.03
O.OU
0.05
Aldehydes
0.05
0.05
0.07
0.08
0.10
Organic
Acids
0.10
0.11
0.15
0.19
0.21
Table 21.20 AUTOMOBILE EMISSIONS. MM POUNDS
Year
1968
1970
1980
1990
2000
Vehicle
Miles ,
Billions
990
1,068
1,UUO
1,760
2,000
HC8-)
35.2UO
32,060
7,260
2,U20
2,7^0
Aldehydes*^
762
8U8
l,lU2
1,398
1,582
Organic
Acids
28U
306
Ul2
50U
571*
b)
Except for the year 2000, the HC estimates are provided
by Mr. J. H. Southerland, EPA. The year 2000 data are
obtained by extrapolating the EPA estimate for 1990.
Aldehyde and organic emissions are based upon the
emission factors given by McGraw and Duprey, and assume
no controls.
S-Ik121
-------
311
Table 21.21 EMISSION FROM FUEL OIL COMBUSTION. MM LBS
Year
1968
1970
1980
1990
2000
Fuel Oil
Consumption, MM bbl
15^3
1732
2100
2700
UUOO
HC
227
255
309
397
6U7
Aldehydes
97
109
132
170
277
S-1U121
-------
313
Chapter 22. RESEARCH RECOMMENDATIONS
The objective of this section is to recommend specific research and
development plans to provide technology where deficiencies exist, to generate the
data required to predict and improve process performance and economics and extend
the application of afterburners to additional air pollution sources. Each
recommendation contains a description of the purpose, scope, approach and an
estimated time and cost schedule. A priority ratio has been assigned to each
recommendation based on importance. Two research and development programs are
outlined. Plan A is bassd on an expenditure of one million dollars and Plan B,
three million dollars, both for five-year periods. Some projects are commc^ to
Plans A and B and some are not. Summaries of programs are contained in Tablt 22-1
(Plan A) and Table 22-2 (Plan B). The research program is split about &5% on
thermal afterburners and 15$ on catalytic units. This reflects the more complex
engineering questions in thermal units. The levels of fundamental studies on
the thermal and catalytic units are about the same.
22.1 Thermal Afterburners
The research and development program is aimed at the combustor with
fundamental and engineering studies included. Research effort on auxiliary
equipment like heat exchangers and fans is not considered necessary.
22.1.1 Field Test Data
Plan A
The objective of this project is to select appropriate instrumentation
and to obtain field data on 10 or so representative types of thermal afterburners.
These data would include careful measurements of efficiency as a function of
operating conditions on various feeds and types of equipment. This data will
provide a basis for correlating efficiency with design parameters and identifying
areas for equipment improvement. The system study showed that there was a mar> .-d
lack of quantitative performance data.
The first part of this project is the selection of simple, low cost,
rugged analytical instrumentation that can be used for field monitoring of
afterburner performance. This need is common to thermal and catalytic afterburners
and is perhaps more critical for catalytic units since performance monitoring is
more important because of catalyst deactivation. The species to be measured are
hydrocarbons (typical of fuel streams), partially oxygenated con^stion products,
carbon monoxide, oxygen, carbon dioxide and particulates. Dirtct read out of
results is highly desirable since afterburner feed conditions are often quite
variable. This indicates instrumental methods rather than wet chemical methods.
Two sets of instruments to monitor input and output simultaneously would be a
marked improvement. A great variety of analytical instrumentation is becoming
available and components should be selected from them. A system of instruments
should then be field tested for ruggedness, stability and interferences.
Preceding page blank
S-1^121
-------
The second part of this project is the collection of field data on
reprecentative units (10 or so are envisaged). The units tested should be
representative of the various equipment types and the various waste streams
burned. Particular attention should be paid to measuring as many of the operating
and performance variables as possible. The primary operating variables are flow
rate, temperature and residence time. The performance variables are feed
concentration and exhaust concentrations with particular attention paid to
partially oxygenated combustion products. These data will show in a quantitative
way the merits of the various afterburner design features. These data wi_l also
provide a basis for correlation of performance on an empirical basis. This
project would require 18 months to execute and would cost about $130 M and has
a priority 1.
Plan B
With the extended level of effort in Plan B, five additional units would
be tested and measurements of temperature profiles and velocity distributions in
related units would be undertaken. The additional measurements of temperatures
and velocities will permit a better correlation of the field data. The extended
program would require a total of 18 months to execute and would cost a total of
$175 M and has a priority of 1.
22.1.2 Kinetics of Combustion
Plan A
The objectives of the kinetic study are to measure chemical decomposition
rates of the typical waste materials being burnt. These experiments can be made
on a laboratory scale in which mixing can be removed as a rate limitation. The
chemical decomposition rates are fundamental parameters for a rational design
procedure for afterburners. As the system study showed, the controlling rate in
many cases is mixing rate in the combustor rather than the chemical oxidation step.
However, there are some cases where the chemical rate is limiting, and in most
cases mixing and oxidation rates can be made the same order. Furthermore,
operating temperatures could be reduced in many cases where the chemical rate is
currently many times faster than the mixing rates. The reduced operating temper-
ature will result in fuel economy, increased equipment life and less chance of
overheating damage due to feed surges. Measurements of oxidation rate should be
made in a laboratory apparatus in which conditions are carefully controlled and
measured. Mixing rates of contaminant with hot flue gas can be made very fast in
the laboratory so that chemical oxidation rates can be measured as a function of
temperature. Much of the literature data now available has mixing and oxidation
rates combined. The feeds used should be the hydrocarbons representative of
industrial emissions. Rate expressions should be developed relating oxidation
rate to hydrocarbon and oxidant concentrations. Correlation methods should be
developed to extend the data to a broader range of fuel components. This project
would need about one year for execution at a cost of about $110 M with a
priority 1.
S-11H21
-------
315
Plan B
With the extended level of effort in Plan B, this kinetics study should
be expanded to include more hydrocarbons, odorous compounds, liquid droplets,
solid particulates and materials containing heteroatoms like S, Cl, P, etc. The
odorous compounds have the special problem of often being in low v-oncentrations
to begin with and requiring high conversions to effect odor nuisance abatement.
The liquid droplets have a vaporization step which must occur first which may
require additional residence time. If complete vaporization can occur without
coking then the rates may not be much lower than for vaporized feevls. "i.f coking
occurs or solid particulates are present in the feed then the rates of oxidation
of these particules will have to be measured. It is expected that the particulate
oxidation rates would be low so that significant increases in residence time vould
be needed. The oxidation rates of feeds containing heteroatoms like Cl, for
example, are expected to be quite low so that kinetic data will be very important
in a design. The combustion products must also be measured carefully since they
will effect the downstream equipment requirements. The additional work on
kinetic studies would take about three years to execute at a most of $290 M.
The kinetics of decomposition of hydrocarbons odors and particulates has a
priority 1 and heteroatoms a priority 3- The total kinetic study would take about
four years at a total cost of $400 M.
22.1.3 Research Afterburner - Design, Construction and Operation
Plans A and B
Design procedures and equipment improvements cannot be developed until
engineering experiments are performed on a commercial scale piece of equipment.
These experiments will provide the data on efficiency as a function of the measured
fundamental parameters, reaction rate, temperature profiles, velocity profiles and
mixing rates.
The purpose of this project is to design and build a versatile experi-
mental afterburner. Provision for measurement of the internal variables like
temperature, velocity and concentration profiles is required. The ability to
change burner arrangements and combustion chamber geometry is to be incorporated
into the design. The size of the unit should be selected so that internal
conditions are representative of commercial equipment (a size in the range of
2 - 5 MM Btu firing rate would appear reasonable). The design of the unit should
be oriented towards convenience in obtaining the required data. Operation of
the unit without contaminant feed should be made first and data obtained on
temperature and velocity profiles. Also careful measurement of th* mixing rates
should be made. This provides the fundamental fluid mechanical information
necessary for the analysis of the performance efficiency experiments described in
subsequent sections. This project will require about one year for execution and
will cost about $250 M (includes $75 M capital experditure) and has a priority 1.
-------
316
22.1.j.l Experimental Study Hydrocarbon Feeds - Research Afterburner
Plan A
Experiments would be run with various hydrocarbon feeds over a range of
operating conditions in the reseaich afterburner described in Section 22.1.3-
Operating conditions to be investigated should be feed concentrations, feed
temperature, oxygen content, afterburner temperature, and feed flow rate. The
analytical techniques developed in Section 22.1.1 would be utilized. Particular
attention should be paid to measurement of concentration profiles throughout the
afterburner. This project should require about six months to execute for a
limited range of hydrocarbon at a cost of $70 M with a priority 1.
Plan B
This would be an extension of the number of hydrocarbons used to a broad
range and would require about six months to execute and a total cost about $88 M
(priority l).
22.1.3.2 Experimental Study Odorous Feeds - Research Afterburner
Plan A
This is a similar study to that described in 22.1.3.1 on a range of
odorous materials. The special features of odorous materials are that the feed
concentrations may be quite low and the required conversion to eliminate odor
may be quite high. Characterization of the concentration levels of specific
odorous materials that constitute a nuisance should be developed. Additional
analytical techniques may be needed. This project should require about six months
to execute for a limited number of odorous feeds at a cost of about $70 M
(priority l).
Plan B
This would be an extension of the number of odorous feeds investigated
and should require six months of effort at a total cost of $87 M (Priority 2).
22.1.3.3 Experimental Study Particulate Feeds - Research Afterburner
Plan A
A method of adding liquid sprays and solid participates should be added
to the research afterburner described in Section 22.1.3. Analytical apparatus for
sampling and particle size analysis should be added to the experimental equipment.
A limited range of feed materials would be injected into the afterburner feed at
various levels of dispersion. The volatility should range from easily vaporized
through low volatility heavy liquids up to non-volatile (but combustile) solids.
This project will require six months to execute and will cost about $70 M
(priority 2).
Plan B
This will be an extension to a broad range of feed compounds and will
require six months at a total cost of $87 M (priority 2).
S-lkl2I
-------
317
22.1.3.4 Experimental Study of Oil Firing - Research Afterburner
Plan B
The previous experimental work was all based on gac firing as the
supplemental fuel. This is the supplementary fuel that is coinvonly used in
existing afterburners, however, natural gas supplies are becoming limited and the
cost of natural gas will increase. Therefore, in the future, there will be
increasing use of oil firing in afterburners. A gun for oil firing will be added
to the research afterburner described in Section 22.1.3. Experiments will then
be conducted with representative feeds from the hydrocarbons odors and particulates
to determine the difference between oil and gas firing. This project could be
executed in a period of one year at about a cost of $18? M (priority 3).
22.1.3.3 Experimental Study Heteroatom Containing Feeds - Research Afterburner
Plan B
The afterburner effeciency will be measured for effluent streams
containing heteroatoms like chlorine, sulfur, phosphorous, which result in flame
inhibition, increased residence time requirements, corrosive conditions and
requirement for stack gas scrubbing. The research afterburner will need additional
equipment for stack gas scrubbing and for corrosion protection in the combustor.
The experiments would cover a range of feeds containing heteroatoms. The
operating problems will be quite difficult with these feeds so that a period of
about two years would be required at a cost of about $350 M (including $50 M
capital) with a priority 3«
S-HH21
-------
22.1.^- Research Afterburner - Gun and Control System Development
Plan A
A common problem in operation of afterburners is the variations in feed
rate and composition occurring in the process generating the effluents. This
causes instability in the burner operation, overheating of the equipment and erratic
performance. New burner designs and integrated control systems are needed to obtain
a much better turndown ratio than is currently achievable. An experimental attack
on this problem is proposed using the research afterburner described in Section
22.1.3- This project would require about one year for execution and would cost
about $75 M (includes 25 M capital) and is priority 1.
Plan B
This would be an expanded version of the above project with the objective
of achieving a larger turndown ratio. This project would require one year for exe-
cution and would cost a total of $150 M (includes $25 M capital) and is priority 1.
22.1.5 Mathematical Modelling of Afterburners
Plan A
The kinetic data obtained in Section 21.1.2 and the engineering data
obtained in Sections 21.1.3.1, 21.1.3.2, and 22.1.3-3 will be a reliable base upon
which to build a mathematical model. The model should include the interaction of
the chemical kinetics, and the physical processes of mixing and heat transfer.
The measured values of concentration, temperature and velocity profiles will provide
a proper test of the validity of the model. Existing methods of solving the fluid
mechanical problems can be utilized. A successful model development will provide
the basis for a design method which can be used to design cheaper, more
efficient afterburners. This initial plan will cover hydrocarbon, odor and
particulate feeds. This project can be executed in about 2 years at a cost
of $50 M and has a priority of 1.
Plan B
Extension of the model to include oil firing and heteroatom containing
feeds. This could be executed in four years at a total cost of $400 M and has a
priority of 1.
22.1.6 Field Implementation of New Designs
Plan B
Utilizing the design methods developed in Section 22.1.5 prepare designs
for two afterburners utilizing new ideas developed in the earlier work. These
units should have a lower cost, higher efficiency than units currently in use and
would hopefully be close to optimum design. These units should then be con-
structed and installed in a process environment. This project would cos>t about
$250 M (include $100 M capital) and would require about one year for execution
(priority of 3).
S-14121
-------
319
22.1.7 Field Test New Afterburners
Plan B
Field tests on the two new afterburners will determine the performance
under realistic industrial conditions. This will demonstrate Lh° applicability of
the new technology developed in this overall research program. This project
would cost about $150 M and could be executed in about one year and has a
priority 3-
22.2 Catalytic Afterburners
The research program here focusses on the fundamental kinetics sir,".e
it has been shown in the handbook that the engineering design methods are available
already.
22.2.1 Field Test Data
Plan A and B
The systems study showed a marked lack of quantitative performance data
so the objective of this project is to obtain this data using the analytical
techniques of Section 22.1.1. The good and bad features of the operating
catalytic afterburners can then be properly assessed with the data. Correlation
and analysis of this performance data can be made on the basis of the design
methods discussed in the handbook. This project could be executed in one year at
a cost of about $75 M and has priority 1.
22.2.2 Kinetics of Catalytic Oxidation
Plan A
The objective is to measure the chemical rate of oxidation for typical
catalyst and waste streams. The handbook shows that mass transfer and chemica]
kinetics interact closely in the performance of catalytic units. A more ratjonal
design can be made if the fundamental kinetic rate of oxidation is known. A
laboratory reactor should be built to measure the rates of oxidation over
selected catalysts commercially available for selected feed materials. The
laboratory reactor must be designed such that the chemical rate is limiting rather
than the mass transfer rate. Kinetic correlations should be developed relating
the oxidation rate to contaminant concentration, oxygen concentration and
temperature. This project could be executed in one year at a COP^ of $50 M and
has a priority 1.
Plan B
Extension to a wider range of commercial catalysts and feed materials.
Execution in one year at a total cost of $100 M with a priority of 1.
S-14121
-------
320
22.2.3 Catalyst Poisoning and Deactivation Rates
Plan A
Catalyst life is an important factor in catalytic afterburner costs. A
laboratory reactor will be built for purposes of measuring catalyst life. This
should be an automated, unattended apparatus so that long run times can conveniently
be obtained. These life tests would use feeds studied in Section 22.2.2. In
addition, the poisoning rates of trace materials commonly found in effluent streams
should also be determined for commercially available catalysts. While these trace
materials may not constitute an effluent requiring control they will have to be
removed prior to catalytic oxidation if they poison the catalyst. This project
could be executed in about one year at a cost of $50 M and has a priority of 1.
Plan B
Extend range of possible catalyst poisons, feed streams and commercial
catalysts. Execute in one year at a total cost of ;|;100 M with a priority 1.
22.2.U New Catalyst Development
Plan B
The cost of the noble metal catalysts is a significant portion of the
total cost of a catalytic unit. Noble metal catalysts are also poisoned rapidly
by certain contaminants. The objective of this project is to find base metal
catalysts that are lower cost and are more resistant to poisoning. These measure-
ments should be made in the laboratory apparatus described in Section 22.2.2.
The research effort in automobile exhaust catalysts should provide some catalysts
that may be applicable to stationary afterburners. This project could be executed
in one year at a cost of about $150 M and has a priority 2.
S-14121
-------
Ol
f\)
Table 22-1. RESEARCH PROGRAM, PLAN A, $1 MM
Project
Thermal
22.1 Field test data
22.1.2 Kinetics of combustion
22.1.3 Research afterburner - Design, construction
and operation
22.1.3.1 Experiments, hydrocarbon feeds
22.1.3.2 Experiments, odorous feeds
22.1.3.3 Experiments, particulate feeds
22.1 A Research afterburner - Gun and control system
development
22.1.5 Mathematical modelling of afterburner
Catalytic
22.2.1 Field test data
22.2.2 Kinetics of catalytic oxidation
22.2.3 Catalyst poisoning and deactivation rates
Cost per year
Total cost
Cost
$ M
130
110
250
70
70
70
75
50
75
50
50
1000
Year
1
-^
180
2
->
235
3
~*
250
k
->
->
~*
2kO
5
->•
^
95
Priority
1
1
1
1
2
2
1
1
1
1
1
to
H
-------
Table 22-2. RESEARCH PROGRAM, PLAN B,
MM
ro
ro
Project
Thermal
22.1 Field test data
22.1.2 Kinetics of combustion (hydrocarbons, odor
particles)
Kinetics of combustion (heteroatoms)
22.1.3 Research afterburner - Design, construction
and operation
22.1.3.1 Experiments, hydrocarbon feeds
22.1.3-2 Experiments, odorous feeds
22.1.3.3 Experiments, particulate feeds
22.1.3.^ Experiments, oil firing
22.1.3.5 Experiments, heteroatoms
22.1.U Research afterburner - gun and control system
development
22.1.5 Mathematical modelling of afterburner
22.1.6 Field implementation of new designs
22.1-7 Field test new afterburners
Catalytic
22.2.1 Field test data
22.2.2 Kinetics of catalytic oxidation
22.2.3 Catalyst poisoning and deactivation rates
22. 2. k New catalyst development
Cost per year
Total cost
Cost
$ M
175
200
200
250
88
88
87
187
350
150
koo
250
150
75
100
100
150
3000
Year
1
2
— >•
613
->
->
— >
562
3
-*
-
-*
675
k
->
750
5
->
^00
Priority
1
1
3
1
1
2
2
3
3
1
1
3
3
1
1
1
2
-------
323
BIBLIOGRAPHY AND REFERENCES
3-l) Danielson, J. A. (ed.), Air Pollution Engineering Manual, U.S. Public
Health Service Publication 999-AP-W U9&7J.U revised edition is
scheduled for publication.)
3-2) Canovali, L. L., Sida, S., "Case History of Selection and Installation
of a Kraft Recovery Odor Reduction System", TAPPI 53, p. 1^88 (August
1970). ~~
3-3) Heilman, W., Welling, J. C., "Odor Control in Rendering by Direct-Flame
Incineration", Paper at 62nd Air Poll. Cont. Assoc. National Meeting
(June 22 - 26, 1969).
3-*0 Mills, J., et al, "Design of Afterburners for Varnish Cookers", J. Air
Poll. Cont. Assoc. 10, pp. 161 - 168 (April 1960).
3-5) Sandomirsky, A. G., et al, "Fume Control in Rubber Processing by Direct
Flame Incineration", Paper at 59th Air Poll. Cont. Assoc. National
Meeting (June 20 - 2k, 1966).
3-6) Sullivan, J. L., et al, "An Evaluation of Catalytic and Direct Fired
Afterburners for Coffee and Chicory Roasting Odors", J. Air Poll. Cont.
Assoc. _15, p. 583 - 586 (1965).
3-7) Waid, D. E., "Incineration of Organic Material by Direct Gas Flame for
Air Pollution Control", Am. Ind. Hyg. Assn. J. 30, 291 - 7 (1969).
3-8) Essenhigh, R. H., "Discussion ASME National Incinerator Conference",
p. 96 (1970).
3-9) Mills, J. L. et al, Emissions of Oxides of Nitrogen from Stationary
Sources in Los Angeles County; Report No. 2 Oxides of Nitrogen Emitted
by Small Sources, Sept., I960] Report No. 4 Final Report, July, 1961,
Los Angeles County Air Pollution Control District.
3-10) Cosden, W. B.; Technical Committee, National Coil Coaters Association;
private communication, January, 1972.
3-ll) Martin, G. B., and Berkau, E. E. (EPA), paper presented at AIChE
Meeting, August 30, 1971, Atlantic City, New Jersey.
3-12) Husain, D. and Norrish, R. G. W., Proc. Royal Soc. (A), 273, 1^5, 1963.
4-l) Acres, G. J. K., "Platinum Catalysts for the Control of Air Pollution:
The Elimination of Organic Fume by Catalytic Combustion", Platinum
Metals Review, Ik, No. 1, pp. 2-10 (January 1970).
k-2) Bolduc, M. J., Severs, R. K., and Brewer, G. L., "Test Procedures for
Evaluation of Industrial Fume Converters", Air Engineering, pp. 20 - 23
(February 1966).
S- Ik 121
-------
324
HIHLIOGUAP11Y AMD INFERENCES (Con Id)
4-3) Franzky, V., "Emission Measurements on Drying Ovens and Jellying Channels
with Secondary Waste-Gas Purifying Plants for Odor Abatement", Staub-
Reinheit-Luft, 29, No. 1, pp. 33-41 (January 1969).
4-4) Hardison, L. C., "Gaseous Waste Disposal", Industrial Gas, 1968,
pp. 16 - 23 (July, 1968).
4-5) Hardison, L. C., "A Summary of the Use of Catalysts for Stationary
Emission Source Control", Proc. 1st Nat'l. Symp. on Heterogeneous
Catalysis for Control of Air Pollution, November 21 - 22, 1968,
Philadelphia, Pennsylvania, pp. 271 - 296, published by U.S. Dept. of
HEW, Nat'l. Air Pollution Control Admin.
4-6) Houdry, J. H. and Hayes, C. T., "Versatility of Oxidation Catalysts for
Industrial Air Pollution Control", J. Air Poll. Cont. Assoc., $>, No. 3,
182 - 6 (195T).
4-7) Lunche, R. G., "Fume and Odor Destruction by Catalytic Afterburners",
Proc. 1st Nat'l. Symp. on Heterogeneous Catalysis for Control of Air
Pollution, November 21 - 22, 1968, Philadelphia, Pennsylvania, published
by U.S. Dept. of HEW, Nat'l. Air Pollution Control Admin., pp. 297 - 320.
4-8) Miller, M. R. and Sowards, D. M., "Solvent Fume Abatement by Ceramic
Honeycomb Catalyst Systems", ibid., pp. 321 - 348.
4-9) Miller, M. R. and Wilhoyte, H. J., "A Study of Catalyst Support Systems
for Fume-Abatement of Hydrocarbon Solvents", J. Air Poll. Cont. Assoc.
17, No. 12, 791 - 5 (December 1967).
4-10) Romeo, P. L. and Warsh, A., "Catalytic Incineration Design Parameters
and Operating Practices", in Combustion Evaluation - Sources and Control
Devices, U.S. PHS Training Manual, (February 1967).
4-ll) Ruff, R. J., "Catalytic Fume Combustion of Organic Chemical and
Petrochemical Wastes", Industrial Wastes, 2, No. 3, PP- 67 - 70
(May - June, 1957).
4-12) Suter, H. R., "Range of Applicability of Catalytic Fume Burners", J. Air
Poll. Cont. Assoc. J?, No. 3, 173-184 (1955)-
4-13) Thomiades, L., "Why Catalytic Incineration?", Pollution Engineering,
(May/June 1971) pp. 32 - 33.
4-14) Volheim, G., "The Catalytic Afterburning of Industrial Effluents",
Staub-Reinheit-Luft 25_ in English, pp. 20 - 26 (November 1969).
4-15) Werner, K. D., "Catalytic Oxidation of Industrial Waste Gases", Chemical
Engineering, November 4, 1968, pp. 179 - 84.
4-16) Brewer, G. L., "Odor Control for Kettle Cooking", J. Air Poll. Cont.
Assoc. 13, No. 4, 167 - 169 (1963).
S-14121
-------
325
BIBLIOGRAPHY AND REFERENCES (Contd)
4-17) Fawcett, R. L., "Air Pollution Potential of Phthalic Anhydride
Manufacture", J. Air Poll. Cont. Assoc. 20, No. 1, 46l - 5 (1970).
4-18) Gadomski, R. R., David, M. P., and Blahut, G. A., Evaluation of Emissions
and Control Technologies in the Graphic Arts Industries, Final Report
under Contract No. CPA 22-69-72, U.S. Dept. HEW, NAPCA (August 1970).
4-19) Horn, J. and Warner, J. I., Jr., "Union Carbide Solves Ethylene Emission
Problem with Fume Abater", Petro/Chem Engineer, 37 - 40 (February 1970).
4-20) Stresen Reuter, J., "Catalytic Incinerator Controls Hydrocarbons-end
Odors", Plant Engineering, April 17, 1969, p. 142.
4-21) , "Oxidation System Solves Odor Pollution", Farm
Chemicals (May 1968;.
5-l) Danielson, J. (ed.), Air Pollution Engineering Manual (see Ref. 3-l)«
7-1) Norton, F. H., Refractories, (4th edition) 450 p. McGraw-Hill (1968).
7-2) Johnson, R. C., "Construction Techniques for Refractories and Insulation",
Chem. Eng. Progress, 66, No. 8, pp. 40 - 42 (1970).
7-3) Wallace, R. W. and Criss, G. H-, "Thermal Conductivity of Castable
Refractories in Relation to Bulk Density", Bull. Am. Ceram. Soc., kl,
No. 2, pp. 176 - 179 (1968).
7-4) Wygant, J. F. and Crow ley, M. S., "Designing Monolithic Refractory
Vessel Linings", Bull. Am. Ceram. Soc., 43, No. 3, pp. 173 - 182 (1964).
7-5) Deeson, A. F. L. and Deeson, E., "Thermal Insulation in the Chemical
Industry", Brit. Chem. Eng., 15_, No. 5, PP- 621 - 626 (1970).
7-6) Smoot, T. W. and Cobaugh, G. D., "Monolithic Refractories for Process
Equipment", Chem. Eng., pp. 105 - 110 (August 1965).
8-l) Committee on Industrial Ventilation, Industrial Ventilation, Am. Conf.
of Gov. Ind. Hygienists, tenth edition, 1968.
8-2) Danielson, J. A., Air Pollution Engineering Manual ('-ee Ref. 3-1 )•
8-3) Perry, J. H., Chemical Engineers' Handbook, (4th edition) Chapt. 20,
published by McGraw-Hill Book Company U963).
8-4) Brink, J. A., and Contant, C. E., "Experiments on an Industrial Venturi
Scrubber", Industrial and Engng. Chemistry 50, 1157 (August 1958).
8-5) Gilbert, N., Wet Collectors, Engng. Report prepared for American
Petroleum Institute, Dept. of Chem. Engng., Univ. of Cincinnati (lfi59).
S-14121
-------
326
BIBLIOGRAPHY AND REFERENCES (Contd)
8-6) Doyle, H., and Brooks, A. F., "The Doyle Scrubber", Industrial and Engng.
Chem. 49, No. 12, p. 57A (December 1957).
8-7) Harris, L. S., "Fume Scrubbing with the Ejector Venturi System", Chem.
Engng. Process 62, No. 4, p. 55 (April 1966).
8-8) Gieseke, J. A., "Pressure Loss and Dust Collection in Venturi Scrubbers",
PhD dissertation Dept. of Chem. Engng. Univ. of Washington (196^).
8-9) Calvert, S., "Venturi and Other Atomizing Scrubbers Efficiency and
Pressure Drop", AIChE Journal 16, No. 3, p. 392 (May 1970).
8-10) U.S. Public Health Service Publication, Control Techniques for
Particulates Air Pollutants, NAPCA Publication No. AP-51 (January 1969).
8-11) Brink, J. A., "Air Pollution Control with Fibre Mist Eliminators",
Canadian J. Chem. Engng. kl, p. 131*- (June 1963).
8-12) Gas Purification Processes, Chapter 15, "Mist Removal" by Lewrie Fairs.
8-13) Kaiser Engineers, Cyclone Dust Collectors, Engineering report prepared
for the American Petroleum Institute (1955 ) •
lO-l) American Society for Testing and Materials, Annual Book of Standards,
Part 23, Water, Atmospheric Analysis (l97l)- Reprints of specific
methods are available from ASTM, 1916 Race Street, Philadelphia, PA 19103.
10-2) Lundgren, D. and Calvert, S., Am. Ind. Hygienic Assn. J., 28, 208 (1967).
10-3) Dietz, W. A., J. Gas Chromatography 5>, 68 (1967). "Response Factors
for Gas Chromatographic Analysis".
10-U) Weiss, F. T., Determination of Organic Compounds, J. Wiley and Sons,
New York (1970T
a- P- 397- The Flame lonization Detector.
b. p. 100. Determination of Aldehydes.
c. p. 91. Determination of Trace Carbonyl.
d. p. 104. Determination of Formaldehyde.
e. p. 120. Colorimetric Determination of Esters.
10-5) Sawicki, E., et al, Anal. Chem. 33, 93 (l96l). Colorimetric Determin-
ation of Aldehydes.
10-6) Lappin, G. R. and Clark, L. C., Anal. Chem. 23, 5kl (1951). Colorimetric
Method for Carbonyl Compounds.
10-7) .Cohen, I. R. and Altshuler, A. P., Anal. Chem. 33, 726 (1961).
Determination of Ac role in.
S-14121
-------
10-8)
10-9)
10-10)
10-11)
10-12)
10-15)
10-14)
10-15)
10-16)
10-17)
BIBLIOGRAPHY AND REFERENCES (Contd)
Selected Methods for the Measurements of Air Pollutants, Public Health,
Service Publication 999-AP-ll. Robert A. Taft Sanitary Engineering
Center, Cincinnatti, Ohio (1965).
Stern, A. C. (ed.), Air Pollution, Vol II, Academic Press, New York
(1968). Sampling Procedure Outline for Some Specific Contaminants,
p. 520.
Goddu, R. F. et al, Anal. Chem., 27, 1251 (1955). Spectrophotometrio
Determination of Esters.
EPA-APCO Method for C02.
Hanson, N. W., Reilly, D. A., Stagg, H. E., Editors, The Determination
of Toxic Substances in Air, W. Heffer and Sons, Cambridge U965J.
a. p. 91- Measurement of Organic Chlorides.
b. p. 107- Method for Chlorobenzene.
c. p. 137- Method for Hydrogen Chloride.
Source Testing Manual, Air Pollution Control District, Los Angeles
County, California (1965).
Federal Register 36, (159), 15703 - 15722 (August 17, 1971).
Society of Automotive Engineers, Tech. Rep. J177, Determination of NOX
in Diesel Exhaust.
West, P. W., and Gaeke, G. C., Anal. Chem. 28, l8l6 (1956).
O'Keefe, A. E. and Orman, G. C., Anal. Chem. 3_8, 760 (1966). Use of
Permeation Tubes for Calibrating Atmospheric Analyzers.
10-18) Sources of Permeation tubes:
Reliable tubes may be purchased from Metronics Associates, Inc.,
3201 Porter Drive, Stanford Industrial Park, Palo Alto, California
94304. Office of Standard Reference Materials, National Bureau of
Standards, Washington, D. C. 20234 will supply a calibrated and
certified SOa tube; order SRM 1625. Other sources are Polyscience
Company, P. 0. Box 65, Kenilworth, Illinois 60043 8'xi Analytical
Instrument Development Company, West Chester, Pennsylvania 19380.
10-19) Cooper, H. B. H. and Rossano, A. T., Source Testing for Pollution
Control, Environmental Research and Applications, Inc., Wilton,
Connecticut 06897.
10-20) Smith, Walter S., et al, "Stack Gas Sampling Improved and Simplified
with New Equipment", Paper 67 - 119, Annual Meeting of Air Poll. Cont.
Assoc. Cleveland, Ohio (June 13, 1967) (see Reference 19, p. 106).
S-14121
-------
328
BIBLIOGRAPHY AND REFERENCES (Contd)
10-21) British Standard 2811:1957- British Standards Institution, British
Standards House, 2 Park Street, London, W 1.
10-22) Devorkin, H. et al, Source Testing Manual, edited by R. G. Holmes,
Los Angeles County Air Pollution Control District, 434 So. San Pedro
Street, Los Angeles, California (1963).
10-23) Federal Register 36, 8191 (April 30, 1971). Calibration Using
Permeation Tubes.
10-24) MacPhee, R. D. and Kuramoto, Mutsuo, "Recommended Test Methods 1'or
Organic Solvents and Vapors, (Rule 66)" Los Angeles Air Pollution
Control District (April 1968).
10-25) Jacobs, M. B-, The Chemical Analysis of Air Pollutants, Interscience
Publishers, Inc., New York U960J.
10-26) Leithe, W., The Analysis of Air Pollutants, Ann Arbor-Humphrey Science
Publishers, Ann Arbor (1970).
10-27) Ruch, W. E., Quantitative Analysis of Gaseous Pollutants, Ann Arbor-
Humphrey Science Publishers,Ann Arbor (1970).
10-28) Byrd, J. F., and Phelps, A. H., in Air Pollution, Vol. II, 2nd edition,
ed. A. C. Stern, Academic Press, New York (1968), pp. 305 - 327-
10-29) ASTM Designation D-1391-57 (Reapparoved 1967).
10-30) Benforado, D. M., Rotella, W. J., and Horton, D. L., J. Air Poll. Cont.
Assoc. 1-2, 101 (1969).
10-31) Los Angeles Air Pollution Control District, Source Testing Manual - 1967*
Chapter 6 - Odor Measurement.
10-32) Danielson, J. A. (ed.), Air Pollution Engineering Manual, (see Ref. 3-1).
Appendix G - Odor Testing Technique, pp. 861 - 864.
10-33) Mills, J. L., Walsh, R. T., Luedtke, K. D., and Smith, L. K., J. Air
Poll. Cont. Assoc. 13, 467 (1963).
10-34) Air and Industrial Hygiene Laboratory, State of California Department of
Public Health, Presented at the 12th Conference on Methods in Air
Pollution and Industrial Hygiene Studies, Univ. of Southern California,
April 6, 7, and 8, 1971.
10-35) Cederlof, R., Edfors, M. L-, Friberg and Lindvall, T., J. Air °oll. Cont.
Assoc. 16, 92 (1966).
10-36) Hemeon, W. C. L-, J. Air Poll. Cont. Assoc. 18, 166 (1968).
S-14121
-------
329
BIBLIOGRAPHY AND REFERENCES (Contd)
10-37) Turk, A., in Air Pollution Control - Guidebook for Management, ed.
A. T. Rossano, Environmental Science Service Division, Stamford,
Connecticut, pp. 115 - 127 (1969).
10-38) Nader, J. S., J. Am. Ind. Hyg. Assn., 19> 1 (1958).
10-39) Adams, D. F., Young, F. A., and Luhr, R. A., TAPPI, £1, 62A (1968).
10-4o) Sullivan, D. C-, Adams, D. F., and Young, F. A., Atmospheric Environment
2, 121 (1968).
10-4l) Huey, N. A., Broering, L. C., Jutze, G. A., and Gruber, C- W., J. *:r
Poll. Cont. Assoc. 10, 44l (1960).
10-1*2) Stockham, J. O'Donnell, A., and Dravnieks, A., Illinois Institute of
Technology Research Institute, Report No. IITRI C8150-5 to Coordinating
Research Council, Chemical Species in Engine Exhaust and their
Contribution to Exhaust Odor, (June 30, 1969)'
10-43) Public Health Service Publication No. 999-AP-X5, U.S. Dept. of HEW,
p. 85 (1965). Method for S03.
10-44) Goksoyr, H. and Ross, K., J. Inst. Fuel, 35_, 177 (1962). Method for S03.
10-45) Benedict, Robert P., Fundamentals of Temperature, Pressure, and Flow
Measurements, John Wiley and Sons,Inc., (1966).
10-46) Fristrom, R. M. and Westenberg, A. A., Flame Structure, McGraw-Hill
Book Co., (1965).
10-47) Blackshear, P. L., NACA Sonic-Flow-Orifice Temperature Probe in High-
Gas-Temperature Measurement, ASME Transactions, 75, 51 - 58 (1953)-
10-48) Marks, Mechanical Engineer's Handbook, 6th edition.
10-49) Beer, J. M. (Univ. of Sheffield, England), "Modelling of Con.bustion
System", a paper presented at the New Developments in Combustion
Engineering Seminar, Pennsylvania State University, (July 26 - 30, 1970).
10-50) Clarke, A. E., Gerrard, A. J., and Holliday, L. A., "Some Experiences In
Gas Turbine Combustion Chamber Practice Using Water Flow Visualization
Technicques", 9th Symposium (international) on Combustion, pp. 878 - 891
(1963).
10-51) Dean, Robert C., Aerodynamic Measurements, Gas Turbine Laboratory, MIT
(1953).
10-52) Gebhart, B. and Hollasch, K., "Calibration of Constant Temperature Hot
Wire Anemometers at Low Velocities in Water with Variable Fluid
Temperature", Paper No. AIChE 17, Cornell University.
S-14121
-------
330
BIBLIOGRAPHY AND REFERENCES (Contd)
10-53)
10-510
10-55)
10-56)
10-57)
10-58)
10-59)
11- l)
12-l)
12-2)
12-3)
I2-k)
12-5)
12-6)
12-7)
12-8)
Leving, Donald L., and Katzoff, S., "The Fluorescent-Oil Film Method
and other Techniques for Boundary- Layer Flow Visualization", NASA
Memo 3-1T-59L.
ASME, Symposium on Flow Visualization, Presentation Summaries (November
30, I960).
Clayton, B. R. and Massey, B. S., "Flow Visualization in Water: A
Review of Techniques", J. Sci, Instrum, kk, p. 2 - 10 (1967).
Hazen, D. C. and Lehnert, R. F., Smoke Flow Studies Conducted at
Princeton University, Report No. 290, Contract N6 onr 27016 U955?).
Hemey, James and Davis, Wendall (Delaval Turbine, Inc. and Univ. of
Connecticut, respectively), "Further Contributions to Velocity
Determination Using the Hydrogen Bubble Technique", presented at the
Flow Symposium (May 10 - Ik, 1971 ), Joint Technical Session with ASME
Fluids Engineering Division.
Etzold, F., "Measurement Techniques For Near Wake Flows", Paper No.
5-2-50 Flow Symposium (May 10 - Ik, 197l), Joint Technical Session with
ASME Fluids Engineering Division.
Roschke, E. J., "Flow Visualization Studies of a Confined Jet Driven
Water Vortex", Jet Propulsion Laboratory Technical Report No. 32-100^.
Hougen, 0. A., et al, Chemical Process Principles, Part I, 2nd edition,
Wiley and Sons, Inc., p. 259
Spaulding, D. B., Some Fundamentals of Combustion, Academic Press (1955)-
Bradley, J. N., Flame and Combustion Phenomena, Methuen (1969).
Minkoff, G. J. and Tipper, C. F. H-, Chemistry of Combustion Reactions,
Butterworths (1962).
Vulis, L. A., Thermal Regimes of Combustion, McGraw-Hill (1961).
Yuster, S. T., et al, Afterburner Studies as Applied to Automobile
Exhaust Systems, U.C-L.A. Report 58-55 I June 1958;.
Hilado, C. J., "How to Predict if Materials Will Bum", Chem. Eng.,
p. 17^ (December Ik, 1970 ).
Loncwell, J. P., and Weiss, M. A., "High Temperature Reaction Rates in
Hydrocarbon Combustion", Ind. and Eng. Chem. kj, p. I6}k - l6k$ (l955)«
Koslov, G. I., "On High Temperature Oxidation of Methane", 7th Symposium
(international) on Combustion, p. 1^2 (1958).
S-14121
-------
BIBLIOGRAPHY AND REFERENCES (Contd)
12-9) Surface Combustion, "A New Approach to Industrial Air Pollutiontoon-j
Heating and Ventilating Engineer and Journal of Air Conditioning,
pp. 80 - 85, (August 19TO).
12-10) Morgan, A. C-, "Combustion of Methane in a Jet Mixed Reactor", D. Sc;
Thesis, MIT (1967)-
12-ll) Fristrom, R. M. and V/estenberg, A. A., Flame Structure, McC^aw-Jlill,
1965.
12-12) Field, M. A., et al, Cor.-.bustion of Pulverized Coal, Chapters 5 - (.,.
British Coal Utilization Research Assoc. (1967).
12-13) Williams, A., "The Mechanism of Combustion of Droplets and
Liquid Fuels", Oxidation and Combustion Reviews 3_, p. 1 -
12-Ik) U.S. National Air Pollution Control Administration, Control Techniques
for Particulate Air Pollution, Publication AP-51, p. 150 U969J-
12-15) Niessen, W. R., et al, Systems Study of Air Pollution from Municipal
Incineration, Clearinghouse for Federal Scientific and Technical
Information, (1970).
12-16) Lee, K. B., et al, "On the Rate of Combustion of Soot in a Laminar Soot
Flame", Combustion and Flame 6, p. 137 - 1^5 (1962).
12-17) Fenimore, C. P., and Jones, G. W., "Oxidation of Soot by Hydroxyl
Radicals", J. Phys. Chem. 71, pp. 593 - 597 (1967).
12-18) Magnussen, B. F., "The Rate of Combustion of Soot in Turbulent Flames",
15th Symposium (international) on Combustion (1970).
12-19) Walls, J. R., and Stickland-Constable, R. F., "Oxidation of Carbon
Between 1000 - 2^00°C", Carbon I, p. 333 (l96U).
12-20) Creitz, E. C., "A Literature Survey of the Chemistry of Flame
Inhibition", J. of Research of Nat. Bureau Standards 7j*A, p. 521
(July 1970).
12-2l) Rosser, W. A., et al, "Mechanism of Combustion Inhibit-5 _>n by Compounds
Containing Halogen", 7th International Symposium on Combustion, p. 175
(1959).
12-22) Fenimore, C- P., and Jones, G. W., "Flame Inhibition by Methyl Bromide",
Combustion and Flame J, p. 323 (1963).
12-23) Edmondson, H., and Heap, M. P., "The Burning Velocity of Methane-Air
Flames Inhibited by Methyl Bromide", Combustion and Flame 13, p.
(1969). ~~
S-
-------
BIBLIOGRAPHY AND REFERENCES (Contd)
Kaesche-Krisher, B., "Flame Speed and Burning Mechanisms of Chlorinated
Hydrocarbons", Chemie Ing. Technik 55., p. 856 (1963).
25) Hochstim, A. R. (ed.), Bibliography of Chemical Kinetics and Collision
Processes, IFI/Plenum U969).
-26) Brokaw, R. S. and Bittkerj D. A., "Carbon Monoxide Oxidation Rates
Computed for Automobile Exhaust Manifold Reactor Conditions", NASA
Technical Note TN-D-7024 (1970).
Avery, W. H., and Hart, R. W., "Combustor Performance with Instantaneous
Nixing", I and EC 45, p. 1634 (1953).
&M Chemical Company, JANAF Thermochemical Tables (1964).
Streeter, V. L., Handbook of Fluid Dynamics, McGraw-Hill, Chapter 10
"Turbulence",
lh.4-2) Forstall, W. and Shapiro, A. H., "Momentum and Mass Transfer in Coaxial
Gas Jets", Trans. ASME, Journ. Appl. Mech., pp. 399 - 408 (December
1950).
14-3) Thring, M. W-, and Newby, M. P., "Combustion Length of Enclosed Turbulent
Jet Flames", 4th Symposium (international) on Combustion, pp. 789 - 796
(1953).
14-4) Chigier, N. A. and Beer, J. M., "The Flow Region Near the Nozzle in
Double Concentric Jets", ASME Journal of Basic Engineering (December
1964), pp. 797 - 804.
14-5) Curtet, R. and Ricou, F. P., "On the Tendency of Self Preservation in
Axisymmetric Ducted Jets", Journal of Basic Engineering, ASME
(December 1964), pp. 765 - 776.
14-6) Beek, J. and Miller, R. S., "Turbulent Transport in Chemical Reactors".
Chem. Eng. Prog. Symp. Series, 55, No. 25, pp. 23 - 28.
14-7) Patrick, M. A., "Experimental Investigation of Mixing and Flow in a
Round Turbulent Jet Injected Perpendicularly into a Main Stream",
Inst. of Fuel Journal, 40, pp. 425 - 432 (September 1967).
14-8) Ruggieri, R. S. and Callaghan, E. E., "Penetration of Air Jets Issuing
from Circular, Square, and Elliptical Orifices Directed Perpendicularly
to an Air Stream", NACA TN 2019 (February 1950).
14-9) Baines, W. D. and Peterson, E. G., "An Investigation of Flow Through
Screens", Trans. ASME (July 1951 ).
14-10) Norris, R., Petersen, E. E., and Prausnitz, J. M., "Mixing and Chemical
Reaction in Turbulent Flow Reactors", AIChE Journal, 11, No. 2,
pp. 221 - 227.
S- 14121
-------
BIBLIOGRAPHY AND REFERENCES (Contd)
14-11) Kanbour, S. I., "An Investigation of Thermal Mixing of Stratified
14-12)
14-13)
14-14)
14-15)
14-16)
14-17)
14-18)
14-19)
14-20)
14-21)
15-1)
15-2)
Streams Resulting from the Introduction of a Rigid Barrie* Nor^B to*
the Flow", PhD Thesis, University of Maryland, Department^f
Enginee ring, (1970).
Priddy, M. H., "Modern Approach to Industrial Air PoTl»t.i«n-rfcMfrni'
Midland Ross Co., Toledo, Ohio, Sales Brochure.
Weske, J. R., "Experimental Investigation of Velocity
Downstream of Single Duct Bends", NACA TN 1471 (January 193
Rao, S. T. R. and Essenhigh, R. H., "Experimental Determinatioroi
Stirring Factors Generated by Straight aid Swirling Jets l^j Jsothftrinal
Combustion Models", 13th Symposium (International) on • Combustion
pp. 549 - 561. '
Kastner, L. J. and Spooner, J. R., "An Investigation of the Performance
and Design of the Air Ejector Employing Low Pressure Air as the
Fluid", Inst. of Mech. Engrs., 160, No. 2, pp. 149 - 166, pp. 46JT
480 (1950).
Chigier, N. A. and Chervinsky, A., "Experimental Investigation ,
Swirling Vortex Motion in Jets", Journ. Appl. Mech. (June 1967),
pp. 443-451.
Gossman, A. D., Pun, W. M., Runchal, A. K., Spaulding, D. B., and
Wolfshtein, M., Heat and Mass Transfer in Recirculating Flows,
Chapter 4, Academic Press
Mathur, M. L. and MacCallum, N. R. L., "Swirling Air Jets Issuing from
Vane Swirlers", Journal Institute of Fuel, (1967), "Pt. 1, Free Jets"
(May 1967), pp. 214 - 225, "Pt. 2 Enclosed Jets", (June 1967),
pp. 238 - 245.
Ramsey, J. W., and Goldstein, R. J., "Interaction of a Heated Jet with
a Deflecting Stream", ASME Journal of Heat Transfer, Nov. 19(1,
pp. 365 - 72.
Flatten, J. L. and Keffer, J. F., "Deflected Turbulent Jet Flows",
Journal of Applied Mech., Trans. ASME, Dec. 1971, pp. 75^ -"8.
Kerr, N. M., and Frazer, D., "Swirl, Pt. 1: Effect on Axisymmetrical
Turbulent Jets", Jour. Inst. Fuel, 38, Dec. 1965, pp. 519 - 26.
Sovren, G. (ed.), Fluid Mech. of Int. Flow, Elsevier, art. by J. Ackeret,
p. 14 (1967).
Baines, W. D. and Peterson, E. G-, "An Investigation of Flow Through
Screens", Trans. ASME, Vol. 73, pp. 467 - 480 (July 195l).
S-14121
-------
f
BIBLIOGRAPHY AND REFERENCES (Contd)
g, M. W. and Newtay, M. P., "Combustion Length of Enclosed Turbulent
Jet Flames", 4th International Symposium on Combustion, pp. 789 - 796
skestad, G., "Two Dimensional Miter-Bend Flow", Jour, of Basi-~ Engrg,
" Sept. 1971, pp. *O3 - *O-
, J., "Design of Packed Catalytic Reactors" in Advances in Chemical
i i ii'l Vol. 3, Academic Press, New York, pp. 203 - 271 (1962).
s, W. M. and London, A. L-, Compact Heat Exchangers, 2nd edition,
~'3 aw- Hill Book Co., New York
m, A. L., Young, M. B. 0-, and Stang, J. H., "Glass-Ceramic
aces, Straight Triangular Passages - Heat Transfer and Flow Friction
.racteristics;, TASME, J. Engr. for Power, pp. 38l - 9 (October 1970).
4cAdams, W. H., Heat Transmission, 3rd edition, McGraw-Hill Book Co.,
York, pp. 162 - 3, 265, 271 - 5 (1954).
Satterfield, C- N., Mass Transfer in Heterogeneous Catalysis, MIT Press,
Cambridge Massachusetts (1970).
Satterfield, C. N. and Cortez, D. H., "Mass Transfer Characteristics of
oven-Wire Screen Catalysts", Ind. Eng. Chem. Fund., 9_, No. 4, 613 - 20,
1970).
16-7) Wheeler, A., "Reaction Rates and Selectivity in Catalyst Pores", in
Advances in Catalysis, Vol. 3, Academic Press, New York, pp. 249 - 327
(1951).
l8-l) Kays, W. M. and London, A. L., Compact Heat Exchangers, 2nd edition,
McGraw-Hill, 1964.
18-2) McAdams, W. H., Heat Transmission, 3rd edition, McGraw-Hill, 1954.
18-3) , Handbook of Fundamentals, Ainer. Soc. of Heating,
Refrigerating, and Air Conditioning Engineers (ASHRACE), New York, 1967.
18-4) Bell, K. J., Final Report of the Cooperative Research Program on Shell-
and-Tube Heat Exchangers, Univ. of Delaware Engrg. Expt. Station,
Bull. No. 5.
21-l) McGraw, M. J. and Duorey, R. L., Compilation of Air Pollutant Emission
Factors - Preliminary Document, Washington, EPA (1971).
21-2) Southcrland, Jtm.tJ H., private communication, August 3; 1971-
21-3) U.S. Bureau of Census, Current Population Reports, series P-25, No. Mi-8
(1970).
S-14121
-------
BIBLIOGRAPHY AND REFERENCES (Contd)
21-4) Landsberge, H. H., et al, Resources in America's Futui
Press, Baltimore (1963).
21-5) Kahn, H. and Wiener, A. J., The Year 2000, Macmilliar., Ne
21-6) Anon., Predicasts, issue No. 40, 7 (1970).
21-7) Anon., Threshold Limit Values of Airborn Contaminants ani
Changes, 1970» American Council of Governmental Industri
Cincinnati U970).
21-8) Copley International Corporation, National Survey of the
A Study of the Social and Economic Impact of Odors, Phase
APTD-05BO-1, NAPCA, Washington U970;.
21-9) Altshuller, A. P., "An Evaluation of Techniques for the Detj
Photochemical Reactivity of Organic Emissions", J. Air Poll
16 (5), 257 (1966).
21-10) Laity, J. L. et al, "Photochemical Smog and the Atmospheric Re
Solvents", presented at ACS l62nd National Meeting, Washing*
(1971).
21-11) Brunelle, M. F. et al, Effectiveness of Organic Solvents in Photoi
Smog Formation, Los Angeles Air Pollution Control District, Los
U966J.
21-12) Levy, A. and Miller, S. E., Final Report on the Role of Solvents in
Photochemical-Smog Formation, National Paint, Varnish and Lacquer
Association, Inc., Washington (1970).
21-13) Finklestein, H., Preliminary Air Pollution Survey of Aeroallergens, A
Literature Review, APTD-69-23, NAPCA, Washington U9&9J.
21-lU) Hindawi, I. J., Air Pollution Injury to Vegetation, U.S. Government
Printing Office, Washington (1976).
21-15) U.S. Bureau of the Census, Census of Manufactures, 1967> U.S. Government
Printing Office, Washington (1911).
21-16) Stanford Research Institute, Chemical Economics Handbook, Stanford
Research Institute, Palo Alto, California (various dates).
21-17) Oil, Paint, and Drug Reporter, Chemical Profiles, Schnel, New York
(various dates).
21-18) Environmental Sciences Services Corp., Solvent Emission Control Laws and
the Coatings and Solvents Industry, Environmental Sciences Services
Corp., Stamford, Connecticut (no date).
S-14121
-------
AI!U iUSl-KitL'HCiaJ (ConUl).
21-19 '•dptao.ijfe'Control Techniques for Partioulate Air Pollutants, NAPCA
Ubl^ation >.'o. AP-51," NAPCA, Washington U9&9J-
21-20; Dimi'iric.des, B. and Wesson, T. C., "Reactivities of Exhaust Aldohydor.",
at tine 6^th Annual Meeting of the Air Pollution Control
Atlantic City, Hew Jersey (l97l).
21- 211 AftaJJIler, A. P., "Photochemical Reactivity of Aldehydes", Paper prepared
£ Government Printing Office, Washington (1968).
------- |