EPA-450/1-78-001
CONTROL TECHNIQUES
FOR NITROGEN OXIDES EMISSIONS
FROM STATIONARY SOURCES -
SECOND EDITION
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
January 1978
-------
This report has been reviewed by the Emission Standards and Engineering Division, Office of Air Quality Planning
and Standards, Office of Air and Waste Management, Environmental Protection Agency, and approved for
publication. Mention of trade names or commercial products does not constitute endorsement or recommendation
for use. Copies are available free of charge to Federal employees, current contractors and grantees, and nonprofit
organizations - as supplies permit - from the Office of Library Services, Environmental Protection Agency, Research
Triangle Park, North Carolina 27711; or copies may be purchased from the Superintendent of Documents, U.S.
Government Printing Office, Washington, D.C. 20460.
Publication No. EPA-450/1-78-001
-------
PREFACE
This document is the second edition of the EPA document entitled: Control Techniques for
Nitrogen Oxides Emissions from Stationary Sources. This document was first published in 1970 as
National Air Pollution Control Administration Publication No. AP-67. Some sections of the second
edition have been substantially modified from the original, and others have required only minor up-
dating. For example, Section 6 on NOX control of nitric acid plants has been extensively rewritten.
Additionally, Section 8 of the original edition, "Nitrogen Oxides Emission Factors" has been incor-
porated into Section 2. Section 9, "Possible New Technology" has been included in Section 3. This
revision incorporates reviewers' comments from drafts of the second edition and adds new material on
the energy and environmental impacts of the control techniques as required by Section 108 (b) (1) of
the 1977 Clean Air Act.
The Energy and Environmental Division of Acurex Corporation has prepared this document for
the Environmental Protection Agency. The EPA Project Officer was G. H. Wood, who was assisted by M.
Davenport. The Acurex Program Manager was H. B. Mason and the Project Engineer was R. M. Evans;
principal contributors were A. Balakrishnan, C. Castaldini, R. Schreiber, W. Toy, and L. R. Waterland.
This document has been reviewed by the Environmental Protection Agency, the National Air Pollu-
tion Control Techniques Advisory Committee (NAPTAC), and many individuals associated with other Federal
agencies, State and local governments, and private industry. The members of NAPTAC are listed on the
following page. In addition, Acurex acknowledges the valuable assistance provided by the following
individuals and their organizations: J. Copeland, G. Crane, M. Davenport, K. Durkee, R. Iversen, T.
Lahre, A. Trenholm, R. Walsh, G. Wood and K. Woodward of the Office of Air Quality Planning and Stan-
dards; J. S. Bowen, R. E. Hall, D. G. Lachapelle, W. S. Lanier, G. B. Martin and J. Wasser of the
Combustion Research Branch, Industrial Environmental Research Laboratory (IERL); R. D. Stern of the
Process Technology Branch, IERL; Don Carey of the Division of Stationary Source Enforcement, IERL;
John Pierovich of the U. S. Forest Service; W. Skidmore of the U. S. Department of Commerce; Wes
Pepper and J. Mulloy of the Los Angeles Department of Water and Power; J. Peregoy and W. Barr of the
Pacific Gas and Electric Co., R. E. Levine of Southern California Edision and J. Johnson of Babcock and
Wilcox Co.
iii
-------
U. S. ENVIRONMENTAL PROTECTION AGENCY
NATIONAL AIR POLLUTION CONTROL TECHNIQUES ADVISORY COMMITTEE
Chairman and Executive Secretary
Mr. Don R. Goodwin, Director
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
COMMITTEE MEMBERS
Dr. Lucile F. Adamson
1344 Ingraham Street, N. VI.
Washington, D. C. 20011
(Howard University Professor,
(School of Human Ecology)
Mr. 0. B. Burns, Jr., Director-
Corporate Environmental Activities
Westvaco Corporation
Westvaco Building, 299 Park Avenue
New York, New York 10017
Mr. Donald C. Francois, Assistant Director
Division of Natural Resources Management
Department of Conservation and Cultural Affairs
Post Office Box 578
St. Thomas, Virgin Islands 00801
Mr. Waldron H. Giles, Manager
Advanced Material and Space Systems Engineering
General Electric Company
Reentry and Environmental Systems Division
3198 Chestnut Street, Room 6839B
Philadelphia, Pennsylvania 19101
Mr. James K. Hambright, Director
Department of Environmental Resources
Bureau of Air Quality and Noise Control
Post Office Box 2063
Harrisburg, Pennsylvania 17120
Mr. W. C. Hoi brook, Manager
Environmental and Energy Affairs
B. F. Goodrich Chemical Company
6100 Oak Tree Boulevard
Cleveland, Ohio 44131
Mr. Lee E. Jager, Chief
Air Pollution Control Division
Michigan Department of Natural Resources
Stevens T. Mason Building (8th floor)
Lansing, Michigan 48926
Dr. Joseph T. Ling, Vice President
Environmental Engineering and Pollution Control
3M Company
Minnesota Mining & Manufacturing Company
Box 33331, Building 42-5W
St. Paul, Minnesota 55133
Mr. Marcus R. McCrayen
Assistant Vice President
of Environmental Engineering
United Illuminating Company
80 Temple Street
New Haven, Connecticut 06506
Mrs. Patricia F. McGuire
161 White Oak Drive
Pittsburgh, Pennsylvania 15237
(Member of the Allegheny County
Board of Health, Pennsylvania)
Dr. William J. Moroz
Professor of Mechanical Engineering
Center for Air Environment Studies
226 Chemical Engineering, Building II
Pennsylvania State University
University Park, Pennsylvania 16802
Mr. Hugh Mullen, Director
of Government and Industry Relations
I. U. Conversion Systems, Inc.
3624 Market Street
Philadelphia, Pennsylvania 19104
Mr. C. William Simmons
Air Pollution Control Officer
San Diego Air Pollution Control District
9150 Cheasapeake Drive
San Diego, California 92123
Mr. E. Bill Stewart, Deputy Director
Control and Prevention
Texas Air Control Boad
8520 Shoal Creek Boulevard
Austin, Texas 78758
Mr. Victor H. Sussman, Director
Stationary Source Environmental
Control Office
Ford Motor Company
Parkland Towers West, Suite 628
Post Office Box 54
Dearborn, Michigan 48126
iv
-------
TABLE OF CONTENTS
Section
Page
LIST OF FIGURES ..... ~^
......................... ix
LIST OF TABLES ....
.......................... xi1
SUMMARY ...................
......................... xvi
1 INTRODUCTION . . .
.............................. 1-1
2 CHARACTERIZATION OF NOY EMISSIONS . .
* ..................... 2-1
2.1 Definitions and Formation Theory
2.2 Sampling and Analysis Methods . . .................... ;'J
2.3 Equipment Descriptions, Emissions Estimates! Emission "Factors ........
and Fuel Usage by Application Sector .......... ....... 2_3
2.3.1 Utility Boilers .....
2.3.2 Industrial Boilers ...... ...................... 2'6
2.3.3 Commercial and Residential Space Heating .' .' ! .............. I'M
2.3.4 Internal Combustion ......... . .....'.'.'.' ......... ?~2?
Combust1on E"9ines .......... 2.22
............. 2-25
2.3.5 Industrial Process Heating . .
2.3.6 Incineration 2"2'
2.3.7 Noncombustion Sources 2"31
2.3.8 Other NOY Emissions ...'!.' 2"31
x 2-34
2.4 Summary of 1974 NO Emissions and Fuel Consumption .... ? o,
2.5 NOX Emission Trends and Projections ............'. 2 34
REFERENCES FOR SECTION 2
2-48
3 CONTROL TECHNIQUES
*"** ,j |
3.1 Combustion Modifications
*"** ,3~ |
3.1.1 Factors Affecting NOX Emissions from Combustion 3_-,
3.1.1.1 Thermal NO ....
3.1.1.2 Fuel NO x ....'.'.'. 3'2
3.1.1.3 Summary xof Process Modification'concepts' ' .' .' .' .' .* .' .' .' .' .' .' .' .' .' ; 3^5
3.1.2 Modification of Operating Conditions ... -> -,<-
J-lb
3.1.2.1 Low Excess Air Combustion .... , lc
3.1.2.2 Off-Stoichiometric Combustion '. t ,q
3.1.2.3 Flue Gas Recirculation ,1;:
3.1.2.4 Reduced Air Preheat Operation ... 0*07
3.1.2.5 Load Reduction ] '*'
3.1.2.6 Steam and Water Injection . o"
3.1.2.7 Ammonia Injection ,f[j
3.1.2.8 Combinations of Techniques .'.".'."!.'.'.'.'!."!!.''"'" 3!34
3.1.3 Equipment Design Modifications 3_34
3.1.3.1 Burner Configuration , -,.
3.1.3.2 Burner Spacing "'.'.'.' 1 .'!.*.'.'!.'!! I.'"'' ^37
3.1.4 Fuel Modification 0 00
o-oo
3.1.4.1 Fuel Switching ,
3.1.4.2 Fuel Additives 3"3°
3.1.4.3 Fuel Denitrification .'
-------
TABLE OF CONTENTS (Continued)
Section ^3^
3.1.5 Alternate Processes 3-42
3.1.5.1 Fluidized Bed Combustion 3-42
3.1.5.2 Catalytic Combustion 3-44
3.1.5.3 Repowering 3-45
3.1.5.4 Combined Cycles 3-46
3.2 Combustion Flue Gas Treatment 3-47
3.2.1 Dry Flue Gas Treatment 3-47
3.2.2 Wet Flue Gas Treatment 3-49
3.3 Noncombustion Gas Cleaning 3-51
3.3.1 Plant Design for NOX Pollution Abatement at New Nitric Acid Plants . . . 3-52
3.3.1.1 Absorption Column Pressure Control 3-52
3.3.1.2 Strong Acid Processes 3-52
3.3.2 Retrofit Design for N0y Pollution Abatement at New or Existing
Nitric Acid Plants . 3-53
3.3.2.1 Chilled Absorption 3-54
3.3.2.2 Extended Absorption 3-54
3.3.2.3 Wet Chemical Scrubbing 3-54
3.3.2.4 Catalytic Reduction 3-57
3.3.2.5 Molecular Sieve Adsorption 3-59
REFERENCES FOR SECTION 3 3-59
4 LARGE FOSSIL FUEL COMBUSTION PROCESSES 4-1
4.1 Electrical Utility Boilers 4-1
4.1.1 Control Techniques 4-2
4.1.1.1 Combustion Modification 4-2
4.1.1.2 Flue Gas Treatment 4-19
4.1.1.3 Fuel Switching 4-23
4.1.1.4 Fuel Additives 4-24
4.1.2 Costs 4-24
4.1.2.1 Combustion Modification 4-25
4.1.2.2 Fuel Gas Treatment 4-33
4.1.3 Energy and Environmental Impact 4-33
4.1.3.1 Energy Impacts 4-35
4.1.3.2 Environmental Impact 4-36
4.2 Industrial Boilers 4-48
4.2.1 Control Techniques 4-48
4.2.2 Costs 4-55
4.2.3 Energy and Environmental Impacts 4-55
4.3 Prime Movers 4-62
4.3.1 Reciprocating Internal Combustion Engines 4-62
vi
-------
TABLE OF CONTENTS (Continued)
Section
Page
4.3.1.1 Control Techniques .
4.3.1.2 Costs ........."!.'.".".'.' 4'62
4.3.1.3 Energy and Environmental impact 4"70
4-76
4.3.2 Gas Turbines ....
" ' ' 4-83
4.3.2.1 Control Techniques . .
4.3.2.2 Costs . . . ; ." .' 4-88
4.3.2.3 Energy and Environmental impact' .......'.'.'.', '* ?"92
4.4 Summary ....
4-98
REFERENCES TO SECTION 4 .
4-101
5 OTHER COMBUSTION PROCESSES
" 0"|
5.1 Space Heating
** 0** I
5.1.1 Control Techniques . .
5.1.2 Costs . . . .'.'.'.'." 5'4
5.1.3 Energy and Environmental impact ........ f"7
5-8
5.1.3.1 Energy Impact
5.1.3.2 Environmental Impact . 5"8
5-8
5.2 Incineration and Open Burning
5-10
5.2.1 Municipal and Industrial Incineration
*"* 0" i \j
5.2.1.1 Emissions
5.2.1.2 Control Techniques 5-1]
5.2.1.3 Costs . . 5-11
5-14
5.2.2 Open Burning ....
5-14
5.2.2.1 Emissions ....
5.2.2.2 Control Techniques ..'.'.'!.*.'.'!.' 5'14
5-15
5.3 Industrial Process Heating
5-18
5.3.1 Petroleum and Natural Gas
5-19
5.3.1.1 Process Description
5.3.1.2 Emissions and Control Techniques 5"19
5-20
5.3.2 Metallurigical Process .
5-25
H'H I?"55 Description and Control Techniques ... , «
5.3.2.2 Emissions . . 5-25
' " 5-32
5.3.3 Glass Manufacture .
5-34
5.3.3.1 Process Description .
5.3.3.2 Emissions .' 5-34
5.3.3.3 Control Techniques . '. " 5"38
5-40
5.3.4 Cement Manufacture . . .
5-40
5.3.4.1 Process Description . .
5.3.4.2 Emissions ... 5-40
5.3.4.3 Control Techniques .....'. 5'42
5-43
5.3.5 Coal Preparation Plants ...
5-44
REFERENCES FOR SECTION 5
5-46
vi i
-------
TABLE OF CONTENTS (Concluded)
Page
NONCOMBUSTION PROCESSES , 6'1
6.1 Nitric Acid Manufacture 6-2
6.1.1 Dilute Nitric Acid Manufacturing Processes . . . 6-2
6.1.1.1 Single Pressure Processes 6-4
6.1.1.2 Dual Pressure Processes 6-4
6.1.1.3 Nitric Acid Concentration 6-7
6.1.1.4 Direct Strong Nitric Acid Processes 6-7
6.1.2 Emissions 6-10
6.1.3 Control Techniques for NOX Emissions from Nitric Acid Plants 6-11
6.1.3.1 Chilled Absorption 6-14
6.1.3.2 Extended Absorption 6-16
6.1.3.3 Wet Chemical Scrubbing 6-16
6.1.3.4 Catalytic Reduction 6-25
6.1.3.5 Molecular Sieve Adsorption 6-29
6.1.4 Costs 6-33
6.2 Nitric Acid Uses 6-36
6.2.1 Ammonium Nitrate Manufacture 6-36
6.2.1.1 Process Description 6-36
6.2.1.2 Emissions 6-38
6.2.2 Organic Oxidations 6-38
6.2.2.1 Process Description 6-38
6.2.2.2 Emissions 6-39
6.2.2.3 Control Techniques 6-39
6.2.2.4 Costs 6"40
6.2.3 Organic Nitrations 6-40
6.2.3.1 Process Description 6-40
6.2.3.2 Emissions 6-46
6.2.3.3 Control Techniques 6-47
6.2.3.4 Costs
6-49
6.2.4 Explosives: Manufacture and Use 6-49
6.2.4.1 Process Description 6-49
6.2.4.2 Emissions 6-50
6.2.4.3 Controls °'^
6.2.4.4 Costs 6"5Z
6.2.5 Fertilizer Manufacture 6-52
6.2.6 Metals Pickling 6-54
REFERENCES FOR SECTION 6 6-55
APPENDIX A - SELECTED TABLES IN ENGLISH UNITS A-l
APPENDIX B - PREFIXES FOR SI UNITS B-'
APPENDIX C - GLOSSARY --'
VI11
-------
LIST OF FIGURES
Figure
Page
2-1 Stationary sources of N0₯ emissions
x t-4
2-2 Summary of 1974 stationary source NO emissions 9 -,7
A *« t"~O/
2-3 Nationwide annual NOV emission trends 1940-1972 . o /.n
* £-*tU
2-4 Annual stationary source NOX emission trends 2_41
2-5 Annual stationary source NOX emissions projections - low nuclear 2-46
2-6 Annual stationary source NOX emissions projections - high nuclear 2-47
3-1 Kinetic formation of nitric oxide for combustion of natural qas
atmospheric pressure . .
3-4
3-2 Nitrogen and sulfur content of U.S. coal reserves 3_10
3-3 Percent conversion of fuel nitrogen to NO in
combustion .... x
3-11
3-4 Possible fate of fuel nitrogen contained in coal particles or oil
droplets during combustion
3-5 Conversion of nitrogen in coal to NO
X «. O " I H
3-6 Corner windbox showing overfire air system 3_21
3-7 Two-stage combustion
"""** o L.C.
3-8 N0x vs. theoretical air, overfire air study .............. 3_23
3-9 N0x vs. tilt differential, overfire air study .............. 3_23
3-10 N0x vs. theoretical air, biased firing study, maximum load ........... 3_24
3-11 Effect of FGR on NO emissions ..... 0 or
................... o-£o
3-12 Reduced air preheat with natural gas firing, 320 MW corner-fired unit ...... 3-28
3-13 Correlation of NOX emissions with water injection rate for natural qas
fired gas turbine (Houston L&P Wharton No. 43 unit) ........ ..... 3_31
3-14 Comparison of NOX emissions with pulzerized coal firing, circular
burner vs. dual register burner ....... , ,,
3-15 Extended absorption system on existing nitric acid plant ............ 3.55
4-1 NOX emissions from gas, tangenti ally-fired utility boilers ........... 4_8
4-2 Effects of NOX control methods on a gas, wall-fired utility boiler ........ 4_]0
4-3 NOX emissions from residual oil, tangenti ally-fired utility boilers ....... 4.12
ix
-------
Page
Effects of NOX controls methods on an oil, wall-fired utility boiler 4-13
NOX emissions from tangential, coal-fired utility boilers 4-17
Effect of burner stoichiometry on NOX production in tangential,
coal-fired boilers 4-18
4-7 1975 capital cost of overfire air for tangential, coal-fired boilers 4-26
4-8 S02 conversion vs. excess oxygen in coal-fired utility boilers 4-45
4-9 Effect of combustion modification methods on total nitrogen oxides
emissions and boiler efficiency 4-49
4-10 The influence of flue gas recirculation on NO emissions from a firetube
boiler 4-51
4-11 The influence of flue gas recirculation on NO emissions from a watertube
boiler 4-51
4-12 Effect of NO controls on solid particulate emissions from, industrial
boilers . .x 4-60
4-13 Effects of NO controls on particulate size distribution from oil-fired
boilers . . ? 4-61
4-14 Effect of NOV emissions level on fuel penalty for light-duty trucks 4-77
A
4-15 Effect of derating on 1C engine HC emissions 4-80
4-16 Effect of retarding ignition on 1C engine HC emissions 4-80
4-17 Effect of air-to-fuel ratio on 1C engine HC emissions . 4-81
4-18 Effect of decreased manifold air temperature on 1C engine HC emissions 4-82
4-19 Effect of water injection on 1C engine HC emissions 4-82
4-20 Smoke levels versus NOX level for large-bore diesel engines 4-84
4-21 NOV emissions from large gas turbines without NOV controls 4-86
X X
4-22 NO emissions from small gas turbines without NO controls 4-87
X **
4-23 NO emissions from gas turbines having NO controls and operating on
liquid fuels 4-91
5-1 General trend of smoke, gaseous emissions, and efficiency versus
stoichiometric ratio for residential heaters 5-3
5-2 Effect of excess air on NOX emissions from a 45.3 Mg (50 ton) per day
batch-feed incinerator 5-12
5-3 Effect of underfire air on NO emissions from a 227 Mg (250 ton) per day
continuous-feed incinerator 5-13
5-4 Effect of process rate on NO emissions from a process heater 5-22
5-5 NO emissions as a function of time for an open hearth furnace 5-36
-------
Figure
Page
5-6 The effect of cement kiln temperature on NO emissions 5_45
6-1 Single pressure nitric acid manufacturing process
****** D"0
6-2 Dual pressure nitric acid plant flow diagram 6_6
6-3 Nitric acid concentrating unit
6-8
6-4 Process flow diagram for direct production of highly concentrated nitric acid .... 6-9
6-5 Schematic flow Sheet of the CDL/VITOK NOX removal process 6_15
6-6 TVA chilled absorption process ... , ,.,
6-17
6-7 Grande Pariosse extended absorption process for NO abatement 6_18
6-8 Flow diagram of the MASAR process
6-9 Process flow diagram for the Goodpasture process 6_23
6-10 Nonselective catalytic reduction system 6_26
6-11 Molecular sieve system c ~n
o-ju
6-12 Batch process for the manufacture of nitroglycerin (NG) 6-42
6-13 Schmid-Meissner continuous-nitration plant 6_43
6-14 Biazzi continuous-nitration plant
6-15 Recovery of spent acid c ...
o-4b
6-16 Trinitrotoluene (batch process) manufacturing diagram 6.51
XI
-------
LIST OF TABLES
Table
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
2-9
2-10
2-11
2-12
2-13
2-14
2-15
2-16
2-17
2-18
2-19
2-20
2-21
2-22
2-23
2-24
2-25
2-26
2-27
2-28
Quality of Emission Factors for Oxides of Nitrogen . .
Emissions, Emission Factors, and. Fuel Usage for Steam
Generation, 1974 -Utility boilers . .t
Annual NO Emissions and Fuel Consumption Comparison for
Utility B8ilers, 1974 ,
Annual Fuel Usage for Utility Boilers (EJ)
Annual NOX Emissions for Utility Boilers (Tg)
Emissions, Emission Factors, and Fuel Usage for Steam Generation,
1974 - Industrial Boilers
Annual Fuel Consumption for Industrial Boilers .
Annual NO Emissions from Industrial Boilers
Emissions, Emission Factors, and Fuel Usage for Commercial Boilers, 1974 ....
Emissions, Emission Factors, and Fuel Usage for Residential Space Heating, 1974.
Summary of Annual NO Emissions and Fuel Consumption for Commercial and
Residential Space Heating, 1974
Annual Fuel Usage for the Commercial/Residential Sector, (EJ)
Annual NOX Emissions from Commercial Boilers (Gg)
Annual NOX Emissions from Residential Space Heating (Gg)
Emissions, Emission Factors, and Fuel Usage by Equipment Category for
Internal Combustion Engines, 1974
Annual Fuel Consumption by Internal Combustion Engines (PJ)
Annual NO Emissions from Internal Combustion Engines (Gg)
Summary of Annual Emissions for Industrial Process Heating Equipment (Gg) . . .
Summary of Annual NO Emissions from Incineration .
Summary of Annual Emissions for Noncombustion Sources
Estimates of Annual NOW Emissions from Other Sources
X
Summary of Total Annual NOX Emissions from Fuel User Sources, 1974
Summary of Annual Fuel Usage, 1974
Comparisons of Annual NO Emissions Data
Annual Fuel Consumption Comparisons
Annual Nationwide NOX Emissions Projected to 2000
Estimated Future NSPS Controls
Annual Nationwide NO,, Emissions to 2000
Page
2-5
2-9
2-10
2-11
2-12
2-15
2-17
2-18
2-20
2-21
2-21
2-23
2-24
2-24
2-26
2-28
2-28
2-32
2-33
2-33
2-33
2-35
2-36
2-38
2-39
2-42
2-44
2-45
xn
-------
Table
Page
3-1 Factors Controlling the Formation of Thermal NO
3-2 Analyses of Typical U.S. Fuel Oils
3-8
3-3 Analyses of Typical U.S. Coals and Lignite
3-4 Summary of Combustion Process Modification Concepts
************ «3"" I /
3-5 Summary of Results with Ammonia Injection
3-6 NOX Formation Potential of Some Alternate Fuels
4-1 Range of Uncontrolled Utility Boiler NO Emissions
4-2 Summary of Combustion Modification Techniques for Large Boilers 4.4
4-3 Major Japanese Dry FGT Installations (Selective Catalytic Reduction) 4.2Q
4-4 Major Japanese Wet FGT Installations
4-27
4-29
3-33
3-40
Los Angeles Department of Water and Power Estimated Installed 1974 Capital
Costs for NOX Reduction Techniques on Gas and Oil-Fired Utility Boilers
4-7 nsae apta
4-30
4-8 1975 Differential Operating Costs of Overfire Air on New and Existina
Tangential Coal-Fired Utility Boilers ............... \ 4_31
4-9 Impact of NOX Control Techniques on Major Utility Boiler Components ....... 4.32
4-10 Cost Estimates for Combustion Flue Gas Treatment Processes ............ 4.34
4-11 Effects of Retrofit Combustion Modification NO Controls on Utility
Boilers Efficiency .............. x .........
4-12 Representative Effects of NOX Controls on CO Emissions from Utility Boilers . . . 4-38
4-13 Effects of NO Controls on Particulate Emissions from Coal-Fired
Utility Boilers .
4-14 Effects of NO Controls on Emitted Particle Size Distribution
from Coal-Firld Utility Boilers 4_43
4-15 SOX Emissions Summary for Utility Boilers 4_46
4-16 Summary of POM Emission Tests for a Coal-Fired Utility Boiler 4.47
4-17 Effects of NOX Controls on CO Emissions from Industrial Boilers 4.57
4-18 Representative Effects of N0y Controls on Vapor Phase Hydrocarbon
Emissions from Industrial Boflers 4_5g
4~19 CoSJstion EnSnesSSi°n C°ntro1 Techniclues for Reciprocating Internal
4-63
4-20 Effect of N0x Controls on Large-Bore Internal Combustion Engines 4.65
1975 California 13.4 G/KWHR (l^G/HP-HR) combined NOX and HC°Levels *° *?**. . . . 4-67
4-22 1975 Vehicle Emission Limits . . cn
H-DO
xiii
-------
Table
4-23
4-24
4-25
4-26
4-27
4-28
4-29
4-30
4-31
4-32
4-33
4-34
4-35
4-36
5-1
5-2
5-3
5-4
5-5
5-6
5-7
5-8
5-9
5-10
6-1
6-2
6-3
6-4
6-5
Emission Control Techniques for Automotive Gasoline Engines
Emission Control Systems for Conventional Gasoline Internal Combustion Engines. .
Cost Impacts of NOX Controls for Large-Bore Engines
Typical Baseline Costs for Large (> 75 kW/cylinder) Engines
Typical Controls Costs for Diesel-Fueled Engines Used in Heavy-Duty Vehicles
(>2700 kg or 3 tons)
Estimates of Sticker Prices for Emissions Hardware from 1966 Uncontrolled
Vehicles to 1976 Dual-Catalyst Systems
Representative Effects of NOX Controls on CO Emissions from Internal
Combustion Engines
Relationship Between Smoke, EGR, and Retard
Gas Turbine - Summary of Existing Technology - Combustion Modifications
Impact of NOX Emission Control on the Installed Capital Cost of Gas Turbines. . .
Water Injection Costs, Mills/kWh
Representative Effects of NOX Controls on CO Emissions from Gas Turbines
Summary of the Effects of NO Controls on Vapor Phase Hydrocarbon
Emissions from Gas Turbines
Summary of NOV Controls Technology
X
Nationwide N0y Emissions from Space Heating Projected to 1990
Comparison of Mean Emissions for Cyclic Runs on Residential Oil-Fired Units . . .
Effect of Low-NO Operation on Incremental Emissions and System Performance
for Residential Warm Air Furnaces
Annual Emissions of Nitrogen Oxides from Open Burning
Effects of NOV Controls on NOV Emissions from Petroleum Process Heaters
X X
NOX Emissions from Petroleum Refinery CO Boilers
Estimated NO Emissions from Steel Mill Processes and Equipment
Effects of NOX Controls on Steel Industry NOX Emissions
NOX Emissions from Glass Melting Furnaces
Recommended Programs for Reducing Emissions and Energy Consumption in the
Glass Industry
NOX Abatement Methods on New or Existing Nitric Acid Plants
Performance of Hercules Purasiv N Unit During Three-Day Run
Performance of U.S. Army-Holston Purasiv N Unit During Three-Day Run
Capital and Operating Costs for Different NOX Abatement Systems in a 270 Mg/d
Nitric Acid Plant
Annual Energy Requirements (TJ) for NO Abatement Systems for a 270 Mg/d
Nitric Acid Plant
4-68
4-69
4-72
4-72
-73
4-74
470
4-85
4-89
4-93
4-94
4-97
4-97
4-99
5-2
5-5
5-9
5-15
5-21
5-24
5-33
5-35
5-39
5-41
6-12
6-31
6-32
6-34
6-35
xiv
-------
Table
Page
6-6 Basis for Tables 6-4 and 6-5
'«' 6-35
6-7 Annual Nitric Acid Consumption In the United States, 1974 6_37
6-8 Estimated NOX Emissions from Organic Nitrations 1n 1970 6_48
6-9 Emission Factors for Manufacture of Explosives 6_53
xv
-------
SUMMARY
In this document, the term "nitrogen oxides" or "NO " refers to either or both of two gaseous
oxides of nitrogen, nitric oxide (NO), and nitrogen dioxide (NOo). These substances are important
in air pollution control because they are involved in photochemical reactions in the atmosphere and
because, by themselves, they have harmful effects on public health and welfare.
CHARACTERIZATION OF NOX EMISSIONS
Manmade oxides of nitrogen are currently emitted at a rate in excess of 20 Tg (22 million
tons/yr) in the United States. Stationary sources account for approximately 60 percent of these
emissions, of which 98 percent are due to combustion sources. Combustion generated NO is derived
from two separate formative mechanisms, thermal NO and fuel NO . Thermal NO results from the
thermal fixation of molecular nitrogen and oxygen in the combustion air. This is the dominant
mechanism with the firing of clean fuels such as natural gas and distillate oil. Fuel NO results
from the oxidation of organically bound fuel nitrogen compounds. This can be the dominant mechanism
with the firing of coal and high nitrogen residual oils. The rate of formation of both thermal NO
and fuel NO is strongly dependent on the combustion process conditions. The emissions due to both
mechanisms are increased by intense combustion resulting from rapid mixing of the air and fuel
streams. Additionally, the emissions due to thermal NOX are sharply increased by increased local
combustion temperatures.
Since equipment process conditions and fuel type are so important in determining NO emissions,
the characterization of emissions and the evaluation of control potential requires detailed classi-
fication of stationary sources according to factors known to influence NO formation. Over 100
combinations of equipment type and fuel type are identified as having significantly different poten-
tial for NOX emissions and/or NOX control. The emission compilation for these sources for the year
1974 shows, however, that the 30 most significant equipment/fuel combinations are responsible for
over 80 percent of stationary source emissions.
xvi
-------
The total nationwide emissions in 1974 for stationary sources, grouped according to applica-
tion sector, are shown on Figure S-l. On an uncontrolled basis, utility boilers accounted for over
40 percent of stationary source emissions. These boilers fired 61 percent coal, 18 percent oil, and
21 percent gas. For all stationary sources, the firing of coal yielded 35 percent of total NO
emissions while comprising only 28 percent of stationary source fuel consumption. Conversely, natural
gas comprised 42 percent of stationary source fuel consumption, but generated only 34 percent of sta-
tionary source NOX. Although some sectors shown on the figure, such as noncombustion sources, are
small on a nationwide scale, they may be crucial in local NO abatement programs.
CONTROL TECHNIQUES
Current and advanced methods for stationary source N0x control operate either through suppres-
sion of NOX formation in the process or through physical or chemical removal of NO from the stack
gases. Suppression of NOX formation is most effective with combustion sources. Candidate approaches
include combustion process modification through alteration of operating conditions on existing sys-
tems or alternate design of new units, fuel modification through fuel switching, fuel denitrifica-
tion, or fuel additives, and use of alternate combustion concepts such as catalytic combustion and
fluidized bed combustion. Removal of N0x from stack gases is most effective with noncombustion
sources of NOX, chiefly chemical manufacturing. Candidate approaches include catalytic reduction,
with wet chemical scrubbing, extended and chilled absorption, and adsorption with molecular sieves.
A summary of general stationary source NOX control techniques is given on Table S-l.
Combustion process modifications have been extensively implemented on existing gas and oil
fired utility boilers to comply with local emission standards. External control techniques such
as low excess air firing, biased burner firing, overfire air and flue gas recirculation have
yielded emission reductions up to 60 percent of uncontrolled, baseline operation. A summary of
combustion modification concepts is given in Table S-2.
With coal firing, the most effective combustion modification technique for utility boilers
is a combination of low excess air firing and off-stoichiometric combustion through biased firing,
overfire air, or use of delayed mixing burners. Utility boiler manufacturers are currently includ-
ing these procedures in new unit designs to comply with the Standard of Performance for New Sta-
tionary Sources of 301 ng/J (0.7 Ib N02/106 Btu). Retrofit implementation of low excess air and
off-stoichiometric combustion has shown that a level of 258 ng/J (0.6 Ib N02/106 Btu) is achievable
with some unit designs. Emission levels as low as 189 ng/J (0.44 Ib N02/106 Btu) have been demon-
strated on a tangentially fired unit equipped with factory installed overfire air. Current
xvi i
-------
Commercial/
residential
space heating
9.0%
Utility boilers
41.9%
Reciprocating 1C
engines 19.8%
Industrial
boilers 18.2%
Incineration 0.3%
Gas turbines 2.0%
Others 3.6%
Noneombustion 1.7%
Industrial process
heating 3.5%
Source
Utility Boilers
Industrial Boilers
Reciprocating 1C Engines
Coranercial/Residential Heating
Industrial Process Heating
Noncombustion
Gas Turbines
Incineration
Other
TOTAL
Estimated N0x Emissions
Fg 106 Tons
105
218
628
444
2.413
1.090
0.432
0.203
0.236
0.039
0.435
12.171
13.416
Figure S-l. Summary of 1974 stationary source N0₯ emissions.
xvili
-------
UJ
o
cc.
X
o
O
>-
i
co
CO
<:
o
IO
f
*Q.
V
C
4>
t
4J
i
_
C
C
1
n-
o
vt
2
«
c
o
7
§.
o
0
o
c
u
0.
§
a-
I
u
Ol
E
4
u
g
£
1
I
?
«
&
S £
3'
N S
II
in
t- i
>> 0) E
^ "o - c
i- -o in en
3 c in *J
IO *i QJ 3 01
4-> ID m o i v^
Cl C o c.a
Degree of control
limited by opera-
tional problems
A " u
? 0 S. I ^
iO O T> v»
4J ^ go 4J
C 01 O CL
t-i VI -O Ol
O « VI C C 0
o. o a> o « e
O <- 3 10 6
H- o 4-> o a» e:
in -t- i oi vi
§CL Ol t- p v> O
i -o C oi t>
> « O O.T3 3
* «n o E O
* O o'itg 01 *"
V. C r- 01 C 0 10
0) 3 C O C Ol
S. O 3 f- - <0 U
U O 1 f
3 X U
2* Jj{ _* S
e .*£-
x: 01
X*-> «*- Ol Ol
0 10 3.
Z t- M <4-
tvt CM
§1. C? 01
o.-<- -o e
4-» O. >i Ol O
V 3 >0 O N
t > Ol t- C
lltl?^
§
73
M «r-
3 H-
11
°-2
U X
4-> 10 O
C 0 Z
o> ~>.
C *- x
Ol - O
"j^a*"
0. E m
3 3
**5 8
XI tn 4-» iti
t- 3 i >
sill
O. o tn L.
o
j$~
4^ °e 4V
3 *"* 5
e oicL
O U
o c-o
C 3 -
O O O
z i/i «
f- tn i in
i O VI 3 C
Oi O. o O in 3
> 10 f O O
a c x.c *oj E*
i- i Ol 3 O
in ip i- H- o
01 01 43 f.
tn o to >> o
in c o oi x> i-
Ol Ol > f
0 -r- T3 10 01 r-
p (- x: o to
C Ol O C +»
o. a. vi oi aj
x » oi t. E
*J oi vi in oi
$C in "*- U
Oi O Oi 1- O
C-r- 0 Ol
Xf- 4-» P *i 4.
5 a. a C c 3
ZOO 0.^-M-
in
4-> at
o>c
c ia 10 t- c
7IQ.Si£5
VI 4->
X t- O Ol 3
Ol 0 r- v- X»
<0"- T3 g
t- 0. 3 O
O O 4-> O
t- JO i
aSS?i
0 r- -°
- X oii 4-» in
4-> S c x> c E
to ct- t- oi (D
fc. 4-» in > *J
01 TJ Ol IO C V>
S.C aj oi 0 X
O 10 E <*- o vi
O
«» CM
o p
H- X
o xo
O Z 4J
§z*.g
7 4J
XI in o
*+*
sg1-^
5U^5
*J ..Q 10
F- mat o
32 ^>»
» «»r-
VI 0 Ol IO
** X> « 7
C Ol C
»- C 4J 01
O f l~ -O
CLXI oi «-
g > in
llii
Ol
M
ai
_T
5
I
**
7 x
cox:
Ol (SI 4-1
Ol X
<- X
<«- O v>
3"si
vt Ol
O O
u x o
01 t-
01 >> c 10
3 a O O
u. E o o
c
w § X
2 SI?
c
<0 > Ol
i Ol 3
o. oi-o
Ol in 4->
3 C
V- OJ «
1 1 l/l
O IO O.
*J in r CO
0* 1 <0 Ot
x: 4J - ^
"p2 t-o
t^cl 21 "i
c
o Cf-
Xr i^
C 10 4J
U» i- 7 0
oi oi ac
in > x: 1/1
3 C 4-1
§o c -
0 >>-
in m c
5^..8&
i-'SpI
?
i!
Ol
c
M
1
a.
01
3
01
_
1
f
M- 1 Of 1
O - U 3
4J C tA "o
,0,0,0!
vt CL tn
§-«...-
i o 4-> ai
oi t- o >
-X Ol -^
SOI >«- 4->
> »- ^~
r- O> -O
Ol 4J T3
O>-r- 4-1 Ol
o oi vt
10 t- VI
CO Ol
in c
*1
O<*-
tt-
vi O
VI
£S
CL-1-
a.*j
3 O
Vt
4J
"i -c
£3
o
7
VI
g
x»
1
0
c X>
13
f 1*-
CL^-
O.-O
3 0
VI S
vt
Ol
X)
"St
Ol
Ol
c
Effectiveness for coa
doubtful ; no effect o
thermal NOX
IO
N r
L. *r-
3 4J C
^ .°
in 10
Oi in o
O "x,4-»
A
8
fuel nitrogen c
retreatment
^ °-
°x?
53
i§
^cl
o
33
m
o
iT
L.
4J
"oi *C
,f«S
1 f
Of fc C
ill
O Ol 3
H- C 4->
Ol 4^ VI
tn to to
3 01 Ol
Ol -
VI C f-
£-0.8
g
Ol
1 **
II
1 C
1- «
Limited retrofit appl
cations; requires cle
fuels
1.
£
tn
+>*- 13
VI C 3
01 Ol 4J
** 0) vi
tt) 0 >>
i M 4J
IO *^-
O 4» I
Vt VI f*
o a vt
I 4J 10
-»- 10 Ol
a. u (*-
i i
?|l
SIS**
Jo *
S»5
3-2 .
KSS
in Oi 4->
?l >» !
Heterogeneoi
reactions y
tion tempen
mal NOX
u o
77
Bs
s!
i **=
.0 .* u o
'^ _ O +^
!lil
Vt ^- CL
3 014 GL
» *c 8»-
^flrS^
Isll
Ol
s
,
i
i
^
1- HI 1 *>
Fuel nitrogen convers
ay require control
(staging); may requln
large make-up of lime
stone sulfur absorben
o
1 Ol i-
4-> N 3
IO *ff 14
<»- 3 3
O 11 VI
VI
>, 4) C l/>
o t. o oi
30 0
VI T3 3 ">
C O Ol
0) 10 o TJ
8 " .. §
in i- vi i-
i Oi E 4->
4-> x: 35 c
O CL4-> 01
r VI in 4J
J- 0 >, 91
Q. E in u.
"S *
tion 1n solid b<
temperature, lo
i^
5«
r i
2 "-I1
3T.S
o S
n7
r- VI
13 b
LZ 5
xlx
-------
I/)
cl
UJ
o
i
I/I
4-*
C
o
k?
Ol
3
u
r
O.
<
£
°E
Ct.
if
« «o
^ 4^
kM-
£
r i/i
5
CIS
ex 5*
o<
5 ^
o
y~
V p^
£3
ii «-
UJ
X
C C3
o ^z
4J
U «
Ol E
<*- 41
UJ JC
1
Z>^=
J-5
s
o
3
U
k
SoT
L.C9
u.
s.1
o
us
r IO
U-r-
°IO
I/I
I/I
41
U
X Ol
o> c
££
-JH-
IfL
I/I O 01
O k f-
CX4-> S.
X v- 10
41 C -r-
o
01 *4> p
U 3 C CM
3 «*- Ol O
411- C O
DC 0-«- 4J
« Wl
.C 4->
U Ol
k U 01
1 O E
CM CX IO
O
in O
Ol Z ttl
U 1 -C
3 .C 4J
o o>
ttl t- C
ce. .c t-
vi
01
u
0
1
1^
I/I
Ol
^w
P*
o
CQ
IS
u
01
o
01 Ol
k i
U
So"
^^
k C
Ol Ol
1 x
ctf-
VI
i.
i
°
sx
Q
_J
4J
J
0) VI
ii
CO IO
c
z '*~ II-
r CM O
01 2C
«»- O O
4-> C
01 Ol
r- I/I I/I
4^ O IO 41
* 3 Ol
r- -0 41 >»
O 4> -C X
> k 4-> O
o
c
10 01 1 0.
c oi E
Olf- t- 01
C *- 4->
t- 3 X
§*"" "E 10
C Ol
o o n ex
f Ol
CU 4J >> I/I
E 3 C
£5
ts
(/)
41
IO
^
U
C
1 1
^J
O 1
O 41
f f- > C
10 k O 10
0 4-> 41
O $ C» C
I/I t- v> 41 4J
Ol .C C 3 I/I
0 0 OM- 3
O "5 Olr E
41 4J 01 i O
OH I/I t. IO O
I/I
C
^
I
4->
I/I
IO
^
01
3
Ol
o c
01 1-
Vt X
10 t-
£e
U k
c ^«
c
1 O
urner/flrebox
eslgn for two
tage combustl
CO "O I/I
o
Ol
t^.
k T>
^ IO
10 I-
ttl O £
O CX U
o
<4- 01
O in Ol
10 C
".£
Ol »- 01
3 41 3
CQ I/I ^3
C
i CM O
Ql SBT
«*- O 0
4J C
01 Ol
p I/I I/I
* Ol ^3 C
4-> U IO Ol
IO 3 Ol
r 73 Ol 55
0 41 J= X
> k 4-> O
s
C CX 1
r- E 01 0.
01 k E
O>4-> 41
C 1 41 4->
O i H » ex
CMk
IO $ -t- O
r- 0 k 3
U. r- Q.T3
I/I
Ol
U
0. 0»T-
CX.C C
3 4-> 10
in JE
«*- 0
4-> O ttj
O*M Z
I/I
Ol
(J
ID
3*01
l»- Irt C
41 i-
« C J3
01 C 4->
f 01
r- I/I
O 0 10
CO i Ol
g
IO
5I£
U 3
01 T- 4->
in 4-t 10
« IO k
« .0 Ol
S42&
a-a
X k
O Ol
£> C
«^
o o>
oi in oi
Ol IO C
* *» * n
IO k U ^
r- O IO
C C CX
UJ i- I/I
c
o
o
o
IO
o
_J
4-> 01
u c
£*x
4J f- f-
u *o E
ttl C
k f- C
t- O
o
k U U
O 41 41
C <4- <*-
t- >*- H-
Z Ol Ol
IO
i 01 i-
«*- C 41
o^ o a.
Ol O E
to O Ol
10 o in 4->
4» -0
i. Ol -*
U C Ol 10
C O t- 41
' « 1 Nl >> Q.
in
Ol
u
IO
c
3
1*-
c
Ol
^^
fBI
o
CO
01 f 4J
ie in in
01 3 C
k ^ CU
o E *->
Ol O C
O (J v
4- §
Si- .0 in
3 ttl O
f. in k -
C kiT §
Ol Ol C
i-«4- >,
in in in ~o
Ol C Oi 0
O 10 u k
Oi k ie 01
Q: 4J <4-
U
«*-
Ol
c
ie
f O) 1-
<*- C ttl
o 1^ o a.
Ol Oi E
Ul O ttl
ie o in 4->
41 -0
C cui ^
u c tti n
C O i- 41
HI N >, CX
in
41
u
. Ol
r- OIT3
«»- C-^
t- in
O i 01
Ol 0 k
in o
IO O Ol
£ 41 3 4)
U C *O E
C 0 4)-*-
l M k 4-»
£
Ol 4->
u ie to E
1 k
U O
41**-
> C
ceo
00---
O-i- 4->
4-> U
C U 41
^ I/) C
in i-
4< 4)
T3 > rO
CK4JZ
li
Zi 0
U -^ IO
x« 1
i O Oi
f Z 1 41
10 4-> t.
u oi en
§"*" 3 CX 1:
T3 IO
£ 4> C i
O k i- ««-
XX
-------
developmental activity is focusing on identifying and, if required, correlating operational problems,
such as increased waterwall corrosion with boiler tubes of conventional chemical composition when
exposed to the reducing conditions at the surface resulting from combustion modifications.
Retrofit combustion process modifications have also been extensively applied to gas
turbines. Water injection has been successfully implemented to achieve emission levels of 75 ppm
at 15 percent excess oxygen. Current activity is focusing on development of dry controls using
premixing, prevaporization and controlled mixing for application to new combustor can designs.
There has been only limited field implementation of combustion process modification for other
stationary combustion equipment e.g., industrial and commercial boilers, residential and commercial
space heating equipment, reciprocating internal combustion engines and industrial process furnaces.
The following sequence is being pursued for N0x control development for these sources: control from
operational fine tuning (e.g., low excess air firing, burner tuning), minor retrofit modifications
(e.g., biased burner firing), extensive hardware changes (e.g., new burners) and major new equipment
redesign (e.g., optimized heat transfer surfaces and burner aerodynamics).
Fuel switching for N0x control is not currently practiced due to the supply shortage of
clean fuels. A number of alternate fuels such as methanol and low-heating-value gas have low NO -
forming potential and may be utilized in the 1980's. The economic incentive for alternate fuel use
usually depends on factors other than N0x control, e.g., desulfurization cost tradeoffs, system
efficiency.
Fuel oil denitrification, usually as an adjunct to oil desulfurization, shows promise for re-
ducing fuel N0x. This concept may be effective for augmenting combustion modifications for NO con-
trol with the firing of residual oil. Fuel additives are not directly effective for suppressing
N0x emissions. Their use to suppress fouling and smoke emissions, however, may permit more exten-
sive use of combustion control methods than would otherwise be practical.
Alternate combustion concepts under development include catalytic combustion and fluidized
bed combustion. Lab-scale tests of catalytic combustion have demonstrated extremely low NO emissions
with clean fuels (1-5 ppm). This concept may see application in the 1980's to stationary gas tur-
bines and space heating systems. Fluidized bed combustion pilot plants have demonstrated NO emis-
sions of the same order as conventional coal-fired power plants using process modifications for NO
control (170 ng/J, or, 0.4 Ib N02/106 Btu). The potential for replacement of conventional utility
and industrial boilers by FBC depends on a number of other factors such as S0x control cost tradeoffs
and operational flexibility, e.g., load following.
xxi
-------
Stack gas treatment for NO removal has been implemented 1n the U.S. only on noncombustlon
sources. Here, an additional incentive is the recovery of NC^ as a feedstock material. The most
widely tested technique is catalytic reduction with selective or nonselective reducing agents. The
short supply of reducing agents (methane, ammonia) coupled with the loss of tail gas NCL as a poten-
tial feedstock is causing interest to shift to alternate processes such as molecular sieve absorp-
tion and extended absorption.
Flue gas treatment (FGT) of combustion sources has been at a low level of development in the
U.S. due largely to the lack of regulatory incentive. The developmental activity has recently ac-
celerated, however, as a result of increased emphasis on stationary source NO controls in the na-
tional NO abatement program. Flue gas treatment could be effective in the 1980's to augment combus-
tion process modifications on large sources if stringent emission control is required, for example, to
comply with a potential short-term N0£ air quality standard. Current developmental activity includes
transferring FGT technology from Japan where stringent NO controls are enforced. Processes being
considered include selective catalytic reduction, selective homogeneous reduction and wet scrubbing.
The dry systems show most promise for NO removal alone. For simultaneous NO /SO removal, several
wet and dry processes are effective but the cost tradeoffs have not been identified.
A sunmary evaluation of NO control techniques for combustion sources is given in Table S-3.
LARGE FOSSIL FUEL COMBUSTION PROCESSES
The three largest stationary emitters of NOX are electric power plant boilers (42 percent of
the total), industrial boilers (18 percent) and prime movers, such as gas turbines and I.e. engines
(20 percent). The most successful NOX reduction technique is modification of operating conditions.
For utility boilers, techniques such as lowering excess air, off-stoichiometric combustion, and, for
gas- ans oil-fired units, flue gas recirculation have resulted in NOX reductions of up to 60 percent
making it possible for them to meet emissions regulations at costs of $1 to $10 per kW (electric
output). Ongoing performance tests are investigating potential side effects of the modifications,
such as increased corrosion and particulate emissions with coal firing.
sources are able to decrease NO by up to 50 percent with no efficiency Impairment or increase in
particulate formation. The most successful techniques are lowering excess air, staged combustion,
and flue gas recirculation.
The energy impacts of applying combustion modification NOX controls to utility and industrial
boilers occur largely through the effects on unit fuel-to-steam efficiency. This is usually
-------
_
p
M C
CO
UJ
o
QC
§
oo
frlsl
u u
Ol 01 *»
fc.'o
a. 4-» C
oo o»
ry empl
pllcat
d with
r far-1
Secon
term
comb
mods
il
VI
_ o «
Q--»- 01 c c
v» u vi »- a>
Ol - « i-
* *»
- . «j a>
01 o «o-o
1 O
n *j
«
01 »-
k- VI
X> -M C
»- V» O>
VI 3 ^
«/» "O «/»
O C Ol
o. -^ ta
U ID
O Ol
*«
01 0> 0. 0
e i- s. c
5.285
0» C --4J
o>-^ c oi
k. O> k.
»- Ol «/» L.
C *rt 0> O
UJ 3 T3 «»-
O VI k. C 3v
4-> C <» 0> H- JO
c oi vi *o g
^ -^- 3 »- VI O
VI "O VI VI U
01 oi c 01 oi
4") "O "^ kl O
« O f
k. T3 « k. O O
O Ol >, VI Q. VO
L^^^t -v7
85^^-5g*5
5^^5ZS-5
Operation
Impact
o
u
*- F k. UJ
>- o
.""Si
£ S
Oj I Z
l^C
c..e
i . S
ai u +>
C C -0 (0
U 0> 0>
S--S
^- -r- I/I
«/).£> 10
VI «4- >>
«»- "O *» 0
*J C S- <*- Ol O *-
o o> 01 vi C «
«»- t. C o -0
01
. _ -^ Ol -0 L.
k. 3K (O » -O 0» »- 4-»
O Ol -^ f- (ft «*.--
>, |Q i H-
4-> 4-> (A Ol U O 4->
.
c E u 5
^ -O »- « E
C r- Q. C
Ol 10 10 O
Vt OH-
IO >>4J
Ol U 3 01 Ol
C C Q. C .K
O 0) -4-» IO Q
0> -^ 3 0. E
Q O O «/> W>
No major Impact
with new design;
retrofit use con
X
2
Cv,
rained by
aracteris
si
5..8
X>4-
o z c
ID O
O * **
4-> E u
X 01 H- X
O f <*- o
CVI 4-> 01 Z
fi^
O «
O "«
SOI X
-CO
CM -was
0 *»
Ex
ppl
Retrofit and n
utility boiler
some use In 1n
trial boilers
^ «/l
CO 13
O C T)
O >i O C VI
Ol C 4-> E *- k.
t/> *o ja
<4- CD L Ol
**- E vi
2- »- S c k.
X -U
3 W» O »
u. 25 c
*»«»- Ol Ol 3
E^- 01
^ «- ** t V) M
VI *> C 3
c c c
(/> O 3 Ol O
New and retrofit
use on utility
boilers; demon-
strated on resi-
dential furnaces
fli
C
§<
IS
u"
Ol«*-
4- ceo
O 0-»- CD
xxiii
-------
c
o
o
o
o
O
O
ii
_§
££
|S
51
UJ
I/I
"Si
«*
^* !
l|
_
(C
C ^j
£ *
£JS
0.
o
VI
VI
4>
C
4)
*^
U
t^
UJ
VI
en o
» 4J
4J C
P U
CJ 41
»-
EVI
4) C
41 C 1 6
4-* » VI «^-
i xi 3 4->
k. k.T»
« 1 VI
XT VI k. £
i §* S
4J 41 E
u* S S &
S ^ "vilo
J **g^
o. « O.4J
1 4> i/i
k VI 41
3 3 U
4> 3
VI f" k>
I/I
U
c p
»- vi C
41 CL
4) C
VI t- C
=> XI -
.^
C 4>
4>*E c
V» »- O
10 » T-
4) 4-> 4J
k. « 3
u >i E o.
4) U k 4J VI
O C O 3 41
41 *- O VI
OE 0 O k. 4)
^cn-j- o jj ^
-«»->, 0 C
00 O) JD O.-«-
k
0
«»-
X
** o
oz
^^
0 "g
E
»« 41
(*) 4->
VI
IO
o
«*-
o
4>
l/l
3 VI
^J:
§ "k
5 5
E
S
"J
4-1
k. (J
4) Q)
i?
VI
VI
IO
JC
1
k.
^
o
3
c
41
S
I/I
TJ 41
$^
c en
»- c
4-1 41
§0
«_>-<
^,
5 S
41 1 3
VI C O
IO "t-
41 U P
T5 C O.
4)
JZ U VI
?«£ S
«*- C
t/1 4) U
k
O
1^
X
** o
oz
f
o s
4-> E
^
»* 4)
0 J=
C ^
41 <0
VI JZ
Q
o j5 vi
« k. 41
4> 3 C
k 4J »-
CL en
4> O» 4)
5^«
S"
|Z
O 41
3 f
I*
S. a.
Pl4->
k. U
10 «
1^ ^J
V O **"
^^ AJ
vi c *J en
"- P « C
VI ^- 3 ^«
« 4-> f k.
JC « ID T-
O. U > «t-
^ *o. .-lo
X O. VI P
k? « k 0
1^1?
§^
^
4» VI-
?c§
« »- en
C k.*"'
r- 41 l/l
» IO
4) » 4)
I/I P k
,_
1
4-> P C
»- O. 4)
0 ' 4) U
4-> U k Q.
4) 4-> 41 £
0£ *- > f-
»
k. 3
O ««-
<4-
^P
*« C
O 10
o *«
4-> E
^
»« 4) X
O f O
« 4J Z
^2 k
C f
i "o
4) 10
4-> cT
VI IO 4^
C I/I
Jl 3
il
o
4-*
8
C
«r»
C
J
1
*o
> r
3
vT**"
«- 41
t|
ki
V ID
o-
I/) 3
c v»ac
^^ *O ^J
1
*|fc2
r- 4^ r- 4->
»- « -r- C
O U P 4j
H- «^ 41
O »- 41
k p 0. J=
41 4J k. Q. U
VI t- ID 3 4)
S C r VI 4-1
VI
£
H-
4)
41
VI
k.
41
T>
_
*
k.
O
«*-
O
*» x
pi
^v
X O>
2^
C 41
O 1 E
4J *3 V)
1C VI
U 41 k
-^- "O O VI
«*- «*- k.
v I/I 41
k. 41 C r
4J »- P f-
»- C f- O
C 10 4J XI
$154.
T P*
iio^^
§
r-
4J
U
4J
p*» .^
51
I/I
1 41
C >
8 4J
^
*3Z
VI ^
VI
Jg'B
!*
C u
k. E
51*"
c k.
O 4)
85
to vi
JC
1 4» U
k. ^ 41
O P 4) 1 4J VI
o en E -o »- k
k.-»- > 4> 3 , k
4J O f £ 4J «
*O VI k. U U r
0 P 10 4J C-»- C
< k ex 41 *- xi o
I/I
o
VI
VI
i
4-1
u
11
ex C
£0«4-
1
k. C
^- o
^ «M
1 k. U
COS
">,4l t
p^ ^
» «- X
« 4JO
k 0 Z
4) 41
C «*- 4J
4I««- U
CO 41 41
k
VI
4)
> T3
- 41
4-» VI
5
^O 4->
«P
^
Six
VI
41
£
i
^»
u.
k I 4J
4» «*- -
4^ f- r 3
^ 41 i/i
4|» 4^
« *- k
VI 1o 4J V> .
g > 0 C
JC 41 IO 1e c k
IO IO T- -r- 4>
"O U 4-» JZ JZ
§»- C U 4-»
41 4-» O
U 0. k^-
« CX 41 Jt l»-
(0 « It- VI O
^,
c en
V C
VI ^ 4->
t- « 4>
u x:
>>
u « »
- IO
ISKi
»- ^ 41 41
XJ E 4J
E i/i E vi
O 4) P >>
(_> k u un
4>
k.
10
VI
4)
10
4)
ft 4-*
P IO
»«- 4)
O *-» X>
i en 41
C 10 C 4J
O f 10
»- C 41 k
4J 4> X> 4J
vi en vi
3 5 V) C
O *- 3 4>
«_> C ««- "O
^
« VI
»**
4V 3
k V
4> 4»
4-> X
El
XXIV
-------
o
01
o
o
m
o
o
a:
o
II
<:
o
z
o
«o vi
t-o
i
f §3$
« t- Xt <*-
£322
c
u
VI JZ OI VI
- 4J C "O
to § c
u o
VI f
C O 4-1 .
« 4J «3 a.
o» to o E^
u .a >,4-> >,
E 3 3
o TJ -r-
0> I- t-
f 01 b
01 >, i- vi
e 4-» 4J -
f f VI O
O ^- 3 00
6 4-» C r-
O 3 ->-
u°
o> a> ** o
t/1 «- VI 4->
VI
I Si 0*
01 *-»-
"o. oi ^
CXH- 25
9 TD
V>3E C f-
«t)*42^
O >» I- VI
»- »- VI O
VI 4-» i 3 00
VI C f- T> O>
£i^=c
C 4-> 10
O CO.
u - oi 01 E
M 0> Vl^
VI C « k.
Q. >» 0>
fc f O O> >
*J 4J -^ * C
CO
I
to
o
VI
VI
£
Ol
"^
u
UJ
VI
g
0) O
»- Tj
vi u
X »
1. 1 r>
0) U -0 Ol
«» 0) ^3i-
-^- 4J ^
3 f »- 4J (O
cr o> E 1/1 a.
Ol 3 t- >, 10
o: «*- Z- u
r-
E
j?
t.
«£
i2x
A aE
*
^ rz
Ol 4-> VI
4-» C V-
VI OJ O
^f ^
01 O
C 4J
»- Q.
3 Ol
O" O
Ol U
QC 10
1
U VI
«*- t.^
o^"£
in x
o -o
o ar 01
»« *0) 4.
O 3 3
^
o
^
c a» £
5. o
>.4->
O 4-» VI
01 n-^^- C
t- »X> » XI O
»- VI -O X -^- U
3 r- 3 O VI f-
o- p J- vi >
Ol t U U O C
QC 4J VI O Q- Ol
L.
^^ 3C
A "O
oS|
o
C VI
cl?
- 4J
vi u
X
1. 1 r>
&
ef
*J C
C JZ
O 0
O Ol
Ol 4-> VI
4-» C V-
VI OJ O
ft 3
O Ol U
c
u o
>1 VI
r y
Si
58
»-
C 01 L?
5. o
>>+»
0 4-» VI
Ol O 3
+* 4-> ja
VI p g
01 £ o
t Q. 0
c
^ o
N 4-1
f- VI
^3 3
11
Ol
o o>
Sfe
£
u.
^_f
VI C
IO O>
0>E
4^
0» 10
5e
U. 4J
XXV
-------
expressed as an increase or decrease in fuel consumption for a constant output. Generally, flue gas
recirculation and off-stoichiometric combustion have very little effect on efficiency. In some cases,
taking burners out of service may result in reduced capacity. Low excess air and reduced air pre-
heat have a slight impact, usually less than 1.5 percent increase in fuel use; although, significant
reductions in air preheat (~ 150-200K) can have a much greater impact (~ 3-4 percent increase in fuel
use). New designs should significantly reduce any adverse efficiency impacts.
Emissions of other pollutants, CO, HC, particulates, sulfates, and organics, can be altered
by the use of NO control. Generally, these changes have been acceptable. In some cases specific
consideration of other emissions has been given in the design or method of application of the NOX
control technique.
Prime movers include stationary reciprocating internal combustion engines and gas turbines.
For the former, "dry" methods such as spark retard, air/fuel ratio change, and derating work well,
providing NO reductions of 10 to 40 percent while fuel consumption increases 2 to 15 percent.
Water injection ("wet" control) is currently the most effective technique for gas turbines, reducing
NO up to 90 percent at costs of 0.4 to 14 mills/kWh, depending on the turbine's application. "Dry"
control techniques show potential, but it will be a number of years before their development will be
complete and they will be ready to be applied to large production turbines.
The energy impacts of applying NO control to internal combustion engines and gas turbines
are manifested almost exclusively through corresponding increases in fuel consumption. Since both
types fire mainly clean fuels, the impact on other emissions is confined primarily to HC, CO, and
particulates (smoke).
OTHER COMBUSTION PROCESSES
Space heating, incineration and open burning, and industrial process heating are additional
combustion sources of NO . Residential and commercial space heating contributes 9 percent of the
nation's stationary NO emissions. Emissions of CO and smoke from the major equipment types,
residential and commercial warm air furnaces, can be controlled by burner maintenance, tuning, or
replacement. These techniques are ineffective for NOX reduction, however. The most promising pros-
pect for NO control in space heating systems is for new equipment applications. New low NOV sys-
X X
terns are available at a cost of 10 percent or more above conventional systems. These systems are
capable of reducing NO emissions by more than 50 percent, while increasing operating efficiency
by more than 5 percent.
xxvi
-------
There has been negligible application of combustion modification to incineration and open
burning and to industrial process heatfng equipment.
NONCOMBUSTION PROCESSES
Noncombustion-generated NOX, only 1.7 percent of stationary emissions, is produced mainly
during nitric acid manufacture. NOX control methods include extended absorption, wet scrubbing,
and catalytic reduction. Catalytic reduction was initially practiced but because of catalyst costs,
fuel costs and changes in the operating conditions of nitric acid plants, greater use of the extended
absorption and wet scrubbing processes have been employed more recently. Other minor noncombustion
sources are mainly those that use nitric acid as a feedstock. Control methods are similar to those
used for nitric acid manufacturing. Table S-4 gives a summary of tail gas abatement processes and
applications.
Xxvii
-------
to
«t
o.
o
II
u
a:
o
o
o
V)
41
₯
IO
X
LU
I/I
L.
O
C
o>
u
^
J
Comments
xi
o
*->
y
I/I
4>
U
o
i.
Q_
,c
IO
IO IO
|g
r- ^
X: 10
f- 4->
- 'Z
^ .
-o c
v» O
VI 1
C XI
o xi
4-> IO
O C
in o
CO t-
u
-O 3
CD 4)
a: u
CO U
CM t
^^
0
^H
J»
1
O
0
0)
u
I/I
^
«/l
^J
c
IO
"ex
*
"O-
^/r
c
o
in CM
in r-»
CD *
0*10
m o
1 -C
CM l/)
^
5»
l
V*
Usually cannot meet
NSPS without other ad(
on technology or low-
ered add product
concentration
c
3
r- X i.
O O 41
> 41
41 -Pr-
Or- f-
C t- XT
-. .0 o
C
o
1
ex
0 i-
41 O
r^ Vt
""" fc>
^*n «c c $
x r- c -a
en c 41 -r-
^j ^j ^
C * 4/) * ^C
IO to r -»J XI T-
«C CX
x-^. ^-«. «/) 41 -Q
>» >>cn u c
"O 4^ *^l ? t»
to O in 10 to
C C 41 r
O O O 4J CO U.
4-> O 4-> C
N IO «
O IO Of 4-> «
VO > O Q. CO
n ro 10 E
""^ * *-' £ rr «
M O Q. (_
«O ^^4J O
cn o .cn-r- o u
^1 * § ^ S
r«. 41 CM E E t-
CM x: r«* E O «
CO O CM «? O i2
"(aT
XI «/l
%- to
«O i-
x: o
u s.
4J IO
T- a.
u
CX 4)
TJ
C
U- 10
O3
T-"
4/1
c?
!
4^
C ^"*
IO «/»
r- «/>
ex oi
c o
* J-
cx ex
"D C
o
t «r
4-*
4-» O
C (A
to xt
ex
a
CO r-
!c
en «j
>^
^«"
f^
4)
a
4)
3E
CX
o
e
I/I
'^^
S
^^ ^^
ccn
O
O 10
IO
'v.'o
S> «J
2 4>
S. 41 £1
3 -0 S-f-
to 4) CX 3
w> J- E cr
41 -r- O 41
t. 3 U C
a. cr
41 r- 41
4J t. 10 X)
41 C
f (0 O >,
J=CLf- «
o
o«
=
^^
c
ex in
t>%
*x?
TJ
«/) 41
C Z
O
in «
CM O
*-> IO
Cn O
O 4)
CM O
4)
X>
P*
=
Requires additional
compressor
01
4->
, i a.
W) ^3 jQ "^»
XI fO 3
10 x cr
Or- 41
XI Z IO
OI C C
10 <4- O O
<0 O -r- f-
£4J 4->
C f- CL4->
O O X» ET c
C »- t} O 4»
l-l 4-» 10 tO E
§
^
o ex
41 S-
XI 0
C VI
41 XI
X
LU
w
l""1 «
tA ^^
4>
r- OI
10 U
1. Q.
:8
4-> JC
i O
0.
e
c
gn£
££
4JTJ
_
r^ L.
ES
IO
t O)
3 C
^1
OCX)
< O
^
^^
IO
§
z^
c
c
3
O^
e
Ck.
e
"0
^£
£
2
O*
c
^2
e^
o t-
-^ IO
O 3
ZT ^ '
L. 10
4-> 4) 41
-C fO > i- 1
CD O 3 4)
r- 4) 0 U
-C r- 4) "O
J3 CC C W>
4-> IO IO 41
10 L. l_
41 41 -r
$- CX to *J 3
4) O 4> (0 CT
-> t. L. 41
M 4J 3 *J OC .
41 3 «/> i- C
-O J3 10 C O
41 . T-
W> 41 U E C 4-»
E I- O. 3 O 10
C 3 -r- -r- 1_
O W J- C 4-» 4*
*»- «/> 4) O 3 CT
O) i- O 1 *O H
O_ CXr- 10(^*4-
S
r \o -o
S^
fO
CX r- 10 r-
ttl CT> d)
2 t 5^
IO IO O IO i-
CX O 4-» E
a z o 10
P * ** c
o "O 10
CD C 4-> «/) >,
Q C 4J 0
«-» E 0 c
>>£ E «
10 to u S> T
T5 x »* w ex^<
-^ 4) ee u- >,
c*~ - ***§
o ut k «>.
4^ 4-> O OI O 10
4-> CJ t- «4- C
O -t- U o
^i- 0 3 4>0
o TJ.OO
a c c x) vo
^. O i-« 3>
a> . ». ».
3E ^- > O T5
o en .C * en
en r- <_> o CM E
OI
i_ *-«
3 cn
4J 10 C
«/> f- T-
IO C XI
O E 3
o E J-
O 10 U
C3 '** irt
u -o i
O 4> T3
*-» 4J O
May require an evapora
to produce a concentra
ammonium nitrate by pr
uct. No refrigeration
required.
IO
r- 4>
f- t.
2 3 i_ x
5O 0 O
4J Z
^Q "^ *» Q}
241 3 O 0
W> i £ o
u 10 o E 4)
t/> en «/> 10 u.
*io en
u c
§""£
jO
f 3
O 1.
3T
xxviii
-------
TJ
Ol
TJ
3
U
§
U
a:
oo
o
o
CQ
<
10
Ol
i
10
X
UJ
in
I_
O
c/>
C
u
_J
Comments
I
t/>
t/)
01
u
o
a.
£i
01
JC
IO
^
o
in
10
10
_i
01
10
ai
IO
c c
i- 10
0*0.
^_
ai
TJ
CD
cQ
»
10 jc
IO CD
cn-r-
j= +J
*io «4- 0)
i. f- -P
4Ji C
IO IO U
c u
i- cv
«o E O
in u
C Ol X
o c o
0 3 Z
O CVJ
cvj -MO
o o
CVJ
TJ 3C «
C 0
10 1- CVI
O I
X
Z Z CVJ
u z
10
J. 5 C
3-^-0
00 X ««-
ai
u
u ai
>, <0
iO O
IO
o
I/)
M O 10
f TJ
r ^*. C
CD 01
at en
«O f-* L.
00 T-
z,
o . «
CD C JD
c o E
Ol O. O
t U. O
M O
JQ
10 r^.
O.TJ
^ Ol i^
O TJ
O C >
0 +>'>, .
£ x 10 10
M 01 TJ CD
^>
.
Ol
f
^J
O) in t- o>
co o C
UJ Q. _K 0-
TJ
1
IO
10
r
«
C
E , «o
10 i- i
&.OIO.
f- «f- O
Z 10 r-
<4_
CD
Q) 1
10 X
C $- 01 C
IO O) C
0 Q. J= i-
O *> +
* T- C
IO O £1.
i- +J C
C TJ W
O 01 0) X
10 in T:
C CO
IO Ol Ol "C
01 0- -4J C
w) x «»- a
=> ai o *J
sis1
X » Oi
O O *>
X4-> CVJO
O Z Ol
Z 10 «4-
M C C CO
TJ IO +J
Works at low or high
pressure
Energy recovery
usually not possible
Can achieve very low
emission 1f desired
C
o
ai
>
ai
TJ
01
TJ
C
^C
O
I-H
^_
a
o
ilted success
e
_J
Oxidation of NO -
N02 catalyzed by
heterogeneous ca-
talysis before
absorption
3
O 10
Ol »-
C IO
CT>r^
e*
Ol 10
l->0
TJ
Ol
J_
«^
^^
o
^>
c
t!
3
U
4-*
O
C
c
10
0-
IO
-3
C
IO
C 00
IO
"a.
c
cvj f-
r
IO
U
^"
E
ai
JC
o
in
01
Oxidation with
KMn04 (KMn04
electrolytlcally
reclaimed
o
^
^>
15 £"
U 0
- in
v ^C
U
O T?
M 01
10 I
U O C
C Z 10
» IO
10 E TJ
0) U
*?» *
O r
t- oo rj
Ol 10
Z => l_
^ 4»
>> >> 0
10 «* «O C
TJ fx & »r-
IO r «O
O 0 C
4J IO 4-> C
r Ol
in <; in |
IT) If)
TJ Oj TJ O
cr> 01 o> to
Ein E o>
O Ol O !-
ir> CD m :*£
^^
Ol
TJ
yQ
l-
Z 10
o
f- C
to O
J- "c
3 rs
O- * ^
1
O. CD IO
1C C 1 &-
O T- C Ol
o o.
TJ «J U O
c >,
IO (J O 4->
-» c
>> m +> c re
CDTJ T- f- r
1. C «+- Q.
Ol IO (LI
c E O > in
Ol Ol 4J Z C3 -M -U
«
Absorption by
molecular sieve,
regeneration of
the sieve by
thermal cycling
w
ID
r- Ol
3 >
U 01
Q) *^
t~~ I/1)
*
xxix
-------
-------
SECTION 1
INTRODUCTION
Manmade oxides of nitrogen (NOX) are currently emitted at a rate in excess of 20 Tg (22 mil-
lion tons) per year in the United States. Over 98 percent of manmade NOX emissions result from com-
bustion with the majority due to stationary sources. Combustion generated oxides of nitrogen are
emitted predominantly as nitric oxide, NO, a relatively harmless gas, but one which is rapidly con-
verted in the atmosphere to the toxic nitrogen dioxide, N02> Nitrogen dioxide is deleterious to
human respiratory functions and, with sustained exposure, can promote an increased incidence of
respiratory ailments. Additionally, N02 is an important constituent in the chemistry of photochemi-
cal smog. The N0/N02 conversion in the atmosphere promotes the formation of the oxidant ozone, 0,,
0
which subsequently combines with airborne hydrocarbons to form the irritant peroxyacetylnitrates
(PAN). Nitrogen dioxide is also a precursor in the formation of nitrate aerosols and nitrosamines,
the health effects of which are under study by the EPA. Because of the quantity generated and their
potential for widespread adverse effects on public health and welfare, nitrogen oxides are among the
atmospheric pollutants for which standards and regulatory controls have been established both by the
U.S. Environmental Protection Agency (EPA) and by State and local agencies.
As part of the regulatory control program, the U.S. Environmental Protection Agency (nee the
National Air Pollution Control Administration) published "Control Techniques for Nitrogen Oxide
Emissions from Stationary Sources" (AP-67) in March 1970, as one of a series of documents summariz-
ing technology for the control of air pollutants. Since the issuance of AP-67, there has been con-
siderable activity in both regulatory control of NOX and development of emission control techniques
for stationary sources. Under provisions of the 1970 Clean Air Act Amendments, the EPA promulgated,
in 1971, a National Ambient Air Quality Standard for N02 of 100 Mg/m8 annual average. To achieve
and maintain this standand, a number of State and local agencies have established NOX emission con-
trol standards for new and existing large stationary combustion sources and nitric acid plants.
Additionally, Standards of Performance for New Stationary Sources were promulgated by the EPA in
1971 for steam generators with thermal input greater than 73.2 MW (250 x 106 Btu/hr) and nitric acid
plants. Standards of performance for stationary gas turbines were proposed on October 3, 1977.
-------
Standards for stationary large bore reciprocating engines are in preparation. The standard for
large steam generators is under review to determine if additional stringency is appropriate.
The NO control technology development to support the implementation of these standards has
shown widespread advancement since the publication of the AP-67 document. Efforts have proceeded
on methods which suppress NO formation, through combustion process modification, and on methods
which remove NO from the flue or tail gases, through stack gas treatment.
Combustion process modification is the preferred method for control of stationary combustion
sources accounting for 98 percent of stationary source NO . Process modifications have been exten-
sively applied to retrofit of existing utility and industrial boilers and gas turbines firing gas
and oil. The significant role of fuel bound nitrogen in NO formation with the firing of coal and
X
heavy oils was shown early in the control development effort. Current activity is concentrating on
refinement of fuel NO control methods for application to advanced designs of coal-fired combustion
equipment. Progress has also been made in the design of low-NO residential and commercial space
A
heating systems.
Stack gas treatment is the preferred method for control of NO emissions from stationary non-
combustion sources. These sources, primarily nitric acid plants, contribute less than 2 percent of
nationwide stationary sources NO emissions, but can present a serious local hazard. Several con-
trol techniques, including extended absorption, catalytic reduction, wet scrubbing, and molecular
sieve absorption, have been developed and implemented on existing and new equipment. Reductions in
NO in excess of 95 percent have been demonstrated.
The purpose of this report is to update and revise the original AP-67 document by incorporat-
ing improved emissions estimates and NO control technology developments since 1970. Emphasis is
given to identifying the significant stationary sources of N0x emissions based on the most recent
emission factors and fuel consumption data (Section 2), summarizing the developmental status of
candidate NOX control techniques (Section 3), and reviewing the effectiveness, cost and user
experience with the implementation of NOX controls on large combustion sources (Section 4), other
combustion sources (Section 5), and noncombustion sources (Section 6). Also included in these
sections is information on the energy and environmental impacts of the various control techniques
as required by Section 108 (b)(l) of the 1977 Clean Air Act.
This r port is concerned only with quantifying and controlling stationary source NO emis-
X
sions. The effects upon health and welfare of nitrogen oxides and their secondary atmospheric reac-
tion products are considered in two related documents, AP-63, "Air Quality Criteria for Photochemi-
cal Oxidants," and AP-84, "Air Quality Criteria for Nitrogen Oxides". Both of these documents are
under revision by the Office of Air Quality Planning and Standards.
1-2
-------
SECTION 2
CHARACTERIZATION OF NOX EMISSIONS
This section presents a nationwide inventory and projection to the year 2000 of stationary source
emissions of oxides of nitrogen. Section 2.1 defines NOX and summarizes the basis of its occurrence
in stationary source combustion. Section 2.2 describes the standard EPA method for analysis of
source and ambient NOX concentrations. Specific stationary source equipment types are described in
Section 2.3. The NOX emission factors, fuel consumption rates and annual N0x emissions for each of
these source types are also tabulated in Section 2.3. A summary of nationwide NOX emissions and
fuel consumption by equipment application sector is given in Section 2.4. Projections of these
emissions to the year 2000 are given in Section 2.5.
2.1 DEFINITIONS AND FORMATION THEORY
Seven oxides of nitrogen are known to occur: NO, NO,,, N03> N20, N^, N^ and N^. Of
these, nitric oxide (NO) and nitrogen dioxide (NO,,) are emitted in sufficient quantities in fuel
combustion and chemical manufacturing to be significant in atmospheric pollution. In this document,
"NO/ refers to either or both of' these two gaseous oxides of nitrogen. Nitrogen dioxide is dele- '
terious to human respiratory functions and is a key participant in the formation of photochemical
smog. Nitric oxide, taken alone, is relatively less harmful but is important as the main precursor
to N02 formation in the atmosphere.
Approximately 95 percent of oxides of nitrogen emanating from stationary combustion sources
are emitted as nitric oxide. Two separate mechanisms, thermal NOX formation and fuel NOX formation,
have been identified as generating N0x during fossil fuel combustion.
Thermal NOX results from the thermal fixation of molecular nitrogen and oxygen in the com-
bustion air. Its rate of formation is extremely sensitive to local flame temperature and somewhat
less so to local concentration of oxygen. Virtually all thermal N0x is formed at the region of the
flame which is at the highest temperature. The NO, concentration is subsequently "frozen" at the
level prevailing in the high temperature region by the thermal quenching of the combustion gases.
The flue gas NOX concentrations are therefore between the equilibrium level characteristic of the
2-1
-------
peak flame temperature and the equilibrium level at the flue gas temperature. This kinetically con-
trolled behavior means that thermal N0x emissions are dominated by local combustion conditions.
Fuel NO derives from the oxidation of organically bound nitrogen in certain fuels such as
coal and heavy oil. Its formation rate is strongly affected by the rate of mixing of the fuel and
airstream in general and by local oxygen concentration in particular. The flue gas N0y concentra-
tion due to fuel nitrogen is typically only a fraction (e.g., 20 to 60 percent) of the level which
would result from complete oxidation of all nitrogen in the fuel. Thus, fuel NOX formation, like
thermal NO formation, is dominated by the local combustion conditions. Additionally, fuel N0x
emissions are dependent on the nitrogen content of the fuel. The NOX emissions characterization
detailed in this section, therefore, takes account of variations in equipment operating conditions
and in fuel type which influences the emissions as well as the potential for control. Additional
discussion on thermal and fuel NOX formation mechanisms is given in Section 3.1.
Oxides of nitrogen emitted in the byproduct streams of chemical manufacturing (nitric acid,
explosives) are predominantly in the form of N02- The N02 concentration in the flue gas is typi-
cally at the equilibrium level characteristic of the chemical compositions and temperatures required
in the manufacturing process. The NO emissions from noncombustion sources are then much less sen-
sitive to minor process modifications than are combustion generated NOX emissions.
2.2 SAMPLING AND ANALYSIS METHODS
The standard EPA method for compliance testing of Nt>x from stationary sources is the phenol-
disulfonic acid (PDS) method. This method was developed for the measurement of nitrate in solution
by Chamot around 1910 (Reference 2-1). The specifications for the PDS method are given in
Reference 2-2. Briefly, the method requires that a grab sample be collected in an evacuated flask
containing a dilute sulfuric acid-hydrogen peroxide solution which absorbs the nitrogen oxides,
except nitrous oxide (N20). The sample is then processed following the procedures of Reference 2-2.
The absorbance of 420 nm wavelength light by the treated samples is then measured. A calibrated
relationship between absorbance and NOp concentration is used to relate the measurement to the sample
NOp concentration.
The advantages of the PDS method include the wide concentration range, minimum number of
sample handling steps, and lack of interference with sulfur dioxide in the flue gases. The disad-
vantages are the long time elapsed between samples, a possible interference from halides, and the
inherent problems with grab sampling.
2-2
-------
Chemiluminescence, the Federal Reference Method (Reference 2-3) for ambient NO sampling, is
also a popular source testing technique. Although it cannot be used for compliance testing, its
continuous electronic measurement feature is advantageous for use in emissi'on control development
programs.
2-3 c2rinoENT DESCRIPTIONS> EMISSIONS ESTIMATES, EMISSION FACTORS, AND FUEL USAGE BY APPLICATION
oLL1 UK
The rate of emission of oxides of nitrogen from stationary combustion sources is dominated
by equipment design characteristics (combustion intensity, fuel/air mixing pattern, combustion gas
temperature history) and fuel characteristics (combustion temperature, fuel nitrogen content). A
previous NOX emissions inventory for 1972 (Reference 2-4) classified stationary combustion sources
according to the design characteristics known to influence NOX emissions. A total of 137 combina-
tions of equipment type and fuel type were identified as having significantly different potential
for NOX emissions and/or NOX control. The emissions data cited in this section are an update, to
1974, of the 1972 inventory of Reference 2-4.
An overview of stationary sources of NOX emissions is provided in Figure 2-1. The first
division is by application and the second by use sector. The six applications encompass all major
sources and the cited sectors include all those of importance within each sector. Steam generation
is by far the largest application on a capacity basis for both utility and industrial equipment while
space heating is the largest application by number of installations. Internal combustion engines
(both reciprocating and gas turbines) in thi petroleum and related products industries have gener-
ally been limited to pipeline pumping and gas compressor applications. Process heating data are
not as readily available, but the main sources appear to be process heaters in petroleum refineries,
the metallurgical industry and the drying and curing ovens in the broad-ranging ceramics industry.
Incineration by both the municipal and industrial sectors is a small but noticeable source, pri-
marily in urban areas. Noncombustion sources are largely within the area of chemical manufacture,
more specifically nitric and adipic acids and explosives. The final description level in Figure 2-1
is the important equipment types. Although these equipment categories do not include all the pos-
sible variations or hybrid units, the bulk of the equipment is included in the breakdown.
The emissions inventory from Reference 2-4 for the significant stationary source equipment
types was updated to 1974 using the most recent emission factors and fuel consumption data. These
emission factors were obtained from EPA publication AP-42 (Reference 2-5), its three supplements
(References 2-6, 2-7, 2-8) and recent field test studies (References 2-9 to 2-13). A rating of the
quality and general applicability of these emission factors for each sector is given in Table 2-1.
2-3
-------
APPLICATION
SECTOR
EQUIPMENT TYPES
STATIONARY
SOURCES OF <
NOX
STEAM
GENERATION
-PRIME MOVER
. SPACE .
HEATING
-INCINERATION
PROCESS
HEATING
ELECTRIC POWER.
GENERATION
INDUSTRIAL
PROCESS STEAM
ELEC POWER GEN.
OIL AND GAS
PIPE LINE PUMPING
NATURAL GAS
PROCESSING
RESIDENTIAL-
COMMERCIAL-
MUNICIPAL
INDUSTRIAL
PETROLEUM
REFINING
METALLURGICAL-
L-NONCOMBUSTION-
CERAMICS INDUSTRY,
BRICKS, CEMENT,
GLASS
CHEMICAL
'MANUFACTURING
FIELD ERECTED
"WATERTUBE BOILERS r FIELD ERECTED
.WATERTUBE BOILERS-^
I PACKAGED
-t
PACKAGED
FIRETUBE BOILERS
RECIPROCATING
1C ENGINES
GAS TURBINES
-FURNACES
CAST IRON BOILERS
BOILERS
WATERTUBE
.FIRETUBE
CAST IRON
FURNACES
PROCESS HEATERS
.FLUID CATALYTIC CRACKERS
HEATING AND
'ANNEALING OVENS
L-COKEOVEN UNDERFIRE
-OPEN HEARTH FURNACE
-SINTERING OVENS
KILNS
FURNACES
NITRIC ACID
ADIPIC ACID
EXPLOSIVES
Figure 2-1. Stationary sources of NOX emissions.
2-4
-------
TABLE 2-1. QUALITY OF EMISSION FACTORS FOR OXIDES OF NITROGEN,
Sector
Utility
Industrial
Reciprocating 1C
Engines
Gas turbines
Residential
Commercial
Incineration
Process heating
Noncombustion
Fuel
Bituminous
Lignite
Anthracite
Oil
Gas
Coal
Oil
Gas
Gas
Oil
Gas
Oil
Gas
Oil
Coal
Gas
Oil
Quality3
A
B
B
A
A
B
B
B
B
B
A
A
A
A
B
A
A
C
D
C
A - good, based on high quality field measurements
B - average, based on limited number of field measurements
C - marginal, sparse data base
D - Inadequate data base
2-5
-------
A grading of "A" means the quality of the emission factor is good, i.e., based on high quality field
measurements of a large number of sources. "B" indicates average quality or based on a limited num-
ber of field measurements. "C" refers to a sparse data base, or to data which is of marginal quality
and "D" indicates an inadequate data base.
The emission factors cited herein are for the baseline operating conditions without the use
of NO controls. "Baseline" refers to nominal settings of process variables such as unit load,
excess air levels, and combustion air preheat as well as to the most representative design type with-
in a given equipment class. It should be noted that stationary source NOX emissions are strongly
dependent on small variations in design types, fuel composition or process operating conditions.
Thus, for those equipment sectors given a rating other than "A" in Table 2-1, the sparseness of the
available emission data may preclude the specification of a true baseline emission factor for all
significant design types.
2.3.1 Utility Boilers
Emissions and Fuel Use
Utility boilers are field-erected watertube boilers ranging in thermal capacity from 30 MW
(100 x 106 Btu/hr) to around 3000 MW. This equipment category includes the large majority of utility
and industrial electric power generating boilers. Field-erected watertube boilers operate at steam
temperatures up to 840K (1050 F) and steam pressures up to 26 MPa (3800 psi). Depending upon manu-
facturer, units greater than about 2250 MW operate at supercritical steam pressures above 24 MPa
(3500 psi) (Reference 2-14). In general, utility boilers recover up to 90 percent of the heat with
waterwalled combustion chambers in combination with superheaters, reheaters, economizers and air
preheaters. Approximately half of this heat energy is absorbed by radiant heat transfer to the
furnace walls.
Although there are some differences among utility boiler designs in such factors as furnace
volume, operating pressure, and configuration of internal heat transfer surface, the principle dis-
tinction is firing mode. This includes the type of firing equipment, the fuel handling system, and
the placement of the burners on the furnace walls. The major firing modes are: single- or opposed-
wall fired, tangentially fired, turbo fired, and cyclone fired. Vertically fired units and stoker
units are used to a small extent in older steam generating stations. All of the major firing types
can be designed to burn the principle fossil fuels - gas, oil and coal - either singly or in com-
bination. However, the cyclone unit is primarily designed to fire coal as the principal fuel.
2-6
-------
In addition to differences in firing mode, coal, depending on its ash characteristics, is
burned in either a dry-bottom or wet-bottom (slag tap) furnace. Dry-bottom units operate at tem-
peratures below the ash-fusion temperature, and ash is removed as a solid. Wet-bottom furnaces
melt the ash and remove slag through a bottom tap. Although wet bottom units were once used exten-
sively in burning low ash-fusion temperature coals, they are no longer manufactured due to opera-
tional problems with low sulfur coals and because their high combustion temperatures promote NO
formation.
In single-wall firing (front-wall firing) burners are mounted normal to a single furnace
wall. Furnace wall area generally limits the capacity of these units to about 1200 MW. When
greater capacity is required, horizontally opposed wall-firing furnaces are normally used. In these
units burners are mounted on opposite furnace walls. Generally, capacities for these units exceed
1200 MW (Reference 2-14). Burners on the single-wall and opposed-wall firing designs are usually
register type where fuel and combustion air are combined in the burner throat.
Turbo-fired units are similar to the horizontally opposed-wall-fired units except that
burners are mounted on opposed, downward inclined furnace walls. Fuel and combustion air are intro-
duced into the combustion zone where rapid mixing occurs.
In tangential firing, arrays of fuel and air nozzles are located at each of the four corners
of the combustion chamber. Each nozzle is directed tangentially to a small firing circle in the
center of the chamber. The resulting spin of the four "flames" creates high turbulence and thorough
mixing of fuel and air in the combustion zone.
In the cyclone furnace design fuel and air are introduced circumferentially into a water-cooled,
cylindrical combustion chamber to produce a highly swirling, high temperature flame. The cyclone
was originally developed as a slagging furnace to burn low ash-fusion temperature coals, but has
recently been used successfully on lignite. Relatively high levels of thermal N0x formation
accompany the high temperatures of slagging operation. Due to the inability of this design to
readily adapt to low NOX opeartion, this type of furnace is no longer being constructed.
Vertical-firing furnaces were developed for pulverized fuels prior to the advent of water-
walled combustion chambers. These units provide a long-residence time combustion which efficiently
burns low-volatile fuels such as anthracite. Vertical-fired boilers are no longer sold, and rela-
tively few of these units are found in the field.
Stoker-fired units are designed for solid fuel firing. Unlike liquid, gaseous or pulverized
fuels which are burned in suspension, the stoker employs a fuel bed. This bed is either a station-
ary grate through which ash falls or a moving grate which dumps the ash into a hopper. The main
2-7
-------
types of stokers are overfeed and underfeed designs. Spreader stokers are overfeed designs and dis-
tribute fuel by projecting the fuel evenly over the fuel bed. Other overfeed stokers generally
deposit fuel on a continuously moving grate. Underfeed designs introduce fuel beneath the fuel bed
as ash is pushed aside by the newly introduced fuel.
Tangential firing, single-wall and horizontally opposed-wall firing and turbofurnace firing
account for 40 and 36 and 14 percent of the fuel consumed by utility botlers (Reference 2-15). In
terms of units, their distribution is 19, 59 and 8 percent, respectively. Cyclone, vertical and
stoker designs make up the remaining 14 percent.
Recent trends indicate a continued strong movement toward pulverized coal-fired boilers.
Many previously ordered oil-fired units are being converted to coal firing during the design phase.
The trend of the last 10 years to increasing capacities appears to have slowed with many utilities
electing to install two small boilers rather than a single larger unit (Reference 2-14). Industry
sales were particularly depressed in 1976 and 1977. Uncertainty about the nation's energy policy,
environmental regulations, mild summer load peaks, increased energy costs, and a 1975 reserve capa-
city of about 35 percent have combined to produce this situation.
Estimates for the uncontrolled NOX emissions from utility boilers were derived from the
1974 utility fuel consumption compilation for coal, oil and natural gas published by the Federal
Power Commission (FPC) in Reference 2-16. The consumption rate of each fuel type was prorated
according to firing type based on the procedures used irwReference 2-4. Emission factors for each
equipment type were obtained from AP-42 (Reference 2-5), AP-42 Supplement No. 5 (Reference 2-7),
and a field test survey of utility boilers (Reference 2-10). These factors were applied to the
individual fuel consumption rates to arrive at the annual NOX figures presented in Table 2-2.
The nominal heating values of the fuels were as follows: gas - 37.3 MJ/Nm3* (1000 Btu/scf), oil -
39 GJ/m3 (140,000 Btu/gal), and coal - 27.9 MJ/kg (12,000 Btu/lb).
A summary of the emissions and fuel usage with respect to firing type is presented in
Table 2-3. The resultant total nationwide annual NOX emissions by utility boilers in 1974 is esti-
mated to be 5.1 Tg (5.63 million tons). The corresponding total annual fuel consumption for 1974
is 16.3 EJ (15.4 quadrillion Btu).
Fuel consumption data from References 2-4, 2-15, 2-17, and 2-18, and the present report are
compared in Table 2-4. The corresponding NOX emissions are compared in Table 2-5. The NOX emis-
sions for both oil and gas firing compare very well between the present report and the recent study
of Reference 2-15. The discrepancy in emissions with coal firing is primarily due to the use of
more recent emission factors for lignite combustion and for cyclones in this report.
* The symbol Nm3 is used to denote cubic meters at standard temperature and pressure.
2-8
-------
ro
00
ae.
z
o
CO
cc
o
CQ
<
CO
0
CO
'i CM
LU O
Z
X
o o)
Z CO
4->
0
O co -».
r** t» CM
CO O O
CO +J Z
- 0
S ro O)
UJ U- C
"Tu **T*i--j
CU
OL
>ij
H-
r»
CU
"
r
CU
3
LL>
CL
1
cn
c
nr*
t.
r»
tir
0)
CL
H-
4J
C
i
CL
3
CT
LU
CM
O
CM
O
CM
in
r«.
CM
O
ej-
00
vo
ro
CO
3
O
C
'i
3
M
CO
CM
0
CM
in
r>»
CM
O
«sj-
CU
p.
c
Ol
_J
*rO
O
O
r*.
in
ro
in
CM
r^
r^
*j-
cu
ro
P_
lp_
1 *
CO
a
VO
P
in
ro
in
ro
ro
vo
«*
P_
ro
3
-a
CO
cc
P_
r-
O
00
j
8
CM
r>.
CM
in
cn
i
CO
ro
CD
O>
.JE
4-
LU
,_
ro
4^
C
CU
o>
c
1
CO
r»
"^
S
CM
CM
ro
ro
CO
VO
en
CM
CO
3
O
c
'i
3
CQ
in
0
»
ro
ro
vo
CM
cu
>r_
c
O)
_J
O
4-^
O
CO
£>
a
ro
O
O
^_
fi^
ro
O
in
ro
in
o
CM
CO
r--.
in
CO
3
O
C
f*
E
3
M
CQ
VO
O
CO
ro
0
£
cu
4.)
pi-
c
O)
_J
E
^
-M
O
CQ
CU
»
ro
O
p_
p»
0
in
CM
CM
ro
o
en
ro
0)
ra
p_.
p-
4^
CO
Q
cn
VO
CM
o
m
CM
CM
ro
CO
cn
m
ro
CO
pM>
(O
3
-a
CO
CD
cc
P_
P»
O
^
ro
*!
O
O
O
ro
ro
o
*3-
n
^
i
CO
rO
CD
>>r-
p ro
4-> 1
c -a
O CU O)
N CO C
r- O T-
S_ Cu S_
0 Q-'i-
n: o u-
00
^^
*t
O
in
CM
CM
ro
ro
CO
VO
cn
CM
to
3
O
C
r~
^
3
*->
cc
in
g
co
r^.
ro
ro
CM
CU
4^
r
C
01
-1
1
4->
O
ca
£>
a
"
ro
O
CJ
^
O
ro
O
in
h-
ro
in
en
en
VO
ir>
3
O
C
r-
E
4->
CO
in
0
CO
r^
ro
£
CO
cu
4-*
r-
C"
01
-1
O
+J
o
ca
cu
3
*
ro
O
O
,_
r**~
O
in
CM
CM
ro
O
cn
sr
ro
cu
4*
CM
r"^
o
in
CM
CM
ro
r>.
VO
cn
en
ro
3
a
c:
rr~
E
3
+*
CO
,
O
in
CM
CM
ro
in
ro
Q_)
4_>
r-
C
cr
_j
o
4->
O
CQ
^>
O
*
'ro
O
O
o>
c
s_
LU
p_
ra
u
^"
4->
$_
cu
^
CD
^o
o
0
cn
in
in
in
r*.
vo
O
^3
0
C,
"r~
E
("j
ca
in
o
*a-
CO
ro
«*
ro
ro
a;
4-j
Bp»
C
o-
_J
o
4J
O
ca
cu
3
*
rO
O
O
,
0
in
CM
CM
ro
CO
o
ro
4~>
0!
en
r
o
m
CM
CM
ro
S
CO
uo
p""~
* ' fCI
f- ' : H}
p
T3
^.
ro
0
m
^i-
CM
CM
cn
CO
r~
in
f*"
o
in
*
CM
r^.
cn
in
-t~* i i
t/ll (/)! 1 i
a
eg
i
r~-
0
CU
c
o
u
^k
CJ
ra
o
CJ
i.
eu
T3
ra
cu
L.
ex
CO
rO
0
o
T3
CU
CU
<+-
4.
CU
-o
c
ID
CO
4. 4-
eu cu
-o S- -o ^
cu o cu o
M CO) 4-> CO
IJ (J
cu cu cu cu
4- -Ct 1- -Q
LU 3 LU 3
1 -t-> 1 -M 4-
o 4. -a i- cu
r- CU r- CO J*
CU 4-> CU 4-> O
p« flj »p» (Q +J
U_ 3 U. 3 CO
CU
CL
a.
a
cu
T3
0)
In
O
cu
o
CL
eu
s-
OJ
rO
CO O
C Z
o
r- CO
co ro
to
M C
C O
CU *r-
>> O
CQ U
-D
2-9
-------
o
00
O
CO
o
o
CO
o
^_^
-l_>
c
CD
0
i-
CU
O.
C
E
r~
o
u
JO
-3
LLJ
*
c
o
1 1
Q.
E
3
CO
C
o
CJ
3
U.
+J
E
O)
CJ
t-
O)
Q.
c
E
3
^
O
CJ
JO
h-
*
C
o
CO
r
E
LLJ
X
o
're
0
I
^
re
0
o
^1
O
CO
re
CD
n~
re
0
(
r~
re
0
c_j
r
o
re
&
>)
t
en
c
r
tZ
o"
o
2-
CM
CM
LO
VO
^^
^>
^~
^j-
«- -
CM
i
**
^
_
co"
cr>
co
" '
«3-
Cn
r
r-*
p
P*V
C\J
***
VO
^
en
O
r^
CD
CM
*
LO
r^
LO
-'
co"
^f
CO
«*_«'
*5j"
I
C\J
r
O*
^J1
CM
CO
00
r~
o
CO
CO
"
^-"
00
CJ
re
r
M
0)
CD
C
re
P
CO
CM
CM
r-.
f*^
CO
.^^
r
cr>
^>
^ -
LO
CD
00
r-1
LO
O
cr>
CM
""""'
r
p^
00
o
00
O*i
CM
*^-^»
vo
0
0
-
oo"
LO
CM
00
^
CO
-1
CM
1"
CM
*_*>
LT)
ro
fN^
0
^
^iD
CO
o
CO
CM
O
CO
LO
CO
^" ^
co
o
CO
CD
re t^-
u re
N 4->
r- C
I- 0
0 i-
1C <*-
CT>
LO
CM
0
CO
CM
^
^-^
r-
CM
^-*
>- -
LO
0
,_!
LO
O
o^
CM
*""*
,_
f^.
00
o
^
CO
^*
^_^
^«
LO
^t
-
LO
00
CM
^
LO
«*
'-L
CM
r*~
CNJ
»* rf-
o
f*s^
CD
-
*
vo
co
o
CO
CM
o
"o\
o
LO
*~*
CO
o
1
§
CO
O r
Q.I
Q. re
o i
p
^^
^_,
CM
LO
~
rl.
^*^
o
*
I
f
^^^
0
o>
o
rJ
r«.
CM
"*~*
CM
VO
o
0
1
co
CM
r
VO
CM
VO
O
«J-
r^»
P..
s^.^
vo
O
vo
0
vo
CM
*-*
O
CM
O
o
0)
c
o
"o
'-
CM
^_,
^.
CO
o
^*^
^J-
^^
x^^
^J-
co
*
o
1
p
CM
'
0
r
O
o"
^J-
^.^.x
o
r-»
0
r
re
CJ
^J
>
C\J
r
>«^
Cft
cr>
Q
^^^
0
CM
^^^
CD
o>
0
1
1
o"
p..
en
^^
o
o
^J-
^v
«- ~.
en
0
o
52
O)
1
CO
o
o
en
o
CO
vo
^ ^
0
o
^^^
LO
CO
en
en
^-^
o
o
"""""^
CO
en
en
CM
o"
O
i^^f
VO
r*.
CO
CO
o
O
LO
O
LO
, ^
O
o
^_ ^
^j-
CO
co
c
o
CO
vo
.
o
^_^
o
o
*""*
00
LO
oo
CJ
^^
re
0
a>
jr
4J
d)
o
L.
Q.
LO
en
>>
cu
X
O
Q.
a.
0)
Q.
Q.
o
0)
T3
O
a.
a>
ai
i-O
re z
(/) CO
c re
o
r C
P O
re !-
c o
ai u
C 0)
o u
U i-
2-10
-------
IO
"3
CO
a:
at
0
j-
3
O
CO VO
LU Ol
Zr-
U 1
at at CM
o «->
to o at
Q. i- O
io CL c
o at
at in at
o> at
+,
fc-O?
o
in
i
CM
at
jrj oj at
at
at
OL
\
at
» CM O
co r>- c
co 01 at
3£ « l~
co
s-
oo*
CM
at
i at
i i Ct
at r^. c
4-> 01 ai
Mr- J.
ia at
00 <«_
at
en
r
at
3
LU
21 r^ CM
CO VO Ot
Ol CO VO
O CO ^P
VO O» 00
SCO in
vo
Ok CM CO
in oo vo
CO Ol iv.
O< Ol CO
O> CM CO
O VO VD
t*» O CO
Ol CM «3-
OO CO CO
00 CO ^.
oo rv. o>
00 CM CO
CM in ot
r- CM r-
00 CO O
in CM *t
IO r- (/)
O H- 10
0 O CS
00
a*
oi
CO
0
in
o
CO
vo
CM
r
VO
in
^~
in
Cft
in
in
VO
in
CM
i->
0
VO
r»«
Ol
X
o
ia
o
at
IO
_ 10 in
at o rv
CL Ol
Q. w> i
< §^
c - c
r- W> IO
O £ <*
at E ^*
TJ LU w
u 10 s.
c co
r- O ««-
W 4->
-------
ee.
£
00
o
I
«c
00
0)
1
0
CO
CO VO
o r-
LU O>
Z
Aerospace
1975, projected
(Reference 2-17)
Is
3 T~
O
in
r
1
CM
gS|
0)
C£
CM
0)
1 CM U
|£|
C£
CO*
1
CM
0)
i 0)
t i 0
0) r»» c
Z2 *
o>
ce.
"oJ
u.
W r-^ CP
CM in o
^r to to
if O O
s s s
** r«. oo
co o o
co in r
to co i*«.
CM r^ oo
^r o o
to r- o
co en r-
*a- vo o
CO O
C) CO CO
CM O r
O -^
to
r*.
Ot
2
CL
o.
CM
O
cro
o>z
w v>
to n
Q.E
a> a>
S**
o «/>
§3
O
r- V)
C (O CO
O C t-
t- O
o
o
ot
0)
o.
o.
(O
-i-> in
re r**
o o«
v> *
u E
O i-
M (O
« -M
Z -
£15
4-> C
<§8
u
2-12
-------
2.3.2 Industrial Boilers
Equipment Description
This equipment category is comprised of industrial boilers ranging in capacity up to 73.2 MW
(250 x 106 Btu/hr). Industrial boilers are either field-erected or packaged units. The field-erected
units are only the very large capacity units and are-quite similar in design to utility boilers (See
Section 2.3.1). Packaged boilers, which are equipped and shipped complete with fuel burning equip-
ment, are mainly watertube and firetube designs. Other designs such as cast iron, and shell type are
also used. Each of these designs has a fairly distinct capacity range. Packaged boilers far out-
number field-erected units, but their combined fuel consumption is slightly less than that of field-
erected boilers.
In watertube boilers, hot gases pass over tubes which are water or steam filled. The tubes
line the combustion chamber walls and gain heat mainly by radiative heat transfer from the flame.
Downstream the combustion chamber heat is absorbed convectively with tubes mounted across the hot
gas flow. Almost all package boilers greater than about 8.8 MW (30 x 106 Btu/hr) are watertube boilers
Sales statistics for 1975 (Reference 2-19) indicate that 86 percent of the packaged water-
tubes were burner-fired (usually a single burner) and the remainder were stoker-fired. Of the burner-
fired units, 50 percent fired either residual or distillate fuel oil, 40 percent have dual fuel (oil
-natural gas) capability and 10 percent fire only natural gas. In general, natural gas and distil-
late oil firing units are more prevalent at the lower capacity ranges of watertube boilers.
In firetube boilers hot gases are directed from the combustion chamber through tubes which
are submerged in water. Firetube boilers burn fuel oil and natural gas because the design is parti-
cularly sensitive to fouling with ash-containing fuels. Residual oil and natural gas are the most
common fuels in the larger firetubes and natural gas and distillate oil are the main fuels for the
smaller units. Firing is by single burner. Recent sales statistics indicate that the firetube has
diminished in sales in the past 5 years (Reference 2-20).
Emission and Fuel Use
The 1974 NOX emissions from industrial boilers were estimated by essentially the same proce-
dure as used for utility boilers. The fuel consumption data for the total industrial boiler sector
were obtained from Reference 2-18. These 1971 data were updated to 1974 by an annual growth rate
estimate from Reference 2-21. The process gas consumption data were also obtained from Reference
2-18.
2-13
-------
The fuel usage for each specific equipment type was derived from the total fuel consumption
data based on the procedures used in Reference 2-4. The following were the basic assumptions made in
that report in formulating the emission estimates:
Field-erected watertube boilers larger than 29 MW (100 x 106 Btu/hr) are indistinguish-
able from utility boilers and have the same firing distribution
Field erected watertube boilers smaller than 29 MW (100 x 106 Btu/hr) are single-wall-
fired units
Packaged watertube boilers are single-wall-fired units
Packaged firetube boilers do not fire pulverized coal in significant quantities
Alternate fuel usage (e.g., coal-derived gas) is negligible
The emission factors were largely derived from a recent field test survey of industrial
boilers, Reference 2-9. The test data were screened for application to a nominal baseline operating
i»
condition. Where baseline data were available for more than one unit of a specific design type, the
data were averaged. When test data for a specific firing type/fuel wer . unavailable, emission fac-
tors were estimated based on data for similar units. The updated emission factors based on this
rdcent data are generally 15 to 30 percent lower for gas and oil-fired industrial boilers than those
based on earlier data (Reference 2-5). The resultant estimated 1974 NOX emissions from industrial
boilers are presented in Table 2-6.
Comparisons of fuel consumption data and NO emission estimates from recent emission inven-
tories for the industrial boiler sector are given in Tables 2-7 and 2-8, respectively. There are
substantial differences in the fuel consumption data used in the previous inventories which cause
wide discrepancies in NO emission estimates. These could be due to the difficulty encountered in
separating the total fuel usage in the industrial sector into industrial boilers, direct heat, feed
stock, internal combustion and other categories. The data used for the current estimates are
regarded as the most extensive and reliable to date.
2.3.3 Commercial and Residential Space Heating
Equipment Description
This category is made up of commercial and residential warm air furnaces and boilers. Warm
air furnaces are space heaters, where the unit is located in the room which it heats, or central
heaters which use ducts to transport and discharge warm air into the heated space. Space heaters
comprise less than 10 percent of the nation's heaters. Central heaters make up the remainder of the
2-14
-------
CO
Q
a:
O
O
Z
CO
O£.
O
CO
Z
o
LO
I
CM
UJ
CO
-d
CO
c
O
CO
CO
'i CM
UJ O
z
X
O CD
Z C3
"re
o
C-Q «-3
O CO ^
r L- CM
CO O O
CO 4-> Z
'£ U
fc c
re c
x
«3-
"3-
cn
CM
o
CO
«3-
r~
*O
&_
3
re
z
c/
fC
cr
>
>
K
n
M
A
in
cn
CO
cn
T*
CO
LO
CO
CO
CU
o
o
CL.
1
t
J
in
LO
CM
CM
CO
r>
«*
i
E
O
P
O
CO
L.
a
-
o
o
>
I-
r
If
4-
C
C
N
p-
S.
C
X
LO
LO
LO
in
cn
CO
1
o
[-*
o
CO
4->
cu
3:
re
0
>
i
«:
3
> l
X
a
OL
o
a
c
r-
L L.
L-r-
U.
CO
LO
*f
cn
CM
LO
00
CO
*s-
!«_
03
i_
3
re
z
>a-
IO
\
(V
5
cn
i
E
O
4->
M
O
CO
i-
o
(_-
D
?
^
J
J
D
^
LO
CO
on
co
CM
LO
LO
1
E
0
P
P
O
CO
1)
2
p_-
D
a
c
c
r
C
CJ
cn
*3-
CM
LO
^?
r_
IT
T
CO
CU
a:
^
>
**"
LO
st-
LO
CO
00
co
cu
re
^^
^«
^ ^
CO
a
1^
c
a
cu
4->
o
2!
UJ
o
01
^
LO
in
CO
«
a
,_
re
3
o
CO
cu
a:
>
c
c
r-
S.
u.
r~*
|HIH
1C
3
cn
L.
O)
O
CQ
ai
JT
J-
0)
4->
re
r
s
cn
CO
cn
10
S
CO
re
z
(/
n
cr
D
2
cn
CM
i
cn
CM
~
c
cn
00
cn
00
CO
CO
CO
cu
o
O
Q.
)
in
pM.
CO
cn
00
r£
CO
1
re
o
L,
U
o
u
1
CO
I
4J
t^
a
LU
-i
r
01
1
u.
co
CO
co
cn
CM
CM
cn
LO
""
1
re
0
o
«
V
V
^
1)
;
13
I/
a
I-
i-
a
js
3
4^
01
4-
3
cn
LO
CO
CO
LO
CM
VO
CM
1
re
o
o
o
cu
u
jj
>
o
I
cn
J-
01
^
c
->
ai
LO
CO
o
1
re
o
0
"S
£
CO
CO
i re
1° ""
V
D o
cu z
o
cu
TJ
C CU CO
cu i- re
CL ro
a. -a
«C co CU
C 4J
C O -M
a co cu
cu »-
a E co
3 CU f-
u xc
c o o
c o
CU CJ
c cu
o u
U t-
3
>> o
co co
_a
2-15
-------
z: cu
«
fj
o
h~
CX» -}
o to
f~ f CM
to O O
to 4-> z:
f- O
E re cn
tu ia- n
r- tU
3 re '-3
u_ 10 o.
^
Q.
^>»
<-~
m
o
CM
CM
CO
co
r--.
sfr
F ii
C.' » I
3 J
u. 5
4 H
5
|;
IS
i
j; ! ^
u_
cu
D.
>^
(
O'
e
r*
k.
=^-
t i
O)
Q.
>sj
1
4-)
C
CU
n
r~
3
cr
LU
o
o
!«.
r
in
vo
00
00
«=}
in
r
01
ro
r~
r^
r-
4^
I/I
ii
j.
co
in
f
in
co
f-~
O
00
,_
CO
CO
r-"
ro
3
-a
i/>
a;
D;
.. _j
f--
,r
O
in
vo
CM
cn
00
cn
CM
CM
in
f-.^
VD
CM
r -
ro
i.
3
4->
re
z.
sj"
r~>
cr>
00
cn
r^
cn
co
r
I/)
01
QJ
O
O
t-
0-
i.r>
1W
- 1 ._
cn
c.
r
i-
r
LU
r
r
re
3
Cu'
XI
3
-M
j..
CU
4-3
ro
3
o
CU
cn
re
^^
o
re
O-
cn
co
cn
r^
r
cn
ro
CO
CO
1
ro
O
O
t_
a>
^5
ts
a
c
3
in
CM
00
00
CM
C\J
**
m
cn
i
re
o
o
-o
cu
QJ
<4_
^_
cu
>
o
r-*
r-.
cn
^
r-
CM
VO
CM
vo
i
re
0
-a
cu
r
4-
f-»
i/)
" 10
i re
re
i- O
cu
g 4J
cu o
C5 Z
CJ
jQ
3
4~>
t-
OJ
4_J
T3
3*
"O
OJ
cn i_
ro a>
IS ^
0 O
ro 4->
Cu 00
!««.
r
Lf>
VO
CO
00
l«T
t.O
r
a.)
ro
i^~
r~"*
r"-
4-^
l/l
*r~
a
CO
f "
LO
s
fHH
CO
in
f-^
r^»
,_
re
3
-o
t/)
CU
a:
P
0
CO
CO
cn
00
cr.
r--
VO
"^J"
CO
^-»
ro
t.
3
4J
re
^
CM
cn
CO
cn
0
0
CM
to
t/)
CU
o
0
a.
re
0
cn
c
r*
t-
.p.
LL.
r
r
re
3
cu
y*t
3
4J
CU
SH
r
U.
o
cu
re o
^^ ^>
0 O
re o
a. oo
^S"
in
r~
vo
in
in
cn
in
a>
ro
i
^«
»^-
4>>
{/}
r-
Q
vo
in
in
00
r~-
cn
CM
vo
o
CO
r
re
3
"^
r
to
CU
a:
r
o
CO
^
cn
CO
cn
vo
CO
,
CO
r*^
r
re
i.
3
4_)
re
z
CM
cr>
co
a>
o
o
CM
to
to
cu
0
o
i_
a.
to
re
O
cn
c
^
S-
p
LL.
f
i-~
ro
3
CU
XI
3
4_>
cu
L
r*
U_
-0
CU X
cn o
re xi
_* a>
0 J-
re T-
D- Lu
<^
CO
cn
r~»
r~
CTi
CO
r~.
P.-
i
re
o
o
i-
cu
TD
re
cu
a
oo
vo
co
cn
r--.
r»
VO
«*
cn
CD
|
re
0
0
-a
0)
cu
*4
Sw
CU
T3
C
=
*^
00
00
VO
CM
CO
CT)
f
r«_
1
re
o
o
T3
CU
CU
4-
S-
cu
0
cu
XI
3
4-> i.
CU QJ
i- -*
r- O
Lu -4->
oo
o
CU X
cn o
re xi
\X QJ
U S-
re >-
CX Lu
CM
in
L£>
CO
cn
CM
ai
ro
r.
r""*
r
4^
to
r~
Ci
0
CO
in
cS
r
cn
r
VO
i
re
3
T3
r~
l/l
d)
*!«
O
00
CO
cn
00
Ch
cn
CO
«d-
CX3
CO
1
re
cs
cn
c
[I
r~
LL.
,__
r
re
3
CU
XI
3
CU
s-
f
^^
T^J
cu
cn
re
O I
re a:
D. a:
CM
CO
cn
r»»
P
in
cn
00
i
re
o
o
^
cu
o
re
cu
s_
a
oo
00
CO
cn
i-^
r"»
CO
r*»
^a-
i
re
o
CJ
T3
CU
CU
L.
CU
C
=
CM
CO
CO
VD
CM
VD
cn
in
i
ro
o
^^
cu
cu
4-
CU
0
cu
XI
3
CU
s_
p*
Lu t-
CU
T3 ^i
0) O
CD4J
re oo
U 1
re ce.
a. 2:
C
Lul
C
«t
X
TJ
CU
o
cu
XJ
to o
C Z
o
E cu
O) +->
o "E
c to
O 3
(J XI
>> o
CO U
2-16
-------
CO
QC
§
oi
O
CO
o
Source
co to
UJ en
cu r-.
U 1
V CU CM
U'O
<0 0 CO
Q. i- U
tO CL C
2 .C
o> in
f C7>
"o> c
C O
C IO
OJ O
Q.
o. to
*
"8 1
TD UJ
U
Q)
Q.
.a
re
4J
cu
u
o
Q.
O
X
2-17
-------
.a
re
01
cs
CO
QL
LU
O
CO
1
1^1
fyt
\-
co
1
~
o
CC
u_
CO
z
o
1 «
CO
CO
s:
LU
x
o
1
12
z
z
00
1
CM
LU
CD
1
01
o
o
CO
U
CO VO
LU O>
Z i
TO "»
O) U 1
O
0. O QJ
l/> &» O
O Q. C
t- OJ
a; - t.
^c in QJ
Of QJ
r a:
4~>
£*
3^
O
^T-
^*
1
CM
Q)
4-
QJ
a:
^*
i
CM
QJ
1 CM U
co r>. c
/) CTt ai
*2-
AJ
a:
vo
i
CM
W
QJ
i t O
W t^ C
-> CTl Q>
o a>
CO («_
oo
in co co
r1**. ^0 o^
vo o> ^j-
o in a>
CO VO «*
,_
re i w
o re
o o cs
00
00
CM
CM
00
CM
CM
VO
CO
CM
*~
2
re
4^
o
a>
in
0 &
4J '~
5 1
H *
S-
QJ ^j-
Q. r^
un ,
"aj «t
^_>
re. re
. +j
X TO
o
t- i/)
o. c
Q. -,-
^ re
c
o
QCM o
Z
vo
1 -
ro . .
> to vo
r~ 4-^ r~*
CTO C »->
3
l/> 10 jc O)
10 *° 4^ rf
T3 T3 i
OJ QJ CD C
+-><-> C 0
t, 4J LU
O 'I- £
Q. E C QJ
QJ QJ .r- 4_>
1- to
QJ-£ "* oo4
I- X
re c -f- 10
O T3 4J W
«/> T- c re re
C 4-> QJ Q O>
O (/> CL
'-3 Q. to jrt
<^-G «t c w
co E O a;
E o !- Lo o
W co t.
> !- o xj
c re co -M en
O C -i- (O O
r o sz tZ.
"c !J3 *" °°
W re .0 o (/>
> -M re QJ
C «/> +J 4-> T3
S E 33
O -r- 4J (j
re JQ o TO"*
2-18
-------
warm air heater equipment sector. Combustion chambers are cylindrical for distillate oil firing or
sectional for natural gas firing. Combustion products pass through flue gas passages of the heat
exchanger and exit through a flue to the atmosphere. The commercial packaged boilers are very simi-
lar to industrial packaged boilers (See Section 2.3.2). Boilers used for residential space heating
are generally cast iron designs. Residential warm air furnaces and cast iron boilers are available
in sizes up to 0.12 MW (4 x 10s Btu/hr). Larger units are mainly confined to the commercial and
institutional sector.
Commercial and institutional systems are used for space heating and hot water generation.
The equipment consists mainly of oil-fired warm air furnaces and firetube boilers. The rated heat
input, or fuel consumption, of this equipment ranges from 0.12 MW (4 x 10s Btu/hr) to 3.6 MW (12.5 x
106 Btu/hr).
Fuels burned for residential and commercial space heating are residual and distillate oil,
natural gas and occasionally coal. Typically residual oil firing is limited to the larger commer-
cial or institutional boilers.
Although there has been a continuing trend in the recent past toward space heating equipment
which uses natural gas, this trend is expected to reverse itself in the near future (Reference 2-22).
'Furthermore, the use of fossil fuels of all types is expected to drop drastically by the year 2000.
Emissions and Fuel Use
The fuel consumption data for commercial and residential space heating equipment were
obtained from the National Gas Survey (Reference 2-23). The fuel consumption data were subdivided
by equipment type based on the procedures of Reference 2-4. The basic assumptions made in that
study were:
Coal is burned in commercial units (stokers) but not in residential systems
t No pulverized coal is burned in commercial or residential units
Residual oil consumption is negligible in residential units
t Commercial fuel usage is directly proportional to installed capacity
LPG, wood, or producer gas have negligible use in space heating
The emission factors used were obtained from AP-42 (Reference 2-5), AP-42 Supplement No. 6 (Refer-
ence 2-8), and Reference 2-24.
The NO emissions estimates by equipment type for the commercial and residential sector are
presented in Tables 2-9 and 2-10, respectively. A summary of these results by sector is given is
Table 2-11.
2-19
-------
10
*»
r^
Cft
LU
8
to
to
a:
o
to
o
c
o
r-
1 CM
UJ O
z
X
Z CO
lo
»*
o
C.Q --3
o to ^
i- I. CM
CO O O
to 4-> zr
i- O
E 10 en
UJU. C
r- CO
£3".
Oi
Q.
*o>
"oJ
U-
Ol
Q.
&
o»
c
"^
-
Ol
ex
c
.5.
3
CT
UJ
00
VO
CO
in
vo
o
in
«*
in
Ol
IO
to
0
co c
8S
O) C
«S- C
SS
n f
** o
in «
3
;/» i
OJ
QC
r V
O C£
en
c
r-
U_
*""
fm
C
in c
en u
Su
in i-
CM C.
r^ <*
CM «!
IO
0
LO 1
Ol
on
o cr
en
c
r-
ii
^^
IO
2
O
i <
to
IO
0
Packaged
Boilers
Q
C\
in
jQ
in
in
en
CM
01
IO
D
^ r
vo r
V7> f
*± ^
VO U
CO *1
IO
3
D
o;
i V
i- 1C
o e
en
c
r-
i.
r
1 1
p«
IO
2
Ol
3
Ol
5
en
t
^ <
cnc
S^
O r-
«a- G
CM r-
3
o
^
U
cc
r V
51- «
0 Ci
en
c
Z
1 1
pM
IO
2
01
jQ
3
Ol
4J
2
Packaged
Firebox
^^
CM
in
0
in
in
M
CO
o
!/)
a
r^ ^"
vo c\
en co
^± «;
^^
r^ en
vo vo
co «a-
IO
D
LA 1
CO
OL
r CO
5 3
en
c
TI
i i
p.*
IO
o
Ol
w««-
01 to
J3 to
3 10
s. G
Ol
4-> 4->
«o
Packaged
General ,
Ol
a.
in
en
x
o
Q.
Q.
CM
O
c cro
3 01 z
C.
«f.
S- 4->
O i-
Q. E
Ol O)
IO C
o
to i-
-o
CO
o. o to
Q. -r- 3
T3 CO
CO XU
O O J-
3 Z 3
O r CO
C r-
i- IO >>
co c: 10
r- o C
CO 4-> -r-
I C 4J
JD CO IO
IO > +J
4-> C CO
O
2-20
-------
en
T-
ft
O
i
LU
CO
<
>i
H-
LU
a
co
LU
a:
o:
o
CO
=3
u.
a
z
CO
oc
o
o
JQ
VI
C
O
1-
l/l
'E CM
UJ O
X
o o>
y tD
IO
4->
O
CjQ O
O V) -x.
r- t. CM
vi O Cj
I/I 4-> Z
r- O
E 10 0>
UJ U- C
r §7,
3 "* 2
Lu ^
^3
CO
Q.
r
CO
U.
r
CO
3
U.
0)
Q.
t
01
c
t
$_
u.
0>
O.
f*
+1
C
is
ex
^
3
O-
LU
i
O
o
in
in
o
en
CO
f*n.
CM
r~
CO
1C
r
M
to
O
r
o
s.
0)
£
3
CO
CO
*OI
c
CO
i-
co
IO
3
4^
o
a:
t.
0
E <
IO H
0)
+J I
CO
CM
fN,
CO
VO
CM
in
CO
CM
in
in
o
r~
i
in
(Q
0
ft
.
u
J
o
u
c
CO
^*
>,
cS
in
in
en
in
o
*t
CO
*->
(0
r*
_l/l
O
>^B
o
i
<
!
c
<
1
!
c
1
<
1
1
L
J
i
«
4
j
en
oo
o
r**
VO
CM
in
CO
«a-
en
o
o
CO
i
i/i
.
i
to
ia
0
L.
I)
f
3
O
V
7>
O
S2
a
10
a
(/I
0)
^>
IO
£
i/i
o
p*
o
i
1
c
1
0
.c
r-
"^
fc.
CO
+^
(Q
0)
o
a:
o
vo
CM
M
in
CO
VO
in
00
CO
r«^
i
1/1
IO
CD
L
L*
f
3
Q
1)
7)
=
O
OJ
P*"*
U.
f>^
VO
CM
in
CO
oo
CM
,_!
o
CM
0)
fj
ITJ
^
.*"
a
r
r~
0
CO
^
3
CO
CO
'en
c
^
Q
f
r-
Tg
t.
CO
fO
^
^^*
§ 3
O r-
ecu.
ID
*»
r-.
CD
i
S
o
2
CO
«c
tI
o
o
o
O
O
O
£
a
«
|
3
i
O
O
C
0
1
E uj
^i
1/1
c
o
o
r""*
CO
3
Lu
J3
|
3
"o
O CM
« O
C
O C7>
I/I
VI
r-
g
LU
X
O
z
fH»
IO
*>
o
ia
o
o
V)
S
r_
o
r»
IO
>
O
o
o
I/I
IO
u
r
O
u
o
u
CO
r en
-^ 00
^- in
*-,_ - s^^
en co
00 O
CM O
vo en
"o
o
**
CM
00
CM
O
en r-
en o
CM r^
'**** sw
^ ^
oo en
co in
CM in
in in
i-^ 00
in *»
o vo
CM O
vo «*
CO CO
in «»
VO CO
* »
en in
en co
O 00
1^ CO
8
**
CM
en
^^
CO* ^
o en
in ««
^ ' *^_
CO CM
co r-.
o en
CM i
^j- 10
i-^ oo
hv CM
^- co
^^ CO
in oo
r
t- "£
U C
t- 0)
g **~
o co
o ec
o"
o
^v
--^ 1
CM
en
CM
in
r
"o"
0
*~*"
CM
00
CM
0
_
o
o
*"**
^
CO
en
^
'o'
o
10
CM
0
^
^J-
o
o
o"
o
*
CM
en
^^
o^
o
-T_J
in
O
**
8
!>.
O
*
O f
O.E
Q) 0)
to
0) C -M
a. o i/i
ex t- 3
o co
a> xo
o o t-
3 Z 3
^- O
O r VI
C !->
c to
8§
>ȣ
2-21
-------
Thirty percent of the total fossil fuel used in the United States in stationary sources is
consumed in space heating. According to the 1970 United States Census, 57.7 percent of residential
heating equipment was gas-fired, 28.3 percent was oil-fired, and the remaining 14 percent used other
fuels such as propane, coal, and wood. Fuel consumption data for the combined commercial/residen-
tial sector obtained from several sources are compared in Table 2-12. The fuel usage used in this
report compares very well with those from the FEA for 1974. The NOX emission estimates by fuel are
compared with several recent sources for the commercial and residential sector in Tables 2-13 and
2-14. The estimates are in reasonable agreement.
2.3.4 Internal Combustion
2.3.4.1 Stationary Reciprocating Internal Combustion Engines
Reciprocating 1C engines for stationary applications range in capacity from 15 kW (20 hp) to
37 MW (50,000 hp). These engines are either compression ignition (CI) units fueled by diesel oil or
a combination of natural gas and diesel oil (dual), or spark ignition (SI) fueled by natural gas or
gasoline.
Stationary reciprocating 1C engines use two methods to ignite the fuel-air mixture in the
combustion chamber. In CI engines, air is first compression heated in the cylinder, and then diesel
fuel is injected into the hot air where ignition is spontaneous. In SI engines, combustion is spark
initiated with the natural gas or gasoline being introduced either by injection or premixed with
the combustion air in a carburetted system. EUher 2- or 4-stroke power cycle designs with various
combinations of fuel charging, air charging, and chamber design are available.
Because reciprocating 1C engine installations characteristically have a low physical profile
(low buildings, short stacks, and little visible emissions), they are frequently located in or adja-
cent to urban centers where power demands are greatest and pollution problems most acute. These
units are used in a variety of applications because of their relatively short construction and instal-
lation time and the fact that they can be operated remotely. Applications range from shaft power for
large electrical generators to small air compressors and welders.
By capacity, 73 percent of the 1C engines are fueled by natural gas, 16 percent by diesel oil
and 11 percent by gasoline. In terms of installed capacity, the oil and gas industry is the leading
user of stationary 1C engines for pipeline and production applications, followed by general industri-
al users, electric power generation, and agriculture. In terms of annual energy consumption, oil and
gas industry applications again come first, followed by general industrial and electrical generation
applications.
2-22
-------
oc
o
o
UJ
CO
o
1-4
to
0
QC
cc
o
z
<
CM
CM
0)
co*
CM
«*-
ce.
00
CM
^^l*-'
0)
ce.
|g
o
tn"
gg~
e
co*
CM
2^^
0)
1 CM CM
O)
DC
ik. r*»*
0 ^ ~
|£" «
in CM
J (O CM
1 Ot
Q-
Of.
in
CM
MM. 1
^^ J2
O£
"cw
u.
CTl
r
CM
O
CO
CO
o
CM
00
CM
O
r-.
to
CO
o
CO
o
in
«*
CM
O
o
CO
0
00
CM
O
§
in
o
19
O
0
s
vo
«0
s
o
~
CM
O
r"
in
*
vo
it
$
CM
VO
0»
(O
CM
§
CO
*
o
O)
o
oo
s
"
1
^
to
r«.
cy>
CM
§
00
r*.
^
CO
,
vo
IQ
Ul
c
Q.
C
^
"S
o
jQ
5
2-23
-------
a\ b
TABLE 2-13. ANNUAL NOY EMISSIONS FROM COMMERCIAL BOILERS (Gga)
Fuel
Coal
Oil
Gas
Total
Source
Battell e
1971
(Reference 2-18)
119
134
23
276
MSST
1972
(Reference 2-4)
26
192
109
327
GCA
1973
(Reference 2-15)
27
570
100
697
Current
1974
49
457
203
709
aBy convention all NOX emissions are reported as equivalent NO;?. Approximately
95 percent of the NOX from stationary source combustion is emitted as NO.
n"his table is included in Appendix A in English units.
a,b
TABLE 2-14. ANNUAL NOX EMISSIONS FROM RESIDENTIAL SPACE HEATING (Rga)
Fuel
Coal
Oil
Gas
Total
Source
Battell e
1971
(Reference 2-18)
-
153
162
315
MSST
1972
(Reference 2-4)
-
230
192
422
GCA
1973
(Reference 2-15)
11
89
190
290
Current
1974
-
183
197
380
aBy convention all NOX emissions are reported as equivalent NOg. Approximately
95 percent of the NO from stationary source combustion is emitted as NO.
This table is included in Appendix A in English units.
2-24
-------
The emissions estimates for reciprocating 1C engines, given in Table 2-15, were derived from
Reference 2-11. This is the most recent and complete survey of emissions from reciprocating 1C
engines. Since the horsepower, speed, cycle, fuel, air charging, and fuel charging combinations are
so numerous, emission estimates for each combination would be impossible considering the data avail-
able. In view of this fact, the 1C engine emissions are categorized into spark-ignition (gas-fired),
and compression ignition (diesel or dual-fueled). The fuel data of Reference 2-11 were updated to
1974 by the FPC data from Reference 2-16.
2.3.4.2 Gas Turbines
Gas turbines are rotary internal combustion engines fueled by natural gas, diesel or distil-
late fuel oils, and occasionally residual or crude oils. These units range in capacity from 30 kW
(40 hp) to over 74 MW (100,000 hp) and may be installed in groups for larger power output. The basic
gas turbine consists of a compressor, combustion chambers, and a turbine. The compressor delivers
pressurized combustion air to the combustors at compression ratios of up to 20 to 1. Injectors
introduce fuel into the combustors and the mixture is burned with exit temperatures up to 1,370K
(2.000F). The hot combustion gases are rapidly quenched by secondary dilution air and then expanded
through the turbine which drives the compressor and provides shaft power. In some applications,
exhaust gases are also expanded through a power turbine.
While simple-cycle gas turbines have only the three components described above, regenerative-
cycle gas turbines also use hot exhaust gases (700K to 870K, 800F to 1.100F) to preheat the inlet air
between the compressor and the combustor. This makes it possible to recover some of the thermal
energy in the exhaust gases and to increase thermal efficiency. A third type of turbine is the
combined-cycle gas turbine. This is basically a simple-cycle unit which exhausts to a waste heat
boiler to recover thermal energy from the exhaust gases. In some cases, this waste heat boiler is
also designed to burn additional fuels to supplement steam production, a process which is referred
to as supplementary firing.
Gas turbines have been extremely popular in the past decade because of the relatively short
construction lead times, low cost, ease and speed of installation, and low physical profile (low
buildings, short stacks, little visible emissions, quiet operation). In addition, features like
remote operation, low maintenance, high power-to-weight ratio, and short startup time have added to
their popularity. Primary applications of gas turbines include electricity generation (peaking and
baseload), pumping, gas compression, standby electricity generation, and miscellaneous industrial
uses.
The fuel data for gas turbines were obtained from the Shell Report (Reference 2-29) and updated
to 1974 by data from the same FPC source mentioned above. Emission factors were obtained from AP-42
Supplement No. 4 (Reference 2-6). NOX emissions for gas turbines are shown in Table 2-15.
2-25
-------
JO
c
o
i
oco
s
2
JO
vt
J.
0
u o
«o *>x
U. CM
ffi
O
In c
*i
UJ
O*
0)
10
V)
*!w
u.
*o>
3
u.
Q.
>,
t
o>
c
T
iZ
a
>>
t
c
u
s.
"5
cr
UJ
§
o
CM
oo
J2
CO
I
V)
c
^
M
0>
Q. C
L> O>
CD C
oe uj
0
CM
00
oo
00
CM
vo
*p
o
.*:
CO
CO
A
p_
O)
c
p
3
""
to
o
o
r
in
in
vo
CO
m
CO
o
CO
r~
o
91
JC
01
o
0)
Q.
X
o
Q.
Q.
CM
O
<:
x
O)
CL
a.
a:
T3
o>
TJ
a>
»-> u>
s.
s. *->
s- E
>o a>
o >r"
E co
a> 2
i a»
r O
a J-
c o
O co
a> 10
> c
c o
o f-
2-26
-------
Comparisons of fuel data and NOX emission estimates from several recent surveys are given in
Tables 2-16 and 2-17, respectively. The lower emissions from gas turbines from this report, compared
to those from the earlier inventory, Reference 2-4, are mainly due to the smaller and more recent
emission factors obtained from AP-42 Supplement No. 4 mentioned previously.
2.3.5 Industrial Process Heating
Significant quantities of fuel are consumed by industrial process heating equipment in a
wide variety of industries, including iron and steel production, glass manufacture, petroleum refin-
ing, cement manufacture, sulfuric acid manufacture, and brick and ceramics manufacture. In addition,
there are dozens of industrial processes that burn smaller amounts of fuel, such as coffee roasting,
drum cleaning, paint curing ovens, and smelting of metal ores, to name only a few. Brief process
descriptions for some of these are given below.
Iron and Steel Industry
The iron and steel industry is one of the major contributors to combustion-related process
NO emissions. The most important combustion processes are sinter lines, coke ovens, open hearth
furnaces, soaking pits and reheat furnaces. The remaining combustion-related processes (pelletiz-
ing, heat treating, and finishing) are less important because they use relatively small amounts of
fuel (Reference 2-9).
Sintering machines are used to agglomerate ore fines, flue dust, and coke breeze for charg-
ing of a blast furnace. The use of this operation is presently declining at the rate of about 3.4
percent annually because of its inability to accommodate rolling mill scale which is contaminated
with rolling oil.
Coke ovens produce metallurgical coke from coal by the distillation of volatile matter pro-
ducing coke oven gas. The fuels commonly used in this process are coke oven gas and blast furnace
gas. Although NO emissions are minimized by slow mixing in combustion chambers, they are nonethe-
less substantial because of the very large quantity of fuel consumed in this process. Present pro-
jections show a 5.7 percent annual increase in fuel consumption for coke ovens.
Open hearth furnaces are now being replaced in the U.S. steel industry by the basic oxygen
furnace, but are still an important source of NOX emissions because of the very high combustion air
preheat temperature, high operating temperatures, and the practice of oxygen lancing. Fuel consump-
tion in open hearth furnaces is presently decreasing about 8 percent per year.
Soaking pits and reheat furnaces are used to heat steel billets and ingots to correct work-
ing temperatures prior to forming. Current trends are toward continuous casting of molten metal,
2-27
-------
TABLE 2-16. ANNUAL FUEL CONSUMPTION BY INTERNAL
COMBUSTION ENGINES (PJ)a
Fuel
Oil and Dual
Gas
Source
Shell
1971
(Reference 2-29)
503
1579
MSST
1972
(Reference 2-4)
548
1716
Current
1974
569
1706
This table is included in Appendix A in English units.
TABLE 2-17. ANNUAL NOX EMISSIONS ROM INTERNAL COMBUSTION ENGINES (Gg3)1
Equipment
Reciprocating
Engines
Turbines
Fuel
Oil and
Dual
Gas
Oil
Gas
Total
Source
Shell
1971
(Reference 2-29)
360
1580
30
70
2040
MSST
1972
(Reference 2-4)
287
1697
108
156
2248
Current
1974
399
2014
109
127
2649
By convention all NOX emissions are reported as equivalent N02. Approximately
95 percent of the NOX from stationary source combustion is emitted as NO.
This table is included in Appendix A in English units.
2-28
-------
and the need for these units 1s being eliminated. At present, however, soaking pits and reheating
furnaces still consume more fuel than any other single process 1n the steel industry. In spite cf
the fact that soaking pits and reheat furnaces are being phased out, consumption of process fuel
continues to increase at an annual rate of about 2.8 percent in the Iron and steel Industry as a
whole.
Glass Industry
In the glass industry, melting furnaces and annealing lehrs are the two fuel combustion pro-
cesses of greatest importance. Melters in the glass industry are continuous reverbatory furnaces
fueled by natural gas and oil. Coal is not suitable for these furnaces because of its inherent
impurities. Annealing lehrs control the cooling of the formed glass to prevent stains from occur-
ring. Some lehrs are direct-fired by atmospheric, premix, or excess-air burners. About 80 percent
of the total industry fuel consumption goes for melting, while annealing lehrs consume about 15
percent. There is a current trend in the glass industry towards electric melters, or at least elec-
trically assisted conventional melters. But until it becomes clearer which fuels are going to be
available in the future, no definite trends will emerge. Present trends toward fuel oil in place
of natural gas have begun as a result of natural gas shortages and price increases.
Cement Industry
Cement kilns are the major combustion processes in the cement industry. These kilns are
rotary cylindrical devices up to 230 m (750 feet) in length which contain a feedstock combination
of calcium, silicon, aluminum, iron, and various other trace metals. This mixture of elements in
the form of various combinations of clay, shale, slate, blast furnace slag, iron ore, silica sand,
limestone, and chalk slowly moves through the kiln as products of fossil fuel combustion move in an
opposite direction. Temperatures of the material during the process may reach 1.756K (2.700F).
Coal, fuel oil, and natural gas are the main fuels used in cement kilns. Natural gas
accounts for 45 percent of the fuel consumed, coal for 40 percent, and fuel oil for 15 percent. The
major effluent stream for this process is the exhaust gas which passes through the entire length of
the kiln and may entrain additional particulate or trace metals from the kiln feedstock. Cement
industry figures show that the industry has grown an average of about 1.9 percent annually over the
past 20 years. Industry projections, however, predict a greater growth in the next few years of
between 2.6 to 4.1 percent per year (Reference 2-30).
2-29
-------
Petroleum Refineries
A wide variety of process combustion takes place in the petroleum refining industry, includ-
ing catalyst regenerating in the catalytic cracker, catalytic reforming, delayed coking, and hydro-
treating and flaring of waste gases. Catalytic cracking is required for a large portion of gasoline
production. Fuel is consumed in this operation in the catalyst regeneration procedure which removes
coke and tars from the catalyst surface. Temperatures during this process are moderate, ranging
from 840 to 922K (1,050 to 1.200F), but fuel requirements are on the order of 829kJ/l (125,000 Btu/
Bbl) feedstock. Catalytic cracking capacity increased about 1.7 percent per year between 1960 and
1973. Future growth will depend on energy and environmental policy and particularly the demand for
low sulfur fuel oil. Present estimates of future growth are from 1 percent to 3.0 percent per year
(Reference 2-30).
Catalytic reforming, where certain saturated ring hydrocarbons are converted into aromatic
compounds, typically utilizes oil, gas, or electricity as its primary fuel. Delayed coking is an
energy extensive process which uses severe cracking to convert residual pitch and tar to gas, naptha,
heating oil and other more valuable products. Hydrotreating is a process designed to remove impur-
ities such as sulfur, nitrogen, and metals to prepare cracking or reformer feedstock.
Process heating fuels used by the refinery industry are primarily natural gas and refining
gas, along with some residual oils and petroleum coke. Projections are for a 2.7 percent annual
increase in process heating to 1980, and 2.9 percent per year to 1985 (Reference 2-30). The fuel
mix for the future is highly dependent on both availability and costs of the preferred fuels, and
is therefore very difficult to project until national energy priorities are established and the ques-
tion of natural gas price regulations is settled.
Brick and Ceramic Kilns
Brick and ceramic kilns for curing clay products are another major user of process heating
fuels. Products of these kilns include structural bricks, structural and facing tile, vitrified
clay pipe, and other related items. Typically a kiln is operated in conjunction with a drier which
recovers part of the heat contained in the exhaust gases. Kilns are fueled by coal, oil, or gas
(depending on the availability of fuel and the product being cured) for batch runs of 50 to 100
hours at temperatures around 1,367 K (2,000 F). Combustion products are ducted from the kiln to a
drier, where wet clay products undergo an initial drying process. Occasionally, when higher
temperatures are needed for drying, a secondary combustion process is used in the drier itself.
2-30
-------
Emissions
Emissions from the industrial process equipment sector are regarded as the most difficult to
quantify of all stationary sources. This is largely due to the extreme diversity of equipment types
in use and is compounded by the common practice of reporting industrial fuel use by sector rather
than by equipment type. The annual nationwide NOX emissions estimates for the significant emitters
in the industrial process heating sector are given in Table 2-18. A number of minor equipment types
are excluded from this table, as insignificant on a national scale, but could be important from the
standpoint of localized pollution potential. The data sources used to generate Table 2-18 include
References 2-4 and 2-31 through 2-35. The nationwide emissions for the industrial process sector are
estimated to be 432.2 Gg (0.476 x 106 tons) per year which comprises 3.5 percent of the nationwide
total from all stationary sources. Table 2-18 also shows estimates for some process equipment types
from a recent output from the National Emissions Data System (NEDS) and from a recent IGT study
(Reference 2-36). Major discrepancies exist in the estimate for glass melting furnaces and for
heating/annealing ovens. Further study should therefore be made before these data are used to
evaluate the need for control measures for these sources.
2.3.6 Incineration
NOX emissions estimates due to incineration are taken from the OAQPS survey, Reference 2-32.
NOX emissions from open or prescribed burning are not included in this category. OAQPS reported a
total NOX production due to incineration of 37 Gg (41,000 tons). The 1971 estimate was updated to
1974 using population growth and GNP data obtained from the Bureau of Census, Reference 2-35, and
is presented in Table 2-19. The total NOX emissions from incineration are thus estimated to be
39 Gg (43,000 tons) per year and may be compared to the AP-115 (Reference 2-26) estimate for 1969
of 64 Gg (70,000 tons) NOX and the 1976 NEDS National Emissions Summary value of 42 Gg (46,000 tons)
NOX- In view of the broad spectrum covered by the industrial incineration sector, these discrep-
ancies are not surprising.
2.3.7 Noncombustion Sources
NOX emissions for the chemical industry dominate this category. Again, OAQPS, Reference
2-32, data are used exclusively for nitric acid production, sulfuric acid production, and explosives
manufacture. The emissions for these sources were updated to 1974 by production data obtained from
the Bureau of Census, Reference 2-35, and are presented in Table 2-20. Personal communication with
GCA Corporation yielded emission estimates for adipic acid plants. The total NO emission from
these noncombustion sources amounts to 203 Gg (0.224 million tons) per year. Nitric acid production
2-31
-------
JQ
en
X
o i-
"
C «/>
9) 9)
i~ 4->
1- <2
LU
Lu
C
O
^J
ID
U
lo.
QL
£
U)
3
C
t <
O
i i i o.
en
CO LO 00
i
co co
CO i
r
r*^ to *^ o> CA
vo o LO to co
LO ^- ^- ^ 9>
f^
«
r V> CD r
f- «A
0 «0 C
Ol
O -C f-
CA 4) ^^ i/k
C 0 >» «/> 0>
r- 10 r- 0) C
M C «0 0 t-
i t 4J 0 >»
0> 3 0 C
i/i L o>
<0 4J E
» 0» 0»
CD IX (_}
Ol
^O
CVJ
(ft
CO
<
in
t/>
CD CD M
->. 1 1 'v. «O
f CD
0 O
c
0)
o
O) l/t
c c
w- 9)
(O O
0)
C J=
C 4-1
C 4-> i- t.
«t
2- x
«0 C f-
O t3
V»T- C
c 4J a>
o to o.
r- 3 Q.
«/>.Q <
0)
xo
O i-
"8
2-32
-------
TABLE 2-19. SUMMARY OF ANNUAL NOX EMISSIONS FROM INCINERATION1
Industry
Incineration
Total
Application
Industrial
Municipal
Total NO Emissions
Gg
21.8
17.2
39.0
TABLE 2-20. SUMMARY OF ANNUAL EMISSIONS FROM
NONCOMBUSTION SOURCES3
Industry
Acid
Explosives
Total
Application
Sulfuric
Nitric
Adi pic
NOX, Ggb
10.9
127.0
14.5
50.9
203.3
TABLE 2-21. ESTIMATES OF ANNUAL NOX EMISSIONS FROM OTHER SOURCES
Source
Solid waste disposal
Forest wildfires
Prescribed burning
Agriculture burning
Coal refuse fires
Structural fires
Misc. (welding, grain silos, etc.)
Total
NO , Ggb
x y
150
138
30
13
53
6
45
435
These tables are included in Appendix A in English units.
By convention all NOX emissions are reported as equivalent NO-.
Approximately 95 percent of the N0x from stationary source combustion
is emitted as NO.
2-33
-------
is by far the largest source of noncombustion NOX emissions, contributing nearly 62 percent of the
total. Although NO emissions from the manufacture of sulfuric acid are a result of.combustion of
sulfur in the feedstock with gas or oil, this source is included here rather than with the combus-
tion sources.
2.3.8 Other NOX Emissions
Other sources of NO emissions include forest fires, prescribed burning, and structural fires.
Estimates of emissions from these sources are very inconsistent. A composite of estimates and data from
several sources (References 2-32, and 2-35, and the 1976 NEDS National Emission Report) is given in Table 2-21
2.4 SUMMARY OF 1974 NOX EMISSIONS AND FUEL CONSUMPTION
This section presents a summary of the 1974 estimated NOX emissions and fuel consumption by
sector and fuel. This will be followed -by comparisons with other sector inventories, primarily
References 2-4 and 2-15.
A summary of total NO emissions by fuel and sector compiled from the best available data is
presented in Table 2-22. Table 2-23 summarizes the 1974 fuel consumption by sector. The NOX emissions
estimates are further summarized in Figure 2-2. Tables 2-24 and 2-25 compare these data with pre-
vious estimates: MSST, Reference 2-4; GCA, Reference 2-15; ESSO, Reference 2-31; AP-115, Reference
2-26; and OAPQS, Reference 2-32. Exact comparison to other sources is virtually impossible since
each chose to present NO sources grouped under different headings. These tables demonstrate that
A
the present set of data, while based on much more detailed breakdowns, are in reasonable agreement
with previous estimates.
2.5 NOX EMISSION TRENDS AND PROJECTIONS
Nationwide NO emission trends from 1940 to 1972 as compiled by the EPA (Reference 2-37) are
illustrated in Figure 2-3. In general, stationary sources comprise between 60 and 70 percent of the
total NOX production, as shown in the figure. Figure 2-4 compares the EPA figures with the ESSO
(Reference 2-31) estimates published in 1968. The slight downward trend in 1971 of the EPA data is
due to revised emission factors. As can be seen from the figure, 1972 emissions have already attain-
ed the 1978 ESSO estimate.
Projections for nationwide NO emissions from stationary sources have been made by the
National Academy of Sciences (Reference 2-37) based on several assumptions, including consideration
for various control options. These projections with extrapolation to the year 2000 are presented in
Table 2-26. Assumptions made for these projections are:
2-34
-------
CJ
on
o
o
HH
CO
JS
o
£
I
*ji
,_
g
3
CJ
t-
0
J %
CJ
O
*" *
r*~
re
o
<*-
o
M
(U
u
>
>
;
>
CM
,_
**"
CM
?
in
0
r~
»
in
£?
VO
*
CO
VO
^
C3
.^^
VO
CO
CM
** *
CO
*
CO
X ^
4^
r
^
ID
r>» r^
t CO
vo vo
CO O
»
cn CM
CO VO
r- CO
«* CM
CM 0
co C
(O -r-
(/> U -O
0> O W.
C &- 3
"~ .? "~
CUM
O
cn
,^
CO
CM
*
CO
00
CM
CM
CO
VO
*~*
in
vo
I"*;
0
^-^
co
in
*w
o
o
f^
0
^^>*^
CM
vo'
VM^
CO
in
f*«
O
to
l.
V
r*
o
CO
f-BB
(O
j-
4-)
CO
3
o
i <
Ol
0
cn
0
cn
O
o
^
CO
in
r«
^^
vo
C3
^^
«*
CD
%«^
cn
O
O
^ ^
CO
CO
^H^
8
CD
re
c
(U
f~
CO
OJ
^^
^B
re
u cn
L. C
t re
o o»
oz
*S"
^J.
cn
in
CO
CM
CO
O
,
1
1
cn
c
r-
^j
o
. .
o
o
ps,,.
f^
%
C\J
p
CVJ
OJ
*"*"*
s^
p^
VO
CM
^_^
CO
^
CO
CO
CO
CM
^!
I
^
r
5
CM
in
*
T
c co
8E
2-35
-------
TABLE 2-23. SUMMARY OF ANNUAL FUEL USAGE3, 1974
Sector
Utility Boilers
1C Engines
Reciprocating
Turbines
Industrial Boilers
Commercial Boilers
Residential Heating
TOTAL
Fuel Usage6 - EJ
(percentage of total)
Gas
3.376 ( 6.95)
1.063 ( 2.19)
0.642 ( 1.32)
7.230 (14.87)
2.387 ( 4.91)
5.597 (11.51)
20.295 (41.75)
Coal
9.935 (20.44)
-
-
3.294 ( 6.78)
0.282 ( 0.58)
-
13.511 (27.80)
Oil
2.998 ( 6.16)
0.268 ( 0.55)
0.301 ( 0.62)
4.208 ( 8.66)
3.620 ( 7.45)
3.406 ( 7.01)
14.801 (30.45)
Total
16.309 (33.55)
1.331 ( 2.74)
0.943 ( 1.94)
14.732 (30.31)
6.289 (12.94)
9.003 (18.52)
48.607 (100.0)
aThis table is included in Appendix A in English units.
Excludes process fuel.
2-36
-------
INCINERATION
GAS TURBINES
-i BOTHERS
NONCOMBUSTION
12 O/ oc / Vv,-INDUSTRIAL PROCESS
HEATING
9.0
OMMERCIAL/
RESIDENTIAL
SPACE HEATING
RECIPROCATING 1C
ENGINES
ESTIMATED NOX EMISSIONS
SOURCE
UTILITY BOILERS
INDUSTRIAL BOILERS
RECIPROCATING 1C ENGINES
COMMERCIAL/RESIDENTIAL HEATING
INDUSTRIAL PROCESS HEATING
NONCOMBUSTION
GAS TURBINES
INCINERATION
OTHER
TOTAL
Tg
5.105
2.218
2.413
1.090
0.432
0.203
0.236
0.039
0.435
12.171
1fl6 tons
5.628
2.444
2.660
1.202
0.476
0.224
0.260
0.043
0.479
13.416
Figure 2-2. Summary of 1974 stationary source NOX emissions.
(Pie chart units in percent).
2-37
-------
to
Z
o
o
CO
o
CO
a:
CO
LO
0
LO
O
LO
«*
LO
00
CO
CM
^
00
ro
to
CO
^
0
CO
^fc
~f->
r
^"*
r
M
CO VO
f CO
*S- CM
CM O
en LO
«s- <
r O
LO en CD
CM en CM
CM i O
<+- <*-
4- 4-
o
en
o
O^
c to
f- CO
4J C
re t-
O JD
v> o i-
Q> t. 3
c ex H-
en o
c 0) re
LU OC O
o
14
00
CM
CM
,
co
r~
^
r-
LO
CO
^
^
LO
LO
CM
to
L.
CO
r~
o
CO
la
p
L.
M
(0
3
TJ
C
» i
O
CD
O
r~-
0
CO
co
o
CO
LO
CD
CM
CD
en
o
re
r~
(J
i.
Q}
Es
^
O
o
0
CO
co
o
en
CM
O
CM
0
co
LO
CD
CM
LO
0
en
o
r »
re
Jj
c
CO
o
r«
to
0)
O£
CM
CO
O
I
1
LO
CO
0
0
CO
o
o
en
c
1
-M
re
0)
to
to
Q)
O
o
a.
CO
o
CM
CD
I
1
to
3
,,Q
F
O
O
C
o
z
en
co
0
CD
LO
r
O
CD
0
O
O
O
0
0
0
c
o
-J3
re
s.
CO
c
r"
o
LO
co
st
CD
CO
CO
CO
^
CO
^z
4^
O
^
CM
LO
~;
CD
CO
LO
CD
O
cn
ro
cn
en
cn
_)
^^
(«
O
1
to
c
3
-C
to
c
LU
C
X
o
c
CO
ex
Q.
>
S-
r re
i C
re o
C 4J
o re
4-> 10
c
CO E
> o
C i-
O «4-
0
X
^">^^
CO Z
.a
to
L.
CO
r*~
^
0
j^
0)
N
r
to
pw
re
f
s-
4^
to
3
TD
r
c
T3
O)
TO
3
O
C
|_H
O
.
>^
r
C
O
to
4-)
C
re
^
Q.
to
re
cn
O
c
re
CO
c
^
1
CO
a.
r
a.
o
ro
4-*
CO
-o
C
^
^3
a;
TJ
3
i
a
c
^>
4_>
O
g^
0)
depending on use.
r*~
re
r~
^.
4->
i/)
3
"O
C
^
T3
c
rO
>^
4_>
^
f
r"
4_>
rs
c
»^-
-a
-------
ro
CO
O
CO
ti
C£
Q.
O
O
tt
r-
D_
2:
ID
CO
z
o
<"
°c:
r- C\?
co r<-.
CO C^
2! r-
LO j-r
[H vo
j£
II
S-
0
M
0
0)
CO
CT»
O
CO
!£
UD
LO
O
VO
LO
p
00
CM
00
to
t»
QJ
p
p-
O
CO
>^
^_>
p-
p
p
4-)
ID
CO
CO
r
CM
»
CM
CO
CO
.
1
1
^>
O4
O
p
^J"
r
cr>
CO
1
1
cn
c
p- O»
M C
to *p
O JD
to o s-
0> S- =J
C Q. r-
C7) CJ to
C 0) «0
LU OH O
0
CM
CO
f*»
P^
CM
f^-
00
^t"
CM
JD
0
O
""
JQ
£
to
L.
0)
p
O
CO
r
«3
r-
t^
1 »
to
3
TD
c~
i t
CO
CM
t.f^
^^
^O
^
^o
CO
CO
o
o
en
vo
CO
CO
CM
O
r
CM
O
r
r*
0
c
p-
r
to
o
p-
S-
F
O
0
*»k
0
0
r
10
r
«4_»
C
o
p-
to
cu
Di
00
O
*o
CO
cr>
o
cr»
5
CM
O
CM
LO
i
re
to
tO
OJ
C
-P-
cn
c
a)
to
O)
2-39
-------
O TOTAL EMISSIONS
DSTATIONARY FUEL COMBUSTION
AROAD VEHICLES
ELECTRICITY GENERATION
INDUSTRIAL FUEL COMBUSTION
AINDUSTRIAL PROCESS LOSS
x
o
1940
1950
1960
YEAR
1968 1970 1972
1969
Figure 2-3. Nationwide annual NOX emission trends 1940 - 1972 (Reference 2-37).
2-40
-------
25
20
20
15
X
o
10
15
10
UJ
X
o
J
EPA
ESSO REPORT
Ol
1950
1960
1970 1980
YEAR
1990
2000
Figure 2-4. Annual stationary source NOX emission trends.
2-41
-------
o
o
o
CVJ
o
LU
O
a:
D_
co
o
ii
CO
CO
I-H
LU
X
o
z
LUCT3
3 co
Z I
O CVJ
I-H
J VD
VD
S-
z VD
z o:
VD
CVJ
I
CVJ
LU
_1
00
.a
h-
to
0
r
to
to
E
LU
X
0
z
O
o
o
CVJ
O
en
en
in
CO
en
o
00
en
CVJ
en
>J
s-
o
CD
VD
(->
to
_«
O
CO
^
CO
-
c
0
r-
+J
to
3
O
E
O
o
r
VD
3
U_
>5
S_
to
c
o
r
+J
«3
4->
CO
xx
CO
in
in
^^,
o in i
P^. 00 VO
en en
CO CO
in ^j-
c
o
r
M
to
l-
VD
C
VD
CD
U -
r- S_
i- !->
4-> CO
O 3
VD "O
r- C
LU i i
en
vo
CD
en
in
CD
r-
C
0
r-
^_>
3
^
<4_)
to
C
1
r
to
r
U
s-
Q^
£
P
O
o
CO
CD
CVJ
o
1
to
*f*
4_>
c
VD
r
CO
VD
OL
vo
CO
o
^.
o
to
VD
to
to
O
_J
to
to
VD
O
0
a.
f
to
r
S-
4_>
CO
3
o
C
t4
0
CVJ
o
vo
CD
f
to
to
0
a.
to
r-
Q
VD
to
to
*^e
"0
r-"
O
CO
vo
CD
un
CD
to
3
O
VD
C
to
r
i_
VD
U
to
r~
^
r
CVJ
VO
r
^
^~x
VD
CVJ
^
«.
j-
CVJ
_J
1
O
1
.
to
40
r-
C
3
t~
to
r-
n~
CD
C
LU
C
r~
to
^
.C
O
i.
VD
Q.
in
en
>,
F
VD
4_>
fO
^E
x
o
S-
Q.
O.
^^
.
CVJ
O O
y* ^^
4-> tO
C fO
VD
'tO VD
> 4_)
r- 4->
3 -r-
CT" ^~
VD VD
to to
fO T-
^3 C»
CD O
4-> !
t- -M
O CO
CL 3
VD Q
j^ F
O
VD U
S-
u
to J-
C 3
o o
i- tO
CO
w h
r- S-
£ co
VD C
O
XT-
o *->
Z (O
4_>
f tO
r
to E
O
c i-
0 M-
P
Jj S^
c o
VD Z
>
C VD
O .C
O 4->
^*+-
to
VD
to
VD
f
4-J
C
VD
S_
to
CL
c
r
c
VD
r-
cn
VD
t-
to
LO
en
r
S.
VD
4-)
t^.
fO
to
P
C
fO
ri
a.
£_
VD
2
0
Q.
t.
fO
VD
f
O
3
C
*
CD
C
O
c
S-
o
CO
c
o
to
to
>r~
E
VD
X
0
p^.
CO
1
CVJ
CD
O
C
VD
VD
4-
VD
a:
c
r-*
t.
O
i.
L.
VD
cn
c
r
s-
o
a.
VD
s_
CVJ
en
r
J-
O
^3
VD
U
VD
S-
c
O
0
to
VD
to
to
o
to
to
VD
0
0
a.
r
to
r
j_
4J
to
3
a
c
IO
CQ O
U
2-42
-------
Implementation of NSPS (1972) for utility boilers and nitric acid plants
Electrical demand grows at about 6.5 percent per year
No increase in oil consumption after 1975
t The 1940 to 1972 growth rate of NO emissions from industrial, commercial, and
institutional sources will be reduced over the next 30 years to 2.63 percent
per year due to a shift to electricity
Two cases for electric power generation are considered. One assumes that most new electric
power plants will be nuclear; the other assumes no new nuclear plants after 1975. Neither of these
are realistic but were considered, at the time they were made, to bracket the possible cases. The
uncertainty of projections of this nature is compounded by several emerging trends:
t There will be a significant increase in the utilization of coal in power generation,
leading to an intensified NO problem unless stringent controls are adopted
t Industrial area sources may be switching from gas to oil or coal, resulting in
larger potential NO emissions
The potential application of alternate fuels is difficult to quantify at this
time (probably 10 years away)
The recent emphasis on energy conservation has produced lower than expected energy
growth rates in the industrial and utility sectors
In view of these trends, the confidence level of any speculations on growth rates of specific
nt/fuel combina
include the following:
equipment/fuel combinations 1s very low. Other significant factors affecting future NO emissions
0 Major technological developments 1n equipment design, fuels and fuel treatment, combus-
tion control and exhaust gas cleanup
Uncertainty concerning the future of nuclear energy as a major source of electrical
power
The degree to which NOX emissions will be regulated by both local and federal agencies
More recent projections of stationary source NOX emissions have been made in Reference 2-38.
Several growth and control scenarios are considered. The energy use patterns are based primarily
on FEA (Reference 2-39) and ERDA (Reference 2-40) reports. Emission factors consider retirement
rates of old equipment and various levels of NSPS (Including existing, proposed, and possible) for
2-43
-------
TABLE 2-27. ESTIMATED FUTURE NSPS CONTROLS (Reference 2-38)
NOX Source
Utility and Large
Industrial Boilers
(>73 MW)a Coal
Oil
Gas
Large Packaged Boilers
(>7.3 MW)a Coal
Oil
Gas
Small Packaged Boilers
£7.3 MW)a Coal
Oil
Gas
Small Commercial and
Residential Units
Oil
Gas
Gas Turbines
1C Engines Dist Oil
Nat Gas
Gasoline
Process Combustion
Date Implemented
1971
1977
1981
1985
1988
1971
1971
1979
1985
1990
1979
1979
1979
1979
1979
1983
1983
1977
1983
1972
1985
1979
1985
1979
1985
1981
1990
Standard (ng/J)
300
258
215
172
129
86
129
258
215
172
129
86
50% reduction
86
129
30
17
129
86
1390
1040
1240
930
950
710
20% reduction
40% reduction
aThermal input
2-44
-------
fO
CM
CU
O
C
CU
CU
»*-
CU
a:
o
o
o
CM
00
o
tI
to
II
s
X
o
UJ
o
00
CM
CM
00
JD
I-
C
O
r-
U
J
3
D
)
O
O
O
^V 1
vNI
LO
00
^
o
o
CM
LO
00
en
h
**
%
i
vo r^ cr» CM o vo CM
CM LO CM Kf O 1"-. CM
O 00 CO CM CO O CO
VO r r O O O O
O CO CT> r O CO CTt
CM ^f CM VO VO LO CO
«3- O VO CM CM O CM
VO CM r O O O O
^- VO C7> CM O VO CM
O «5f CM *» O r-. CM
i VO CO CM CO O CO
o o o o o
o CM cr» i o co cr»
r- CO CM VO VO LO CO
** CO VO CM CM O CM
r>. CM r- o o o o
VO LO r^ r CM O CO
VO ^J" CT» CM CO ^" CT>
LO CO CM CO «5f O r
LO CM CM O O O O
C C CU
O O to
4-> «/> 4-> en cu
(O S- to C O
1- CU 3 -r- O
CU i XI 4-> t. C
C v- E to D_ C O
CU O O CU O T-
CD CO O 31 r »- 4->
O XJ r 1- -r- fO 3
r- CU «r- 1- i- .d
J- C7> C CU E
4-> «T3 J- W C O
o .*: cu E 3 - u
CU U 4J J- -O 0 C
o
r
o
^jl
^J-
00
'
CM
cy>
-
_J
I
o
1
to
r-
c
3
J^
to
r-
'o>
C"
UJ
c
1
<:
X
r-
"2
CU
CL
Q.
^^
c
o
CU
"T3
3
r^
u
c
r
(/)
r-
a>
t
^
o
to
r-
^1
1
c
o
4->
to
3
JO
CME
O 0
z u
4-> CU
c u
CU J-
r 3
(O O
^> to
'3 >»
cr s-
QJ 03
tO O
ro T-
CU 4->
4-> tO
s_
0 E
Q. O
CU i-
S- H-
a> x
$- o
to CU
C JC
0 4->
p-
tO M-
tO O
r-
^* I_J
0) C
CU
X U
o s-
Z CU
CL
f^
i LO O
to cr> z
C >j U)
O i 03
r- CU
4-> 4-> TO
C fO CU
CU E 4J
> *r~ 4->
C X f-
O O E
(J S- CU
CL.
>) CL. to
OQ cC «i-
2-45
-------
20
V)
<
CM
o
X
O'
OTOTAL NOX - PRESENT CONTROLS
D TOTAL NOX - NSPS CONTROLS
AUTILITY BOILERS-PRESENT CONTROLS
OUTILITY BOILERS - NSPS CONTROLS
1974
2000
Figure 2-5. Annual stationary source NOX emissions projections - low nuclear
(Reference 2-38).
2-46
-------
15
O TOTAL NOX - PRESENT CONTROLS
D TOTAL NOX-NSPS CONTROLS
AUTILITYBOILERS -PRESENT CONTROLS
OUTILITY BOILERS - NSPS CONTROLS
1974
2000
Figure 2-6. Annual stationary NOX emissions projections - high nuclear
(Reference 2-38).
2-47
-------
new equipment. The most stringent NSPS considered are given in Table 2-27. Table 2-28 presents a
breakdown by end use sector for two of the cases considered in Reference 2-38. The main assumptions
are:
NSPS as shown in Table 2-27
Growth in electrical demand of 4.4 percent per year
Continuation of current consumption patterns
The low nuclear case considers 35 percent of new electrical capacity to be supplied by nuclear
power and the remaining 65 percent by coal-fired boilers. The high nuclear case reverses these per-
centages. Both of these cases are shown graphically in Figures 2-5 and 2-6. For comparison the same
growth cases with current control levels only are also shown. The potential reduction through a
vigorous control program is evident. Comparison of Figures 2-5 and 2-6 and Table 2-26 shows a size-
able difference in projected emissions for the various assumptions. A large part of the difference
is due to a downward revision in the growth of electrical demand in Reference 2-38 to reflect recent
trends. These results further illustrate the difficulty of projecting emissions very far into the
future.
REFERENCES FOR SECTION 2
2-1 Chamot, E.M., D.S. Pratt, and H.W. Redfield, "Journal of the American Chemical Society,"
33, 366, 1911.
2-2 Code of Federal Regulations, Title 40, Part 60, Appendix A, Method 7. See also: Hami1, H.
and R. Thomas, "Collaborative Study of Method for the Determination of Nitrogen Oxide
Emissions from Stationary Sources," SwRI-EPA Contract 68-02-0626, May 8, 1974.
2-3 Federal Register, Vol. 41, No. 232, December 1, 1976.
2-4 Mason, H.B. and A.B. Shimizu, "Definition of the Maximum Stationary Source Technology (MSST)
Systems Program for NOX," (Draft Report) Aerotherm Final Report 74-123, Acurex Corporation,
Aerotherm Division, October 1974.
2-5 "Compilation of Air Pollution Emission Factors (Second Edition)," Publication No. AP-42,
Environmental Protection Agency, Research Triangle Park, North Carolina, April 1973.
2-6 Supplement No. 4 of Reference 2-5, January 1975.
2-7 Supplement No. 5 of Reference 2-5, December 1975.
2-8 Supplement No. 6 of Reference 2-5, April 1976.
2-9 Cato, G.A., H.J. Buening, C.C. DeVivo, B.C. Morton, and J.M. Robinson, "Field Testing:
Application of Combustion Modification to Control Pollutant Emissions from Industrial
Boilers - Phase 1," KVB Engineering Inc., EPA-650/2-74-078a, Research Triangle Park, N..C.,
October 1974.
2-10 Crawford, A.R., E.H. Manny, and W. Bartok, "Field Testing: Application of Combustion Modi-
fications to Control NO Emissions from Utility Boilers," Exxon Research and Engineering
Company, June 1974.
»
2-11 Offen, G.R., et al., "Standard Support Document and Environmental Impact Statement - Sta-
tionary Reciprocating Internal Combustion Engines" (Draft Report). Acurex Corp./
Aerotherm Division, Mountain View, California, Project 7152, March 1976.
2-48
-------
2-12 Environmental Protection Agency, "Standard Support and Environmental Impact Statement for
Standards of Performance: Lignite-Fired Steam Generators," Final Draft, OAPQS, March 1975.
2-13 "Standard Support and Environmental Impact Statement, Vol 1: Proposed Standards of Perfor-
mance for Stationary Gas Turbines," EPA-450/2-77-017a, September 1977.
2-14 Personal communication with H.J. Melosh III, Foster Wheeler Corporation.
2-15 Surprenant, N.F. et. al., "Preliminary Emissions Assessment of Conventional Stationary Com-
bustion Systems,"~GCA Corporation, EPA Report No. 600/2-76-046b, March 1976.
2-16 FPC News, Federal Power Commission, Washington, D.C., June 6, 1975.
2-17 Dykema, O.W., and Kemp, V.E., "Inventory of Combustion Related Emissions from Stationary
Sources," (First Update). Aerospace Corporation, EPA-600/2-77-066a, March 1977.
2-18 Putnam, A.A., Kropp, E.L., and Barrett, R.E., "Evaluation of National Boiler Inventory,"
Battelle Columbus Laboratories, October 1975.
2-19 Personal communication with R.R. Vosper, Coen Company, January 1977.
2-20 "Current Industrial Reports, Steel Power Boilers," 1968-1975, U.S. Department of Commerce,
Bureau of the Census.
2-21 "Patterns of Energy Consumption in the United States," Stock No. 4106-0034, Stanford Research
Institute, Menlo Park, California, January 1972.
2-22 Dupree, W.G., and J.S. Corsentino, "Energy Through the Year 2000 (Revised)," Bureau of Mines,
December 1975.
2-23 "National Gas Survey," Preliminary Draft, Federal Power Commission, 1974.
2-24 Barrett, R.E., Miller, S.E., and Locklin, D.W., "Field Investigation of Emissions from Com-
bustion Equipment for Space Heating," Report No. EPA-R2-73-084a, Prepared by Battelle Memo-
rial Institute, Columbus, Ohio, June 1973.
2-25 "Nationwide Inventory of Air Pollutant Emissions 1968," Pub. No. AP-73, Environmental Pro-
tection Agency, Research Triangle Park, North Carolina, July 1971.
2-26 Cavender, J.H., Kircher, D.S. and Hoffman, A.I., "Nationwide Air Pollutant Emission Trends
1940-1970," Pub. No. AP-115, Environmental Protection Agency, Research Triangle Park, North
Carolina, January 1973.
2-27 Crump, L.H. and Reading, C.L., "Fuel and Energy Data -United States by States and Regions,
1972," Information Circular 8647, Bureau of Mines, Department of Interior.
2-28 "Monthly Energy Review," Federal Energy Administration, July 1976.
2-29 McGowin, C.R., "Stationary Internal Combustion Engines in the United States," Report No.
EPA-R2-73-210, Prepared by Shell Development Company, Houston, Texas, April 1973.
2-30 Foley, G., "Industrial Growth Forecasts," Stanford Research Institute, Contract No. 68-02-
1320, September 1974.
2-31 Bartok, W., et. al_., "Systems Study of Nitrogen Oxide Control Methods for Stationary Sources -
Vol. II," prepared for National Air Pollution Control Administration, NTIS Report No. PB-192-
789, Esso Research and Engineering, 1969.
2-32 "OAQPS Data File of Nationwide Emissions, 1971," Office of Air Quality Planning and Standards,
Environmental Protection Agency, May 1973.
2-33 Goldish, J. et. al_., "Systems Study of Conventional Combustion Sources in the Iron and Steel
Industry," Report No. EPA R2-73-192, Prepared by Walden Research Corporation, Cambridge,
Massachusetts, April 1973.
2-34 Oil and Gas Journal, Volume 73, No. 12. The Petroleum Publishing Company, Tulsa, Oklahoma.
2-35 "Annual Survey of Manufacturers 1974 - Fuels and Electric Energy Consumed," U.S. Department
of Commerce, Bureau of the Census.
2-49
-------
2-36 Ketels, P.A., J.D. Nesbitt, and R.D. Oberle, "A Survey of Emissions Control and Combustion
Equipment Data in Industrial Process Heating," Institute of Gas Technology, Final Report
8949, October 1976.
2-37 National Academy of Science, "Air Quality and Stationary Source Emission Control," Prepared
for the Committee on Public Works, United States Senate, Serial No. 94-4, March 1975.
2-38 Salvesen, K.G., et ^1., "Emissions Characterization of Stationary NO Sources," Aerotherm
Draft Report, TR-77^72, October 1977. x '
2-39 "1976 National Energy Outlook," Federal Energy Administration, FEA/N-75/713, February 1976.
2-40 "A National Plan for Energy Research, Development & Demonstration: Creating Energy Choices
for the Future," ERDA-48, Volume 2 of 2.
2-50
-------
SECTION 3
CONTROL TECHNIQUES
This section presents a survey of the general principles and developmental status of poten-
tial techniques for NO control for stationary sources. It is intended to provide a broad perspec-
tive on the various suggested concepts for NO control by combustion process modification and
flue gas treatment for combustion sources and by tail gas cleanup for noncombustion sources. A
more detailed review of the effectiveness and cost of control implementation on specific equipment
types is given in Sections 4, 5, and 6.
3.1 COMBUSTION MODIFICATIONS
Modifying the combustion process is the most widely used technique for reducing combustion
generated oxides of nitrogen. This section describes the four most popular methods: modification
of the operating conditions, equipment design modification, fuel modification, and use of alternate
combustion processes. The section begins by describing the factors which affect the generation of
NO during combustion.
3.1.1 Factors Affecting NOX Emissions from Combustion
Oxides of nitrogen formed in combustion processes are usually due either to thermal fixation
of atmospheric nitrogen in the combustion air, leading to "thermal NO ", or to the conversion of
A
chemically bound nitrogen in the fuel, leading to "fuel NOX". For natural gas and light distillate
oil firing, nearly all N0y emissions result from thermal fixation. With residual oil, crude oil,
and coal, the contribution from fuel-bound nitrogen can be significant and, under certain operating
conditions, predominant.
A third potential mechanism of NO formation arises in processes such as glass manufacturing,
where the raw materials in contact with the combustion products contain nitrogen compounds. Little
is known about the extent of conversion to NO of these nitrogen compounds, or of the effects of
combustion modifications on this mechanism.
3-1
-------
3.1.1.1 Thermal NOX
The detailed chemical mechanism by which molecular nitrogen in the combustion air is con-
verted to nitric oxide is not fully understood. In practical combustion equipment, particularly
for liquid or solid fuels, the kinetics of the N2-02 system are coupled to the kinetics of hydro-
carbon oxidation and both are influenced, if not dominated, by effects of turbulent mixing in the
flame zone. It is, however, generally accepted that thermal NOX forms at high temperatures in an
excess of air. The usually stable oxygen molecule dissociates to oxygen atoms which are very
reactive. These atoms react with the otherwise stable nitrogen molecule to form NO .
The most widely accepted reactions that describe the formation of thermal NO are those of
the extended form of the Zeldovich chain mechanism (Reference 3-1):
02 + M^O + 0 + M (3-1)
N2 + 0 t NO + N (3-2)
N + OH 2 NO + H (3-3)
02 + N ^ NO + 0 (3-4)
Equation (3-1) is considered to be in equilibrium, and M is a "third body", normally taken to be
molecular nitrogen. For thermal NOX, reaction (3-2) is much slower than reaction (3-3) and, there-
fore, controls the rate of NO formation. The creation of an NO molecule from reaction (3-2) is
accompanied by the release of an N atom, which rapidly forms another NO molecule from reaction
(3-3) and (3-4). Reactions (3-2) and (3-4) are the chain-making and chain-breaking mechanisms,
and the oxygen atom is the chain carrier.
Experimental measurements of NO formation in heated mixtures of N2> 02 and Argon at
atmospheric pressure have been correlated with an equation of the form (Reference 3-2):
[NO] = k, e-k2/T [N2] [02]1/2 t (3-5)
Where: [] = mole fraction
T = absolute temperature
t = residence time
k,, k2 = constants
This expression reflects the strong dependence of NO formation on temperature. It also shows that
NO concentration is directly proportional to N2 concentration and to the residence time, and varies
3-2
-------
with 02 to the one-half power. A rate expression such as this one does not fully describe the
thermal NO reaction mechanism, but it does give some valuable qualitative trends.
The temperature and time dependencies of NO formation are illustrated in Figure 3-1 for
idealized conditions (References 3-3 and;3-4). The results at 0.01 sec for three values of the
stoichiometric ratio (S.R.) show, as expected, that NO formation is suppressed by reduced avail-
ability of oxygen. In a practical combustor, departure from S.R. = 1 would result in reduced
temperatures which would further suppress NO formation. It is precisely these factors of high
sensitivity to temperature, oxygen concentration level, and time of exposure which make the forma-
tion of thermal NO susceptible to combustion modification.
Ideally, then, the formation of thermal NO could be reduced by four tactics: (1) reduce
nitrogen level at peak temperature, (2) reduce oxygen level at peak temperature, (3) reduce peak
temperature, and (4) reduce time of exposure at peak temperature. In typical hydrocarbon-air
flames, [N~] is of the order 0.7 and is relatively difficult to modify. Therefore, field practice
has focused on reducing oxygen level, peak temperature, and time of exposure in the NO -producing
region of the combustor. (Reference 3-5.) These parameters are in turn dependent on secondary
combustion variables such as combustion intensity and internal mixing in the flame zone - effects
which are ultimately determined by primary equipment and fuel parameters over which the combustion
engineer has some control. A hierarchy of effects leading to thermal NO formation is depicted
in Table 3-1. Although causal relationships between the four categories shown in Table 3-1 are
not firmly established, combustion modification technology is, nevertheless, confronted with the
task of reducing thermal NO through modification of equipment and fuel parameters. This task
has been approached with efforts ranging from the short-term testing of equipment modifications
on commercial units, in order to determine the effect on NO emissions, to long-term fundamental
studies and pilot testing directed at achieving a basic understanding of NO formation.
Combustion modification techniques such as lowered excess air and off stoichiometric or
staged combustion have been used to lower local 0? concentrations in boilers. Also, staged
combustion in the form of stratified charge cylinder design has been used successfully in 1C
engines. Since gas turbines typically operate at excess air levels far greater than stoichiometric,
lowering excess air levels in this equipment class does not control thermal NO .
Flue gas recirculation and reduced air preheat have been used in boilers to control thermal
NO by lowering peak flame temperatures. Analogously, exhaust gas recirculation (EGR), reduced
manifold air temperature (1C engines) and reduced air preheat (regenerative gas turbines) have been
3-3
-------
2800
6000
3000
TEMPERATURE, °F
3200 3400
3600
3800
4000
PRIMARY ZONE-ADIABATIC
NO FORMATION
2376 K (3800 °F)
FOR 0.05 sec-^150 ppm
^ADIABATIC
TEMP.
RECIRCULATION ZONE
NO FORMATION
1922 °K (3000 °F)
FOR 2.0 sec-^-50 ppm
S.R. - STOICHIOMETRIC RATIO
10
1800
1900
2000
2100 2200
TEMPERATURE, °K
2300
2400
Figure 3-1. Kinetic formation of nitric oxide from combustion of natural gas at
atmospheric pressure (References 3-3 and 3-4).
3-4
-------
to *"
l«
i i
-o «
3 *
if °-
O
z:
o
QC
O
LU
:r
CJ3
O
o:
o
o
CO
Of
o
i
CO
CO
to
i-
cu
CU
S-
to
QL
C
o
to
3
O
o
to
-t-> J-
C OJ
CU -M
3 S-
CT" (O
LU D_
S- CU
E U.
£ -O
O. C
tO
E X
i- O
CU Z
(U
CU O. CU
i E CU O- Q-
CU S- E
C +-> 3 4-J CU
CU CO to -M
O
X CU X
O O- LU
CU
4->
C
C
o
CU +> O
+J tO 4J
fl3 3 C
S- XI -
i O to
tO O ->
> O
O C
r- O
4_> .,_
tO -M CU
S- t. C
o o
S- +J N
<0
I
CU
E
(O
4J CU O 03 4- CU
tO i- Cn S- r- i <4-
3 C Q.I*- i 3 O
_Q 4-> -r- to -O
E fO X O S-
O CU -r- O 3
03:2: __J h-
CU E S-
S- C S- »r- CO CO
3 O CU tO <-> CU
-> C -r- -4-J CU i
(O O>-M4->o8r-O CO
CU -M CO CO Q.r O -4-> S- 03 O
D.T- CU O CU S- S- -r- E '«-
CUO EO«4- O_ CO i T-
H->i XO'i- 4- «OT3
CUOO-M4- OS-l-C-O
+J > JD UO OCUS-rQ
CU CU r CU CU CO)
r- l_ CU '«-5 Nl i. 4J
C - 3 C T- 3 X
ii U. U. « CO CO LU
3-5
-------
applied to 1C engines and gas turbines. Other techniques designed to lower peak temperatures in
prime movers include water injection and altered air/fuel ratios.
Techniques which reduce residence time at peak temperature have been more easily applied to
prime mover equipment classes. Although flue gas recirculation (and EGR) reduces combustion gas
residence time, it acts as a thermal NO control primarily through temperature reduction. Tech-
niques which specifically reduce exposure time at high temperatures include ignition retard for 1C
engines and early quench with secondary air for gas turbines.
It is important to recognize that the above-mentioned techniques for thermal NO reduction
alter combustion conditions on a macroscopic scale. Although these macroscopic techniques have all
been relatively successful in reducing thermal NO , local microscopic combustion conditions ulti-
mately determine the amount of thermal NO formed. For example, recent studies on the formation of
thermal NO in gaseous flames have confirmed that internal mixing can have large effects on the total
amount of NO formed (References 3-6, 3-7). Burner swirl, combustion air velocity, fuel injection
angle and velocity, burner divergent angle and confinement ratio all affect the mixing between fuel,
combustion air and recirculated products. Mixing, in turn, alters the local temperatures and
species concentrations which control the rate of NO formation.
x
Unfortunately, generalizing these effects is difficult, because thp interactions are complex.
Increasing swirl, for example, may both increase entrainment of cooled combustion products (hence
lowering peak temperatures) and increase fuel/air mixing (raising local combustion intensity). The
net effect of increasing swirl can be to either raise or lower NO emissions, depending on other
system parameters.
In summary,a hierarchy of effects depicted in Table 3-1 produces local combustion conditions
which promote thermal NOX formation. Although combustion modification technology seeks to affect
the fundamental parameters of combustion, modifications must be made by changing the primary equip-
ment and fuel parameters. Control of thermal NOX, which began by altering inlet conditions and
external mass addition, has moved to more fundamental changes in combustion equipment design.
3.1.1.2 Fuel NOX
The role of fuel-bound nitrogen as a source of NOX emissions from combustion sources has
been recognized since 1968 (Reference 3-8). Although the relative contribution of fuel and thermal
NOX to total NOX emissions from sources firing nitrogen-containing fuels has not been definitively
established, recent estimates indicate that fuel NO is significant and may even predominate. In
A
one recent study (Reference 3-9), residual oil and pulverized coal were burned in an argon/oxygen
3-6
-------
mixture to eliminate thermal NOX effects. Results show that fuel NOX can account for over 50 per-
cent of total NO production from residual oil firing and approximately 80 percent of total NOX
from coal firing. Therefore, as coal is increasingly used as a national energy source, the control
of fuel NO will become more important.
A
Fuel-bound nitrogen occurs in coal and petroleum fuels. The nitrogen containing compounds
in petroleum tend to concentrate in the heavy resin and asphalt fractions upon distillation (Refer-
ence 3-10). Table 3-2 gives analyses of typical fuel oils. Fuel nitrogen is less than 0.01 per-
cent for distillate oils; however, it ranges from 0.1 to 0.5 for residual oils.
The classes of nitrogen compounds in fuel oil include indoles, quinolines, pyradines, and
carbazoles. Their quantities in the distilled fractions vary with the origin of the crude oil.
From one California crude, pyradines dominated in the distillate fraction and carbazoles dominated
in residual oil (Reference 3-11).
Table 3-3 gives analyses of four ranks of U.S. coals. Nitrogen content of most U.S. coals
lies in the 0.5 to 2 percent range (Reference 3-12); anthracite coals contain the least and bitu-
minous coals the most nitrogen. Although the structure of coal is not known with certainty, it
is believed that coal-bound nitrogen occurs in aromatic ring structures such as pyridine,
picoline, quinoline, and nicotine (Reference 3-10). Figure 3-2 illustrates the nitrogen content
of various U.S. coals, expressed as ng N02 produced per Joule for 100 percent conversion of the
fuel nitrogen. The figure clearly shows that if all coal-bound nitrogen were converted to NOX>
emissions for all coals would exceed New Source Performance Standards. Fortunately, only a frac-
tion of the fuel nitrogen is converted to NOX for both oil and coal firing, as shown in Figure 3-3.
Furthermore, the figure indicates that fuel nitrogen conversion decreases as nitrogen content
increases. Thus, although fuel N0x emissions undoubtedly increase with increasing fuel nitrogen
content, the emissions increase is not proportional. In fact, recent data indicate only a small
increase in NO emissions as fuel nitrogen increases (Reference 3-14). From observations such as
A
these, the effectiveness of partial fuel denitrification as a NOX control method seems doubtful.
The precise mechanism by which fuel nitrogen is converted to NOX is not understood; however,
certain aspects are clear, particularly for coal combustion. In a large, pulverized coal utility
boiler, the coal particles are conveyed by an airstream into the hot combustion chamber, where
they are heated at a rate in excess of 10"K/s. Almost immediately volatile species, containing some
of the coal-bound nitrogen, vaporize and burn homogeneously, rapidly (-10 ms) and probably detached
from the original coal particle. Combustion of the remaining solid char is heterogeneous and much
slower (~300 ms).
3-7
-------
CM
I
CO
CU
O
CU
CU
o:
oo
__i
14
O
__l
UJ
ID
U.
00
oo
UJ
00
CM
I
CO
00
vo
d
z
r-
O
LO
=3 O
-a z
to
cu
C£
-
d
CM
d
i Z
^
0
cu
re
r II
cr>
LO
CDr
VO CM
COf
CD
CM
CM
CM
O CM
1 1
VO CO
CD
O
LO
O CM
1 1
O CO
r CM
CD
C0'~*
LO
O CO
1 1
OO O
oo co
r- n
J? 1 °°^
O » H CM
D O ^f
d B *a- LO
z n co co
1 CD ^
tt
|
1 -M*
i **""
fO
1 C»
II
o
r
M- - .
r 1 i
O O.
cu <:
CLO
00 '
00 O*
«d- O
CT> O
LO O
cr> o
^^
o
LO O
1 1
CM CD
CO LO
r
LO
VO CM
CM i
1 1
VO LO
CO CO
1 1
O CO
CO
CM '
1
(/)
v^
04 *-*
E 0
CU
> tO
>>
Viscosit
(Say bolt
^
, .
*3- LO
«* r LO
1 1 1
CD 00
^2,
CM
«* i CO
O VO r
CM ^J-
*"
O
o
LO *fr
*d- r OO
1 1 1
^00 r
CM cr>
^
00
. 0> 00
LO CO
«d- r- 0
LO CM CM
^- vo o
^j- co
d-
CM
. VO CO
vo co
«3- r O
vo r^ i
LO CM CD
^- co
CD
0
«
~J^
r (0 ^
>, ^^ ^
13
C 00 iT
i- 13
tO O r
CU r 13
nr oo
CM 00
00
1 1
O LO
r 00
CM 00
r CO
1 1
O VO
i CO
CM CO
r- 00
I
1 1
r VO
r~ 00
*f CO
00
| I
CM VO
r CO
«a- co
00
1 1
CM VO
CO
^^W^MMMBMMHM^MMMM
4J
1 '
n ^
c
cu
0) c
0 0
i- -0
o $-
^> rt3
3C 0
*^*
CO LO . .
> . . CM OO
O O r CM
1 1 II
o r "^
O CM LO
O x_x
^"""**»
O &\
i LO . .
CM r>v
o o r CM
< i it
5 r ^
O CM LO
CD ^_^.
r>. CM
i LO . .
O O r^ CM
5 r *^
O CM LO
o
f f CM r
O O CM LO
o o o o
V V V
i p CM r-
O O CM LO
O O O CD
V V V
cu
VH-
^~
o
r_)
*J ^*
5 a*^
^-*
^§ c c
M CU CUco
5 cn CTO
0 0 r-
^r *~ "^
tO -t- 'r- r
^c z z ^^
3-8
-------
CVI
r
CO
cu
cu
cu
M-
cu
oe:
CD
CO
o
o
CO
3
CO
LU
CO
CO
CO
CO
r_
n
1
.O
CO
CO
3
o
C
f^
3
r
CO
cu
£f
o
^_
.C
c
CO
r-
CO
i
c
in
in
CVJ
o
CO
i
in
r
J£
i
CVI
in
i
CVJ
^^
4-*
2
CU
&-
4->
CO
r
^
O
CO
i
in
CVJ
O
i
o
CO
o
1
00
CVI
i
in
+J
*
i_
cu
^3
4_>
CU
(O
r
O
o
CO
o
CVI
in
i
in
CO
in
i
o
o
cr>
o
fN^
55
^
A
c
o
J-
o
O
X
I'-
LL.
O
CO
|
in
in
CVI
i
CO
in
CVI
CO
0
CVI
i
CO
5-s
£
"
CO
<
^_
. o m
oo
r CO CVJ
i i i
co in in
CVJ U> O
«3- LfT
. .
Kf O
CVI r CO
1 1 1
vo o in
00 CO
1
in o
CVJ ^1-
co r in
i i i
cvi o m
CO O O
CVI r
^
. .
co ««d-
CO
1 1 r
co cr> v
f^_ r
CVI r
OT
3
*
^^ ^^^
3 jQ
la *Z. 4J
> 3 2
^J
^u^ co **
c s-
r- O 3
MO M-
(O O i
CU r- 3
1C CO
O CVJ
in r*. r-.
r J^ CVJ
1 1 1
in o i
o co
o
CO*
in o co
r-^ CO r-^
1 1 1
r O VO
<^ CVI
in
^^,
o
0 O
UD
CVI 00 CVI
1 1 1
i O O
CO O
^"
x-x
O 00
vo
i CO O
1 1 1
in o cvi
00 «3-
0
o
cu
M-
0
"3
o^ ^**s«
4-* cr**-*-
^ ^-p
». A CO
c c
CU CU w
O) C7)O
O O r-
t. s- *>^
4-> 4-> JD
r «r~ r
z z-
3-9
-------
e
r
_
c
0
"*
c=
c
CO
^
en
c
Z
NVERSIO
o
o
c
S
a
= H
oo <
iij ° "JO
"J. co Z >
-^ ^ DC ^
CJ O S UJ 2
< Z p uj co
K i A z u!l
z t = £ of-
rv ^" " * O ^ ^
v iii ^ co o 2 -C
^u "^~"' ^j ^^ o^ ^^ . .
f ^ h "^ *""" S* ^
" Jx < > S S c5
< ^
H- 3
a
^
D
D
a
<
L
en
C
_]_
.
J
a
o
a
D
a ,_,
a
Q&
o CK <
a vo
^o«F° ° £
< < ml
CO
3
63
O)
QO ^V%
*** f*V
CO
o
c
a>
i_
CD
DC
w
_ 03
1 £
^
(0 ]o
O
Z CJ
o
CO ^
DC >
LU '
> H-
z o
LO O 4_>
" c
*- a>
B -M
09 f-
0 £
a> °
n. 0
0 »-
0 ^
« M
* CM D
0 <«
CO -^
c
03
C
0)
O5
0
co *-
n »-<
2
CN
CO
CD
«M a)
u_
a
NOISH3ANOO
3-10
-------
100
g 80
x
o
60
u- 40
o
V)
cc
o
o
20
o
OMARTIN AND BERKAU
DTURNER, ETAL. ~
ATURIMER ANDSIEGMUND
OFENIMORE
HAZARD
MCCANN,ETAL. ~
APERSHING, ET AL.
FLAGAN AND APPLETON
NOTE: OIL FUEL UNLESS _
OTHERWISE INDICATED
COAL
i
0.2 0.4 0.6 0.8 1.0 1.2 1.4
NITROGEN IN FUEL BY WEIGHT, percent
1.6
1.8
2.0
Figure 3-3. Percent conversion of fuel nitrogen to NOX in laboratory scale combustion
(Reference 3-13).
3-11
-------
Figure 3-4 summarizes what may happen to fuel nitrogen during this process. In general,
nitrogen evolution parallels evolution of the total volatiles, except during the initial 10 to 15
percent volatilization in which little nitrogen is released (Reference 3-16). Both total mass
volatilized and total nitrogen volatilized increase with higher pryolysis temperature; the nitro-
gen volatilization increases more rapidly than that of the total mass. Total mass volatilized
appears to be a stronger function of coal composition than 1s total nitrogen volatilized (Reference
3-17). This supports the relatively small dependence of fuel NOX on coal composition observed in
small scale testing (Reference 3-9).
Although there is not absolute agreement on how the volatiles separate into species, it
appears that about half the total volatiles and 85 percent of the nitrogeneous species do not evolve
as permanent light gases. However, prior to oxidation, the devolatilized nitrogen may be converted
to a small number of common, reduced intermediates, such as HCN and NHL, in the fuel regions of the
flame. The existence of a set pf conmon reduced intermediates would explain the observations that
the form of the original fuel nitrogen compound does not influence its conversion to NO (e.g.,
References 3-10, 3-18). More recent experiments suggest that HCN is the predominant reduced inter-
mediate (Reference 3-19). The reduced intermediates are then either oxidized to NO, or converted
to N2 in the post-combustion zone. Although the mechanism for these conversions is not presently
known, one proposed mechanism postulates a role for NCO (Reference 3-20).
Nitrogen retained in the char may also be oxidized to NO, or reduced to N2 through hetero-
geneous reactions occurring in the post-combustion zone. However, it is clear that the conversion
of char nitrogen to NO proceeds much more slowly than the conversion of devolatilized nitrogen.
In fact, based on a combination of experimental and empirical modeling studies, it is now believed
that 60 to 80 percent of the fuel NOX results from volatile nitrogen oxidation (References 3-16 and
3-21). Conversion of the char nitrogen to NO is in general lower, by factors of two to three, than
conversion of total coal nitrogen.
Regardless of the precise mechanism of fuel NOX formation, several general trends are evi-
dent, particularly for coal combustion. As expected, fuel nitrogen conversion to NO is highly
dependent on the fuel/air ratio for the range existing in typical combustion equipment, as shown
in Figure 3-5. Oxidation of the char nitrogen is relatively insensitive to fuel/air changes, but
volatile NO formation is strongly affected by fuel/air ratio changes.
In contrast to thermal NOX, fuel NOX production is relatively insensitive to small changes
in combustion zone temperature (Reference 3-18). Char nitrogen oxidation appears to be a very
3-12
-------
OIL DROPLET
OR
kCOAL PARTICLEj
+ HEAT
VOLATILE FRACTIONS
(HYDROCARBONS, RN ETC.)
/
RN.
x x
PATH A PATH B
NO
TO FLUE GASES RELEASE ZONE
OXIDATION AT
PARTICLE SURFACE
1
NO
/\
ESCAPE FROM REDUCTION IN
BOUNDARY LAYER BOUNDARY LAYER
Figure 3-4. Possible fate of fuel nitrogen contained in coal particles
or oil droplets during combustion (Reference 3-15).
3-13
-------
X
o
100
90
80
70
60
50
40
30
20
10
0
I
WALL TEMP. 1500 °K
FLAME TEMP. 1600 °K
O LIGNITE 75-90 jum
O LIGNITE 38-45 jum
A BITUMINOUS 3845 Aim
D
12345
FUEL EQUIVALENCE RATIO (INVERSE OF STOICHIOMETRIC RATIO)
Figure 3-5. Conversion of nitrogen in coal to NOX (Reference 3-22).
3-14
-------
weak function of temperature, and although the amount of nitrogen volatiles appears to increase as
temperature increases, this is believed to be partially offset by a decrease in percentage conver-
sion. Furthermore, operating restrictions severely limit the magnitude of actual temperature
changes attainable in current systems.
As described above, fuel NO emissions are a strong function of fuel/air mixing. In general,
any change which increases the mixing between the fuel and air during coal devolatilization will
dramatically increase volatile nitrogen conversion and increase fuel NO. In contrast, char NO forma-
tion is only weakly dependent on initial mixing and therefore may represent a lower limit on the
emission level which can be achieved through burner modifications.
From the above modifications, it appears that, in principle, the best strategy for fuel NO
abatement combines low excess air firing, optimum burner design, two-stage combustion and high air
preheat. Assuming suitable stage separation, low excess air may have little'effect on fuel NO, but
it increases system efficiency. Before using LEA firing, the need to get good carbon burnout, and
low CO emissions must be considered.
Optimum burner design ensures locally fuel-rich conditions during devolatilization, which
promotes reduction of devolatilized nitrogen to N2- Two-stage combustion produces overall fuel-rich
conditions during the first 1 to 2 seconds and promotes the reduction of NO to N« through reburning
reactions. High secondary air preheat also appears desirable, because it promotes more complete
nitrogen devolatilization in the fuel-rich initial combustion stage. This leaves less char nitrogen
to be subsequently oxidized in the fuel-lean second stage. Unfortunately, it also tends to favor
thermal NO formation, and at present there is no general agreement on which effect dominates.
3.1.1.3 Summary of Process Modification Concepts
In summary of the above discussion, both thermal and fuel NOX are kinetically or aerodynami-
cally limited in that their emission rates are far below the levels which would prevail at equilib-
rium. Thus, the rate of formation of both thermal and fuel NOX is dominated by combustion condi-
tions and is amenable to suppression through combustion process modifications. Although the
mechanisms are different, both thermal and fuel NOX are promoted by rapid mixing of oxygen with the
fuel. Additionally, thermal NOX is greatly increased by long residence time at high temperature.
The modified combustion conditions and control concepts which have been tried or suggested to combat
the formation mechanisms are as follows:
Decrease primary flame zone 02 level by
- Decreased overall 02 level
- Controlled mixing of fuel and air
- Use of fuel-rich primary flame zone
3-15
-------
Decrease time of exposure at high temperature by
- Decreased peak temperature:
- Decreased adiabatic flame temperature through dilution
Decreased combustion Intensity
- Increased flame cooling
- Controlled mixing of fuel and air or use of fuel-rich primary flame zone
- Decreased primary flame zone residence time
Chemically reduce NOX in post-flame region by
- Injection of reducing agent
Table 3-4 relates these control concepts to applicable combustion process modifications and
equipment types. The process modifications are further categorized according to their role in the
control development sequence: operational adjustments, hardware modifications of existing equipment
or through factory installed controls, and, major redesigns of new equipment. The controls for de-
creased &2 are also generally effective for peak temperature reduction but have not been repeated.
The following subsections review the status of each of the applicable controls.
3.1.2 Modification of Operating Conditions
The modification techniques described in this subsection include low excess air, off stoichi-
ometric combustion, flue gas recirculation, reduced air preheat, load reduction, steam or water
injection, and ammonia injection.
3.1.2.1 Low Excess Air Combustion
Reducing the total amount of excess air supplied for combustion is an effective demonstrated
method for reducing NOX emissions from utility and industrial boilers, residential and commercial
furnaces, warm air furnaces, and process furnaces. Low excess air (LEA) firing reduces the local
flame zone concentration of oxygen, thus reducing both thermal and fuel NO formation. LEA firing
is furthermore easy to implement and increases efficiency (slight decrease in fuel consumption).
It is, therefore, used extensively in both new and retrofit applications, either singly or in com-
bination with other control measures. The ultimate level of excess air is generally limited by
the onset of smoke or carbon monoxide emissions which occurs when excess air is reduced to levels
far below the design conditions. Fouling and slagging may also increase in heavy oil- or
coal-fired applications at very low levels of excess air, thus limiting the potential of this
technique.
Low excess air firing is the most widespread NOX control technique for utility boilers. It
was initially implemented to increase thermal efficiency and reduce stack gas opacity due to acid
mist. A number of studies have shown LEA firing to be effective in reducing NO emissions without
3-16
-------
o
o
8
a:
r
O
O
O
CO
^_
jD
IO
U
JH
CX
CX
^c
>,
1
s.
o.
en
I- T-
O «/>
!-> O)
£"8
oc
c
o
O* t-
^ *^*
IO IO
S u
o v-
i- <*-
(O v
Z "O
^3
^ t/)
S*f
C
o 02
^ c
4-> 4-»
IO I/)
» 3
CU >-}
CX-O
°ox
-U ^^
^"*
,
f 3
X
C 0
0 Z
4-> f
u " .,
^
CX °"
<
4->
O CX
J- 01
4-> U
c c
o o
o o
C (/>
o c
T- O
4J i-
t/> 4->
£ c
o o
o o
1
3
o
U
O ^"^
AJ rv^*
i. U?
U-
IO
en c
0
O) T-
3 4->
f IO
U. f
^
T-
IO
i/)
(/)
o*
u
x en
o> c
lJ
_J <*-
01 c
J- 0)
3 en «/i
«/> O 4)
O i. i-
CX+J S-
X -r- 10
CU c 'r~
^3
«/> r O)
CU O) E
O 3 E CM
3 «- 01 O
am- c o
DC O - 4->
> to
JC 4->
U O)
J- O O)
i 0 E
CMCX Z O)
U 1 JC
3 JC 4->
*o en
01 n- C
0£ JC v-
a>
u
IO
S-
«l
.
t.
Ol
r"
CO
1o
i-
O)
o
Irt'cO
01 01
S- f
u
Ol CM
00
*>.
i. C
ai en
£5
3 Ol
43 O
E X
3 O
- Ol
CX-i-
o «*-
(/)
J-
Ol
c
I_
3
JQ
X
f-"i
35
§
_J
I/)
4J
c
QJ
J- 4->
ai to
i- -»
3 "O
CO IO
c
T. CSJO
Ol Z
3 01
<4- O 0
*-> c
ai cu
^^ (/) »
O O) JC X
> t. 4-> O
c
10 en i Q.
c 01 E
C i. 4->
t- 3 X
-O f- J*
0 E > W
E 3 IO Ol
IO i O
- 01 3
U. -O T3 TJ
to
O)
O
IO
c
s.
3
>
L>
ai
r-
o
CO
i-
cn-r-
C IO
"x -o
T3 i
01 01
>» 3
10 «4-
0 O
M *
C CX
en
M Ol
01 t-
o ex
c >
10 X
O i-
Z CX
l/>
ai
V)
IO
ai
u
c
t 1
0 1
JC i.
U Ol
i i- > C
a i- o 10
CJ 4-> Ol
O Of C i C
r- E i- O
O i-
V) -i- W 01 4->
Ol JC C 3 VI
U O O <4- 3
O o 5>» E
ai 4-> ai r o
Q£ tO J- IO U
v>
Ol
c
t.
3
**
V)
C9
^.
"oJ
3
en
to x
IO *
u t.
fH tO
c
1 O
X 0-r-
O jt 4J
.a 4-> i/i
Ol 3
s_ i- .a
o
s- c
Ol Cn Ol
c t- en
L. to IO
3 a* 4->
CO T3 1/1
"S
jT"
t l
k 4-*
Ol
4- w> en
i. 4J l-
01 J- 10
> 0 JC
o ex u
«1_ "S
o «/» en
IO C
3 JD t.
o -
«»*-
to ai
l~ U i-
Ol t Ol
c > c
J- S- t.
3 Ol 3
CO tO JQ
C
*^ «4-
^- CMO
01 Z
3 ai
«*- o o
4-> C
01 Ol
I »/) tO
i- Ol J3 C
4J O 10 Ol
>
O 01 JC X
=> 1- 4J 0
C CX 1
i- E Ol CX
4-> Ol
c i ai 4->
- i c
t 6 0 J£
Or- IM IO
O Ol
0 >» CX
CMS*
Ol O 10 «/>
E i E fl»
IO jt f- O
O i- 3
U- f CX-O
o
1 1
«
M ai
t- u
^01 > N Ol
Ol IO Ol Ol
J- E Er-
t> -i- IO
Ol J- r- CM
O Q.M- O
_
o
^J
Ol
c
s.
CO
4-> QC
IO O
^
pB
IO
O 4J
01 IO
O Ol
3 JC
"82
OC CX
Ol
^v
^_)
o
01
01
1 1
1 IO
y F E
oi C t/>
i. Ol i-
CXJC C
3 4J 10
««- 0
4-> O 01
u E
01 C
J- 0 X
F- i- O
O U) Z
1
u to
3 01
M- « C
O '~
I/I S.
S. * 3
Ol «/» 4J
r (J T- !-
S^E
i. -i- c
i- O
*J4J
S. U U
O Ol Ol
f- «t- «*-
ac oi oi
8
IO
en j.
<4- C Ol
^"o i E
in O CO
IO (J t/> 4->
Ol T3
t. Oli J*
U C Ol ID
C 0 t- 01
-H IM >> CX
to
CO
U
IO
C
i.
3
^~
to
J_
01
IJZ
<§
i
Decrease com
bustlon
intensity
x
4-> 1 0
IO i. JO tO
CO 3 CO U
_c to l_ -i
1
Redes igg 1
transfer :
faces, fii
aerodynam
Ol
r
il
«*-
Ol
C
8
10
en i_
"*- c 01
o 1^ o Q.
01 Or- E
to o oi
IO O 10 4->
01 -0
i- Olr- J*
U C CO (C
C O f- Ol
-H IM >, CX
I/)
Ol
u
IO
3
(/)
*
01
^7
o
CO
Ol
I«^
Increased fL
zone cooling,
reduce resi-
dence time
Ol
I_
Ol 4->
v> to
10 t.
01 0) Ol
t. it E ex
oi co i 5
C^ CX<^ 4J>
1 L-
O O
> c
ceo
0 0 T-
4J 0
C U Ol
en oi -o
'r- «/) C
10 -f-
01 01
o > m
Ol - Z
oe 4J z
c
O Ol
4J O
U M
01
o c
C O
Ol
o >-;
i- JO
C *
O C
-------
significantly increasing CO or smoke levels (References 3-14, 3-23 through 3-27). NO reductions
averaging between 16 and 20 percent are achieved on gas- and oil-fired boilers when the excess air
is reduced to levels between 2 and 7 percent. Low excess air firing below 5 percent is now standard
practice on most oil and gas utility boilers. NOX reductions of 20 percent on the average are
achieved on coal-fired utility boilers when excess air is reduced to the 20 percent level or lower.
However, the minimum excess air levels achievable with satisfactory performance are 8 to 12 percent.
In some existing units, excess air levels below 15 to 18 percent present operating problems
(Reference 3-25).
The minimum practical level of excess air which can be achieved in existing boilers, without
encountering operational problems, depends upon factors in addition to the type of fuel fired.
These factors include low load operation, nonuniformity of air/fuel ratio, fuel and air control lags
during load swings, use of upward burner tilt to increase steam superheat (for tangentially-fired
boilers), and coal quality variation and ash slagging potential (for coal-fired boilers). They tend
to increase the minimum excess air level at which the boiler can operate safely.
Other factors such as secondary air register settings and steam temperature control flexibil-
ity also affect the excess air levels. The boiler combustion control system must be modified so
that the proportioning of fuel and air is adequate under all operating conditions. Uniform distri-
bution of fuel and air to all burners is increasingly important as excess air is lowered. Excess
air levels are also affected if other NO control techniques are employed. Staging and operating
at reduced load increases the minimum excess air levels whereas switching from eastern to western
coals could decrease the levels (References 3-24, 3-28, and 3-29).
LEA firing is a very effective method for controlling NO in industrial boilers. Although
it is not in widespread use as a NO control technique for industrial boilers, LEA is generally con-
sidered as part of an energy conservation program. LEA is also a feasible NO control technique for
residential and commercial furnaces; however, the trend in NO control for these sources has been in
improved burner design in order to obtain low excess air levels without extensive CO emissions.
LEA is not a very promising technique for 1C engines and gas turbines. When the air/fuel
ratio is reduced, CO and HC emissions increase sharply for 1C engines. In gas turbines, the
overall air/fuel ratio cannot, be modified to control NO , since the ratio is determined by the
turbine inlet temperature.
In summary, changing the overall air/fuel ratio to control NO emissions is a simple, feasi-
ble, and effective technique for stationary sources of combustion, with the exception of gas turbine
3-18
-------
engines and 1C engines. For certain applications such as utility boilers, LEA firing is presently
considered a routine operating procedure and is incorporated in all new units. Since it is effi-
cient and easy to implement, LEA firing will see increasingly widespread use,in other applications.
Most sources will require additional control methods, in conjunction with LEA, to bring NO emis-
sions within statutory limits. In such cases, the extent to which excess air can be lowered will
depend upon the other control techniques employed. However, virtually all developmental programs
for advanced NOX controls are placing maximum emphasis on operation at minimum levels of excess air.
LEA will thus be an integral part of nearly all combustion modification NO controls, both current
and emerging.
3.1.2.2 Off-Stoichiometric Combustion
Off-stoichiometric combustion (OSC) is a NOX control technique in which the mixing of fuel
with the combustion air is altered so that substoichiometric conditions prevail locally in the
primary combustion zone. Complete combustion occurs downstream of the primary zone. OSC is effec-
tive for retrofit implementation on large boilers having multiple burners arranged in rectangular
matrices mounted either on one boiler wall (front-fired) or on opposite walls (horizontally
opposed-fired). This method can also be used on corner-fired boilers (tangentially fired), however
in the case of OSC with burners out of service these boilers require that all four burners on any
level be "taken out" simultaneously. Front-wall and horizontally opposed firing types are more
flexible in the location and number of burners that can be set on air only. For new units, OSC is
an attractive control technique to be included in the design of both single and multiple burner
units of all design types.
Off-stoichiometric combustion appears to be an effective technique for control of both
thermal and fuel NOX due to its ability to control the mixing of the fuel with the combustion air.
The resulting fuel-rich regions in the primary flame zone are cooled by flame radiation heat trans-
fer prior to completion of combustion with the remaining combustion air. Thus, although the overall
air/fuel mixture is near-stoichiometric, the primary N0x-forming region of the flame is operated at
a substoichiometric, low NOX condition. The NOX control effectiveness with OSC depends on burner or
primary stage stoichiometry which in turn is limited by convective section fouling, unburned
hydrocarbon emissions or poor ignition characteristics which occur at excessively rich operation.
An additional limitation of fireside corrosion may arise with the firing of some coals and heavy
oils.
In off-stoichiometric firing, the flame is long, yellow, and smokey, as opposed to the short
and intense flame observed on normal firing. Fuel combustion also extends further into the furnace,
3-19
-------
sometimes causing excessive superheater (convective section) temperatures. On some units, increased
operator vigilance is required to surmount decreased effectiveness of the flame detector system.
In practice, OSC consists of operating some burners (usually the ones located in the lower
part of the pattern) fuel-rich while the burners in the upper part of the pattern operate on pure
air. Off-stoichiometric combustion is a generic term and several modes of operation are asso-
ciated with it.
"Two-stage" combustion is based on the same principles as off-stoichiometric combustion
except that the fuel-rich burner operation is achieved by diverting a portion of the toLu. required
air through separate ports located above the burner pattern. This is also known as "overfire air/
NO port" operation and is the method used for several new multiburner designs and for use on single
burner units such as industrial boilers. Figure 3-6 shows the overfire air system on a corner
windbox of a tangentially fired boiler. So-called "simulated overfire air" operation results when
the top row of burners operate on pure air. In certain boilers, NO reduction optimization requires
A
that the burners operate either fuel- or air-rich in a staggered configuration. This is sometimes
called "biased" firing or, in the extreme where some burners are operated on air only, "burners
out of service" (BOOS).
The two-stage combustion technique is shown in Figure 3-7. A vertical cross section of a
utility boiler burner is shown schematically. Two-stage combustion of natural gas (methane) is
depicted, and a few of the global reaction mechanisms associated with the primary and secondary
combustion zones are identified.
The effect of two-stage combustion on NO emissions from three tangential coal-fired utility
boilers is shown in Figures 3-8, 3-9 and 3-10 (Reference 3-31). In these tests, NOX diminished
steadily while first stage air (burner combustion air) was decreased and routed to the overfire air
ports. Ninety percent of stoichiometric air supplied to the first stage resulted in a 58 percent
NO reduction (Figure 3-8). This reduction was obtained with overfire air ports tilted approxi-
mately 40 degrees away from the burners (Figure 3-9). NO emissions were reduced approximately 40
percent when two-stage combustion with burners out of service was applied on the same boilers
(Figure 3-10).
On existing large boilers, a load reduction will result with BOOS firing if the active fuel burners
or pulverizers do not have the capacity to carry the fuel required for full load. Most utility boilers
constructed after 1971 are, or have been, designed with overfire air ports so that all fuel burners
are active during off-stoichiometric operation.
3-20
-------
WINDBOX
SECONDARY
AIR DAMPERS
SECONDARY AIR
DAMPER DRIVE UNIT
OVER-FIRE AIR
NOZZLES
SIDE IGNITOR
NOZZLE
SECONDARY
AIR NOZZLES
COAL NOZZLES
OIL GUN
Figure 3-6. Corner windbox showing over- fire air system (Reference 3-31
3-21
-------
SECONDARY OXIDIZING
ZONE -N
CO + 0
C02 \
-C02 |
{ C + 0
\
\ /
\-- --./
+ 4H20
,' CH4 + 202>-C02 + 2H20
1 CH4 + 02>»C + 2H20
\ C + 02>-CO +0
\ 02M) + 0
^ N + 02 -^-NO + 0
\ No + 0 >-NO + N
\ L
^ PRIMARY REDUCING ZONE
( NOZZLE
FURNACE
WALL
AIR REGISTER
Figure 3-7. Two-stage combustion (Reference 3-30).
3-22
-------
O ALABAMA POWER CO.
BARRY #2
3/4 LOAD
D WISCONSIN POWER & LIGHT CO
COLUMBIA #1
FULL LOAD
A UTAH POWER & LIGHT CO
HUNTINGTON#2
FULL LOAD
100
90 100 110 120
THEORETICAL AIR-TO FUEL FIRING ZONE, percent
Figure 3-8. NOX vs. theoretical air, overfire air study (Reference 3-31).
320
300
280
260
240
220
200
180
160
140
120
NSPS
60
A
D
O ALABAMA POWER CO.
BARRY #2
D WISCONSIN POWER &
LIGHT CO.
COLUMBIA #1
A UTAH POWER & LIGHT CO.
A HUNTINGTON#2
40 20 0 20 40
TOWARD AWAY
OFA REGISTER AND FUEL NOZZLE TILT DIFFERENTIAL, degrees
Figure 3-9. NOX vs. tilt differential, overfire, air study (Reference 3-31).
3-23
-------
O ALABAMA POWER CO.
BARRY #2
D WISCONSIN POWER & LIGHT CO
COLUMBIA #1
A UTAH POWER SLIGHT CO.
HUNTINGTON#2
95 100 105 110 115 120
THEORETICAL AIR-TO-FUEL FIRING ZONE, percent
Figure 3-10. NOX vs. theoretical air, biased firing study, maximum load
(Reference 3-31).
3-24
-------
With OSC, excess air cannot be generally maintained as low as with normal firing. This is
because OSC does not achieve the intimate mixing of fuel and air that is required for low excess
air operation.
In early work with OSC, fairly significant results had been obtained for gas-fired utility
boilers by Southern California Edison Company and Pacific Gas and Electric Company, and from coal-
fired subscale combustion in tests performed by the U.S. Bureau of Mines in 1966. This modifica-
tion technique has been more thoroughly investigated during the last several years, and subsequent
sections of the present document review the recent developments for specific equipment and fuel
types.
3.1.2.3 Flue Gas Recirculation
A portion of the flue gas recycled back to the primary combustion zone reduces thermal NO
formation by acting as a thermal ballast to dilute the reactants. This reduces both the peak flame
temperature and the partial pressure of available oxygen at the burner inlet.
Some large steam boilers are designed for recirculation of.a portion of the flue gases in
order to control superheat temperatures. Normally, as boiler load decreases, steam temperatures
tend to drop unless some method of control is employed. By recirculating an increasing portion
of the flue gas as the boiler load decreases, it is possible to maintain steam temperature at a
constant level over a wider load range. .Where this type of control is used, the flue gases are
injected through the hopper bottom to reduce the effectiveness of the furnace heat absorption sur-
face without interfering with the combustion process.
It has been concluded that recirculation for steam temperature control is relatively ineffec-
tive in suppressing NO . The flue gas must enter directly into the combustion zone if it is to be
effective in lowering the flame temperature and reducing NOX formation.
A typical performance of flue gas recirculation (FGR) is shown in Figure 3-11. These results
were obtained on three similar 320 MW tangential, gas-fired utility boilers at full load. The data
show a substantial reduction in NO up to 20 percent recirculation and diminishing returns thereafter.
Similar results were obtained at reduced load operation (Reference 3-32).
Operational problems are sometimes associated with large rates of FGR. Possible flame
instability, loss of heat exchanger efficiency, and, for packaged boilers, condensation on internal
heat transfer surfaces, limit the utility of FGR on some units.
3-25
-------
DATA FROM DIFFERENT
UN ITS OF SAME TYPE
50
Figure 3-11
20 30
RECIRCULATIONr wRG/Wf + wa, percent
WHERE> = MASSFLOWRATE
RG= RECIRCULATEDGAS
f = FUEL
a = AIR
Effect of FGR on NO emissions (Reference 3-32),
3-26
-------
Although it has been concluded that FGR reduces thermal NOX, recent experience has cast
doubts on its capability to reduce fuel NOX. This method will, therefore, probably be restricted
to low-nitrogen fuels, such as natural gas, distillate oil, and low nitrogen residual oils.
Flue gas recirculation requires greater capital investment than LEA and OSC methods because
of the need for high temperature fans and ducts and large space requirements for the modifications.
However, for those boilers originally designed with FGR (for superheat control), costs of retro-
fitting are reasonable (Reference 3-30).
With moderate rates of recirculation ( 20 percent), FGR can generally be implemented without
significantly increasing emissions of CO or HC. At high rates of recirculation (30 percent), how-
ever, flame instabilities accompanied by increased CO and HC emissions can result. There is a
slight decrease in unit efficiency with FGR due to the recirculation power requirements.
3.1.2.4 Reduced Air Preheat Operation
Reducing the amount of combustion air preheat lowers the primary combustion zone peak tempera-
ture, generally lowering thermal NO production as a result. It has been used only sparingly because
of the energy penalty. It is applicable to utility steam generators and large industrial boilers
which employ heaters to impart about 280K (500F) incremental heat to combustion air. Figure
3-12 shows the NO reduction effect of reduced air preheat temperatures on 320 MW corner-fired
boiler burning natural gas. NO emissions were reduced 15 percent at full load with a 45K (80F) re-
duction in combustion air temperature (Reference 3-32).
With present boiler designs, reducing air preheat would cause significant reductions in
thermal efficiency and fuel penalties of up to 14 percent. This technique would be feasible for
thermal NO control if means other than air preheat were developed to recover heat from 423K to
A
698K (300F to 800F) gases. Reduced air preheat appears relatively ineffective in suppressing fuel
nitrogen conversion (References 3-30, 3-33).
This technique is also applicable to turbocharged internal combustion engines and regenera-
tive gas turbines. The turbocharged 1C engines have normally an intercooler to increase inlet
manifold air density permitting higher mean flowrates, and consequently higher power output. The
reduced air temperature also reduces NO emissions.
Regenerative gas turbines recover some of the thermal energy in the exhaust gas (tempera-
tures ranging from 700K (800F) to 867K (110F)) to preheat the combustion air. Any reduction in
air preheat causes severe fuel penalties unless other means of recovering the heat in the exhaust
can be implemented.
3-27
-------
400
AIR TEMPERATURE, °K
450 550
650
AIR TEMPERATURE, °F
Figure 3-12. Reduced air preheat with natural gas, 320 MW corner-
fired unit (Reference 3-32}.
3-28
-------
3.1.2.5 Load Reduction
The term "load" is defined as the percentage of the rated capacity at which the furnace or
boiler is being operated. Increasing boiler load causes an increase in primary combustion zone
volumetric heat release rate which generally increases the temperature and rate of thermal NOX
formation. Reducing boiler load, or derating, is accomplished by reducing the reactant flow
rate (fuel and oxidizer) into the furnace. Both the heat release rate (also known as combustion
intensity) and peak flame temperature are lowered.
Apart from the obvious drawback of limiting boiler capacity, load reduction can lead to
operational problems. Higher levels of excess air are typically required to suppress CO or smoke
emissions thus leading to an overall reduction in efficiency. The increased residence time of the
combustion gases at the reduced load can cause steam temperature imbalance in the convective sec-
tion. Higher excess air or flue gas recirculation may be needed to maintain superheat temperatures.
Also, operation at greatly reduced load may exceed the practical turndown limit of the burners.
Some burners may need to be taken out of service to maintain good firebox mixing and steam tempera-
ture control.
Most of the above problems can be avoided when the unit is designed to operate at low com-
bustion intensity. Here, the use of enlarged fireboxes on new units produces NOX reductions simi-
lar to load reduction on existing units. Some of the last gas- and oil-fired utility boilers sold
were equipped with enlarged fireboxes. New coal-fired utility boilers use fireboxes typically 30
percent larger than was the practice in the 1960's (Reference 3-34). This practice is partly in
response to the New Source Performance Standards set in 1971 and partly to facilitate combustion
of lower grade western coals. With coal-firing, the NOX reduction due to an enlarged firebox is
largely indirect through the change in firebox aerodynamics.
As mentioned in Section 3.1.2.2 the retrofit of off-stoichiometric combustion to achieve
significant NO reductions often requires derating of the boiler. Derating becomes necessary when
the desired first stage burner stoichiometry cannot be obtained with the number of burners out of
service (BOOS) at full load conditions. The reduced load, thus, permits additional burners
out of service and consequently lowers first stage stoichiometries. Load reduction is therefore
also effective in reducing fuel NO when this technique is implemented with staged combustion.
3-29
-------
3.1.2.6 Steam and Water Injection
Flame temperature, as discussed above, is one of the important parameters affecting the
production of thermal NO . There are a number of possible ways to decrease flame temperature via
thermal means. For instance, steam or water injection, in quantities sufficient to lower flame
temperature to the required extent, may offer a control solution. Water injection has been found
to be very effective in suppressing NO emissions from internal combustion engines and gas turbines.
Figure 3-13 shows NOX emission reductions from a gas turbine as high as 80 percent (Reference 3-35).
Since steam and water injection reduce NO by acting as a thermal ballast, it is important
that the ballast reach the primary flame zone. Combustion equipment manufacturers vary in their
methods of water or steam introduction. The ballast may be injected into the fuel, combustion air,
or directly into the combustion chamber.
Water injection may be preferred over steam in many cases, due not only to its availability
and lower cost, but also to its potentially greater thermal effect. In gas- or coal-fired boilers,
equipped for standby oil firing with steam atomization, the atomizer offers a simple means for in-
jection. Other installations will require special rigging so that a developmental program may be
necessary to determine the degree of atomization and mixing with the flame required, the optimum
point of injection and the quantities of water or steam necessary to achieve the desired effect.
The use of water injection may entail some undesirable operating conditions, such as de-
creased thermal efficiency due to the high heat capacity of water compared with that of flue gas
or other inert diluents, and increased equipment corrosion. This technique has the greatest
operating costs of all combustion modification schemes with a fuel and efficiency penalty typically
of about 10 percent for utility boilers and about 1 percent for gas turbines. It is, therefore,
an unpopular NO reduction technique for all combustion equipment except for stationary gas tur-
bines (References 3-30 and 3-33) which, in addition to the lowest reduction in efficiency, showed
no major operational problems or reduced equipment life. Water injection for NO reduction does
not appear to have a significant effect on stack opacity and emissions of CO and HC.
3.1.2.7 Ammonia Injection
The post-flame decomposition of NO by reducing agents has recently shown promise as a
method for augmenting combustion modifications if stringent emission limits are to be met. Exxon
has patented a process for the homogeneous gas phase selective decomposition of NO by ammonia
(Reference 3-36). The gas phase reaction in the temperature range of 978K (1,300F) to 1,368K
3-30
-------
0.4 0.8 1.2 1.6
WATER INJECTED TO COMBUSTION AIR, percent
2.0
Figure 3-13. Correlation of NOX emissions with water injection rate for natural
gas fired gas turbine (Houston L&P Wharton No. 43 unit) (Reference 3-35).
3-31
-------
(2.000F) converts nitric oxide, in the presence of oxygen and ammonia, into nitrogen and water
(Reference 3-37).
Results of lab scale tests show that the level of NOX reduction depends on the combustion
product temperature, initial NOX concentration, and quantity of ammonia injected (Reference 3-38).
Based on the available results, ammonia injection appears to be most effective between 978K
(1.300F) and 1.368K (2.000F), which corresponds to conditions in the convective section of large
boilers. Maximum NO reductions, as much as 90 percent, were obtained at 1.233K (1.750F) with
molar ratios of ammonia to initial nitric oxide ranging from 1.0 to 1.5.
Field tests were conducted on a gas-fired furnace rated at 147 MW (500 x 106 Btu/hr) and on an
oil-fired boiler rated at 41 MW (140 x 106 Btu/hr) with both the units retrofitted for NH, injection
O
(Reference 3-37). A reduction in NOX of nearly 70 percent was obtained with a NH3/initial NOX ratio
of 4.5. Table 3-5 summarizes the available test results to date.
Although ammonia injection is a promising technique, there are a number of developmental
questions which must be answered before its full potential can be assessed. The first is the
applicability of ammonia injection to existing utility boilers and to systems other than steam
generators. Ammonia injection appears to have potential for new utility boilers and large indus-
trial boilers since the required temperature range is compatible with current convective section
design. New units could conceivably be*designed to include ammonia injection cavities in the
convective sections. Applications to existing units may be limited by the absence of the precise
residence time-temperature conditions required for the process. Additionally, ammonia injection
seems to be limited for other equipment types such as gas turbines and 1C engines because the re-
quired time-temperature constraint cannot practically be met.
The second question concerns the ability to maintain adequate convective section temperatures
required for selective reduction during boiler load changes. Normally, during load reduction, the
convective section temperature will reduce substantially below the base load level. The tempera-
ture excursions during load reduction could easily move out of the range where ammonia injection is
effective. Load following capability may thus be a limitation on ammonia injection for nonbase-
loaded units.
The third question concerns the effectiveness and environmental impact of the process, par-
ticularly with coal firing. The process has been demonstrated on oil- and gas-fired units but is
just starting to be studied in coal-fired pilot scale units. Environmental concerns with ammonia
injection include the presence of ammonia as a primary pollutant in the stack gas and potential
3-32
-------
«^^
in
4-> ^
c 1-
cu ic
E cu
o ce:
o
«4- X
O O
0>r- IT
C ro CL
(O C CL
U_
C
X O
O t-
O 'O
r~ L» *'"*
E
cn-r- c: CL
C 4-> Q.
o» c CL
en o»
c o
10 C
rv o
o
3 '*"'
*~ a2
p-oT"'
i IO
^ o:
^. c
4J O
fl) 1 *
E CL
^ w
3 O
CT V)
1 1 Q)
O
CO X
i. 4J O
3 to -C Z 1/1
0 Q 4J C
o E - O
XO X IO v
O « >- 4->
Z C ^ 0> 4-> (O
0 0 > -r- S-
E i- CO -r- C 4->
§4-> CVJ 4-> - C
O*O C
10 ^: o
O
O
0
O
O
O
A
7
0
o
1
o
uo
CVJ
LO
1
o
o
00
1
o
CVI
IT
.c
^ s.
i 3 ^:
4J > CO 4J i
C I- -i- -i-
gj o « o» o
E 4J O
r- to 3 i C7> 1-
L. 3-^ CO
0> ^J X -r-
Q. £LO S_ Irt
X O 00 CVI »r- IO
LU (J LT> ~ '"4- O)
^-^ C
U> 1 C O C
i. ^ 0 i- 0
3 4J -r- +J i<£ -i-
O f- 4J O O 4J
O $ O 3 00 -i-
X O **"^ fl^ ^O O^ "O
O T-> 0» TJ
z c ^ c j- 4-> re
O 0 -r- «0
§-r- If) E CVI 1
4-> CVJ CVJ 3 to 1C
E o »DC E 1-
r- 3 i i- 3 -C
X ^O 4-> X (J 4->
3 IO O -r-
^ i- (o o s: o x
c
o o a« o
o LO in -
r CVJ VO 4->
1 1 10
O O O 3
i i CO TD
0)
s-
0 0
Lf> Lf> 1
cvj cvj
^.
_o
r~
O 0 II
0 O
*5t" *^ ^"~
10
^
CO -i-
n: c
Z «r-
z
0 0
O 0
co co
T T i
0 O
r^» r*1*
CO 00
J. *b
O r-
r- 0 X
IO IO
4-> O) O
C i- T3 «*
f- «0 -""""iT
l-r- M- 3 J=
0> 3 1 3C ^^
CL^S r- 3
X 3 f- ->
Ul 4-> O * CO
C
*e o
0 -r-
00 4->
1 O
0 3
CO T3
O)
,
LO
O
^~
II
p*
*o
r~
CO -
Z !-
Z
IT
.c
^>^
3
4-*
CO
o t
OJ r
I-
-------
reactions of ammonia with the fly ash and sulfur compounds in coal firing. Since low temperature
stack gas reactions are important here, pilot scale tests will be of limited use.' Full quantifica-
tion of potential adverse impacts of ammonia injection will await full scale demonstrations with
coal firing.
In addition to the above operational concern, there is also the strategic question of whether
sufficient ammonia would be available in the 1980's and 1990's for widespread application in utility
boilers (Reference 3-39).
In summary, ammonia injection has near-term application for NO control in gas- and oil-fired
boilers; also, it shows promise for far-term applications to coal-fired boilers.
3.1.2.8 Combinations of Techniques
Since 1969, it has been demonstrated that several of the previously discussed modification
techniques can be effectively utilized in combination since they reduce NO by different mechanisms.
Most often, off stoichiometric combustion is used with low excess air, or load reduction on all
fuel-boiler type configurations. For oil and gas fired units flue gas recirculation is used in
conjunction with the above techniques. Flue gas recirculation and load reduction lower peak
combustion temperatures, while off-stoichiometric operation reduces the amount of fuel burned at
peak temperature. For the most part, combining control techniques has been shown to be comple-
mentary but not additive for NOX reduction (Reference 3-30).
3.1.3 Equipment Design Modification
3.1.3.1 Burner Configuration
Burner or combustor modification for NOX control is applicable to all stationary combustion
equipment categories. The specific design and configuration of a burner has an important bearing
on the amount of N0x formed. Certain design types have been found to give greater emissions than
others. For example, the spud-type gas burner appears to give a higher emission rate than the
radial spud type, which, in turn, produces more NO than the ring type.
During the early 1970's specially designed "low-NOx" burners were produced for thermal NO
control. For the most part, they are designed for utility and industrial boilers and employ in-
flame LEA, OSC, or FGR principles. The aim is to strike a balance between minimum NO formation
and acceptable combustion of carbon and hydrogen in the fuel.
There are currently several commercial low-NO gas and oil burner designs in operation and
development (References 3-40 through 3-44). Full scale test results in Japan show reduction in
x
3-34
-------
No emissions from 40 to 60 percent with low-NO gas burners. Sub-scale tests with single burners of
the type normally used in utility boilers have indicated that simple changes in burner block and
nozzle geometry and in swirl vane angles can decrease NO production by up to 55 percent (References
3-41 and 3-45). Some of the more innovative methods for oil burners include: flame splitting
distributor tips which cause a flower petal flame arrangement, and atomizers with fuel injection
holes of different diameters which create fuel-rich and fuel-lean combustion zones (References 3-15,
3-40, 3-43). Up to 55 percent reductions in NO emissions are reported with the use of these nozzle
tips. However, the change in flame shape may cause problems due to impingement on walls and effectiveness
may be reduced as flames interact in multiburner furnaces.
Other air-fuel modifications include a low-NO burner (offered by at least one company in
the U.S.) for oil- and gas-fired package boilers. This burner uses shaped fuel injection ports
and controlled air-fuel mixing to create a thin stubby ring-shaped flame (References 3-40, 3-42).
With this modification, reductions in NOX from 20 to 50 percent are claimed. The most extensive
air-fuel modifications involve the self-recirculating and staged combustion chamber type of
burners, used in industrial process furnaces. These burners are equipped with a prevaporization
or a precombustion chamber in the windbox. In the chamber, the fuel is vaporized and premixed with
part of the combustion air, or is allowed to undergo partial combustion under oxygen deficient
conditions before being discharged into the furnace. NO reductions of about 55 percent are
typical for these devices.
Similar reductions are being demonstrated on prototype coal-fired units. One major utility
boiler manufacturer has recently fabricated and tested a dual register pulverized coal burner, de-
signed to produce a limited turbulence, controlled diffusion flame. The manufacturer claims NO
reductions of 50 percent (Reference 3-46). Figure 3-14 shows field test results on three existing
boilers equipped with the new low-NO burners (Reference 3-47). Emissions were 55 percent less NO
* x
than identical units operating under similar conditions with old circular burners (Reference 3-26).
The new low-NOx burners are designed to attain controlled mixing of fuel and air in a pattern
that keeps the flame temperature down and dissipates the heat quickly. Burners can be designed to
control flame shape for minimizing the reaction at peak temperature between nitrogen and oxygen.
Other designs internally recirculate part of the combustion gases or have fuel-rich and fuel-lean
regions within a burner to reduce flame temperature and oxygen availability.
3-35
-------
1.0
0.9
0.8
0.7
"3 0.6
00
CO
5 0.5
CM
o
^ 0.4
X
o
2 0.3
0.2
0.1
0
CIRCULAR BURNER
DUAL REGISTER BURNER
90
250
UNIT CAPACITY, MW
EPA NOX
EMISSION LIMIT-
400
300
200 o
X
o
100
700
Figure 3-14. Comparison of NOX emissions with pulverized coal firing, circular burner vs dual
register burner (Reference 3-47).
3-36
-------
Burner design modifications have the major advantages of not requiring redesign of boilers
or combustion chambers, not necessitating load reduction, and possible applicability to many types
of boilers. The disadvantages are that some burners may have to be custom designed for specific
fuels and that some burner designs optimized for low-NO may only be applicable to a limited number
of boiler configurations. However, improved burner design may in the early 1980's be used in con-
junction with some already proven external combustion modification such as OSC. This combination
of low-NOx burners and OSC may lead to significantly lower N0x emissions. It is also possible that
with more advanced burner designs currently under development, the external combustion modifica-
tions might be entirely replaced with the low-NOx burners (References 3-33, 3-48).
3.1.3.2 Burner Spacing
The interaction between closely spaced burners, especially in the center of a multiple-
burner installation, increases flame temperature at these locations. There is a tendency toward
greater NOX emissions with tighter spacing and a decreased ability to radiate to cooling surfaces.
This effect is illustrated by the higher N0x emissions from larger boilers with greater multiples
of burners and tigher spacing. During a field test program conducted by KVB Inc. two 215 MW units
were tested for NOX reduction by combustion modification. These two units are identical in design
except for burner spacing. At reduced load operation the closeness of burner spacing for one of
the units resulted in higher NOX levels by as much as 25 percent (Reference 3-32).
In most new utility boiler designs, vertical and horizontal burner spacing has been widened
to provide more cooling of the burner zone area. In addition, the furnace enclosures are built to
allow sufficient time for complete fuel combustion from slower and more controlled heat release
rates, such as that associated with the off-stoichiometric operating mode. Furthermore, furnace
plan areas have been increased to allow for larger heat transfer to the cooling walls. This in-
crease in the burner zone dimensions creates more wall area thus increasing the distance between
evenly spaced burners.
Horizontal burner spacing is largest for tangentially fired boilers with the burners
being located at each corner of the furnace. Flames in a corner-fired unit interact only at the
center of the furnace in the well known spiral configuration. As a result the flames radiate
widely to the surrounding cooling surfaces before interacting with one another. Also, the tangen-
tial firing configuration results in slow mixing of fuel with the combustion air. For these
reasons, tangentially-fired boilers show baseline, uncontrolled emissions below those for other
utility boilers firing configurations. It has been observed, however, that for many tangentially-
3-37
-------
fired boilers, the NOX response to operating modifications has been less impressive than from
boilers of other designs, even though the magnitude of the initial, uncontrolled emission level
was lower (References 3-27, 3-30).
3.1.4 Fuel Modification
Another alternative for controlling NOX through process alteration is through modification
of the fuel. Three candidate techniques are fuel switching, fuel additives, and fuel denitrification.
3.1.4.1 Fuel Switching
This method usually entails the conversion of the combustion system to the use of a fuel
with a reduced nitrogen content (to suppress fuel NOX) or to one that burns at a lower temperature
(to reduce thermal NOX). Sulfur control is usually a dominant cost incentive for fuel switching.
Natural gas firing is an attractive NO control strategy because of the absence of fuel NO in
x x
addition to the flexibility it provides for the implementation of combustion modification tech-
niques. Despite the superior cost-effectiveness of gas-fired NOX control, the economic considera-
tions in fuel selection are dominated by the current clean fuel shortage. Indeed, the trend is
toward the use of coal for electric power generation and larger industrial processes. Fuel
switching to natural gas or distillate oil is not a promising option for widespread implementation
(Reference 3-49).
Western coals constitute one abundant alternate source of potentially low-NOx fuels. The
direct combustion of western subbituminous coals in large steam generators generally produces lower
NOX emissions than with combustion of eastern bituminous coals. Three mechanisms are responsible
for lower NOX emissions: first, western coals in general contain less bound nitrogen than eastern
coals on a unit heating value basis; second, the excess 02 in a steam generator burning western
coal can be maintained at very low levels; and third, the high moisture content of western coal
produces lower flame temperatures.
The NO emissions for a 59Mg (130,000 Ib) steam/hr industrial boiler firing pulverized western
coal at baseline conditions were 24 percent lower than for eastern coals (Reference 3-29); NO,
emissions remained unchanged when firing western coal in stokers. However, the slope of the NO
vs. excess 02 curve for a water-cooled vibrating grate stoker firing western coal (Wyoming Bighorn)
was 12 ppm/percent excess 02, compared to 35 ppm/percent excess 02 for eastern coal (Kentucky
Vogue).
3-38
-------
Some specific problems associated with burning low sulfur, high moisture content coals in
combustion equipment designed for higher quality coals are listed below (Reference 3-50):
Poor ignition
0 Reduced boiler load capacity
Increased carbon loss
Boiler fouling
t High superheat steam temperature
Flame instability
t Increased boiler maintenance
t Reduced boiler efficiency
Reduced collection efficiency of electrostatic precipitator (ESP)
However, most of these operational problems can be solved with current boilers specifically
designed to burn these lower grade coals.
Formerly, a major incentive for switching to western coals was the low sulfur content of
these fuels. Economic conditions made fuel switching from high sulfur eastern bituminous coals to
low sulfur western subbituminous coal competitive with the cost of gas scrubbing for SCU removal.
Therefore, low sulfur, low nitrogen, western coals represented a promising short-range option in
fuel switching for large industrial and utility boilers. However, the 1977 Clean Air Act requires
that NSPS be based on a percentage reduction in the pollutant emissions which would have resulted
from the use of fuels which are not subject to treatment prior to combustion. This deemphasizes
fuel switching.
A promising long-range option is the use of clean synthetic fuels derived from coal. Candi-
date fuels include low to high Btu gas (3.7 to 30 MJ/Nm3, or 100 to 800 Btu/scf) and synthetic
liquids and solids. Process and economic evaluations of the use of these fuels for power genera-
tion are being performed by, among others, the EPA, DOE, the American Gas Association, and the
Electric Power Research Institute (EPRI). Two alternatives for utilizing low- and intermediate-Btu
O
gases (up to 26 MJ/m , or 700 Btu/scf) are firing in a conventional boiler or in a combined gas and
steam turbine power generation cycle. For both systems, economic considerations favor placement of
both the gasifier and the power cycles at the coal minehead. The most extensive use of these systems
would probably be for replacement of older conventional units upon their retirement (Reference 3-51).
The NOX emissions from lower-Btu gas-fired units are expected to be low due to reduced flame
temperatures corresponding to the lower heating value of the fuel. The effects on NO formation of
3-39
-------
the molecular nitrogen and the intermediate fuel nitrogen compounds, such as ammonia, in the lower-
Btu gas have not yet been fully determined and require further study.
The synthetic fuel oils or solid solvent refined coal (SRC) may be expedted to be high in
fuel nitrogen content even though some denitrification may occur in the desulfurization process.
This high nitrogen, carried over from the parent coal, would promote high NO emissions. Other
potential alternate fuels that might be considered and their potential for fuel or thermal NO
are listed in Table 3-6.
TABLE 3-6. NOX FORMATION POTENTIAL OF
SOME ALTERNATE FUELS
FUEL
Shale oil '
Coal -oil mixture
Methanol
Water-oil emulsion
Hydrogen
THERMAL NOX
Low
Low
Low
Low
High
FUEL NOX
High
Moderate
Low
Unchanged
Low
Shale oil ranks second to coal as the most abundant source of nonpetroleum fossil fuel in the
United States (Reference 3-52). Even though there are six proven recovery processes, current shale
oil production is very limited. The combustion of shale oil will cause higher levels of fuel NO
because this fuel generally contains bound nitrogen in excess of 2 percent. Distillation of shale
oil would reduce fuel nitrogen content, however.
Coal-oil mixtures have recently become of interest as an alternate fuel which could stretch
the domestic oil supplies and reduce our dependence on foreign oil. NO from combustion of this
fuel will depend on the quantity of nitrogen present in the coal and oil and the percentages of
coal and oil used to make the mixture. However, NO emissions are expected to be lower than
emissions obtained from combustion of coal only.
Methanol is currently produced from the synthesis of methane from natural gas. However, due
to the shortage of natural gas, future production will have to come from synthetic gas generated
from coal and biomass. Baseline NOX emissions from the combustion of methanol in an experimental
3-40
-------
hot wall furnace system were reported at 50 to 70 ppm, compared to 240 to 300 ppm for distillate
oil. With flue gas recirculation, the NOX emissions from methanol combustion were reduced to 10
ppm or 15 percent of the baseline level (Reference 3-53).
In gas turbines 74 percent less NOX was produced using methanol, compared to distillate oil.
The hot wall experimental furnace showed a 20 percent increase in stack heat loss (SHL), compared
to SHL of 14 percent for distillate oil (based on 115 percent theoretical air at a 473K (390F) stack
temperature). For natural gas, turbine efficiency levels increased by 6 percent due to higher inlet
temperatures.
Since water-oil emulsions affect only thermal NOX these alternate fuels have a definite NOX
reduction potential when distillate oil is used (Reference 3-54). NOX emission levels from emul-
sions with approximately 50 mass percent water in distillate oil approached the levels obtained
from methanol combustion (Reference 3-55).
Hydrogen as a fuel is used in high energy production concepts such as rocket engines. The
high levels of thermal energy released make this fuel attractive for other energy conversion sys-
tems. Thermal NOX levels are, however, high when hydrogen meets with oxygen in the presence of
atmospheric nitrogen. Water represents an abundant supply of hydrogen with the use of electrolysis.
Low NO application of these fuels may require development of additional control technolo-
A
gies. Many of the current combustion control strategies may, however, be applicable, especially
for the case of thermal NOX<
The feasibility of synthetic fuel firing as a NOX control option is contingent on the cost
tradeoff between synthetic fuel production and the total control costs for NOX, SOX, and particu-
lates in conventional coal firing. In the case of coal derived fuels, there is preliminary
evidence that gasification may be more costly than flue gas cleaning of conventional systems.
3.1.4.2 Fuel Additives
For purposes of this document, a fuel additive is a substance added to any fuel to inhibit
formation of NOX when the fuel is burned. The additive can be liquid, solid, or gas. For liquid
fuels, the additive should preferably be a liquid soluble in all proportions in the fuel, and it
should be effective in very small concentrations. The additive should not in itself create an air
pollution hazard nor be otherwise deleterious to equipment and surroundings.
In 1971, Martin, et al_., tested 206 fuel additives in an oil-fired experimental furnace,
and four additives in an oil-fired packaged boiler. None of the additives tested reduced NO
3-41
-------
In general, NO control in FBC is a matter of good management of the normal process
variables. If more stringent standards are enacted, conventional NO controls, such as flue gas
recirculation and off-stoichiometric combustion, may be used. Exploratory results indicate that
two-stage combustion could be advantageous for;both NOX and SOX control. The flexibility of the
FBC process is a technical and cost advantage to the implementation of new control techniques.
From a NOX control standpoint, fluidized bed combustion appears to be competitive with
control of conventional combustion methods. At the present time, however, FBC must be viewed as
a medium risk concept. The economics of the basic process have not yet been fully established
relative to conventional boilers or low-Btu gas combined cycle units. Also, the versatility of
the FBC concept to a wide variety of equipment applications needs to be shown.
3.1.5.2 Catalytic Combustion
Catalytic combustion refers to combustion occurring in close proximity to a solid surface
which has a special (catalytic) coating. A catalyst accelerates the rate of a chemical reaction,
so that substantial rates of burning should be achieved at low temperatures, avoiding the forma-
tion of NOX. Moreover, the catalyst itself serves to sustain the overall combustion process,
thereby minimizing the stability problems (References 3-64 and 3-65). However, the overall success
of a catalytic combustion system in reducing CO and UHC to low levels is a function of both hetero-
geneous and gas phase reactions; surface reactions alone appear to be unable to achieve the desired
low levels.
Emissions from catalytic combustion experiments have typically been: NO < 2 ppm, UHC - 4 ppm,
and CO = 10 to 30 ppm. Both gaseous and distillate fuels have been used and combustion efficiencies
above 95 percent have been obtained (Reference 3-65).
The catalyst bed temperature must be held below 1,81 IK (2.800F) to minimize the formation of
NOX. At high temperatures, above 1.273K (1,830F), catalyst degradation can be significant. Excess
air can be used to lower the bed temperature; but except for gas turbines excess air is unattractive
since it also reduces thermal efficiency. Further research is underway to consider other systems,
such as catalyst bed cooling, exhaust gas recirculation and staged combustion to maintain a low bed
temperature.
Recent tests evaluated the applicability of catalytic combustors for gas turbines. Test
fuels used were No. 2 distillate oil and low Btu synthetic coal gas, for a range of pressure,
temperature, and mass flow conditions. Test results show that the catalyst bed temperature profile
at the bed exit was very uniform for low Btu gas, but not as uniform for No. 2 oil. Exceptionally
3-44
-------
A 30 MW AFBC pilot plant began operation in late 1976 (Reference 3-61). Pressurized systems
are still being tested, with a pilot plant planned for the early 1980's. Results of recent work in
FBC, the status of FBC development, and EPA, ERDA and EPRI FBC programs can be found in Reference
3-61.
Suggested advantages for fluidized bed combustion compared to conventional boilers are: (1)
compact size yielding low capital cost, modular Construction, factory assembly and low heat transfer
area, (2) higher thermal efficiency yielding lower thermal pollution, (3) lower combustion tempera-
ture resulting in less fouling and corrosion and reduced NO formation, (4) potentially efficient
sulfur oxides control by direct contact of coal with an SC^ acceptor, (5) fuel versatility, (6)
applicable to a wide range of low-grade fuels including char from synthetic fuels processes, and
(7) adaptable to a high efficiency gas-steam turbine combined power generation cycle. The general
validity of these suggested advantages have yet to be demonstrated in field application. The
principle disadvantages of FBC are: (1) potential large amounts of solid waste (the sulfur acceptor
material) and (2) heavy particulate loading in the flue gas.
The feasibility of the FBC for power generation and utility boilers depends in part on the
following: (1) development of efficient methods for regeneration and recycling of the dolomite/
limestone materials used for sulfur absorption and removal, (2) obtaining complete combustion
through fly ash recycle or an effective carbon burnup cell, (3) development of a hot-gas particulate
removal process to permit use of the combustion products in a combined-cycle gas turbine without
excessive blade erosion.
Oxides of nitrogen emissions from fluidized bed combustors have been shown to be predominately
fuel-derived. Seven to ten percent of fuel nitrogen is converted to NO (References 3-62 and 3-63).
Experiments with nitrogen-free fuels resulted in NOX concentrations in agreement with equilibrium
values at the bed temperature. However, coal-fired experiments resulted in NO concentrations in
excess of the equilibrium values. Furthermore, experiments using nitrogen-free gases with coal
yield substantially similar NOX levels as combustion in air (Reference 3-62).
NOX emissions have been found to be slightly dependent on coal particle size, the type and
amount of sulfur acceptor, the amount of excess air and the design of the combustor itself. Emis-
sion levels from pressurized fluidized bed combustors are significantly less than from atmospheric
combustors. This is probably a result of greatly increased NO decomposition rates at elevated
pressures. Even at 100 percent excess air, NOX emissions from a PFBC are well below the current
standards of 300 ng N02/J (0.7 lb/108 Btu). Results of 160 ng/J (0.37 lb/106 Btu) have been
reported (Reference 3-61).
3-43
-------
References 3-67 and 3-68 describe in detail the application of repowering to boiler, gas
turbine, and steam generating plants; savings in capital and operating costs are anticipated.
Repowering of two steam turbine units in the City of Glendale, California increased power output
by 75 MW and reduced power cost to the consumer by 8 percent (Reference 3-69). Under
contract from the Electric Power Research Institute, Westinghouse Electric Corporation is evaluat-
ing repowering conventional steam power plants without replacing the boiler. Earlier pilot scale
work for EPRI by KVB Inc. shows a low NO potential for repowering. The boiler is fired fuel-rich
using approximately 85% of the NO bearing gas turbine exhaust as the combustion air. The remain-
ing gas turbine exhaust provides the boiler second stage air which is injected through overfire
air ports above the fuel-rich primary stage. Up to 55% of the NO in the gas turbine exhaust is
chemically reduced by the fuel rich primary stage of the boiler. Also, the use of overfire air
reduces the NO formed in the boiler by up to 50%. The present use of repowering is very limited.
It may see extensive use in the 1980's if significant increases in generating capacity are needed.
3.1.5.4 Combined Cycles
Combined cycles may, in the long term, reduce emissions of sulfur oxide, nitrogen oxide,
particulate matter, and waste heat while generating power at efficiencies higher than conventional
fossil- and nuclear-fueled steam stations (Reference 3-70).
The combined gas and steam turbine system consists of a gas turbine using a coal-derived fuel,
which exhausts into an unfired waste-heat-recovery boiler. At the gas turbine inlet, the most
economical large scale steam system would operate at 16.6 MPa (2,400 psig) with 811K (l.OOOF)
throttle steam and 81 IK reheat temperatures. In this system, roughly 66 percent of the power
would be generated by the gas turbine; the remaining 34 percent would be generated by the steam
boiler system (Reference 3-71).
Combined cycle efficiency improves significantly as the gas turbine inlet temperature is
increased. At turbine inlet temperatures of 1,478K (2.200F), an efficiency improvement of 2 per-
centage points per 55K (100F) increase in turbine inlet temperature is found.
The current status of combined cycles has been reviewed by Papamarcos (Reference 3-72) who
concludes that, before combined cycles are commercialized, efficient fuel conversion processes and
high temperature gas turbines that can use coal-derived fuels must be developed. He estimates that
these developments will take place in some 15 to 20 years, and current DOE projections concur with
his estimate.
3-46
-------
low emissions (2 to 3 ppm NO , 20 to 30 ppm CO) were achieved for both fuels, and unburned hydro-
carbons (UHC) were less than 1 ppm (Reference 3-66). However, much additional work is needed
before catalytic combustion can be applied to gas turbines in the field.
Catalytic combustion has been demonstrated to be effective in removing pollutants such as
N0x, CO, and UHC, but at present, catalytic combustors are limited by the catalyst bed temperature
capability. Various government agencies and private industries are developing catalysts that will
withstand high temperatures, retain high catalyst activity, and last longer. Catalytic combustion
systems are also under development; it appears that during the next 5 to 10 years, catalytic com-
bustion concepts may be incorporated into new gas turbine and residential, commercial, and indus-
trial heating designs.
3.1.5.3 Repowering
Repowering adds a combustion turbine to an existing steam plant, providing additional capa-
city at lower initial costs and lower energy costs than other spare capacities available to a
utility.
Repowering includes: (1) steam turbine repowering, in which gas turbines and new heat re-
covery boilers are added to an existing steam electric generating plant; (2) boiler repowering in
which gas turbines are added to the existing steam generating facilities for power generation, re-
quiring the conversion of existing conventional boilers to heat recovery type boilers; and (3) gas
turbine repowering in which a steam generating plant is added to an existing gas turbine plant
(References 3-67 and 3-68).
Depending on the system and power needs, repowering of existing facilities offers the
following advantages:
There is no need to acquire and develop a new plant site
Repowering generally requires smaller increments of investment, saving on fixed charges
since major investment on new plants is deferred
Repowering improves heat rate, which lowers fuel consumption
0 The environmental impact is reduced, with improving schedules for environmental and site
related approvals
For boiler and steam turbine repowering, there is no increase in cooling water requirements
0 Gas turbines may be operated independently as peaking units, which provides greater plant
flexibility
3-45
-------
Each process developer utilizes different catalysts, catalyst supports and bed configura-
tions. Also, differing applications require substantially different catalysts^and operating condi-
tions depending upon the S02 content and dust loading of the specific flue gas. The inlet NOX
concentrations being treated in Japan range from 150 ppm to 250 ppm with NOX exit concentrations
of 10 ppm to 50 ppm (Reference 3-74).
Compared to the wet process, the dry process is simple, requires less space, generates no
troublesome byproducts and requires no tail gas reheating. However, the dry process has yet to
prove itself on a dirty gas stream of commercial scale. If the sulfur and particulate laden flue
gas is scrubbed to remove S02 before FGT, it will have to be reheated from about 31 IK to 643K
(100F to 700F) (Reference 3-74). If the S02 is not removed from the flue gas excess ammonia may
combine with S03/S02 and cause a visible plume. This byproduct is also corrosive to mild steel.
Large amounts of ammonia may be required which will cause an increased consumption of natural gas
currently used to produce the ammonia. Ammonia requirements are proportional to the quantity of
NO removed. Thus, combustion modifications are likely to be used to reduce NO levels as much as
x *
possible before treatment in FGT systems.
A selective noble metal catalyst process using ammonia was recently explored on a pilot scale
by an EPA contractor. The pilot plant, using natural gas, accumulated about 2000 hours of testing
and achieved NO reductions of 90 percent with essentially no catalyst degradation. Further tests
have been conducted using fuel oil and/or sulfur-containing flue gases. These tests indicate that
platinum is not satisfactory for flue gases containing S02 (Reference 3-74).
Another study for EPA has been conducted for the technical and economic assessment of various
catalytic processes for NO control (Reference 3-75). Lab scale tests on simulated flue gas inves-
tigated several operating variables and catalysts. The major emphasis was on selective reduction
of NOX with ammonia using nonnoble metal catalyst systems. These parametric studies showed NO
reductions of 60 to 85 percent at inlet concentrations of 250 to 1000 ppm.
Although selective catalytic reduction has been the most widely used dry process, selective
noncatalytic NO reduction is under investigation in both the U.S. and Japan. The technique in-
volves the homogeneous decomposition of N0x by injecting a gaseous reducing agent into the post-
flame region. Ammonia is the most common reducing agent, although the two U.S. research firms in-
vestigating this concept have considered proprietary agents as well. Injected ammonia is most
effective for NO reduction when the combustion products are between 1.200K and 1.311K (1,700F and
1.900F). The concept has been demonstrated commercially on a 41 MW (140 x l^Btu/hr) oil-fired
3-48
-------
3.2 COMBUSTION FLUE GAS TREATMENT
Combustion modification is a demonstrated and'effective method for achieving reduction of
NOX from stationary sources. It is, however, somewhat limited both in emission reduction efficiency
and range of applicability, particularly for coal-fired sources. Removing the NOX directly from
the flue gas can be used in addition to combustion modification when high removal efficiencies
are required.
Flue gas treatment (FGT) processes reduce NOX emissions from combustion sources either by
decomposing N0x to nitrogen and water or oxygen, or by removing NOX from the gas stream. Work on
these systems in the United States is being funded by EPA, but the major work is being conducted
in Japan (References 3-40 and 3-73). At present, FGT is commercially available only for oil and
natural gas firing. For S02 and particulate laden gas streams, FGT for NO removal is still in the
A
developmental stage.
For convenience of discussion, the two FGT process routes can be categorized as dry pro-
cesses (reduction) and wet processes (oxidation followed by scrubbing). Dry systems are operated
at about 644K (700F) and generally employ flue gas additives and catalysts. Wet systems employ a
wider variety of chemicals and are operated at 313K to 323K (100F to 120F), the same temperatures
scrubbers use to remove SO,,. These processes are described separately below.
3.2.1 Dry Flue Gas Treatment
Dry processes are the most fully developed. They are mainly applicable to flue gas
streams free of S02 and particulates; that is, gases from the burning of gaseous fuels or distil-
late oils. Large dry FGT systems have been in operation since 1974 in Japan. Some systems applied
to S02 and particulate laden gas streams have been piloted successfully and several prototype plants
are now being constructed to treat gases from residual oil and coal-fired boilers (References 3-40
and 3-48).
Although there are many theoretical dry process variations (e.g. , nonselective reduction,
selective reduction using NHg. and molecular sieves), only selective catalytic reduction using
ammonia has achieved notable success in treating combustion flue gases for removal of NOX- The
presence of oxygen in concentrations many times greater than N0x in the flue gas precludes consider-
ation of nonselective reduction. However, selective reduction of NOX using ammonia is readily
accomplished using any of a number of catalysts.
3-47
-------
absorber may be so great as to preclude the use of liquid phase oxidation processes on combustion
flue gases where large volumes of gas and low NO concentrations are involved.
The use of ozone or chlorine dioxide to oxidize NO in the gas stream prior to the scrubber
appears to be the more successful approach. Although the required scrubber is still quite large,
substantial removal of NOX can be obtained in a scrubber designed for S02 removal. Unfortunately,
chlorine dioxide is expensive and its use introduces the problem of disposing of chloride-
containing liquid discharges. In addition, the production of ozone requires expensive equipment
and a great deal of electrical energy. For coal-fired flue gas with its higher NOX concentrations,
the oxidant cost is very likely to be prohibitive. Also, where ozone is utilized, additional
equipment may be required for removal of excess ozone from the final gas stream.
Although wet NOX FGT systems may not see widespread use in this country, two deserve addi-
tional mention because thej are relatively simple extensions of well established flue gas desulfuri-
zation (FGD) technology. The Chiyoda 101 F6D process has been modified by the inclusion of an
ozone generator for NO oxidation. The absorbed NO is removed from the system as a dilute calcium
nitrate solution requiring disposal. The process has been designated Chiyoda Thoroughbred 102.
This
yet.
This simultaneous SO /NO process has been piloted in Japan but has not seen commercial service as
Mitsubishi Heavy Industries (MHI) also is developing a wet NO FGT process which is a fairly
A
simple extension of their limestone FGD process. In this case, two additional pieces of equipment
are required: one for ozone generation and injection into the flue gas, and one for treatment of
the tail gas to remove unreacted ozone prior to release to the atmosphere. The MHI process differs
from most other wet FGT processes in that the captured nitrogen oxides are reduced to elemental
nitrogen by reaction with calcium sulfate in the circulating scrubber liquor. A proprietary
catalyst present in the scrubber liquor promotes this reaction. This simultaneous SO /NO process
is presently in the pilot plant state in Japan.
Both the Chiyoda and MHI processes are attractive from the standpoint of having simultaneous
SO /NO control capability and from the developmental standpoint since they involve presently com-
mercial FGD technology. However, the high energy consumption associated with the required production
of ozone is likely to render wet simultaneous SOX/NOX processes impractical. For example, the oxida-
tion of NO to N02 by ozone for a 300 MW power plant is estimated to require 3 to 9 percent of the
plant power output (Reference 3-77).
It appears that wet FGT systems cannot compete with dry selective catalytic reduction where
simple N0x control is involved. For coal-fired applications where the dust loading and S02
3-50
-------
boiler and a 147 MW (500 x 10« Btu/hr) gas-fired furnace. Reductions of 70 percent were achieved
with a 1.5:1 molar ratio of ammonia to NO (Reference 3-37).*
Although selective noncatalytic N0x reduction holds promise, further tests and process
studies are needed before its application in, the field. The major requirements are: (1) determina-
tion of its applicability to systems other than steam generators, such as gas turbines and combined
cycles, (2) identification of reducing agent injection rate requirements for retrofit field appli-
cations, (3) evaluation of techniques for maintaining adequate convective section temperatures
required for selective reduction during boiler load changes, (4) assessment of possible byproduct
emissions, and (5) assessment of the impact on reducing agent markets (References 3-37, 3-39, 3.75).
Another dry FGT process that has been identified as a possible NOX control technique is
molecular sieve adsorption. In general, this is not applicable to the water-containing effluents
from combustion sources due to preferential absorption of moisture and resultant loss of active
sites. It holds promise primarily for specialized noncombustion applications where N02 concentra-
tions are high (i.e., nitric acid plants).
In general, the considerable experience in Japan qualifies N0x flue gas treatment by selec-
tive catalytic reduction as commercially available for application to gas- and oil-fired sources
in the U.S. from a technical standpoint. There are, however, several factors and questions which
must be considered in determining the potential for widespread use of this technology in the U.S.
Of particular importance is the applicability of this technology to coal-fired sources. EPA's
research and development program is aimed at resolving these questions (Reference 3-74).
3.2.2 Wet Flue Gas Treatment
The chemistry and process steps involved in wet processes are considerably more varied than
in dry processes (Reference 3-74). All Of those systems which have advanced beyond bench scale
involve the use of a strong oxidant such as ozone or chlorine dioxide to convert the relatively
inactive NO in the flue gas to N02 or N20 for subsequent absorption. Unfortunately, nearly all
wet processes result in a troublesome byproduct which may be little commercial value. Some of
these byproducts are nitric acid, potassium nitrate, ammonium sulfate, calcium nitrate, and
gypsum.
The required oxidation for wet FGT processes can take place either in the liquid or gas
Phase. Those processes that utilize liquid phase oxidation require extensive liquid/gas contact
in order to absorb the inactive NO. It appears that the size and pressure drop of the NO
*Amnonia injection is discussed in more detail in Section 3.1.2.7.
3-49
-------
These designs will be described in Section 3.3.1. Techniques suitable for retrofit abatement for
older plants or add-on controls for plants built using old technology include catalytic reduc-
tion, extended absorption with and without refrigeration, wet chemical scrubbing, and molecular
sieve adsorption. These techniques will be described in Section 3.3.2. The techniques used for
other noncombustion sources, such as explosive plants and adipic acid plants, are basically the
same as those used for nitric acid plants, but vary with choice depending on economies of scale
and throughput.
3.3.1 Plant Design for N0y Pollution Abatement at New Nitric Acid Plants
Nitric acid is manufactured in the United States by the catalytic oxidation of ammonia
over.a platinum or palladium catalyst with the subsequent absorption of the product gases, primarily
N0? and NO, by water to make nitric acid. A more detailed discussion of the chemical process is
given in Section 6. Each of these two catalytic processes have optimum conversions at different
operating conditions. Moderate pressures of 300 to 500 kPa allow longer catalyst life by lowering
operating temperatures in the initial oxidation reaction. Higher pressures in the range of 800 to
1100 kPa (116 to 160 psia) allow higher absorption rates in the absorption columns with smaller
equipment sizes and lower costs. The higher conversions of N02 to HN03 allow for smaller equipment
for both the main process plus any tail gas treatment required to meet emission standards. Cur-
rently most existing plants operate at low or moderate pressures throughout the process. Sections
3.3.1.1 and 3.3.1.2 will discuss how the design of new nitric acid plants has taken these factors
into account to increase conversion and decrease emission control costs.
3.3.1.1 Absorption Column Pressure Control
By designing a new plant so that the inlet pressure at the absorption is 800 to 1000 kPa
(116 to 145 psia)> the efficiency of the absorber can be increased so that an effluent of less
than 200 ppm NO is emitted. A high inlet gas pressure at the absorber can be achieved either by
running the ammonia-oxygen reaction at high pressure, or by running the ammonia-oxygen reaction
at low pressure, with compression of the gas stream before introduction to the absorber. Higher
absorption pressures will increase the convertion of NOp to nitric acid and minimize NOX emissions.
However, there are economic penalties in the form of increased equipment cost, thicker walls and
compressors, and increased maintenance costs.
3.3.1.2 Strong Acid Processes
Nitric acid is usually produced at strengths of 50 to 65 percent by weight in water due to
azeotrope limitations. Azeotropic conditions result in a constant composition in both vapor and
3-52
-------
concentrations are high, it is not clear whether dry FGT combined with conventional FGD processes
will be cheaper than the wet simultaneous SOX/NOX systems such as Chiyoda 102 or MHI. Dry simul-
taneous S0x/N0x systems such as the Shell and the Sumitomo Shipbuilding processes may also prove
to be cheaper than the wet simultaneous processes. The Shell process is being commercially applied
on a 40 MW* oil-fired boiler in Japan and is being applied in the U.S. on the flue gas streams from
a 0.6 MW* coal-fired boiler. The Sumitomo Shipbuilding process will be tested at the prototype
level on an oil-fired boiler.
In general, wet processes are less well developed and show higher projected costs than dry
FGT processes. Considering their cost and complexity, it is doubtful that wet processes would be
receiving any development attention in Japan were it not for the potential for simultaneous SO
and NOX removal. For both processes, a number of important questions concerning NO FGT costs,
secondary effects, material use, reliability, and waste disposal remained to be answered. EPA
is proceeding with a coordinated program of experimental work, technology assessment, and engineer-
ing studies to answer these questions (References 3-48, 3-74).
3.3 Noncombustion Gas Cleaning
Emissions from noncombustion sources as industrial or chemical processes are small relative
to the total emissions from stationary sources (1.7 percent)'. Nationwide NOX emissions from nitric
acid manufacturing are estimated, for the year 1974, at 127 Gg (140,000 tons) uncontrolled emissions,
which is about 1.0 percent of the total stationary source emissions. The Environmental Protection
Agency issued standards (under the authority of the Clean Air Act) that new nitric acid plants
constructed after December 23, 1971, have a maximum permitted nitrogen oxide effluent of 1.5 kg
(measured as N02) per Mg of acid (100 percent basis) produced (3 Ib/ton). This is equivalent to
approximately 210 ppm NOX- For existing plants the maximum nitrogen oxides permitted has been set
at 2.75 kg/Mg (5.5 Ib/ton) of acid or approximately 400 ppm N0x in several states. These standards
were established in consideration of the then available technology, which was catalytic reduction
of NOX to t\2 and water using methane or hydrogen.
Several economic factors, discussed in Section 3.3.2.4 have stimulated development of
Improved processes for tail gas cleaning and improvements in the nitric acid process itself. One
of the major considerations is that much of the residual oxides of nitrogen formed in the manu-
facture of nitric acid can be recovered and converted into nitric acid, thus increasing the plant
yield. Also, new plants can be designed to have low NOX emissions without add-on control equipment.
*electric output rating.
3-51
-------
3.3.2.1 Chilled Absorption
The basic principle involved is that the amount of NOX that can be removed from the process
gas by the absorber (water) increases as the water temperature decreases. Therefore, this method
of retrofit provides for chilling of the water prior to entry into the absorption tower or by direct
cooling of the absorption trays. This method of NOX reduction has only provided marginal results
and has had problems in continuously meeting the NSPS, especially in warm weather. Refrigeration re-
quirements can prove costly, both in equipment and energy use.
3.3.2.2 Extended Absorption
One of the most commonly used retrofit processes, which has been used effectively to meet
the NSPS, is extended absorption. Figure 3-15 shows the flow diagram of a nitric acid plant after
addition of the extended absorption system, which consists of an additional absorber and a pump.
This method is offered by several licensors both with and without other features such as compres-
sion of the tail gas before entry to the additional tower or a supply of chilled water to the
absorption column trays. Because of the additional pressure loss in the second column an
inlet pressure of at least 700 kPa (101 psia) is preferred to make the economics of this method
attractive.
3.3.2.3 Wet Chemical Scrubbing
Wet chemical scrubbing removes NOX from nitric acid plant tail gases by chemical reaction.
Liquids such as alkali hydroxide solutions, ammonia, urea, and potassium permanganate convert N02
to nitrates and/or nitrites. These techniques produce a liquid effluent which needs disposal.
For three recent techniques urea scrubbing, ammonia scrubbing and nitric acid scrubbing the
effluent is a valuable byproduct which can be reclaimed and sold as fertilizer.
Caustic Scrubbing
In this process, NO in the tail gas reacts with sodium hydroxide, sodium carbonate, or
ammonium hydroxide to form nitrite and nitrate salts. Although caustic scrubbing removes NO from the
tail gas, it has not found extensive use in the industry because of the difficulties encountered in
disposing of the spent solution. The alkali metal nitrite and nitrate salts contained in the spent
solution become a serious water pollutant if released as a liquid effluent, and their concentrations
are too dilute for economic recovery.
3-54
-------
liquid phases. With higher operating pressures nitric acid up to 68 percent can be obtained.
Further concentration is sometimes accomplished by dehydration of the acid or further distillation
with sulfuric acid addition.
However, nitric acid of high strength can be made directly from ammonia by the Direct Nitric
Acid (DSNA) process. Ammonia is burned with air near atmospheric pressure, and the nitrogen oxides
are oxidized to nitrogen dioxide in a contact tower. The nitrogen dioxide is then separated from
the gas stream by physical absorption in chilled high-concentrated nitric acid, stripped by distilla-
tion and then liquified as NpO..
The liquid dinitrogen tetroxide is pumped to a reactor together with aqueous nitric acid.
Pure oxygen is added and the dinitrogen tetroxide reacts at a pressure of approximately 5200 kPa
(756 psig) directly to highly concentrated nitric acid. Variations on the process can produce
both strong (98 to 99 percent) nitric acid and weak (50 to 70 percent) nitric acid at the same
plant (Reference 3-78). Tail gas emissions from this process are within the 1.5 g/kg (3 Ib/ton) NO
regulation. This occurs primarily by ensuring oxidation to N02 and physical absorption with the
concentrated nitric acid at low temperature.
Concentrated nitric acid has also been made by the SABAR (Strong Acid By Azeotropic Reacti-
vication) process. Ammonia combustion occurs at near atmospheric pressure and at 1.123K (1.560F)
with the usual waste-heat boiler, tail gas p'reheater, cooler/condenser effluent train. By mixing
the combustion gases with feed air and recycled nitrogen dioxide and compression nearly all the
NO is converted to NO.,. Chemical absorption with an azeotropic mixture of about 68 percent (by
weight) nitric acid produces a superazeotropic mixture. A 99 percent (by weight) overhead prod-
uct is produced by vacuum distillation.
3'3'2 Profit Design for NOX Pollution Abatement at New or Existing Nitric Acid Plants
Most existing nitric acid plants were not designed with the present NOX emission standards
in mind. Abatement methods for these plants are installed on a retrofit basis. The available abate-
ment methods include chilled absorption, extended absorption, wet scrubbing, catalytic reduction,
and molecular sieve adsorption. In this section, these various control techniques for NO are
described. These same procedures are also used on new nitric acid plants using the earlier low or
moderate operation pressure design where the abatement facility is designed to process the tail gas
to meet the 1.5 g N02/kg of acid product (3 Ib/ton) emission standard.
3-53
-------
Urea Scrubbing
Urea can be used to treat tail gases for NOX control since it reacts rapidly with nitrous
acid. Nitrogen dioxide, N(L reacts with water to form both nitric acid (HNOo) and nitrous acid
(HNOp) in equal proportions. Nitrous acid will rapidly decompose to form NO and NOo. Urea
(CO (NHpJo) wnen contacted with the tail gas will absorb N0£ indirectly as nitrous acid to form
ammonium nitrate, NhLNO- and free nitrogen, N«. By depleting the liquid phase of nitrous acid
the equilibrium conversion of nitric oxide, NO, to nitorgen dioxide occurs to remove NO also.
The result is conversion of NOp to either free nitrogen which is vented to atmosphere or ammonium
nitrate which is sold as fertilizer.
Ammonia Scrubbing
Ammonia, a weak base, can be used to scrub the oxides of nitrogen (weak acids) from the
nitric acid plant tail gas. The- product of this scrubbing reaction is an ammonium nitrate solu-
tion (NH.NO.J which can be recovered and sold as fertilizer. This process can be applied to tail
gas concentrations up to 10,000 ppm and requires 1 to 1.5 percent excess oxygen.
Nitric Acid Scrubbing
Nitric acid scrubbing of tail gas has been commercially applied by one licensor. The pro-
cess uses both physical absorption and stripping and chemical oxidation absorption. The process
uses only water and nitric acid and coverts nitrogen oxides in the tail gas to nitric acid at
concentrations which can be commercially utilized (Reference 3-79).
Potassium Permanganate Scrubbing
A potassium permanganate scrubbing process has been used to reduce NO emissions from
1800 ppm to 49 ppm at a nitric acid concentration plant in Japan. The process reacts potassium
permanganate with nitrogen oxide and sodium hydroxide to form potassium sodium manganate, sodium
nitrite and potassium nitrite. The potassium permanganate is regenerated by oxidizing the
potassium sodium manganate electrolytically (References 3-80 and 3-81). However, the process
is presently considered to be too expensive to be competitive (Reference 3-82). It has not
been tried on any plants in the United States, and is not presently offered by any licensor.
3-56
-------
COMP.
AIR
AMMONIA
i
1
CONVERTER
!
HEAT
RECOVERY
-^
? ?
CONDENSER
PROD
f~
*.
1
1
1
1
1
r m \
"^"^ i
oc
LU
00
o
<
^ ,
ur.Tf
PUMP \^
^ 1
1 1
1
1
1
1
I
^^^^^^
-
V
LU uj;
Q OQ:\
Z QC i
LU O
1- eo^
LU
ACID
Figure 3-15. Extended absorption system on existing nitric acid plant.
3-55
-------
NO abatement using nonselective catalyst is more difficult technically than decolorization,
and commercial results have been less satisfactory. Provisions must be made to control the heat
released in reacting all the tailgas oxygen. The thermal control must be done before extensive
NO reduction proceeds.
In Section 6 the success of the various types of catalytic abaters in coping with the
problems of temperature rise and high space velocities will be discussed. In general, nonselective
catalytic reduction is not likely to be used in the future for NOX control. The availability and
cost of natural gas, increasing catalyst cost and poor performance have led to a decline in inter-
est in this process.
Selective catalytic reduction
In selective catalytic reduction, ammonia is reacted with the NOX to form Ng. The oxygen
in the tail gas does not react with the ammonia, so stoichiometric amounts of ammonia are used.
In contrast to nonselective techniques, selective catalyst abatement must be carried out
within the narrow temperature range of 483K to 544K (410F to 520F). Within these limits, ammonia
will reduce N02 and NO to molecular nitrogen, without simultaneously reacting with oxygen. The
overall reactions are shown in the following equations:
8NH3 + 6N02 + 7N2 + 12 H20 (3-9)
4NH3 + 6NO - 5N2 + 6H20 (3-10)
Above 544K, ammonia may oxidize to form NO ; below 483K, it may form ammonia nitrate.
Selective oxidation with ammonia has several advantages over nonselective reduction:
0 The reducing agent, ammonia is usually readily available since it is consumed as
feed stock in the nitric acid process
Temperature rise through the reactor bed is only 20K to 30K (36F to 54F) so that
energy recovery equipment, such as a waste heat boiler or high temperature gas
turbine, is not required
Lower raw material costs since the amount of ammonia required is approximately
equal to the molal equivalent amount of NO abated
Heterogeneous Catalysis
One wet scrubber process uses heterogeneous catalysis in a packed column to oxidize NO to
N02 (References 3-83 and 3-84). This system is currently in the development stage.
3-58
-------
3.3.2.4 Catalytic Reduction
There are three types of catalytic reduction processes used for NO control: nonselective
reduction, which removes both NO and oxygen; selective reduction, which removes only NO , and
x x
heterogeneous catalysis used in conjection with wet scrubbing. Each of these will be discussed
in the following paragraphs.
Nonselective catalytic reduction
The nonselective reduction process reacts NOX with H2 or CH4 to yield N2, C02 and H20.
The process is called nonselective because the reactants first deplete all the oxygen present
in the tail gas, and then remove the NOX- Prior to the large increases in natural gas prices the
excess fuel required to reduce the oxygen did not Impose a heavy economic penalty. The reactions
were exothermic, and much of the heat could be recovered with a waste heat boiler.
The nonselective reduction process is used for decolorization and energy recovery, as well
as for NOX abatement. Decolorization and power recovery units reduce N02 to NO and react part
of the oxygen, but their capacity to reduce NO to elemental nitrogen is limited. The nonselective
abatement units carry the process through to NO reduction as well. In nonselective reduction,
the tail gases from the absorber are heated to the necessary catalyst ignition temperature, mixed
with a reducing agent, such as hydrogen or natural gas, and passed into the reactor and through
the catalyst. The main chemical reactions that take place are:
CH4 + 4N02 + 4NO + C02 + 2H20 (3-6)
CH4 + 202 + C02 + 2H20 (3-7)
CH4 + 4NO -» 2N2 + C02 + 2H20 (3-8)
Similar equations can be written substituting hydrogen for methane, in which case two moles
of hydrogen are needed to replace one mole of methane. The reaction kinetics are such that reduc-
tion reaction (3-6) is faster than reduction reaction (3-7), but abatement reaction (3-8) is much
slower than reaction (3-7). Thus, decolorization can be accomplished by adding just enough fuel for
partial oxygen burnout. If NOX abatement is required, however, sufficient fuel must be added for
complete oxygen burnout.
Both catalyst and nitric acid manufacturers report satisfactory performance for decoloriza-
tion units. The reduction of total NO is limited, but ground-level NO, concentration in critical
x c
areas near the plant is reduced substantially.
3-57
-------
3-10 Sarofim, A. F., ejt aJK, "Mechanisms and Kinetics of NO Formation: Recent Developments,"
presented at 65th Annual AIChE Meeting, Chicago, November 1976.
3-11 Snyder, R., "Nitrogen and Oxygen Compound Types in Petroleum," Analytical Chemicstry 41:
314-323, February 1969.
3-12 Martin, G. B., and E. E. Berkau, "An Investigation of the Conversion of Various Fuel Nitro-
gen Compounds to Nitrogen Oxides in Oil Combustion," presented at AIChE meeting, August 30,
1971, Atlantic Civ--. August, 1971.
3-13 United States Senate, Committee on Public Works, "Air Quality and Stationary Source Emission
Control," Serial No. 94-4, March 1975.
3-14 Habelt, W. W. and B. M. Howell, "Control of NO Formation in Tangentially Coal-Fired Steam
Generators," in Proceedings of the NOX Control Technology Seminar, EPRI SR-39, February 1976.
3-15 Heap, M. P., elt al., "The Optimization of Burner Design Parameters to Control NOX Formation
in Pulverized CoaT and Heavy Oil Flames," in Proceedings of the Stationary Source Combus-
tion Symposium, EPA-600/2-76-152b, June, 1976.
3-16 Pohl, J. H., and A. F. Sarofim, "Devolatilization and Oxidation of Coal Nitrogen," presented
at 16th International Symposium on Combustion, M.I.T., August 1976.
3-17 Blair, D. W., et al_., "Devolatilization and Pyrolysis of Fuel Nitrogen from Single Coal
Particle Combustion," 16th Symposium (International) on Combustion, Cambridge, Mass., 1976.
3-18 Pershing, D. W., "Nitrogen Oxide Formation in Pulverized Coal Flames," PhD Dissertation,
University of Arizona, 1976.
3-19 Axworthy, A. E., Jr., "Chemistry and Kinetics of Fuel Nitrogen Conversion to Nitric Oxide,"
AIChE Symposium Series. No. 148, Vol. 71, 1975, pp. 43-50.
3-20 Axworthy, A. E., ejt al_., "Chemical Reactions in the Conversion of Fuel Nitrogen to NOX," in
Proceedings of the Stationary Source Combustion Symposium, Volume I, Fundamental Research,
June 1976.
3-21 Pershing, D. W., and J. 0. L. Wendt, "The Effect of Coal Combustion on Thermal and Fuel
NOX Production from Pulverized Coal Combustion," presented at Central States Section,
The Combustion Institute, Columbus, Ohio, April 1976.
3-22 Pohl, J. H. and A. F. Sarofim, "Fate of Coal Nitrogen During Pyrolysis and Oxidation," In:
Proceedings of the Stati
EPA 600/2-76-152a, June
Proceedings of the Stationary Source Combustion Symposium. Volume I. Fundamental Research,
3-23 Barr, W. H., and D. E. James, "Nitric Oxide Control -A Program of Significant Accomplish-
ments," ASME 72-WA/Pwr-13.
3-24 Barr, W. H., ejt al_., "Retrofit of Large Utility Boilers for Nitric Oxide Emissions Reduc-
tion Experience and Status Report."
3-25 Crawford, A. R., et al_., "Field Testing: Application of Combustion Modifications to Control
NOX Emissions from Utility Boilers," Exxon Research and Engineering Co., EPA-650/2-74-066,
June 1974.
3-26 Crawford, A. R., e_t a_l_., "The Effect of Combustion Modification on Pollutants and Equipment
Performance of Power Generation Equipment," Exxon Research and Engineering Co., EPA-600/2-
76-152c, prepared for the Stationary Source Combustion Symposium, September 24-26, 1975.
3-27 Blakeslee, C. E., and H. E. Burbach, "Controlling NOX Emissions from Steam Generators,"
C.E. Inc., APCA 72-75, 65th Annual Meeting of Air Pollution Control Association, June 18-22,
1972.
3-28 Hollinden, G. A., et al_. , "Evaluation of the Effects of Combustion Modifications in Con-
trolling NOX Emissions at TVA's Widow's Creek Steam Plant," EPRI SR-39, February 1976.
3-29 Maloney, K. L., "Western Coal Use in Industrial Boilers," Western States Section/The
Combustion Institute, April 19-20, 1976, Salt Lake City, Utah.
3-60
-------
3.3.2.5 Molecular Sieve Adsorption
One method of NOX control involves the adsorption of N0x onto a solid followed by regenera-
tion of the adsorbent. Materials such as silica gel, alumina, charcoal, and commercial zeolites
or molecular sieves have been employed for this method. Molecular sieves have been found to be
the most effective medium for this method of control, since they adsorb N02 selectively.
Special sieves have been developed which incorporate a catalyst to simultaneously convert NO to
N02. This process operates best only when low concentrations of oxygen are present, which is true
of most tail gas streams. The abatement bed is usually provided with a dehydration section prior
to contact with the abatement sieve to improve overall performance.
The adsorbent bed is regenerated by thermally cycling the bed after it is loaded with NO .
The required regenerating gas is obtained by using a portion of the treated tail gas stream to
desorb the adsorbed NO,, from the bed. This gas stream is then recycled to the nitric acid plant
absorption tower. No other liquid, solid or gaseous effluents are produced by this process.
Two plants using this system were in operation and had experienced difficulties. The pro-
cess has become unattractive for future installations because of the cost of the catalyst bed,
the energy cost of thermal cycling, and the operational difficulties of using a cycling adsorption
process with a steady state nitric acid plant.
REFERENCES FOR SECTION 3
1n CombuSt1on and Explosions," Acta Physiochim,
MacKinnon, D. J "Nitric Oxide Formation at High Temperature," Journal of the Air Pollution
Control Association, Vol. 24, No. 3, March 1974. niuciuri
3-3 Rawdon, A. H. , and R. S. Sadowski, "An Experimental Correlation of Oxides of Nitroapn
***** '" ^ Data'" ^ ASME °°Ur"a! " W»X for
3-2
3-4
3-5
Breen, B. P. ."Control of the Nitric Oxide Emissions from Fossil Fueled Boilers " The Fourth
Westinghouse International School for Environmental Management, July 15-18, 1973.
Bartz, D. R. , et ah , "Control of Oxides of Nitrogen from Stationary Sources in the South
Coast Air Basin," California Air Resources Board Report No. ARB 2-1471? Septeiber, 1974*
3"6 !Sf:hi;«PA; S£ ' D"?Urn?r 5r1Jer?a for N0x Control» Volume I. Influence of Burner
Variables on NOX in Pulverized Coal Flames," EPA 600/2-76-061 a, March 1976.
Drtllll. ... ^-. of Interaction Between Fluid Dynamics on Chemistry of
S^nn^m v^ ? 'S ^^IT^ Proceedings of the Stationary Source Combustion
Symposium, Volume I. Fundamental Research. EPA 600/2-76-152a. .i..np iQ7fi
3~8 nJScinJ:/*; !ud ?' C' Ih°mas' "Oxides of Nitrogen in Relation to the Combustion of Coal,"
presented at the Seventh International Conference on Coal Science, Prague, June, 1968
3"9 FuPlShNn9fr;mWRocTJ,,!T'A-iInf]UrnC? ?f ?esi?n Vari'ables on the Production of Thermal and
Fuel NO from Resi^uaT Oil and Coal Combustion," AIChE Symposium Series. No. 148, Vol. 71,
iy/0, pp. iy-^9.
3-59
-------
3-49 "Analysis of the Proposed National Energy Plan" - Office of Technology Assessment, Congress
of the United States, August 1977.
3-50 Ctvrtnicek, T. E., e_t aj_., "Evolution of Low-Sulfur Western Coal Characteristics, Utiliza-
tion and Combustion Experience," Monsanto Research Corp., EPA 650/2-75-046, May 1975.
3-51 Shimizu, A. B. , et al_. , "No Combustion Control Methods and Costs for Stationary Sources,"
Environmental Protection Technology Series, EPA-600/2-75-046, September 1975.
3-52 Bartick, H. A., ejt a_l_., "The Production and Refining of Crude Shale Oil into Military
Fuels," Final Report by Applied Systems Group, submitted to Office of Naval Research,
Arlington, Va., August 1975.
3-53 Martin, G. B., "Evaluation of NOX Emission Characteristics of Alcohol Fuels in Stationary
Combustion Systems," presented at Joint Meeting, Western and Central States Sections,
The Combustion Institute, April 21 and 22, 1975, San Antonio, Texas.
3-54 Hall, R. E., "The Effect of Water/Residual Oil Emulsions on Air Pollutant Emissions and
Efficiency of Commercial Boilers," ASME 75-WA/APC-l, July 14, 1975.
3-55 Martin, G. B., "Environmental Considerations in the Use of Alternate Clean Fuels in
Stationary Combustion Processes."
3-56 Martin, G. B., D. W. Pershing,E. E. Berkau, "Effects of Fuel Additives on Air Pollutant
Emissions from Distillate Oil-Fired Furnaces," EPA, Office of Air Programs, AP-87, June 1971.
3-57 Shaw, H., "Reduction of Nitrogen Oxide Emissions from a Gas Turbine Combustor by Fuel Modi-
fications," ASME Transactions, Journal of Engineering for Power, 95, 4, October 1973.
3-58 Altwicker, E. R., et al_., "Pollutants from Fuel Oil Combustion and the Effects of Additives,"
Paper No. 71-14, 64th Annual APCA Meeting, Atlantic City, N. J., June 1971.
3-59 Barrett, R. E., e_t aj_., "Field Investigation of Emissions from Combustion Equipment for
Space Heating," EPA R2-73-084a, June 1973.
3-60 Frey, D. J., "De-Ashed Coal Combustion Study," Combustion Engineering, Inc., October 1964.
3-61 Energy Research and Development Agency, "Proceedings of the Fourth International Conference
on Fluidized Bed Combustion," McLean, Va., December 1975.
3-62 Jonke, A. A., et^ al_., "Pollution Control Capabilities of Fluidized-Bed Combustion," AIChE
Symposium Series No. 126, Vol. 68, 1972.
3-63 Chronowski, R. A., and B. Molayem, "NOX Emissions from Atmospheric Fluidized-Bed Boilers,"
ASME 75-PWR-4, October 1975.
3-64 Pfefferle, W. C., el^ §J_., "CATATHERMAL Combustion: A New Process for Low-Emissions Fuel
Conversion," presented at the 1975 ASME Winter Annual Meeting, Houston, Texas, ASME
Paper No. 75-WA/FU-l.
3-65 Kesselring, J. P., et a]_., "Catalytic Oxidation of Fuels for NOX Control from Area Sources,"
EPA Report, EPA-600/2-76-037, February 1976.
3-66 DeCorso, S. M., et al., "Catalysts for Gas Turbine Combustors - Experimental Test Results,"
paper presented at A"SME Gas Turbine Conference and Products Show, New Orelans, March 1976,
ASME Paper #76-GT-4.
3-67 Gerstin, R. A., "A Technical and Economic Overview of the Benefits of Repowering," paper
presented at the Gas Turbine Conference and Products Show, Houston, Texas, March 2-6, 1975,
ASME Paper #75-GT-16.
3-68 Ahuja, A., "Repowering Pays Off for Utility and Industrial Plants," Power Engineering,
pp. 50-54, July 1976. a
3r62
-------
3-30 Brown, R. A., H. B. Mason, and R. J. Schreiber, "Systems Analysis Requirements for Nitrogen
Oxide Control of Stationary Sources," Environmental Protection Technology Series EPA-650/2-
74-091, Sepgember 1974.
3-31 Selker, A. P., and R. L. Burnington, "Overfire Air Technology for Tangentially Fired Utility
Boilers Burning Western U.S. Coals," in Proceedings of the Second Stationary Source Combustion
Symposium Vol II, Utility and Large Industrial Boilers. EPA-600/7-77-073b. July 1Q77,
3-32 Bagwell, F. A., et. ah, "Utility Boiler Operating Modes for Reduced Nitric Oxide Emissions,"
Presented at the 64th Annual Meeting of the Air Pollution Control Association, June 1971.
3-33 U.S. Environmental Protection Agency, "Draft - NOX SIP Preparation Manual Volume II -
Support Sections," Office of Air Quality Planning and Standards, Research Triangle Park
N.C., April, 1976.
3-34 Copeland, J. 0., "Standards Support and Environmental Impact Statement: An Investigation
of the Best Systems of Emission Reduction for Nitrogen Oxides from Large Coal-Fired Steam
Generators," (Draft) EPA, October 1976.
3-35 Crawford, A. R., E. H. Manny and W. Bartok, "Field Testing: Application of Combustion
Modifications to Power Generating Combustion Sources," in Proceedings of the Second Stationary
Source Combustion Symposium, Volume II. Utility and Large Industrial Boilers, EPA-600/7-77-
073b.
3-36 Lyon, R. L., "Method for the Reduction of the Concentration of NO in Combustion Effluents
Using Ammonia," U. S. Patent No. 3,900,554, assigned to Exxon Research and Engineering
Company, Linden, New Jersey, August 1975.
3-37 Lyon, R. K. and J. P. Longwell, "Selective, Non-Catalytic Reduction of NOX by NHa,"
Proceedings of the NOX Control Technology Seminar, EPRI SR-39, February 1976.
3-38 Muzio, L. J., and T. K. Arand, "Homogeneous Gas Phase Decomposition of Oxides of Nitrogen,"
EPRI Report FP-253, August 1976.
3-39 Teixeira, D. P., "Status of Utility Application of Homogeneous NOX Reduction," Proceedings
of the NOX Control Technology Seminar, EPRI SR-39, February 1976.
3-40 Ando, J. and T. Heiichiro, "NOX Abatement for Stationary Sources in Japan," Environmental
Protection Technology Series, EPA-600/2-76-013b, January 1976.
3-41 Shoffstall, D. R., "Burner Design Criteria for Control of Pollutant Emissions from Natural
Gas Flames,1 Institute of Gas Technology, EPA-600/2-76-152b, June 1976.
3-42 Koppang, R. R., "A Status Report on the Commercialization and Recent Development History
of the TRW Low NOX Burner," TRW Energy Systems Group.
3-43 Tsuji, S., et al_., "Control Technique for Nitric Oxide - Development of New Combustion
Methods," IHI Engineering Review, Vol. 6, No. 2.
3-44 Ando, J., et aJL, "NOX Abatement for Stationary Sources in Japan," August 1976 (Preliminary
uraTt).
3-45 Shoffstall, D. R., "Burner Design Criteria for Control of NOX from Natural Gas Combustion,
Volume I,1 Institute of Gas Technology, EPA-600/2-76-098a, April 1976.
3-46 Brackett, C. E., and J. A. Barsin, "The Dual Register Pulverized Coal Burner -a NOv
Control Device," EPRI SR-39, February 1976.
3-47 Campobenedetto, E. J., "The Dual Register Pulverized Coal Burner - Field Test Results,"
presented at Engineering Foundation Conference on Clean Combustion of Coal, Franklin Pierce
College, New Hampshire, July 31-August 5, 1977.
3-48 Bowen, J. S., D. G. Lachapelle, and R. Stern, "Overview of EPA's NOX Control Technology for
Stationary Sources," presented at 67th Annual AIChE Meeting, December 1974.
3-61
-------
-------
3-69 Stambler, I., "Repowering Gives Glendale Extra 75 MW and Lower Rates," Gas Turbine World,
September 1977.
3-70 Robson, F. L., and A. J. Giramonti, "The Use of Combined-Cycle Power Systems in Nonpolluting
Central Stations," JAPCA, Vol. 22, pp. 177-180i 1972.
3-71 Amos, D. J., et al.., "Energy Conversion Alternatives Study (EGAS), Westinghouse Phase I
Final Report, Volume V - Combined Gas Steam Turbine Cycles," NASA CR-134941, Volume V, 1976.
3-72 Papamarcos, J., "Combined Cycles and Refined Coal," Power Engineering, December 1976, pp.
34-42.
3-73 Ando, J. and B. A. Laseke, "S02 Abatement for Stationary Sources in Japan," EPA-600/7-77-
103a, September 1977.
3-74 Stern, R., "The EPA Development Program for NOX Flue Gas Treatment," In: Proceedings of the
National Conference on Health, Environmental Effects, and Control Technology of Energy Use,
EPA Report 600/7-76-002, February 1976.
3-75 Koutsoukos, E. P., et al_., "Assessment of Catalysts for Control of NOX from Stationary Power
Plants, Phase I," Volume I, EPA-650/2-75-001-2, January 1975.
3-76 Muzio, L. L., J. K. Arand, and D. P. Teixeira, "Gas Phase Decomposition of Nitric Oxide in
Combustion Products," In: Proceedings of the NOX Control Technology Seminar, EPRI Special
Report SR-39, February 1976.
3-77 "Technology and Economics of Flue Gas NOX Oxidation by Ozone," EPA 600/7-76-003, December
1976.
3-78 "Nitric Acid from Ammonia," Hoechst-Uhde Corp. brochure (FWC 11 619), Englewood Cliffs, N.J.
3-79 Mayland, B. J., "The CDL/VITOK Nitrogen Oxides Abatement Process," Chenoweth Development
Laboratory, Louisville, Ky.
3-80 "New System Knocks NOX Out of Nitric," Chemical Week, September 3, 1975, pp. 37-38.
3-81 "NOX Removal System Now Available," Wet Scrubber Newsletter, September 30, 1973, pp. 3-4.
3-82 Personal communication, Mr. Kenneth Ficek, Technical Service Manager, Carus Chemical,
November 1977.
3-83 Mayland, B. J., "Application of the CDL/VITOK Nitrogen Oxide Abatement Process," presented at
Sulfur and Nitrogen Symposium, Salford, Lancashire, U.K., April 1976
3-84 Mayland, B. J., and R. C. Heinze, "Continuous Catalytic Absorption for NOX Emission Control,"
Chemical Engineering Process, Vol. 6, May 1973, pp. 75-76.
3-63
-------
The total N0x emitted in 1974 by the electric utility industry was 5.1 Tg (5.6 x 106 tons)
or 41.9 percent of the total stationary source emissions. Coal-fired boilers accounted for approxi-
mately 68 percent of the total utility emissions. A more detailed emission breakdown is presented
in Section 2. For reference, the ranges of uncontrolled NOX emissions for three of the firing types
are given in Table 4-1. Cyclone-fired boilers typically have the largest uncontrolled emissions,
and tangentially-fired have the lowest (Reference 4-1).
4.1.1 Control Techniques
The NOX control options for utility boilers include combustion modification, flue gas treat-
ment, and fuel modification. The former has been the most successful and widely used option, and
is described below for gas, oil, and coal-fired units. The other less popular and less developed
options are discussed after combustion modification.
4.1.1.1 Combustion Modification
The general concept of combustion modifications as potential NO control techniques for sta-
tionary sources was discussed in Section 3.1. These techniques have been developed and refined in
numerous laboratory test installations and in many successful field applications to commercial
utility boilers.
Utility boilers, due to their importance as NOX sources and their control flexibility, are
the most extensively modified stationary equipment type. The selection and implementation of effec-
tive N0x controls for a specific utility boiler is uniquely dependent on the furnace characteristics
(i.e., geometry and operational flexibility), fuel/air handling systems and automatic controls, and
to the potential for operational problems which may result from combustion modifications. The
following discussion is, therefore, not intended to provide application guidelines, but rather to
give a broad overview and evaluation of tested procedures.
Table 4-2 summarizes the status of combustion modification technology for NO control in
utility boilers. The references cited in the table are bases for the remainder of the discussion
in this section. The table also lists typical values of controlled emissions for the major modifi-
cation techniques and two major firing types, tangential firing and wall firing.
Retrofit N0x control implementation by combustion modification usually proceeds in several
stages depending on the emission limits to be reached. First, fine tuning of combustion conditions
by lowering excess air and adjusting the burner settings and air distribution is employed. If NO
emission levels are still too high, the minor modifications, such as biased firing or burners
out of service (BOOS) are implemented. Increased frequency of boiler washing increases flame heat
4-2
-------
SECTION 4
LARGE FOSSIL FUEL COMBUSTION PROCESSES
Fossil fuel combustion in utility and industrial boilers and internal combustion (1C)
engines account for about 80 percent of NOX emissions from stationary sources (see Section 2).
The large boiler category encompasses application to utility power generation and industrial
process steam generators. Large 1C engines are used predominantly for power generation and
for pipeline pumping and encompass large bore reciprocating engines as well as continuous
combustion gas turbine engines. This section summarizes the effectiveness, cost, user exper-
ience and energy and environmental impact of the implementation of NOX controls on these
equipment categories.
4.1 ELECTRICAL UTILITY BOILERS
Most of the nation's electricity is generated in large fossil-fueled central station
power plants, which consist of high-pressure watertube boilers in the 100 to 1000 MW* range
serving turbine-generators. Firing capacities of individual burners in utility boilers
commonly have thermal inputs as high as 58 MW (200 x 106 Btu/hr). A 1000 MW* opposed
wall-fired unit may require as many as 60 separate burners.
Although there are some differences among utility boiler designs in such factors as
furnace volume, operating pressure, and configuration of internal heating transfer surface,
the principal distinction is firing mode. This includes the type of firing equipment, the
fuel handling system, and the placement of the burners on the furnace walls (see Section
2.3.1).
*Electrical utility boilers are commonly described in terms of electrical output rating,
This convention will be used throughout Section 4.1.
4-1
-------
o
00
a:
o
o
o
0)
u
c
0)
^
0)
tt)
a:
TJ
-
co m
0 +J
r- X
4J tt)
IO Z
O
tQ.c2
D.
to
Existing
Application
o
to
O
+J
IO
-M
'i
i
,_^
xo
O CM
Z O
^~*&«
to co
0)
oi to
C CO
O IO
r- QQ
00
to CM
r- 0
0
r
M
iO
0)
O.
O
CJ-
Principle o
tt)
cr
c
.c
u
OJ
r
CM
VO i
1 1
** **
i
(- C
tt) -r-
> C
O CO O>
«f- i- CO
O O tt)
Q.TJ
O S- -M
CO (O C
3 3
r tt)
0 S- 2
C - OJ
1 1 J-
1 C i 0)
<*-!- O 0) T-
O O .Q
M S_ r- r-
i- >> 10 ro r- >,
«*- +J T- 10 +J
0 -- J- .r-
1.1 .»-> >,
4-> -r- CO CO CO -r-
tt) 4J i. 3 I_ -M
Q: 3 tt) TJ 0) 3
(
U O 1 1
tt) O O) S- OJ
> 0-"- 0 r-
C 4- -O C
O " tt) M- O
O C O> .+J CO T-
O IO O CO +->
f- -r- +-> O O O
O +* CO to CL 3
0 TJ
O> 0) >,.«>.« CO
c to i_ c c s-
r- ro O O
tt) E -r- -r- TJ
3 > -i- 4-> +-> IO
O «- S- T- IO O
U_ 4-> Q. C E
E E E
Q. O. Q.
O- O. O.
o m o
in CM o
r CM «±
o in o
o CM in
r r- CM
r
CO r IO
<0 -r- O
CJJ O <_)
TJ 1 1 1
O) tt) -i- to
4-> S- S. 0)
(O Q. i- TJ
S- O) Q. C
tt) C C Q. IO
CL-r- -r- 3
O TJ tO r
CO tt) tt) TJ E
0) T- > C i.
i >> tt) (0 tt)
N r- JZ
N ^: O) 4->
O 0 CM C X
C T- O O <4- O
, S- is* 0 Z
5- TJ
0) tt) >, C i
5 tt) O S- O 0)
O 3 3 IO T- 3
1 M- TJ E to »_
r-
>r.
S- 4->
r- C
OJ c
4 c
i- J= -r-
tt) 4J S-
> T- !-
^ s * '
CM in
r r
1 1
1
S- C
a,-c
o to o>
M -r-
>t- J- 10
O O tt)
Q-TJ
0 J- 4->
3 3
*3 £ $
C - OJ
< <4 C
1
1 C r
<4- £ "~ *O
O O tt) JD
JD Ol
4J I. r- *
r- >, (0 (0 tt) >> CO
«4- +J -f- C !-> t-
0 -r- $. n- ft)
+> V CO CO tO 'I- >-
tt) *-> J- 3 S_ 4J O
Or 3 tt) TJ tt) 3 X)
1
r tt)
3 >
O -r-
-C 4- 4->
4J 0
r- « tt)
5 01 >
C T- 0
O S- 0 .
CO «*- If- O
0 0-r-
S- r -M
I- IO O) O
O O C tt)
O O T- to
E E E
0. Q. 0.
Q. Q. Q.
O O O
O m o
co co in
000
o in o
CM CM ^1-
to i (o
IO -r- O
CJ3 O C_>
CO
C S- C J-
r- O -r- tt) tt)
J- Q. C -C
r- r- O «-> X
M- t- tt) N O
r- > to Z
t- (O tt) tt) 0)
0) E CO i
CO) IO to 0)
i- 1- CMr tt) 3
3 T- 0 «*- S_ «4-
J3 <4- O.
1- CO >, Q.TJ
TJ tt) tt) 1- 3 C
tt) > o ig co 10
to o 3 E
O TJ -r- TJ
i- i. tt) i- C (0
03 O t. Q. 10 E
L.
^
< ^
i- *io
r 12 O>
J- .C !-
tt) 4-> S-
*> *r~
i- C O
C
O 0) '"I
r- O> CO IO
to U L. O
3 10 tt) O
73 '" ^ t.
C «t- O O
« O .Q M-
M- to
0 -i- -r-
TJ 0
r- TJ tt) >» CO
"4- C -(-> +J S_
O CO r^ ^ ^
0) IO »-»-> O
Q£ O>4-> 3 ^1
> 1
M 10 C
CJ CD -r-
«f- » ^
tt) r 10 >,
Or- 4?
TJ O M- !-
tt) U i
O «!-
3 -C to .O
TJ 4J 10
tt) !- -r- +->
o; $ o to
I O) r O>
C T- C T
O tt)Olf-O tt)O<4-
CM oiin CM otin
r C CO i CM CCOr
1 IO 1 i 1 IO 1 r
O-I->O IOO-MO IO
oo ir> Sin* in $
CO i
a -i-
CD O
.
tt) i
S- tt) CL. X
E 30
to ro CO Z
IO r-
om- TJ i
C IO
r 10 tt) tt)
«*- E J- ^=
r- 3 +J
TJ $- +->
tt) Q. IO CO
i S- tt)
U CO tt) CO
>» tt) 0. 10
u u E a>
0) 3 tt) l-
O£. TJ 4-> Q.
'
a: o
to +3
Ly r~*
3
Q) O
J- O
CM in
i i
*r , O)
*->!- co o> o
(- S. U
c TJ tt) tt) a.
0 C i C
r- «O -r- tt) C
+J O O
r ! r"" nj
r U IO 4-> >
O. t- ! L. i.
Q. tt) L. fO 0)
> CO
3 -M 31
!-> CO
tt)r- CO" TJ 3 CO
C -r- i. O) TJ i-
r- 4-> tt) -M C r- O)
3 ^ E *~ T- r
O C O -i- C J. O
QC T- J3 r- T- 4J ^
3t CO
1 o to
er- tt) TJ
U 0)
>>O ra tt) ro
-CO a)
to <4- C
TJ > C O O
tt) CO O CO)
C C ! tO ! C
1- O CO r !-
3 JD CO tt) «"r
JD J- T- > t. 3
C ro E tt) -r- o
=> U 0) r (Q (f.
E_TOECETOE_i>E_«Ec
Q--l-> O.S- Q.4J Q-t. Q.4J Q.S.
C -r- C T- C !-
O tt)O»f-O tt)O«4-O O)Ot)_
in o>in in o>m in o*^
CM C CO i CM CCOr ^J- C<£>r
O+JO o'io
o o 5o*-o 2 in in 2
CM CO^ CM CO* CO *3-^
CO r 1o
(O -r- O
C5 0 <_>
JC C
O>-r- OJ
3 C
Of- 0
i. OJ N
4-» tt) OJ
r E
i ro
O CMr
+J
C TJ >>
0 OJ S-
O U fO
O 4) T
z s- a.
S-
"r~
"i
CO
CO
OJ
o
X 0>
LLJ C
r"
f S»
O "r~
-C
o
ve effects
T3
TJ
^j
(0
tt)
C
rO
to
CO
OJ
IO
i-
tt)
c
tt)
o>
CO
to
c
0
+->
IO
c_>
I.
OJ
.0
o
JC
4J
*o
i.
+J
c
0
o
X
0
0)
-C
+*
TJ
0)
TJ
3
'o
X
tt) *0>
t- C
ro -r-
CO
C TJ
O tt>
(O Q.
r- IO
r- CO
TJ C
O O
+J
TJ (O
tt) U
C *r
.5 -r-
E TJ
O O
0 E
J-
tt)
Q.
O
T3
C
(O
IO -t->
C TJ
r- OJ C
§!-> O
S- -r-
C O +->
QL IO
tt) tt) U
ro £
to -i-
TJ tt) TJ
co 1.
to tt)
r- TJ
S-
r- o
2 z:
u
4-4
-------
CO
o
00
CO
O
CO
O
O
O OJ
UJ <1)
CD M-
Z 0)
CVJ
o
C rJ c?«
o **^ co
P- CVJ
CO O 4->
CO Z 03
til c Q.
Q.
.0
03
8
JO
0
jQ
CO
CD
o»
Q.
>»
r-
cn
c
p-
S-
r-
U_
^
r>. o
i LO
LO CO
i i
CO CO
LO CVJ
CVJ «tf-
N-rf"
,
CO 0
O LO
CO LO
II
1
0 O
«3- LO
r CVJ
s~+
r-. o
LO O
co r>s
i i
CO O
CO CO
r_
^^
P
r
03
3
4-*
c
o
i.
u_
r
f
03
3
^
LO O
** o
LO en
, ,
CVJ O
«a- o
CVJ «tf-
v_^
^-^
CO O
CVJ LO
CVJ O
r- 0
r CVJ
**,
«5j- O
CO LO
«3* en
«sl- O
o o
CM «d-
+^*
r t
r
o3
"^g
o
0)
CO
O
o.
Q.
O
r- 0*
vo O
CO CO
1 1
O O
CO O
i CO
v_««
. ,
CO O
en LO
r CO
i i
CO O
LO O
^^
CO O
r- LO
r CO
1 1
r- O
LO O
r-"
V_X
p
03
r
4J
C
0)
C7)
C
03
H-
O
LO O
CM CM
r-* r
1 1
CO O
o o
CO LO
*~^
^-^
o o
CO O
CM LO
1 1
CO O
CO O
r CO
1
1
0)
c
o
r
o
>)
o
,_
3
M-
o
0)
03
m
c
p-
4-*
J
S-
0)
g"
CD
1
-C
c
o
o
0)
CO
03
<*^
4_>
3
Q.
C
p
4J
03
0)
OC
""^
CO
*-
a
03
-o
03
4-1
CO
CO
0
o
CO
E
"s,^
r-^
s:
II
CO
03
cu
, ,
CO
OJ
3
P
03
4->
03
r
rc
13
^^
jQ
r
s*^
3
1^
CO
O
o
o
CO
C7>
-^
CM
II
O
CO
O
J3
3
CO
O
0
0
«
CO
2
o
CO
II
'to
o
o
CO
o
c
^
E
3
4->
p
rr)
CO
OJ
O
03
C.
E
O
+>
-M
O
d)
CO
c
o *~*
CO Q.
CO Q.
'i o
OJ O
LO
X r-
CO
O)
o
03
E
O
O
X)
O
o
en
CO
03
0)
CO
03
C
03
O
4-3
-------
removal thereby decreasing thermal NOX (Reference 4-2). If still necessary, these minor modifications
are followed by the major retrofits, including overfire air ports (OFA), flue gas recirculation (FGR),
and new burners.
At present, more of these stages of control have been implemented for control of thermal NO
than for fuel NOX control. That is, FGR and staging are now commonly found in gas and oil-fired
boilers, while coal-fired units are only just now entering the second stage of control.
The feasibility, effectiveness, and application technique of the modifications within each
stage of control depend heavily on the fuel and firing type. For example, testing has shown that
FGR does not significantly reduce fuel N0x, so this technique is usually not cost-effective for
coal-fired units. Also, such techniques as BOOS or OFA are implemented differently on wall-fired
than on tangentially-fired units due to burner configuration and hardware differences.
The practical limits on the modifications are based initially on three subjective criteria:
emission of other pollutants (i.e., CO, smoke, and carbon in flyash), onset of slagging or fouling
and incipience of flame instability at the burner. When problems are encountered, implementation
is halted and the situation reevaluated. Stack gas sampling for NO , CO, and 0- is usually car-
A £
ried out concurrently during compliance tests. In the long term, the effects of the modification
on such factors as burner condition, furnace slagging and corrosion, ability to change fuels, and
boiler load are monitored to varying degrees.
The remainder of this section describes recent combustion modification experience on gas,
oil and coal-fired boilers.
Gas-Fired Boilers
The highest degree of success in reducing N0x by the application of combustion modifications
has been obtained on gas firing. The reason for this effectiveness lies in the fact that all of
these techniques reduce thermal NOX, which is the only NOX formation mechanism in gas combustion.
Low excess air operation has been shown to be extremely effective in lowering NO emissions
from gas-fired boilers. An extensive 1971 study of NOX reduction techniques applied to six wall-
fired utility boilers showed reductions of 25 to 60 percent at full load. The NO reduction mag-
nitude depends not only on final excess air level, but also on furnace design and firing method
(Reference 4-3).
x
4-6
-------
o
O)
o
u
o
o
O
<:
Q
O
00
CO
o
CM
*!
CO
«t
cu
u
1
£
TJ
CU
c co
TO TO
i~~ CU
CX >~
co tr>
o 4J
4-> CU
t\ *^
r- i.
a.
^
CO
en o
i +*
+* TO
CO U
X r^
LU Q.
Q.
^
CO
c
o
£
TO
.|_>
'i
Jj
xo
O CM
Z O
^^3*P
co ro
CU
TO
cx: co
C CO
O TO
r- CO
CO
CO CM
c 1^
c ^-
LU
C
o
^
4J
TO
CU
Q.
O
O
CU
"a.
o
c
si
Q.
cu
cr
c
o
cu
»
00
10 r"
sj- sf
cu
c
o
cu
o
z
»-> 1
TO £
cx cu
si ^
3 i_ C
TJ CU CU
aj.c i-
OH +J 0
TO
4J
TO
o
73
^"^
"'""
J
J- 3
i- o
TO r CU
t*
Reduced combustion
temperature yields
er flame temperatu
and lower N(L
1
0.
J-
f^
^
2 TO
O CU
_J IZT
to r~s
i i
cu
c
o
cu
0
TO CU U
E i to > cu
i- TO E CU «-»
CU « i- CU C
-C >> CU r -r-
4-> o cx .a ^:
c o o cn s-
TJ CU 1- i CU
CU ! CU Q--C »->
U U S- TO
3 -i- CU C -C 2 C
TJ ««- > O 4-> O
CU <4- CU !-!- ««- T-
a: cu co 4J s o +>
TO
TO
cu
+->
1
,
TO
i~ 1
cu
CX 3
cu cu
4J
CU
p to O
*4- CO CU
OM-
cu cu
o >
3 CU C
TJ I- O
CU 3 -
0£ +-> 10
0
+J
S- U
«'l?
m VXD oo
i i i
£= TJ
O CU
+J -
TO «4-
O
r- CO
r TO
CX CD
CX
TO TJ
C -t->
TJ TO C
CU CU
*J r E
0-^ CX
QJ O -r-
r 3
CU O O"
OO 4-> CU
CU C
CX 0
-M -4->
O TO
+J S-
TJ O 4J
C S- to
TO CX C
i CO E
i- TO CU
O CT>TJ
^
U i
TO »-> cn TO
M i- C O
to E ! -4-> cj
c^ggu
^» . * I*
TO »
C & -i- CU
O "TJ -t->
§CO TO 'r- CO
TO o .a cu
c
0
-M
o
cu
S-
o
00
1
o
n
o cu.
3 TO
cu £
^ cx
TO
0 C C
^ f- O
o xai
J= o cu
<_> Z J-
c
o
r «4->
C 0
§cu
c
>
*£
CU r
4-> TO
C 3 -i-
i- i-
C O CO
OM- 3
r- TD CO
«/> C C S-
3 CD-i- CU
O CO TJ !-
C CU C Q
TJ TO .5
2
CU 4-*
C f-
cu o
S- .. i-
1- CO -M
«4- 1- CU
1 CU S-
TO O C
3: -Q TO
i cu c:
o c o
S.T- 0
4J TO
O CU S- X C
S- I- 4-> O O
!-> CO JD T-
C -C C
O cn o s_ TO
0 3 CJ v- J-
O «4- 3
xs- -4-> cn
O .C - >,-
Z »->«*- jQ «4_
E E E
5L cx cx
cx a. cx
O 0 0
CM in LO
CM CM ^J-
1 1 1
000
if) O*t
r- CM CM
.. .. ,11
CO i TO
TO - O
C3 O O
^^
<*- 1 CU
0 C 3
0 «4-
O) O
C >
i- CO r
X t3 TO
i Ol
o >,-c
CU 4->
1^ -r- «»-
O TO O
c "cu *o x
O 3 1_ O
O «4- +J Z
i_
CU
1- co
3 C
CD CD
X CO
CU CU
Z O
CU
.C
<*-
o
4J
U
CU
H
«4
CU
CU
4J
TJ
CU
^
^
TO
P
CO
CO
cu
>»
TO
i.
CU
c
cu
CO
CO
C
o
-4_>
TO
O
C
TJ
CU
C
"I
0
-C
2
£
-t->
c
o
o
X
o
z
^
4-*
**
cu
TJ
3
U
X
cuo,
TO !-
CO
CO
O OJ
M F^
TO CX
U CX
i TO
i CO
TJ C
0 0
4->
TJ TO
CU 0
<4-
Si
-(->
TO
TJ
CU
CU
O)
TO
CU
>
TO
i-
O
TO
C TJ
r- CU
§+>
I_
C O
a.
cucu
re
CO
TO
S- C
TO
c .c
O 4->
4-5
-------
HUU
E
n.
n.
evj
O
oo 300
CO
L_J
CJ
X
LLJ
f-
LU
0
5 200
a.
CO*
CO
00
^
g 100
X
O
0
[] NORMAL OPERATION
[2 OVERFIREAIR
^ FLUE GAS
i***i
RECIRCULATION
EPA STANDARD
FOR NEW
GAS-FIRED
BOILERS
__,
11
11
1
i
*ww
_
«---»
il
___
mm
"Is!
i
|
P
is
11
73 78 82 105 110 121 160 160 160 180
FURNACE SIZE (ELECTRICAL), MW
230
250
418 550
Figure 4-1. NOX emissions from gas, tangentially-fired utility boilers (Reference 4-6).
4-8
-------
In 1972, a West Coast utility obtained a 23 percent NOX reduction on a 750 MW* horizontally
opposed unit as a result of lowering excess air. Off-stoichiometric firing and flue gas recircula-
tion were subsequently implemented on this unit to achieve over 50 percent further reduction
(Reference 4-4). In other early work a 33 percent NOX reduction on a 250 MW* tangentially-fired
utility boiler was obtained when the flue gas oxygen content was decreased from 3.9 percent to 0.6
percent (Reference 4-5). In most cases, LEA was implemented without serious flame stability pro-
blems, and a slight increase in thermal efficiency was noted.
From both full and pilot scale results, flue gas recirculation (F6R) has been proven effec-
tive for lowering NOX formation from gas combustion. In general, NOX reduction figures range from
20 to 60 percent for various boiler designs and load conditions. NOX reductions are substantial up
to 20 percent FGR; further recirculation yields only marginal additional reduction. Subscale test-
ing has shown that the magnitude of NOX reduction is mainly dependent on the amount of gas recircu-
lated up to the point of incipient flame instability and other undesirable operating conditions.
Some references which may be scanned for further details on FGR are References 4-3, 4-6, 4-7, and
4-8.
With gas firing, off-stoichiometric combustion (OSC) has been shown to be one of the most
effective means of N0x control and also one of the easiest to implement. Biased firing and burners
out of service (BOOS) are the most frequently used and most effective OSC methods. Overfire air
port operation achieves less reduction, particularly where biasing has already been implemented.
N0y reduction figures of 25 to 58 percent were obtained on wall-fired utility boilers ranging from
80 to 480 MW* when two-stage combustion was applied (Reference 4-3). Similar results were reported
for gas-fired boilers in Southern California. The effectiveness of off-stoichiometric combustion
with gas firing has been well validated and documented (References 4-3, 4-6, and 4-9).
The effectiveness of overfire air and flue gas recirculation on existing gas, tangentially-
fired boilers is shown graphically in Figure 4-1. The bar chart shows the variation in uncontrolled
emissions even among boilers of the same capacity and design. Emissions from five of the boilers
that exceeded the EPA's standard of performance for new gas-fired sources, 86 ng N02/J (0.2 lb/106Btu)
were reduced below the standard by either overfire air or FGR. The data also show a trend toward
higher emissions in larger units. This is attributed to the increase in thermal NO formation due
to the higher temperatures resulting from the higher volumetric combustion intensity used in larger
boilers (Reference 4-6).
*
electrical output rating
4-7
-------
1600
1400
1200
oc
a
u
cc
1000
800
600
400
200
O ORIGINAL FIRING METHOD
O REDUCED EXCESS AIR FIRING
A TWO STAGE COMBUSTION
OTWO STAGE COMBUSTION PLUS GAS
RECIRCULATION THROUGH BURNERS
200 400 600
LOAD (ELECTRICAL), MW
800
1000
Figure 4-2. Effects of NOX control methods on a gas, wall-fired utility boiler
(Reference 4-4).
4-10
-------
Figure 4-2 shows results from the modification of a 750 MW*, horizontally-opposed, wall-fired
unit firing gas. At full load, a 90 percent NO reduction was obtained using the combination of
staging and FGR. These results also show that load reduction is effective for NO control, but is
A
not favored because of economic considerations (Reference 4-4). It should be noted that this 90
percent reduction is more a result of the extremely high uncontrolled NO emissions (1400 ppm) for
this unit, than of any special control procedures used. For units with more moderate uncontrolled
emissions, a 50 to 60 percent reduction is usually the upper limit.
Water injection into gas-fired utility boilers has been tested to a limited extent. A 50
percent maximum NOX reduction was demonstrated at full load for a 250 MW* tangentially-fired unit
when water was injected at a rate of 20 kg per GJ (45 lb/106 Btu) fired. Boiler convective section
temperature increased by 139K (250F) and boiler efficiency dropped 5 percent. The economic penal-
ties resulting from this method, as from reduced load or air preheat, make such techniques unattrac-
tive (Reference 4-5).
A significant amount of work has been done on optimizing gas burner design for low NO pro-
duction. Of the three types of burners - spud, radial spud, and ring - the latter produce the
least NO , while the spud type yields the highest NO formation. In addition, burners which pro-
/» X
duce controlled, low turbulence, flames have been found to form lower quantities of NO .
In summary, the effectiveness of NO controls for gas-fired utility boilers has been ade-
quately explored. Further investigation may not be warranted because the number of large gas-fired
utility and industrial boilers, small to begin with, is now declining rapidly due to the present
natural gas supply shortage. For example, several West Coast utilities, the nation's largest users
of gas for this purpose, reduced from about 70 percent gas-firing in 1972 to less than 10 percent in
1974. Low sulfur residual oil is the predominant fuel to replace gas on the West Coast.
Oil-Fired Boilers
Compared to gas-fired boilers, a generally poorer record of NO reduction has been compiled
for oil-fired units. This is due largely to reduced operating flexibility. When firing residual
oil, fuel NOX becomes an important contribution to the total NOX emission from a given unit, and the
individual modifications are less effective and more complicated to implement. Nevertheless, sub-
stantial reductions have been achieved, in some instances as high as 60 percent on utility boilers.
The most popular N0y reduction techniques for both new and existing oil-fired boilers include
overfire air ports, BOOS, flue gas recirculation, and combinations of these techniques. Lowering
*
electrical output rating
4-9
-------
600
500
400
u
X
u
cc
C/5
00
cc
a
X
o
300
230
200
100
NORMAL OPERATION
OVERFIRE AIR
GASRECIRCULATION
GAS RECIRCULATION AND
LOW EXCESS AIR
EPA STANDARD
FOR NEW
OIL-FIRED
BOILERS
71 78 79 84 84 89 160 160 160 180 230 289 300 378 380 400
FURNACE SIZE (ELECTRICAL), Mw
Figure 4-3. NOX emissions from residual oil, tangentially-fired utility boilers
(Reference 4-6).
4-12
-------
excess air is now considered a routine operating procedure and is incorporated in all new units.
Overall response to the control techniques among boilers, even of the same size and design, can
differ significantly.
For wall-fired units use of overfire air ports alone results in NO reductions of about 15 to
20 percent. For both wall- and tangentially-fired units, BOOS is implemented by removing from ser-
vice several burners in the upper part of the firing pattern. This technique results in NO reduc-
tions of 25 to 35 percent. Flue gas recirculation, in which 15 to 25 percent of the combustion air
is recirculated flue gas, has given NOX reductions of 10 to 45 percent. However, control effective-
ness is usually extended when F6R is combined with the other techniques. BOOS and F6R give total
N0y reductions of 40 to 60 percent, although derating is sometimes necessary to reach these levels
(References 4-3 and 4-10).
The combined use of overfire air and BOOS operation reduces NO only marginally. Smoke thres-
holds are higher and excess air levels must be slightly increased, which cancels the effects of the
overfire air. However, if boiler load is reduced, lower first stage stoichiometries are permitted
and further N0x reductions are achieved with BOOS and OFA.
The effectiveness of overfire air, flue gas recirculation, and their combination on existing
oil, tangentially-fired boilers is shown graphically in Figure 4-3. Emissions from four of the
boilers that exceeded the EPA's performance standard for new oil-fired sources, 129 ng N02/J,
(0.3 lb/106 Btu) were reduced below the standard by use of either overfire air or FGR.
Figure 4-3 also shows the influence of fuel NOV on the total NO production. Unlike the data
* x
for gas-fired units shown previously in Figure 4-1, there is no discernible trend toward higher
emissions from larger units. Apparently, emissions are largely dependent on fuel nitrogen content.
For example, the 160 MW (electrical) unit with an emission rate of 600 ppm was fired with a high
nitrogen (1 percent) California residual oil, while the other two 160 MW units used oil with a nitro-
gen content of only 0.3 percent. The 45 percent difference in emissions can be attributed to higher
fuel NOX formation. In addition, the figure shows that FGR reduced total NO from the oil-fired
boilers by only 30 percent, compared to 70 percent from gas-fired units. This is because FGR
reduces thermal NO but is relatively ineffective on fuel NO .
** x
Figure 4-4 shows results with oil firing from modifying the same wall-fired unit depicted in
Figure 4-2. The combination of staging and FGR produced only a 50 percent NOX reduction, compared
to 90 percent on gas. Again, this difference is attributed mainly to the influence of fuel NO
(Reference 4-4).
4-11
-------
Some experimental work has been performed with injecting water into the combustion air of oil-
fired boilers. Spraying water at a rate of about 0.6 kg per kg of oil reduced emissions about 40
percent. The effectiveness of FGR is increased when combined with water spraying, but the latter
increases the minimum excess air requirement while FGR alone does not. In addition, the energy loss
is significantly greater for water injection as compared to FGR to obtain equal NO reduction. For
these reasons, water injection is not a popular NOX control method.
Operational problems associated with NOX control techniques on some oil-fired boilers include
degraded flame detection, flame instability, boiler vibration, and limited load capability. Combin-
ing FGR with BOOS has in some cases made existing flame scanning equipment inadequate (Reference 4-6).
BOOS and FGR can also cause flame instability. Increasing the fuel flow through the burners left in
service causes significant changes in flame quality and stability. Flame stability is further de-
graded by the increased burner throat velocities resulting from the addition of flue gas recircula-
tion. These factors have been largely responsible for flame pulsations that cause boiler vibration
in some units using large rates of gas recirculation.
Limited load capability can result from the retrofit application of BOOS and FGR due to the
limited fuel/air handling capacity of existing burners and distribution equipment. Load reductions
of 10 percent have been experienced with burner modifications. Additional capacity requirements in
the form of a forced draft fan are also imposed by FGR.
There are several subtle factors that influence NOX emissions regulation compliance. Among
them are operational flexibility, fuel properties, and boiler cleanliness (Reference 4-2). Since
boiler operating conditions are variable, the chosen low NOX operating mode must be flexible enough
to allow some latitude during periods of adverse operating conditions. Equipment problems may occur
somewhat more frequently due the the fine tuning needed for NO control (Reference 4-11).
Variations in fuel supply are the second important factor influencing regulation compliance.
Residual oil fuel nitrogen content can vary between 0.2 and 1.0 percent. Typically, the conversion
of fuel nitrogen is in the range of 20 to 40 ppm NOX per 0.1 percent fuel nitrogen. Other fuel oil
properties influence operating conditions such as smoke threshold, atomization characteristics and
excess air level for stable combustion.
Boiler cleanliness appears to be another important factor influencing NO emissions. Indica-
tions are tha. up to a 50 ppm increase in NOX emissions can be attributed to furnace deposits in the
radiant section. This is attributable to higher flame temperatures needed as a result of the low
radiant heat transfer condition incurred with deposits (Reference 4-2).
4-14
-------
400
300
cc
a
r>g
o
o
oc
100
O ORIGINAL FIRING METHOD
O TWO STAGE COMBUSTION
A TWO STAGE COMBUSTION PLUS GAS
RECIRCULATION THROUGH BURNERS
200
400 600
LOAD (ELECTRICAL),MW
800
Figure 4-4. Effects of NOX control methods on an oil, wall-fired utility
boiler (Reference 4-4).
4-13
-------
In field tests, existing wall-fired boilers under full load baseline operation generally pro-
duced NO emissions that exceeded the NSPS for new boilers. However, under modified operation using
low excess air and staged firing, NOX was reduced about 20 to 40 percent and the boilers were able
to meet the new unit standard. Additional reductions were possible in some cases when the load was
reduced about 20 percent. One 270 MW (electrical) wall-fired boiler, after being fitted with a
specially-designed "low NOX" burner, obtained a NOX reduction of 35 percent (Reference 4-12).
Flue gas recirculation has also been tested on a wall-fired boiler (Reference 4-13). Apply-
ing 15 percent windbox FGR to a 560 MW (electrical) unit resulted in a 13 to 17 percent NOX reduc-
tion under normal, air-rich, conditions. When applied in conjunction with OFA, FGR yielded a 7 per-
cent NO reduction to augment the 33 percent reduction from OFA alone. The reduction obtained from
FGR alone is, however, less than half of that normally obtained on oil firing. It appears that,
due to the influence of fuel NOX formation, FGR is generally less effective for coal firing than
for gas and oil firing. FGR may be justified if minor emission reductions (i.e., "trimming") are
necessary to achieve compliance with a stringent emission standard for existing units.
Under baseline conditions, tangentially-fired boilers usually emit less NOX than wall-fired
boilers. For example, of the 16 units shown in Figure 4-5, 10 meet the NSPS with no modifications.
For those that do not meet the regulation, overfire air, burner staging, and low excess air tech-
niques can be used to reduce NOV by an average of 40 percent.
A
A measure of the degree of control implemented with off-stoichiometric combustion is given
by the value of the first stage burner stoichiometry. Figure 4-6 shows the relationship of NOX
reduction with stoichiometry to the active burners for 13 tangential, coal-fired boilers (References
4-3, 4-12, 4-14, 4-15, 4-16, and 4-17). The best-fit line illustrates that NOX emissions in general
are reduced approximately 140 ppm for each 10 percent reduction in burner stoichiometry. It was
also discovered that an optimum burner tilt angle exists from a NOX formation standpoint. Hori-
zontal burner operation reduced NO emissions by 18 percent, while lowering burner tilt to -26
percent increased NO emissions to 9 percent above baseline operation (Reference 4-12).
Several utility boiler tests have been conducted using the combined firing of gas/coal and
oil/coal as a NOX reduction strategy. Tests on the former mixture were conducted on a 130 MW
(electrical) tangential unit. Firing with 80 percent of the heat release from coal reduced NO by
an average of 30 percent, while firing with 60 percent coal and 40 percent gas resulted in a 32
percent NOX reduction from 100 percent coal firing. The data indicate that replacing coal with gas
fuel lowers NO in the direction of 100 percent gas firing, but the relationship is not linear.
4-16
-------
In general, an estimate of the boiler's actual operating conditions should be made in order
to assess all factors that may influence regulation compliance. It is best to have at least a 25
to 35 ppm margin under average conditions (References 4-2, 4-3, 4-4, and 4-10).
Coal-Fired Boilers
The retrofit implementation of NOX controls on coal-fired boilers is currently less wide-
spread than on gas- and oil-fired units. Due to the continuing clean fuels supply shortage, however,
the development of control for coal-fired units is receiving primary emphasis in the research and
development programs of the Environmental Protection Agency's Industrial Environmental Research
Laboratory at Research Triangle Park. Major developmental activity to date has been focused on
achieving the level of control for new units mandated by the Standards of Performance for New Sta-
tionary Sources - 300 ng N02/J (0.7 lb/10* Btu). By contrast, the major activity for gas and oil-
fired units has been on retrofit compliance with emission standards for existing units in N0,,-sensi-
tive Air Quality Control Regions. Nearly all new utility boilers currently being installed, or on
order, in the U.S. are designed to use coal as the primary fuel.
The combustion characteristics of a solid fuel such as coal are vastly different from either
gas or oil, and the NOX control strategy varies accordingly. For the cleaner fuels, thermal NO
formation mechanism dominates; certain sets of combustion modification methods have been found to be
well suited to suppress this problem. For coal, however, up to 80 percent of total oxides of nitro-
gen comes from fuel-bound nitrogen. Researchers have found that combustion modification methods
that were effective on gas and oil firing either do not work as well or must be applied differently
on coal-fired boilers. In addition, operational problems associated with the modifications, such
as slagging, fouling, and carbon burnout, are more pronounced.
The most extensive series of tests performed on coal-fired boilers has been sponsored by the
EPA and the Electric Power Research Institute. The combustion modification methods tested include
lowering excess air, off-stoichiometric combustion, biased firing, a "low N0x" burner, and flue gas
recirculation. Coals tested include both Eastern and Western bituminous and Western sub-bituminous.
The sequence of combustion modification implementation is similar to that for gas and oil
firing. First, the boiler is fine tuned by minimizing excess air to the threshold of excessive CO
and unburned hydrocarbon formation. If the NOX reduction obtained by such a procedure is inadequate,
off-stoichiometric techniques, such as biased firing and burners on air only, may be utilized. The
hardware retrofit methods are the last to be used. These include overfire air ports and "low NO "
burners.
4-15
-------
700
600
500
CM
o
o
X
400
o
cc
£2
CO
a 300
x
o
200
100
60
80 100 120 140
STOICHIOMETRY TO ACTIVE BURNERS, percent
160
Figure 4-6. Effect of burner stoichiometry on NOX production in tangential, coal-fired boilers.
4-18
-------
II NORMAL OPERATION
TOP ELEVATION NOT FIRING - NO OVERFIRE AIR
TOP ELEVATION FIRING - OVERFIRE AIR
EPA STANDARD
FOR NEW COAL
FIRED BOILERS
52
100 110 122 170 206
(80)* (157)*
215
(158)
250 250 265 370 375 378 426 485 565
FURNACE SIZE (ELECTRICAL), MW
*REDUCED RATING WHEN TOP ELEVATION NOT FIRING
Figure 4-5. NOX emissions from tangential, coal-fired utility boilers (Reference 4-6).
4-17
-------
TABLE 4-3. MAJOR JAPANESE DRY FGT INSTALLATIONS
(Selective Catalytic Reduction) (Reference 4-21)
Process
Developer
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Sumitomo Chemical
Hitachi Shipbuilding
Hitachi Shipbuilding
Hitachi Shipbuilding
Tokyo Electric-
Mitsubishi H.I.
Kurabo
Kurabo
Kansai Electric-
Hitachi Ltd.
Chubu-IHI-Mitsui Toatsu
Chubu-MKK
Mitsubishi H.I.
Kobe Steel
Mitsui Toatsu
Mitsui Toatsu
Mitsui Toatsu
Mitsui Toatsu
Hitachi Ltd. -
Mitsubishi P.C.
Hitachi Ltd.
Ube Industries
Mitsui S.B. - M1tsu1 P.C.
Mitsui S.B. - Mitsui P.C.
MKK - Santetsu
MKK - Santetsu
MKK - Santetsu
Seltetsu Kagaku
Japan Gasoline
Japan Gasoline
Asahi Glass
Capacity
(Nm'/hr)
30,000
200,000*
250,000*
100,000*
200,000*
200,000*
250,000
300,000
5,000
350,000
440,000
10,000*
5,000
30,000
4,000
8,000
100
4,000
600
1 ,000*
3,000
4,000*
8.000*
150,000
350,000
10,000
200,000
240,000
1,500
1,000
15,000
15,000
50,000
70,000
70,000
Source of Effluent
011 -fired boiler
Heating furnace
Heating furnace
Gas-fired boiler
Gas-fired boiler
Heating furnace
Oil -fired boiler
Oil-fired boiler
Oil-fired boiler
CO-fired boiler
Oil-fired boiler
Gas -fired boiler
011 -fired boiler
Oil -fired boiler
Oil-fired boiler
Oil-fired boiler
011 -fired boiler
Oil-fired boiler
Sintering plant
Gas-fired boiler
Oil -fired boiler
Gas-fired boiler
Gas-fired boiler
Oil-fired boiler
Coke oven
Oil -fired boiler
011 -fired boiler
Oil -fired boiler
011 -fired boiler
Coke oven
011 -fired boiler
011 -fired boiler
Heating furnace
CO boiler
Glass furnace
Completion
Date
Jul 1973
May 1974
Mar 1975
Feb 1975
Feb 1975
Mar 1975
May 1976
May 1976
Nov 1973
Nov 1975
Nov 1975
Jan 1974
Nov 1973
Aug 1975
Jan 1975
Oct 1974
Oct 1974
Dec 1974
May 1974
Oct 1973
Oct 1974
Oct 1974
Jun 1974
Dec 1975
Oct 1976
Jan 1975
Sep 1975
Aug 1976
Dec 1974
Mar 1975
Jun 1976
Jun 1975
Nov 1975
Mar 1976
Apr 1976
*C1ean gas; others are for dirty gas
4-20
-------
Similar results were obtained with the oil/coal mixed fuel. Further NOX reductions were possible
when mixed fuel firing was combined with techniques such as lowering excess air, off-stoichiometric
combustion, and reduced load (Reference 4-12).
Throughout the developmental field tests, attention was given to the potential side effects
of low NOX operation. Excessive smoke and CO levels generally limit the extent to which the burners
are fired fuel-rich. The fuel-rich conditions can lead to flame instability, and the reducing atmos-
phere in the primary combustion zone can accelerate tube corrosion and slagging (Reference 4-18).
One utility company reported experience with retrofit biased firing on a coal-fired boiler. The
problems included increases in carbon losses, decreases in boiler efficiency of about one percent-
age point at all load levels, and increases in tube wastage on the sidewall near the biased burners
(Reference 4-19). The EPA is conducting long-term field tests to determine the extent to which OSC
accelerates tube wastage. Corrective measures to suppress tube wastage are also being examined.
One utility boiler manufacturer uses a "curtain air" oxidizing atmosphere at the tube walls to sup-
press wastage (Reference 4-20).
4.1.1.2 Flue Gas Treatment
The major NOX control emphasis in the United States has been on process modification since
it permits the construction of new equipment that can meet existing emission standards. Due mainly
to economic penalties, a less intensive effort has been devoted to removing nitrogen oxides direct-
ly from flue gases. However, faced with emission standards 20 to 40 percent more stringent than
those in the U.S., Japanese industry has been much more interested in flue gas treatment (FGT) and
has several major pilot plants and full-scale plants in operation. The major application of flue
gas treatment in Japan has been to utility boilers and the larger combustion and noncombustion
industrial point sources of NOX. More inexpensive alternatives, such as process modifications,
will continue to be used for smaller combustion sources of NOX, although regulations could eventu-
ally require the use of FGT as well.
As described in Section 3.2, the two FGT process routes can be categorized as dry processes
(reduction) and wet processes (oxidation followed by scrubbing). In Japan, dry processes are the
more popular of the two types, and these usually involve the use of the selective catalytic reduc-
tion process (SCR). Table 4-3 shows the SCR processes on conmercial and pilot plants in operation
or under construction in Japan.
Most of the larger installations treat only "clean" exhaust gas from the combustion of
gaseous fuels. However, two large plants being constructed by Sumitomo Chemical and Hitachi
4-19
-------
TABLE 4-4. MAJOR JAPANESE WET FGT INSTALLATIONS (Reference 4-21)
Process Developer
Sumitomo Metal
Fuji Kasul (Moretana)
Sumitomo Metal
Fuji Kasui (Moretana)
Sumitomo Metal
Fuji Kasui {Moretana)
Chiyoda
Osaka Soda
Shirogane
Mitsubishi H.I.
Ishikawajima H.I.
Tokyo Electric
Mitsubishi H.I.
Tokyo Electric
Mitsubishi H.I.
Kawasaki H.I.
Mitsubishi Metal
MKK, Nihon Chem.
Kobe Steel
Kobe Steel
Hissan Engineering
Hissan Engineering
Hodogaya
Chisso Corporation
M1tsu1 S.B.
Asa hi Chemical
Kureha Chemical
Type of Process
Redox
Redox
Redox
Redox (102 process)
Redox
Redox
Redox
Redox
Oxidation
absorption
Oxidation
absorption
Oxidation
absorption
Absorption
oxidation
Absorption
oxidation
Absorption
oxidation
Absorption
oxidation
Absorption
oxidation
Absorption
oxidation
Reduction
Reduction
Reduction
Reduction
Capacity,
NmVnr
62,000
100,000
39,000
1,000
60,000
48,000
2,000
5,000
2,000
100,000
5,000
4,000
1.000
50,000
1,800
3,000
4,000
300
150
600
5,000
Source of Effluent
011 -fired boiler
Metal heating furnace
01l-f1red boiler
01l-f1red boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Gas-fired boiler
Gas -fired boiler
Coal -fired boiler
Oil -fired boiler
Iron ore sintering furnace
Iron ore sintering furnace
Pickling
HNOa plant
Iron ore sintering furnace
011 -fired boiler
011 -fired boiler
011 -fired boiler
011 -fired boiler
Completion
Date
Dec 1973
Dec 1974
Dec 1974
1973
Mar 1976
Aug 1974
Dec 1974
Sep 1975
Dec 1973
Oct 1974
Dec 1975
Dec 1974
Dec 1973
Mar 1976
Jul 1973
Mar 1975
Oct 1975
1974
1974
1974
Apr 1975
By-product
NaN03, Nad
Na2S04
NaN03, NaCI
Na£S04
NaN03. NaCI
Na2S04
Gypsum,
Ca(N03)2
NaNOs, NaCI
Na2S04
Na2S04,
NaN03
Gypsum, N2
Gypsum, N2
HN03
HN03
Gypsum,
Ca(N03)2
KN03
Gypsum, N2
Gypsum, N2
NaN02
NaN03, NO
NaN03, NaCI
(NH4)2S04
H2S04, N2
Gypsum, N£
Na2S04, N2
4-22
-------
Shipbuilding companies will treat "dirty" gas (containing S0x and participate) from oil-fired sources
A pilot plant treating dirty gas (oil-fired) has been operated by Sumitomo for over 4,000 hours
reportedly without serious problems. Electrostatic precipitators remove dust and prevent contamina-
tion of the catalyst. More than 85 percent NOX removal has been achieved (Reference 4-21).
Several EPA contractors have investigated SCR on both the lab and pilot scales. Emphasis was
on technical and economic assessment of various catalytic processes, using both noble and nonnoble
catalyst systems. These tests achieved NOX reduction of 60 to 95 percent at inlet concentrations
of 250 to 1000 ppm. They also indicated that platinum is not a satisfactory catalyst for flue gases
containing SOp.
Although selective catalytic reduction has been the most widely used and investigated dry FGT
process, selective noncatalytic reduction of NOX using ammonia has also been demonstrated commer-
cially in Japan. NOX reductions of 70 percent have been reported. Investigation of this technique
is underway in the U.S. as well. Barriers to its use on steam generator exhaust include reducing
agent injection problems, load following, byproduct emissions, and high reducing agent use and cost.
The attractiveness of this technique may improve as more accurate assessment of these problems are
developed (References 4-22 and 4-23).
Wet systems generally have not been as popular as dry processes. Major disadvantages of wet
systems are (1) the need for expensive oxidizing agents and/or energy input in proportion to the
quantity of N0x removed', (2) generation of unmarketable byproducts, (3) waste water production, and
(4) requirement of prior S02 removal to reduce the consumption of N0x-removal chemicals. Its poten-
tial major advantage, however, is simultaneous NOX and SO removal.
As with the dry process, most of the research, development and demonstration of wet pro-
cesses has been conducted in Japan (Reference 4-21). In 1975, there were 12 different wet pro-
cesses being developed in Japan at pilot plants and small commercial plants (100 to 25,000 cubic
meters per hour). The largest systems reportedly are treating 32,000 to 1,000,000 cubic meters
per hour of flue gas. No firm data are available as to NOX removal efficiencies, but the range
appears to be from 60 to 90 percent. Table 4-4 lists plants in Japan using wet FGT systems.
Two of these processes, the Chiyoda 102 and the Mitsubishi Heavy Industries Systems, are
relatively simple extensions of well-established flue gas desulfurization (FGD) technology. Both
processes are attractive from the standpoint of simultaneously removing SO and NO . However, both
X X
require the use of ozone. Ozone production for application to coal-fired flue gas is expected to
require up to approximately 10 percent of the power plant electrical output. When added to the
4-21
-------
the use of coal as well. Fuel switching to natural gas or low nitrogen oil is, therefore, not a
promising short-term option.
A promising long-range option is the use of synthetic fuels derived from coal. Candidate
fuels include lower Btu gas (3.7 to 30 MJ/Nm3, or 100 to 900 Btu/scf) and synthetic liquids or
solids. Although these fuels would have all the emission control advantages of conventional clean
fuels, there are several disadvantages associated with them. First, economic considerations favor
the placement of both the gasifier and the power cycle at the coal minehead. This fact eliminates
existing utility boilers without an onsite coal supply as users of lower-Btu gas. Second, the cost
of required equipment modification to lower-Btu gas is high, ranging from $5 to $15/kW*. Third,
although synthetic oil can be transported like conventional oil, the coal conversion process is
highly, and perhaps prohibitively, expensive (Reference 4-24, 4-25).
In general, the feasibility of coal-derived fuel switching is dependent on the cost tradeoff
between the coal conversion route and more conventional means of controlling the criteria pollu-
tants (i.e., gas cleaning). The economics of the former alternative are not well defined at pre-
sent and will not be clear until the ongoing studies and pilot projects are completed (Reference
4-26).
4.1.1.4 Fuel Additives
The basis of this control technique was covered previously in Section 3.1.4. In general, fuel
additives are not effective. Most of the additives that have been tested do not decrease NO
emissions, and some that contain nitrogen actually increase NOX formation. Several additives con-
taining metallic compounds were found to promote the catalytic decomposition of NO and N2. However,
serious operational difficulties, high cost, and the presence of the additive as a pollutant in the
exhaust made these additives unattractive (Reference 4-27, 2-28).
Other fuel additives investigated recently are intended to prevent boiler tube fouling. Use
of these additives could conceivably allow a further decrease in excess air which would reduce NO
formation. However, the emission reduction from this method is quite limited and the cost-effec-
tiveness is likely to be poor (Reference 4-29).
4.1.2 Costs
The cost of implementing the preceding NOX reduction techniques is basically the sum of the
initial capital cost and annual operating cost (which includes any cost savings). The following
*
electrical output
4-24
-------
power requirement of the FGD portion of these processes, this energy consumption may render wet simul-
taneous SO /NO processes impractical for commercial use.
In general, therefore, wet NOX FGT systems cannot compete with dry selective catalytic reduc-
tion where simple NOX control is involved. For coal-fired applications where high dust loadings
and S02 removal are involved, it is not as yet clear whether dry FGT combined with conventional FGD
processes will be cheaper than wet simultaneous SO/NOv systems. Other dry simultaneous SO /NO
AX XX
systems, such as the Shell and the Sumitomo Shipbuilding processes, may also prove to be cheaper
than the wet simultaneous processes. At present, the Shell process is being commercially applied
on a 40 MW (electrical) oil-fired boiler in Japan and is being piloted in the U.S. on a 0.6 MW
(electrical) flue gas stream from a coal-fired boiler. The Sumitomo Shipbuilding process is being
tested on an oil-fired boiler in Japan.
In summary, wet FGT processes are more expensive and less well developed than dry processes.
Considering their cost and complexity, it is doubtful that wet processes would be receiving any
development attention in Japan were it not for the potential for simultaneous SO and NO removal
(Reference 4-65).
4.1.1.3 Fuel Switching
The aim of this technique is to switch the combustion system to a fuel with a reduced nitro-
gen content (to suppress fuel NOX) or to one that burns at a lower temperature (to reduce thermal
N0x). Switching decisions are based on the knowledge that solid fuels generally contain more organ-
ically-bound nitrogen than liquid fuels, and gaseous fuels are usually nitrogen-free. Coal-fired
utility boilers, unless they already have a dual fuel capability, can be converted either to oil or
gas. Likewise, oil-fired boilers can be switched to gas fuel. Due to design limitations, however,
the reverse order of these conversions is generally not practical.
During the 1960's and early 1970's many Eastern and Midwestern U.S. coal-fired utility
boilers were converted to oil and/or gas in response to tightened particulate and S02 emission
standards. This trend was attractive from a NOX control standpoint for two reasons. First, liquid
and gaseous fuel firing provides more flexibility for implementing combustion modification techniques.
Second, fuels containing less sulfur generally contain less nitrogen also, which serves to reduce
fuel NOX.
Despite the superiority of oil- and gas-fired NOX control, the economic considerations in
fuel selection are dominated by the current clean fuel shortage. Existing utility boilers are cur-
rently returning to coal, and the trend for new utility and industrial boilers is strongly toward
4-23
-------
1.00
0.75
10.50
CO
o
o
1.50
1.25
1.00
0.75
CO
O
CJ
A. NEW BOILER INSTALLATION
4WINDBOX FURNACES
8WINDBOX FURNACES
200
400 600
UNIT SIZE (ELECTRICAL), MW
800
1000
0.50
0.25
B. EXISTING BOILER MODIFICATION
400 600
UNIT SIZE (ELECTRICAL), MW
1000
Figure 4-7. 1975 capital cost of overfire air for tangential, coal-fired boilers (Reference 4-27).
4-26
-------
discussion will center on the costs of reducing NOX from utility boilers by combustion modification
and flue gas treatment. In several cases the costs presented are for combined "NOX controls. Gener-
ally, the effectiveness of combined NOX controls is not equal to the sum of the individual effects
of each control. Likewise the cost of combined controls is not the sum of the costs of single con-
trols. The cost of fuel additives is not discussed due to its status as an unattractive option and
to a general lack of cost data. Fuel switching economics are similarly not treated in this dis-
cussion.
4.1.2.1 Combustion Modification
Much of the pioneering work on evaluating the cost effectiveness of combustion modification
in full-scale combustion equipment has been performed on utility boilers. Correspondingly, the
related costs of these modifications have been fairly well documented compared to other source types.
One of the earliest efforts of this kind was attempted by Esso Research Labs in 1969 (Reference
4-30). Based on estimates for the capital, annual, and operating costs, the Esso report presented
the results of a cost effectiveness study performed for NO control on utility boilers by means of
combustion modification. Since 1969, however, it has been revealed that a wide variation in the
effectiveness of the control techniques among boilers exists. This problem will require that con-
tinuing cost-effectiveness evaluations be done on an individual boiler basis.
Data from Combustion Engineering
The most recent cost data were published in Reference 4-31 for new and existing tangential,
coal-fired utility boilers. These data are summarized in Figures 4 - 7a and b. The cost range curves
were derived from estimates developed under an EPA-sponsored contract involving the reduction of NO
from both new and existing tangential coal-fired utility boilers. The costs are for the com-
bined use of overfire air ports and low excess air firing, as this is the preferred control system
for tangential,coal-fired boilers. Capital costs were projected over a unit size range of 25 to
1000 MW*. The corresponding annual operating costs for 500 MW* units was 0.006 mills/kWh for a new
unit and 0.021 mills/kWh for existing units. Figure 4-7a applies to new unit designs with heating
surfaces adjusted to compensate for the resultant changes in heat transfer distribution and rates.
Figure 4-7b applies to existing units with no change in heating surface, as these changes must be
calculated on an individual unit basis.
*
electrical output rating
4-25
-------
Data from the Pacific Gas and Electric Company
As an example of the manner 1n which the costs for combustion modification may vary among
individual existing units, several case studies are presented in Table 4-6. The numbers shown are
the costs incurred by the Pacific Gas and Electric Company during a program to bring six units into
compliance with local NO emission regulations. For the most part, the conversions involved the
combination of windbox flue gas recirculation and overfire air ports. The average cost of the modi-
fications is about $10/kW* (Reference 4-32).
Data from the Los Angeles Department of Water and Power
Another West Coast electric utility company, the Los Angeles Department of Water and Power
(LADWP), has had extensive experience in implementing NOX control techniques on its gas- and oil-fired
boilers. The techniques currently utilized by the Department include burners out of service (BOOS),
overfire air ports, and low excess air. Although the units are operated with the lowest excess air
possible, it has been found that when LEA is combined with other reduction methods, excess air levels
must be increased beyond those normally required.
The Department's data indicate a unit efficiency decrease of approximately one percent attri-
butable to BOOS operation. As has been found by other operators, LEA tended to increase efficiency
slightly: a one percent decrease in excess oxygen increased efficiency by about 0.25 percent. Pro-
perly retrofitted, overfire air had no effect on efficiency.
The NOX control costs incurred by LADWP are shown in Table 4-7 for four different units. The
figures for the BOOS techniques reflect the R&D costs that necessarily precede the retrofit. All
costs include the labor required to implement the control methods, and are, therefore, installed equip-
ment costs. The very low expense associated with overfire air on the B&W 235 MW* unit is due to the
base year of the estimate (1964 to 1965) and to the fact that this modification was included in the
original boiler design.
The overfire air costs for the B&W 235 MW* unit lie in the low range of the appropriate band
of costs in Figure 4-7b. The LADWP boilers were, for the most part, modified without much difficulty,
and the associated costs probably represent the lower limits of the costs for the three NO reduc-
tion techniques implemented (Reference 4-33).
electrical output rating
4-28
-------
It is readily observed that the cost ranges for existing units vary more widely than for new
units. This is due to the variations in unit design and construction which can either hinder or aid
the installation of a given NO control system.
Above approximately 600 MW*. single cell-fired boilers exceed a practical size limit and
divided furnace designs are utilized. Since a divided tangentially-fired furnace has double the
firing corners of a single cell furnace, the costs increase significantly.
It should be kept in mind that although these cost data for utility boilers were developed
for tangentially coal-fired boilers, it is felt that the range of costs presented should also be
applicable to wall-fired boilers burning coal. Additionally, the cost for similar combustion modi-
fication on gas and oil-fired utility boilers should be no higher than for the coal-fired units.
The cost of reducing low excess air was not investigated since there is generally no signi-
ficant additional cost for modern units or units in good condition. However, some older units may
require modifications such as altering the windbox by addition of division plates, separate dampers
and operators, fuel valving, air register operators, instrumentation for fuel and air flow and
automatic combustion controls.
Data from EPA
Table 4-5 shows estimated investment costs for low excess air (LEA) firing on utility boilers
requiring modifications (Reference 4-1). These costs can vary depending on the actual extent of the
required modification and are only provided as guidelines. As unit size increases, the cost per kW
decreases since the larger units typically have inherently greater flexibility and may require less
extensive modification.
The use of low excess air firing reportedly increases boiler efficiency by 0.5 to 5 percent.
Additional savings may result from decreased maintenance and operating costs. Consequently, invest-
ment costs may be offset in fuel and operating expenses.
TABLE 4-5. 1974 ESTIMATED INVESTMENT COSTS FOR LOW EXCESS AIR
FIRING ON EXISTING BOILERS NEEDING MODIFICATIONS
Unit Size
(MW*)
1000
750
500
250
120
Investment Cost
($/kW*)
Gas and Oil
0.12
0.16
0.21
0.33
0.53
Coal
0.48
0.51
0.55
0.64
0.73
electrical output rating
4-27
-------
TABLE 4-7. LOS ANGELES DEPARTMENT OF WATER AND POWER ESTIMATED INSTALLED 1974
CAPITAL COSTS FOR NOX REDUCTION TECHNIQUES ON GAS- AND OIL-FIRED
UTILITY BOILERS (Reference 4-33)
Unit
Capacity
(MW)
180
235
235
350
Unit
Type
C.E.
wall -fired
C.E.
wall -fired
B&W Opposed-
fired
B&W Opposed-
fired
NOX Reduction
Technique
BOOS
LEA
BOOS
LEA
BOOS
Overfire air
LEA
BOOS
Overfire air
LEA
Implementation
Method
Retrofit
Retrofit
Retrofit
Retrofit
Retrof i t
Original Design
Retrofit
Retrofit
Retrofit
Retrofit
Estimated
Cost
($)
67,400
28,900
75,200
28,900
75,000
14,000a
28,900
266,000
100,600
28,900
$/kW
0.38
0.16
0.32
0.12
0.32
0.06
0.12
0.76
0.29
0.08
1964-65 base year
Operating Cost Data
In addition to the increased capital costs from including a NO reduction system in new or
existing units, the increased unit operating costs may be considered. These differential operating
costs were defined for 500 MW (electrical) new and existing units and are shown in Table 4-8
(Reference 4-31). The costs are given in 1975 dollars, and the equipment costs shown are determined
from Figures 4-7a and b. To put these operating costs in perspective, they can be compared to the
percent increase in generating costs shown at the bottom of Table 4-8. Except for the case of older
units, the difference in operating cost is below 0.1 percent of annual cost.
Table 4-9 shows the impact on major system components, efficiency, and capacity when employing
the major combustion modification NOX control techniques. The relative changes in unit design or
efficiency are shown to increase (or require addition) by a plus sign (+) or decrease by a minus sign
(-). If the item is unchanged, or is altered to a negligible extent, it is indicated by a zero (0).
Heat transfer surfaces remain unchanged in all cases (Reference 4-1).
The following are the major economic considerations that the boiler operator or designer may
face:
The lowest cost method for reducing N0y emission levels on new and existing units is
the incorporation of low excess air firing. Minimal additional costs are involved.
For most utility boilers, the second lowest cost NOX control method appears to be staged
combustion by biased firing, "burners out of service" (BOOS) or addition of an overfire
air system. Although lowering excess air (LEA) alone is less expensive than off-stoi-
chiometric combustion, one utility company has found that when LEA is implemented con-
currently with other control techniques, the excess air levels must be increased be-
yond those normally required.
4-30
-------
O
LU
or
o
I
CO
CD
CJ
ti
Lu
<i
CJ
a.
CD
CO «3-
O
CJ O>
-J M
- XJ
5
C
o
*j"^ ^_^
ro 4~* **
O CO O
r- O r
*»- CJ V»
XJ
'ro >>
U 4->
II O 3
j-> m y
<_) CX '
O) ro
> CJ
LU
O>
(O 'P
Q) r
>- C
O
OJ
ex
j
o>
CO
"
i
rO
+j
C
33
CO
C XJ
ro 5-
S- M- x-v C ro
- Q: o 3
CD u_ *-> OJ
OJ LL. ro <4-
J_ ^ ro
i- ex r*. " ">
<*- x: ,_ ro
J- , 4J i-
ai o co oj
> o o> c c
O 0 C -r- $-
in -f- 3
-M -M XI
a: X 0 i.
CD O) 3 O 2 E
LL. c XJ Q. Q> tt)
C -l->
x o o: ,
XI 1 U- O Z CO
XJ
c
3
in
px»
CT.
CO
^
IO
O
co
r-.
CM
r^»
crt
p""
i
r^ (U
ra XJ
r" !"
P >
C -r-
1) -O
a>
c «
t-
LU -r-
0 «4-
cn
^
(/) ^L
-M "^
r-
a.
i.
E 0
O «4-
s_
<4_ XJ CO CJ1 C J-
O OJ +J C t- v- CX 4J S-
ct: a> 3 u ex s. 3
CD «4- 3 0 O XJ
u- co i- xj x: ex
CO) -C
x ro x: cc S cr <->
o i- «-> CD
00
r-
E
CO
x:
o
O)
0
CO
CO
vo
CT.
r^
1
1
TD
0)
CO
o
CX
CX
0
f
00
CO
cn«>
3
CO C
+> ro
r- in
CX =fe
d)
«« ** E
§XJ ro
t. i
+J ro M-
co O
CO CO
1 i-
XJ O D
j- i. ex
rO 4-> E
3 C ra
CT) O XJ
«t- c
rO X O
CO O
Z CO
i. r
a> > o
C 5- 1- CO
$- a> -M t.
3 4-> C OJ
X> 3 O C
CL u c
S. E ro
a> o CMO
Z U O CO
^_^
a: o>
CD C
U- -I-
fr? U
t XJ
J- ^ » C
«t c -
ro 4->
OJ OJ ^ ro
$- O .
««- 4-> ra
J- C -M
OJ CO 3 CO
> C O C
O ro E X -r-
«+- ro O
" X> +J
Qi Q£ r XJ t-
CD O ro C O
U. Li. c -i- Q.
x * i 5 <
O OJ O O | '
X) Z Z +J O
XJ
3 »
in
CT.
co
c-
5
CD
x:
u
ro
C
3O^
=«=
i.
FGR, Overfire
X
XJ
c
3
in
-
-
in
co
CO
CM
CM
cn
^
XJ
«r
1
o
X)
J_
3
>,
f-
.
^
co
e
CD
0
CL
ro
C «
'E E *"
O -M -M
C CO 3
>> CX
* co E
'5 ^, °,
rv* O ^^ ^
g '-13 « .
"~ ro CD 3 X)
^j, i C CD (.
rj r -r- OJ ro
^ ro 4J <4- O
r", »-> 0 ro X)
^^ CO 3 CO
C C XJ T~
ro -r- l-O
i+_ 4 D l_
MO c -»
0£ J. J. C
CD O »-> 3 O
U_ CL C XI U
$ <: o xx
OJ LL. E OJ O
Z 0 ro Z z
. . .
»
r OJ
O CD
4-> ra
O "~
J- 4->
r- OJ -r-
<: s- 2
3
FGR, Overfire
ting temperati
fans replaced
CO CO
X -i- C£ C
O X O ro
jQ LU Lu M-
c
3
-^^
CT.
00
JZ
[ 1
0
X}
"
cc
CM
t
u
"C
1
(1)
(/)
0
ex
0
^£
CD
OlT
c r-»
XJ
C XJ
ro C
CO i
CO 1
"£ =«=
(A
flame scanner;
5
4-29
-------
o
CO
O
o
ce
o
X
o
o>"°
0 + i
0 + + + o 0 + i
+ + + + O O + i
+ + + + 0 0 + |
Ol
o o «t <: o o o o
i.
r- Ol
«- to i_ oj
>> CO MJ fc. 01 > *J
»' O N «^ «O
<» XCO 4-> f- 4-> OJ tt» c» x u. u. QO OQ. oco coco
< <:
z z ' '
^ * 0 I
| | 0 ,
^^
_. -^ o i
«c <:
z z ° °
l + 10
1 + 0 O
+ 00
" + 0 0
o o o
N c >>
+> o *i S $. .2 2
IO 1-3
o «
4-> 01
-M
c:
O *^* g"
r- Q. Ol
O. CT>
^
2 Z o
CO CL*" **
CO CO
01 C
-> 'i
f* ^*-f
cn o>
3 J= LO
O -Mi.
i- 0)
.C O Nl
4J tl
EC CO)
o o; o >
f- £ ?-
E 4-> -M -M 3
oi » 0 E 0 <0
I/) i- O t. O
t- O - U
J- O U
r- O) X O) O)
«O i- O S- -C
f^ I *
Ol CO -O CO
i- <0 C (0 «4-
f- W'-- CD O
U 3 T3 3lT
> i C t -r-
D U-
-------
00
<
CO o
I f-
x VK
Z CO
si
a: y
OS CO
i os:
J_
i- .E -o cu
o *p" 5 il
i2 * *
o
J_
fell 2
T5 ** «*~ .^
** l/> 4-> ,!_
o:'r"£ »~
"5|
i_
*J _ CO
.s c £ 4-
CU 10 ."Jl i-
Z5l:*t
cu
o
t.
+* 2 cu
x c g 2.
CU 10 £ T-
ol ^ i.
3 cu
>
o
o
o
in
CM
o
o
g
CM
O
O
in
o
CM
o
O
in
8
§
2
^^
|^
i/)
+J
i/)
O
»
IO
f
ex
IO
0
o
8
vo
in
o
o
CM
JO
0
o
o
o
o
CM
O
o
o
vo
in
O
o
o
o
o
J£
o
CD
IO
8
o
o
o
o
o
u
o
o
CL
IO
o
^B
10
c
c
o
o
o
er>
o
o
o
8"
o
CO
o
o
o
8
o
CO
o
o
o
o
o
o
o
CO
^>
irt
o
o
r>
cu
3
u.
^B
IO
c
c
8
o
8
CO
t
8
^~
oo
o
8
8
1
CO
o
8
«t
o
o
00
o
8
8
00
«t
cu
<^>
c
r
10
^p^
B
o
10
O CM
§vo
r^*
t
VO CO
in
f^r
CO
O r-
O «9-
o r^
8 J2
^~
re
CO
O CM
o o
o in
vo W
in CM
t
vo
vo
CD ^^
o oo
o **
M
i CM
*~^
VO
vo
O 00
o* W
O CM
VO
VO
«*-
4J O>
t/l ^>
O I/I
0 O
0
f~
IO >»
C f JZ
C 03
O
IO
CO
el
II
cr
o
II O
c co
O l/l
o o
in u
J3
i
S-
o
IO
c
o
o
CO O
en in
i.
10 X
o 2
i A *» CO
o
CO *«*
^- O CD
Irt «/) C £
O O IO
O O C O
CO O
r- t 4-> in
IO
O
o
O t-
o 2
CM *« V* IO .C
10 X <
3 VO
c >« vo
c vo
< f O
IO .O O
CM i- O
CO O O 4->
CO £) ^- W>
(O > O
o i m o
a; «*_ o>
4-31
-------
m
S
a:
o>
u
cvj cvj
O* l l
0>
- 0i 0>
o
S i
*g> CO ^
*°I
Ills
** 00
5 -M 00 O
o
0) n
f- f- f- O.
55 5 5
^ TJ "O O
t t £ *
C «T «T £
ii i j.
r «O r- 4J
o cS o z
0
5 £
w m
r~ U ^~ O
CVJ CVJ
1 1
«t vo
o. r*.
0> 0i
"g "g
at 3
v£» «
2
«
t- O)
OJ >
I s
1 1
C *^
i i
5 J
f
U C
ID O
*-> X
X LU
0) O>
^
"5 5^ sJ 5
T ° S - >»£ S^ "°
OI*^-O O* f O C (A O
Vt *»-^- «0 10-^ 01 0 C
>>4-> 4-> *J O> O.
f o «e t> o E H
" « 3 .. 0 3 f 5
^»*J7. ?"»C o 0* f oj e
out. o c t.C^-0 10
4-34
-------
Gas recirculation 1s significantly more costly to implement than overflre air and
requires additional fan power. In existing units, the necessity t'o reduce unit
capacity to maintain acceptable gas velocities through the boiler conventive
sections may impose an additional penalty.
t For coal-fired units, gas recirculation to the coal pulverizers would cost approxi-
mately 15 percent less than windbox FGR; however, this may require increased excess
air to maintain adequate combustion. In any case FGR is not particularly effective
in reducing total NOX emissions from coal-fired systems.
t Water injection involves moderate initial equipment costs, but due to high operating
costs resulting from losses in unit efficiencies, it is the least desirable of the
NO reduction techniques evaluated.
In general, the cost of applying any of the control methods to an existing unit will
be approximately two to three times that of a new unit design.
Attention must be given to the base year in which control cqst estimates were made.
Figures on comparative electric power equipment costs from the most recent Marshall
and Swift Equipment Cost Index (1974) indicate that such costs have increased 19
percent from 1972 and 16 percent from 1973. It is safe to assume that costs will be
correspondingly higher in subsequent years.
4.1.2.2 Flue Gas Treatment
The flue gas treatment methods described in Section 3.2 included wet (oxidation followed by
scrubbing) and dry (reduction) methods. Since most of these processes are still in the early
stages of development, definite costs are, for the most part, not yet available. However, preli-
minary cost estimates have been made and are presented in Table 4-10. These estimates indicate
that the capital and operating costs for some of these processes are comparable to existing flue
gas desulfurization (FGD) systems.
4.1.3 Energy and Environmental Impact
In addition to affecting the cost of operating electrical generating combustion equipment,
implementing NOX control techniques can also impact overall plant efficiency and emissions levels
of pollutants other than NOX. These energy and environmental impacts are discussed below. The
discussion emphasizes potential Impacts due to applying combustion modification controls as these
have been the most extensively studied and offer the greatest potential for widespread use in the
4-33
-------
TABLE 4-11. EFFECTS OF RETROFIT COMBUSTION MODIFICATION NOV CONTROLS ON
UTILITY BOILER EFFICIENCY (Reference 4-34) x
Control
Effect on Efficiency
Comment
Low Excess Air
Flue Gas Recirculation
Off-Stoichiometric
Combustion
Water Injection
Reduced Air Preheat
Up to 1.5% increase
Insignificant
Little effect with oil/
gas firing
Possible 1% decrease with
coal firing
About 10% decrease
About 1% efficiency loss
per 30K decrease in air
preheat temperatures
Reduced stack gas heat
loss
Small increase due to
fan requirement
Possible increase in
overall excess air
needed for earlier
burnout
Heat of vaporization of
injected water lost
Increased stack gas
heat loss
The situation with flue gas treatment is likely to be different; however, data are insuffi-
cient to allow any quantitative assessment of the potential penalties. For the case of oxidation/
absorption wet processes it has been estimated that generation of the oxidizing agent (ozone) will
require approximately 10 percent of the power plant electrical output (Reference 4-35). When this
is added to the reheat requirements associated with all flue gas wet scrubbers, energy impacts may
be quite significant.
4.1.3.2 Environmental Impact
Modification of the combustion process in utility boilers for NO control reduces the ambient
levels of NOg, which is both a toxic substance and a precursor for nitrate aerosols, nitrosamines,
and photochemical smog. These modifications can also cause changes in emissions of other combustion
generated pollutants. If unchecked, these changes, referred to here as incremental emissions, may
have an adverse effect on the environment, in addition to effects on overall system performance. How-
ever, since the incremental emissions are sensitive to the same combustion conditions as NOX, they may,
with proper engineering, also be held to acceptable levels during control development so that the net
environmental benefit is maximized. In fact, control of incremental emissions of carbon monoxide, hydro-
carbons and particulate has been a key part of all past NOX control development programs. In addition,
recent control development has been giving increased attention to other potential pollutants such as
sulfates, organics, and trace metals.
4-36
-------
near term. Due to a virtual lack of data, the potential effects of flue gas treatment techniques are
only briefly mentioned. For the same reason, fuel switching and fuel additive effects are not treat-
ed at all.
4.1.3.1 Energy Impacts
The energy impacts of applying combustion modification NOX controls to utility boilers occur
largely through effects on unit fuel-to-steam efficiency. Although applying flue gas recirculation
requires additional forced draft fan capacity, the additional energy penalties imposed to drive the
fan are generally insignificant. Thus, effects on unit efficiency tend to dominate energy effects.
The efficiency effects of the combustion modifications for retrofit application are listed in
Table 4-11. As the table shows, applying low excess air firing results in unit efficiency gains.
For this reason the technique is gaining acceptance and becoming more a standard operating procedure
than a specific NOX control method in both old and new units.
The other commonly applied combustion modifications, F6R and off-stoichiometric firing,
generally have little energy impact on utility boiler operation. In certain instances, higher over-
all excess air levels are required when using these techniques (especially for coal-firing) to pre-
vent combustible losses. However, adverse effects are generally small.
A special point of concern relates to taking burners out of service on coal-fired boilers.
Since, in a typical installation, each coal mill supplies a given set (generally a row or an eleva-
tion) of burners, applying BOOS generally involves removing a mill from service. However, the
remaining mill capacity is usually insufficient to allow overfiring the remaining burners to main-
tain rated load. Thus implementing BOOS in a coal-fired unit may require derating the unit 10 to
20 percent. Of course, such derating represents a capacity loss, not an efficiency loss. But it is
an energy related adverse impact nonetheless.
The remaining combustion modification techniques listed in Table 4-11, water injection and
reduced air preheat, can impose quite significant energy penalties on utility boiler operation. As
a consequence, these techniques are quite unpopular, and have found little acceptance.
Table 4-11 applies only to retrofit application of the cornnon NOX combustion controls. These
same combustion modifications (LEA, FGR, off-stoichiometric combustion), in addition to low-NO
burners, almost never adversely affect unit efficiency when designed in as part of a new unit. This
illustrates that with suitable care during engineering and development, combustion modification NO
controls can be incorporated into new unit designs with no adverse energy impacts.
4-35
-------
TABLE 4-12. REPRESENTATIVE EFFECTS OF NOx CONTROLS ON CO
EMISSIONS FROM UTILITY BOILERS
(References 4-12, 4-16, 4-19)
NOX Control
Low Excess Air
Staged Combustion
Flue Gas Recirculation
Fuel
Natural Gas
Oil
Coal
Natural Gas
Oil
Coal
Natural Gas
Oil
CO Emissions (ppm)a
Baseline
14
86
12
8
14
19
85
15
19
42
20
24
27
27
14
86
12
14
19
85
15
28
24
27
17
31
175
21
NOX Control
68
74
61
8
34
42
53
20
19
93
60
283
81
225
16
67
13
14
21
85
21
37
20
26
40
45
65
9
{3% 02, dry basis.
4-38
-------
This section presents data obtained to date on the demonstrated effects of combustion modi-
fication NOX controls on incremental emissions from utility boilers. Attention is focused on flue
gas emissions as no data exist on incremental effects on liquid and solid effluents. Emphasis is
placed on incremental CO, vapor phase hydrocarbon, and particulate emissions; although incremental
sulfate and condensed phase organic emissions are briefly discussed. Lack of data precludes any dis-
cussion on the incremental effects of flue gas treatment, fuel switching, or fuel additive approa-
ches to NO control.
Carbon Monoxide Emissions
Since large quantities of CO in the flue gas of utility boilers mean decreased efficiency,
these boilers are operated to keep CO emissions at a minimum. Furthermore, if flue gas CO levels
reach concentrations in excess of 2,000 ppm, severe equipment damage can result from explosions
in flue gas exit passages. Thus, the degree to which a NOX reduction technique is allowed to
increase CO is limited by other than environmental concerns. In general, a NO control method is
applied until flue gas CO reaches about 200 ppm. Further application is then curtailed.
NOX control effects on CO emissions are highly dependent on the equipment type and the fuel
fired. In utility boilers of newer design, it is generally possible to achieve good NO reduction
without causing significant CO production. This is possible because newer burner and furnace designs
allow for better combustion air control and longer combustion gas residence time. In addition, oil
and coal-fired boilers usually emit very low CO levels during low NOX combustion because smoke and
soot production generally occurs with these fuels before significant CO levels are attained. Since
boiler operators strive to keep combustible losses to a minimum, conditions which result in soot
formation are avoided, resulting in correspondingly low CO levels. A summary of the field data on
the effects of the more extensively implemented modifications on CO emissions are shown in Table
4-12. These data are discussed below for each combustion NO control.
As the data in Table 4-12 illustrate, lower excess air levels in utility boilers can have pro-
found effects on CO emissions. In virtually all instances CO emissions increased significantly
when excess 02 levels were reduced 30 to 60 percent. Gas-fired boilers showed emission increases
up to 400 percent when excess 02 was lowered over this range, while oil-fired boilers were less
sensitive, and showed CO emission increases from 0 to 120 percent. However, coal-fired boilers
were the most sensitive to excess air reductions. Reducing excess 02 by 40 to 60 percent gave 100
to 1,000 percent increases in CO emissions.
4-37
-------
operating efficiency, and NOX controls which significantly decrease efficiency have found little
acceptance.
Particulate Emissions
Although gas-fired units produce negligible amounts of particulate, oil- and coal-fired
utility boilers currently emit approximately 38 percent of the nationwide particulate and smoke
(Reference 4-34). Potential adverse effects on these particulate emissions from NO combustion con-
trols could therefore have significant environmental impact. Unfortunately the optimum conditions
for reducing particulate formation (intense, high temperature flames as produced by high turbulence
and rapid fuel/air mixing), are not the conditions for suppressing NO formation. Therefore, most
attempts to produce low NO combustion designs have been compromised by the need to limit forma-
tion of particulates. This compromise has generally produced designs which maintain a well con-
trolled, cool flame, while still providing sufficient gas residence time to completely burn carbon
containing particles.
The NOX combustion controls currently receiving Lhe mos. widespread application in utility
boilers are low excess air, off stoichiometric combustion, and flue gas recirculation (for gas and
oil). The altered combustion conditions resulting from these modifications can be expected to
influence emitted particulate load and size distribution. For example, smoke and particulate emis-
sions tend to increase as available oxygen is reduced (soot emissions increase and ash particles
contain more carbon). Thus the degree to which excess air can be lowered to control NO is usually
limited by the appearance of smoke, especially in oil-fired units. Of course, the extent to which
excess air can be limited depends on equipment types and design. Many modern burners can operate
on as little as 3 to 5 percent excess air.
Similarly, the degree to which staged combustion can be employed is frequently limited by
the degree to which the primary flame zone can be stably operated fuel-rich, how well the second
stage air mixes with primary stage combustion products, and the residence time for combustion in
the second stage. Soot and carbon particles formed in the fuel-rich primary stage tend to resist
complete combustion downstream of that stage.
On the other hand, flue gas recirculation on oil-fired units can serve to decrease particu-
late emissions by providing more intimate mixing. Kamo, et al. (Reference 4-36) have demonstrated
that recirculation rates of 40 to 50 percent on a heater-sized oil-fired furnace reduced the smoke
number significantly.
4-40
-------
Off-stoichiometric, or staged combustion has proven to be a very effective NO reduction
technique for large steam generators. It can be implemented in a variety of ways including burners
out of service, overfire air ports, and biased firing. In all cases, the effectiveness of staged
combustion in reducing NO emissions depends in large part on the fraction of total combustion air
A
that can be introduced into the second combustion stage. It is in this second stage that complete
combustion of the fuel is achieved. CO emissions arise when this second stage combustion does not
go to completion prior to quenching in the convective section. This is caused by a combination of
the first stage being too fuel rich and the mixing of second stage air being too slow for the resi-
dence time provided. During development of retrofit or new design controls, these parameters are
usually selected so that CO emissions are acceptable.
The effectiveness of staged combustion in reducing NO formation while keeping CO emissions
low is highly dependent on specific equipment type. New utility boilers with multiburner furnaces
are especially amenable to this technique because it is generally not difficult to adequately dis-
tribute secondary air and assure complete combustion in these sources. Consequently, implementing
staged combustion in utility boilers is expected to elicit little effect on incremental CO emissions,
This conclusion is certainly borne out by the representative data presented in Table 4-12.
The use of flue gas recirculation (FGR) for NOX control has, in practice, been restricted
to gas- and oil-fired units. This technique is ineffective in reducing fuel NO production, the
predominant source of NO in coal firing. When FGR is implemented, 10 to 30 percent of the total
burner gas flow is recycled flue gas from the boiler exhaust. Further FGR increases can cause
flame instability due to reduced flame temperatures and oxygen availability. Theoretically, FGR
can lead to increased CO emissions, but unacceptable flame instabilities usually occur before the
onset of CO or smoke production. Thus, as Table 4-12 shows, the use of FGR has not caused increased
CO emissions. On the contrary, CO emissions have generally decreased.
Hydrocarbon Emissions
Field test programs studying the effectiveness of NOX controls often monitor flue gas HC
emissions as a supplementary measure of boiler efficiency. Therefore, some data on the effect of
these controls on HC emissions are available. Two recent test programs on utility boilers rou-
tinely measured flue gas HC (Reference 4-12 and 4-16). However, in virtually all tests, both base-
line and low NOX emissions were less than 1 ppm (or below the detection limit of the available
monitoring instrument). Thus, it was concluded that HC emissions are relatively unaffected by
imposing preferred NOX combustion controls on large utility boilers. However, this conclusion is
not altogether unexpected. The presence of unburned HC in flue gases implies poor boiler
4-39
-------
O
0:03
Li. '-i-
VO
OO r
Z I
O «d-
I I
CO "
CO CM
i i r
s: i
LU tO
| (U
«=C O
_J C
^D OJ
t_> s-
i i CU
»- «*-
C£ O)
<: ce:
Q. ». __
Z CO
O CK:
LU
co _i
_j >-i
O O
o: CQ
H-
O H-
X5 i
O H-
Lu Q
o uu
co
U- O
LU O
CO
LU
CO
03
"
U
r
4.)
S-
03
Q_
X
o
"^
5
Q
_J
O)
c
r
r
CU
to
03
CO
X
o
z:
2
o
i
o>
c
f
'cu
to
03
CO
CU
0
2E
CD
C
r*
S-
r
U_
^r
CM
r
1
VO
^J-
00
cr>
CM
VO
1
o
en
in
CM
«^(-
CM
1
in
VO
cr>
CO
CO
1
o
0
CM
Zl
03
^g
1
4_>
C
O
t_
u_
CM VO
00 ^1-
i O
1
\ \
CO CO
f*^ r~"
VO O
en
in vo
LO O
1 1
CO
co in
CM O
CO O
i vo
r CO
1 1
*3- O
co *a-
r CM
LO CO
VO CO
r CO
1 1
o cr>
CO CM
r CO
>,
'03
4->
c -o
O CU
N tO
-i- O
i- O.
0 0.
3C O
CO
co in
CO r~^
r
1 1
CO CO
«* o
1
CO
co cr>
in r
CM
1 1
CM
CM Cn
.
«5fr 0
CM
o m
^^ o^
CM CM
1 1
VO CO
CO CM
.
CM r
i *tf-
CM CO
CM r-
, ,
VO 00
CO O
.
O r
03
r
^ '
C
CU
CD
C
03
h-
CD
C
o
c:
CD
Q.
Q.
-a
a>
c.
to
CU
03
4->
to
03
4-42
-------
Published data on the effects of NO reduction techniques on particulate emissions from
A
utility boilers are scattered and insufficient for indepth analysis."" Table 4-13 summarizes the
particulate emissions data obtained during two recent field test programs which studied coal-fired
utility boilers (References 4-12 and 4-16). During the studies, particulate measurments were re-
corded under baseline and low NOV conditions. Since these NCL conditions were generally produced
A A
by a combination of low excess air and staged combustion, the individual effect of each technique
on particulate emissions cannot be determined. Nevertheless, the data do show that particulate
emissions are relatively unaffected by low NOX firing in front wall- and horizontally opposed-fired
boilers. Tangentially-fired boilers, on the other hand, exhibit slightly increased particulate
emissions under low NO conditions.
A
The effects of low NOX firing on carbon (or combustible) content of the particulate are also
shown in Table 4-13. Although the data are quite scattered, it appears that carbon losses increase
for front wall- and horizontally opposed-firing under low NOX conditions, but decrease slightly for
tangential firing. However, the changes are small and may not be significant.
The effects of low NOX conditions on emitted particle size distribution have also been
investigated to a limited extent (Reference 4-12). The data from a study of particle size distri-
bution in three boilers are summarized in Table 4-14. As the table shows, no significant changes
were noted in two of the boilers, both of which were tangentially fired. In the third, a hori-
zontally opposed boiler^ a distinct shift to smaller particles was noted, but the author reported
problems with the sampling and particle sizing equipment in this test, so the data may not be
significant.
Sulfate Emissions
Ambient sulfate levels have recently become a matter of increasing concern in regions with
large numbers of combustion sources, notably boilers, firing sulfur-bearing coal and oil. Although
the direct health effects of high ambient sulfate levels are currently unclear (References 4-37 and
4-38), recent thought suggests that sulfates may be more hazardous than S02. For this reason, con-
trol of primary sulfate emissions is becoming a concern even though primary sulfates (directly
emitted) comprise only 5 to 20 percent of ambient sulfate on a regional basis (Reference 4-38).
Since approximately 98 percent of the sulfur introduced into a utility boiler appears in flue
gas as an oxide, applying NOX controls would have essentially no effect on total SOX emissions. How-
ever, effects on the emitted (S03 + particulate sulfate)/S02 ratio can be significant. Specifically,
combustion conditions which limit local oxygen concentrations would be expected to decrease the
4-41
-------
extent of S02 to S03 oxidation. Thus applying low excess air firing and off-stoichiometric combus-
tion to control NOX should also lower S03/sulfate emissions.
Confirming data, though sparse, do exist. Recent measurements have demonstrated the expected
dependence of sulfate emissions on boiler excess air levels. Bennett and Knapp (Reference 4-39) have
shown that particulate sulfate emissions increase with increasing boiler excess 0~ in oil-fired
power plants. Homolya, et al. (Reference 4-40) report a similar increase in sulfate emissions as a
percentage of total sulfur emissions with increasing excess 02 in coal-fired boilers. Their data,
shown in Figure 4-8, show a linear relationship between the sulfate fraction of emitted sulfur and
boiler excess 02. Other data (Reference 4-41), shown in Table 4-15, also show that SO., emissions
decrease when staged combustion is used to control NO .
Organic Emissions
The term organic emissions as used here is defined to mean those organic compounds which exist
as a condensed phase at ambient temperature. Thus they are organics which are either emitted as "car-
bon on particulate" or condense onto emitted particulate in the near-plume of a stack gas. These com-
pounds, with few exceptions, can be classified into a group known variously as polycyclic organic
matter (POM) or polynuclear aromatic hydrocarbons (PNA).
POM production is generally only a minor concern in gas-fired systems, of some concern in
oil-fired sources, and of greater concern in coal-fired equipment. Like CO and HC emissions, POM
emissions are the result of incomplete combustion. Since NOV combustion controls can lead to inef-
A
ficient combustion, if not carefully applied (especially low excess air and off-stoichiometric com-
bustion), applying these controls can potentially lead to increased POM production.
Supporting data, However, are very limited, largely because of the difficulty of sampling
flue gas streams for POM and of accurately assaying samples for individual POM species. Thompson
et al., recently reported the effects of staged combustion and flue gas recirculation on POM emis-
sions from a coal-fired utility boiler (Reference 4-13). Their data, shown in Table 4-16, seem to
indicate that POM emissions do increase with two-stage combustion. However, they state that the
sampling and laboratory analysis procedures used in obtaining the data varied over the sample set.
Thus, they conclude that POM emissions are not significantly affected by firing mode. In another
study, Bennett and Knapp (Reference 4-39) attempted to investigate the effects of boiler excess 0?
on POM emissions from an oil-fired utility boiler. They found that particulate carbon content
increased with decreasing excess 02. However, because POM assay data varied widely, even for base-
line condition analyses, no conclusion regarding POM emissions was possible.
4-44
-------
CO
tI
CtL
CO
i«
O
L4J
II
CO
Uul
II
or
D_
O
CO
Z O
O CO
CO >-
Q CM
XUJ r-
o ctL i
LJL. I CU
O J O
< C
oo o cu
H- O J-
o cu
LU S «4-
u_ o a>
LL. a: oc
*d-
I
00
<
CU
N
r
CO
<4-
o
10
CU
0
^
4->
S-
^
H-
4-
C
C
E
C
r-
C
U
E
^
in
CD
V
E
in
o
E
;3.
O
i
E
in
-
E
l~t
O
CM
E
3.
in
CM
A
C
0
r-
T3
C
O
o
C"
si
r
U.
J
X
. cu
f °
lt^"__
)
O)
X. C
m «r~
3 $-
r«r-
J U.
in
un
CM
CT>
CM
^^
VO
CM
,_
O
CM
CM
r
O*>
00
f^
,_!
00
1
cu
c
«-
'cu
CO
03
CO
1
n
r
4-
C
a
c
c
o
1-
00
00
in
CM
CO
CM
cr»
CM
00
CM
CM
^~
cr>
00
r**-
CD
00
X.
o
^
o
1
3
3
J
71
3
00
CO
r
CM
^«
p*»*
C7^
O
O
o
p-»>
CT>
CM
in
f^
CM
O">
cu
c
r
'cu
CO
n3
CO
r~
n
r
4-
c
a
c
c
n;
H
VO
o
1
^_
VO
00
CD
O">
in
o
CTt
CO
r
CO
CTv
X
o
^
0
3
J
3
p
3
s
in
^*
^_
vo
en
CD
CM
vo
CD
o^
in
CM
vo
in
CM
cu
c:
"-
'cu
co
n3
CO
i-
as
c
O
N
<^
O
31
VO
CO
in
in
.
C7"1
C7^
00
in
oo
o
**
P>^
^^
.
o
CO
Q^
in
X
o
2
^
o
TQ
cu
co
o
Q.
a.
o
4-43
-------
CO
C£.
O
CO
OL
O
QL
CO r-
CO
CO
CO
o
c
O)
$_
UJ QJ
if-
X 0)
o ct:
co
CO
CO
O
CO
CM
O
CO
^_^
S.
CM
O
fr<
CO
o
M
T3
CO
cu ^
i-
o
0
E
o.
o.
o
1 !»
r- X
O UJ
CO
00
CM
i
00
vo
c^
CM
O)
c
'
. >
1*^
fs.
in
in
in in
r r
UJ
_J
ooj
^ CO)^*
r- i (r-
1 C7)
1^
1^"
1
1
O 00|O>
r VO CO
CO O^'O%
* 1
*"" 1 >
l<
CM
^^% ^^*
^^^ ^^^
t~~ ^f
m
CO CO
o
CO
o
4-46
-------
0.09
'« 0.08
o
*J 0.07
o* 0.06
0.05
1550
£1500
0*1450
1400
0.2
0.4
0.5
0.6 0.7 0.8 0.9
BOILER EXCESS 02, percent
1.0 1.1
1.2 1.3
1.4
Figure 4-8. S02 conversion vs. excess oxygen in coal-fired utility boilers (Reference 4-40).
4-45
-------
4.2 INDUSTRIAL BOILERS
Industrial boilers range in capacity from 3 to 73 MW*, (10 to 250 million Btu/hr), and com-
prise a wide diversity of firing types and fuels. In 1974, industrial boilers consumed about 30
percent of the fossil fuel used by stationary sources. The industrial boiler capacity distribution
was approximately 49 percent gas-fired, 29 percent oil-fired and 22 percent coal-fired. About a
third of the industrial units are small packaged boilers. Fifty-three percent are- in the 5 to 30
MW (17 to 100 x 106 Btu/hr) range and are primarily packaged watertube units. A more complete des-
cription of this equipment category is given in Section 2.3.2.
NO emissions from industrial-size boilers amounted to 2.2 Tg per year (2.44 x 106 tons) in
1974, or 18.2 percent of the nationwide stationary source emissions.
The following discussion of NO control techniques centers on the most promising method, com-
bustion modifications.
4.2.1 Control Techniques
The data on applied combustion modification technology for industrial boilers are limited.
The most extensive results were derived from a recent EPA-sponsored study (Reference 4-49, 4-50).
This study involved the field testing of a representative sample of industrial boilers to determine
their NO reduction potential. Ten different combustion modification techniques were implemented.
The effects of these techniques on NOX emissions and boiler efficiency are summarized in
Figure 4-9 for 73 separate boiler tests. The ten techniques are listed at the top of the figure.
The graph is divided into quadrants. The criterion for the best quadrant is that the modification
technique should simultaneously reduce NOX and increase efficiency. In general, the study showed that
total NO emission reductions of up to 47 percent were possible by using one or a combination of six
different methods. These were: excess air reduction, burner-out-of-service (BOOS), flue gas recir-
culation (FGR), overfire air addition, burner register adjustment, and reduced air preheat. Only
with the first three methods was boiler efficiency generally unimpaired.
Of these three, lowering excess air was the preferred method because boiler efficiency was
usually maintained or improved, and particulate emissions did not increase, as they do with most of
the other techniques. FGR is the next most promising technique, since particulate emissions
Approximately 1 percent of the industrial boilers are greater than the 73 MW (250 x 10 Btu/hr)
classification for which new source performance standards have been established (Reference 4-42).
These boilers are essentially the same as utility boilers.
4-48
-------
o
CO
Q
LU
or
o;
o
LU
I
z.
O
tI
00
CO
o
O. -Vi-
CO
> OJ
JX $-
in o
^"^ 9fm 4M
_« r O")
. 3-
i .
^^ CO ^^ C^
1
O 00 CO VO |
CM
1
to u ' n
CM C£
C
O
fll.
r T- ^,
cu
C£
c
o
r-
4->
co n
j? 2,1
<*> ^t r^ co
^3- 00 CM CM
E «"> | co in co r
° *o> 1 °^ °° ^ °^ '
cr> *3 I *~" *""
CO <-- *
0
^^^ ^^^
^^ *^ J
CM* ^
oo 2i.
CM
CM ^-*
O CO
CT> i
CO
to
CNJ
oo
«+-
'T. E R gp °^ *" "^ O
c
cu
*-
^> ^^
C <^-x
IB C~ -*
^ _^
O
00 O
CM r-
CM v_-
_
0)
c
r^
CQ
<
R c
n ^^
£
fl ^
I Anthracel
*
O)
^J
c
O)
f~
1 Fluorantl
-------
increased only five percent. Staged air addition and BOOS show some potential, but difficulties
exist in distributing the air to avoid an increase in particulate emissions.
The remainder of this subsection is devoted to a discussion of combustion modification
experience on gas, oil, and coal-fired boilers. This information is derived mainly from References
4-49 and 4-50.
Gas-Fired Boilers
NO emissions under base load conditions for gas-fired units varied from 26 to 190 ng/J
(50 to 375 ppm) (Reference 4-51). The level was most strongly dependent on combustion air tempera-
ture since NO from gas firing is predominantly thermally produced. For example, emissions from
base loaded watertube boilers varied from 37 to 60 ng/J (70 to 116 ppm) when operated on ambient air.
NO emissions from watertube boilers with preheated combustion air varied from 47 to 190 ng/J (90
to 375 ppm).
Boiler NO emissions with gas firing generally decreased with decreases in excess air level
although significant exceptions were noted. The decrease in NOX emissions with low excess air fir-
ing was generally more pronounced when preheated combustion air was utilized.
Off-stoichiometric combustion was demonstrated to be successful for NOX control in multi-
burner, gas-fired boilers. NOX reductions of 12 to 40 percent were achieved by terminating the
fuel flow to an individyal burner and using that burner port as an air injection port. Simultaneous
limiting of excess air also showed a 24 percent reduction in NOX emission in one instance. In the
tests where these combustion modifications were utilized, the percentage of units with emissions
less than 86 ng/J (0.2 Ib N02/106 Btu) was increased from 75 percent to 82 percent.
Cichanowicz et al. (Reference 4-52) have compared the influence of FGR for both watertube
and firetube boilers for natural gas at constant excess air. These results are shown in Figures
4-10 and 4-11. Forty percent FGR reduced NOX emission by approximately 70 percent for both these
boilers. Flue gas recirculation per se was found to have minimal effects on gas fueled boiler effi-
ciency (Reference 4-50).
Reducing firing rate, in general, does not have a strong effect on NOX emissions. The N0y
reduction achieved with lower load was nullified by the increase in excess air at the reduced load
that was required for adequate boiler performance. This resulted in an insignificant NOX decrease
or even an increase at the lower firing rate. Watertube gas-fired boilers were relatively insensi-
tive to load changes unless they were equipped with preheaters. In this case, NOX reductions of
4-50
-------
+7
+5
«- +2
i
x
o
C3
O
cc
LU
CO
25
-50
-75
WORST QUADRANT
COMBUSTION MODIFICATION METHOD
O FLUE GAS RECIRC.
O AIR REGISTER ADJ.
A OIL VISCOSITY
O BURNERTUNEUP
V ATOMIZATION PRESSURE
ATOMIZATION METHOD
REDUCED EXCESS AIR
AOVERFIREAIR
REDUCED AIR PREHEAT
V BURNER-OUT-OF-SERVICE
o J-
Is
7*ffT7
I1-':
i: £
BEST QUADRANT
-
-10
-5
+5
+10
CHANGE IN EFFICIENCY, percent
Figure 4-9. Effect of combustion modification methods on total nitrogen oxides
emissions and boiler efficiency (Reference 4-50).
4-49
-------
about 20 percent were obtained as the firing rate was dropped from 100 percent to 50 percent of
name plate capacity.
Oil-Fired Boilers
Base load NO emissions for these boilers ranged from 36 to 101 ng/J (65 to 180 ppm) with
No. 2 oil, 112 to 347 ng/J (200 to 619 ppm) for No. 5 oil, and 107 to 196 ng/J (190 to 350 ppm) for
No. 6 oil. The influence of fuel nitrogen was clearly shown by the variation from 50 ng/J (90 ppm)
for 0.045 percent nitrogen oils to about 180 ng/J (325 ppm) for 0.5 percent nitrogen fuel oils.
The oil-fired firetube boilers showed little dependence of NO emissions on excess air or
A
load. The watertube boilers were more sensitive, showing decreasing NOX with decreased excess air.
There was also a stronger effect of excess air when the heavier oils were fired.
Off-stoichiometric combustion on a multiburner oil-fired boiler produced NOX reductions of
17 to 49 percent. These reductions were achieved by stopping fuel flow to selected burners and
using the burner port as an air injection port. It was also shown that using upper burners as air
injection ports gave better results than using lower burners. This is not a general recommendation,
because this kind of result is highly dependent on the individual internal flow patterns of a parti-
cular boiler.
Emissions of NO were found to be relatively independent of the fuel oil atomization method,
but dependent on the characteristics of the individual burner. For a given oil burner, the atomiza-
tion method that produced the lowest nitrogen oxide emissions also generally produced the greatest
quantity of particulate. Boiler efficiency was essentially unaffected by atomization method.
Atomization pressure, however, did effect NO emissions to a slight extent. In one series
of tests, an increase in steam atomization pressure increased NO by 6 percent.
In most tests, NO emissions decreased as the excess air was reduced. An average change in
NO of about 11 ng/J (20 ppm) for a one percent change of excess oxygen level was observed for
A
watertube boilers. Firetube boilers averaged about 3 ng/J (6 ppm) for each one percent change in
oxygen. Emissions of NO for No. 2 oil were less affected by excess oxygen level than were those
for Nos. 5 and 6 oil. For all oils, the sensitivity of NO to excess oxygen was small when the total
NO concentration was less than 100 ng/J (180 ppm).
Reducing air preheat temperature on No. 6 oil firing did not have as large an effect on N0y
emissions as it did on gas firing. This technique primarily affects thermal NOX formation; the con-
tribution of fuel NO in residual oil firing is considerable. Reducing air preheat resulted in an
4-52
-------
LOAD-2.72Mg/h STEAM
(6000 Ib/hr)
FUEL -NATURAL GAS
02 -3.9 PERCENT
20 30 40
FLUE GAS RECIRCULATION, mass percent
Figure 4-10. Influence of flue gas recirculation on NOX emissions from a
firetube boiler (Reference 4-52).
LO AD-3.12Mg/h STEAM
(6900 Ib/hr)
FUEL -NATURAL GAS
0? -3.5 PERCENT
20 30 40
FLUE GAS RECIRCULATION, mass percent
Figure 4-11. Influence of flue gas recirculation on NOX emissions from a
watertube boiler (Reference 4-52).
4-51
-------
Reducing excess oxygen in coal-fired boilers was found to be more effective in reducing N0x
than in either oil or gas-fired boilers. Each one percent change in excess oxygen resulted in
approximately a 37 ng/J (60 ppm) change in N0x emission, regardless of air preheat temperature.
This strong, consistent, decrease in N0x with decreasing excess air was unique to coal; the N0x
occasionally increased when excess oxygen was decreased on oil and gas firing. Also, the average
amount of excess air fired for coal averaged 8.7 percent, which was significantly higher than for
oil or gas.
The overfire air concept was tested on several stoker fired boilers, and the results were
disappointing. The above-grate air injection ports (conventional equipment on all stokers) re-
duced NO by only 8 percent. It was concluded that the ports were located too far from the pri-
X
mary flame zone or that the overfire air system lacked the capacity to alter the air/fuel ratio
sufficiently for NO reduction.
The effect of firing rate was investigated by varying the boiler load about the base load
point of 80 percent of nameplate capacity. Generally, coal-fired watertube boilers showed an
increase in NO when operating below 60 percent capacity. This increase usually coincided with an
excess air increase.
Another EPA-sponsored study is assessing the potential of substituting western U.S. sub-
bituminous coals for eastern bituminous coals as an industrial boiler fuel (Reference 4-53). The
results of the program are significant from a NO control standpoint since switching to this more
A
abundant coal may become widespread in the near future. The western coals were found to be com-
patible with the industrial boilers of current design, although two units of older design (underfed
and traveling grate stokers) had some combustion difficulties. The coals were superior to eastern
coals in terms of lower NO , SO , particulate, and unburned hydrocarbon emissions. In addition,
they could be fired at lower excess air than eastern coal, and produced much lower combustible
losses.
To summarize, it appears that significant NOX emission reduction can be obtained in most indus-
trial boilers with minor modifications in operating conditions. Additional emission reductions are
possible with boiler redesign which would permit cost-effective implementation of off-stoichiometric
or staged-combustion in units burning heavy fuel oils or coal. Very few existing units possess the
necessary flexibility for this type of major retrofit. Problems which must be considered in the
design of new units, and particularly in the modification of existing units, include corrosion and
deposits on boiler tubes, flame instability, and combustion noise. The level of NOX reduction that
4-54
-------
efficiency decrease of about 2.5 percent per 50K increase in stack temperature. For general appli-
cation, this technique would require an increase in economizer area to maintaia overall efficiency.
Flue gas recirculation was implemented on a watertube boiler firing a No. 6 oil. At a mini-
mum excess oxygen level of 2 percent, adding 20 percent FGR to a steam-atomized flame lowered NO
x
by 50 percent. For an air-atomized burner, less dramatic N0x reduction effects were experienced
at both nominal and minimum excess oxygen levels. For oil firing in general, FGR rates greater
than 27 percent caused flame instability and blowout.
The effect on N0x emissions of tuning the burner was also determined. Tuning involved nozzle
examination, spray angle adjustment, and flame length adjustment. The chief effect of burner modi-
fication was a reduction in carbon monoxide rather than N0x> Particulate generally increased after
tuneup. The most effective method of reducing N0x by tuning was to reduce excess oxygen and accept
some increase in CO emissions.
Coal-Fired Boilers
The baseline N0x emissions from coal-fired boilers were generally higher than those from gas
and oil-fired units. Emissions ranged from 100 to 550 ng/J (165 to 900 ppm). Although the fuel
nitrogen contents of the test coals were high, ranging from 0.8 to 1.5 percent by weight, the field
studies indicated no strong dependence of N0x emissions upon fuel nitrogen content. Other factors
are apparently more important in determining N0x production, such as furnace geometry, excess air,
firing rate, burner type, and other fuel properties.
It was found that pulverized and spreader stoker-type boilers produced the highest - 550 ng/J -
baseline N0x emissions. Chain grate and underfed stokers had the lowest - 100 ng/J - emissions. In
these latter units, the combustion air fed up through the grating is insufficient for complete oxida-
tion, so additional air must be introduced above the grating through overfire air ports. The combus-
tion is, therefore, effectively staged, and the N0x emissions were quite low (100 ng/J or 165 ppm).
Spreader stokers, in which the fuel is introduced with the air flow above the grate, had
intermediate emission characteristics. Some of the fuel is burned in suspension, and the remainder
is combusted on the grate as in the underfeed stoker. The resultant combustion is only partially
staged. The combustion intensities are also higher than for underfed stokers, possibly increasing
thermal NO formation.
The pulverized coal units, especially the cyclone-fired types, produced the highest baseline
emissions due mainly to the very high combustion intensity.
4-53
-------
boilers to date (References 4-49 and 4-50) were presented in Figure 4-9. As the figure indicated,
only with low-excess air, BOOS, and FGR was boiler efficiency generally unimpaired. For low excess
air firing, efficiency gains of from 1 to 3 percent were typical. Taking burners out of service and
FGR generally did not affect efficiency. On the other hand, other tested controls generally imposed
efficiency penalties. Both overfire air and reduced air preheat gave efficiency losses of 1 to 2
percent.
The effects of NOX combustion modification controls on incremental pollutant emissions from
industrial boilers should also be analogous to those described previously in Section 4.1.3.2 for
utility boilers. Unfortunately, the only currently circulated data are on incremental flue gas
CO, HC, and particulate effects.
Carbon Monoxide Emissions
The bulk of the data on incremental CO emissions due to NOX controls applied to industrial
boilers was obtained in the two previously cited field test programs (Reference 4-49 and 4-50). In
these studies, CO emissions were reported for both baseline and for low NOX firing. Baseline emis-
sions were recorded with the boiler operating at 80 percent of rated capacity under normal (or as-
found) conditions. Low NO testing was implemented until CO emission levels reached 100 to 200 ppm,
then it was curtailed.
The data obtained during these studies are summarized in Table 4-17. As indicated in the
table, baseline CO emissions for industrial boilers are generally insignificant. However, the applv
cation of NO combustion controls in most cases adversely affected CO levels because each control
was implemented until CO levels became unacceptable.
As noted for utility boilers, CO emissions from industrial boilers are also adversely
affected by lowering excess air levels. As observed in the field study and shown in Table 4-17,
CO emissions from gas- and oil-fired boilers can be significantly increased when excess oxygen is
reduced 20 to 50 percent. Coal-fired boilers showed lower residual CO emission increases.
Two methods were used in the industrial boiler study to effect staged combustion: overfire
air and burners out of service. In these tests baseline CO emissions were always low. Combustion
staging by both methods generally resulted in unchanged to slightly increased CO emissions.
The data in Table 4-17 show that FGR has little effect on CO emissions. This conclusion
substantiates what was noted in the utility boiler testing discussed in Section 4.1.3.1. In addi-
tion, the data in Table 4-17 illustrate that varying combustion air temperature has almost no
4-56
-------
is achievable on industrial boilers is close to, but generally not as great as, that attainable
with utility boilers. The Industrial Environmental Research Laboratory -RTF of the EPA is continuing
to develop cost-effective NOX reduction methods for industrial boilers.
4.2.2 Costs
Cost data for combustion modifications on industrial boilers are virtually nonexistent. Only
the most broadly based estimates are available to the boiler owner and operator at the present time.
The most recent information of this kind was published in Reference 4-5Q. In that industrial boiler
field study, a 5.1 MW (17.5 x 103 Ib steam/hr) D-type watertube boiler was modified by adding staged
air and flue gas recirculation capability. The windbox depth was increased and a second set of
registers to control the recirculating flue gas was installed inside the extension. The cost of
these modifications was estimated at $5000. (The current cost of new boilers of this type is about
$60,000.) The cost of a similar modification on other modern D-type boilers could be as high as
$7500 if the existing burner registers cannot be used.
Manufacturers of industrial boilers in the 90 MW (300 x 103 Ib steam/hr) size range and one
million dollar cost category estimated that a staged air installation in general would add 2 to 4
percent to the boiler's cost. Specifically for A-type boilers, the incremental cost would be about
2 percent, and for D-type about 3 percent. Another booster air fan, if required, would increase
the modification cost by about one percent (Reference 4-50).
In a recent study, costs for retrofitting an existing unit to accept flue gas recirculation
were estimated (Reference 4-54). Approximate costs, which include design, installation and equip-
ment costs associated with the retrofit of FGR systems were, in 1975 dollars, $20,340 for a 3.51 MW
firetube boiler and $21,190 for a 3.51 MW watertube boiler. However, these costs would be consid-
erably less for a new boiler. Reference 4-54 estimates that for a new boiler of the size mentioned
above, the cost of including an FGR system will be about $6,900.
Research and development, including field testing and application of NO control methods to
this equipment category, is still in its early stages. More accurate cost estimates for these
techniques are being developed as part of ongoing and planned EPA studies.
4.2.3 Energy and Environmental Impact
As was the case for utility boilers, the energy impacts of applying combustion modification
NOX controls to industrial boilers occur almost exclusively through effects on boiler efficiency.
Data on these efficiency effects of NOX control from the most extensive field study of industrial
4-55
-------
effect on CO emissions. These observations suggest that effects of peak flame temperature on CO
emissions were also insignificant.
Hydrocarbon Emissions
The field test program investigating NOX controls applied to industrial boilers also reported
data on incremental HC emissions. These data are summarized in Table 4-18 and show that the use
of NOX combustion controls generally do not affect flue gas HC levels. Some tests show a slight
increase in HC emissions, yet others indicate slight reductions. Based on these data, it seems
fair to conclude that HC emissions from boilers are unaffected when implementing NOX combustion
controls.
Particulate Emissions
In addition to the above data, the industrial boiler test program reported some particulate
emissions and size distribution data showing the effects of several N0y combustion controls. These
particulate emissions data from several oil- and coal-fired boilers are summarized in Figure 4-12.
The figure shows changes in particulate emissions versus changes in NOX emissions from baseline con-
ditions as a function of the applied NOX control.
As Figure 4-12 shows, the effects of NOX controls on particulate emissions are mixed. For
example, both forms of staged combustion tested increased particulate emissions by 15 to 90 percent,
while flue gas recirculation increased emissions by 10 percent. In contrast, reducing air preheat
decreased particulate emissions by 45 percent. Furthermore, low excess air firing generally lowered
particle emissions 25 to 60 percent.
These observations are in general agreement with those of Heap, et al. (Reference 4-54) who
studied FGR and staged combustion applied to two oil-fired packaged boilers. They found that
smoke emissions increased slightly when both FGR and staged combustion were applied.
Cato, et al. (Reference 4-50) also reported some very limited particle size distribution
data, shown in Figure 4-13. This figure shows that, in a distillate oil-fired boiler, as excess
air levels are lowered, the emitted particle size distribution shifts slightly to larger sizes. A
more pronounced shift to larger particle sizes was observed with reduced load in a residual oil-
fired boiler. However, these data are much too limited to allow any definite conclusions to be
made regarding the effects of combustion modifications on flue gas particle size distribution.
4-58
-------
TABLE 4-17. EFFECTS OF NOx CONTROLS ON CO EMISSIONS
FROM INDUSTRIAL BOILERS
(References 4-49 and 4-50)
NOX Control
Low Excess Air
Staged Combustion
t Overfire Air
Burners Out of
Service
Flue Gas Recirculation
Variable Air Preheat
Fuel
Natural Gas
Distillate Oil
Residual Oil
Coal
Natural Gas
Distillate Oil
Residual Oil
Coal
Natural Gas
Residual Oil
Natural Gas
Residual Oil
Natural Gas
Residual Oil
Coal
CO Emissions (ppm)a
Baseline
10
0
0
0
50
47
90
0
0
0
0
0
0
25
70
0
0
25
10
10
0
0
0
0
0
10
0
0
10
0
0
10
10
0
0
10
0
322
0
0
0
NOX Control
10
0
11
900
129
485
150
17
45
100
9
28
205
20
60
0
22
25
10
20
0
0
30
80
49
10
0
43
20
0
10
0
75
0
0
0
30
320
0
0
10
a3% 02, dry basis.
4-57
-------
+50
+40
+30
^+20
CO
LLJ
S+10
x
o
CO
C9
2
X,
-10
-20
-30
-40
-50
COMBUSTION MODIFICATION METHOD
O AIR TEMP. REDUCTION
D REDUCED FIRING RATE
A FLUE GAS RECIRC.
O REDUCED EXCESS AIR
STAGED AIR
BURNERTUNEUP
A BURNER-OUT-OF-SERVICE
O
BEST QUADRANT
WORST QUADRANT
b
-200
100 0
CHANGE IN PARTICULATES, percent
+100
+200
Figure 4-12. Effect of NOX controls on solid paniculate emissions from industrial
boilers (Reference 4-50).
4-60
-------
TABLE 4-18. REPRESENTATIVE EFFECTS OF NOX CONTROLS ON VAPOR
PHASE HYDROCARBON EMISSIONS FROM INDUSTRIAL
BOILERS (References 4-49 and 4-50)
NOX Control
Low Excess Air
Staged Combustion
t Overfire Air
Burners Out of
Service
Flue Gas Recirculation
Variable Air Preheat
Fuel
Natural Gas
Distillate Oil
Residual Oil
Coal
Natural Gas
Distillate Oil
Residual Oil
Natural Gas
Residual Oil
Residual Oil
Natural Gas
Residual Oil
HC Emissions (ppm)a
Baseline
42
10
17
7
3
8
35
11
21
5
0
0
0
12
35
0
10
15
35
NOX Control
34
0
13
8
9
13
25
18
7
0
0
0
0
14
15
0
0
13
25
13% 02, dry basis,
4-59
-------
4.3 PRIME MOVERS
4.3.1 Reciprocating Internal Combustion Engines
Stationary reciprocating engines account for nearly 20 percent of the NOX from stationary
sources, or 2.4 Tg per year (2.66 x 106 tons). There are presently no Federal regulations for gase-
ous emissions from these engines. Some local areas, such as the South Coast Air Pollution Control
District of Southern California, have set standards for internal combustion engines.
A 1973 study by McGowin (Reference 4-55) provides a good overview of emissions from station-
ary engines, particularly the large bore engines used in the oil and gas industry and for electric
power generation. An EPA-sponsored Standards Support and Environmental Impact Study (SSEIS) for
these engines (Reference 4-56) will be completed in 1978 and will be the most comprehensive study of
stationary reciprocating engines to date.
4.3.1.1 Control Techniques
The NO control techniques for 1C engines must be effective in reducing emissions over a broad
range of operating conditions from continuous operation at rated load to lower utilization appli-
cations at variable load. In general, large natural gas spark ignition engines running at rated
loads have the highest NOX emission factors. Gasoline engines, in contrast, frequently operate at
lower loads (less than 50 percent of rated) and produce substantially higher levels of CO and HC. The
NO control techniques for these engines often involve HC and CO control since these emissions fre-
quently increase as NO is reduced. Divided chamber diesel-fueled engines produce low levels of
NO (accompanied by greater fuel consumption than open chamber designs). In general, all diesel-
fueled engines have relatively small HC and CO emissions (less than 4 g/kWh*).
The following paragraphs will discuss NO control techniques in general followed by a tabula-
tion of specific NOX reductions, by engine group. A lack of emission data precludes any discussion
of natural gas engines less than 75 kW/cylinder (100 hp/cylinder).
Table 4-19 summarizes the principal combustion control techniques for reciprocating engines.
These methods may require adjustment of the engine operating conditions, addition of hardware, or a
combination of both. Retard, air-to-fuel ratio change, derating, decreased inlet air temperature,
or combinations of these controls appear to be the most viable control techniques in the near term.
Nevertheless, there is some uncertainty regarding maintenance and durability of these techniques
because, in the absence of regulation, very little data exists for controlled engines outside of
laboratory studies, particularly for large stationary engines. In general, increases in fuel
consumption, as much as 10 percent, are the most immediate consequence of the application of these
*
shaft output
4-62
-------
. BASELINE
LOW EXCESS
AIR
AERODYNAMIC DIAMETER,jum
Figure 4-13. Effects of NOX controls on particulate size distribution fromioil-fired
industrial boilers (Reference 4 50).
4-61
-------
techniques (excluding inlet air cooling). All control techniques involve only operational adjust-
ments with the exception of (1) derating which may require additional installed capacity to compen-
sate for the decreased rating, (2) inlet manifold air cooling which involves the addition of a heat
exchanger and a pump, and (3).catalytic conversion, which requires adding a catalytic reactor.
While exhaust gas recirculation (EGR) yields effective reduction of NO , this technique
requires additional development to overcome fouling of flow passages and increased smoke levels. In
general, recirculated exhaust is cooled in order to be effective. This practice promotes fouling.
EGR has not been field tested for large engines, and has been rejected by one manufacturer of heavy-
duty diesel truck engines and limited by another manufacturer. EGR has potential application in
naturally aspirated engines if full load EGR cutoff is provided to prevent excessive smoke (<10
percent opacity). EGR, however, has been applied successfully in combination with other techniques,
such as retard, in gasoline-fueled automobile engines (References 4-56, 4-57).
Water injection, similarly, has serious maintenance and durability problems associated with
mineral deposit buildup and oil degradations. Despite use of demineralized water and increased
oil changes, the control problems associated with engine startup and shutdown persist. This
factor, coupled with the need for a water source, has led manufacturers to reject this technique
(Reference 4-56).
Combustion chamber modifications such as precombustion and stratified chambers have demon-
strated large NOX reductions, but also produce substantial fuel consumption increases (5 to 8 per-
cent more than open chamber designs). With the rapid increases in the price of diesel fuel and
gasoline, manufacturers have been reluctant to implement this technique. In fact, one manufacturer
of divided chamber engines is vigorously pursuing development of low emission open chamber engines
(Reference 4-56).
Table 4-20 summarizes emission reductions achieved with large bore engines by use of retard,
air/fuel ratio changes, derating, and reduced inlet manifold air temperature (MAT). This table
includes only those techniques from Table 4-19 which could be readily applied by the user. The
cited emission reductions are based on results obtained from engines tested in manufacturers'
laboratories. Therefore, some uncertainty exists concerning durability and maintenance over longer
periods of operation. In general, the greatest NOX reductions are accompanied by the larger
increases in fuel consumption. This is a direct result of reducing peak combustion temperatures
and, thus, decreasing thermal efficiency.
4-64
-------
TABLE 4-19. SUMMARY OF NOX EMISSION CONTROL TECHNIQUES FOR
RECIPROCATING INTERNAL COMBUSTION ENGINES
CONTROL
RETARD
Injection (CI)a
Ignition (SI)b
CHANGE AIR-TO-FUEL (A/F)
RATIO
DERATE
INCREASE SPEED
DECREASE INLET MANIFOLD
AIR TEMPERATURE
EXHAUST GAS RECIRCULATION
(tGR)
External
Internal
valve over! ap
or retard
exhaust back
pressure
CHAMBER MODIFICATION
Precombustion (CI)
Stratified charge (SI)
WATER INDUCTION
CATALYTIC CONVERSION
PRINCIPLE OF REDUCTION
Reduces peak temperature
by delaying start of
combustion during the
combustion stroke.
Peak combustion tempera-
ture is reduced by off-
stoichiometric operation.
Reduces cylinder pres-
sures and temperatures.
Decreases residence time
of gases at elevated
temperature and pressure.
Reduces peak temperature.
Dilution of incoming com-
bustion charge with inert
gases. Reduce excess
oxygen and lower peak
combustion temperature.
Cooling by increased
scavaging, richer
trapped air-to-fuel
ratio.
Richer trapped air-to-
fuel ratio.
Combustion in ante-
chamber permits lean
combustion in main cham-
ber (cylinder) with less
available oxygen.
Reduces peak combustion
temperature.
Catalytic reduction of
NO to Ng.
APPLICATION
An operational adjustment. Delay
cam or injection pump timing (CI);
delay ignition spark (SI).
An operational adjustment. In-
crease or decrease to operate on
off-stoichiometric mixture. Reset
throttle or increase air rate.
An operational adjustment, limits
maximum bmepc (governor setting).
Operational adjustment or design
change.
Hardware addition to increase
aftercooling or add aftercooling
(larger heat exchanger, coolant
pump).
Hardware addition; plumbing to shunt
exhaust to intake; cooling may be
required to be effective; controls
to vary rate with load.
Operational hardware modification:
adjustment of valve cam timing.
Throttling exhaust flow.
Hardware modification; requires
different cylinder head.
Hardware addition: inject water into
Inlet manifold or cylinder directly;
effective at water-to-fuel ratio *
1 (kg H20/kg fuel).
Hardware addition: catalytic con-
verter installed in exhaust plumbing
or reducing agent (e.g. ammonia)
injected into exhaust stream.
BSFCd
INCREASE
Yes
Yes
Yes
Yes
No
No if EGR
rates not
excessive
Yes
Yes
Yes
No
No
COMMENTS LIMITATIONS
Particularly effective with moderate amount
of retard; further retard causes high exhaust
temperature with possible valve damage and
substantial BSFC increase with smaller NOX
reductions per successive degree of retard.
Particularly effective on gas or dual -fuel
engines. Lean A/F effective but limited by
misfiring and poor load response. Rich A/F
effective but substantial BSFC, HC, and CO
increase. A/F less effective for diesel-
fueled engines.
Substantial increase in BSFC with additional
units required to compensate for less power.
HC and CO emission increase also.
Practically equivalent to derating because
bmep is lowered for given power requirements.
Compressor applications constrained by vibra-
tion considerations. Not a feasible tech-
nique for existing and most new facilities.
Ambient temperatures limit maximum reduction.
Raw water supply may be unavailable.
Substantial fouling of heat exchanger and flow
passages; anticipate increased maintenance.
May cause fouling in turbocharged, aftercooled
engine. Substantial increases in CO and smoke
emissions. Maximum recirculation limited by
smoke at near rated load, particularly for
naturally aspirated engines.
Not applicable on natural gas engine due to
potential gas leakage during shutdown.
Limited for turbocharged engines due to
choking of turbocompressor.
5 to 10 percent increase in BSFC over open-
chamber designs. Higher heat loss implies
greater cooling capacity. Major design
development.
Deposit buildup (requiring demineralization) ;
degradation of lube oil, cycling control
problems.
Catalytic reduction of NO is difficult in
oxygen-rich environment. Cost of catalyst
or reducing agent high. Little research
applied to large-bore 1C engines.
'compression ignition
Spark ignition
cbmep -- brake mean effective pressure
dBSFC -- brake specific fuel consumption
4-63
-------
Numerous investigators have studied control techniques to reduce NO in diesel-fueled auto-
motive truck applications. Many of these studies are summarized in Reference 4-57. Retard, turbo-
charging, aftercooling, derating and combinations of these controls are techniques that are current-
ly utilized by manufacturers to meet California heavy-duty vehicle (>2700 kg, or 6000 Ib) emission
limits for diesel-fueled engines.
Table 4-21 lists five samples of NOX control techniques currently implemented by truck
manufacturers to meet the 1975 California 13.4 g/kWh* (10 g/hp-hr) combined NO and HC emission level
Manufacturers indicate that greater reductions will require (1) increasing degrees of application
of these controls (and incurring additional fuel penalties) or, (2) application of techniques that
need further development to overcome maintenance, control, and durability problems. Controls in
this second category include EGR, water injection, and NO reduction catalysts.
Gasoline engine manufacturers, in response to Federal and State regulations, have also con-
ducted considerable research of emission control techniques to reduce NO , as well as HC and CO,
X
levels. Efforts in this area have been directed at reducing emissions to meet (1) Federal and
California heavy-duty vehicle limits, and (2) Federal and California passenger car emissions limits.
Table 4-22 lists Federal and State emission limits, and Table 4-23 lists the various controls that
are used in several combinations by manufactures to meet these limits. Table 4-24 gives specific
examples of control techniques recently applied to meet Federal light duty vehicle emission limits.
Based on the preceding discussion, potential N0x emissions reductions for stationary recipro-
cating engines can be summarized as follows:
t Controls such as retard, air-to-fuel ratio change, turbocharging, inlet air cooling (or
increased after cooling), derating and combinations of these controls have been demon-
strated to be effective and could be applied with no required lead time for development.
Fuel penalities, however, accompany these techniques and may exceed 5 percent of the
uncontrolled consumption.
Exhaust gas recirculation, water injection, catalytic conversion and precombustion or
stratified charge techniques involve some lead time to develop as well as time to address
maintenance and control problems.
*
rated shaft output
4-66
-------
UJ
o;
o
ca
i
LU
o
O H-.
o
CO Z
_J UJ
o
a: z
H- o
z -
O h-
o co
n>
XOQ
o :E
z o
o
o
o
OJ
co
o
to
o
+3
o
-o
cu
s-
cz
cu
o
s-
cu
a.
ro
E
S-
o
ro
0)
I/I
V
f
o
r_
0)
u-
(O
o
3
3
^.
c^
^.
CM
CM
to
O)
o
c
s
.
(I)
.O
Z
U
U
^~
oo
CO
o
^~
o
0
r"
^5
^
r-
co
CO
m
r
CO
r
1
*^
r**
VO
CM
1
00
f
t^
1
0
r
CM
O
CO
1
r*,.
r
(T>
CM
1
^j-
CM
£
0
CM
CO
r"
p
CU
(/)
fO
oo
^ CO
W ^ C3
1 1
r-. o -
CM C3
"*,,<=>
II
CO CM
,
in o
CM
o
f CM O
T ^ , W.
W ^- -
^O Ot O^
O CO f O
I'll
P-" CO ^ ^
CO
o
CO
in oo i
i
r- r- in
CM
O
r- in vo co
CO ^- CM
cn
in » CM o»
CM O VO O
^IW
0)
u.
i
T> O CO
1. 4J *J
a.
E
3
CO
c
O
cu
a
CM
OJ
0
CO
rn
1
°CD
^.
CO
CM
VO
co
CO
*
CM
in
o
1
1 *°
1 a; °
1 1 ^
o
1 ^ 1
o , in
^ ?
vo °0 ^r
CM r*. o
O
CO f
O
CM r
a
O CM
^ 00* '
in r in
vo o
O VO CO
CM CM P
3
Lk.
1
O CO
1 <
C o £
o -
i- fO
-M
o u.
o o
c
O i
r
<-> $-
o o
13
-O "
CD a^
J- C
CD
Qj
-a cu
CO
i. ro
CU (U
JC S-
-«-> o
o c
ftJ O
-t-> ro
^
O 4->
C
!-> CU
M- O
ro s-
^: cu
co Q.
CU
4J
ro
Q.
co
ro
ro
c
CU
en
4->
I
* . "O
"O CU CU
S- tO CD
ro ro c
cu cu
S- >
o ro
cu u
roo ~o co cu
4-> r r3
CU
ro
cu
CO S-
r- ^
cu ro
CU -MO
c i ro i
r- O S- -Q
r S- CU
CU -4-> Q- I
co c E
ro o cu co
CO U 4-> CO
ro .a u
cu
en
ro
cu
>
4-65
-------
TABLE'4-22. 1975 VEHICLE EMISSION LIMITS
Passenger Car,
g/kWh (g/mi)a
California
Federal
Light duty truck^
g/kWh (g/mi)
California
Federal
Heavy duty vehicles,
g/kWh
California
Federal
NOX
6 (2.0)
9 (3.1)
6 (2.0)
9 (3.1)
HC
3 (0.9)
4 (1.5)
6 (2.0)
6 (2.0)
13
21
CO
26 ( 9)
44 (15)
59 (20)
59 (20)
40
53
Emissions limits are estimated in g/kWh from g/mi assuming an average of 38.4
km/hr requiring 8195 W (11 bhp) for the 7-mode composite cycle.
TABLE 4-23. EMISSION CONTROL TECHNIQUES FOR AUTOMOTIVE GASOLINE ENGINES
Control
NOX:
Rich or lean A/F ratio
Ignition timing retard
Exhaust gas recirculation
(5 to 10 percent)
Catalytic converters
(reduction)
Increased exhaust back pressure
Stratified combustion
HC, CO:
Thermal reactor
Catalytic converter (oxidation)
Exhaust manifold air injection
Positive crankcase ventilation
Comment
Increased bsfc, HC, and CO
Increased bsfc, HC, and CO, amount
of control limited by potential
exhaust valve damage
Increase bsfc and maintenance
related to fouling, smoking limits
degree of control
In developmental stage
Increase bsfc
Requires different cylinder head,
increased bsfc
Very effective in reducing HC, CO
Requires periodic catalyst element
replacement
Increased bsfc to power air pump
Reduces HC evaporative losses
4-68
-------
TABLE 4-21. CONTROL TECHNIQUES FOR TRUCK SIZE
DIESEL ENGINES [<375 kW (500 HP)]
TO MEET 1975 CALIFORNIA 13.4 G/KWHR .
10 G/HP-HR) COMBINED NOY AND HC LEVEL1
Control
Percent
bsfc^ Increase
Retard, modify fuel
system and turbocharger
Retard, modify fuel
system and turbocharger,
add aftercooler
Add turbocharger and
aftercooler0
Retard0 (naturally
aspirated version)
Precombustion chamber
3
3
5 - 8
lBased on Federal 13 mode composite cycle
'bsfc = brake specific fuel consumption
'Stationary versions of this engine would
require a cylinder head with four exhaust
valves rather than existing two valves.
4-67
-------
NO control technology for automotive applications can be adapted to stationary engines;
however, NO reductions and attendant fuel penalties for automotive applications are
closely related to the load cycle, which in some cases may differ from stationary
applications
Viable control techniques may involve an operational adjustment, hardware addition, or
a combination of both
0 Additional research is necessary to
Establish controlled levels for gaseous-fueled engines (<75 kW/cylinder, or
100 hp/cylinder)
- Establish controlled levels for medium-powered diesel and gasoline engines based
on stationary application load cycles
- Supplement the limited emissions data available for large bore engines
4.3.1.2 Costs
As discussed earlier, stationary engines are unregulated for gaseous pollutants. Consequently,
few data are available for field-tested controlled engines, particularly for large (>375 kW or 500
hp) engines. Sufficient data exist, however, to give order or magnitude NOX control costs for the
following engine categories:
Natural gas-, dual-, and diesel-fueled engines above 75 kW/cylinder
(100 hp/cylinder)
Small to medium (<75 kW/cylinder) diesel-fueled engines
Gasoline-fueled engines (10 kW to 375 kW)
Costs for large stationary engines can be estimated based on Reference 4-58 and information
supplied by Reference 4-56. These costs, however, relate to emission reduction achieved by engines
tested in laboratories rather than to field installations. Reference 4-59 indicates, nevertheless,
that these data are representative.
In contrast to the large stationary engines, more published cost data exists for smaller
(<375 kW, 500 hp) gasoline and diesel engines which must meet State (California) and Federal
emission limits for mobile applications. Stationary engines in this size range are versions of these
mobile engines. Therefore, costs can be estimated based on a technology transfer from mobile appli-
cations to stationary service, keeping in mind that in some cases mobile-duty cycles (variable
4-70
-------
CO
\l 1
z
1 1
o
z
LU
Z
o
1*
I
CO
13
CO
r>
O
o
1
<
z
o:
LU
H-
Z
H- 1
UJ
z.
* t
_l
o
to
3
_J
z
o
14
H-
Z
LU
>
Z
0
o
x-^
o£ r^. ..
O in
U- 1
«±
co
s: LU
LU CJ
1 Z
CO LU
>- ct:
CO LU
LJ_
_l LU
o o:
OL
1 z
z o
O Ct:
O Lu
Z 0
O LU
*-* I
CO D-
CO CC
i Q
z: j
^.j
r
ra
c
CU
Q-
fj _
0)
3
U-
E
5
+j
00
CO
y_
ra
CU
s_
CU
i
3
CD"
s~~**
- X
_J _l _l 0 O
-^
s: _j
0 O
O r r
r*
0 0
CD
O -H +1
CM +1
*x> ^o
+1 <£> O O
r- O O O
m co
co i o
m CM
r^ -oo
CO O
+1 +i
O +1
LO LO
+i o in i
r- i O O
O LO LO
CO t O
LO
r^ o o o
CO
+1 +1 +1
o
LO in co
fl CO VO i
i r- O O
CO CM CM
| -H +1 +1
l-» LT> CO
/j
o
o: t-i
CD
O r CM CO
D
CU
0
-o
cu
(.
o
c
ra
*
"cu
3
M-
1
O
4->
1
s_
r
ra
c
ra
cu
-o
s_
ra
cu
j-
cu
L-
ra
a
C r-
o c
4-* ^
J
^
^
}
ra
S_ CD
3 E
CD ;
c ra c
cu s- c
CO C T
CO 0
r- ! 0
i CO (T
CU CO -C
CO CU
ra J-
JD O- CO
cu o o
_E 0 -
4-> CO
* CO
C i -r-
i- cu E
3 CU
"O ^4~
CU I E
o o cu
r 1 CO
US- >1
C -r- CO
r- ra
CZ i
CO > O O
C T3 -r- S-
O i- 4-» »->
r- ra u c
4J 4J CU C O
ra cu -i-5 o o
U i- C -r-
i- -r- 4J
<4- .» ra
r- CO S- r C C
X> C C !- 3 O H
§00 ra o -i- »->
r- .- CU J- »-> V) T
C4J4J -*-O-'-CU>> <
o ra-i- ra i o s- 1 c
M-OC -4->oQJ3ra T
O) +J-1-CD CM-i-^i-M >t
c ra <4- *- -i-T- t. ra c
i- o !- c: co ra u T
CDO -r- -oocu-4->rareo
C'r-<«- O'r-J^raECD CD i
CU4->-<-ECOCU T3C C
ra-t-> ojrx:4J-»->cU''- »
cut- c ai t- u COCO>NI c
C CUC-»-> ^J33O-r- (1
r-c 73 -i-o+Jureras-TS *
r-o HH CDCUCO-I-.C.C: o.-r-
cu-i- cr rasxxEx f
COCO +> LULULuCD-LULUi >O C
ra co c T
-Q CU CU 4-
S- C 1 1 1 1 1 1 1 1 t
CM a. o :
- cu
S-
3 3
"O
( CU
u
c o
J S-
Q.
~ ^J3
3 »-> O
CO CO
3 CO
J -(-> II
0
3 re re
3 CJ
1 "
co &s
CO > O
c cj co
o
r- LO 1
co r-.
co en o
r r r
j
CO O II
4«)
cu s:
C "O
r- CU -
r 4-> ^-5
CU 0 O
co cu i
ra L.
J3 S- II
0
CM O _J
1*^
en $- .«
i O CO
E
CD CU
c: -M
-I- CO
O CO >>
9 3 CO
3 C »->
u cu c
I H CU
ra co
-MO)
U £-
3 ra o.
4_>
- ra «+-
3 T3 0
J
J CO C
3 C O
O -r-
r- 4-)
r co ra
3 co i-
- ,- 0
J E -r-
J CU i-
3 CU
cnoE *-> ce.
-------
TABLE 4-25. COST IMPACTS OF NOX CONTROLS FOR LARGE-BORE ENGINES
Cost Impact
Retard
Air-to-fuel changes
Derate
Manifold air cooling
Combinations of above
Control techniques
Increased fuel consumption, more frequent
maintenance of valves
Increased fuel consumption, more frequent
maintenance of turbocharger
Fuel penalty, additional hardware, and increased
maintenance associated with additional units
Increased cost to enlarge cooling system, and
increased maintenance for cooling tower water
treatment
Initial, fuel, and maintenance
Increases as appropriate
TABLE 4-26. TYPICAL BASELINE COSTS FOR LARGE
(>75 kW CYLINDER) ENGINES3
Costs
1. Initial,6 $/kW
2. Maintenance,
$/kWh
3. Fuel and lube,
$/kWh
Total Operating,
2 + 3
Gas
174
4 x 10'3
10 x TO'3
14 x 10'3
Dual Fuel
174
4 x 10~3
10 x 10-3
14 x TO'3
Diesel
174
4 x 10-3
23 x 10"3
27 x 10'3
Based on Reference 4-58 and information supplied to
Reference 4-56 by manufacturers.
Includes basic engine and cooling system.
4-72
-------
load) can differ from stationary-duty cycles (rated load). Hence, costs (e.g., fuel penalties)
associated with a control technique used in a stationary application may vary from the mobile case.
Control costs for the three categories discussed above may include:
Initial cost increases for control hardware and/or equipment associated with a particular
control (e.g., larger radiator for manifold air cooling or more engines as a result of
derating)
Operating cost increases which consist of either increased fuel consumption and/or in-
creased maintenance associated with NO control system
Combinations of initial and operating cost increases
Control Costs for Large Bore Engines
Table 4-25 lists cost impacts for control techniques available to users of large stationary
engines. These cost impacts may be related to actual installations using baseline data presented
in Table 4-26. In practice, these figures vary depending on the application, but, in general,
they are representative of the majority of applications. Basically, these controls involve an operat-
ing adjustment with the exception of derating and manifold air cooling, which would require hardware
additions. Derating is not a viable technique for existing installations unless additional units
can be added to satisfy total power requirements.
The impact of the above control costs may vary considerably given the following considera-
tions:
Standby (<200hr/yr) application control costs are primarily a result of initial
cost increases due to the emission control, whereas continuous service (>6000 hr/yr)
control costs are largely a function of fuel consumption penalties
Controls which require additional hardware with no associated fuel penalty (e.g.,
manifold air-cooling) may be more cost effective in continuous service (>6000 hr/yr)
than operating adjustments which impose a fuel penalty (e.g., retard, or air-to-
fuel change)
The price of fuel can affect the impact of a control which incurs a fuel penalty.
For example, a control which imposes a fuel penalty of 5 percent for both gas and
diesel engines has more impact on the diesel fueled engine because diesel oil costs
about 40 percent more per Joule than natural gas. This impact will diminish if gas
prices increase more rapidly than oil prices.
4-71
-------
TABLE 4-28. ESTIMATES OF STICKER PRICES FOR EMISSIONS HARDWARE FROM 1966 UNCONTROLLED
VEHICLES TO 1976 DUAL-CATALYST SYSTEMS (Reference 4-57)
Model
Year
1966
1968
1970
1971-
1972
1973
Configuration
PCV-Crank Case
Fuel Evaporation
System
Carburetor Air/Fuel Ratio
Compression Ratio
Ignition Timing
Transmission Control
Sys tern
Total 1970
Anti-Dieseling
Solenoid
Thermo Air Valve
Choke Heat Bypass
Assembly Line Tests,
Calif (1/10 vol)
Total 1971-1972
OSAC (Spark Advance
Control)
Transmission Changes
(some models)
Induction Hardened Valve
Seats (4 and 6 cyl )
EGR (11 - 14%)
Exhaust Recirculation
Air Pump Air
Injection System
Quality Audit, Assembly
Line (1/10 vol)
Total 1973
Typical Hardware
Value
Added
1.90
9.07
0.61
1.24
0.61
2.49
3.07
2.49
2.74
0.18
' 0.48
0.63
0.72
5.48
27.16
0.23
Price
2.85
14.25
0.95
1.90
0.95
3.80
4.75
3.80
4.18
0.57
0.95
0.95
1.90
9.50
43.32
0.38
Excise
Tax
0.15
0.75
0.05
0.10
0.05
0.20
0.25
0.20
0.22
0.03
0.05
0.05
0.10
0.50
2.28
0.02
Sticker
Price
3.00
15.00
1.00
2.00
1.00
4.00
8.00
5.00
4.00
4.40
0.60
14.00
1.00
1.00
2.00
10.00
45.60
0.40
60.00
4-74
-------
Control Costs for Small and Medium Gasoline- and Diesel-Fueled Engines
Control costs for these engines can be characterized by analogy to those incurred to meet
State and Federal emission limits for automotive vehicles. Again, these costs consist of initial
purchase price increases for control hardware and increased operating costs (fuel and maintenance
cost increases).
Table 4-27 lists typical costs for techniques implemented for 1975 diesel-fueled truck
engines. These costs are presented to indicate order of magnitude effects. More research is
required to relate specific emission control reductions to initial and operating cost increases for
stationary engine applications.
Table 4-28 gives control hardware costs to meet gasoline-fueled passenger vehicle emission
limits through 1976. Note that cost increases correspond to increasingly more complex controls to
meet more stringent emission limits.
TABLE 4-27. TYPICAL CONTROL COSTS FOR DIESEL-FUELED ENGINES USED IN HEAVY-DUTY
VEHICLES (>2700 kg OR 3 tons)
Vehicles'
Initial
baseline
engine
cooling system
turbocharger
aftercooler
EGR
$40-$67/kW ($30-$50/hp)
8X-14X engine
$4/kW ($3/hp)
6%-10% engine
$3-$4/kW ($2-$3/hp)
Operating
Fuel:
Maintenance:
Fuel penalties range from 3 to 8 percent for various techniques.
Typical present fuel cost: $0.095/1iter ($0.36/gallon) #2 diesel
or $2.13-$2.37/GJ ($2.25-$2.50/106 Btu).
EGR system will require periodic cleaning. Note that turbo-
charged, aftercooled engines require additional maintenance for
the turbocharger and aftercooler compared to a similarly rated
naturally aspirated version.
Based on information supplied to Reference 4-56 by manufacturers.
4-73
-------
Figure4-14 illustrates the effect of various control techniques on fuel economy. Fuel
cost increases can be easily derived from typical gasoline costs, presently $0.55-0.75/gallon.
In addition to this operating expense, control techniques utilizing catalysts and EGR require peri-
odic maintenance.
Manufacturers, in addition, incur certification costs for gasoline and diesel-fueled engines
which must meet State and Federal regulations. These costs are passed on to the user in the form
of increased initial costs. Manufacturers of diesel-fueled engines report these costs range from
$50,000 to $100,000 for a particular engine family. This can result in a $125 cost per engine
based on a low sales volume family.
4.3.1.3 Energy and Environmental Impact
The energy impacts of applying NO controls to stationary reciprocating 1C engines are mani-
fested almost exclusively through corresponding increases in fuel consumption (bsfc). Typical
percentage increases as a function of applied control were discussed in detail previously in
Sections 4.3.1.1 and 4.3.1.2.
Potential adverse environmental impacts occur through increases in emissions of combustion-
generated pollutants other than NO attendant to applying a NOX control. Since 1C engines emit
only an exhaust gas effluent stream, impacts through liquid and solid effluents need not be con-
sidered. In addition, since 1C engines fire "clean" fuels (natural gas and distillate oil) incre-
mental effects on the emissions of such pollutants as SO and trace metals are relatively unimpor-
tant. Thus, the following discussion will focus on the measured effects of specific NOX control
techniques on incremental emission of CO, HC, and particulate (smoke). Again, all available data
were obtained in tests on laboratory engines. Nevertheless, such data should be representative.
Carbon Monoxide
As discussed in Section 4.3.1.1, the most common NOX reduction techniques applied to 1C
engines include derating, ignition retard, altering air/fuel (A/F) ratios, reducing manifold air
temperatures (MAT), and water injection. The effects of each of these N0y controls on engine CO
emission levels are summarized in Table 4-29.
As indicated, baseline CO emissions from two-cycle engines are generally lower than those
from four-cycle engines. However, derating two-cycle engines increases CO emissions 50 to 100 per-
cent, while derating four-cycle engines actually gives a 60 to 100 percent decrease in CO levels.
4-76
-------
TABLE 4-28. ESTIMATES OF STICKER PRICES FOR EMISSIONS HARDWARE FROM 1966 UNCONTROLLED
VEHICLES TO 1976 DUAL-CATALYST SYSTEMS (Reference 4-57) (Concluded)
Model
Year
1974
1975
1976
Configuration
Induction Hardened
Valve Seat V-8
Some Proportional EGR
(1/10 vol at $52)
Precision Cams, Bores,
and Pistons
Pretest Engines -
Emissions
Calif. Catalytic Con-
verter System (1/10 vol
at $64)
Total 1974
Proportional EGR
(acceleration-
deceleration)
New Design Carburetor
with Altitude
Compensation
Hot Spot Intake Manifold
Electric Choke (element)
Electronic Distributor
(pointless)
New Timing Control
Catalytic - Oxidizing-
Converter
Pellet Charge (6 Ib at
$2/lb)
Cooling System Changes
Underhood Temperature
Materials
Body Revisions
Welding Presses
Assembly Line Changes
End of Line Test Go/No-Go
Quality Emission Test
Total 1975
2 N0x Catalytic Converters*
Electronic Control3
Sensors9
Total 1976
Typical Hardware
Added
0.72
3.21
2.44
1.80
4.02
20.07
7.52
2.87
2.67
4.35
1.40
18.86
12.00
1.17
0.63
0.67
0.13
1.85
1.22
22.00
28.00
3.00
List
Price
1.90
4.94
3.80
2.85
6.08
30.02
14.25
4.75
4.75
9.50
2.85
34.20
20.52
1.90
0.95
1.90
0.95
2.85
1.90
37.05
47.50
5.70
Excise
Tax
0.10
0.26
0.20
0.15
0.32
1.58
0.75
0.25
0.25
0.50
0.15
1.80
1.08
0.10
0.05
0.10
0.05
0.15
0.10
1.95
2.50
0.30
Sticker
Price
2.00
5.20
4.00
3.00
6.40
20.60
31.60
15.00
5.00
5.00
10.00
3.00
36.00
21.60
2.00
1.00
2.00
1.00
3.00
2.00
138.20
39.00
50.00
6.00
134.00
1976 most common configuration
4-75
-------
o
oo
t/0
.
LU ai
l>0 1-
LU (U
^*"x
^s
CT>
C
c
o
(/>
(/)
«r*
E
LU
o
O
s_
4-J
c
o
o
X
0
a
£
"a
i/
a
a.
c.
o
i_ "_J3
2 °
(0
-^» "*""5
C
(1)
o ,
3 *J,
"S 5
ex:
co
O) U_
t. "*^
0 *£
S
(O
^
0 <
C
11
*Tjj tr
w '*~
+j .t!
CD
O)
4^
(0
s_
0)
o
) ^
* O x^>
I - "3
, Jfl\
!-|,5
' ' ' f
a»
a.
^
' '
m o
co co
o
«3- in
OPi
i i
O *d-
^a- in
0
in
O CO
*3- CO
i i
in in
ai (U
U 0
i i
CM «a-
v>
CJ3
03
s_
^J
**
z
0
VO
vo i
^^
o
VO
*
o
m
in
L^ |
^
r~
in
i r*.
i vo
en
i co
i ^^
CO «3-
CM m
vo vo
i i
0 0
«3- VO
i CM
r*\
00 '
0
o
m vo
CM «*
co m
i i
CM «sj-
CU O)
U O
1 1
CM *3-
^
9)
(ft
0)
a
0
in
i i
i
co
O
in
^ CM
in °°
*° VO
««* vo
^^ en
CM CM
00
1 CO
1 CM
O
r-. r-.
vo o
CM i
i i
5 S
CM VO
*d- en
** co
CM CM
O
VO
in
O
o
CM
CU CU
o o
1 1
CM **
f__
LU
fBM
(Q
a
-------
1.5
1.0
0.5
VARYING EGR AND
SECONDARY AIR RATES
10 20
INCREASE IN BSFC, percent (OVER UNCONTROLLED VEHICLE)
3.0
2.5
2.0
cc
=)
Q
LLJ
1.5"
cc
D_
C/J
C)
X
1.0°
0.5
30
X
o
GENERAL CORRELATION
ESTIMATED FOR ADDITION OF NOX CATALYST
BED AT 75 PERCENT EFFICIENCY
VARYING DRIVING CYCLES
AND CONTROL TECHNIQUES
1 1
5 10 15 20 25
INCREASE IN BSFC, percent (OVER UNCONTROLLED VEHICLE)
cc
=3
a
LU
o
o
cc
00
>
o
x
O
30
Figure 4-14. Effect of NOX emissions level on fuel penalty for light duty trucks
(Reference 4-59].
4-77
-------
O 2 CYCLE, BLOWER SCAVENGED
D 2 CYCLE, TURBOCHARGED
A 4 CYCLE, NATURALLY ASPIRATED
O4CYCLE, TURBOCHARGED
G NATURAL GAS
DF DUAL FUEL
0 DIESEL
20 30
POWER DERATE, percent
Figure 4-15. Effect of derating on 1C engine HC emissions (Reference 4-56).
1000
O 2 CYCLE, BLOWER SCAVENGED
D2CYCLE.TURBOCHARGEO
4 CYCLE NATURALLY ASPIRATED
<>4 CYCLE,TURBOCHARGED
G NATURAL GAS
DF DUAL FUEL
D DIESEL
20
4 6
TIMING RETARD, degrees
Figure 4-16. Effect of retarding ignition timing on 1C engine HC emissions (Reference 4-56).
4-80
-------
Retarding ignition generally causes increased CO output for all engines. This is somewhat
expected, though, since retarding ignition decreases both peak combustion temperature and combustion
gas residence time, which can lead to incomplete combustion. Both increasing A/F ratios and reduc-
ing manifold air temperature (MAT) has little effect on CO levels. However, decreasing A/F causes
50 to 100 percent increases in CO emissions. Water injection seems not to affect CO emissions from
gas and dual fuel engines, but increases diesel engine CO emissions by 60 to 130 percent.
Hydrocarbons
The use of N0y combustion controls on 1C engines can also have significant effects on HC
emissions, with different N0x reduction techniques eliciting different effects.
As shown in Figure 4-15, derating causes HC emissions to increase, with the increase becom-
ing more pronounced as load is further reduced. As the figure illustrates, derating can cause a 20
to 130 percent increase in HC emissions. Figure 4-16 shows the effect of ignition retard on incre-
mental HC emissions. In contrast to the effects of engine derating, ignition retard tends to
decrease slightly or not affect emissions of HC. However, in cases where retarding ignition initi-
ally reduces HC emissions, increasing the degree of ignition retard seems to have little further
effect. The data in the figure indicate that HC emissions decrease on the average of 30 percent
when ignition is retarded 3 to 8 degrees.
Changing the air-to-fuel (A/F) ratio, decreasing manifold air temperature (MAT) and water
injection can all result in increased HC emissions. As shown in Figure 4-17, both increasing and
decreasing the A/F ratio by 10 percent increases HC levels 20 to 65 percent. Larger percentage
increases occur in engines with high baseline emissions. Figure 4-18 shows analogous effects when
MAT is decreased. Decreasing 10 to 20 K (20 to 40 F) increases HC emissions 5 to 50 percent. HC
levels increase as MAT is further reduced. Turbocharged engines exhibit the largest percentage
emissions increases. Water injection also increases HC emissions from 1C engines regardless of the
baseline HC level, as shown in Figure 4-19. Average increases of 16 to 25 percent have been experi-
enced for water/fuel (W/F) ratios of 0.1 to 0.25.
Particulates
Virtually no data are available specifically on particulate emission rates from stationary
1C engines because it is difficult, time consuming, and expensive to measure particulate emissions
from these engines directly. Instead, exhaust gas opacity readings have been used as a substitute
measure of particulate emissions. These readings effectively measure particulate since a relation-
ship between visible smoke and particulate mass emissions has been established for medium power
4-79
-------
1000
ODF
O 2 CYCLE, BLOWER SCAVENGED
D 2CYCLE.TURBOCHARGED
O4CYCLE.TURBOCHARGED
G NATURAL GAS
OF DUAL FUEL
D DIESEL
MAT DECREASE. UK
Figure 4-18. Effect of decreased manifold air temperature (MAT) on 1C engine HC emissions
(Reference 4-56).
1000
O 2 CYCLE, BLOWER SCAVENGED
O4CYCLE.TURBOCHARGED
G NATURAL GAS
DF DUAL FUEL
D DIESEL
0.4 0.6
W/F RATIO
Figure 4-19. Effect of water injection on 1C engine HC emissions (Reference 4-56).
4-82
-------
1000
O 2 CYCLE, BLOWER SCAVENGED
CD 2CYCLE,TURBOCHARGED
O4CYCLE,TURBOCHARGED
G NATURAL GAS
DF DUAL FUEL
D DIESEL
-10 0 10
CHANGE IN A/F RATIO, percent
Figure 4-17. Effect of air-to-fuel ratio on 1C engine HC emissions (Reference 4-56).
4-81
-------
NOX LEVEL, g/kWh
12 16
20
24
OQ
s
x
cj
<
QC
CO
ec
o
CO
o
OQ
CONTROL CODE
AIR-TO-FUEL RATIO I
REDUCE COMP. RATIO M
DERATE R
EXTERNAL EGR
INDUCTION S
INTERNALEGR
MANIFOLD AIR TEMP.
RETARD IGNITION
TIMING
INCREASE SPEED
R #5 I
*^\FUEL
TYPE^\^
2 STROKE
BLOWER
SCAVENGED
2 STROKE
TURBO-
CHARGED
4 STROKE
NORMALLY
ASPIRATED
4 STROKE
TURBO-
CHARGED
DIESEL
0
D
A
DUAl
FUEL
O
3.
4.
ENGINE CODE NUMBER (#) DENOTES INITIAL POINT.
CONTROL CODE DENOTES LEVEL AFTER CONTROL.
BACHARACH AND BOSCH METERS ARE FILTER-TYPE
INSTRUMENTS.
SMOKE LEVELS FOR ENGINES #8-12 WERE MEASURED
WITH A BOSCH METER.
FINAL SMOKE LEVEL IS AT END OF LINE HAVING
CONTROL CODE.
s 10
0.
>*
<
a.
O
#34
#12
CR#8 #9
10 12
NOX LEVEL, g/hp-hr
14
16
18
20
Figure 4-20. Smoke levels versus NOX levels for large bore diesel engines.
4-84
-------
diesel engines (Reference 4-60 and 4-61). Therefore, 1C engine smoke emissions are generally re-
ported as percent plume opacity, or as Bosch or Bacharach smoke spot numbers.
The plumes from most large-bore engines are nearly invisible when the engine is operating
at steady-state. However, applying NOX combustion controls can significantly affect smoke emissions.
Figure 4-20 shows the relationship between smoke emissions and NO reduction as a function of NO
x x
control for those engines where data were reported on both pollutants. As the figure shows, NO
controls, other than derating, generally increase smoke emissions, while derating decreases smoke
levels. Ignition retard and exhaust gas recirculation (EGR) cause the most significant increases
in smoke level.
Since N0y controls which caused smoke levels to exceed 10 percent opacity were considered
unacceptable in the tests summarized in Figure 4-20, none of the data points for controlled engines
are above this value. However, the effect of progressively applying ignition retard and EGR on
smoke emissions is best demonstrated by data which include higher smoke levels. Such data are pre-
sented in Table 4-30 for two-cycle diesel engines, and clearly show that smoke emissions increase
progressively as percentage EGR or degree of retard is increased.
In summary, experimental data have shown that applying conventional combustion modification
NOX controls to 1C engines can cause increases in CO, HC, and particulate (smoke) emissions. This
is so because the combustion conditions required to prevent NOX formation generally lead to less
complete combustion.
4.3.2 Gas Turbines
Gas turbines contributed only 2 percent of the annual stationary source NO emissions in
1974, or 236 Gg (2.6 x 10s tons). They do, however, comprise a very rapidly growing source with
increasing application in intermediate and base load power generation, pipeline pumping, natural gas
compression, and onsite electrical generation. The increasing application of gas turbines carries
with it the potential for increasing the N0y emission contribution from these sources. In response
to this, the frequency of control technique demonstration and implementation has increased in the
past several years.
Uncontrolled NOX emissions are a function of turbine size (or efficiency) and fuel type. In-
creasing the turbine size (or efficiency) increases the N0y concentrations primarily due to higher
combustion temperatures and to increased residence time at high temperatures. Oil-fired turbines
generally have higher NOX concentrations than gas-fired units. Typical uncontrolled NO emissions
from gas turbines are illustrated in Figures 4-21 and 4-22 for large and small units, respectively.
4-83
-------
250
200
CM
o
o
QC
150
cc
h-
Z
LU
U
z
o
U
X
o
100
50
O GAS-FIRED UNITS
D OIL-FIRED UNITS
NOTE: DATA NOT ADJUSTED FOR
GAS TURBINE EFFICIENCY
D
D
D
O
PROPOSED NSPS
10
15 20
TURBINE SIZE, MW
25
30
35
Figure 4-21. NOX emissions from large gas turbines without NOX controls!Reference 4-62).
4-86
-------
TABLE 4-30. RELATIONSHIP BETWEEN SMOKE,
EGR, AND RETARD
(Reference 4-56).
Engine Type
2-cycle, Blower
Scavenged Diesel
2-cycle,
Turbocharged Diesel
Control a
None
10% EGR
20% EGR
39% EGR
4° advanceb
None
4° retard
None
4.9% EGR
8.4% EGR
12.1% EGR
Opacity, %
4.7
12
27.5
59
2.7
4.6
10
7.5
10.0
11.5
14.8
All EGR data based on hot EGR.
Injection advance is not a control; data included to show trend
4-85
-------
Imposed on these figures is the proposed NSPS of 75 ppm for these sources. Very few units meet these
standards in the uncontrolled state.
4.3.2.1 Control Techniques
Combustion modification techniques for gas turbines differ from those of boilers since turbines
operate at a lean air/fuel ratio with the stoichiometry determined primarily by the allowable turbine
inlet air temperature. The turbine combustion zone is nearly adiabatic and flame cooling for NOX
control is achieved through dilution rather than radiant cooling. The majority of NOX formation in
gas turbines is believed to occur in the primary mixing zone, where locally hot stoichiometric flame
conditions exist. The strategy for NO control in gas turbines is to eliminate the high temperature
stoichiometric regions through water injection, premixing, improved'primary zone mixing, and down-
streamed dilution.
Combustion modifications for ^as turbines are classified into wet and dry techniques. Wet
methods, such as water and steam injection, presently provide substantial reductions. As yet, no
combination of dry methods has been successful on field units in reducing emissions below a typical
standard of 75 ppm NO at 15 percent oxygen. Presently available wet and dry methods for NO reduc-
" X
tion are aimed at either reducing peak flame temperature, reducing residence time at peak flame
temperature, or both. These techniques, along with their reduction potential and future prospects,
are shown in Table 4-31.
Wet techniques are the most effective methods yet implemented with reduction potentials as
high as 90 percent for gas and 70 percent for oil fuels. With wet control, water or steam is intro-
duced into the primary zone either by premixing with the fuel prior to injection into the combustion
zone, by injection into the primary airstream, or by direct injection into the primary zone. The
effectiveness of each method is strongly dependent on atomization efficiency and primary zone resi-
dence time. In the case of water injection, peak flame temperatures are reduced further through
vaporization of the water.
Although NOX reduction is quite effective, numerous difficulties offer incentive to the
development of dry controls. If dry controls are developed as expected, the long-term future of
wet control does not appear promising based on the following inherent problems of wet controls:
0 Requirements for "clean" water or high-pressure steam
Hardware requirements which increase plant size
t Delivery system hardware which results in increased failure potential and overhaul/
maintenance time
4-88
-------
200
150
CJ
oc
100
o
o
X
o
O GAS-FIRED UNITS
D OIL-FIRED UNITS
NOTE: DATA NOT ADJUSTED FOR
GAS TURBINE EFFICIENCY
D
D
D
PROPOSED NSPS
50P_
P
D
0
0.5
1.0
2.0
2.5
1.5
TURBINE SIZE, MW
Figure 4-22. NOX emissions from small gas turbines without NOX controls (Reference 4-62).
3.0
4-87
-------
Uncertainty regarding long-term control effects on turbine components.
Although few combinations of presently available dry controls have the NO reduction poten-
tial of the wet methods, many dry techniques are used in conjunction with water or steam injection,
particularly on larger units. On the smaller units, dry controls may be sufficient to meet stan-
dards. The dry controls now available are:
Lean out primary zone - Reduces NOX levels up to 20 percent by lowering peak flame
temperatures. This option allows less control over flame stabilization and reduces
power output but is an attractive control to be built into future low NO combustors.
Increase mass flowrates Possible NO reductions up to 15 percent by reducing
residence time at peak flame temperature. This control essentially increases
the turbine speed at constant torque and is not feasible in many applications.
t Earlier quench with secondary air - This is a minor combustor modification which
entails upstream movement of the dilution holes to reduce residence time at peak
temperatures. This is a promising control which is generally employed in advanced
combustor research.
Reduce inlet air preheat -A control applicable only to regenerative cycle units.
It is not attractive due to reduction in efficiency.
t Air blast and air assists atomization - Use of high-pressure air to improve atomiza-
tion and mixing requires replacement of injectors and addition of high-pressure air
equipment. This control is considered an excellent candidate for incorporation into
new low NO design combustors.
Exhaust gas recirculation - Possible NOX reduction of 30 percent. A candidate dry
control for future design, though it has limited application in some online units.
EGR requires extensive retrofit relative to other dry controls and also requires a
distinct set of controls for the EGR system.
Other minor combustor modifications are generally aimed at providing favorable interval flow
patterns in the primary zone and fuel/air premixing. The bulk of these modifications are combus-
tor-specific and are being investigated by the manufacturer. In general, dry controls available
for immediate implementation have not exceeded 40 percent NOX reduction and as such may be insuffi-
cient controls for the larger units at present. Since dry techniques approach NO reduction dif-
ferently than do wet controls, their effects are complementary and, consequently, can be used toge-
ther. Figure 4-23 illustrates the effect of dry and wet controls used separately and in combination
4-90
-------
-
oo
I
UJ
z.
I
OQ
OO
CO
I
LU
_j
OQ
;
S
JLJ
1
O
1
E
Ol
J-
IO
u.
E
Ol
i
O 1o
4-> 4J
u c
3 01
* O
ee a.
s1
^£
CJ 4^
10 C
£3
|
u
i
in to to 10
i i i i
Ol
i C
5 in t-
- Ol
^ in >
4- 4-> 3 fc. *
U «O O 4->
C t- C l«- U
- 01 O Ol
O. 4-> 1*-
o oi c <*-
c -5 S «
Ol M C t.
<- 0 OI-»-
0. O. 3
f 01 cr >>
I*. O XI fit
H- 4-> i. Q.
Oi » CL
Cki >, 3 X
IA 3 A IA O
oi vi ae
o m 10 xi h.
3 4-> 01 01 i
O VI X fc- 4-> Ol
Ol O O Ol 10 3
ee o« x> x <-
* C
5"°
l/l"*" IA
IA IO
|£
Ol 01
4J C
IO 3
0 JA
V
Ol
1
g
IO
g. >i
10 C « X^
10 N OL
C f- Ol
O >v fc. > O
«- -f- 4J O t- 4J
4-> -^- O.4-> C
X « W 10 V-^- Ol
9 N IO > 01 C
Z a. <*-xi o
CLr- « H- H- O> IM
^ 0 0 0»4-»
Ol 3 4J 4J 4-» Ol E
|
V, »
£ H
4»*
§ i
0 4-»
XI I
i e u E
^ » oi 3
C C 4-> X
^"2^ IA O
4-> I. Ol
>> 10 01 i-
Ui 4J 3 Ol
§"« 5 IA«2
T- *> 01
U VI VI t. Ol
- C 10 0. C
<- 0) f *U
Ol E 01 3
IO f XI
-3"1*!
Increases overa
creasing flowra
operating costs
tlon. Requires
Ineffective 1n
Ol
i
to
IA
10
i
IO
CO
Ol
^
u
U- 4->
<- s
Ol X
4-> U
IA Ol
§£
IO C
o
Ol OI-»-
4J U 4-»
10 3 U
X) IA Ol
o «->
i
O
(/)
|
in
IO
%
e
o
5 *
« "g
! ^
£
o
1^ 10
4_> ^_
O
f Ol
Not an attract
term option du
flexibility
<4-
ractlve option
slble
tig
t <*-
2
o
*~
aiv,
Reduce residence t1
at peak temperature
VI
M
2
« «
»-4 1^.
h«. evi in
in to v«
i i i
L. in i
o c oi
Oj O XI XI
4J X IO
t 10 oi
o u c m
<*- t- Ol
>»- o *->
»- 4J 4J >,4->
a. 3 L. 3
An attractive o
term minor comb
and for Incorpo
signs. Limited
Incomplete comb
CL
01 in
U t-
KC 0
U VI
An attractive
to be employed
advanced combu
i
i. Q.4J
O Ol
sl*
4 3 4->
O 10 4J 4J
C U C C
** Ol O
X ,
U i.
J.§
Ol IS>
uj x«e
CVI
to
*
c
Ol
IA
01
XI
a
1
Ol
a
Generally consi
concept
o c
o
XI «-
5 e
Promising meth
Incorporated 1
low NOX design
c
1ted appHcatlo
retrofit
L£
§
0
S o>u
4-> C 5
!** XI
10 VI Ol
ctS1-^
*g|'s
oe ^i «-><
t
°4-§
4-> m t-
10 "vi «
«<
S
«
^
Reduces efflclei
Ol
o
Q
**-
01
> c
S°
U 4->
2 =
4J O
4-» IA
:E
0 01
z *->
o
S>>
u
o c
01 5
> u
4-> <*-
0«^-
IO Ol C
4^.2
4J « 4->
10 E 0
4-* Ol XI
i5 u
Ol
.A
IO
IO
<
Q
1
i
IO
1
o.
Ol
1
ee
4- ?
1 "^
*C 10
s~i
IcI
Qe°-%-'
(VI
to
i
+
^
c
Ol
u
With proper des
unimpaired
in
.c
«
=3 *J
E ,
Ol 3
4J .
01 Ol
t- C
0 """
10 C
i- IO C
4-> O
4J VI «-
< 10 4J
gl
o B
1-0
OJ 1
Reduce peak flame t
through premlxing,
secondary air 1njec
tlon, primary zone
flow reclrculatlon
1'
i."
f o e 4J
X 4J 0-^
in T- <-
1- 3 4J O
Ol A 10 1.
J= g 0 4J
4-1 O *- Ol
o o<«- ee
XI IO
4-> O> t-
0 J- Ol
IO 3 4->
t- 4-1 C in
4-> 3 f- t-
4->
U 4-> VI
C O i- 3
C
Ol T3
Ion has seen us
or combustor mo«
ations
4-1 C U
g
o
Reduce peak flame
temperatures
ws
3 «
IA U
3 t.
LU ee
S
*
_ 1^
C «4-
i ^~
"8 x
t» i *
4-> ^
10 3
o «
Oj C
^- IO
IO 4->
C Q.
Current researcl
reliability, mai
time and start-i
i. VI
0 C
**- O)
^ "w*
O OJ
^- T^
Attractive metl
new combustor <
~
"lO
o
o
o
*o
-C Ol
U i
Ol .O
t IO
oo
Ol
o
Complete combustion
at lower peak
temperatures
uj
** 4->
^S
51
10 O
o o
4-89
-------
for liquid fuels (Reference 4-62). The figures show dry controls to be not sufficiently developed
to meet the standards, whereas wet controls are sufficient.
Future NOX control in gas turbines is directed toward dry techniques with emphasis on com-
bustor design. Medium term (1979-1985) combustor designs incorporate improved atomization methods
or prevaporization and a premixing chamber prior to ignition. These developmental combustors are
projected to attain emission levels of 20 ppm NO at 15 percent oxygen. A possible long term option
is catalytically supported combustion. Laboratory tests have given NOX reductions of up to 98 per-
cent while maintaining stable, complete combustion. This concept - described in Section 3.1.5.2 of
this report -will probably require a new combustor design to accomodate it (Reference 4-57 and
4-62).
4.3.2.2 Costs
The most recent cost study of N0x controls for gas turbines has been performed by the EPA
(Reference 4-62). Based on information presented in this study, the best available system of emis-
sion reduction considering costs are the wet systems. Wet systems can be applied to turbines imme-
diately and their cost impact is minimal. Although dry control techniques may be preferable
because of their minimal impact on efficiency, their complete development and application to large
production turbines is still several years away. Cost considerations for dry methods are, therefore,
not discussed.
Table 4-32, derived from Reference 4-62, shows the expected increase in installed turbine cost
that will result from using water injection to control NOX to the proposed standard of 75 ppm. The
impact varies from 0.8 percent in the case of the 820 kW (1100 hp) standby unit to 7.1 percent for
the unit requiring extensive water treatment equipment.
Table 4-33 presents a summary of the costs in mills/kWh which would be incurred for 11
simple cycle turbine plants to meet the 75 ppm standard. This analysis was part of a cost model
developed in an EPA report (Reference 4-62). The results for each case are explained below.
Standby Units
The first two cases, S-l and S-2, differ only in the number of hours operated annually. Unit
S-l operates 80 hours and S-2 200 hours per year. These units show the highest percentage impact in
terms of the incremental costs per net kWh of power generation. The low number of hours operated
each year tends to increase the cost of producing power because fixed costs are spread over a rela-
tively small base. The estimated impact in both cases was roughly 2.4 percent.
4-92
-------
1ZU
100
E
o.
ci
o
g 80
0
cc
LLJ
a.
in
*~
h-
J 60
o
\
cc
H-
2
"J dO
2
O
X
o
2
?n
0
A'
0 J|
y
o
0
o
PROPOSED NSPS
1 KEY
-S-AVERAGE
O
O EPA TEST METHODS
0 OTHER TEST METHODS
NOTE: NO ADJUSTMENT FOR GAS
TURBINE EFFICIENCY
O
fl
! i
i i
-*r
A U
p-
-------
21
1
«fl r-
«- T
<4- !
oc
£
§
IO
1/1
3
o
C
f
CO
1
to-
- o
- <*-
- IO
5 cT
^.
1
co
1
ID
CM
1
=3
1
=
CO
1
»-H
CM
^
^
CO
CO
1
CO
CM
1
CO
1
CO
«
.c
^^ o co in co o*
3O CM CO , «« , O
o \
00 O O O O
*~
3 0 VO t- »- « r-
3E O O t O CO f
vo oo o o d d d
O VO r- i *
VO O CM t O CO .
vo O
CM O O O O
£ CO r r- «
O O CO
vo O . | |
VO O r O O O
in
3 CO i ^~
£ in t o co
o |
VO O CM O O O
VO CM
O CM O i CO V CH
3O r- O « VO O
5: o ... .
CO O O O O CD O
m
X O CM O r- CO O>
0 .- . 0 «» 0
CO O
CM O O O O O
2 o oo o p co
ZI C^ ^f1 r^ O ^f
o | |
CO CM O O O O
3 o co r^ i
* o r^ co o in
CM * . .
0 0 0 0
CO
3 o CM r-~ f t
-^ CO co co o in
0 ^ d d d '
00
VO IO r r
-^ O * * . **! | |
_ CM in o o o
o
VO
CM
in vo
_ o vo «» o in
3 CO
-^ CO O O O
o "~
VO
CM
c
S *- -2 "c
f- -O O -> 0)
4-> 41 *- u »-> E
10 x *J c o> >> t- S3
fc. «F- W»4)-^>*»p U
tt) I*- OECf-21 C
O. (3 +3 10 10 10
O T) 10 c c JT
41 W C7> 0) 1- O> - T-( «»_ h-
CO 4) r 4J >, 4->
i/i>- b- 3
4-> i. 34-> J. O) 41 f~ 4l4-> 25.
5001 co 0.10 10 c 10 o a
Z Q. <0 O3 3 UJ 3 (_> <5
CM CO CO
c^ in co
O CM CM
CO
en vo
t» co in
d vo i-^
CM
CM O 00
t«» CM 00
O CO
CO
en in in
^ *» r».
t in i-^
00
»» O Cft
o o vo
CO d r-^
co
^- CO r
CM in r*.
i in co
CO
r~ co in
in in r>»
O CM r
CO
CM f^ CM
O r-^ co
r CO CM
CM o\ 01
vo r-- tn
CM «» O
vo
CM
i cn in
CM CM CO
in r o
VO
^ W CO
. *
VO *T CVJ
CM
VO CM CO
W f CM
VO
4-> U
ix> 10
o o.
0 ^
Ol
C 4->
10 41 O
4-> 1/1 i.
O IO 41
(mo.
4-94
-------
00
o
o
D_
O
Q
CO
CM
S
o
O
CU
o
C
CU
O CU
»-« M-
00 CU
oo o:
00
LU
o -<
Z OO
QC
l^ =>
O h-
I 00
o <
. CM CM
to r>. in in
r CO CO
r
CO
U
Q.
^^"^ O 1 Qj
Q. ^I t.
^Z ^ N o
o a. .c
o o ^r to
1 If) i M_
CO r O *»-
' O O
o
CO fO *d-
"9 -gp ^^ '
-i* ^S ^» |X
CO CO cO
O O «i- 3; 3E
>> vo CM t. r: s:
.a CM co -i->
"O CO CO CO
C 3
JO
3:
4-> VO
r- VD
r
1 < .
ID
O
Q.
3
O
CO
fO
JJ
13
a.
4J
3
o
fO
o
U
CU
CU
4-93
-------
In summary, the resulting estimates showed that, except for standby units, the total change
in costs will probably fall in the range of 0.4 to 1.5 mills per kWh for turbines used in industrial
and utility applications. This cost is equivalent to about a 2 percent increase in operating costs.
Control costs for standby units were much higher, ranging from 2 to 14 mills per kWh. This is pri-
marily due to their low use factor. This cost is equivalent to approximately a 2.5 percent increase
in operating costs.
4.3.2.3 Energy and Environmental Impact
As was the case for reciprocating 1C engines, the energy impacts of applying NO controls to
gas turbines occur almost solely through effects on unit fuel consumption, which were noted in the
foregoing discussion. Dry controls, except for reduced air preheat applied to regenerative cycle
turbines, have insignificant effects on unit efficiency. On the contrary, wet controls can impose
energy penalties. Water injection at the rate of 1 kg H20/Kg fuel reduces turbine efficiency by
about 1 percent. If waste steam is available, steam injection can increase turbine efficiency by
increasing turbine power output at constant fuel input. But, if a fuel debit is taken for heat
needed to raise injection steam, overall plant efficiency losses comparable to those experienced
with water injection will occur.
Again, as with 1C engines, gas turbines emit only an exhaust gas effluent stream and fire
"clean" fuels. Thus the potential environmental impacts of NO controls applied to gas turbines
will occur through incremental effects on emissions levels of exhaust gas CO, HC, and particulate
(smoke). Effects through liquid and solid effluents need not be considered, and incremental
impacts on SOX, trace metal, and, to some extent, higher molecular weight organic emissions are
insignificant.
The effects of some commonly applied NOX control techniques on CO emissions from gas turbines
are summarized in Table 4-34. From the table, it is apparent that dry controls, notably leaning
the primary zone and air blast (or air-assist atomization) reduce CO levels. This is expected since
the additional air introduced into the combustor when applying these techniques allows more complete
fuel combustion. On the other hand, wet control techniques, such as water injection, tend to quench
combustion and give lower combustor temperatures. This leads to incomplete combustion and increased
CO levels as shown in Table 4-34.
The very limited data on incremental hydrocarbon emissions due to NOX combustion controls
applied to stationary gas turbines are summarized in Table 4-35. As the table shows, the effects
of dry NOX controls are mixed. Air blast tends to increase HC emissions while leaning the primary
zone tends to decrease HC levels. Increased combustion efficiency due to higher combustion
4-96
-------
Cases S-3 and S-4 are 820 kW (1100 hp) units operating the same number of hours, respectively,
as the smaller 260 kW units. These units can use exactly the same water purification system as the
smaller units. Since the costs of producing power independent of the water injection system (the
baseline cost) are identical between cases S-l and S-3 and S-2 and S-4, the percentage impact of
NOX control is decreased to less than one percent.
Industrial Units
Case 1-1 represents a normal, single shaft gas turbine application. The unit is operated
2000 hours per year and is slightly oversized. This negates any benefits that might be derived from
improved unit output. For Case 1-2, also a baseload turbine, a credit was taken for the improved
capacity of the unit.
The highest cost impact was recorded in Case 1-3, which represents a remote turbine applica-
tion in an arid climate in which water must be transported fifty miles at a cost of 2tf per gallon.
The impact in such cases, including water storage facilities, is approximately a 3.7 percent in-
crease in the average cost of generating power. Since water injection results in a slight increase
in the power output capacity of the unit, a credit of 0.05 mills per kWh was taken for the output
enhancement.
Utility Applications
The first unit operated 200 hours, the second 500 hours, the third 2000 hours, and the fourth
8000 hours annually. A credit for enhanced output was taken in the last case, since the unit is
baseloaded. In all four cases, the impact is less than 2 percent.
Offshore Drilling Platform
Initially, it was thought that this case would evidence the highest cost impact. The unit
was assumed to use sea water to fuel the water purification system, resulting in a substantial
increase in the capital and operating cost of the system. The installed cost of the water treat-
ment equipment was $27,000, compared to $14,000 for an onshore application. Despite these higher
costs, the availability of water offset the costs associated with transporting water to the remote
gas compressing station application (1-3). The total cost of water injection for the offshore plat-
form was 0.92 mills/kWh compared to 1.21 mills/kWh for the remote site.
In the EPA cost model, no attempt was made to provide detailed estimates of the control costs
for regenerative and combined cycle gas turbines. The cost impacts, in absolute terms, are not
expected to be much greater than for simple cycles. Indeed, the percentage impacts will be less,
given the higher cost per kW of generating capacity of these units.
4-95
-------
temperatures tends to support this latter observation. The effects of applying wet controls are
also mixed. As indicated in the table, with water injection at a water-to-fuel (W/F) weight ratio
of 0.5, HC emissions increased for turbines having high baseline HC emissions, but decreased for
turbines which emitted low baseline HC levels.
The data on particulate emissions from gas turbines resulting from applied NO controls are
also very limited and are as inconclusive regarding the increment in particulate emissions from
NOX controls as those for incremental CO and hydrocarbon emissions. For example, the effect of
water injection on particle emissions seems to be related to the specific injection method used
(Reference 4-62). Some tests show smoke level reduction of 1.5 to 1.75 smoke spot numbers when
water injection is used. Other tests, however, indicate increased particulate emissions with water
injection at peak load.
In summary, the limited data available on the incremental effects of NO controls on CO, HC,
and particulate emissions suggest that the control techniques do not significantly affect these
emissions. While dry techniques appear to decrease CO emissions and wet controls seem to increase
CO levels, even these data, as well as those on effects on HC and particulate levels, are inconclu-
sive.
4.4 SUMMARY
Table 4-36 summarizes current and emerging NOX control technology for the major source cate-
gories. These results show that current technology is dominated by combustion process modifica-
tion. Emerging technology is also centered around combustion modifications. Other approaches,
such as flue gas treatment, may be used in the 1980's to augment combustion modification if
required by stringent emission standards.
The level of combustion modification control available for a given source depends on the
importance of that source in regulatory programs. Utility boilers have been the most extensively
regulated and accordingly, the technology is the most advanced. Available technology ranges from
operational adjustments such as low excess air and biased burner firing to inclusion of overfire
air ports or low NOX burners in new units. Some adverse operational impacts have been experienced
with the use of combustion modification on existing equipment. In general these have been solved
through combustion engineering or by limiting the degree of control application. With factory-
installed controls on new equipment, operational problems have been minimal.
The technology for other sources is less well developed. Control techniques shown effective
for utility boilers are being demonstrated on existing industrial boilers. Here, as for utility
4-98
-------
TABLE 4-34. REPRESENTATIVE EFFECTS OF NOX CONTROLS ON CO EMISSIONS FROM GAS TURBINES (Reference 4-62)
NOX Control
Lean Primary Zone
Air Blast/
Piloted Air Blast
Water Injection
Fuel
Natural Gas
Kerosene
Diesel
Kerosene
Diesel
Natural Gas
Diesel
CO Emissions (ppm)a
Baseline
102
102
53
195
969
53
147
252
99
135
93
NOX Control
51
96
99
59
110
36
1134
1512
144
162
30
13% 02, dry basis. Emissions levels at full load.
TABLE 4-35. SUMMARY OF THE EFFECTS OF NOX CONTROLS ON VAPOR PHASE HYDROCARBON
EMISSIONS FROM GAS TURBINES (Reference 4-62)
NOX Control
Air Blast
Lean Primary Zone
Water Injection
Fuel
Jet-A
Natural Gas
Diesel Fuel
Kerosene
Natural Gas
Diesel Fuel
HC Emissions (ppm)a
Baseline
18
9
33
30
3
27
234
141
36
24
N0x Control
41
11
9 - 12
12
7
12
372
246
27
12
Comment
Idle
Full load
Full load
/ W/F = 0.5
3% 02, dry basis.
4-97
-------
o
a>
o
3
O
O
>-
o
_J
o
o
a:
o
o
o
>-
ro
i
3
0
5
o
o
u
Of
o»
I
en
o
**
w
c
x:
u
Of
H-
C
y
^*
o
«».
Ei
2~
*T 1
2 §
ES
Of O»
*TT
« P*
Of
Operational
Impact
Of ^3 O
*° s
^ ^
I* * 5
I/) :*" ^
UJ ^ C
0"^ <
e ~-+
at o r» 3
,2 .**
0 vt o»co
IE-O
* ui o> *
_ ^ ^^
"ot*'5
Or Of
XI O V
r- ***C
"5 0 U
i - i
s. 9 »
»- 3 S
3 ** «
2" °
J
i
if,
«*»
ESs
3? Of
U WO
o
g»
oo e
Of O
kJ * * M ^M
^i ^* IQ ^^
^F» Ml ^^ ^^
TJ 3 C U
zlis
o c
X 0»
e «>-
o c M
Of
3
c u. "c
*- O 3
Of 1
s§ .."s
Of o e -w
Of U COW i-
Ur- 0 C H-
XI C Of «« -^- O
'v*'*" «2 0, k. OI-M **
«/» M E ^ "O «O Of
o»- c S o» i- »-
O. ( «/» O «» O>
^
0^ jg
F-* O&
I .^
^^ ^^
~+
*?
fc
^^
O 1
» in
o
m o
^^
o
3
m*
l^m ^"» ^Q
Of w> 1- Of
^< 3 Of
O ^ ** »
^535
If
§0
5?**
*li
t.^"oi
3^«
X v
2|
"SC* §
fc. C*^
C-?Of
** U M -^>
"^It**
c?0^ e
Z 3 Ol
O Of Of Of
_i c c-o
01 1
S"§ S*c£
C Of XI t. 0 0
»« g <0 VI "0 Of
^-c5< cout.
1 M <*- XI XI '
^
^^ fm
^& ^^
Ot W
1 J4
f^ ^^
C*)
o
t CM
1
egS
oo
oca
:sc
UJ O
xu. S
oo c-o
» Of U. 3
01 1 1
VI C «
« o ^**
«
m x:
^9 ^^
1 Ol
2ii
in
CM
C9
O
f t"»
1
in o
oe«j
Ik O
* 4.
Ul UJ
*
** 3 C
V> ** 01
5C3
1
e
a
uS
3»
t
5 .
o-o
« e
u «
.* u
** C M
>» o T» e
0 4* O ^~
4-* M C l/t
« 3 « if
^ y XI ^ ^3
i. §
OJ U U
o-o o
Of S- X
01. o
C O VI Z M
«4J C »
!!£££
Of 1 ^i
v> e *-
« o e
Of O O
fe^gv,^
C Oft- «» «
^ 3 *» Cf E
«*- Q. Of C
C § «- -C
at
^^
CM
1
^T
^^^
in
en
o
SJ,
1 CM
O
-O
§
^
tSiJ
*3 oi -^»
5tJ-
M
I
1
s
5
n
e^a
W
U M
Of Of
1- **
^-
4rt-CM «w»^
So*
CM
*"^
i i
o
^^ tf)
0
CM
Ol
c
c*c .
01
i
o
o
o.
E
S
4-100
-------
ro
i
O)
u
c.
0)
0)
>-
o
_J
o
z
3:
o
O
O
o
>
CQ
O»
t
t. r*.
S
C U *4
O 41 41 X
oo
C
41 41 O O»
» VI C
ja - c jc
41 O C7> O Wl
M ^ S» C7)^3 *O
O C O C C
TJ X> TJ O-^
41 4) »
"C -0 "5 3 "c
44 fc "3 g g O
1 M
Q »-
f
» j§
U»*
O 6 44 .. I*
^J 4> C C C O> «
O C (/» f crt E *~
Z Q.-M T3 O T3 O O
wC
o
CM
1
O
^~
^^*
(O
0
o
VO 1
CM
i in
0
CM
u.
0 M
U
<: » C
UJ 41 3
i c.2
o
e
»
^-
ise
U *> 41
j5i
Z 3 2
c
^11
44 4> 3
- F- U >»
§§'sl
?* °
4> V* 44 C
Sill
u <
41 O»
? 41
r- 3 44
Ai t^ Aft
U
JC <0
*!£
TI
^- c
XVI C O
0 *- 4»-«-
Z 41-0 *4
i C «
X U^ U
0 3 t-
_l JO OH-
..
44* 1
41 i^">
Sa S v. o
"»/» 6 ** » +*
!o 2 i. o
"JJ S 2 ^ 44
en m u^ <0
O C O t.
a. <»- JQ xj
2
^^
CM
I
p^
**
^.^
CM
O
1
in m
CD **
in o
*o
o
1
O 41 C
^ *<^ ^^
~ ^» ^^
O "O *4
< u.
cS-^^i^
54,-i
o «- o o
c -
.. 3 ..i>_
XVI l/» »-
O U X C L.
Z If 4> C7I44
i c c »- e
X 1- vi C O
O 3 C at 41
_1 JO T3 -0 44
4> 1 I
cA C *O
10 0 44 t ~~*
cu y vi 44
i. C C > ^>
o» o c «»-
C 4) *- >> fc. O O
"*"»- CX g »- 44 44
J2 C 3^ "5 £ 2
I »- VI "4- 03 .0 ^^-*
AJ
^^ JJC
^^ ^^
Ol Ov
1 Jtf
»^
^^
m
o
CO '
CM
1
in o
ao
o
w»
o
< u. a tZ*
uj o o on
^~
£ J3
k 3 y»
»- 44 44 L.
H- VI V. 41
i^sl
O
a.
.*:
IB
4-99
-------
4-16 Crawford, A. R., e_t aJL, "Field Testing: Application of Combustion Modifications to Control
NOX Emissions from Utility Boilers," EPA-650/2-74-066, June 1974.
4-17 Selker, A. P. and R. L. Burrington, "Overfire Air Technology for Tangentially Fired Utility
Boilers Burning Western U.S. Coal", Proceedings of the Second Stationary Source Combustion
Symposium. Vol. II. Utility and Large Industrial Boilers, EPA 600/7-77-0735, July 1977.
4-18 Jain, L. K., et al., "State of the Art for Controlling NOX Emissions, Part I: Utility
Boilers," EPA-R2^72-072a, September 1972.
4-19 Hollinden, G. A.", et al_., "Evaluation of the Effects of Combustion Modifications in Controlling
NOX Emissions at TVA's Widow's Creek Steam Plant," EPRI SR-39, February 1976.
4-20 Friedrich, J. L., e_t al_., "Nitrogen Oxides Reduction," Foster Wheeler Energy Corporation,
EPRI SR-39, February 1976.
4-21 Ando, J., and H. Tohata, "NOX Abatement for Stationary Sources in Japan," Environmental Pro-
tection Technology Series, EPA-600/2-76-013, January 1976.
4-22 Lyon, R. K., and J. P. Longwell, "Selective, Noncatalytic Reduction of NOX with Ammonia,"
In: Proceedings of the NOX Control Technology Seminar, EPRI Special Report SR-39, February
1976. ~ '
4-23 Muzio, L. J., J. K. Arand, and D. P. Teixeira, "Gas Phase Decomposition of Nitric Oxide in
Combustion Products," In: Proceedings of the NQV Control Technology Seminar. EPRI Special
Report SR-39, February 1976.
4-24 Schreiber, R. J., e_t a\_., "Boiler Modification Cost Survey for Sulfur Oxides Control by
Fuel Substitution," Environmental Protection Technology Series, EPA-650/2-74-123, November
1974.
4-25 Frendburg, A., "Performance Characteristics of Existing Utility Boilers When Fired with
Low-Btu Gas," EPRI Conference Proceedings, Power Generation, Clean Fuels Today, April 1974.
4-26 Agosta, J. e_t al_., "Status of Low Btu Gas as a Strategy for Power Station Emission Control,"
presented at the 65th Annual Meeting of the American Institute of Chemical Engineering,
November 1972.
4-27 Martin, G. B., D. W. Pershing, and E. E. Berkau, "Effects of Fuel Additives on Air Pollutant
Emissions from Distillate Oil-Fired Furnaces," EPA, Office of Air Programs AP-87, June 1971.
4-28 Shaw, H., "Reductions of Nitrogen Oxide Emissions from a Gas Turbine Combustor by Fuel
Modification," ASME Trans., Journal of Engineering for Power, Vol. 95, No. 4, October 1973.
4-29 Kukin, I., "Additives Can Clean Up Oil-Fired Furnaces," Environmental Science and Technology,
Vol. 2, No. 7, July 1973.
4-30 Bartok, W., e_t aj_., "Systems Study of Nitrogen Oxide Control Methods for Stationary Sources -
Vol. II," prepared for National Air Pollution Control Administration, NTIS Report No.
PB-192-789, Esso Research and Engineering, 1969.
4-31
Blakeslee, C. E., and A. P. Selker, "Program for the Reduction of NOX from Tangential Coal-
Fired Boilers, Phase I," Environmental Protection Technology Series, EPA-650/2-73-005,
August 1973.
4-32 Kelley, D. V., data submitted at the EPRI NOX Control Technology Seminar, San Francisco,
by Pacific Gas and Electric Company, February 1976.
4-33 Letter from the Los Angeles Department of Water and Power to Acurex Corporation, May 5,
1975.
4-102
-------
boilers, the emphasis in emerging technology is on development of controls applicable to new unit
design. Advanced low NOX burners and/or advanced off-stoichiometric combustion techniques are the
most promising concepts. This holds true for the other source categories as well. The R&D emphasis
for gas turbines and reciprocating 1C engines is on developing optimized combustion chamber designs
matched to the burner or fuel/air delivery system.
REFERENCES TO SECTION 4
4-1 Lachapelle, D. G., J. S. Bowen, and R. D. Stern, "Overview of the Environmental Protection
Agency's NOX Control Technology for Stationary Combustion Sources," presented at the 67th
AIChE Annual Meeting, December 1974.
4-2 Teixeria, D. P., R. E. Thompson, "Utility Boiler Operating Modes for Reduced Nitric Oxide
Emissions with Oil Fuel Firing," presented at the 66th Annual Meeting of the Air Pollution
Control Association, Chicago, June 1973.
4-3 Bartok, W., e_t aj., "Systematic Field Study of NOX Emission Control Methods for Utility
Boilers," Esso R & E, Report GRU.4GNOS.71, December 31, 1971.
4-4 Barr, W. H., "Nitric Oxide Control -A Program of Significant Accomplishments," ASME Paper
72-WA/PWR-13, 1972. M
4-5 Blakeslee, C. E., "Reduction of NOX Emissions by Combustion Modifications to a Gas-Fired
250-MW Tangential Fired Utility Boiler," presented at Conference on Natural Gas Research
and Technology, Atlanta, Georgia, June 5-7, 1972.
4-6 Blakeslee, C. E. and H. E. Burbach, "Controlling NOX Emissions from Steam Generation,"
JAPCA, Volume 23, No. 1, January 1973.
4-7 Halstead, C. J., et al_., "Nitrogen Oxides Control in Gas-Fired Systems Using Flue Gas Recircu-
lation and Water.Injection," IGT/AGA Conference on Natural Gas Research and Technology,
Atlanta, Ga., June 1972.
4-8 Bagwell, F. A., et aj_., "Utility Boiler Operating Modes for Reduced Nitric Oxide Emissions,"
JAPCA, Volume 21, No. 11, November 1971.
4-9 Habelt, W. W., and A. P. Selker, "Operating Procedures and Prediction for NOx Control in
Steam Power Plants," presented at Central States Section of the Combustion Institute, Spring
Meeting, March 1974.
4-10 Norton, D. M., K. A. Krumwiede, C. E. Blakeslee, and B. P. Breen, "Status of Oil-Fired NOX
Control Technology," In: The Proceedings of the NQ₯ Control Technology Seminar, EPRI
Special Report, SR-39, February 1976.
4-11 Barr, W. H., F. W. Strehlitz, and S. M. Dalton, "Retrofit of Large Utility Boilers for
Nitric Oxide Emission Reduction - Experience and Status Report," presented at AIChE 69th
Annual Meeting, November 1976.
4-12 Crawford, A. R., et a].. "The Effect of Combustion Modification on Pollutants and Equipment
Performance of Power Generation Equipment," In: Proceedings of the Stationary Source Combus-
tion Symposium, EPA-600/2-76-152c, June 1976.
4-13 Thompson, R. E., M. W. McElroy, and R. C. Carr, "Effectiveness of Gas Recirculation and Staged
Combustion in Reducing NOX on a 560 MW Coal-Fired Boiler," In: Proceedings of the NOX Control
Technology Seminar, EPRI Special Report SR-39, February 1976. '
4-14 Selker, A. P., "Program for Reduction of NOX from Tangential Coal-Fired Boilers, Phase II and
I la," EPA 650/2-73-005a and b, June 1975.
4-15 Hollinden, G. A., "NOX Control at TVA Coal-Fired Steam Plants," Proceedings of Third National
Symposium, ASME Air Pollution Control Division, April 24, 1973.
4-101
-------
4-54 Heap, M. P., ejt al_., "Reduction of NO emissions from Package Boilers," Revised Draft Final
Report by Ultra Systems, Inc., Irvine, California.
4-55 McGowin, C. R., "Stationary Internal Combustion Engines in the United States," EPA-R2-73-210,
April 1973.
4-56 "Standard Support and Environmental Impact Statement Stationary Reciprocating Internal Com-
bustion Engines," (Draft Report). Acurex Corp./Aerotherm Division, Mountain View, California,
Project 7152, March 1976.
4-57 Aerospace Corporation, "Assessment of the Applicability of Automotive Emission Control Tech-
nology to Stationary Engines," EPA-650/2-74-051, July 1974.
4-58 The American Society of Mechanical Engineers (ASME), "Power Costs, 1974 Report on Diesel and
Gas Engines," March 1974.
4-59 Calspan Corporation, "Technical Evaluation of Emission Control Approaches and Economics of
Emission Reduction Requirements for Vehicles Between 6000 and 14000 Pounds GVW," EPA-460/73-
005, November 1973.
4-60 Bascom, R. C., et al., "Design Factors that Affect Diesel Emissions," SAE Paper 710484, July
1971.
4-61 Hills, F. J., et a]_., "CRC Correlation of Diesel Smokemeter Measurements," SAE Paper 690493,
May 1969.
4-62 "Standards Support and Environmental Impact Statement, Volume I: Proposed Standards of
Performance for Stationary Gas Turbines," EPA-450/2-77-017a, September 1977.
4-63 Shaw, H., "The Effects of Water, Pressure and Equivalence Ratio on Nitric Oxide Production
in Gas Turbines," ASME Paper 73-WA/GT-l.
4-64 Hilt, M. B. and Johnson, R. H., "Nitric Oxide Abatement in Heavy Duty Gas Turbine Combustion
by Means of Aerodynamic and Water Injection," ASME Paper 72-GT-53.
4-65 Stern, R. D., "The EPA Development Program for NOX Flue Gas Treatment," In: Proceedings of
the National Conference on Health, Environmental Effects, and Control Technology of Energy
Use, EPA 600/7/76-002, February 1976. " ~ ~~ "
4-104
-------
4-34 "Preliminary Environmental Assessment of Combustion Modification Techniques," Vol. II,
EPA 600/7-77-1196, February 1977.
4-35 "Technology and Economics of Flue Gas NOX Oxidation by Ozone," EPA 600/7-76-033, December
1976.
4-36 Kamo, R., e_t aJL , "The Effect of Air-Fuel Mixing on Recirculation in Combustion," Paper CP-
62-12, API Research Conference on Distillate Fuel Consumption, June 1962.
4-37 Hegg, D. A., ejt aj_., "Reactions of Nitrogen Oxides, Ozone, and Sulfur in Power Plant Plumes,"
EPRI EA-270, September 1976.
4-38 Richards, J. and R. Gerstle, "Stationary Source Control Aspects of Ambient Sulfates: A
Data Base Assessment," Pedco Final Report, EPA Contract No. 68-02-1321, Task 34, Pedco
Environmental, Cincinnati, OH, February 1976.
4-39 Bennett, R. L., and K. T. Knapp, "Chemical Characterization of Particulate Emissions from
Oil-Fired Power Plants," presented at the 4th National Conference on Energy and the Environ-
ment, Cincinnati, OH, October 1976.
4-40 Homolya, J. B., ejt al., "A Characterization of the Gaseous Sulfur Emissions from Coal and
Coal-Fired Boilers,^presented at the 4th National Conference on Energy and the Environment,
Cincinnati, OH, October 1976.
4-41 Hall, R. E., CRB, IERL, U.S. EPA, personal communication.
4-42 Locklin, D. W. , et^ aj_. , "Design Trends and Operating Problems in Combustion Modifications
of Industrial Boilers," NTIS PB235-712/AS, 1974.
4-43 Krippene, B. C., "Burner and Boiler Alterations for NOX Control," Central States Section,
The Combustion Institute, Madison, Wisconsin, March 1974.
4-44 Heap, M. P., e_t al_., "Burner Design Principles for Minimum NOX Emissions," EPA Coal Combus-
tion Seminar, Research Triangle Park, North Carolina, EPA 650/273-021, June 1973.
4-45 Lyon, R. K., "Method for the Reduction of the Concentration of NO in Combustion Effluents
Using Ammonia," LL. S. Parent No. 3,900,554, assigned to Exxon Research and Engineering
Company, Linden, New Jersey, August 1975.
4-46 Lyon, R. K. and J. P. Longwell, "Selective, Non-Catalytic Reduction of NOX by NH3," Proceed-
ings of the NOX Control Technology Seminar, EPRI SR-39, February 1976.
4-47 Campobenedetto, E. J., "The Dual Register Pulverized Coal Burner -Field Test Results,"
presented at Engineering Foundation Conference on Clean Combustion of Coal, Franklin Pierce
College, New Hampshire, July 31-August 5, 1977.
4-48 Teixeria, D. P., "Status of Utility Application of Homogeneous NOX Reduction," Proceedings of
the NOX Control Technology Seminar, EPRI SR-39, February 1976.
4-49 Cato, G. A., et^ al_., "Field Testing: Applications of Combustion Modification to Control
Pollutant Emissions from Industrial Boilers Phase 1," Environmental Protection Technology
Series, EPA-650/2-74-078-a,
4-50 Cato, G. A., e_t al_., "Field Testing: Application of Combustion Modification to Control
Pollutant Emissions from Industrial Boilers Phase 2," Environmental Protection Technology
Series, EPA-600/2-76-086a, April 1976.
4-51 Heap, M. P., et_ al_., "Application of NOX Control Techniques to Industrial Boilers," Ultra-
systems, Inc., presented at the 69th Annual Meeting of the AIChE, Nov. 28 to Dec. 2, 1976.
4-52 Cichanowicz, J. E., et_ aj_., "Pollutant Control Techniques for Package Boilers, Phase I
Hardware Modifications and Alternate Fuels," (Draft Report) Ultrasystems and Foster Wheeler,
November 1976.
4-53 Maloney, K. L., "Western Coal Use in Industrial Boilers," Presented at the Meeting of the
Western States Section of the Combustion Institute, Salt Lake City, Utah, April 1976.
4-103
-------
LU
Ou r
CO I
to
2:
O UJ
a; o
CO C£L
"Z. UJ
o u_
(4 UJ
co D:
co
cr.
Xr
o
z: o
a o
o uj
t-H rD
h- O
< cc:
2: D.
I
LO
to
r-
CO
00
CM
O
z
CO
z.
o
CO
CO
t I
LU tO
c
X O
o +->
z.
to
1 0
;D *-'
r^
Z. CJ>
< C3
o
en
LO
00
f_
0
00
o^
1
CM
£
f-*~
en
c
'£)
fD i
Z 2
o
CM
o
CO
CO
CM
(O
»-
c
0)
Q
r-
CM
LO
o
0
p
'
o
r>»
CTi
^
en
0
CM
LO
00
r
fl3
4->
o
h-
5-2
-------
SECTION 5
OTHER COMBUSTION PROCESSES
Significant amounts of the total fuels burned and NO emissions released in the United
States are associated with small-scale combustion processes. These include important nonindustrial
uses in domestic and commercial heating, hot water supply, a wide variety of incinerators, and open
burning of solid wastes. The contribution to ambient N02 can be significant, particularly in local-
ized, residential areas. Control techniques, costs, and energy and environmental impacts are dis-
cussed for those systems where .data are available.
5.1 SPACE HEATING
Residential and commercial space heating generated about 1.1 Tg (1.2 x 106 tons) of NO
during 1974, which accounts for approximately 9.0 percent of the total national stationary source
N0x emissions (see Section 2). Projections for nationwide emissions due to space heating have
been made by the National Academy of Sciences (Reference 5-1). These projections, shown in Table
5-1, assume that approximately half of new housing units will use electricity for space heating.
This assumption may be somewhat optimistic, from the standpoint of NO , due to the recent high
A
rate of increase in the cost of electrical heat compared to oil or gas firing. The decline over
the past 30 years in residential combustion, due to increased use of electrical heat, may reverse
if electrical heating costs continue to increase faster than fuel costs.
Hall, et al_. (Reference 5-2) studied the factors that affect emission levels from residential
heaters. This project, which concentrated on an oil-fired warm air furnace, showed that excess air,
residence time, flame retention devices, and maintenance are major factors in the control of
emissions.
As shown in Figure 5-1, emissions of CO, HC, smoke, and particulates pass through a minimum
as excess air is increased from stoichiometric conditions. By contrast, both thermal efficiency and
NO emissions pass through maximum points as excess air is increased. The experimental results showed
that increased residence time of the combustion products reduces emissions of CO, caseous HC, and
smoke but has no effect on NOX emissions. Combustion chamber material was found to affect all
5-1
-------
emissions. Furnaces with steel-lined chambers required higher excess air levels to reach Optimum
emission levels, thus reducing efficiency. The shape of the combustion chamber had little effect
on pollutant generation.
A specially designed flame retention device meant to decrease particulate emissions was
found to increase NO emissions, but such a device also increased furnace efficiency. Poor furnace
condition also yielded higher NO emissions.
5.1.1 Control Techniques
In a recent study of space heating equipment (Reference 5-3), emission levels were found to
be dependent upon boiler size, design, burner type, burner age, and operating conditions. The
type of fuel used in the combustion equipment for space heating is important since 40 to 60 percent
of the fuel nitrogen present was converted to NOX-
Currently available emission reduction techniques for space heating units are: (1) tuning:
the best adjustment in terms of the smoke-CO^ relationship that can be achieved by normal cleanup,
nozzle replacement, simple scaling and adjustment with the benefit of field instruments, (2) unit
replacement: installation of a new, advanced low-NO unit, and (3) burner replacement: installation
of a new low-emission burner.
Reference 5-3 indicates that tuning has a beneficial effect on all pollutants with the excep-
tion of NO . In the field program, oil-fired units considered in "poor" condition were replaced and
A
all others were tuned, resulting in reductions in smoke, CO, HC, and filterable particulate by 59,
81, 90, and 24 percent respectively, with no significant change in NO levels.
Hall, e_t a]_. (Reference 5-4) determined that gas-fired units exhibit emission levels similar
to an equivalent size high-pressure atomizing gun oil burner. Table 5-2 shows mean emission levels
prior to and after replacement or tuning. Although tuning or replacement has been shown to have
little effect on NO levels, yearly inspection accompanied by one of these techniques is highly
recommended since other pollutant levels are so greatly reduced.
Significant emission reduction can be achieved by burner retrofit replacement. Reference
5-3 found this procedure to produce significantly lower levels of CO and filterable particulate.
In general, recently developed burners have not demonstrated the ability to consistently reduce
NO levels while many, in improving combustion efficiency and reducing emissions of other pollutants,
actually increase NO emission over the standard burner.
5-4
-------
\OPTIMUM SETTING FOR MINIMUM
EMISSIONS AND MAXIMUM
SMOKE EFFICIENCY
(10TH
X
u
X
CJ
CJ
<
QQ
o
s
CO
o
CJ
1.6 1.8 2.0
STOICHIOMETRIC RATIO
Figure 5-1. General trend of smoke, gaseous emissions, and efficiency versus stoichiometric ratio
for residential heaters (Reference 5-4).
5-3
-------
A number of commercially available burners were tested by Hall (Reference 5-2) wherein pol-
lutant levels were determined under typical operation conditions. Combustion-improving devices
yielded higher NO levels than the standard burner, but demonstrated a potential for reducing
levels of one or more pollutants and improving combustion efficiency. Flame retention burners were
shown to be capable of operating at low excess air levels, resulting in increased combustion effi-
ciency with accompanied reduction in emission levels with the exception of N0x-
An advanced residential warm air oil furnace has been developed in an EPA-funded program
(References 5-5 and 5-6). The furnace is said to increase the fuel utilization efficiency by up
to 10 percent. In addition, a 65 percent reduction in NO emission levels was realized.
The advanced oil furnace design consists of an optimized oil burner and firebox combination.
The system has completed a 500 hour laboratory performance test. The tests evaluated the effects
of combustion air swirl angle, nozzle spray angle, and axial injector placement on NO emissions
levels for various oil flowrates and overall excess air combinations. The optimum burner was a
nonretention gun-type with six swirl vanes set at a 26-degree angle. The firebox design selected
was a cylindrical fin cooled firebox. The optimum burner/firebox combination yielded emissions of
0.6 g NO/kg of fuel (1.2 Ib/ton) at 10 percent excess air compared to 2-3 g/kg (4-6 Ib/ton) for the
baseline conmercial burners. Operation at these low excess air levels combined with use of outside
air for combustion produced up to 10 percent increase in system efficiency (Reference 5-6).
In a related study, Combs and Okuda (Reference 5-7) investigated the commercial feasibility
of an optimum low-NO distillate oil burner head. They reported that sheet metal stamping was the
best fabrication method for commercial production of the burner head. They also investigated retrofit
possibilities and found that the optimum burner heads were operationally satisfactory and had long
life potential.
Emissions of NO from natural gas-fired furnaces were measured by the American Gas Associa-
A
tion Laboratories (Reference 5-8). Measurements indicated that water-backed heat transfer systems
emitted higher levels of NO compared to gas-to-air systems. Also, multiport burners emitted
higher NO levels than single port burners. In another test, it was found that addition of a
radiant screen placed above a water heater burner resulted in a net reduction of NO by about 55
percent.
Another advanced burner/furnace design is the "Blueray" system (Reference 5-9). This system
consists of a "blue flame" oil burner integrated with the firebox of a warm air furnace package.
Two sizes are currently available: 0.63 cm3 oil/sec (0.6 gph), and 0.789 cm3 oil/sec (0.75 gph).
The efficiency of the burner is reported to be about 84 percent and the NO emission level is
5-6
-------
LU
Q
ii
OO
LU
a:
o
oo
o
ii
_i
o
u.
-------
discussed in Section 3.5, offers the potential for extremely low levels of N0x (1-10 ppm) when firing
natural gas or distillate oils. Catalytic combustion is still in the exploratory stage of develop-
ment and no reliable cost estimates are available for residential heating systems.
The prospects for cost-effective NOX control in existing space heating units are not promis-
ing. Furnace tuning and, if required, burner head replacement are strongly reconmended for reduc-
tion of carbon monoxide and smoke and for improving unit efficiency. The impact on NOX is negli-
gible, however. Furnace tuning (cleaning, leak detection, sealing and burner adjustment) costs a
minimum of $40 for the average residential unit. Burner head retrofit replacement costs an addi-
tional $25 less installation. These are usually cost effective in view of the fuel savings and
increased safety derived from the maintenance.
5.1.3 Energy and Environmental Impact
5.1.3.1 Energy Impact
All three NO emission reduction techniques (tuning, unit replacement and burner replacement)
result in improved system efficiencies and, consequently, reduced fuel consumption. The exact
amount of improvement varies widely depending on the type of equipment. The most promising method,
unit replacement, appears to offer in excess of 5 percent fuel savings. On a national basis, this
represents a potential savings of 0.6 percent of annual fuel consumption if all space heating equip-
ment were replaced with'new designs.
5.1.3.2 Environmental Impact
The effect of lower excess air on CO, HC, and particulate emissions was discussed previously
and is illustrated in Figure 5-1. By constraining incremental emissions during control development,
however, it has been possible to achieve low-NOx combustion conditions without adverse incremental
emissions (Reference 5-6). Table 5-3 shows a comparison of typical uncontrolled units and a proto-
type unit with an optimized burner/firebox. Incremental emissions were held constant or reduced
at the low-NO , high efficiency condition. Table 5-3 also shows incremental emissions with a com-
mercially available oil emulsifier burner. Again, low N0y operation was achieved with no adverse
effects on incremental emissions (Reference 5-13).
Over 90 percent of residential and commercial warm air furnaces fire either natural gas or
distillate oil. Emissions of sulfates and trace metals from these units are thus of minor concern
compared to coal-fired boilers. About 3 percent of U.S. warm air furnaces still fire coal. For
these, sulfates, trace metals and especially ROM's could cause severe localized environmental
5-8
-------
o
IISSIONS
:URNACES
2. U_
UJ
(__
P^
z 2:
uj a:
s: <
UJ 3
<_3 i
Z <
1 t t 1
h-
z z
O UJ
o
^J* > 1
O CO
i ct:
a; a:
LU O
DL. U.
o
UJ
>0
o z
z «=c
i s:
0 O
III
^ J LL.
ct:
U_ 1 1 1
O Q_
H- s:
O UJ
UJ 1
U. CO
U. 5-
UJ CO
JC
0) «
ac J.
J«o
J=
o
CO
1
x en
CD
1
-*
en
$
^-^ *^
x en
j^
en
x-^
a;
^> ^~*
_. U 10
*JI C 4-»
g 0) CO
0} o >»
2» *^* ^3
z <*- *
^ S- 0)
UJ +>
CO
«rt «
gi
x ^*>
UJ
CO
LO
UJ
_J
CO
,! cd Q.CC!
^^
t
^^^iM^Mi^MV^HHI
ro
O
in
CO
o
g
00
CM
m
p-
o
o S- O)
OJ O
p~ T" C
X> M- OJ
S- to 3 OJ
-------
problems. However, it is doubtful that NOX controls, except for fuel switching, will be developed
and implemented for these sources, and they will not be considered further here^.
An additional factor in evaluating incremental emissions for warm air furnaces is the cyclic
nature of operation. Warm air furnaces typically undergo two to five on/off cycles per hour. Studies
of emissions without NO controls show that the starting and stopping transients have a strong,
sometimes dominant, effect on total emissions of CO, HC and particulate (smoke) (References 5-2 and
5-3). The effect of NO controls on transient emissions is presently unknown. Incremental steady-
state emissions must eventually be weighed against the transient emissions for this significance
to be shown.
Data on warm air furnace POM emissions under low-NOx operation are apparently nonexistent.
Data on both transient and steady operation with and without NOX controls are needed to form a
general conclusion on the total incremental impact of N0x controls. Additionally, it should be
emphasized that the incremental emissions data shown in Table 5-3 are for well maintained laboratory
operation. Data are needed on long-term field operation with NOX controls.
5.2 INCINERATION AND OPEN BURNING
5.2.1 Municipal and Industrial Incineration
According to a Public Health Service survey conducted in 1968 (Reference 5-14), an average
of 2.5 kg (5.5 pounds) of refuse and garbage is collected per capita per day in the United States.
An additional 2 kg (4.5 pounds) per capita per day are generated by incineration of industrial
wastes, wastes burned in commercial and apartment house incinerators, and backyard burning. The
total per capita waste generation rate is conservatively estimated at about 4.5 kg (10 pounds) per
day (Reference 5-14).
Incineration is economically advantageous only if land is unavailable for sanitary landfill.
Incineration requires a large capital investment, and operating costs are higher than for sanitary
landfill.
The most common types of incinerators use a refractory-lined chamber with a grate upon which
refuse is burned. Combustion products are formed by contact between underfire air and waste on the
grates in the primary chamber. Additional air is admitted above the burning waste to promote burn-
out of the primary combustion products.
5-10
-------
Incinerators are used in a variety of applications. The main ones are municipal and indus-
trial solid waste management. Municipal incinerators consist of multiple chamber units that have
capacities ranging from 23 kg (50 pounds) to 1,800 kg (4,000 pounds).
5.2.1.1 Emissions
Nationwide N0x emissions from incineration in 1974 amounted to 39 Gg per year (43,400 tons
per year) which is 0.3 percent of the total N0x emissions from stationary sources. Fifty-five
percent of these emissions result from industrial incineration with the remainder due to municipal
incineration. A number of other multimedia effluents from incineration may be of greater concern
than NOX. These include metallic compounds in the particulate flyash and hopper ash and chlorinated
organic and inorganic gaseous compounds. Incinerator effluent rates are strongly dependent on the
composition of the solid waste, the incinerator design and specific operating variables such as
excess air and firing rate. The effluent rates can vary considerably from day to day because of
variations in refuse composition. An average emission factor for incineration of 1.5 g N02/ kg
refuse (3 Ib/ton) was reported by Niessen (Reference 5-15). AP-42 (Reference 5-16) specifies the
same value for multichamber industrial and municipal incinerators. For single chamber industrial
incinerators, a lower factor of 1 g N02/ kg refuse (2 Ib/ton) is specified.
Stenberg, et a].., conducted field tests to ^tudy the effects of excess combustion air on NO
x
emissions from municioal incinerators (Reference 5ll7). The nitrogen oxide emissions ranged from
0.7 g/kg (1.4 Ib/ton) to 1.65 g/kg (3.3 Ib/ton) of refuse charged for a 45.3 Mg (50 ton) per day
batch-feed incinerator and a 227 Mg (250 ton) per day continuous-feed incinerator. As shown in
Figure 5-2, NOX emissions increase with increasing amounts of excess air. The amount of underfire
air also has a significant effect on N0x production and is shown in Figure 5-3.
In general, nitrogen oxide emissions from incineration are not a primary source of air pollu-
tion; however, particulate emissions are significant. It is for this reason that incinerator air
pollution control equipment is adopted to the removal of particulate matter rather than NO
x'
Activity in pollution abatement for incinerators to date has focused on particulate control rather
than N0x.
5.2.1.2 Control Techniques
The use of waste disposal methods other than combustion may be the most likely means for
reducing N0x emissions, since the methods normally used for control of other emissions from inciner-
ation, such as particulate matter, organics, and carbon monoxide, tend to increase emissions of
5-11
-------
0.6
0.5
to
V)
CJ
X
o
oc
in
vt
0.4
cc
a
x
o
0.2
0.
OFURNACE - BEFORE SCRUBBER
NOX = 0.081 + 0.00144 (PERCENT EXCESS)
DSTACK -AFTER SCRUBBER
NOX = 0.093 + 0.00156 (PERCENT EXCESS)
100
200
300
EXCESS AIR, percent
Figure 5-2. Effect of excess air on NOX emissions from a 45.3 Mg (50 ton) per day batch-feed
incinerator (Reference 5-17).
5-12
-------
NOX = 0.365 - 0.00183 (PERCENT UNDERFIRE AIR)
UNDERFIRE AIR, percent
Figure 5-3. Effect of underfire air on NOX emissions from a 227 Mg (250 ton) per day continuous-
feed incinerator (Reference 5-17).
5-13
-------
NO . Other disposal methods include dumping, sanitary landfill, composting, burial at sea, disposal
A
in sewers and hog feeding.
One of the first refuse disposal methods used was open dumping of refuse on land. This
method is obviously very inexpensive, but extremely objectionable and offensive in and near popu-
lated areas.
Sanitary land fills are good alternatives, to the extent that land usable for this purpose
is available. Approximately 1233m3 (1 acre-foot) of land is required per 1000 persons per year
of operation for a waste production of 2 kg (4.5 pounds) per day per capita (Reference 5-18). In
addition, cover material approximating 20 percent by volume of the compacted waste is required;
the availability of cover material may limit the use of sanitary landfill.
5.2.1.3 Costs
At present, gaseous emission controls are not applied to incinerators. As described earlier,
only particulate emission controls are employed. Reference 5-19 presents estimated construction
costs in 1966 and operating costs for particulate pollution control.
5.2.2 Open Burning
Open burning includes forest wildfires, prescribed burning, coal refuse fires, agricultural
burning, and structural fires. Open burning for solid waste management is usually done in large
drums or baskets, in large-scale open dumps or pits and on open fields. Commonly, municipal waste,
landscape refuse, agricultural field refuse, wood refuse, and bulky industrial refuse are disposed
of by open burning.
5.2.2.1 Emissions
Emissions from open burning are affected by many variables including wind, ambient tempera-
ture, composition and moisture content of the debris burned, and compactness of the pile. Nitrogen
oxides emissions depend mainly upon the nitrogen content of the refuse. Generally, due to the low
temperatures associated with open burning, nitrogen oxides emissions are low.
Annual emissions from open burning vary from year to year, and the data for the various
sources are not entirely consistent. Table 5-4 shows the estimated NOX emissions from open burning
5-14
-------
sources for 1971 as reported in Reference 5-20. More recent estimates from the 1976 NEDS data file
and Reference 5-21 are also given in Table 5-4. Increasing awareness of air pollution problems
has contributed to a general decline in the quantity burned (and thus the emissions) from those
categories which can be controlled. For example, despite the continuing growth in crop harvest,
NOX emissions from agricultural open burning has declined from an estimated 29 Gg (32,000 tons) in
1969 to 13 Gg (14,300 tons) for 1973 (Reference 5-21).
TABLE 5-4. ANNUAL EMISSIONS OF NITROGEN OXIDES FROM OPEN BURNING
Source
Solid Waste Disposal
Forest Wildfires
Prescribed Burning
Agricultural Burning
Coal Refuse Fires
Structural Fires
NOX Emissions
1971,
Reference 5-20
Gg
150
138
19
29
31
6
103 Tons
165
152
21
32
34
7
1976 NEDS
Gg
95
48
30
13a
53
5
103 Tons
105
53
33
14a
58
6
1973 estimate from Reference 5-21.
5.2.2.2 Control Techniques
Solid Waste Disposal
From the standpoint of air pollution, sanitary landfills are good alternatives to open burn-
ing. In addition to the land necessary for sanitary landfill, cover material approximating 20
percent by volume of the compact waste is required. The availability of cover material may limit
the use of the sanitary landfill method.
Unusual local community factors may lead to unique approaches to the landfill site problem.
For example, Reference 5-22 reports that in a pilot project the refuse is shredded and baled for
loading on rail cars for shipment to abandoned strip mine landfill sites.
Other noncombustion alternatives may have application in some localities. Composting is
now being tested on a practical scale (Reference 5-23). Hog feeding has been used for disposal
of garbage. Dumping at sea has been practiced by some seacoast cities, but is now extensively regulated,
5-15
-------
Elsewhere, refuse has been ground and compressed into bales, which are then wrapped in chicken
wire and coated with asphalt. The high-density bales sink to the bottom in the deeper ocean areas
and remain intact. The practice of grinding garbage in kitchen units and flushing it down the sewer
has been increasing. This in turn increases the load of sewage disposal plants and the amount of
sewage sludge (Reference 5-24).
Forest Wildfires
122
In the United States, forests comprise approximately 3.2 x 10 m (786 million acres), or
34.4 percent, of the land area. Seasonal forest fires are quite prevalent in dry western regions.
Considerable activity has been and is being directed toward reducing the frequency of occurrence
and the severity of these fires. These activities include publishing and advertising information
on fire prevention and control, surveillance of forest areas where fires are likely to occur, and
various firefighting and control activities. Additionally, prescribed burning is being used to
reduce the loading of combustible underbrush and thereby decrease the fire hazard and potential
fire spread rate.
10 2
The U.S. Forest Service estimated that 2.06 x 10 m (5.11 million acres) of land were burned
in 1976 (the World Almanac, 1978). A similar estimate for 1971 (Reference 5-20) was 1.73 x 1010 m2
(4.28 million acres) burned, producing 138 Gg (152,000 tons) of nitric oxides emissions. Emissions
from forest fires are dependent on the local combustion intensity, the overall scale of the fire,
and, to some extent, the nitrogen content of the fuel. These in turn are related to the topography
of the forest, the composition and dryness of the underbrush, the local meteorological conditions,
and the elapsed time since a previous fire. The topography of the forest, the composition and dry-
ness of the underbrush, the elapsed time since a previous fire and the meteorological condition are
all interrelated and dictate the burn rate and spread, intensity of the burn, and the size of the burn.
Prescribed Burning
Prescribed burning is the use of controlled fires in forests and on ranges to reduce the pos-
sibility of wildfire and for other land management goals. Four classes of open burning operations
are traditionally practiced by the Forest Service (Reference 5-25):
Slash disposal resulting from forest harvesting operations
Forest management operations for forest floor fuel reduction, seedbed preparation, pest
control, forest thinning and undergrowth control
t Public works construction operations to clear reservoir and dam-sites, utility and high-
way rights-of-way and building and structure site areas
5-16
-------
t Public works maintenance operations for the disposal of reservoir driftwood and of
rights of way and storm damage debris
In addition, controlled burning is used to reduce unwanted quantities of waste and to improve land
utilization.
Because collection and incineration of these materials would tend to increase NO emissions,
the only current way to control emissions is to avoid combustion. In the future it may be possible
to develop incineration processes that can control N0x and other emissions such as particulate
matter, organics, odorous compounds, and carbon monoxide; or it may be possible to develop equipment
that can burn these materials as substitutes for fossil fuels.
Other alternatives to incineration are abandonment or burying at the site, transport to and
disposal in remote areas, and utilization. Abandonment or burning at the site is practical in cases
where no other harmful effects will ensue. Abandoned or buried vegetation can have harmful effects
upon plant life by hosting harmful insects or organisms, for example. Agricultural agencies such as
the U.S. Department of Agriculture, or state and local agencies should be consulted before these
techniques are employed.
Agricultural Burning
Agricultural burning includes the burning of residues of field crops, row crops, and fruit
and nut copes for at least one of the following reasons (Reference 5-21):
t Removal and disposal of residue at low cost
Preparation of farmlands for cultivation
Clearing to facilitate harvest
t Control of disease, weeds, insects, or rodents
Mitigation of the environmental impact of agricultural open burning is possible by proper
fire and fuel management (for example, single-line backfiring), meteorologically scheduled burning
to optimize dispersion, or by the substitution of other alternatives, such as mobile incineration,
incorporation into the soil, and mechanical removal. Care must be exercised in the choice of alter-
nate methods of disposal since a change in method may have significant adverse effects. For ex-
ample, in. situ burning can provide thermal treatment to the soil which raises the production yield
substantially, incorporation of the residue into the soil may restrict rapid replanting, and residue
decomposition may deplete the soil nitrogen.
Coal Refuse Fires
An estimated 53 Gg (58,000 tons) of NOX is emitted each year from burning coal refuse.
Extinguishing and preventing these fires are the techniques used for eliminating these emissions.
5-17
-------
These techniques involve cooling and repiling the refuse, sealing refuse with impervious material, in-
jecting slurries of noncombustibles into the refuse, minimizing the quantity of combustibles in refuse,
and preventing ignition of refuse. The NO emissions from coal refuse fires are highly dependent on
A
the nitrogen content of the coal.
Structural Fires
There were almost one million buildings attacked by fire during 1971 with losses estimated at
$2.21 billion (Reference 5-20). An estimated 6.3 Gg (7,000 tons) of NOX were emitted during 1971.
Prevention is the best control technique to reduce these emissions. Use of fireproof construction,
proper handling, storage, and packaging of flammable materials, and publishing and advertising infor-
mation on fire prevention are some of the techniques used to prevent structural fires.
Fire control techniques include the various methods for promptly extinguishing fires: use of
sprinkler, foam, and inert gas systems; provision of adequate firefighting facilities and personnel;
provision of adequate alarm systems. Information on these and other techniques for prevention and
control are available from agencies such as local fire departments, National Fire Protection Asso-
ciation, National Safety Council, and insurance companies.
5.3 INDUSTRIAL PROCESS HEATING
Fossil fuel derived heat for industrial processes is supplied in two ways: (1) by direct
contact of the raw process material to flames or combustion products in furnaces or specially-
designed vessels, and (2) by heat transfer media (e.g., steam, glycol or hot water) from boilers and
I.C. engines. NO emissions and control techniques for the latter equipment types have been de-
A
scribed in previous sections of this document. The former equipment types are described in the
present section. Industries covered include petroleum and natural gas, metallurgical, glass, cement,
and coal preparation plants. Much of this section is taken directly from a recent study of indus-
trial process heating performed by the Institute of Gas Technology (Reference 5-28).
There is currently very little application of NO control to industrial process heating equip-
ment. Consequently there are very few data on MO control costs or energy and environmental impact,
and separate sections for these topics will not be included. EPA's Industrial Environmental Research
Laboratory (RTP) is sponsoring a field test program to identify the potential for NO control in a
diversity of process furnaces, ovens, kilns, and heaters. Partial results from that study are given
in Reference 5-26 and are discussed, as appropriate, in the following subsections. The complete
results of that program (scheduled for 1978) will provide a broad data base on which to evaluate
alternate control options.
5-18
-------
5.3.1 Petroleum and Natural Gas
5.3.1.1 Process Description
Oil and gas production, gas plants, and pipeline stations are usually located in remote areas
far from population centers. Emissions do not, therefore, contribute substantially to ambient N0?
levels in populous areas. Petroleum refineries, however, are often located in or near densely popu-
lated areas.
Petroleum refining is the process of converting crude oil into salable products. Crude oil
is charged to an atmospheric pipestill where light products are separated and taken overhead and
light catalytic reforming feed, raw gasoline, kerosene, middle distillate, and heavy gas oil are
taken as sidestream products. The reduced crude is charged to a vacuum pipestill where heavy gas
oil, lube stocks, and residuum are cut.
Atmospheric and vacuum gas oils are charged to catalytic cracking units, which provide light
ends, cracked gasoline, and fractions for blending distillate and residual fuels. Reduced crude is
used in making asphalt or residual fuels, and is often fed to coking units to increase the yield of
distillate products. Catalytic cracking and coking produce propylene and butylene, which are often
alkylated with isobutane to make alkylate. Sometimes the olefins are polymerized for gasoline or
chemical production. Catalytic reforming increases the octane number of naphtha by converting
naphthenes (saturated cyclic hydrocarbons) and paraffins to aromatics. Hydrogen treating is used
to reduce sulfur content, increase stability, and improve burning characteristics of kerosenes and
middle distillates.
The relative volumes of gasoline, kerosene, middle distillate, heavy fuel oil, etc., can be
adjusted by diverting heavy gasoline fractions from gasoline to middle distillate and cat-cracking
feed, by diverting coker feed to heavy fuel, and by other changes.
A fluid-bed catalytic-cracking unit is often the heart of a modern refinery. Preheated gas
oil is charged to a moving stream of hot regenerated catalyst while it is being transferred from
the regenerator to the reactor. The gas oil is cracked in the reactor or the tube inlets to the
reactor; the products then pass through cyclone separators for removal of entrained catalyst and are
cut into products in a fractionator. Coke forms on the catalyst during the reaction.
Spent catalyst is withdrawn from the bottom of the reactor and transferred to the regenerator
where coke is burned off. The regenerator flue gas passes through cyclone separators for catalyst
removal and is discharged through the stack. The hot, regenerated catalyst flows back to the
reactor, supplying heat and catalyzing the cracking reaction.
5-19
-------
The regenerator flue gas contains about 10 percent carbon monoxide. This gas is sometimes
fed to a CO boiler where it is burned in preheated air to generate steam. Auxiliary fuel is
required to maintain satisfactory combustion conditions.
Typical refinery process heaters are the cabin-type furnace, used for heat release rates
above 44 MW (150 x 106 Btu/hr), and the vertical cylindrical furnace, used for heat duties below
23 MW (80 x 106 Btu/hr). Either type may be used in the 23 to 44 MW (80 to 150 x 106 Btu/hr)
range. Combustion boxes are lined with refractory. Fuels may be liquid, gas, or a combination of
both. Gas burners operate with 10 to 40 percent excess air, liquid burners with 20 to 80 percent.
Stack temperatures are 478K to 756K (400F to 900F).
5.3.1.2 Emissions and Control Techniques
Process Heaters
Oxides of nitrogen emissions in the petroleum and natural gas industries result from the com-
bustion of fuel in process heaters and boilers, and from internal combustion engines used to drive
compressors and electric generators. Annual NOX emissions for 1974 from petroleum process heaters
are estimated to be 147 Gg (162,000 tons). N0x control techniques for these sources are described
in Section 4.2.1 of this report.
Recent test data (Reference 5-26) on NO emissions from both natural draft and mechanical
draft heaters are summarized in Table 5-5. Five vertically fired natural draft heaters ranging in
size from 11 to 26 MW (36 to 87 x 106 Btu/hr) were tested. These units had 10 to 32 burners sized
about 940 ± 140 kW. Baseline NO emission factors, which were in agreement with the findings of
Bartz (Reference 5-27), ranged from 39 to 52 ng/J (90 to 120 lb/109 Btu), considerably lower than the
EPA emission factor for this category of 95 ng/J (220 lb/109 Btu). Combustion modifications for these
tests included fuel heat content variation, load variation, burner air register adjustment and BOOS.
Prior work by Bartz (Reference 5-27) had attributed large changes in NOX emissions to fluc-
tuations in fuel gas composition. However, the tests reported in Reference 5-26 indicate that NOX
does increase with increased fuel heating value, although not to a significant degree. The results
are not conclusive, and more tests with different heaters and a wider variation in heating value
are needed.
Two of the natural draft heaters were tested during process rate changes of ±20 percent.
Figure 5-4 shows the observed decrease in NO emissions as the load is increased. The probable
cause for the NO reduction is that excess air was reduced as the load was increased.
5-20
-------
o
u. vo
csj
XLU
o a:
XCJ
o o
Z. OL
o z:
OO LU
E
C
f
X
ro
to U
.0 C
O ^3
00
c
o
<- x
4-> O
u z
=1
o c
01 v
in
ro
to
ro
CD
.
«4-
Ol
01
rocess
0.
+j
i^
ro
i.
O
i.
i ttJ
ro 4J
i. (0
3 Ol
+> 3:
z
L.
Ol 4->
ct: to
^
53
csj
csj
fs^
Cn
O
in
(/)
ro
CD
.
<4-
Ol
IX.
rocess
o.
^j
4-
ro
t.
0
i.
r- 01
ro ->
3 QJ
4-> "T"
ro
O
J
O
T3
ro
O
^^
^
^«
CSJ
VO
f"^
en
oo
to
ro
CD
.
Oi
rocess
Q.
4j
t- ro
2 Ol
-i-> 3:
ro
CO
O
O
CO
VO
csj
00
CO
0
CSJ
in
to
ro
CD
.
a;
C£
rocess
Q.
(->
(4_
ro
i.
o
S-
Oi
ro »->
i- ro
3 Ol
!-» I
ro
I-
01
ij
(/I
*CT
Ol +>
DC V)
13
53
^^
csi
00
CD
(/)
ro
CD
.
0)
OL
rocess
Q.
4J
(^
ro
^
Q
L.
r- Ol
ro 4-1
1- ro
Z3 Ol
+J ^n
ro
CSJ
Oi O
r~ ^ '
r~ (/)
Ol Ol
to 2
ro O
CO _l
0
o
CSJ
CO
co
*
l/>
ro
CD
.
Ol
OL
*
o>
4J
ro
O)
Z 4j
4J ro
«4_ a,
ro -C
i- Ol
a &-
a.
TJ
0) S-
O -i-
J- «t
0
U-
u
ai
*4
UJ
o
csj
o
o
o
csj
CSJ
CSJ
co
^
^.
o
V£>
O
CSJ
O
0
f"^.
WD
^^
csi
CT>
o
to
ro
CD
0)
OL
1-
a;
M
ro
Oi
Z
M
4-
ro
Q
"T5
Ol
u
&.
o
u_
CSJ
0
5
00
f^^
Jf2
oo
00
r_
o
VO
o
z.
5-21
-------
120
1.50
PROCESS RATE, 1000 m3/d
1.75
2.00
100
QC
a
o
QC
X
o
80
60
40
REFINERY GAS FUEL
10
12
PROCESS RATE, X 103 BARRELS/d
14
Figure 5-4. Effect of process rate on NOX emissions from a process heater
[Reference 5-26).
5-22
-------
Staged combustion for NO control was attempted by adjusting the air registers and taking
burners out of service. Although NO emissions were reduced (see Table 5-5), the natural draft
heaters did not respond to burner adjustments as well as expected, based on previous results with
other types of boilers. There are three primary reasons for this. First, in most cases the burner
removal patterns resulted in increased excess air which could not be lowered to baseline levels be-
cause of furnace pressure limits. Secondly, the design of natural draft burners utilizes the fuel
flow as an aid to induce combustion air. This acts to defeat the attempt to achieve staged com-
bustion. Finally, the in-line vertically-fired burner arrangement used for most heaters does not
provide much inter-burner mixing, a necessary feature of staged combustion.
Two mechanical draft heaters, one with air preheat, were tested while firing either process
gas or Number 6 fuel oil. Both were vertical cylindrical types and had only one burner; therefore,
the only possible combustion modification was variation of excess air. As shown in Table 5-5, the
unit with air preheat and higher emissions for both oil and gas firing. For both units, changes in
excess air had little effect on NO emissions when firing oil. For the unit without air preheat,
reductions in NO from 64 ng/J to 36 ng/J were achieved in one test with refinery gas when the
excess oxygen was reduced from 5 percent to 2 percent.
Catalytic Crackers and CO Boilers
NOX is also released from the catalytic-cracking regenerator and from CO boilers, which are
fired by the catalytic cracker off gas. Emission testing in CO boiler stacks, summarized in Table
5-6, has shown results ranging from 100 ppm to 230 ppm of NO . Hunter (Reference 5-26) reported
testing one CO boiler that was equipped with staged air ports. Baseline emissions were 126 ppm.
Lowering excess oxygen from 2.1 to 1.8 percent reduced NOX by 8 percent. Adjustment of the air ports
and BOOS had negligible effect on NOX emissions. CO emissions, however, were very sensitive to
excess air and increased rapidly below about 2 percent excess oxygen. The lack of response of NOX to
combustion modifications is attributed to NO that is formed from ammonia in the CO gas feed acting
similarly to fuel nitrogen in oil or coal.
The average emission factor for NO from fluid catalytic cracking units is estimated in
Reference 5-16 as 0.24 kg N02/liter feed (84.0 lb/103 bbl feed). The total nationwide annual
emissions from fluid bed and thermal cat crackers is estimated in Section 2 to be 45 Gg (50,000 tons)
in 1974. If the regenerator exhaust is burned in a CO boiler, the resulting NO emissions can pre-
sumably be controlled by the classical methods discussed in Section 4.2.1 of this report.
5-23
-------
TABLE 5-6. NOX EMISSIONS FROM PETROLEUM
REFINERY CO BOILERS (REFERENCE 5-28)
Investigator
NOX
(ppm as measured)
Schulz, et. al., (Reference 5-29)
Schulz, et. al., (Reference 5-30)
Shea (Reference 5-33)
Shea (Reference 5-31)
Cowherd (Reference 5-32)
104-116
(average 106)
70-89
(average 78)
96-233
(average 163)
101-159
(average 135)
108-162
(average 129)
5-24
-------
5.3.2 Metallurgical Processes
5.3.2.1 Process Description and Control Techniques
The iron and steel industry is the predominant source of NOX emissions derived from metallur-
gical processes. Other industries, such as aluminum production, extensively use electric melting
furnaces or operate the process equipment at temperatures below the minimum required for formation
of significant quantities of NO . Copper, lead, and zinc smelting require combustion operation in
the reverberatory furnaces and converters (copper) and in sintering machines (lead and zinc). These
combustion emissions are deemed insignificant relative to the emissions from the iron and steel
industry. Emissions from these other industries may become significant as a result of the trend
toward higher melting rates in new equipment designs. This section reviews the equipment types and
available NO control technology for the major sources of NO within the iron and steel industry.
Section 5.3.2.2 summarizes NO emission factors for these equipment types. Major portions of this
A
section are taken from a 1976 IGT study (Reference 5-28) which uses 1971 steel industry data as a
source for fuel consumption and NO emissions estimates.
Pelletizing
Pelletizing of extremely fine low grade iron ore occurs in a specially designed furnace at
or near the iron mine. The cost of shipping the unbeneficiated ore would be almost double that of
the pelletized product.
Previous studies by the Institute of Gas Technology have shown that pelletized ore production
will be about 54 Tg per year (60 million tons/yr) by 1985. The fuel consumed by the pelletizing
furnaces has remained about constant at 0.7 MJ/kg (600,000 Btu/ton). This indicates that annual NO
emissions from pelletizing furnaces will reach about 7.65 Gg (8,500 tons) by 1985. The steel industry
and equipment builders are considering coal firing the pelletizing furnace combustion chambers. If
this is done, it will probably bring about an increase of about 50 percent in NO emissions. There
is no information available concerning NO control techniques for pelletizing furnaces (Reference 5-28)
Sintering
Some of the iron ore and flue dusts are available in particle sizes too small to be charged
directly to the blast furnace. These particles are mixed with flux and coke breeze and loaded onto a
traveling grate-sintering machine. An auxiliary fuel such as natural gas, coke oven gas, or oil is
used to initiate combustion on the surface of the mixture and is referred to as ignition fuel. Com-
bustion is continued over the length of travel by forcing air through the mixture on the grates.
5-25
-------
The mixture is heated to a fusion temperature, which causes agglomeration of the iron-bearing par-
ticles. The discharged sinter is cooled, crushed, and screened prior to transfer to the blast fur-
nace charging oven.
The major source of energy used in the production of sinter is the carbon content of coke
breeze and flue dust. The amount of ignition fuel required is about 140 J/g (0.12 million Btu per
ton) of sinter. The total fuel requirement, including coke breeze, is about 1.74 kJ/g (1.5 million
Btu per ton) of sinter.
The use of sinter machines to agglomerate ore fines, flue dust, and coke breeze has been
declining since 1966 and amounted to 39 Tg (43 x 10s tons) in 1971. If the present rate of decline
continues, the 1985 production of sinter would be about 24.3 Tg (27 x 106 tons). The attitude of
the steel industry is mixed because many steel plants are phasing out sinter lines, while at least
one major producer has replaced several small sinter lines with a large machine designed to meet
pollution control regulations. On the other hand, the use of sintering for recycling iron has
simultaneously been increasing. Therefore, the projected decrease in the number of sinter machines
may not occur. In any case, the IGT estimates (Reference 5-28) show that NOX will continue to be a
major pollutant. There is no information available concerning NOX emission control techniques for
these furnaces.
Blast Furnace
The blast furnace is the central unit in which iron ore is reduced, in the presence of coke
and limestone, for the production of pig iron. The blast furnace itself is normally a closed unit
and therefore has no atmospheric emission. A preheated air blast is supplied to the furnace from
the blast furnace stove, through nozzle-like openings called tuyeres. The subsequent reactions in
the blast furnace are not pertinent to this discussion. Excellent descriptions are available, how-
ever, such as the complete discussion of the process of changing raw ore to finished steel published
by the United States Steel Corporation (Reference 5-34).
The hot blast reacts with the coke to produce heat and more carbon monoxide than is needed to
reduce the ore. The excess CO leaves the top of the blast furnace with other gaseous products and
particulates and is known as blast furnace gas. This gas is cleaned to remove the particulates, which
could later cause plugging. It is then available for heating purposes. Blast furnace gas contains
about one percent hydrogen and 27 percent carbon monoxide; it has a heating value of approximately
3600 kJ/Nm3, or, 92 Btu/ft3 (Reference 5-34).
5-26
-------
Coke Ovens
Coke is an essential component in making pig iron and steel; coke ovens are generally an
integral part of the steel plant complex. One-sixth of the total bituminous coal produced is charged
to coke ovens. On the average, 1.4 kg of coal is required for each kilogram of coke produced.
Conventional coking is done in long rows of slot-type ovens into which coal is charged
through holes in the top of the ovens. The sidewalls, or liners, are built of silica brick, and the
spaces between the chambers are flues in which fuel gas burns to supply the required heat. Each
kilogram of coal carbonized requires 480 to 550 kJ (450 to 520 Btu). Flue temperatures are as high
as 1.753K or 2,700F (Reference 5-35). Much of the remaining heat in the partially spent combustion
gases is accumulated in a brick checkerwork, which releases it to preheat the combustion air when
the cycle is reversed. This is a typical regenerative cycle to conserve fuel and give a higher flame
temperature.
The coal in the coking chambers undergoes destructive distillation during a heating period of
about 16 hours. The noncondensable gaseous product is known as coke oven gas and on a dry basis
has a heating value of about 22 MJ/Nn3 (570 Btu/ft3). Approximately 35 percent of the coke oven gas
produced is used in heating the oven.
The major sources of emissions from coke ovens are the rapid evolution of steam and other
gases when moist coal is charged, the discharge of gases and particulates from the charging openings
during charging, and the emissions during the coke push and subsequent quencing. Recent coke-oven
battery designs have reduced the emissions from charging and pushing by using advanced engineering
features and improved operating procedures. During the coking process, leakage from the push side
and coke side door seals can account for most of the emissions during the coking process itself.
Improved door sealing techniques reduce door leakage substantially.
Although the current practice of firing coke ovens with a mixture of blast furnace gas and
coke-oven gas and slow mixing in the combustion chambers should tend to minimize NO production,
the estimated total is substantial because of the large quantity of fuel consumed.
The reduction in the coke required per kilogram of hot metal achieved during the 1960's will
continue, but steel mills are currently installing new coke ovens because of the increased need for
hot metal due to the high BOF*hot metal-scrap ratio. It is believed that the decline in coke rate
may have been stopped by the increased cost of fossil fuels used as injectants. The 1985 projection
for coke-oven underfiring fuel is 485 PJ (458 trillion Btu). If the NOX concentration remains con-
stant, the resulting total emissions of NOX will reach 57.8 Gg (64,120 tons) per year.
Basic Oxygen Furnace
5-27
-------
Although it may be reasonable to assume that substitution of form coke may result in a sub-
stantial reduction in NOX production, the general opinion in the steel industry,is that form coke
will not be a significant factor in 1985 (Reference 5-28).
Blast Furnace Stove
Between 2.2 and 3.5 kg of blast furnace gas is generated for each kilogram of pig iron pro-
duced. Some 18 to 24 percent of this gas is used as fuel to heat the three stoves which are usually
associated with each blast furnace. Two are generally on heat while the third is on blast.
The blast furnace stove is a structure about 8 to 8.5 m (26 to 28 feet) in diameter and
about 36 m (120 feet) high. A roughly cylindrical combustion chamber extends to the top of the
structure and the hot combustion gases pass through a brick checkerwork to the bottom by reverse
flow and then to the stack. The checkerwork usually contains 25,500 m2 (275,000 ft2) of heating
surface and has about 85 percent thermal efficiency. Unlike the conventional regenerators, which
extract heat from the waste combustion gases, the blast furnace stove is heated by burning fuel.
The stored heat is then used to preheat air for the combustion of fuel in the furnace to be served.
As in the case of coke oven underfiring, the blast stoves require very large quantities of
fuel for heating. However, since the stoves are heated primarily with blast-furnace gas (3.0 to
3.5 MJ/Mn3, or 80 to 95 Btu/ft3) the NOX concentration is lower due to the presence of diluents and
a low flame temperature.
The projected need for hot metal in 1985 is 112 Tg (124 million tons). This amount of hot
metal will require 295 PJ (280 trillion Btu) for blast-stove heating. Assuming no reduction in NO
stack-gas concentration, the NOX emission in 1985 will be 17.7 Gg/yr (19,600 tons/yr). Because of
the low estimated NOX concentration and the presence of inerts in the fuel gas, equivalent to flue-
gas recirculation, the potential for NOX reduction is probably small (Reference 5-28).
Open Hearth Furnace
Steel making by the open hearth process has been decreasing since it reached a peak in 1956,
when it represented 90 percent, or 92.7 Tg (103 million tons), of the total production. The use of
open hearth furnaces is expected to continue to decline and will probably amount to about 10 percent
of total steel production by 1985. Regardless of this dramatic decline due to the inroads of the
basic oxygen furnace (BOF) and electric arc furnace steelmaking processes, its NOX emission poten-
tial deserves consideration.
The open hearth furnace is both reverberatory and regenerative, like the glass melting fur-
naces. It is reverberatory in that the charge is melted in a shallow hearth by heat from a flame
5-28
-------
passing over the charge and by radiation from the heated dome. It is regenerative in that the
remaining heat in the partially spent combustion gases from the reverberatory chamber is accumulated
in a brick filled chamber, or "checker", and released to preheat the incoming combustion air when
the cycle is reversed. Fuel of low calorific value such as blast furnace gas as well as the com-
bustion air may be preheated by the checkers in order to obtain the high temperatures required.
Hot metal from the blast furnace, pig iron, scrap iron, and lime are the usual materials
charged to an open hearth furnace. These are heated over a period averaging 10 hours, at a tempera-
ture as high as the refractories will permit. Fuel oil is the preferred fuel and is burned with
excess air to provide an oxidizing influence on the charge.
NOX emissions from open hearth furnaces are very high because of the high combustion air pre-
heat temperature, high operating temperature, and the use of oxygen lances to increase production
rates. The data available indicate that NOX concentrations will be in the 1000 to 2000-ppm range.
Although many open hearths are being phased out because of emission control difficulties and better
economics of steel production with the BOF process, several steel mills are modernizing open hearth
shops, including pollution control equipment to provide flexibility in the hot metal-scrap ratio,
particularly those mills with a hot-metal deficiency. Therefore, predictions that the open hearths
will be phased on entirely by 1985 are unrealistic, and it is anticipated that about 13.5 Pg (15
million tons) will still be made by the open hearth process in 1985. Fuel consumption has been
decreasing and may reach 2.9 MJ/kg (2.5 million Btu/ton) in 1985. This will require a fuel con-
sumption of 40 PJ (37.5 trillion Btu) for open hearth steel production and result in an NOX emis-
sion level of 14 Gg (15,750 tons) (Reference 5-28).
Basic Oxygen Furnace
In the basic oxygen furnace (BOF), oxygen is blown downward through a water-cooled lance into
a bath containing scrap and hot metal. Heat produced by oxidation of carbon, silicon, manganese, and
phosphorous is sufficient to bring the metal to pouring temperature and auxiliary fuel is not required,
The furnace is an open top, tiltable, refractory-lined vessel shaped somewhat like the old-fashioned
glass milk bottle. Furnace capacities range up to 309 Mg (340 tons). The time required per cycle is
very short - from 45 to 60 minutes.
The BOF has displaced the open hearth as the major steel production process, but is much less
flexible because of the inherent limitation of 25 percent to 30 percent scrap in the charge. The
5-29
-------
amount of BOF capacity in an integrated steel plant is, therefore, closely associated with hot metal
availability. Additional flexibility in scrap use can be obtained by preheating the scrap with an
oxygen-fuel burner. In many steel plants, the open hearth shop is modernized and equipped with
appropriate pollution control equipment so that it can be used in conjunction with BOF shops to
provide the required flexibility to accommodate variations in hot metal-scrap ratio. A combination
of BOF shops and electric furnace shops provides the maximum in flexibility and may represent the
makeup of future steelmaking facilities.
Excluding fuel use for scrap preheating, other uses are for refractory dryout and to keep
the BOF vessel from cooling between heats. Their uses amount to about 232 kg per kg (200,000 Btu
per ton) of steel produced.
Decarburization of the iron charged to the BOF produces about 467 kJ of carbon monixide per
kilogram of steel (400,000 Btu/ton). The off-gases also contain large amounts of particulates,
which must be removed before discharge into the atmosphere. Typical American practice is to burn
the combustible gases in water-cooled hoods mounted above the BOF vessel, cool with excess air or
water sprays, and pass the cooled gases through high-energy scrubbers or electrostatic precipitators.
In most cases, the BOF vessels are equipped with open hoods that admit air for combustion of carbon
monoxide on a relatively uncontrolled basis. If additional steam can be used in the plant, the
combustion hood can be used as a steam generation device, although the steam production will only
be cyclic. Some new plants use suppressed combustion hoods which do not inspire air and burn off-
gases. New BOF capacity is expected to continue this trend, which may cause a decrease in total
NO emissions.
During the combustion of the waste gas, the potential for NO production exists. One steel
manufacturer gives a range of values of from 30 to 80 ppm, or 180 to 500 ng NO per kg (0.36 to 1.0
Ib NOX per ton) of steel produced. There is no information available on NOX control techniques for
the basic oxygen furnace (Reference 5-28).
Soaking Pits and Reheat Furnaces
These are large furnaces with fuel inputs ranging from 1.17 to 4.12 MJ/kg (1.0 to 3.5 x 106
Btu/ton) heated. Fuel efficiency is affected by many factors such as furnace size, design, combus-
tion controls, combustion air temperature, furnace scheduling, and downtime. Improved efficiency
measures, which do not increase flame temperature, will, in general, reduce NO emissions in propor-
tion to the reduction in fuel usage.
5-30
-------
Existing fuel conservation measures in soaking-pit heating include improved scheduling so as
to charge at a higher ingot temperature, programmed input control, improved burner designs, air/
fuel ratio control responsive to stack-gas oxygen content, addition of recuperators to existing
cold combustion air installations, and use of recuperators designed to give higher preheat tempera-
ture. Of these, the use of high-mixing-rate burners and an increase in combustion air preheat are
likely to increase the N0x emission level. At the present time, only experimental information is
available concerning the effect of these parameters on NO levels.
A
Soaking-pit and reheat-furnace operating temperatures are such that the estimated NO levels
should fall in the 250 to 350-ppm range. However, the very large amounts of fuel used result in a
total NOX output estimated at 97 Gg (107,000 tons) in 1971.
A major factor that will reduce consumption of purchased and in-plant fuels and thereby de-
crease NOX output is the trend toward use of continuous casting to replace some ingot casting. In
this process, billets and slabs which are hot-rolled prior to cooling are produced from molten
steel, thus eliminating soaking-pits and most of the reheat requirement. About 20 percent of total
steel production, or 36 Tg (40 x 106 tons), is estimated to be produced by continuous casting in
1985. In spite of this, soaking-pit and reheating furnace steel capacity will have to be increased
during the 1975 to 1985 period to provide for the expected growth in steel production and for the
steel which for process reasons will have to be cast in ingots. According to the IGT projection,
conventional steel processing will account for 144 Tg (160 x 106 tons) in 1985. At present fuel
consumption of 5.4 MJ/kg (4.7 x 106 Btu/ton), the total fuel consumed for soaking-pits and reheat
furnaces in 1985 will be 795 PJ (750 x 1012 Btu). This fuel consumption will result in estimated
NOX emissions of 143 Gg (157,900 tons).
Heat Treating and Finishing Operation
This category includes annealing, hardening, carburizing and normalizing of some of the
steel industry cold-rolled products, as well as production of coated products. Fuel consumption
in 1971 was about 632 PJ (600 x 1012 Btu) for the production of cold-rolled products (about 25
percent of total steel production). NOX emission levels are assumed to be in the 150 to 250-ppm
range. On this basis, total NOX emission in 1971 for this category will be about 7.6 Gg (8,400
tons). Assuming that production of cold-rolled products remains at about 25 percent of total steel
production, the 1985 N0x emission will amount to 10 Gg (11,200 tons) per year. There is no informa-
tion available concerning NOX control techniques for these sources (Reference 5-28).
5-31
-------
Electric Furnaces
Production of steel in electric-arc furnaces has grown rapidly since World War II and is
currently estimated to be about 20 percent of total steel production. Because of the phase out of
open hearth steelmaking, the increase in BOF steel production, and the associated scrap-use limita-
tion, the amount of steel produced in electric-arc furnaces is expected to increase even more.
The combustion of fossil fuels currently plays a very small role in electric steelmaking.
This may change in the future as advances in technology permit the increased use of scrap preheating.
Most authorities agree that scrap preheating will be accomplished outside the electric-arc furnace
in a specially designed charging bucket, probably equipped for bottom discharge. Many of the designs
use excess air burners to limit flame temperature and minimize oxidation of the scrap. Associated
air-pollution problems include particulates from dirty scrap, iron oxide, and oil vapors. The
requirement for both incineration at or above 1,033K (1.400F) and particulate removal has caused
shutdown of several scrap preheating installations because of economic considerations.
The use of electricity for heat in steel productljn transfers the NO emissions to the
utility plant where the problem is easier to control. Electric furnaces are, in any case, a very
minor source of NOX from the steel industry (Reference 5-28).
5.3.2.2 Emissions
Emissions in the steel industry and its related processing have historically consisted of
fumes, smoke, and dust or particulates. The gases usually considered obnoxious have been SO , CO,
and odors. The presence of oxides of nitrogen has been obscured by the heavy emission of particu-
lates and a resulting lack of physical evidence. The NOX emissions observed can be traced largely
to the combustion of fuel oils and gas and, in part, to the burning of carbon monoxide, which is a
product of the processing operations.
The emission of nitrogen oxides from iron and steelmaking and processing equipment does not
appear to have been extensively investigated. However, reasonable estimates can be made by assuming
a relationship between known operating temperatures and N0x concentrations in stack gases (Reference
5-28). This relationship is affected by other variables, such as combustion air preheat temperature
and oxygen enrichment of combustion air.
Table 5-7 shows the estimated N0x concentrations for the major energy-intensive processes
and the resulting total annual combustion-related NOX production based on 1971 steel production
energy consumption data (Reference 5-28).
5-32
-------
to
UJ*-*
O TJ
O CU
en +>
o. o
a.
co
o cu
U_ 00
O
1-1 CO
oo o
xco
o a:
a H-
LU z
i LU
IO
VI
o
I/)
»/»
t
c
LU
X
O
5
3
C
C
IO
to
t.
o
u
Lu.
O
to
to
LU
X
o
c
o
4->
^^
1
c
0
CO
u.
IO
3
C
C
I/)
C
0
4->
Cfl
CD
+J
CO
n-
.O
^~
*>*
O>
C
1
e- to
5. *
^ O)
E
in
IO
3
CO
f4
0
o.
to c
to O)
01 I
0 t- 0.
t~ 3
O. CT
LU
£
to
in
^
in
CM
0
o
CO
^~
o
o
CO
r*.
CM
0>
CM
O>
C
*N
r-
4->
CU
^~
*oi
o.
O CJ>
in co
in c
CM
CO
* i?
cr> c
C\J
O CD
r-~. cu
C
0
i^* O)
O O)
CO C
CU
O O en
O CO O)
in CM c
**
CO TJ
o^ c
00 TJ
Cft C
CO
IO
O) fc.
T- U.
CU 4->
^j (X)
C 10
to co
I §
^3" ^"
CM i
c* «>
to CM
CM
00 ^~
CM t
o o
o o
CM tO
t
CO
in
§00 O
««* o
CM I
O
r"~
^**
CM O
O
CM CM
in CM
CM r
CM CM
CO
0
+J
to
CU
u
c s-
O) 3
> U.
0
CO tf>
^ IO
o
O CO
o
CO
CO
in
CO
CO
oS
0
o
IO
CO
o
o
to
r»
CM
^
in
CO
^«
O)
u
IO
c
3
U.
J=
fc.
IO
0)
c
&
o
T>
TJ
TJ
TJ
C
TJ
C
TJ
C
TJ
C
ftJ
U
IO
£»
3
U.
CO
0
u
10
CO
§
i^
o
r
«
vo
CM
O
o
CO
CO
8 CM
CJ1
CO ^<^^
o
in
r-
in
TJ 10
C CO
10 U
IO
4-> C
t- i.
CL. 3
u_
en
C -M
?- 10
AC CU
10 -C
O 0)
8
00
to
rv-'
CO
CM
o
o
CM
r
O
0
CM
0
to
s
c
f»
JC
in
r-
C
^
^^
TJ
C
ID
O)
C
IO
£
^ >
IO
CO
o
1
to
CM
§
CM
O
TJ
C
TJ
C
in
CM
1
to
TJ
C
TJ
C
TJ
CU
U
IO
c
3
U.
£
4»*
o
CU
LU
CM
O
z
to 10
IO 4->
IO
TJ TJ
CO
to o
in c
0)
I- ii
o.
X TJ
CU C
IO £
I/)
CU
O
to
CO
1
in
O)
u
1
Qi
«+-
CO
cc
^^
*l""l«
r
CU
CO
to
.
=3
C
O
to
o
IO
t_
CU
CU
3
£
4-*
in
in c
o
«n *cu
to CU
*£ +*
E to
CO
TJ
CU C
IO
t- co c
C7) 1 O
i- in j-
C7> CU
CU U C
C C 10
CO U
II fc. f-
o> s-
cn *- cu
co co E
c cc
-------
Other test results provided by the American Iron and Steel Institute (Reference 5-36) indi-
cate different emission factors as shown in parentheses in Table 5-7. The emission levels for the
coke ovens are the result of three separate tests (10, 186, and 485 ppm). There was some concern
about the experimental procedures, and new tests are planned for 1978.
Results of recent tests reported by Hunter, et al_. (Reference 5-26) are summarized in Table
5-8. The open hearth furnace was tested while operating on natural gas and Number 6 fuel oil
(60/40). The wide fluctuations in N0x and CO observed as various operations were performed are
shown in Figure 5-5. Large changes in excess air occurred as the operators opened doors to look at
the steel and to add material or adjust fuel flow to change heating rate. NO emissions varied
from 100 to 3500 ppm and averaged about 1800 ppm or about 950 ng/J (2.2 Ib/MMBtu). NO increased
somewhat linearly with excess 02- Particulate emissions were 2200 ng/J (5.02 IbMMBtu), measured
upstream of the precipitator. Following baseline tests the furnace was overhauled to repair refrac-
tory and fix leaks. A second test cycle was observed on the repaired furnace and the average NO
was 1094 ng/J (1250 ppm), a reduction of about 40 percent. During baseline tests, NO frequently
A
exceeded 2000 ppm but with the excess air controlled, excursions over 2000 ppm occurred only twice.
One steel billet reheat furnace was tested while firing natural gas at heat rates between 13
and 30 MW. Baseline NOX emissions at 24 MW (82 million Btu/hr) were 56 ng/J (110 ppm) and particu-
lates were 17 ng/J (0.04 Ib/MMBtu). This furnace had two heating zones with 13 and 14 burners,
respectively. The row with 13 burners released about 80 percent of the heat input. Combustion
modifications included reduced excess air, resulting in a 24 percent NOX reduction, and burners out
of service which produced a 43 percent NOX reduction with three burners out of service in the row
of 13 burners.
One steel ingot soaking pit was tested (site 16/2) while firing natural gas at about 2.9 MW
(10 MMBtu/hr) through a single burner. Baseline NOX emissions at 2 MW were 52 ng/J (101 ppm).
Reduction of excess air reduced NOX by 69 percent with no adverse effect on the steel.
5.3.3 Glass Manufacture
5.3.3.1 Process Description
The glass manufacturing industry is made up of several basically different types of opera-
tions. They are:
5-34
-------
oo
rD
Q
OO CNJ
o
CU
oo o
_J C
O CU
a: i.
i cu
2: M-
o cu
o a;
X
O OO
O OO
OO
OO i
K- ^
O LU
LU
Li_ X
LL. O
00
I
LO
CO
o
I
4-*
to
13
c:
0
-M
rO
o
inQ t|
E T-
0 T3
0 0
X
o
r
4->
O
s: -a
c
cu
r-
'cu
to
03
rn x
1 1 1 **N
O
cu z
CD
rC
CU
cu
n:
CM
O
E
Q.OO
CD
C
3
LL.
CU
r- Q.
cu t
Q
CM
O
O
O
^t'
O
o
CM
0
«i
CJ3
4_) *
rn ^
S-
cu
c cu
CU 0
Q. ro
0 C
t-
,^ -^
CU LL.
CU
4_>
OO
OO
O
O
CD
CM
CO
00
^
0
£1
^o
LO
rO
CJ3
.
-j_>
^
ro
-C o
CU ro
eel. c
S-
'cu u_
cu
M
OO
CM
0
o
_J
CT)
vo
o
CM
LO
to
ro
CTJ
4
M
r
Q_
CD
sx
ro
O
OO
r~~
CU
CU
K
5-35
-------
4000
3500
BASELINE TEST
NO. 6 OIL AND GAS FUEL
1200 1300 1400 1500 1600 1700
TIMEOFDAY.hr
1800
1900
2000
Figure 5-5. NO emissions as a function of time for an open hearth furnace (Reference 5-26).
5-36
-------
1. Glass container manufacture
2. Fiberglass manufacture
3. Flat glass manufacture
4. Specialty glass manufacture
The largest type is the glass container industry, which produces about 45 percent of the total
amount of glass (by weight) produced by the entire industry.
While the specific processes used within each segment of the industry vary according to the
product being manufactured, glass manufacturing involves three major energy-consuming processes:
melting the raw materials, refining the molten glass, and finishing the formed products. Typically,
about 80 percent of the energy consumed by the glass industry is for melting and refining, 15 per-
cent is for finishing, and 5 percent is for mechanical drives and conveyors. The primary differ-
ences in processes used among the various segments occur in the refining and finishing operations.
The raw materials used in glass manufacture consist primarily of silica sand, soda ash, lime-
stone, and cullet (crushed waste glass). In the production of window and plate glass, for example,
temperatures in the range of 1,783K to 1,838K (2.750F to 2.850F) may be required to melt these raw
materials into a viscous liquid.
The furnaces used are of the pot type if only a few tons of a specialty glass are to be pro-
duced, or of the continuous tank type for larger quantities. By far the larger amount of glass is
melted in furnaces, and only these will be considered in connection with NO control.
Continuous reverberatory furnaces have a holding capacity of up to 1.27 Gg (1,400 tons) and a
daily output of as much as 270 Mg (300 tons). Reverberatory furnaces in this industry are broken
into two classifications according to the firing arrangement used: end-port and side-port melters.
In the operation of a side-port-fired furnace, the preheated combustion air mixes with the fuel in
the port, resulting in a flame that burns over the glass surface. The products of combustion exit
via the opposite port, down through the checkerbricks, and out through the reversing valve to the
exhaust stack. Typically, there are several ports situated along each side of the furnace. In
contrast, there are only two ports in an end-port-fired furnace, located on the rear wall of the
furnace. The flame is ignited in one port, travels out over the glass toward the bridgewall, and
"horseshoes" back to the exit port - the other port in the rear of the furnace. In both types of
furnaces, the firing pattern is reversed every 20 to 30 minutes, depending upon the specific furnace.
During this reversal period, the flame is extinguished, the furnace is purged of combustion gases by
reversing the flow of combustion air and exhaust gases passing through the reversal valve, and
5-37
-------
combustion is then reestablished in what was previously the exhaust port. Both types of melters are
operated continuously throughout a campaign that normally lasts 4 to 5 years, at sustained tempera-
tures up to 1.867K (2,900F).
In addition to the reverberatory-t.ype melters, day tanks, unit melters, and pot melters are
used, mostly in the pressed and blown glass industry. Many of these melters are batch-type, as
opposed to continuous, resulting in a substantial reduction in fuel-utilization efficiency. Much
of the fuel that is wasted is due to the antiquated methods of operation and associated equipment
used with these melters (Reference 5-28).
The combustion gases, on leaving the melting zone, retain a considerable amount of heat. This
is reclaimed in a regenerator or brickchecker chamber. When the firing cycle is reversed, combus-
tion air is preheated by being passed through the brick work. Preheating saves fuel but increases
the flame temperature which promotes NO formation.
Coal is not used in glass melting. Since molten glass is conductive, electrical heating is
used as a booster to supplement fuel firing whenever technically and economically practical. Gas
and, to a lesser extent, fuel oil are the preferred fuels.
5.3.3.2 Emissions
The flue gas from glass-melting furnaces is the major source of NO emission in the glass
industry. The operation of these furnaces is similar to that of open hearth furnaces used in steel -
making; regenerative checkerwork sets absorb heat from the combustion gases for subsequent release
to the incoming combustion air. This is accomplished by a reversing valve which puts each checker-
work set through its heating and cooling cycle in turn. The sequence of intense high-temperature
combustion and quenching in the checkerwork sometimes raises NO emissions to levels higher than
those experienced in a steam boiler of equivalent heat release. For example, during a recently
completed experimental program, NOX emissions were measured during a complete firing cycle of a
glass melter. NOX emissions were highest at the beginning of the firing cycle and then, as the
cycle continued, decreased by about 30 percent. At the beginning of the firing cycle, the combus-
tion air is preheated to a higher temperature, which results in a hotter flame than at the end of
the cycle when the checkerbrick and hence the air have cooled considerably. Other major factors in
NOX formation in a glass melter, such as flame velocity and recirculation patterns of flue gases,
are being studied.
Table 5-9 summarizes the emissions from several glass melters as measured by a number of
investigators.
5-38
-------
to
4->
O
o
to
to
cu
O
X
CU
CO
O)
cu
(J
X
O)
LO
C\J
E
O
E
ID 'r-
to
to
CD
CJ
X
o
o
o
vo
o
LO
o
^d-
co
o
CsJ
LO
£q-
O
2 &
ID JO
00 ~[
(O
oo
tn
O LU
CsJ
QQ
a
4->
o>
to
(U
I
in
-------
5.3.3.3 Control Techniques
According to representatives of the glass industry, the efforts of the industry to reduce
air pollutant emissions are severely hampered by the variations in regulations that exist from
state to state. This lack of uniformity requires that different solutions to the problem be sought,
depending on the location of the specific plant. This, in turn, adds substantially^ the cost of
pollution control. In addition, not only are the regulations variable from one location to another,
but these regulations are constantly changing. As a result, very few air pollution control equip-
ment installations have been made on glass furnaces, and there is very little data available on
the effectiveness and cost of these devices.
In general, SO , NO , and particulates are the primary air pollutants from the glass manu-
A A
facturing processes. The concern is primarily with the melting process because this is the largest
energy consumer and the major contributor to air pollutant emissions. The major pollution problem in
the combustion process is NO emissions.
While the formation of NO in the combustion process is not entirely understood, it is clear
that the goals of reducing NO emissions and reducing energy consumption are seemingly at odds. N0x
A
formation is a temperature-related phenomenon; as temperature increases, NOX emissions increase.
On the other hand, increasing available heat to a process may result in increases in efficiency and
in temperature, which in turn increase N0x emissions. Analysis of the process modifications under
consideration in the glass industry shows that there is a possibility of increasing N0x emissions.
If the implementation is carried out properly, however, this need not occur.
Six recommended modification programs are listed in Table 5-10. The order of listing is
according to programs that afford the greatest potential for solving the problems in the shortest
period of time. The table also presents estimates of improvements that may be obtained, where such
estimates can reasonably be made (Reference 5-28). Cost data for these programs are not available
at this time. Two of the six recommendations are currently being pursued by EPA/IERL-Cincinnati.
5.3.4 Cement Manufacture
5.3.4.1 Process Description
The cement industry includes all establishments engaged in the manufacture of hydraulic cement
(generic name: portland cement), masonry, natural, and pozzuolana cements. This discussion is
limited to the production of portland cement because it accounts for 95 percent of the total
5-40
-------
CD
CXL
LU
UJ CO
CM
O I
Z LO
OO O
O LU
00 LU
OO Ll_
ii LU
s: or
I I I v_^
CD >-
»-" I
O 00
Q Q
C£ CO
O OO
00 O
s:
<: LU
C£ 2C
0 f-
O
Cti Z
a. I-H
o 2:
LU O
Q i i
z. \
LU Q_
O OO
O Z
LU O
I
LO
CO
5-41
-------
cement manufactured in the United States, with the remaining 5 percent split among the other
types.
Raw materials used in the manufacture of port!and cement consist of limestone, chalk or marl,
and seashells. These are combined with either clay, shale, slate, blast furnace slag, iron ore,
or silica sand. The end product is a chemical combination of calcium, silicon, aluminum, iron,
and other trace materials. The raw materials are first ground and blended together. Depending
upon which of the two processes is used, water may be added during blending (the wet process) or
the ingredients can be mixed on a dry basis (the dry process). In general, the moisture content
of the raw materials determines the process used. If the moisture content is greater than 18 per-
cent, by weight, the wet process will be used. If the moisture content is less than 18 percent,
the dry process will be used. The next step is the calcining or burning of the mixed raw material
in a rotary kiln. During this step, the material is heated to approximately 1,755K (2.700F) and
transformed into clinker, which has different chemical and physical properties than the raw
materials had initially. The clinker is discharged from the kiln and cooled. The last step is to
grind the clinker to the desired fineness and add gypsum to control the setting time of the concrete
(Reference 5-28).
5.3.4.2 Emissions
The major air pollutant emission problem in the manufacture of portland cement is particu-
lates, which occur in all phases of cement manufacturing from crushing and raw material storage to
clinker production, clinker grinding, storage, and packaging. However, emissions also include the
products of combustion of the fuel used in the rotary kilns and drying operations; these emissions
are typically NO and small amounts of SO . For both the wet and dry kiln processes, the limited
A A
data shows that nitrogen oxides are emitted at a rate of about 1.3 g per kg (2.6 Ib per ton) of
cement produced.
The largest source of emissions in cement plants is the kiln operation. At present, about
56 percent of the cement kilns in operation use the wet process, and 44 percent use the dry process.
Based on this information, estimates of total NOX emissions from cement plants in 1972 are 42.7 Gg
(4.7 x 104 tons) for the dry process and 54.5 Gg (6 x 10* tons) for the wet process. These estimates,
because of a lack of data, assume the use of no controls by the industry. Without an inventory of
control equipment in use, they cannot be refined.
Future efficiency-improving process modifications that increase flame temperature without
improving heat transfer to the process load will almost certainly result in increased NO emissions.
5-42
-------
Converse^, adequate removal of the additional heat resulting from the applicable process modifica-
tions should maintain NOX emissions at their current Level.
Of the process modifications deemed to be near-term, only the use of oxygen,nHchment has
any great potential of Increasing air pollutant emissions, pr^rlly NO,. , SOTO applicat10ns ,
other industries, for example, glass melting, oxygen enrichment can be used without Increasing NO
-i»1«,. However, due to the different type of load , the cement Industry and the different *
patterns of heat transfer, 1t 1s suspected that NO, would Increase with the Imputation of oxygen
enrichment (Reference 5-28).
5.3.4.3 Control Techniques
There 1s very little Information In the literature regarding co^erclal Installation of equip-
ment for removing NOX from kiln waste gas or of edifications to Mln operations to reduce NO
production. Water scrubbing 1s sometimes used for partlculate rental from waste gas fro. Ill
kiln,. In this operation, the gas contacts a slurry of calcium hydroxide, which should remove a
50/50 mixture of NO and NO, and reduce NO, up to 20 percent. Flue gas regulation, which Is used
to contro! temperature 1n some 1,. kilns, should reduce NO, emissions by lowering « temperature.
Reference 5-26 reports NO, emission test results for both a dry process Kiln and a wet pro-
cess Kiln. The dry process kiln was tested at full capacity while firing a 68/32 mixture of coke
and natural gas. Data for the same Mln firing natural gas and oil separately were also available
for comparison. Emissions of NO, while firing natural gas were 1,050 to 1,800 ng/J (1680 to 2900 ppm)
Operation on oil resulted In a 60 percent reduction (400-710 ng/J). Operation on combined coke
and natural gas produced emissions of 655 to 710 ng/J, a 50 percent reduction.
Lower NO, emissions on solid and liquid fuels compared to gas are attributed to the highly
ad,abat1c nature of the process. Many cement Mln, are currently being converted from gas to solid
fuels. This conversion W1,l be beneficial In reducing NO, and could be pursued as an NO control
method that Is consistent with the reduction of Industrial gas consumption.
The wet process cement kiln was tested only while firing natural gas and had baseline
emissions of 1319 ng/J (2250 pp.,. c^bustlon modifications Investigated Included variation of
combustion air Inlet temperature and excess oxygen, tncrease of combustion ,1r temperature from
644K (700F, to 767K (920F, Increased NO emissions to 15,8 ng/J, and ,5 percent Increase. Reduction
of excess oxygen at baseline a,r temperature reduced NO, to 846 ng/J, , 36 percent reduction The
independent reductions of either excess air or air temperature caused unacceptable reduction of
kiln temperature that can result In a process upset. The NO .missions were found to be a strong
5-43
-------
function of kiln temperature, as shown in Figure 5-6. It was found that simultaneous reduction of
excess air and increase in air temperature could produce a reduction in NO of about 14 percent while
maintaining kiln temperature.
Electric heating eliminates all the pollutants associated with combustion sources, but its use
in kiln operation is very limited. Another means of emission control in kiln operation is the choice
of kiln type. Some NO reduction in limestone calcining is obtained by usipg a vertical instead of
a rotary kiln. The mechanism of operation is such that heat transfer to the load is very high, and
peak temperatures are lower than required to obtain the formation of N0x in large amounts.
5.3.5 Coal Preparation Plants
Coal in its natural state contains impurities such as sulfur, clay, rock, shale, and other
inorganic materials, generally called ash. Coal mining adds more impurities. Coal preparation plants
serve to remove these impurities. Coal cleaning processes utilized by coal preparation plants may be
wet, dry, or a combination of both. Wet processes are a minor source of oxides of nitrogen.
After the coal is wetted by the cleaning process, primary drying is done mechanically by
dewatering screens followed by centrifugal driers. When lower surface moisture is desired (3 to 6
percent) with finer coal sizes, secondary drying is required. Such low moisture levels can best be
accomplished by thermal drying. It appears that new coal preparation plants that install thermal
dryers will use a fluidized-bed type.
In the fluidized bed drier, hot combustion gases from a coal-fired furnace are passed upward
through a moving bed of finely-divided wet coal. As the bed fluidizes, the coal is dried as the
fine particles come into intimate contact with the hot gases.
The major pollutant evolved from the thermal dryer is particulate. Well-controlled thermal
driers emit only minor quantities of NOX- Concentrations of 40 to 70 ppm (0.16 to 0.28 kg/MJ, or
0.39 to 0.68 lb/106 Btu) have been measured (Reference 5-43). These emission rates are below the
NSPS of 300 ng/J (0.7 lb/106 Btu) for large steam generators. In any case, no N0x standards have yet
been proposed since the thermal dryer capacities are generally less than the smallest power plants
required to control NO emissions: 73.2 MW, or 250 x 106 Btu/hr (Reference 5-43).
5-44
-------
1422
3000
1500
KILN TEMPERATURE, °K
1600 170Q
1800
1867
SHADED AREA SHOWS EFFECT OF INLET AIR TEMPERATURE VARIATIONS
NUMBERS REPRESENT EXIT 02 CONCENTRATIONS IN PERCENT
ROTARY CEMENT KILN, WET PROCESS
2300
2500
KILN TEMPERATURE °F
2700
2900
Figure 5-6. The effect of cement kiln temperature on NO emissions (Reference 5-26).
5-45
-------
REFERENCES FOR SECTION 5
5-1 National Academy of Sciences, "Air Quality and Stationary Source Emission Control," Prepared
for the Committee on Public Works, United States Senate, Serial No. 94-4, March 1975.
5-2 Hall, R. E., J. H. Wasser, and E. E. Berkau, "A Study of Air Pollutant Emissions from Resi-
dential Heating Systems," EPA 650/2-74-003, January 1974.
5-3 Barrett, E. R., S. E. Miller, and D. W. Locklin, "Field Investigation of Emissions from
Combustion Equipment for Space Heating," Battelle-Columbus Laboratories, EPA R2-73-084a,
June 1973.
5-4 Hall, R. E., et_ al_., "Status of EPA's Combustion Research Program for Residential Heating
Equipment," presented at the 67th APCA Annual Meeting, June 1974.
5-5 Dickerson, R. A., and A. S. Okuda, "Design of an Optimum Distillate Oil Burner for Control
of Pollutant Emissions," EPA-640/2-74-047, June 1974.
5-6 Combs, L. P., and A. S. Okuda, "Residential Oil Furnace System Optimization -Phase I,"
Rocketdyne Division, Rockwell International, EPA-600/2-76-038, February 1976.
5-7 Combs, L. P., and A. S.^Okuda, "Commercial Feasibility of an Optimum Distillate Oil Burner
Head," Final Draft Report, Rocketdyne Division, Rockwell International, Prepared for U.S.
Environmental Protection Agency, 1975.
5-8 Belles, F. E., R. L. Himmel and D. W. DeWerth, "Measurement and Reduction of NO Emissions from
Natural Gas Fired Appliances," APCA Paper No. 75-09.1. Presented at the 68th Annual Meeting of
the Air Pollution Control Association, June 15-20, 1975.
5-9 National Petroleum News, January 1975, pp. 34-35.
5_10 Lenney, R. J., Blueray Systems Inc., Weston, Massachusetts, Personal Communication, September
1975.
5_H Locklin, D. W., and R. E. Barrett, "Guidelines for Residential Oil Burner Adjustments,"
EPA-600/2-75-069-a, October 1975.
5_12 Locklin, D. W., and R. E. Barrett, "Guidelines for Burner Adjustments of Commercial Oil-
Fired Boilers," EPA-600/2-76-088, March 1976.
5-13 Hall, R. E., "The Effect of Water/Distillate Oil Emulsions on Pollutants and Efficiency of
Residential and Commercial Heating Systems," Air Pollution Control Association Paper 75-09.4,
June 1975
5-14 Black, R. J., H. L. Hickman, Jr., A. J. Muchick, and R. D. Vaughan, "The National Solid
Wastes Survey: An Interim Report," Public Health Service, Environmental Control Administra-
tion, Rockville, Maryland, 1968.
5-15 Niessen, W. R., et_ al_., Systems Study of Air Pollution from Municipal Incineration, Report
to NAPCA under contract CPA 22-69-23, Arthur D. Little, Inc., Cambridge, Mass., 1970.
5-16 McGraw, J. J. and R. L. Duprey, Compliation of Air Pollutant Emission Factors (Revised),
AP-42, EPA, February 1972.
5-17 Stenburg, R. L., et^ al_., "Field Evaluation of Combustion Air Effects on Atmospheric Emissions
from Municipal Incinerators," J. Air Pollution Control Assoc., Vol. 12, pp. 83-89, February
1962.
5-18 Kirsh, J. B., "Sanitary Landfill," In: Elements of Solid Waste Management Training Course Manual
Public Health Service, Cincinnati, Ohio, 1968, p. 1-4.
5-19 Fife, J. A., and R. H. Boyer, Jr., "What Price Incineration Air Pollution Control?,"
Proceedings of 1966 National Incinerator Conference, American Society of Mechanical Engi-
neers, New York. 1966.
5-46
-------
5~21 Chi, C. T. , and D. L. Zanders, "Source Assessment: Agricultural Open Burning, State-of-the-Art,"
5-23 Wiley, J. S. et al_. , "Composting Developments in the U.S.," Combust. Sci. 6(2):5-9, 1965.
5"24 72U"keFebrSary"fe96d9U?in9 Emi'SSi°nS ^ RefUS6 DisP°sal>" J' Al> Pollution Control Assoc., 19: 69-
%% L C°°k' °ffiCe °f F6deral Activities» U' S. Environmental
5-26 Hunter, S. C. , et al_. , "Application of Combustion Modifications to Industrial Combustion
Equipment, "Proceedings of the Second Stationary Source Combustion Symposium Vol III
EPA-600/7-77-0736, July 1977. - ~ - ~ - ! - : - '
5-27 Bartz, D. R. , et al_. , "Control of Oxides of Nitrogen from Stationary Sources in the South
Coast Air Basin of California," PB 237 688/7WP, September, 1974.
5"28 Fnn?IS: E'nV ' ? Ne?b1«. «nd R- D- Oberle, "A Survey of Emissions Control and Combustion
1
5-29 Schultz, E J., L. J. Hellenbrand, and R. B. Engdahl , "Source Sampling of Fluid Catalytic
Jul 1972 standard Oil of California, Richmond, California," Battelle-Columbus Labs,
5"3° SlrMn1 ErnJn'-^ J' "ei!enbrand' and R- B- Engdahl, "Source Sampling of Fluid Catalytic
Cracking CO Boiler and Electrostatic Precipitators at the Atlantic Richfield Company,
Houston, Texas," Battelle-Columbus Labs, July 1972.
5-31 Shea, E P "Source Testing, Standard Oil Company, Richmond, California," Midwest Research
Institute, Kansas City, Missouri, 1972.
5-32 Cowherd C., "Source Testing, Standard Oil of California Company, El Segundo, California,"
Midwest Research Institute, Kansas City, Missouri, 1972.
5-33 Shea, E P. , "Source Testing, Atlantic Richfield Company, Wilmington, California," Midwest
Research Institute, Kansas City, Missouri, January 1972.
5-34 McGannon, H E., The Making. Shaping and Treating of stPPl . 8th ed., Pittsburgh, United
States Steel Co., 1964.
5"35 SH^VI V ^ar5°nifIativn>," In: Kl>k-0thmer Encyclopedia of Chemical Technology. Standen. A.
(ed.), Vol. 4, 2d ed. , New York, Interscience Publishers, Inc.,, 1964, p. 400-423.
5-36 Personal communication, Dr. Walter Jackson, u. S. Steel, November, 1977.
5-37 Personal Communication, Mr. Andrew Trenholm of the Office of Air Quality Planning and Standards
U. S. Environmental Protection Agency, Durham, North Carolina, May 1976; anmng ana btandards>
5"38 mPPHnrAf^ho'l-"^6!?0?051^'0" °f 51ass Furnace Emission." presented at the 63rd Annual
meeting of the Air Pollution Control Association, St. Louis, June 1970.
5"39 JA:c~*St-; HEm1«1?n? ;f°x1des of Nit^gen from Stationary Sources in Los Angeles
N^rpgen Emitted by Medium and Large Sources," Joint District, Federal, State,
Repor ' L°S
5'40 Air Pollution Engineering Mflmml. Danielson, J. A. (ed.). National Center for Air Pollution
Control, Cincinnati, Ohio, PHS Publication Number 999-AP-40.
5-47
-------
5-41 Nesbitt, J. D., D. H. Larson, and M. Fejer, "Improving Natural Gas Utilization in a Continu-
ous End Port Glass-Melting Furnace," In: Proceedings of the Second Conference on Natural
Gas Research and Technology. Session IV. Paper 9, Chicago, 1972.
5-42 Ryder, R. J., and J. J. McMackin, "Some Factors Affecting Stack Emissions from a Glass Con-
tainer Furnace," Glass Ind., 50, June-July, 1969.
5-43 "Background Information for Standards of Performance: Coal Preparation Plants; Volume 1:
Proposed Standards," EPA 450/2-74-021a, October 1974.
5-48
-------
SECTION 6
NONCOMBUSTION PROCESSES
The problem of NOX emissions has been researched in the chemical industry more intensively
than anywhere else because it may represent the loss of a valuable raw material. The following sec-
tions of this report discuss commercial processes developed for NO control in the manufacture and
uses of nitric acid.
The NOX released in vent gases from the manufacture and industrial uses of nitric acid, dif-
fers markedly from that emitted from a combustion flue gas in concentration, total amount, and the
ratio of N09 and NO present. The NO -containing chemical gas is commonly a process stream which
t A '
must be recycled with maximum N0x recovery in order to have an economical process. Vent gas is re-
leased only because it is too impure to recycle or too low in concentration for economic recovery.
The economic limit with a pure gas, as in nitric acid manufacture, is about 0.1 to 0.3 percent NO ,
X
or 1,000 to 3,000 ppm. The limit is higher in organic nitrations, such as the manufacture of nitro-
glycerine, where NOX content of the vent gas may approach 1 percent NO , or 10,000 ppm.
The total amount of NOX emitted from all chemical manufacturing is about 1.7 percent
(203 Gg or 2.2 x 105 tons/yr) of all NOX from manmade sources in the United States. These pro
cesses present problems only in special local areas. The problems have been most serious in military
ordance works, which manufacture large volumes of nitric acid and use it in organic nitrations. A
single plant like the Volunteer Ordance Works has produced, for example, emissions of NO equal to
all nonmilitary uses of nitric acid in the United States.
A high ratio of N02/N0 at high concentrations causes the gases to be visible as a brownish
plume. The visibility limit depends on the total amount of N02 present in the gas volume or layer
observed. A convenient rule of thumb is that a stack plume or air layer will have a visible brown
color when the N0« concentration exceeds 6,100 ppm divided by the stack diameter in centimeters (Ref-
erence 6-1). This means that the threshold of visibility for a 5 cm-diameter stack is about 1,200 ppm
of N02 and for a 30 cm-diameter stack, 200 ppm of N02 (or 2,000 ppm of NO at a 1:10 ratio of NO?:NO).
6-1
-------
The distinction between N0? concentrations and total amount can be quite important in chemi-
cal vent gases, since a short burst of NOp at 10,000 ppm may be visible but less hazardous than
many times as much NO emitted from a large stack at a lower concentration. The total amount in a
short, concentrated emission may be too small to have a detectable effect on NO levels in ambient
air.
A large amount of research with varying degrees of success has been carried out on the devel-
opment of processes for the removal of NO from the off-gas resulting from the manufacture and uses
of nitric acid. The abatement processes are discussed in detail in Section 6.1.3.
6.1 NITRIC ACID MANUFACTURE
Nitric acid plants are divided into two types: those that make dilute nitric acid (50-68 per-
cent nitric acid) and those that make strong nitric acid (over 95 percent nitric). Nitric acid and water
form an azeotropic (constant-boiling) mixture at about 58 percent nitric acid content; this is the limiting
factor in the nitric acid concentration available through distillation and absorption methods.
The acid is concentrated to 98 percent in an acid concentration unit using extractive distillation.
The direct process for making strong nitric acid usually depends on direct formation of nitric acid
in an autoclave where nitrogen oxides react with oxygen and water to form nitric acid. Most (> 95
percent) nitric acid plants presently in operation are of the first kind.
6.1.1 Dilute Nitric Acid Manufacturing Processes
Nitric acid in the United States is made by the catalytic oxidation of ammonia. Air and
ammonia are preheated, mixed, and passed over a catalyst, usually a platinum-rhodium complex. The
following exothermic reaction occurs:
4NH3 + 502 -» 4NO + 6H20
(6-1)
(AH = -906 J/mole)
The stream is cooled to 31 IK (100F) or less, and the NO then reacts with oxygen to form nitrogen
dioxide and its liquid dimer, nitrogen tetroxide.
2NO + 02 + 2 N02 * N204
(6-2)
(AH = -113 J/mole)
The liquid and gas then enter an absorption tower. Air is directed to the bottom of the
tower and water to the top. The NOo (or NpO.) reacts with water to form nitric acid and NO, as
follows:
6-2
-------
3N02 + H20 + 2HN03 + NO
(6-3)
(AH = -135 J/mole)
The formation of 1 mole of NO for each 2 moles of HN03 makes it necessary to reoxidize NO after each
absorption stage since the gas rises up the absorber and limits the level of recovery that can be
economically achieved.
Acid product is withdrawn from the bottom of the tower in concentrations of 55 to 65 percent.
The air entering the bottom of the tower serves to strip N02 from the product and to supply oxygen
for reoxidizing the NO formed in making nitric acid (Equation 6-3).
The oxidation and absorption operations can be carried out at low pressure (~ 100 kPa, 1 atm),
at medium pressure (400 to 800 kPa or 58 to 116 psia) or at high pressure (1000 to 1200 kPa or 145 to
174 psia). Both operations ma^ be at the same pressure or at different pressures.
Before corrosion-resistant materials were developed, the armonia oxidation and absorption
operations were carried out at essentially atmospheric pressure. This also had advantages compared
to the higher pressure processes of longer catalyst life (about 6 months), and increased efficiency
of ammonia combustion. However, because of the low absorption and NO oxidation rates, much more
absorption volume is required, and several large towers are placed in series. Some of these low
pressure units are still in operation, but they represent less than 5 percent of the current U.S.
nitric acid capacity.
Combination pressure plants carry out the ammonia oxidation process at low or medium pres-
sure and the absorption step at medium or high pressure. The higher combustion temperature and gas
velocity at an increased pressure for the oxidation reaction shortens catalyst's lifetime (1 to 2
months) through increased erosion and lowers the anmonia oxidation conversion efficiency (Reference
6-2). Thus lower pressures in the oxidation process are preferred. On the other hand, higher pres-
sures in the absorption tower increase the absorption efficiency and reduce N0x levels in the tail
gas. Of course these advantages must be weighed against the cost of pressure vessels and compressors.
The choice of which combination of pressures to use is very site specific and is governed by
the economic trade-offs between the costs of raw materials, energy and equipment, and process effi-
ciency; and local emissions limits. In the 1960's. combination low pressure oxidation/medium pressure
absorption and single pressure (400 to 800 kPa) plants were preferred. Since the early 1970's, the
trend has been toward medium pressure oxidation/high pressure absorption plants in Europe and single
pressure plants (400 to 800 kPa) in the United States.
6-3
-------
6.1.1.1 Single Pressure Processes
In the single pressure process, both the oxidation and absorption processes are carried out
at the same pressure - either low pressure (100 kPa or ~ 1 atm.) or medium pressure (400 to 800 kPa).
Single pressure plants are the most common type. Figure 6-1 is a simplified flow diagram of a
single pressure process (Reference 6-3). A medium pressure process will be described below.
Air is compressed, filtered, and preheated to about 573K by passing through a heat exchanger.
The air is then mixed with anhydrous ammonia, previously vaporized in a continuous-stream evaporator.
The resulting mixture, containing about 10 percent ammonia by volume, is passed through the reactor,
which contains a platinum-rhodium (2 to 10 percent rhodium) wire-gauze catalyst (for example, 80-mesh
and 75-ym diameter wire, packed in layers of 10 to 30 sheets so that the gas travels downward through
the gauze sheets). Catalyst operating temperature is about 1,023K. Contact time with the catalyst
is about 3 x 10" sec.
The hot nitrogen oxides and excess air mixture (about 10 percent nitrogen oxides) from the reac-
tor are partially cooled in a heat exchanger and further cooled in a water cooler. The cooled gas is
introduced into a stainless-steel absorption tower with additional air for the further oxidation of
nitrous oxide to nitrogen dioxide. Small quantitites of water are added to hydrate the nitrogen
dioxide and also to scrub the gases. The overhead gas from the tower is reheated by feed/effluent
heat exchangers and then expanded through a power recovery turbine/compressor used to supply the
reaction air. The tail gas is then treated by the tail-gas treater for NOX abatement. The bottom
of the tower yields nitric acid of 55 to 65 percent strength.
6.1.1.2 Dual Pressure Processes
In order to obtain the benefits of increased absorption and reduced NOX emissions from high-
pressure absorption, dual-pressure plants are installed. Recent trends favor moderate-pressure
oxidation and high-pressure absorption.
A process flow diagram for a dual-pressure plant by Uhde is shown in Figure 6-2. Liquid
ammonia is vaporized by steam, heated and filtered before being mixed with air from the air/nitrous
oxide compressor at from 300-500 kPa (44 to 72 psia). The ammonia/air mixture is catalytically
burned in the reactor with heat recovery by an Integral waste heat boiler to generate steam for use
in the turbine driven compressor. The combustion gases are further cooled by tail gas heat exchange
and water cooling before compression to the absorber pressure of 800-1400 kPa (116 to 203 psia). The
6-4
-------
WASTE GASES TO
POWER RECOVERY
AND TAIL GAS
TREATMENT
NITRIC ACID
Figure 6-1. Single pressure nitric acid manufacturing process (Reference 6-3).
6-5
-------
CM 3V
CC Z
Z"? |£
< < OQ UJ O <
CC
OQ=
< H
o
OQ
QC 0 m
LU S l_
><<
CC ±
CD
03
O
c
CD
c
_CD
Q.
;g
'o
03
OJ
D.
u_
amon
6-6
-------
absorption tower is internally water cooled to increase absorption by water. Nitric acid up to 70
percent concentration is withdrawn from the bottom of the column and degassed with the air feed to
remove unconverted NO before being sent to storage. The air/NO mixture is combined with reactor
effluent to form the absorber feed. High yields of up to 96 percent conversion and tail-gas emissions
as low as 200 ppm N02 can be obtained by this process.
6.1.1.3 Nitric Acid Concentration
Figure 6-3 illustrates a nitric acid concentration unit using extractive distillation with
sulfuric acid. A mixture of strong sulfuric acid and 55 to 65 percent nitric acid is introduced at
the top of a packed column, and flows down the column counter-current to the ascending vapors.
Nitric acid leaves the top as a 98 percent nitric acid vapor containing small amounts of NO and
oxygen, which result from the dissociation of nitric acid. The vapors pass to a bleacher and a
condenser to condense nitric acid and separate NO and oxygen, which pass to an absorber column for
conversion to, and recovery of, nitric acid. Air is admitted to the bottom of the absorber. Dilute
sulfuric acid is withdrawn from the bottom of the dehydrating tower and is sent to be concentrated
further to be used for other purposes. The system usually operates at essentially atmospheric
pressure.
6.1.1.4 Direct Strong Nitric Acid Processes
Nitric acid of high strength can be made directly from ammonia in direct strong nitric acid
processes. These processes depend upon the formation of nitric acid by reaction of N02 or N20^ with
oxygen and water forming 95 percent to 99 percent nitric acid. In this direct process, the composition
of the product nitric acid is not restricted by the azeotropic limit.
The principal licensors of these direct processes are Uhde and Davy Powergas. Uhde has built
two plants in this country using their direct strong nitric acid process. The Uhde process will be
described in detail below. Davy Powergas has two direct strong nitric acid processes; the CONIA
process and the SABAR process. Davy has not built any plants utilizing these processes in the
United States, but there is a CONIA plant recently constructed in Sweden and a SABAR plant recently
constructed in Spain. How these processes differ from the Uhde process will also be described below.
Figure 6-4 shows a process flow diagram for a direct strong nitric acid plant. Air and
gaseous ammonia are mixed and reacted where steam is generated in a combination burner/waste heat
boiler by the heat of reaction. The reaction products are cooled, and a weak nitric acid condensate
removed. The remaining gases are put through two oxidation columns where the NO is converted to N0.
6-7
-------
VAPOR
98%HN03
FEED
93%
H2S04
60%
HN03*
DEHYDRATING
COLUMN
TAIL GAS TO
ATMOSPHERE (VOLUME %)
COUNTERCURRENT
CONDENSER
VAPORi
i
*
CONDENSATE
BLEACHER
95-99% HN03
TO COOLER AND STORAGE
74.3 N2
20.402
VAPOR
LIQUID
STEAM
COIL
JN-
NSABLE
5ES
ABSORPTION
COLUMN
AIR
BOILER
TO COOLER
55% HNO3
Figure 6-3. Nitric acid concentrating unit.
6-8
-------
QIDVDiailN
u. z oc
CC _J 5 UJ
LU LLJ I 3 J
_l 00 UJ _J O
ejrr^3>C3OH-
O Q7 I O X < o
-------
The overhead vapors are compressed to a pressure that allows the equilibrium di-nitrogen tetroxide
(N204) to be liquefied with the use of cooling water alone. The liquid Np04 is converted to nitric
acid of about 95 to 99 percent by reacting the N204 with oxygen at a pressure of 5000 kPa (50 atm).
The conversion reaction is: 2N204 + 2H20 -» 4HN03> Tail gases from the absorption column are
scrubbed with water and condensed N204 in a tail-gas scrubber before being released. The liquid
from the tail-gas scrubber is mixed with the concentrated acid from the absorption column, which
has been bleached and liquefied. The combined product liquid (containing N204 as well as HNOJ is
reacted with oxygen in the reactor vessel, cooled, and bleached to produce the concentrated nitric
acid.
Both Uhde plants using this process were built in 1973 for the U.S. Government: one in
Joliet, Illinois makes 236 Gg/d (260,000 tons/day) of 98.5 percent nitric acid; the second in
Chattanooga, Tennessee, makes 313 Gg/d (345,000 tons/day) of 98 percent nitric acid. Neither are
in operation at present, although, both were designed to meet the New Source Performance Standards.
In the Davy Powergas SABAR process (Reference 6-4), like the Uhde process, ammonia and air
are reacted at atmospheric pressure, and a 2-3 percent nitric acid is condensed and removed as a
byproduct. Davy Powergas estimates 0.3 kg of this weak acid byproduct is produced per kg of con-
centrated acid. As in the Uhde process, N02 is then produced from the product gases and absorbed
in concentrated nitric acid. However, whereas Uhde forms N204 from this liquid, and reacts the N 0
with oxygen, the SABAR process takes the concentrated HN03 to a vacuum rectification column, where
concentrated HN03 comes off overhead and azeotropic nitric acid is collected at the bottom. Atmos-
pheric emissions are less than 500 ppm nitric oxides, which would not meet the new source standards
in the United States without further treatment.
Davy Powergas developed the CONIA process to meet the more stringent environmental regulations
for its site in Sweden. The CONIA process also depends on the ammonia-air reaction, followed by re-
moval of the water which is generated. The plant produces both 99.5 percent nitric acid and 54 per-
cent nitric acid, with less than 200 ppm NOX in the stack gases, and no other solid or liquid waste
streams. However, Davy Powergas considers this particular plant design to be over-designed and hence
too costly for most applications unless lower emissions limits must be met (Reference 6-5).
6.1.2 Emissions
Absorber tail gas is the principal source of NOX emissions from nitric acid manufacturing.
Minor sources include nitric acid concentrators and the filling of storage tanks and shipping con-
tainers. Nitrogen oxide emissions from nitric acid manufacturing are estimated at 127 Gg
6-10
-------
(140,000 tons) during 1974, which is about 1.0 percent of the NO emissions from stationary sources.
It is estimated that 7.4 Tg (8.2 million tons) of nitric acid (100 percent) were produced in 1974
(Reference 6-6). AP-42 (Reference 6-7) cites an average emission factor for uncontrolled plants of
25 to 27.5 kg/Mg of acid. Typical uncontrolled tail-gas concentrations are on the order of 3000 ppm
of NO with equal amounts of NO and N02 (Reference 6-8). Unde cites emission levels in excess of 800
ppm for low-pressure plants, 400 to 800 ppm for medium-pressure plants, and less than 200 ppm for high-
pressure plants (Reference 6-2). The extent of control for these plants' is not known, although, Uhde
'did state that all three processes could be designed in such a way as to meet State and Federal emis-
sion limits.
In any nitric acid plant, the NO content of tail gas is affected by several variables. Abnormally
high levels may be caused by insufficient air supply, high temperature in the absorber tower, low-
pressure, production of acid at strengths above design, and internal leaks, allowing gases with
high nitrogen oxide content to enter the tail-gas streams. Careful control and good maintenance are
required to hold tail-gas nitrogen oxide content to a minimum.
6.1.3 Control Techniques for NOX Emissions from Nitric Acid Plants
Nitric acid plants can be designed for low NO emission levels without any add-on processes.
Such plants are usually designed for high absorber efficiency; high inlet gas pressures and effec-
tive absorber cooling can lead to low NO emissions. However, many new plants, and all existing
plants, are not designed for NO emission levels low enough to meet present standards. For these
plants, add-on abatement methods are necessary.
The available abatement methods suitable for retrofit include chilled absorption, extended
absorption, wet chemical scrubbing, catalytic reduction, and molecular sieve adsorption. In this
section, these various control techniques for NOV are described. These techniques may also be
.A
appropriate for retrofit of explosive and adipic acid plants.
Many of the retrofit processes are offered by more than one licensor, and many licensors
(such as Uhde) offer more than one process. Table 6-1 lists the major processes, the types of
plants for which the processes are most suitable, and examples of nitric acid plants where the
processes have been applied. (The examples of nitric acid plants are not meant to be inclusive.)
The selection of a control method depends on such things as the degree of control required,
the operating pressure of the plant, and the cost and availability of fuel. For example catalytic
reduction was used to establish the NSPS originally. Since that time fuel costs have risen to the
point where catalytic abatement is not economically attractive for new nitric acid plants but can
be used as an effective secondary treatment to meet the NSPS.
6-11
-------
ft.
_i
Q-
Q
V)
CO
Q.
re
X
UJ
(/)
S-
o
to
c
CO
u
r~
1
Comments
o
O
JC
4**
CO
3E
CO
CO
CO
u
o
i-
Q-
re
r
1 Ll_
re
re re
u o.
E
«t- re
i t
0 £
jc re
i- 4->
^>-zz
>>
T3 C
"». O
CO 1
C "O
O T3
p re
o c
cn o
co i-
* * i %
0
O 3
CD CO
z i.
CO O
co -P
CM*
^^
O
1
1 1
^>
1
t
0
o
*
^~
^ .
J j
CO J- i-
CO CO
EJC Si
P C
P O
o :
C -P C
C 3 C
re o r-
0 JC C
p t
>>«- -t
S v
c
re co +
3 0.
i/> co c
=fz. t.
c
3
r X S
O O C
(/) Z +
»
^O if-
CO O
W> T
re >> <
CO -P i-
i. t- r-
O «
c i- j:
-i JD t
c
0
F
O.
a s-
co o
r tO
r JD
r" ^C
JC
O
CO
"u
4O
.3
?*
1C
CO
+^
c
re
"a.
re
a.
^^
CO
c
o
in CM
in i>^
- cn
-o-
^s^
CD
Z! co
r-~
o re
en o
1 JC
CM CO
^^
^>
I
E
^
ered acid product
concentration
0
J
D
C
3
;
3
o
JC
^^
co E
to t- o>
r; CO «t C o
Z. I-- C ~o
Cn C CO -r-
"r OH- t-
p -p JD
c to « c:
re co i -P JD .f-
r U) O i- CO > - >, CD (J c
re >> re c -i- 10
o 4-> -o - i-
to o to re 10
c c > co i
o o o -P CQ u.
-P O -PC
N re «
o re Or -P «
vo >- o D- c Q
ro co re E
co O CL. i
0 r -0 -r- ^
re -^.4.) o
CD O CD-.- (J t_
z: T- 2: c c co
E 3 -r- u>
r^ co CM E E T-
CM jc r» E o re
CO O CM «t O $£
*o
£ (/)
re -r-
JC 0
0 i.
p re
r- CU
L.
ri fly
T3
C
LJL. re
S-
f^
0
Q
vo
O "O O '"*
re to T3
CO to CO
i. V i.
3 T3 S- -r-
CO CO O. 3
co J_ E cr
>
c D- -i- re
i
ex
t- 1
O >> 1 CL
tO JD JD -i-
JD re 3
re x cr
Or CO
o z re
CO C C
CO >4- O O
re o i- i-
(0 -P -P
I. C - O. -P
o o -o t- c
C i- T3 O CO
i i -p re to E
0
r
o o.
CO t-
T3 O
C «/>
CO JD
P *f.
X
UJ
to
^1
o
^
i-
CO
«H^
+J
re to
i i/>
Q. CO
c o
re s-
O- O.
re
-3 C
o
"~ '-P
> Q.
to S-
P O
C CO
re JD
re
o.
T3
. CO
CO r-
r**
*^ *r-
.C
cn o
r^
£_
CO
4^
re
CO
2
s:
.
Q-
»
O
c
re
(/>
c
£
'-»
re"
o
i/i r*~
c cr>
o
o re
0 i
o u-
re
T3 r
»-^ o
CD O
y^ re
CO
CO C
O CO
cn o.
O
u
r~
E
CO
JC
o
^3
CO
"T~
r«.
5
»
4^
C
re
cxin
«**
re
T3
^x^ <^
to CO
c z
o
JC
in re
CM O
»
TJ r
-»» re
CD O
zr-fi
E
O CO
in jc
CM O
CO
T3
JC
Requires additional
compressor
r~
p
3
JD
CO
C
o
z.
i.
co
T3
I-
P"
O ivj
f.
t i Ll-
eS O
O O
*
I
HH -
fO
« r-
C/) "~
CO
^ CO
re o
«/> re
re
« E
. CO
4-> JC
r O
CL.
oO
C
CO
CD-P
0 0
i- 3
P "O
d.
.
f I-
SS
CO
(U *^^
t- O)
3 C
*-*(
JD
CC JD
< 3
CO J-
«t O
z: to
U)
$_
P CO
jc re >
CD O
r- CO O
JC i CO
Performs better at
pressure but operab
lower pressures. R
re
f CO
- s-
re 3
p i
JC C
CD4J
C t- C
^ 2 C
JD T
JD CO 4-
3 CO =
J- CO r-
o re c
CO CD C
re CD
u c
"E 15
CO JD
JC 3
O S-
o
PCO
CO
^k
re
0
z
c
c
3
CD
CD
e
O-
e
TJ
>^
DT
^^
CO
t-
0
CD
C
*T"
JD
0 JD
S- 3
T3 S-
>> 0
3: to
.* re
co co
i. U
O 3
re
CO
i. i
3 CO
o
ammonium nitrate an
solution. Requires
.friqeration.
X
5 O O
_ 4J Z
5 re s-
- T CO
J C >
3 Q O
-go
3 E CO
t re fc.
^
c
0
*» *
iD TJ
«-> r^ ^- c
p cn ^- re
Q. r in i
^B
CO CD CO
3 «i
4>^ *^ *
10 T r "O
re re o re -i
o. o . 4-> E
o 2: o re
O T3 « re
CJD C 4-> to >,
JC 4^ O
E re
re re u co r
o x T-- t. o^-<
-^ co a: u. >,
w> H- r* re
c -o
o « «/> i- >»
P 4J O CO O Ifl
P 0 T- «4- C
0-r- t- 0
0 E r- 4J W4J
r E f- W» i_
i- O 3 CO O
Q TJ JD O
"O C C JD V£>
^. O - 3
CD J- J-
z: ^- > . o TD
r~- co LI_ w) ^»
o cn jc CD
Cn r O <_> CM E
CO
3 CD
p re c
V) 'r- T-
re c JD
a. Q JD
o 1 s-
o re u
J-T3 1
O CO TO
P -P O
re re i- c
S- J- Q. O
O -P -r-
May require an evap
to produce a concen
ammonium nitrate by
uct. No refrigerat
required.
LO
O
o
-------
o
Oi
o
u
§
o
a.
o
I-H
O
CO
ct
) 4"^
2 (/)
*^O 3
O CT>
*""* O^ 3
10 , ^ .^~
o> c JD
T" » 3
i 4J i
tt) Q. O
k. k. 0
M O
to wi
JD
- re r-^
Q.-O Cf>
k. QJ r-
O T3
O C
CO *'>i
z: x re re
>-t Ol -O O
>}
r~"
L
QJ
JC
^^
re
Ol
^g
z:
o
Oi
JD
re
pv
JC O 4-> JC
C 4-> (J X
re M- oi -r- oi >> k.
k. f- 4-> X C k. O
3 C t- JC O)
*J O VI 4-> > »
re re o 4-> *o - o o
CO U QJ X C O r
f- CM re in o oi
to E O QJ 3 C i- k. QJ -M Ol
§o k. o 4-> re t-
C k. O) -i- O. >>JD 3
O O O JO 4-> k. CD-I- in in
mo in OOk.in.ttin
C Ol Xr- >, C VI Ol I/I k. 01
ocore re 3 jo co Ok.
O 3 Z > Z: -> re UJO.2O.
O CM
CM-»-> O
o o
CM
03:-
c o
re k. CM
o z
X
^^ *T* {>J
in
c jc E
k. 4-> t
3i- O
00 » «4-
O)
^
r
+>
u
T p***.
4J flj
>> v>
i C
re o
re
o
TJ
Ol
r*
f
re
+"*
i/>
C
^
*
«O i/>
rt QJ
i2£
o.
C ^O
r* ^^ *
o =>
4^ ^^ C
C > i-
re -^
'o.o -M
o c
E re
re k. r-
k. Ol O.
+* -4-*
f- <+- 0
z re r-
I^M
f
3
C3
c
re
o.
re
^
c
re
Q^
f%
o
3
UJ
C
tn
c
re
Q.
i
re
4^
re
o
U. i-
00 JC
CO i-
*~^*~*, f*i
tn 3
Ol 4-> I/I
T3 VI 4->
=>*- z:
O) O) JC Ol $
J3 4-> 1 CD i O "O
re x f- jo oi
C k. 01 C JC i- k-
re oi o to >,T-
o ^ ^ »^ k. ^^ v) k. vi
O4->4->O k. O QJ QJ
i- Q. Ol Q. > -O
re o * k. » >
!*-> OO O -t-5 QJ **-
C ^7 c/) ^^ O O ^ **
O QJ QJ JO A) C O)
E> to re -»J QJ k. f-c
E-i-3 TO k. >>JCO
c c oi tn to cDi re to
to QJ QJ ~o -^ (/> k. re i/i
QJ Q. 4-> C k. QJ QJ 3 C-r-
tnxH-oi O C cm re E
r> QJ O (-> 2D.UJ3 O QJ
E
JC C CM
MOO
f- «4- -0
X " QJ
0 O -M
X-M CMO
O Z Ol
x. re «4-
t- TJ «»-
to c c re
c o re
k. E 4->
3 E CMO
co re z c
01
O i-
fw ^>
4-> O
P1^ P""
re oi
4^ OO
re
o
^J
c
i.
o
0)
Ol
o
k.
O)
o
c
=>
^/^
o
i__i
^^
^^^
_j
o
in
Oi
0
o
3
to
o
O)
r-
E
<^-
~J
^^ re
O JQ 0
T5 to k.
«f- O) 3 O
O N O «*-
>> O> Ol C
§i C J3 0
re o> i-
T- 4J O5 «/> -M
M re Of- Q.
re u k. m k.
0 01 >> 0
t- CM4-> t in
x o oi re JQ
o z jc 4-> re
in
3
O l/>
Ol >r-
C VI
O">^~
e|M
^ *
O> IO
*^0
*
o
01
^.
QJ
1^
O
^^
f
^>
C
Oi
k.
3
0
+J
0
c
c
re
ex
re
-3
c
in
C OO
re
'o.
c
CM-i-
^^
re
o
"E
O)
tj
in
3
k.
re
o
i
VI
O>
Q.
4->
0
^
r O>
re k.
O Oi
§"^ M-
o
o >»
Oi-^
Ol ->
c c
13 O)
^^
r
JC i
M re
i- ^- O
O ^ i^^
^ *- o £
re <3-4-> re
0 O 0 r
t- C Ol O
x z: oi
oS2 a* t.
c
0
4->
f Cv
re k.
o o
^ (/)
E JO
JC
w
O T3
4^ QJ
in i
I 4->
0 O C
c z re
*- E
to
m E "O
O> k.
3 Ol
O r
k. oo jo
o> re
Z => k.
Ol
^^ ^"^ O.
>> >> o
re < re c
°o r-« -o -
m r 10
c c
o o c
4-> re 4J c
i Ol
in
CD QJ CD to
E tn E en
k. C
O O) O "i-
in co in ^
^.^
Oi
D
JO
z. re
>
T- C
(/) O
re i-
k. C
3 :D
ex. ^-^
i
f- i
o. CD re
re c i k.
O T- C QJ
O Q.
TJ 0 0 O
c >>
re o o 4->
-> c
>> in 4-> c re
WO i- -f- i
k. C «f- Q.
oi re oi
c E o > vi
QJ CU 4J O) 3
^ -r- O
JC T3 I/) 3 C
CDi k. CO
f- re to **- ~r- i-
Z 4-> Z 0 4-> 4->
* *
It
O) «4- CD
> 0 C
>> Ol i-
JD i- C >>i
to O JO O
C t- >,
O k. 4-> Ol 0
f- re re >
4->f k. O)
o. 3 O) »- re
k. o c to E
O O) O) k.
VI i CD QJ QJ
JD O QJ JC JC
«t E 1- *-> 4->
k.
re
r- 01
3 >
O Ol
OJ-i-
oo
p
6-13
-------
The inlet pressure at the absorber is an important factor in the selection of NO control
equipment. In general, extended absorption equipment cannot be economically installed where the
equipment will have inlet pressures of less than 758 kPa (110 psia). Consequently, extended adsorp-
tion is not usually chosen for older, low pressure nitric acid plants. Wet scrubbing and molecular
sieve absorption are also not as effecive at low pressures. Catalytic reduction, however, does not
require high pressures.
6.1.3.1 Chilled Absorption
This method is used primarily for retrofit of existing plants. Chilling the water used in
a nitric acid absorption tower leads to higher yields of nitric acid and lower NO concentrations
in the tail gas. Both water and brine solutions have been used in a closed loop system to provide
local cooling to the liquid on the trays of the absorption tower. Absorption may be further en-
hanced by heterogeneous catalytic oxidation of NO to N0« upstream of the absorption tower.
CDL/VITOK Process
Figure 6-5 shows a CDL/VITOK process flow diagram. Tail gas enters the absorber, where the
gases are contacted with a nitric acid solution to both chemically oxidize and physically absorb
nitrous oxides. The reaction of NO to N02 may be catalyzed in the main absorber. The upper portion
of the absorber is water cooled to improve absorption. The nitric acid solution from the absorber
is sent to a bleacher where air removes entrained gases and further oxidation occurs. The bleached
nitric acid solution is then either sent to storage or recirculated to the absorber after the addi-
tion of make-up water. The process employes a closed loop system to chill the recirculated acid
solution and tower cooling water by ammonia evaporation.
One variation in this system proposed by CDL/VITOK includes the addition of an auxilliary
bleacher operating in parallel with the primary unit. Another variation uses a secondary absorber
with its own bleacher.
At the Nitram, Tampa, Florida location two 318 Mg/d plants were fitted with the CDL/VITOK
process. NOX tail gas concentrations were reduced from 1500 to 1800 ppm to 600 to 800 ppm. With
the addition of a gulf catalytic abatement system the plant meets local regulations. A second plant
at Nitram fitted with the process showed promise but was shut down and replaced with a new nitric
acid plant.
TVA Process
The Tennessee Valley Authority, at their nitric acid plant in Muscle Shoals, Alabama, designed
and installed refrigeration for NOX abatement pruposes in 1972, in order to meet State standards of 2.75
6-14
-------
PURIFIED
TAIL GAS
COOLING
WATER
RETURN
FEED GAS/
LIQUID
FROM HEAT
EXCH.
ABSORBER
BREACH AIR
RECOVERED
COOLING
WATER
MAKE-UP WATER
PUMP
Figure 6-5. Schematic diagram of the CDL/VITOK NOX removal process (Reference 6-8).
6-15
-------
kg/Mg of nitric acid (5.5 Ibs/ton). A flow diagram of their abatement equipment is shown in Figure
6-6 (Reference 6-10). It consists of a cooler attached to the nitric acid absorption tower, and a
bleacher from which any effluent gases are recycled to the absorption tower. As a result of adding
the NOX control process the concentration of the product acid dropped from 65 percent to 51 to 57
percent.
6.1.3.2 Extended Absorption
The extended absorption process basically consists of a second absorption column to which the
tail gas from the nitric acid plant is sent. The NOX is absorbed by water and forms nitric acid,
which increases the acid yield. Extended absorption can be added in conjunction with pressurizing
the tail gas upstream of the tower or chilling the absorbent in the tower. However, neither of these
options is a necessary part of the absorption process.
This process is offered by several licensors, including J. F. Pritchard (Grande Paroisse
process), D. M. Weatherly, Chemico, Uhde and C and I Gridler (CoFaz process). The economics of the
process generally require the inlet pressure at the absorber to be at least 758 kPa (110 psia). Also,
cooling is usually required if the inlet N0x concentration is above 3000 ppm (Reference 6-6). There is no
liquid or solid effluent from extended absorption; the weak acid from the secondary absorber is
recycled to the first absorber, increasing the yield of nitric acid. In some cases, extended absorp-
tion can be used in conjunction with catalytic tail gas treatment (see Section 6.1.3.4).
Figure 6-7 shows a process flow diagram for the Grande Paroisse process, which is representa-
tive of extended absorption processes. Off-gas from the existing absorber flows into the secondary,
or Grande Paroisse absorber. The tail gas from the secondary absorber goes to an existing tail-gas
heater before being vented to the atmosphere or passing through a catalytic reduction unit. The
liquid effluent is returned to the primary absorber to become part of the acid product.
More than 15 extended absorption plants (by various licensors) are operating in the United
States. In cases where the off-gas must be compressed before going to the secondary absorber, or
where refrigeration is used, maintenance requirements are increased. Power recovery by an air
compressor/tail-gas expander is usually employed when a pressurized absorber is used.
6.1.3.3 Wet Chemical Scrubbing
Wet chemical scrubbing uses liquids, such as alkali hydroxides, ammonia, urea and potassium
permanganate to convert N02 to nitrates and/or nitrites by chemical reaction. Also, scrubbing may
be done with water or with nitric acid. Several of these processes are described below.
6-16
-------
ABSORPTION TOWER AND BLEACHER
DETAIL
COILS SUBMERGED -v ACID FLOW
IN ACID FOR COOLING
OR HEATING c
-v
G }
GAS'FLOW
STEAM
WATER
EXHAUST GASES-
WATER
BLEACHER
AIR
NEW
\IITRICACIE
BLEACHER
COMPRESSED
AIR
WATER
PRODUCT
NITRIC ACID
TO COOLER
T
COOLING
WATER
ABSORPTION
TOWER
SEAL PLATE
TNTE~FWAL~
BLEACHER
SECTION
*
REFRIG-
ERATED
COOLING
WATER
(50 °F)
WATER
\
COOLING
WATER HEADER
7
COOLER CONDENSER
NITRIC ACID GAS FROM AMMONIA
I
OXIDIZER
Figure 6-6. TV A, chilled absorption process (Reference 6-10).
6-17
-------
TO EXISTING
TAIL GAS PREHEATER
1 "
r »_L_ n !
1 EXISTING I ,,
ABSORBER P*] j
1 1 i
i
1 1
I 1
NITROUS j
GAS ^ | 1
1
, -i-
"L
LEVEL
CONTROL
7 K /r
[ r^v.
i. _. i
-n __j_ _,___
> ^
SECONDARY
ABSORBER
TOWER
_ NITRIC ACID ! RE.C.?^E.?ED
AC
ACID TRANSFER
PUMPS
(ONE SPARE)
^T^ C^r
\y ^J
ID
I {-&*
FLOW
CONTROL
/
-------
CO(NH2)2 + N2
+ HN02 t N2
+ *- H
H2o + H + NH]
+ HNCO
+ co2 -
i-
Urea Scrubbing
This process is offered by two licensors: MASAR, Inc. and Norsk Hydro. The mechanisms
given below have been proposed for this process (Reference 6-11).
2H20 (6-4)
20 (6-5)
HNCO + H20 + HT + NH4 + C02 (6-6)
When the concentration of nitric acid is low, reaction (6-6) predominates so that the overall
reaction is
HN02 + CO(NH2)2 + HN03 + N£ + C02 + NH4N03 + H20 (6-7)
As shown in reaction (6-7), half the nitrogen in the reaction will form NH4N03, a valuable by-
product, and half will form N2, a nonpolluting species.
The MASAR process is shown in Figure 6-8. A three-stage absorption column is used with gas
and liquid chillers on the feed gas and recirculated solvents. The process as described by MASAR,
Inc., (Reference 6-12) is given below.
The MASAR process, as applied to nitric acid plants, takes the tail gas from the exit of the
absorption tower and passes it to a gas chiller where it is cooled. During this cooling operation,
condensation occurs with the formation of nitric acid. This chilled gas and condensate passes into
Section A of the MASAR absorber. Meanwhile, the normal feedwater used in the nitric acid plant
absorption tower is chilled in Section C of the MASAR absorber and is then fed to Section A of the
MASAR absorber, where it flows down through the packing countercurrent to the incoming chilled tail
gas to scrub additional NOX from the tail gas. This scrubbing water is recirculated through a
chiller to remove reaction heat and then this weak nitric acid stream is fed to the nitric acid
plant absorber to serve as its feedwater.
The tail gas then passes into Section B of the MASAR absorber where it is scrubbed with a
circulating urea-containing solution. A urea/water solution is made up in a storage tank and
metered into the recirculating system at a rate necessary to maintain a specified minimum urea
residual content. As the solution scrubs the tail gases, both nitric acid and nitrous acids are
formed, and the urea in the solution reacts with the nitrous acid to form CO(C02), N?, and H90. As
the solution is circulated, the nitric acid content rises and some of the urea present hydrolyzes
and forms some ammonium nitrate. To maintain the system in balance, some of the circulated solution
is withdrawn. The recirculated solution is also pumped through a chiller to remove the heat of
reactions and to maintain the desired process temperature in Section B.
6-19
-------
FEED WATER
SPENT MASAR
SOLUTION
(BLOW DOWN)
LIQUID
CHILLER
CONCENTRATED MASAR
SOLUTION
PUMP
TAIL GASES
FROM PLANT
LIQUID
CHILLER
TAIL GAS
CHILLER
I ^
' PUMP
SECTION
/r>
SECTION
A
FEED WATER
TO NITRIC ACID PLANT
ABSORBER COLUMN
MASAR
ABSORBER
Figure 6-8. Flow diagram of the MASAR process (Reference 6-12).
6-20
-------
The tail gases then pass into Section C where they are again scrubbed by the feedwater
stream that is used, ultimately, as the nitric acid plant absorption tower feedwater. The tail
gases then leave the MASAR absorber and pass on to the normally existing mist eliminator and heat
exchanger train of the nitric acid plant. The cooling medium used in the gas chiller can be liquid
ammonia. The vaporized ammonia is subsequently used as the feed to the plant ammonium nitrate
neutralizes For non-ammonia nitrate producers, mechanical refrigeration could be used or the
ammonia vapor can be used in the nitric acid converter directly.
The MASAR process has been reported to reduce NO emissions from 4000 ppm to 100 ppm. The
process could technically be designed for no liquid effluent. In practice, however, liquid blowdown
of 16 kg/h (35 Ib/hr) of urea nitrate in 180 kg/h (396 Ib/hr) of water is estimated for a 320 Mg of
acid/day (350 tons/day) plant (Reference 6-12).
A MASAR unit installed in 1974 for Illinois Nitrogen Corporation on a 320 Mg/d plant regu-
larly operates with between 100 and 200 ppm of NO in the tail gas. According to the Illinois Nitro-
gen plant manager (Reference 6-13), inlet NO concentrations to the MASAR unit are approximately 3500
ppm and outlet concentrations are between 200 and 400 ppm. The Illinois Environmental Protection
Agency has tested this unit, using Method 7, with reproducible results of 57 ppm average emissions.
The unit is reported to operate with good reliability and has increased the net product recovered.
The Norsk Hydro process was developed by Norsk Hydro A/S, the Norwegian state-owned power
generating authority and fertilizer and chemical manufacturer, to reduce NO emissions from 1525 ppm
to 850 ppm. The modifications were made to an older, atmospheric pressure plant and two more recent
medium-pressure plants (300 and 500 kPa) (44 to 72 psia). Basically, the last absorption towers in the
process streams of the older plant were modified to contact the tail gases from all three plants with urea
solution and nitric acid. The result was a net 44 percent reduction in NO emissions, as given
above. On a plant-wide basis, 10.4 kg of ammonium nitrate are produced per Mg of nitric acid (20.8 Ib/ton)
(Reference 6-11).
Norsk Hydro has also used urea addition on three plants producing a total of 5 Gg/d (5500 tons/
day) of prilled NPK fertilizers. This method was used to control NO emissions for lower-grade phosphate
rock. Nitrous oxide is evolved when nitric and nitrous acid oxidizes impurities in the rock such
as sulphides and organic material. The addition of urea to the phosphate rock digester tends to
reduce N0x emissions to 2.5 kg/Mg (5 Ib/ton) phosphate from levels as high as 40 kg NO per Mg phosphate
(80 Ib/ton) by adding 5 to 10 kg urea per Mg phosphate rock (10 to 20 Ib/ton) (Reference 6-11).
6-21
-------
Ammonia Scrubbing (Goodpasture Process)
Goodpasture, Inc. of Brownfield, Texas is the licensor of a process developed in 1973 in
order for its Western Ammonia Corporation nitrogen complex in Dimitt, Texas to meet a 600 ppm
maximum NOX effluent imposed by the Texas Air Control Board. The process which was developed is
suitable to retrofit existing plants for reduction of an inlet concentration of 10,000 ppm to
within the 1.5 kg N02/Mg acid (^210 ppm) standards set for new nitric acid plants.
The process flow diagram for this process is shown in Figure 6-9. Feed makeup streams to
this process are ammonia and water with ammonium nitrate produced as a byproduct. The total process
is conducted in a single packed contact absorption tower with three sections operated in a co-
current flow. Goodpasture states that the key to successful operation is the process' capability
to minimize the formation of ammonium nitrite and to oxidize the ammonium nitrite which does form to
ammonium nitrate.
The Goodpasture process consists of three distinct sections. The first is a gas absorption
and reaction section operating on the acidic side, the second is a gas absorption and reaction
section operating on the ammoniacal side, and the third is principally a mist collection and ammonia
recovery step.
In the first section, a significant portion of the oxides of nitrogen react to form nitric
acid which maintains the acidic condition in this section. The nitric acid formed reacts with the
free ammonia content of the solution from the ammoniacal section to form ammonium nitrate - ^ portion
reacting in the acidic section, and a portion reacting in the ammoniacal section. The feed solution
to the acidic section is the product solution from the ammoniacal section. The ammonium nitrite
content of this solution is oxidized to ammonium nitrate by the acidic conditions existing in this
first section. The product solution from the Goodpasture process is withdrawn from this acidic
section.
In the second, or ammoniacal contacting, section the remainder of the oxides of nitrogen react
to form ammonium nitrate and ammonium nitrite; the proportion of each being dependent on the oxida-
tion state of the oxides of nitrogen in the gas phase. Ammonia is added to the circulating solution
within this section to maintain the pH at a level of 8.0 to 8.3. The liquid feeds to this section
are the product solution from the mist collection section, and a portion of the acidic solution from
the first section.
The third section is incorporated principally to collect the mist, and any ammonium nitrate
or ammonium nitrite aerosols which form in the first two sections. In addition, any free ammonia
6-22
-------
TAIL
GAS IN*
RECORD.
CONTROL.
TREATED
TAIL GAS
OUT
LEVEL
AMMONIACAL
SECTION
CONTROL.
AMMONIA--
LEVEL
CONTROL
HYDRAULIC
CONTROL
VALVE
pH
RECORD.
CONTROL,
STEAM
CONDENSATE
PRODUCT
AMMONIUM
NITRATE
SOLUTION
Figure 6-9. Process flow diagram for the Goodpasture process (Reference 6-14).
6-23
-------
stripped from the solution in the ammoniacal section is also recovered in this third section.
Process water or steam condensate is fed to this section in quantities sufficient to maintain the
product ammonium nitrate solution in the 30 to 50 percent concentration range. A small amount
of the acidic solution is also fed to this section in order to control the pH to approximately 7.0.
The product solution from the abatement process is withdrawn at about 35 to 40 percent
ammonium nitrate concentration, and contains approximately 0.05 percent ammonium nitrite. At the
Dimmitt plant, this solution is heated to 390K (240F), which completes the removal of the ammonium
nitrite, before further processing. Other users have discovered that if the solution sits for a
day in a day-tank, without heating, the ammonium nitrite is removed.
The Goodpasture process has been installed at CF Industries' Fremont, Nebraska plant and
Chevron Chemical's Richmond, California plant. In addition, American Cyanamid Company is installing
the process at one high-pressure and six low-pressure plants in Canada. Existing systems have
given reliable operation and have met the emissions requirements for which they were designed.
One particular advantage of this process is that the pressure losses in the process are only
6.8 to 13.0 kPa (1-2 psi) which allows its application to low-pressure plants. One older, 340 kPa
(49 psia) plant has consistently met its required 400 ppm outlet concentration. Another advantage
of the low-pressure drop is that reheat and power recovery of the effluent train in moderate-
pressure plants is usually economical. However, special precautions must be taken to eliminate
deposition of ammonium nitrate on the turbine blades.
Energy requirements of the process have been less than expected. The original design speci-
fied heating the ammonium nitrate scrubbing solution to facilitate oxidation of ammonium nitrite
to nitrate. However, it has been found that this reaction occurs spontaneously if the solution
is allowed to stand for a day in a holding tank.
The retrofit of a Goodpasture unit may require some additional process modifications beyond
the abatement equipment. For example, modern fertilizer plants use ammonium nitrate solutions in
excess of 85 percent. The Goodpasture byproduct solution is only 35 to 55 percent ammonium nitrate;
therefore, additional evaporators may be needed to concentrate the Goodpasture effluent. Chevron;
however, reports significant overall steam savings without additional evaporators.
Caustic scrubbing
Sodium hydroxide, sodium carbonate and other strong bases have been used for nitric acid
scrubbing. Typical reactions for this process are:
6-24
-------
2NaOH + 3N02 5 2NaN03 + NO + H20 (6-8)
2NaOH + NO + N02 * 2NaN02 + H20 (6-9)
A caustic scrubbing system was installed at a Canadian nitric acid plant in the late 1950's
(Reference 6-15). However, disposal of the spent solution is a serious water pollution problem,
and the concentrations of the salts are too low for economic recovery. There have been no recent
installations of this process.
Potassium Permanganate Scrubbing
Another potential chemical for scrubbing solutions is potassium permanganate. The Carus
Chemical Company (a large producer of potassium permanganate) has developed a process for potassium
permanganate solution scrubbing of NCy However, in the process, permanganate is reduced to manganate,
which must be electrolytically oxidized. The cost for the electrolysis, as well as the permanganate
make up cost, makes the process uneconomical. This process has not been installed at any nitric acid
plant in this country. Two plants are in operation in Japan, but no cost or user information is
available.
6.1.3.4 Catalytic Reduction
This section describes two different catalytic reduction processes. They are nonselective
catalytic reduction and selective catalytic reduction.
Nonselective Catalytic Reduction
In nonselective catalytic absorption, methane or hydrogen reacts with the NO and oxygen in
A
the tail gas to form N2. H20, and C02. A schematic of a typical catalytic reduction unit is shown
in Figure 6-10. The reactions (given in Section 3.3.2.4) in the abater are exothermic; and careful
temperature control is necessary for effective operation. The controls needed for operation as
a decolorizer are much less stringent.
Catalytic reduction units for decolorization and power recovery are used in about 50 nitric
acid plants in the United States. Many plants use natural gas for the reducing agent because of its
easy availability and low cost.* Some plants use hydrogen. When natural gas is used, the tail gas
must be preheated to about 753K (900F) to ensure ignition. A preheat temperature as low as 423K
(300F) is sufficient to ignite hydrogen.
Catalytic reduction is highly exothermic. The temperature rise for the reaction with methane
is about 128K (230F) for each percent oxygen burnout; with hydrogen it is about 150K (270F). For
* £.£" trU6i hOWeVer' the »<« °f «'"« "tur.1 gas have greatly
6-25
-------
TAIL GAS PREHEATER
TAIL GAS
FROM ABSORBER
REGENGAS
OUT
CH4/02
CONTROLLER
SET POINT
CH4 OFF WHEN
NH3 TO CONVERTER
IS OFF
I ON-OFF
j SET POINT-OPEN
A HIGH 02
-C*3 1
T' t
IOLLER I
( f
IV I I I
MIXER
ON-OFF
HIGH TEMP.
SET POINT
TEMP.
RECORDER
LJ :_
MOLECULAR SIEVE
DESULFURIZER
TEMP.
RECORDER/
CONTROLLER
ABATER
POWER RECOVERY
TURBINE
02
ANALYZER
CONTROLLER
CH4
ANALYZER
CONTROLLER
PROCESS
AIR
Figure 6-10. Nonselective catalytic reduction system (Reference 6-16).
6-26
-------
decolorization, the outlet temperature is ordinarily limited to 923K (1.200F), the maximum tempera-
ture limit of turboexpanders with current technology. Increased power recovery may justify adding
sufficient methane to reach the temperature limit of the turbine.
The tail gas must be preheated to 753K (90QF) to insure ignition when methane is used as the
reducing agent. Outlet temperatures would reach 1.088K and 1.138K (1.500F and 1,590F) for 2 and 3
percent oxygen burnout, respectively. These temperatures compare to the 923K (1,200F) maximum
temperature limit for single-stage operation. The oxygen in the tail gas cannot exceed 2.8 percent
to remain within the temperature limit of the catalyst. Cooling must therefore be provided to meet
the turboexpander limit. Older turbines may have even lower temperature limitations.
A somewhat cheaper but less successful alternative is two-stage reduction for abatement.
One system involves two reactor stages with interstage heat removal (Reference 6-17). Another
two-stage system for abatement involves preheating 70 percent of the feed to 753K (900F), adding fuel,
and passing the mixture over the first-stage catalyst. The fuel addition to the first stage is
adjusted to obtain the desired outlet temperature. The remaining 30 percent of the tail gas, pre-
heated to only 393K (250F), is used to quench the first stage effluent. The two streams plus the
fuel for complete reduction are mixed and passed over the second-stage catalyst; the effluent
passes directly to the turboexpander. This system avoids high temperatures and the use of coolers
and waste heat boilers (References 6-18, 6-19, and 6-20).
Honeycomb ceramic catalysts have been employed in two-stage abatement, with hourly gas-space
velocities of about 100,000 volumes per hour per volume in each stage (Reference 6-21).
Nonselective catalyst systems are offered by D. M. Weatherly, C & I Girdler and Chemico.
These systems are not as popular as NOX control methods because of rising fuel costs.
Two or three plants are known to have installed single-stage nonselective abaters. They are
believed to have been designed for natural gas. As noted above, oxygen concentration cannot exceed
about 2.8 percent. The reactors must be designed to withstand 1,088K to 1.118K (1.500F to 1,550F)
at 790 to 930 kPa, which requires costly refractories or alloys. Ceramic spheres are used as cata-
lyst supports, at hourly gas space velocities up to 30,000 volumes per hour per volume. One company
reports that they have been able to maintain NOX levels of 500 ppm or less over an extended period of
time. Operation close to 300 ppm might be attainable. On a plant scale, the effluent gas must be
cooled by heat exchange or quenched to meet the temperature limitation of the turbine. It may be
practical to use a waste heat boiler to generate steam.
Commercial experience with single-stage catalytic abaters has been modestly satisfactory,
but two-stage units operating on natural gas have not been as successful. Two-stage units designed
6-27
-------
for abatement have frequently achieved abatement for periods of only a few weeks, at which point
declining catalyst activity results in increasing NO levels. Recent data indicate that successful
abatement can be maintained for somewhat longer periods. Units that no longer abate NOX emissions
can, however, continue to serve for energy recovery and decolorization.
The success of single-stage abaters compared to the limited success of two-stage units may
result from the following factors: the catalyst is in a reducing atmosphere, the temperatures are
higher, and spherical rather than honeycomb catalyst supports are used. It has not been practical to
change catalyst type in two-stage units because the reactors designed for a space velocity of
100,000 volumes per hour per volume would be too small to accommodate a spherical catalyst, which
effectively removes NO at a space velocity of about 30,000. The failure of the honeycomb catalyst
in NO reduction compared to its success in decolorization may reflect that reaction kinetics make
it much more difficult to reduce NO than N02-
Fuel requirements for nonselective abatement with methane are typically 10 to 20 percent
over stoichiometric. Some hydrocarbons and CO appear in the treated tail gas. Furthermore, not
all methane is converted in decolorization reduction units. Less surplus fuel is required when
hydrogen is used.
Selective Catalytic Reduction
In selective catalytic reduction, ammonia is reacted with the NOX to form N2- No large
temperature rise occurs for ordinary operating conditions, so no waste heat or steam is generated.
The catalyst used in selective abatement units is platinum on a honeycomb support. Many catalytic
systems are installed between the expander and the economizer heat exchanger, and operate at
ambient pressure. This lack of pressure sensitivity is an advantage for retrofitting older low-
pressure nitric acid plants. It is important to control the temperature between 483K and 543K
(410F and 518F) because above 543K, ammonia may oxidize to form N0x; below 483K, it may form ammonium
nitrate.
Gulf Oil Chemicals is the main licensor of selective catalytic abaters in North America.
They have eight systems onstream, and two more planned. Of these systems, nine operate at
ambient pressure, and one operates at 590 kPa (86 psia). Many of these catalyst beds also use a
molecular sieve for N02 adsorption to promote the reaction with ammonia.
Uhde licenses the BASF selective catalyst process and recommends it for tail gas treatment
of 600 kPa (87 psig) nitric acid plants.
User experience with these processes has been good. Catalyst lifetimes of over 2 years have
been reported, and expected lifetime is 5 to 10 years. Catalytic processes have also been used to
supplement chilled absorption units when they have failed to meet emission limits.
6-28
-------
6.1.3.5 Molecular Sieve Adsorption
The main equipment in a molecular sieve adsorption system is in the form of a two-section
packed bed. The first section is packed with a desiccant, since the NOX adsorption sieve material
works best on a dry gas. The second section contains the material which acts as nitrous oxide
oxidation catalyst and NO adsorber.
Figure 6-11 is a schematic of a molecular sieve system added to an existing nitric acid
Plant. N0x removal is accomplished in a fixed bed adsorption/catalyst system. The water-saturated
nitric acid plant absorption tower overheat stream is chilled to 283K (50F), the exact temperature
level being a function of the NOX concentration in the tail-gas stream. It is then passed through
a mist eliminator to remove entrained water and acid mist. The condensed water, which absorbs
some of the NO,, in the tail gas to form a weak acid, is collected in the mist eliminator and either
recycled to the absorption tower or sent to storage. The tail gas then passes through a molecular
sieve bed where the special properties of the NOX removal bed material results in the catalytic
conversion of nitric oxide (NO) to nitrogen dioxide (N02). This occurs in the presence of the low
concentrations of oxygen typically present in the tail-gas stream. Nitrogen dioxide is then selec-
tively adsorbed.
Regeneration is accomplished by thermally cycling (or swinging) the adsorbent/catalyst bed
after it completes its adsorption step and while it contains a high adsorption loading of N02. An
oil-fired heater is used to provide heat for regeneration. The required regenerator gas is obtained
by using a portion of the treated tail-gas stream for desorption of the NO,,. This N02-loaded gas
is recycled to the nitric acid plant absorption tower. The pressure drop in the molecular sieve
averages 34 kPa (5 psi) and N0x outlet concentration averages 50 ppm (Reference 6-22).
This process has been applied to three plants in the United States (Reference 6-6). Tables
6-2 and 6-3 show the performance of the system at two installations. The commercial name for
the process is the Purasiv N process. The unit at the 50 Mg/d (55 tons/day) acid plant of
Hercules, Inc. started up in 1974. Abatement ranged from 95.9 to 98.7 percent averaged over indivi-
dual cycles and was highest at the beginning of a cycle (Reference 6-23). The U.S. Army Holston
Purasiv N unit was started up in August 1974, but has been inoperable for several years.
Both plants have dual-unit NOX adsorbers, operating on a 4 hour adsorption, 4 hour regenera-
tion cycle (Reference 6-22). Initial reports on the operation were very favorable; the effluent
standards were met, and the sieve showed no noticeable deterioration after 6 months. One sieve was
damaged by accidental acid back-up, however, and did not achieve the expected 50 ppm outlet concen-
tration.
6-29
-------
ABSORBER
(EXISTING)
TAIL GAS
CONTAINING NOX
1 ^
1
i
|
1
i
!-»
!
HOT GAS
CONTAINING
DESORBEDN02
S
^.
S
k,
NO OXIDATION TO N02
AND N02-H20
ADSORPTION
REGENERATION
1
X.
/
^
/
:LEAN DRY^ >v
TAIL GAS ( ^
\^J
-^- 1 HEAT
j EXCHANGER
! (EXISTING)
4 '
POWER
RECOVERY
(EXISTING)
HOT GAS
Figure 6-11. Molecular sieve system (Reference 6-22).
6-30
-------
>-
g
UJ
UJ
L_
n^
i i
oc
:D
o
h-
i i
z
:D
z
>
^^
to
S
ID
D_
OO
UJ
__l
_J ^"""^
o «3-
a: CM
1 1 1 i
*T* V£>
ll 1 1 1
O 0
~y
UJ UJ
o a:
Z UJ
r>. in CM
CM in CM r- ^- vo
in i ^-i ...
O * O CM O CM
CM O
CO
CM
o
o
in
o * cy> in <*
!»« O CM r ^- VO
CM 1 «tf- 1 ...
O * O CM O CM
CM O
CO
CM
O
O
O
o * cj> in ^3-
VO O CO CM «d- l*s.
CM 1 VO 1 ...
O «O CM O CM
CM O
O
CM
to to
*v» <">
"E E
z 5
0
2 " """ 2
4-> O 5 «+- 0
C E E r- 0 ^~
O) Q. Q. U. r (/) »4-
3 0. E d. E M- ro
« Q. XJ O. (/) O> «/)
M- 'Q.rjj »>CLrot/) ro
* 1 1 u> > d7),_
rOQj ro>*^ ro
r- (1) C *r- d) C ro >ra S- fO
£
t|_
O
v»
o^
.
00
r
r
CM
II
(/)
^
m
E
r
to
6-31
-------
CD
CO
O
I-
CO
CM
>-
O
UJ ^
CO U. CO
uj o: CD
o Q:
z.>-
co o «3-
o °^o
CO 0
CM
o
o
o o *
co o *±
vo i r^ i
o > o
CO O
in
CO
O
o
o >
CM o in
VI r 1
o > o
*d- o
o
CO
+'
CU Q. Q.
3 0. E Q. E
r- Q. -o Q.
M > Q. QJ < d.
M- (DO)
uj cn * u. cn >
(O QJ fO QJ
c S- cn c J- cn
r- O) C !- fO > ro
x <: a: x
r
O CO
ro QJ E
I o: z
o
VD
CM
0
VO
CM
O
UD
CM
CO
CO
E
^
O
CO
cn
ro
O
H-
o
CO
cr>
CO
CM
It
to
fO
6-32
-------
The process has been successful in meeting emission standards. The principal criticisms
have been high capital and energy costs, and the problems of coupling a cyclic system to a continu-
ous acid plant operation. Furthermore, molecular sieves are not considered as state-of-the-art
technology.
6.1.4 Costs
The most recent cost and energy utilization comparisons of the various abatement processes
are given in Tables 6-4 and 6-5 (Reference 6-6). Direct comparison of these data is rather difficult
since not all the side effects, such as changes in plant yield, and the degree of abatement, are
described.
Chilled Absorption
The cost figures in Table'6-4 for the CDL/VITOK process are in agreement with data provided by
Reference 6-25. According to Reference 6-10, the bottom line costs for the chilled absorption process
used by the TVA is $2.09/Mg ($1.90/ton) of acid, which includes $0.14/Mg ($0.13/ton) credit for additional
product, 8 kWh and 85 kg steam per Mg acid (170 Ib/ton). This cost is higher than the $1.74 Mg ($1.58/ton)
given in Table 6-4 and does not include the reduction in capacity caused by the reduction in the nitric
acid concentration.
Extended Absorption
The Grande Paroisse process is capital intensive; therefore, costs may be dominated by the
assumptions made to calculate return on investment and depreciation. The figures in Table 6-4 reflect
a 20 percent return on capital. The Grande Paroisse literature shows cost of $0.98 to $1.13 per Mg
($0.89 to $1.03/ton) but does not consider a return on investment cost.
Even with high capital cost and unfavorable cost of capital, extended absorption is competi-
tive with other processes. It has low maintenance costs and low energy requirements.
Wet Chemical Scrubbing
The economics and energy use of two wet scrubbing processes, MASAR, urea scrubbing, and Good-
pasture, ammonia scrubbing, are given in Tables 6-4 and 6-5. Costs for the Norsk Hydro process would
be similar if applied to a new plant. Capital cost for the Goodpasture process are estimated as
$425,000 for a 270 Mg/d (300 tons/day) plant (Reference 6-26). No costs estimates are available for
potassium permanganate and caustic scrubbing since they are not in general use.
6-33
-------
H-re
cn
_J S
I O
ii r-»
Q. CM
o <:
^""Y
o
^^
o>
"U
c
re
i-
0
J-
re
0
0)
"o
3E
''"^
Vk
^^
'
1 \
re
Ol
i_
to
CL
o
o
cS
y~
re
re
sr
>f*
o
£
^>
cu
to
to
r-
O
i.
re
a.
cu
>
UJ
to
C
o
r-
4-J
U
3
o
Ol
OO O O LO O O
OO VO O i O O
O O CO CM OO CM *±
LOi CM CM "3-
CMLO
O O O LO O LO
o vo o i o r^
O 00 CM 00 CM Cn
OO CU CM CM LO
VO CU
LO U_
O O O LO O O
O VO O O O
O OO CM CO CM «d-
ft cu * » *
LO C CM CM «d-
r*** o
LO Z
O CD O LO O O
o -a vo o o o
O CU OO CM OO CM ^t
ft "Q * ft ft
O 3 CM CM «*
0 .
O 0
f C
O O O LO O O
O VO O r O O
O 1 00 CM CO CM *
O CM CM "d-
O
CM
r~
O O O LO O O
O VD O i O O
O 1 OO CM OO CM *J-
* I ft ft ft
*t CM CM «*
00
OO
CU
cn
c
^""N ^*""*» Sw
bO- fc. 4_
*"""'* ^>
^V^
XJ i- r
" ^ * °
C >~X lT T-
Q) » O **»^'
E I* o
+-> o re *o
to xi _j re
cu re oi
> J ft) X:
C 0 J-
HH cn c 01
c re >
"re +j *J IT cu IT
Mr- re >>4-> >> s-
i- re J- --» c o
a. >> 01 «/» -I- ** x»
re o cu» - re * re
o a: o z: i
i
^J- 0
1 1 1
III
LO O
1 CO 1
rs>
*± o o o
cn r^ CM vo
LO CD CM
CM CO
oo
1 1 1
1 1 1
" 1
' '
1
1
1
'
1 1 1 1
1 1 1 1
oo oo oo
VO ^f CM
^ r>- CM LO
cn r- ^f
00 t^
s~**
00
c.
^-^. - »- -
O O O O
o «* o o
O VO O LO
CM 00 LO CO
CM 00 00
O VO
cn *±
CM
CM '
*
- ^O O O O 00 CM
oo
^J
w -.
i O «* O LO O *t O
i.
>fc
^^
faO-
*
C
o
r
to
S-
O)
a.
3
*c
I/)
4^
£
Ol
c
Ol
X)
vo cc
i oo cn
i
co st
r
LO O
OO CM
1 i <5J-
r~~ ^*
O CM O
o cn co
vo oo co
LO i r^
^-
0
o
CO 1 1
1 1
f^
^>^
01
Oi
^
oo
^" *»
J- C
re !-
r- E
3 ~^,
(J r-
0J X "
p«>
2 *
"^ Ol
t- 4-*
o re
4.)
to cn^-v
r~ *r- ^)
re r ^.
re o '
o o
CM O VO LO
OO LO CM CM
r^. oo
T~~ ^&
o o
1 1 VD
1 1
!_
00 O CM O
t CM CM LO
i r OO LO
VO CM
LO
1 |
1 1
1 1
1 1
1
1
1
1
0
LO O 1
CM
O
^^
O
LO O 1
*~* ft 1
CM
0
~
VO O O O
^j-
0
1 O VO VC CM 1
1
CM O 00 O CM O CO
00 CM CM
r*** r**. o
^-"00 CM
co
^_^
, f
3E
i- +>*"'
x: !-
^^ TD ^">
cn cu 4->
^ i. -r-
« o o
« i_ T'zr
E >>+-> >>
« u ^^
^ > »*^ p~ *-
C/) LU
CO LO
00 CO
N
CM
f
^ X
C
!"
E
^"^.
^~
^^.x
J-
>
!-*«» O>
O' 3
OO U.
CM
CO
0
CM
r
ft
LO
vo
i_
*^
-* - 1
CM
^^
d*
1 1 1
1 1 1
^^
^.
re
T3
O)
^^
*-* *^
« re
T> -o
o cn
«r z:
r* >N «> &
4J4^ S -^?
r* ^^ ^ ^)*}
Z => '
^^
>>
re
o
*"X^
C7)
*-*
t
Ol
E
4->
^
^^
E
c
P
*
O O O VO
LO 00 VD CM
^1- O CM OO
CM VO 00 r
~ ,-
O O O
o o o
OO O LO
CM LO r
LO i r
r~
O O O
f 0 O
cn o o
o o o
cn o CM
CM
f^ t
r~.
LO
cn
o *s-
co i-^
CO
i >r~
10
o vo
CO LO
r-.
"CM
vo
oo
CM
o o o
cn o o
0 0 O
cn o «=t-
o ** CM
i CM
O 00
00 «3"
cn
oo
^*
^
O O O
o o o
CT> 00 t^.
» ^ ft
LO vo r^
CM r^ CM
CM
f> tf^
00
CM
VO
>«,
^^
<^^K <*"^
Ol CM * .
t^ *~^* c
r- CS) *5 >j
f * CM
I * fOj v,.^
>> CU > cn
i E cu +-) s:
i 4-> O 10 ^
Oi re cj
> S-
C C 3 i «
o -< to re +->
f- C 3 to
4-> C "-. C O
re o co
r- cC <
""^ (J C r
t cu J- to re
>> S- 3 Ol re 3
«v. D-4-> X 4J C
«^ a> oi re o c
« O OC 1 H- «C
to
4->
C
3
x:
to
|H
cn
c.
LLl
C
r-
f.
X
r~
T3
C
0)
a.
Q.
r-
cn
I/I
^
Ol
^
XJ
re
4-1
to
r»
x:
o
o
o
c
re
S
to
Ol
to
o>
a.
X
0)
to
i.
0)
c
^
Q
*
c
o
£
u
3
i.
4->
to
o
0
cn
c
t-
3
o
4->
to
CU
s_
01
4-*
c
r
cu
t3
3
"o
X
Ol
to
o>
4->
re
E
4.}
in
Oi
4->
C
Ol
>s
^
to
Ol
>
c
c
Oi
a.
to
Ol
_c
1 *
c
»^»
c
cu
to
o>
i-
o.
TJ
Ol
u
3
-o
o
s_
Q.
-D
r~
O
re
o
s_
4->
C
it
O
cn
S
re
o>
u
3
<*-
O
cn
2!
r>v
o
o
0
o o
v»- a.
4*t
f- to
TJ
O> O
0*~*
C
Ol O
T3 -
3 4->
i 3
O r
C O
^
C
O)
J*l
re
4.)
4-J
^0
O)
i_
o
Ol
4->
re
o
-o
c
to
to
Ol
x:
4J
c
O)
re
6-34
-------
TABLE 6-5. ANNUAL ENERGY REQUIREMENTS (TJ) FOR NOX ABATEMENT SYSTEMS
FOR A 270 Mg/d NITRIC ACID PLANT (Reference 6-6 and 6-26)*
Steam (Credit)
Electrical
Natural Gas
Oil
Basic Nitric
Acid Plant
(75.2)
172.0
-
96.8
Catalyst
Reduction
(136.18)
11.56
245.12
-
108.94
Molecular
Sieve
2.15
29.08
-
17.20
48.43
Grande
Paroisse
-
8.13
-
-
8.13
CDL/
Vitok
6.14
23.94
-
-
30.08
Masar
11.27
1.80-
-
-
13.07
Goodpasture
-
1.38
-
-
1.38
This table is given in Appendix A in English units.
TABLE 6-6. BASIS FOR TABLES 6-4 AND 6-5 (Reference 6-6)'
(Plant Capacity 270 Mg/day and 92 Gg/yr)
(March 1975 Dollars, ENR Index = 2.126)
1. Operating Labor
2. Maintenance Labor
3. Overhead
4. Cooling Water
5. Boiler Feedwater
6. Natural Gas
7. Oil
8. Depreciation
9. Return on Investment
10. Taxes and Insurance
11. Nitric Acid
12. Urea
13. Ammonium Nitrate
14. 1 kWh = 11.07 MJ
15. Electricity
16. Ammonia
(3 $6.1/hr
(3 $7.0/hr
(3 100% of labor (including fringe
benefits and supervision)
(3 $0.008/1000 1
(3 $0.20/1000 1
(3 $1.90/GJ
(3 $1.90/GJ
@ 11 yr straight line
@ 20% of capital cost
(3 2% of capital cost
(3 $ 99/Mg
(3 $176/Mg
(3 $110/Mg
(3 $0.02/kWh
(3 $173/Mg
JThis table is given in Appendix A in English units.
6-35
-------
Capital and operating costs for these processes are very low and are aided by credit for the
byproducts (ammonium nitrate). In the Goodpasture process approximately 75 percent of the ammonia
is reclaimed as ammonium nitrate.
Catalytic Reduction
The cost and energy data given in Tables 6-4 and 6-5 are for a natural gas fired nonselective
catalytic reduction unit. The process is considerably more expensive than the other processes. Not
only does a catalytic combustor have a high capital cost, but fuel costs are large (and will probably
increase).
Costs for selective catalytic reduction are not included in Table 6-4. Capital costs are
estimated as $100,000 to $125,000 for a 270 Mg per day unit by Gulf (Reference 6-27). Operating
and maintenance costs are expected to be low except possibly for catalyst replacement. The major
operating expense is the cost of ammonia for reaction with NO .
Molecular Sieve
Both capital and operating costs for the molecular sieve process are high. Fuel for the
regeneration phase, high maintenance costs, and catalyst replacement are the primary contributors to the
operating costs. Not included in the cost figures are any extra costs which may result from upsets or
process alterations in the nitric acid plant as a result of the cyclic operation of the abatement unit.
6.2 NITRIC ACID USES
Important uses of nitric acid and the estimated quantities consumed in each are listed in
Table 6-7. Approximately 65 percent of the nitric acid produced in the United States is consumed
in making ammonium nitrate, of which approximately 80 percent is used for fertilizer manufacturing.
Adipic acid manufacture, the second largest use, consumes only about 7 percent. Other uses include
metal pickling and etching, nitrations and oxidations of organic compounds, and production of
metallic nitrates.
6.2.1 Ammonium Nitrate Manufacture
6.2.1.1 Process Description
Ammonium nitrate is produced by the direct neutralization of nitric acids with ammonia:
NH3 + HN03 - NH4N03 (6-10)
6-36
-------
TABLE 6-7. ANNUAL NITRIC ACID CONSUMPTION IN THE UNITED STATES, 1974
(Reference 6-3 and 6-6)
Product
Ammonium Nitrate
Adi pic Acid
Nitrobenzene
Potassium Nitrate
Miscellaneous Fertilizers
Military, other than
NH4N03
Isocynates
Steel Pickling
Other
Total Nitric Acid
Production
Quantity of HNO, used in manufacture
Gg
4830
520
74
37
371
258
in
37
1193
7431
103 tons
5324
573
82
40
409
286
122
41
1315
8192
6-37
-------
About 735 kg (1600 Ib) of nitric acid (100 percent equivalent) and 190 to 205 kg (420 to 450 Ib) of
anhydrous ammonia are required to make 909 kg (1 ton) of ammonium nitrate. In actual practice, 100
percent nitric acid is not used, and typical feed acid contains 55 to 60 percent HN03- The product
is an aqueous solution of ammonium nitrate, which may be used as liquid fertilizer or converted
into a solid product. The heat of reaction is usually used to evaporate part of the water, giving
typically a solution of 83 to 86 percent ammonium nitrate. Further evaporation to a solid may be
accomplished in a falling-film evaporator (Reference 6-28), in a disk-spraying plant (Reference
6-29), or by evaporation to dryness in a raked shallow open pan (graining). The graining process
is no longer used due to hazardous conditions.
A majority of the solid ammonium nitrate produced in the United States is formed by "prilling",
a process in which molten ammonium nitrate flows in droplets from the top of a tower countercurrent
to a rising stream of air, which cools and solidifies the melt to produce pellets or prills (Refer-
ence 6-3).
6.2.1.2 Emissions
No significant amount of NOX is produced in this process; the most likely source of nitric
acid emissions would be the neutralizer. The vapor pressure of ammonia, however, is much higher
than the vapor pressure of nitric acid, and the release of nitric acid fumes or NO is believed to
be negligible (Reference 6-30), especially since a slight excess of NH^ is used to reduce product
decomposition.
6.2.2 Organic Oxidations
6.2.2.1 Process Description
Nitric acid is used as an oxidizing agent in the commercial preparation of adipic acid,
terephthalic acid, and other organic compounds containing oxygen. The effective reagent is probably
NOp, which has very strong oxidizing power.
Adipic acid (COOH-(CH2).-COOH) is a di-basic acid used in the manufacture of synthetic fibers.
It is an odorless white crystalline powder which is manufactured by the catalytic oxidation of cyclo-
hexane, with cyclohexanone and cyclohexanol as intermediates. About 618 Gg (681,000 tons) of adipic
acid were manufactured in 1975 (Reference 6-31). Approximately 90 percent of adipic acid is consumed
in the manufacture of nylon 6/6.
In the United States, adipic acid is made in a two-step operation. The first step is the
catalytic oxidation of cyclohexane by air to a mixture of cyclohexanol and cyclohexanone. In the
6-38
-------
second step, adipic acid is made by the catalytic oxidation of the cyclohexanol/cyclohexanone mix-
ture using 45 to 55 percent nitric acid. The product is purified by crystallization (Reference 6-32).
The whole operation is continuous. The chemistry of the reactions in the two steps is:
cyclohexanone + nitric acid * adipic acid + NO + HLO (6-11)
cyclohexanol + nitric acid -» adipic acid + N02 + H20 (6-12)
The main nitrogen compounds formed in the above reactions are NO, NO,,, and N,,0. The dissolved
oxides are stripped from the adipic acid/nitric acid solution with air and steam. The NO and NOp
are recovered by absorption in nitric acid. The off-gas from the NOX absorber is the major contri-
butor to NOX emissions from the adipic acid manufacturing process.
Nitric acid is used for the oxidation of other organic compounds in addition to the adipic
acid, but none approaches the adipic acid product volume.
Terephthalic acid is an intermediate in the production of polyethylene terephthalate, which
is used in polyester, films, and other miscellaneous products. Terephthalic acid can be produced
in various ways, one of which is by the oxidation of paraxylene by nitric acid (Reference 6-33).
In 1970, the process was used for about a third of terephthalic acid production and accounted for
approximately 20 percent of NO emissions from nitration processes. Since 1975, however, the use
of nitric acid as a feedstock in the production of terephthalic acid has been discontinued (Reference
6-34). No NO is now generated in terephthalic acid plants.
6.2.2.2 Emissions
The off-gases leaving the adipic acid reactor after nitric acid oxidation of organic materi-
als may contain as much as 30 percent NO before processing for acid recovery (Reference 6-35).
One of the principal compounds of the off gas, N20, is not counted as NOX, since it is not oxidized
to NO in the atmosphere and is considered harmless. The seven adipic acid manufacturing plants in
the United States generated about 14.5 Gg (16,100 tons) of NOX in 1975 (Reference 6-31) from a total
acid production of 618 Gg (681,000 tons). This gives an average emission factor of 23.7 kg N02/Mg
(47.4 Ib N02/ton) compared to the nominal value 6 kg N02/Mg (12.0 Ib N02/ton) specified by AP-42
(Reference 6-36).
6.2.2.3 Control Techniques
In commercial operations, economy requires the recovery of NO as nitric acid. It is recov-
ered by mixing the off-gas with air and sending the stream to an absorbing tower, where nitric acid
is recovered as the stream descends and unrecoverable N20 and nitrogen pass off overhead.
6-39
-------
If the resulting emission rates are too high, further reduction could be attempted by stan-
dard techniques such as extended absorption or wet chemical scrubbing. These techniques are
described in Section 6.1.3. A potential, long-range control for eliminating NO from organic oxi-
dation processing is the replacement of nitric acid as an oxidant by catalytic processes using air
oxygen. The laboratory catalytic oxidation of cyclohexanol and cyclohexanone by air to adipic acid
has also been reported, but no commercial process is known (Reference 6-37).
6.2.2.4 Costs
Economy requires that nitric acid be recovered from reactor off-gas in large-scale organic
oxidations using nitric acid as the oxidizing agent. For example, the incentive for acid recovery
for a 270 Mg/d (300 tons/day) adipic acid plant would be about $2.48 x 106 per year. This figure is
based on recovering 0.3 kg of HN03 per kg of adipic acid at a nitric acid cost of $8.6 per 100 kg (Ref-
erence 6-38). The optimum economic recovery level depends upon economic factors at each installation.
6.2.3 Organic Nitrations
6.2.3.1 Process Description
Nitration is the treating of organic compounds with nitric acid (or N02) to produce nitro
compounds or nitrates. The following equations illustrate the two most common types of reaction:
RH + HON02 -> RN02 + H20 (6-13)
ROM + HON02 + RON02 + H20 (6-14)
Examples of products of the first reaction (C-nitration) are compounds such as nitrobenzene, nitro-
toluenes, and nitromethane. Nitroglycerin (or glycerin trinitrate) and nitrocellulose are examples
of compounds produced by the second reaction (0-nitration).
Nitrating agents used commercially include nitric acid, mixed nitric and sulfuric acids
(mixed acids), and N02- Mixed nitric and sulfuric acid is most frequently used. The sulfuric acid
functions to promote formation of N02 ions and to absorb the water produced in the reaction.
Nitrations are carried out in either batch or continuous processes. The trend is toward
continuous processes, since control is more easily maintained, equipment is smaller, system holdup
is smaller, and hazards are reduced. A multiplicity of specialty products such as dyes and drugs,
which are produced in small volumes, will continue, however, to be manufactured by small batch
nitrations.
6-40
-------
Batch nitration reactors are usually covered vessels provided with stirring facilities and
cooling coils or jackets. The reactor bottom is sloped, and product is-»withdrawn from the lowest
point. When products are potentially explosive, a large tank containing water (drowning tank) is
provided so that the reactor contents can be discharged promptly and "drowned" in case of abnormal
conditions.
When the reaction is completed, the reactor contents are transferred to a separator, where
the product is separated from the spent acid. The product is washed, neutralized, and purified;
spent acids are processed for recovery. Figure 6-12 illustrates a batch nitration process for
manufacturing nitroglycerin (Reference 6-39).
Continuous nitration for nitroglycerin is carried out in many types of equipment. Two widely
employed processes are the Schmid-Meissner process (illustrated in Figure 6-13) and the Biazzi pro-
cess (illustrated in Figure 6-14). Both processes provide for continuous reaction, separation,
water washing, neutralization, and purification. The Biazzi process makes greater use of impellers
for contacting than the Schmid-Meissner, which uses compressed air to provide agitation during
washing and neutralizing. Both types of equipment can be used for nitrating in general.
When mixed acid is used, the spent acid is recovered in a system similar to that shown in
Figure 6-15. The mixed acid enters the top of the denitrating tower. Superheated steam is admitted
at the bottom to drive off the spent nitric acid and NOX overhead. The gases are passed through a
condenser to liquefy nitric acid, which is withdrawn to storage; the uncondensed gases are then
sent to an absorption tower. Weak sulfuric acid is withdrawn from the bottom of the denitrator
tower and concentrated or disposed of by some convenient arrangement.
When nitric acid alone is used for nitration, the weak spent acid is normally recovered by
sending it to an absorption tower, where it replaces some of the water normally fed as absorbent.
Nitrobenzene and dinitrotoluenes are produced in large volumes as chemical intermediates.
Explosives such as TNT, nitroglycerin, and nitrocellulose are produced in significant but lesser
volumes.
Nitrobenzene is manufactured in both continuous and batch nitration plants. Mixed acids
containing 53 to 60 percent H2S04, 32 to 39 percent HN03, and 8 percent water are used in batch
operations, which may process 3.785 m3 (1000 gallons) to 5.678 m3 (1500 gallons) of benzene in 2 to
4 hours. Continuous plants, as typified by the Biazzi units (Figure 6-14) also use mixed acids.
The major use of nitrobenzene is in the manufacture of aniline. It is also used as a solvent.
Nitrobenzene production in 1970 was an estimated 188 Gg (207,500 tons). Nitric acid requirements
6-41
-------
r
NITRATING HOUSE
MIXED
ACID
D
MIXED-ACID
STORAGE
CERIN
1
J
1
i '
MIXED-ACID
SCALE TANK
GLYCERIN
V
~\
1
| 4
1
1
1/flTFR. .. i i -
NITRATOR
t
GLYCERIN
SCALE TANK
* t
L
SEPARATOR
/
1
SOUR-
WATER
ffij ^G
PREWASH
WATER +
h
A.
NG
,-1-
GLYCERIN
HEATER
HOUSE
SPENT ACID
CATCH
TANK
CID
USE
WATER
1
i NG
: 1 ,
i
1
1
1
i
L
i-4
i
SODA ASH U
r
SODA
WATER
SODA
WATER
J
_ S
^
/
t
WASTE
WAT El
-I NEUTRALIZ
HOUSE
NEUTRALIZER
/WATER *
t NG
NITRIC ACID
RECOVERY HOUSEl
DENITRATING
TOWER
WASTE
WATER
BUGGY
CATCH
TANK
NITROGLYCERIN
TO POWDER
CATCH
TANK
RECOVERED WEAK
NITRIC SULFURIC
ACID ACID
WASTE
WATER
Figure 6-12. Batch process for the manufacture of nitroglycerin (NG) (Reference 6-39).
6-42
-------
GLYCERIN
FEED
AGITATOR
MIXED
ACID"
WASH
WATER
TO BAFFLED
SETTLING TANKS
WATER
SEPARATOR
NITROGLYCERIN
SODA WATER
CO _l
<0
n
AIR
AIR
TO DROWNING
TANK
WATER
IAIR
(PASSED THROUGH ADDITIONAL NITROGLYCERIN
WASH COLUMNS IF NECESSARY)
Figure 6-13. Schmid-Meissner continuous-nitration plant (Reference 6-39).
6-43
-------
AGITATOR
GLYCERIN
FEED
AGITATOR
WATER
TO
DROWNING
TANK
SPENT ACID
TO
LEVELING DEVICE
SEPARATOR
LEVELING
DEVICE
NITRO
GLYCERIN7
ACID
TO DROWNING
TANK
WASTE
- WATER
WATER
SEPARATOR
AGITATOR
NITROGLYCERIN
LEVELING
DEVICE
SODA ASH
SOLUTION
t
SPENT ACID
TO AFTER-SEPARATOR
AND STORAGE
MECHANICAL
NEUTRALIZER
1
TO ADDITIONAL
WASHERS AND
SEPARATORS
WASH WATER TO SEPARATOR
Figure 6-14. Biazzi continuous-nitration plant (Reference 6-39).
6-44
-------
AIR
GASES TO
ABSORPTION TOWER
S-BEND
CONDENSER
CHEMICAL-WARE
BLOCKCOCK
NITRIC ACID
TO STORAGE
NITRIC
DISTILLATE
SAMPLER
SULFURICACID
COOLING TUB
Figure 6-15. Recovery of spent acid (Reference 6-39).
6-45
-------
are approximately 0.54 kg per kg of nitrobenzene (Reference 6-39). On this basis, nitric acid used
in nitrobenzene synthesis was estimated at 126 Gg (139,000 tons) for 1970.
Dinitrotoluene is manufactured in two stages in both continuous and batch units. The first
stage is the nitration of toluene to mononitrotoluene, which is nitrated to dinitrotoluene in the
second stage. For making mononitrotoluene in the batch process, mixed acid consisting of 28 to 32
percent HNOg, 52 to 56 percent H2S04> and 12 to 20 percent water is used in equipment sized to
handle up to 11.4 m3 (3000 gallons). Operating temperature ranges from 298K (77F) to 313K (104F).
Mononitrotoluene yields of 96 percent are typical (Reference 6-40). The second step, the production
of dinitrotoluene, is carried out separately because it requires more severe conditions.
Dinitrotoluene is made from mononitrotoluene using stronger mixed acid containing 28 to 32
percent HN03, 60 to 64 percent H,,S04, and 5 to 8 percent water. Temperatures are increased to 363K
(194F) after all the acid has been added. Dinitrotoluene yields are about 96 percent of theoretical
(Reference 6-41).
The principal use of dinitrotoluene is as intermediate in making toluene diisocyanate (TDI)
for use in polyurethane plastics. It is usually supplied as mixtures of the 2,4 and 2,6 isomers.
6.2.3.2 Emissions
Relatively large NOX emissions may originate in nitration reactors and in the denitration of
the spent acid. NOX is also released in auxiliary equipment such as nitric acid concentrators,
nitric acid plants, and nitric acid storage tanks.
Nitration reactions per se do not generate NO emissions. NOV is formed in side reactions
A X
involving the oxidation of organic materials. Relatively little oxidation and NO formation occur
when easily nitratable compounds, such as toluene, are processed. Much more severe conditions are
required in processing compounds that are difficult to nitrate, such as dinitrotoluene; more oxida-
tion takes place and, thus, more NO is formed.
Limtied data are available on actual NOX emissions from nitrations. For continuous nitra-
tions, one company has reported emissions of 0.06 to 0.12 kg N02 per Mg of nitric acid (0.12 to 0.24
Ib/ton), with a mean of 0.09 kg N02/Mg (0.18 Ib/ton) at a single location (Reference 6-40). At the
same location, emissions averaging 7 kg of N02 per Mg of acid were reported in manufacturing specialty
products in small batch-type operations. According to Reference 6-42, 0.25 kg of N02 per Mg of nitric
acid (0.5 Ib/ton) are generated in the production of nitrobenzene. In the manufacture of dinitrotol-
uence, 0.135 kg of N02 is estimated to be generated for every Mg of nitric acid used (0.27 Ib/ton).
6-46
-------
Using the Reference 6-42 emission factors as a lower limit, and 7 kg NOX per Mg HNO-, (14 Ib/ton),
(Reference 6-40) as upper limits for nitrations, the NOV emissions in 1970 would have the range indi-
x »
cated in Table 6-8. Even using the upper limit, NO emissions from nitrobenzene and dinitrotoluene
synthesis are relatively small but may present local nuisance problems. Since the upper limit
represents specialty batch operations on a small scale, the emissions are probably much higher than
would be encountered in large volume production of these products in either batch or continuous
equipment.
6.2.3.3 Control Techniques
In large batch or continuous nitrations, operations are carried out in closed reactors.
Fumes are conducted from the reactor, air is added, and the mixture enters an absorption tower for
recovery of nitric acid. If too much N02 remains in the residual gas from the absorber, it may be
further reduced by techniques such as wet chemical scrubbing. Details of the control techniques
are discussed in Section 6.1.3.
Noncondensable gas from acid denitration is treated in the same manner as reactor gas. A
common absorber is sometimes employed.
Small batch nitrators used in manufacturing specialties such as drugs and dyes are small-
volume, high-intensity NOX emitters. In one plant, reaction times ranged from 3 to 12 hours, depend-
ing on the product made. From 3 to 850 batches of each product were made each year. Emissions
ranged from 0.7 to 130 kg of N0x per Mg of nitric acid (0.14 to 260 Ib/ton) with a median of 21 kg per
Mg of nitric acid (21 Ib/ton). The median emission was 7 kg per Mg (14 Ib/ton) when one product was
excluded from the calculations. The emissions, which are vented in individual stacks, are brown in
color for a few hours per batch.
Caustic scrubbing and NO incineration are regarded as the most plausible controls for
A
specialty batch nitrations. Catalytic reduction is usually ruled out because of organic and other
impurities in the gas. Neither control is considered highly efficient in this application.
The intermittent character of emissions makes them difficult to control and contributes to
very high pollution abatement costs per ton of nitric acid consumed. According to DuPont, operating
costs for such equipment would render approximately half of the small batch nitrations so unecono-
mical that the manufacture of these products would be terminated (Reference 6-40). Large batches
may be suitable for conversion to continuous operating, but small batches are not.
6-47
-------
o:
o
o
ct:
oo CM
g?
ii tO
oo
OO UJ
II (_>
XLU
o u.
Z UJ
UJ
H- O
00 Z
UJ ti
00
0 Q- E
r- O-T-
.£ 0) C
E S °
b£ CU *r-
0 3S E
Z ^0 IT-
CO
o c
3= O
-^- »p
"Sf en?
+J § z: o
i| ii
+J °
LU
0
o
3 *~*
"§ ^^
° 2: o
Q. ^^
0
cr>
+*
o
H3
o
0
t-
Q.
CO CO
oo r^.
oo cr>
CT> O
CT» O
O O
^o cy>
CM CO
r r
'
O O
0 0
vo o
^ tf%
co r^.
CO LO
CM CM
CU
c.
CU
N
c
CU
X}
o
*->
z.
00 O
O 00
r 0*
LO O
r LO
r r
0 r-
1 *
CM O
^ 0
in o
*N f>
r- in
CO «^-
r~ ZZ-
0)
c
CU
3
r
Q
4_)
o
[ >
>r-
c
Q
6-48
-------
6.2.3.4 Costs
Fume incinerator investments are quoted at $10,000 to $20,000 by one source (Reference 6-43),
Another suggests that investments of $75,000 to $150,000 are necessary for flame abatement facili-
ties for existing small batch nitrators and $75,000 to $250,000 for existing large nitrators.
Annual operating costs were estimated at $25,000 to $85,000 per product for small batch nitrators
and $25,000 to $40,000 for continuous nitrators (Reference 6-40).
6.2.4 Explosives: Manufacture and Use
6.2.4.1 Process Description
An explosive is a material that, under the influence of thermal or mechanical shock, decom-
poses rapidly and spontaneously with the evolution of large amounts of heat and gas. Explosives
fall into two major categories: high (industrial) explosives and low explosives.
Industrial explosives in the United States consist of over 80 percent by weight of ammonium
nitrate and some 10 percent of nitro organic compounds. During 1975, an estimated 1.4 Tg (3.1 x
109 pounds) of industrial explosives were manufactured, which is about 13 percent higher than the
1974 productions (Reference 6-44). High explosives are less sensitive to mechanical or thermal
shock, but explode with great violence when set off by an initiating explosive (Reference 6-45).
Low explosives, such as nitrocellulose, undergo relatively slow autocombustion when set off and
evolve large volumes of gas in a definite and controllable manner.
Production and consumption data for military explosives are classified. Some of the more
important ingredients in military explosives are known, however: trinitrotoluene (TNT), penter-
ythritol tetranitrate (PETN), cyclotrimethylene-tri-nitramine (RDX), and trinitrophenyl methyl -
nitramine (Tetryl). Nitration is an essential step in the manufacture of each of these.
PETN is most commonly used in conjunction with TNT in the form of pentolites, made by incor-
porating PETN into molten TNT. RDX is used in admixture with TNT, or compounded with mineral jelly
to form a useful plastic explosive. Tetryl is most often used as a primer for other less sensitive
explosives.
TNT (symmetrical trinitrotoluene) may be prepared by either a continuous process or a batch,
three-stage nitration process using toluene, nitric acid, and sulfuric acid as raw materials. In
the batch process, a mixture of oleum (fuming sulfuric acid) and nitric acid that has been concen-
trated to a 97 percent solution is used as the nitrating agent. The overall reaction may be
expressed as: ru
Ln
+ 3HON02 + H2SV °2N o N02 + 3H 0 + H SO (6-15)
NO
2
6-49
-------
Spent acid from the nitration vessels is fortified with makeup 60 percent nitric acid before
entering the next nitrator. Fumes from the nitration vessels are collected and removed from the
exhaust by an oxidation-absorption system. Spent acid from the primary nitrator is sent to the acid
recovery system in which the sulfuric and nitric acid are separated. The nitric acid is recovered
as a 60 percent solution, which is used for refortification of spent acid from the second and third
nitrators. Sulfuric acid is concentrated in a drum concentrator by boiling water out of the dilute
acid. The product from the third nitration vessel is sent to the wash house at which point asym-
metrical isomers and incompletely nitrated compounds are removed by washing with a solution of
sodium sulfite and sodium hydrogen sulfite (Sellite). The wash waste (commonly called red water)
from the purification process is discharged directly as a liquid waste stream, is collected and sold,
or is concentrated to a slurry and incinerated in rotary kilns. The purified TNT is solidified,
granulated, and moved to the packing house for shipment or storage. A schematic diagram of TNT pro-
duction by the batch process is shown in Figure 6-16.
Nitrocellulose is prepared by the batch-type "mechanical dipper" process. Cellulose, in the
form of cotton linters, or specially prepared wood pulp, is purified, bleached, dried, and sent to
a reactor (niter pot) containing a mixture of concentrated nitric acid and a dehydrating agent such
as sulfuric acid, phosphoric acid, or magnesium nitrate. The overall reaction may be expressed as:
C6H?02(OH)3 + 3HON02 + H2S04 -» CgH^ONO^ + 3 H20 + H2$04 (6-16)
When nitration is complete, the reaction mixtures are centrifuged to remove most of the spent acid.
The spent acid is fortified and reused or otherwise disposed. The centrifuged nitrocellulose under-
goes a series of water washings and boiling treatments for purification of the final product.
6.2.4.2 Emissions
The major emissions from the manufacture of explosives are nitrogen oxides and nitric acid
mists. Emissions of nitrated organic compounds may also occur from many of the TNT process units.
In the manufacture of TNT, vents from the fume recovery system, and nitric acid concentrators are
the principal sources of emissions. Emissions may also result from the production of Sellite
solution and the incineration of "red water". Many plants now sell the red water to the paper
industry where it is of economic importance.
Principal sources of emissions from nitrocellulose manufacture are from the reactor pots and
centrifuges, spent acid concentrators, and boiling tubs used for purification.
The most important factor affecting emissions from explosives manufacture is the type and
efficiency of the manufacturing process. The efficiency of the acid and fume recovery systems for
6-50
-------
5
=3
LLJ-
O
a
u
oc
h-
co
o
z
T*
o
to
Q
u
^^
*
J^.
^
*
o
X
s? t
S:
CO
o
z
i
LU
CO
=3
o
X
h-
LU
CO
=3
O
X
UJ
CO
o
X
o
z
o
f
I
1
LU
z
LU
o
H-
CO
3
Z
L
-*
*
^
*
h-
O
h-
O
00
CO
LU '
2 ]
U.
_J
o
o
z
o
H a
to ^
M^
-^
CO
LU
U_
_^.
CO
LU
U_
_^i
OC
NITRIC ACII
CONCENTRAl
CO ^
ii
< 5
a<
QC
O
H
oc
z
o
I
b
<
3
>
U
CO
O
CM
ta
Z
_J
= c?
J LU X
co rr
3
X
z
LU O
M
oc
UJ
a
UJ
oc
Q' '
UJ
LJL ^"
E H
a.
O
CM
0
c
<
LJ
SELLITE
PLANT
VAPORATORS
Q.
U OC
LU m
eo o
00 h-
00
a co
o < o
Hoo t
aS02
>< o °"
oc
rs tl
ELECTR
PRECIP
.5
o <
* CMLU
X H- |_
CO c/j
<
X
x *
le
LU
O
QC
O
3
o
^M -J MM^^M
LU
CO
<
5 ^
CO
o
1-
o
X
h-
"""" ^ uj
n
3
ROTARY
KILNS
UJ
u
QC
U.
a.
"DC
LU LU
go
3 *~
CO
t
CJ ^
CJ OC
= H-
SULFUF
CONCEN
H
CO
* X
X
LU
4
O
t co
n
_J
UJ
u.
1-
* LU
CO
1-
0
X
1
co 52
o ^
x <
o ^
CO Z
FURNACE
LO
CD
O
C
CD
Re
01
CO
O)
C
a>
o
o
D.
_c
o
CD
CD
Q)
6-51
-------
TNT manufacture will directly affect the atmospheric emissions. In addition, the degree to which
acids are exposed to the atmosphere during the manufacturing process affects the NO emissions. For
nitrocellulose production, emissions are influenced by the nitrogen content and the desired quality
of the final product. Operating conditions will also affect emissions. Both TNT and nitrocellulose
are produced in batch processes. Consequently, the processes may never reach steady state and emis-
sion concentrations may vary considerably with time. Such fluctuations in emissions will influence
the efficiency of control methods. Table 6-9 presents the emission factors for the manufacture of
explosives and the effects of various control devices upon emissions (Reference 6-45). Although the
manufacture of explosives is a very small source of NOX emissions nationwide, explosions could be an
intense source in confined underground spaces. Precautions should be taken to avoid chronic exposure.
6.2.4.3 Controls
Explosives manufactured by the commercial industry use ammonium nitrate extensively as the
base material. The ammonium nitrate production process is reviewed in Section 6.2.1. Nearly half
the plants use the catalytic reduction technique for control of NO emissions.
The military explosives which are produced in large amounts include nitroglycerin, nitrocellu-
lose, TNT, and RDX. The molecular sieve abatement system is used at Holston Army Ammunition Plant
in Kingsport, Tennessee. Another Army Ammunition Plant at Radford, Virginia, is constructing two
molecular sieve units to treat vent gas streams from their nitrocellulose plant. The description
of the molecular sieve control technique is included in Section 6.1.3.5.
6.2.4.4 Costs
Costs for controlling NOX from explosives manufacture by tail gas treatment process were
covered in Section 6.1.4.
6.2.5 Fertilizer Manufacture
Sulfuric and phosphoric acids are the principal acids used, in the United States, in acidu-
lating phosphate rock. A few manufacturers produce "nitric phosphate" fertilizers by acidulating
phosphate rock with nitric acid to form phosphoric acid and calcium nitrate. In subsequent steps,
ammonia is added with either carbon dioxide or sulfuric or phosphoric acid, and "nitric phosphates"
are formed. Dibasic calcium phosphate and ammonium nitrate are the useful compounds produced
(Reference 6-48).
U.S. Department of Agriculture statistics do not segregate nitric phosphate fertilizers made
by acidulation of phosphoric rock; but private sources indicate that nitric phosphate fertilizer
6-52
-------
TABLE 6-9. EMISSION FACTORS FOR MANUFACTURE
OF EXPLOSIVES (REFERENCE 6-45)%
Type of process
TNT - batch process
Nitration reactors
Fume recovery
Acid recovery
Nitric acid concentrators
Red water incinerator
Uncontrolled0
Wet scrubber
Sellite exhaust
TNT continuous process6
Nitration reactors
Fume recovery
Acid recovery
Red water incinerator
Nitrocellulose6
Nitration reactors
Nitric acid concentrator
Nitrogen oxides3
(N02)
kg/Mg
12.5(3-19)
27.5(0.5-68)
18.5(8-36)
13(0.75-50)
2.5
4(3.35-5)
1.5(0.5-2.25)
3.5(3-4.2)
7(1.85-17)
7(5-9)
Ib/ton
25(6-38)
55(1-136)
37(16-72)
26(1.5-101)
5
8(6.7-10)
3(1-4.5)
7(6.1-8.4)
14(3.7-34)
14(10-18)
For some processes considerable variations in emissions have been reported
The average of the values reported is shown first, with the ranges given
in parentheses. Where only one number is given, only one source test was
available.
Reference 6-46
^
'Use low end of range for modern, efficient units and high end of range for
older, less efficient units.
Apparent reductions in NOX and particulate after control may not be sig-
nificant because these values are based on only one test result.
Reference 6-47
For product with low nitrogen content (12 percent), use high end of range.
For products with higher nitrogen content, use lower end of range.
6-53
-------
made in this manner was estimated at 450 Gg (500,000 tons) in 1967, and nitric acid consumptions
at 135 Gg (150,000 tons) (Reference 6-49).
NO emissions are dependent on the quantity of carbonaceous material in the rock, since NO
* x
is formed as nitric acid oxidizes the carbonaceous matter. The use of calcined rock avoids the
production of NO .
Air pollution abatement by fertilizer manufacturers' efforts has centered on reducing particu-
lates and fluorides emissions, which are severe problems. The water scrubbing used to.reduce these
pollutants would be expected to reduce NO emissions to only a minor degree. Although no measure-
ments of NO emissions are available, brown plumes are said to occur.
X
One company has found that the addition of urea to the acidulation mixture reduces NO emis-
sions and eliminates the brown plume (Reference 6-49). Urea, as discussed in Section 6.1.3.3
reacts with nitric and nitrous acids to form N2-
6.2.6 Metals Pickling
The principal use of nitric acid in metals pickling is in treating stainless steel. Mill
scales on stainless steels are hard and are difficult to remove. Pickling procedures vary; some-
times a 10 percent sulfuric acid bath at 333K (140F) to 344K (160F) is followed by a bath at 328K
(130F) to 339K (150F) with 10 percent nitric acid and 4 percent hydrofluoric acid. The first bath
loosens the scale, and the second removes it. A continuous system for stainless steel strip con-
sists of two tanks containing 15 percent hydrochloric acid, followed by a tank containing 4 percent
hydrofluoric and 10 percent nitric acid at 339K (150F) to 350K (170F). One effective method is the
use of molten salts of sodium hydroxide to which is added some agent such as sodium hydride. This
may be followed by a dilute nitric acid wash (Reference 6-50).
No measurements were found of emission rates from nitric acid pickling of stainless steel.
Treating equipment should be properly hooded and ventilated and the fumes scrubbed to protect
workers. Urea would probably control the NO emissions.
Nitric acid is also used in the chemical milling of copper or iron from metals that are not
chemically attacked by nitric acid, and for bright-dipping copper. In the latter operation, a cold
solution of nitric and sulfuric acid has been customarily used. It has been reported that copper
can be bright-dipped on cold nitric acid alone when urea is added. A highly acceptable finish is
obtained, and NOX fumes are eliminated.
6-54
-------
Sulfuric acid should not be used with the nitric acid-urea mixture since nitrourea, an explo-
sive, can form. Not more than 62 ml of urea per liter should be added, and satisfactory operation
can be obtained with only 15 ml per liter.
In chemical milling, the addition of 46 to 62 ml of urea per liter of 40 percent nitric acid
will reduce NC^ emissions from 8,000 ppm to levels below 10 ppm, provided a bubble disperser is
used (Reference 6-51).
A small, but intense, source of NOX occurs in the manufacture of tungsten filaments for
lightbulbs. Tungsten filaments are wound on molybdenum cores, and after heat-treating, the cores
are dissolved in nitric acid.
Reference 6-43 describes air pollution equipment for reducing the dense N02 fumes given off
periodically when trays of the filaments are dissolved. The fumes pass over a charcoal adsorber
bed, which adsorbs NOX as fumes are generated and desorbs when no fumes are being generated. This
smooths out peaks and valleys'in NOX content in off-gases, which are then heated and combined with
carbon monoxide and hydrogen from a rich combustion flame. The mixture is then passed through a
bed of noble metal catalyst. A colorless gas is released from the equipment.
REFERENCES FOR SECTION 6
6-1 Manney, E.H. and S. Skopp, "Potential Control of Nitrogen Oxide Emissions from Stationary
Sources," Presented at 62nd Annual Meeting of the Air Pollution Control Association,
New York. June 22-26, 1969.
6-2 Freithe, W. and M. W. Packbier, "Nitric Acid: Recent Developments in the Energy and Environ-
mental Area," presented at AICHE Symposium, Denver, Colorado, August 28, 1977.
6-3 Lowenheim, F.A. and M.K. Morgan, ed., Faith, Keyes. and Clark's Industrial Chemicals. 4th
edition. New York, Wiley Interscience Publication, 1975.
6-4 "Strong Nitric Acid Process Features Low Utility Cost," Chemical Engineering, December 8,
1975, p. 98-99.
6-5 Personal communication. Mr. Dave Kirkbe, Davy Powergas, Houston, Texas, November 1977.
6-6 "Environmental Considerations of Selected Energy Conserving Manufacturing Process Options,
Volume XV, Fertilizer Industry Report," EPA-600/7-76-0340, December 1976.
6-7 "Compilation of Air Pollution Emission Factors (Second Edition)," Publication No. AP-42,
Environmental Protection Agency, Research Triangle Park, North Carolina, April 1973.
6-8 Gerstle, R.W. and R.F. Peterson, "Atmospheric Emissions from Nitric Acid Manufacturing
Processes," National Center for Air Pollution Control, Cincinnati, Ohio, PHS Publication
Number 999-AP-27, 1966.
6-9 Mayland, B.J., "Application of the CDL/VITOK Nitrogen Oxide Abatement Process," Presented
to Sulfur and Nitrogen Symposium, Salford, Lancashire, U.K. April 1976.
6-10 Barber, J.C. and N.L. Faucett, "Control of Nitrogen Oxide Emissions from Nitric Acid Plants,"
Third Annual Air Pollution Control Conference, March 1973.
6-11 "NOX Abatement in Nitric Acid and Nitric Phosphate Plants," Nitrogen, No. 93, Jan/Feb 1975.
6-12 "MASAR Process for Recovery of Nitrogen Oxides," Company brochure, MASAR, Inc.
6-55
-------
6-13 Personal communication, Mr. Feaser, Plant manager, Illinois Nitrogen Plant, Marseilles, 111.
November 1977.
6-14 Service, W.J., R.T. Schneider, and D. Ethington, "The Goodpasture Process for Chemical Abate-
ment and Recovery of NO ," Conference on Gaseous Sulfur and Nitrogen Compound Emissions,
Salford, England, Aprilx1976.
6-15 Streight, H.R.L., "Reduction of Oxides of Nitrogen in Vent Gases," Chem. Eng., Vol. 36, 1958.
6-16 Gillespie, G.R., A.A. Boyum, and M.F. Collins, "Nitric Acid: Catalytic Purification of Tail
Gas," Chemical Engineering Progress, Vol. 68, 1972.,
6-17 Decker, L, "Incineration Technique for Controlling Nitrogen Oxides Emissions," Presented at
the 60th Annual Meeting of the Air Pollution Control Association, Cleveland, Ohio, June 1967.
6-18 Andersen, H.C., W.J. Green, and D.R. Steele, "Catalytic Treatment of Nitric Acid Tail Gas,"
Ind. Eng. Chem., 53:199-204, March 1961.
6-19 Anderson, G.C. and W.J. Green, "Method of Purifying Gases Containing Oxygen and Oxides of
Nitrogen (Englehard Industries, Inc., U.S. Patent No. 2, 970, 034), Official Gazette U.S.
Patent Office, 762(5):969, January 31, 1961.
6-20 Newman, D.J. and L.A. Klein, "Apparatus for Exothermic Catalytic Reactions," (Chemical Con-
struction Corp., U.S. Patent No. 3, 443, 910), Official Gazette U.S. Patent Office, 862
(2):514, May 1969.
6-21 Andersen, H.C., P.L. Romeo, and W.J. Green, "New Family of Catalysts for Nitric Acid Tail
Gases," Nitrogen 50:33-36, November-December 1967.
6-22 Rosenberg, H.S., "Molecular Sieve NO Control Process in Nitric Acid Plants," Environmental
Protection Technology Series, EPA-600/2-76-015, January 1976.
6-23 Chehaske, J.I. and J.S. Greenberg, "Molecular Sieve Tests for Control of NO Emissions from
a Nitric Acid Plant," Volume 1, EPA-600/2-76-048a, March 1971. x
6-24 Rosenburg, H.S., "Molecular Sieve NO Control Process in Nitric Acid Plants," EPA-600/2-76-
015, January 1976. x
6-25 Mayland, B.J. , The CDL/VITOK Nitrogen Oxides Abatement Process," Chenoweth Development
Laboratory, Louisville, Ky.
6-26 Personal communication, Mr. Don Ethington, Goodpasture, Inc., Brownfield, Texas, November 1977,
and February, 1978, and D. F. Carey, EPA-IERL, February 1978.
6-27 "New Unit for Nitric Plants Knocks Out NOX," Chemical Week, July 28, 1976.
6-28 "Ammonium Nitrate," Hydrocarbon Process. 46:149, November 1967.
6-29 Miles, F.D., Nitric Acid - Manufacture and Uses, London, Oxford University Press, 1961.
6-30 Private communication with Esso Research and Engineering Co.
6-31 Durocher, D.F., P.O. Spawn, and R.C. Galkiewicz, "Screening Study to Determine Need for
Standards of Performance for New Adipic Acid Plants," draft report GCA-TR-76-16-G GCA Cor-
poration, Bedford, Massachusetts, June 1976.
6-32 Goldbeck, M., Jr., and F.C. Johnson, "Process for Separating Adipic Acid Precursors," (E.I.
DuPont de Nemours and Co., U.S. Patent No. 2, 703, 331). Official Gazette U.S. Patent
Office. 692(1):110, March 1, 1955.
6-33 Burrows, L.A., R.M. Cavanaugh, and W.M. Nagle, "Oxidation Process for Preparations of
Terephthalic Acid," (E.I. DuPont de Nemours and Co., U.S. Patent No. 2, 636, 99). Official
Gazette U.S. Patent Office. 669(4):1209, April 28, 1953.
6-56
-------
6-34 Durocher, D.F. et^ ^1_., "Screening Study to Determine Needs for Standards of Performance for
New Sources of Dimethyl Terephthalate and Terephthalic Acid Manufacturing" Draft Final Report,
GCA-TR-76-17-G. Submitted to EPA/OAQPS by GCA Corp., Bedford, Massachusetts, June 1976.
6-35 Lindsay, A.F., "Nitric Acid Oxidation Design in the Manufacture of Adipic Acid from Cyclohex-
anol and Cyclohexanone," Special Suppl. to Chem. Eng. Sci. 3:78-93, 1954.
6-36 Compilation of Air Pollutant Emission Factors, Environmental Protection Agency, AP-42,
February 1972.
6-37 "Process for Oxidation of Cyclohexane and for the Production of Adipic Acid (British Patent
No. 956, 779) and Production of Adipic Acid," (British Patent No. 956, 780). Great Britain
Office. J. No. 3918:814, March 19, 1964.
6-38 Oil, Paint, and Drug Reptr. 195(6):l-48, April 21, 1969.
6-39 Crater, W. deC. Nitration. In: Kirk-Othmer Encyclopedia of Chemical Technology, Standen,
A. (ed.). Vol. 9. New York, Interscience Publishers, 1952.
6-40 Private communication with E.I. DuPont de Nemours and Co., March 1969.
6-41 Urbanski, T., "Chemistry and Technology of Explosives," Jeczalikowa, I. and S. Laverton,
(Trs.). Vol. I. New York, MacMillan Co., 1964.
6-42 Processes Research, Inc., "Air Pollution from Nitration Processes," Cincinnati, Ohio.
APTD-1071, 1972.
6-43 Decker, L., "Incineration Technique for Controlling Nitrogen Oxides Emissions," Presented
at the 60th Annual Meeting of the Air Pollution Control Association, Cleveland. June 11-16,
1967.
6-44 Nelson, T.P., and Pyle, R.E., "Screening Study to Determine the Need for New Source Perfor-
mance Standards in the Explosives Manufacturing Industry," Draft Report, Radian Corporation,
Austin, Texas, July 1976.
6-45 EPA, Compilation of Air Pollutant Emission Factors, AP-42, Supplement No. 5, December 1975.
6-46 Air Pollution Engineering Source Sampling Surveys, Radford Army Ammunition Plant. U.S.
Army Environment Hygiene Agency, Edgewood Arsenal, Md.
6-47 Air Pollution Engineering Source Sampling Surveys, Volunteer Army Ammunition Plant and Joliet
Army Ammunition Plant. U.S. Army Environmental Hygiene Agency, Edgewood Arsenal, Md.
6-48 McVickar, M.H. et aj_., "Fertilizer Technology and Usage," Madison, Wisconsin, Soil Science
Society of America, 1963.
6-49 Consumption of Commercial Fertilizers, U.S. Dept. of Agriculture. Statistical Reporting
Service.
6-50 McGannon, H.E., The Making, Shaping and Treating of Steel, 8th ed. Pittsburgh, United States
Steel Co., 1964.
6-51 Kerns, B.A., "Chemical Suppression of Nitrogen Oxides," Ind. Eng. Chem. Process Design Develop.
Vol. 4, pp. 263-265, 1965.
6-57
-------
-------
APPENDIX A
SELECTED TABLES IN ENGLISH UNITS
This appendix contains the English engineering unit version of most of the large tables pre-
sented 1n the text. The tables are arranged sequentially 1n the order 1n which they appear in the
text and have the same table number except for the prefix "A".
A-l
-------
o
CO
I
««
r*»
cn
cc.
o
e
o
to
Q£
O
o
to
to
O
CO
to
CO
c
o
r- CM
«/) O
CO Z
E co
LU c
xS
o
;
H"
c
I
"3
cr
LU
in
CM
CO
r-
VO
o
VO
o
li nous
§
CQ
8
J
<
r
a-*
Q) i
4-> 0
£ '
LU J
CM
r- i
t- I
LL, ;
CO
O
O
VO
O
CM
5
CU
c
o>
Hi
1
f
1
f
L
(
?
H
1
I
1
1
(
f"
fl
\>
Q
V
3
J
5
t
CO
O
o
0
r*.
LO
CO
o
co
CM
LO
CU
CO
O
O
y>
^
L.
t
D
7>
3
cn
o
f*.
LO
o
LO
VO
00
o
3
CO
**
0
co
O
o
CO
o
LO
LO
VO
00
i
CO
3
vo
o
LO
0
CM
CM
cn
CM
CM
unous
3
CO
E
c
o
CO
>
i-
o
o
o
LO
o
o
o
00
oo
0
cn
r-
cu
cn
-1
CO
CO
o
LO
CM
,
VO
0
00
LO
I/I
3
o
c
'i
3
CO
E
o
o
CO
4J
CU
*
re
o
(
r
f
1
o
o
o
00
CO
o
CM
VO
5
c
cn
-1
0
CU
CO
o
o.
Q.
c
c
IO *f
p t.
C t-
O LL.
Isl
r- i
U i-
O 1C
C 3
CM
o
o
LO
o
CO
0
co
co
1
CO
o
0
1
CM
o
LO
0
o
CM
cn
to
3
CO
CU
or
CM
CO
o
0
o
CO
CM
00
co
1
CO
vo
o
LO
o
CM
CM
cn
CM
CM
CO
3
0
c
'i
CO
E
o
4-J
4-i
o
CO
>
L.
o
Id
o
o
LO
0
o
o
§8
0
cn
^
CU
4->
c
cn
-1
oo
co
co
0
LO
CM
r-
CO
CM
o
LO
inous
I
CO
^
o
CO
^J
CU
3
re
o
o
cr
c
u.
^~
re
~^
4J
C
o
LL.
vo
O
0
0
CO
CO
o
CM
CO
^
CU
c
_J
CM
0
O
LO
O
00
o
co
co
CU
4->
re
CO
Q
0
en
CM
0
LO
o
o
CM
cn
re
3
CO
CU
co
co
0
o
o
CM
CM
LO
cn
1
CO
O
0
O
LO
O
VO
00
cr.
CM
CU
re
5
CM
o
LO
o
CO
00
co
inous
E
CQ
E
o
o
CO
^
^
o
re
o
cn
c
^
u-
p_
re
o
t
CU
^
0
o
o
LO
o
cn
cn
CM
CU
c
cn
^
CO
VO
VO
o
o
,_
CM
VO
0
CM
0
CO
3
0
1
p
CQ
o
4->
4-J
o
CQ
CU
3C
c
VO
o
0
0
CO
o
vo
CM
CU
c
cn
^
CU
c
o
>^
w
o
0
0
LO
o
CM
cn
CM
CU
CO
Q
o
CM
o
o
LO
0
CO
in
LO
to
re
3
o
CO
CU
ct:
CO
co
0
o
^
o
CO
cn
co
t
re
o
i_
CU
re
e
Q.
to
<
1-
cu c
4-> 0
u
S-i
1 4
o J
r C
CU 4
1 f
u- a
VO
o
o
^
o
o
VO
VO
LO
1
o
o
^
CU
CU
L.
CU
c
=
J
J
a
V
i u
- CU
u ^
-> 0
0 4->
c to
A-2
-------
oc.
o
C£
O.
o
o
o
a.
o
o
o
CO
oo
<:
~
Ol
u
t.
Ol
0.
1
Q
u
3
CO
CM
O
C
o
1 *
ex
c
"O
u
u_
^ ,^
4->
Ol
OJ
0.
1
3
"o
O
ID
«p
O
C
0
(/>
LU
X
o
p-
£
ID
O
0
I/)
d
,_
a
o
ITJ
O
u
*
o
to
3
tt>
2;
r-
O>
C
r
u.
o co vo *-«. * >
^- CM CM r>» en P
CM in oo CM CM en
oo f- o o* t oo
in o o *
VO CO ^f r
** f CM O * *-*
^J- 0
^- p p P «f CM
CM vo in co CM en
oo en co co p^ oo
p t>^ CO O ^ r
^- p P P
co" ^ ^ ^->
CO CM CM CM
in in II
CM in in en
CO CM CM in
-co co
^^^ ^^ ^^^
p CO t
^«» en co
00 ^" OO
vo in r^
oo en co
f**h 00 VO CO * * ^'^
. . . r^ o
en in co CM
CM CM CM P CVJ P
co r^ p- co ^"
i-». in o en in in
vo ^ vo vo P o
,J r-^ ^ odd
f-^ ^-* * . ^~»
. . o «»
CO CM CM «9-
co o vo en co ^^
co i vo in in
CO CO CO VD O
-^ d d odd
~" £T P ^
vo
CM CO CO CM
CM en en CM | |
O O O CM
CM CO CO O
do d d
r^" "co1
co in P
p co in
CO CO CO
-co *r
do d
P- P-P-
f- *»1o 1
4-> C X T3 0> (D />
C O 1 0) C O t-
0> N 4-* M O *- *
O *- C OP P 4-> Mi
e. up 0.1 o *- o
m o £- ex» o> *J
8
o>
in
*^
Lf>
^~
*»-^«.
8
5
Cft
8
CM
^f
CO
CM
8
5
c^
CM
CO
o"
o
r~
CO
CM
vo
in
8
p
o
CO
8
cvT
CO
d
^-*.
8
»
VD
en
d
^^
5
o
A-3
-------
13
!->
CO
00
o:
o
CO
13
O
CO
ID
r"
01 0 ^
(j Ol
S '^? ci>
*"* ^ t_j
J_? £r c
& - ^
*?!
Is
S- ,
0
LO
CM
Ol
«r< ^^" ^
^^ f* E
Sj en 01
^~ S
O)
«[
Ol
ce:
CM
o»
1 CM O
oo r-- E
oo en 01
S i S-
01
Ol
CO*
Ol CM
^ CD
Ol r- O
4-> r^ c
4-i O) a\
fOr- i.
CO Ol
01
'
u i
s-
3 -O
O C
OO 03
.o
en
o» r~
»-> E
tO O
>»,*-
00 M-
4J 4J
«3
o
o to
3
4-> C7)
3 3
CL
-------
en
c
O
CO
O
CO
00
z
O
00
00
X
O
oc:
CM
t CM O)
oo r^ o
SL r flj
CU
01
00
O> CM
Battell
1971
(Reference
Lu
LO CT» LO
LO CM O
W' i-^ r-^
r^* i^^ co
O"» VD 1*^
CO VO (£>
*± o o
O CM VO
00 00 CTi
CO O O
O O 0
O r VO
P**» OO C7^
^- o o
00 00 ^~
00 VO r-
r** ^» ^~
CO O i-^
^i- co ^d-
CM 1^ r
CM O r
*!
<1J "~"
M E
tO O
>»,*-
00 <4-
Q T3
to to
C C
O «r-
to
to
c en
O i
3
4-> O>
3 3
O O
A-5
-------
o
CO
or
oo
o
or
o
on
o
CO
o
IO
CM
in
§
f- CM
«n O
tn z
g
UJ C
O
. X 4J
i »
O
4->
0
3
CO
o J2»
^ o^
£ ° "CM
5*8
3
"i « "
= o
&
^^
»-
r_
u.
*ttj
u.
CU
o.
l2?
0
c
=
S.
1^^
^"
A t
*^
C
5
f-
j*
»
UJ
o
o
0
a-
vo
o
CM
CO
^
|
o
o
o
CU
4->
u
t
UJ
.£
^
CU
p*
u.
vo
o
o
r-.
in
o
vo
in
CM
(O
P_
10
CU
C£.
0
0
f"
i-
U.
_^
«c
4-
8
C
1C
1-
tn i
k. X
CU -v
r- 4-
o a
CO
1C
CU C
3 ""
4-> >
t.
CU C
4-> C
IO -
in
0
o
3
o
£
cn
CO
^_
i.
3
Z
in
3
>
i
j
>
i
>
c
>
>
§
o
o
CO
CM
O
en
en
in
in
in
CU
u
B
D_
O
O
0
in
O
O
CM
|
§
O
CO
>>
o
(O
o
in
o
o
o
0
in
CM
,_
CO
CO
1
I
o
CO
CU
10
0
o
A
a
V
i
c
c
>
p
p«"
1C
4.
ts
P"
L
C
2
0
CO
in
o
vo
*
*o
3
TJ
CU
OL
O
.
"* C7>
C
f-
t.
ul
*«
o
0
o
f_
co5
o
vo
CM
VO
r.v
i.
3
IO
z
IT
3
0
o
o
co
CM
0
s
co
CM
in
CU
u
E
0.
2
o
o
n
0
o
CM
1
1
O
CO
>>
o
1o*
o
in
o
0
in
CM
,_
CO
CO
1
1
4->
o
CO
4->
CU
"io
O
O
c
f-
s.
f-
u"
1
r
It
i
C
U.
2
0
CO
in
o
VO
*
>
0
1o
o
o
CTi
C
£
^_
U-
IO
(J
i_
CU
>
in
O
0
o
vo
VO
,_
CO
CO
VO
1
1
0
CO
CU
3
IO
o
o
a
c
c
i
c
<_
o
o
o
CO
in
o
CM
in
co
IO
3
a
^-
CU
or
_
0
j
i
>
>>
>
8
o
0
in
o
VO
CO
in
cu
U
CU
i-
UJ
1
o
'cu
u!
CO
VO
0
o
en
CM
o
£
CM
cn
CM
^3
3
^
in
cu
or
a
c
t-
t.
r-
U.
p«
p
C
3
J.
j;
10 "V
CU 4^
cc
o *
XI C
cu "~
O M
3
U C
cu -
4-> 1
10 C
X -
co
en
o
0
0
co
CM
o
^
VO
o
00
p
IO
3
IO
z
I/
3
)
o
o
o
o
CO
CM
O
-
r-.
co
tn
10
cu
u
O
QL.
o
0
r^.
^
O
0
co-
co
1
1o
o
o
J_
cu
(O
0)
a.
r
^3 *"
0)
4-> C
u
cu
1- -
UJ
V
"
cu
u. :
en
0
0
r^
^
o
CO
CM
in
co
r
"io
O
<_>
TJ
CU
,cu
VK
t.
-S
c
=3
n
2
r~
o
33
CU
O
3 I/
H *-
L a
IF* ^1
M C
ro 4-
X I/
in
o
o
in
CM
vo
o
VD
cn
o
CM
1
ia
o
o
o
cu
cu
*4-
i_
cu
o
1
j
f
)
t
>
CM
O
o
r^
5
o
in
o
1
IO
O
a
cu
<*Z
P-
tn
i IO
IO i
i- 0
cu
C 4->
A-6
-------
C£.
CO
ID
Q
00
zr
o
VI
c
o
1- CM
to O
"E to
LU C
O
X +>
o
o
re
o
3
4->
i^r
^ ^ * L-
U) , . ^
r)_ O \-
XI
3
*->
t OJ CD
S£ fO cvJ
t/> '
^ => 0
CU
Q.
>)
1
f
CU
3
U_
cu
3
CO
Q.
1
Cn
C
^
s-
f
U-
co
OL
^^
1
4J
C
CU
£
CL
i
3
CT
UJ
O
O
O
LO
O
O
**
CM
^-
1
re
o
CU
<**»
3
t
CU
4-)
re
*^
o j
CU
cnn
rO
^ <
0
re
a. c
CM
5
0
LO
O
CO
CD
^~
^
OJ
TO
r
^~
r
4_J
tA
O
>r»
o
CU
.a
3
1
U
Q -M
3 .C
- O
r-
p re
c s-
OJ -t->
Q CO
s
o
CM
0
*
00
00
p^
1
re
3
-O
r
00
CU
a:
D
c
r
i.
i
LL.
r
r
(t
*
<^~*.
O
4_>
(D
>r
O
I/}
^/"}
o
CM
CT.
CM
O
o
CO
CM
O
CO
r--
LO
OO
LO
CM
t
re
t-
3
re
z:
to
re
CD
1
LO
O
o
o
OO
CM
O
OO
=!
CM
CO
r
00
00
CU
o
£
D-
CD
O
O
r-
3-
0
ai
OO
CD
oo
1
re
0
CJ
c_
CU
-a
re
CU
S-
CL
CO
CU
n
3
i-
CU
4_>
re
*^
T3
CU
cn
re
\s _
^
re
CL. (
00
o
r^
'S'
o
o
CD
r-^
CD
CO
1
re
0
T3
CU
CU
4-
i-
O)
^3
c
=>
S-
cu
^
o
4->
S)
CO
CM
o
o
CM
CD
o
LO
*t
o
cn
1
re
0
o
QJ
CU
*4
t-
a»
>
o
OJ
o
o
£
"3-
o
OO
cn
LO
1
re
o
CJ
-a
CU
"£
t
OO
« 00
r~ re
re i
i- CJ
CU
c *->
CU O
ro
O
o
LO
o
CO
CD
^~
r-
CU
4_>
rc:
)
r^^
1
4_>
f)
0
r-
c
( i
3
CU
S-
r
t j
"U
CU
cn
re
*^
o
re
O-
00
IT)
0
CTl
CM
«J-
C;
CO
LO
OO
r^^-
f
re
n
T3
.,«
QJ
C
r
S-
i2
i
re
3
C
CJ
4->
O
CJ
JO
tNj
O
C
0
OO
CM
O
o
CD
CM
O
CO
t
re
3
re
2:
00
re
n
Ov.
0
o
o
0
00
CM
o
CO
cn
CO
f
00
00
QJ
(^j
0
S-
Q-
o
o
o
-
1
o
LO
^3-
CD
LO
CU
re
r
.
r-
4.J
o-i
O
t~
O
QJ
-Q
3
+J
CU
1_
r
U.
^J
CU
D
re
\x
(J
re
CL
CM
O
o
en
CM
^
0
CM
OO
O
CM
,
re
3
TJ
r"
00
QJ
cr:
CI
c
.r_
t.
*r
u.
r
I
re
3C
X
> o
JD
CU
S-
tf
u.
o
00
0
o
o
OO
CO
o
0,
CM
on
c >
CD
r-
re
S-
3
re
2:
00
re
i
CM
D
O
O
00
CM
O
Cn
00
r
tn
to
QJ
O
O
§
O
c;
£
"3-
O
CTl
CD
r~-
1
re
o
o
i-
cu
a
re
CU
i.
CL
CU
.a
3
CU
I- _
1 1 .
c
XJ
CU
CD
re j
.*:
o
re
CL. L
...
a;
"
_
^
^t
0
o
CO
.3-
00
i
re
o
T3
CU
CU
>_
CU
"O
c
__
11
^
D
_)
O
X
D
21
D
L_
'5
C--
_
LO
CM
O
00
,
r
I
re
o
Q
QJ
QJ
M-
i_
QJ
>
0
CV.I
O
o
o
r--
LO
o
<~
Cv'
CO
CM
cu
re
/
i>~»
r
4-^
t/1
Q
r
C
CU
3
n;
^
r
1 i
^
CU
CD
re
\s
CJ
re
D_
00
'O
O
O
cn
CM
o
cn
r--
O.I
LO
r
r
re
3
-o
.^-
1/1
Q)
ct:
c
c
t
i.
*I
u.
f
f
re
3:
h-
^
~r
r*j
*
0
o
oo
CM
0
CM
DO
^j
»
:b
t-
cu
*:
o
M
/O
X
r:
CM
O
O
o
LO
CM
VO
0
uo
CO
LO
1
re
o
CJ
-a
QJ
QJ
i_
QJ
O
A-7
-------
OO
oo
O
CQ
cc:
»
oo
o
o
II
Q.
2:
ZD
oo
z
o
o
z
<
CM
<
OO
cu
(J
3
O
oo
s^
r CT>
""
o j^*
0 CU °°
00 £ £
t^ CW
flj ^ /I %
^^ 1^*
e ^^
3 r~"
O
in
i
CM
CO CU
§2 g
*~~ S-
CU
1
CM
l-fM g
OO 1*- ^
;^* O> QJ
^ *" s-
OJ
It
cu
o:
oo
CU CM
^ i CU
CU I"*- O
4-) cy> c
+J r- CU
CO CU
'[
cu
cu
u_
O VO *3-
O1 VO CM -Q
co vo
CM
^J- CM O O
r-» oo CM «sa-
co vo in r**
i CM 00 i
CO CTt r- CM
CM 00 1^ OO
CO CO VO
o o o
f^- O O .
co r^ CM 1
r- i in
oo o^ CM in
»o oo oo o
r*» in i oo
i in «*
oo r^ en CM
o oo CM co
oo in r^-. ^>
CM co in
00
cu
i 0
fO i 00 O CO
O f . o
CT>
r 4J
"O CU
o
CU
CO O
>> S-
00 M-
00 00
C. C
o -
i-
00 C
r- O
E O
LU
ft
i VO
ai r^
C CTi
O i
(O VO
O 00
CD
3
O
CU
Q.
00
00
rtJ
en
oo
oo
CU
O
O
O.
3 C
O O
CM
"o
X
in
CO
.0
A-8
-------
00
a:
o
CO
a:
I
oo
Q
00
2:
o
00
oo
X
o
CO
I
CM
<£.
LU
-J
CO
cu
S r S-
cu
C£
cc?
i
CU CM
i CU
cu r^ o
M CTv C
!->, CU
rO S-
CQ CU
cu
C£
3
U.
(J
IO 1*^ 00
CM *3- 00
O O O
CM CT> CO
cr> in o
in -si- co
o o o
p CO O
I"** <53" CO
r^. co oo
000
000
CT» CO CM
CD CD CD
in oo in
r OS CTt
in o O CT5
Co
-
in
CO
|
CM
O
CO
o
CM
CT>
in
r
CM
r
O
CM
O
2:
4J
C
CU
r
3
cr
cu
to
ro
o
(U
J-.
O
CO-
OT
CU
i-
fO
to
c
o
^
to
to
r-
E
0>
X
o
z:
,
r
(O
c
0
4->
C
cu
>
c
o
o
>>
OQ
e
cD
00
>,
oo
a
to
c
o
r
to
to
'i
L«J
"(O
C
0
r-
4->
^
&
M
o
4->
3
Q.
4->
3
O
un
p^
CTt
r--
T3
C
to
*&
O^
,__
E
O
'M-
03
^_>
ro
o
to
f
r
O
4_)
C
O
o
*
r>-
CTt
A
<£>
'
4->
3
O^
3
<
C
O
«/)
(O
CT>
CO
U)
0)
0
J-
D.
0
4->
CU
3
o
to
c
o
4->
U9
O
r~
X
vo
o^
o
o*
to
cu
a
3
O
C
H 1
A-9
-------
O
o
OL
O
o
00
C
o
i- CNJ
to O
in z
"i to
LU C
O
p
o
ro
O
t~
3
CD
|go
£ " Tg
UJ Lu Z
QJ
O^ 13
CO 4J
J^CG
= N
_,
! I i i
1
o
o
LO
0
LO
" VO
.D
r ~ i
QJ O | ID
QJ
Q.
i^T*
o;
!
QJ
ro
: f
1 T-
J-J
' . -
iZ j ' .
:
1 " '
0
o
o
t -»
C*J
C?
o
0
o
j
\
1
0
ro
0
LO
VD
VO
r~-
li-/
o
ro
CNJ
O
trr
LO
o
LO
LO I UD
LO j r
VO 1 LO
i
j , ,
! ' -
' i -S .'O
fO ' f '
Z3 ' ' i
-f! i j-
^ ' i v"(
OJ ! .,-
cc: j o
QJ >
3
r-
o
o
ro
O
LO
vo
LO
p..
ro
3
I/)
QJ
or
1 ,,, r-
j ro 1 »-
O i O i O
j ';', J S
i fl
o
U)
r-.
o
o
o
CNI
O
f
LO
LO
VO
|
3
QJ *
Q.
*~ $ CT>
CD
cn f! cic
j- ; .,_ .,_
£ : s-
.^ J ij^
S-
HZ
i-s- j! ?
i! p_ ; ,.,.
tj ro :, _ra
,
1
H
r [
QJ 'i JT1
Q. '1 C>
1 " C i
t (-1 I
C j
§ b -0
B. E 3
"3 5 QJ
cr t! *-
LU i *-
n
JO
=3
^
f
U.
Q "O -1 "O
\ rv ( QJ x
| cn i cn o
ro JO
1 -§ u 2
J ro i re T-
B Q. JO. U.
o
ro
O
O
S
O
CNJ
LO
r~
1
ro
O
ro
o
C\J
o
o
r--
UO
0
ro
ro
CO
LO
QJ
ro
r
r-
to
Q
LO
LO
LO
O
O
5
o
ro
00
LO
CM
p_
ro
3
TD
to
QJ
CC.
0
O
CO
ro
o
o
0
CO
CM
o
r
,^
CM
CO
1
to
to
QJ
'£
o
CT
QJ
ro
VJ
P_
P»
**
i-
QJ
6
00
«
X
o
JO
QJ
^
1 1
QJ
JO
3
£
*f
U.
o
QJ
cn
ro
O
ro
n_
cn
r-
*-
ul
r~
ro
3;
O
0
O
r-
0
00
CM
CO
1
ro
O
O
to
QJ
il
O
cn
QJ
ro
,_
P
^
I
t
-*~
QJ
3
?
tp^
tx.
T3
QJ
cn
o
ro
Q_
i-
QJ
O
4-3
OO
*«
h-
CX
3^
QJ
JZ
3
QJ
f
U.
T3
QJ
CO
ro O
Q. Z
CO
o
CNJ
O
o
LO
0
co
co
00
LO
CNJ
QJ
ro
P_
P
r
to
0
LO
LO
LO
O
0
0
CO
o
co
CO
00
LO
CM
r
ro
3
o
to
QJ
tx.
O
0
LO
CNJ
0
O
o
CM
0
CO
vo
CTi
O
1
to
ro
cn
C
p
*-
u_
pZ
ro
*
Y>
i.
QJ
i
p-
O
CO
C
e
I^H
to
ro
O
TJ
QJ
cn
ro
U
ro
Q.
CNJ
CNJ
0
O
O
LO
O
O
00
CM
QJ
(0
f
r
«^-
to
Q
0
O
O
O
CO
O
00
CM
^>
CO
P-*
ro
3
T3
to
QJ
a:
0
co
CNJ
8
0
CO
o
o
CTi
vo
co
^^
1
to
QJ
JO
3
4_>
t.
QJ
ro
"^
o
QJ
cn
ro
U
ro
D-
CO
8
0
LO
o
LO
00
VO
r
QJ
ro
P^
P^
r-
to
O
CTi
O
0
O
0
CO
o
g
CNJ
CM
r
ro
3
o
10
QJ
C£.
o
§
O
O
CO
o
o
CM
CO
f«*
1
S
cn
C
r
i-
u_
pZ
ro
3
X
0
JO
QJ
^
P-
u_
0)
JO
3
4»>
S-
QJ
ro
"f
T3
QJ
cn
ro
O
rO
Q.
CM
CM
0
O
o
LO
o
o
co
CNJ
QJ
*J
ro
P
P
»-
in
0
r*.
8
o
o
CO
0
CO
CNJ
«3-
CO
p
fO
3
T3
to
QJ
a:
O
co
CM
8
o
CO
o
o
VO
CO
^^
1
in
ro
cn
C
r
'-
U.
rZ
ro
*
,_r
ro
t-
QJ
C
QJ
O
QJ
JO
3
4_>
I- X3
QJ QJ
ro *^
"^ -p
to
"O m
OJ ro
7>r*~
(O O
0 »-»
-------
00
UJ
ae
00
o
o
CM
10
§
t- CM
(/) J^
E to
O
X4->
o
Zw
O
IO
3
CO
O fc-w
t- O O
in o ^.
^» (O CNJ
E LL. O
UJ Z
J3
3
"m Oi00
3 ^? c^
LL. J£? «*
' o
1
h-
s
U-
Ol
U.
Oi
O.
^.
f^
C
il
Ol
Q.
>^
}»
+J
C
^
CT
UJ
CO
f^^
O
o
CO
CM
*
o
5
f^
o
CM
10
£
in
O
0
fc.
Oi
c
J^
3
CO
O)
a
c
00
J°
01
IO
Oi
3C
fc.
Oi
4J
IO
!^
.J2
Q
O
fc
f
^
3
CO
Ol
"o
c
oo
to
Ol
u
IO
E
3
LL.
*-
5
o
r^
o
CM
O
CM
CO
o
o
8
Oi
CM
^^
CM
1
in
S
t^*
CM
O
O
co
CM
r
O
£
0>
Oi
^_
Ol
IO
!^I
to
Q
CD
fc.
Ol
C
L.
3
CO
01
c
oo
t
01
10
O)
m
in
a
"aij
a.
SI
t-
o
_
t
IO
fc.
o
o
\L
r^
^^
CM
0
0
CM
CO
o
o
o
LO
p^
VD
1
to
3
j~
Oi
O
O
00
CM
r
O
VO
ao
CM
O>
£
.£
O
o
t.
Ol
^
3
CO
Oi
"cr
c
00
Ol
3
p_
LL.
.c
^
^^
fc.
Ol
^,1
IO
Ol
z
E
§
ae
^N«
CO
CM
O
0
CM
ao
o
o
VO
o
o
o
1
to
3
00
8
o
CM
CO
o
o
5
o
Oi
Ol
IO
£
.£
Q
o
t,
Oi
fc.
3
CO
Ol
C
oo
s
LZ
^J
3
O
JZ
.^
**g
t.
Ol
IO
Oi
z
E
I
o
o.
oo
o
00
Z
o
oo
o
>-
cc
i
o
^^
c
O 3
4-> CO
a.
to o
o *~
o
3
U.
*«
^
r CM
00
O Z
to
c c
to
tO VC
t- O
E
UJ
X
o
Z
IO
o
IO
o
o
in
3
t^
o
IO
+J
o
*
00
i-^ CO
^- LO
^ CO
VD CO
Oi LO
to oo
f
^ '
vo
CM
oT ^
o> o
CM r-»
co to
VO O
CM CO
CM LO
LO LO
LO «3-
"r cc"
CO CM
<«3- CM
CO CO
f Oi
LO W
V£) CO
r%. vo
CM Oi
CO r
O C3
O
O
co
LO
0
0
co CM"
O Oi
LO *J-
^-^ +^^
r LO
*^ ^**
CM i
CM CM
0 CD
i-^ CO
r*» CM
CO i
^j- CM
0 0
LO CM
o' d
r IO
IO f
f- »->
i 1
E "^
E \n
O O)
(_> CC
§;
^^*
^^
"~
r*.
U3
CM
CO
vo
LO
^
Oi
LO
vo
vo
CO
CM
0
CM
T~
co
^3-
LO
0
O
VO
r-~
<0-
^"
0
tj-
VO
0
f^
0
*IO
^A
o
A-n
-------
DC
O
§
CJ
o>
3
O
CO
00
CM
LU £^
1 1 ^^ 9
«
oe
0?
CM
0)
or
3 *~
0
"I
oo"
CM
CO '
0)
1
» CM CM
0 CM
=
VO l*»
in vo
< >"«
o m
r*. r*«
vo vo
en vo
in r-»
O VD
vo f*^
CM CM
00 CO
r en
< vo
CM in
in CM
i m
*r vo
o u
A-12
-------
TABLE A2-13. ANNUAL NOX EMISSIONS FROM COMMERCIAL BOILERS (10s tons)
Fuel
Coal
Oil
Gas
Total
Source
Battell e
1971
(Reference 2-18)
0.131
0.148
0.025
0.304
MSST
1972
(Reference 2-4)
0.029
0.212
0.120
0.361
GCA
1973
(Reference 2-15)
0.030
0.63
0.110
0.77
Current
1974
0.0543
0.5044
0.2241
0. 7828
TABLE A2-14. ANNUAL NOX EMISSIONS FROM RESIDENTIAL SPACE HEATING (106 tons)
Fuel
Coal
Oil
Gas
Total
Battelle
1971
(Reference 2-18)
-
0.1682
0.1785
0.3467
Source
MSST
1972
(Reference 2-4)
-
0.254
0.212
0.466
GCA
1973
(Reference 2-15)
0.012
0.098
0.210
0.320
Current
1974
-
0.2021
0.2175
0.4196
A-13
-------
D;
o
cu
D_
I-H
ID
cr
ui
>-
UJ
CO
cr.
Q
Z
«=C
CO O3
a: z
o UJ
o z
«=c o
Z CO
O Z3
-i CQ
co s:
co o
t-H O
CO C£
Z Ul
O h-
t-H Z
CO HH
CO
LO
I
CVJ
CO
co
c
0
co
CO CO
r- C
E 0
1 1 i 4-J
Xvo
o o
CvJ
C\J
C\J
Z r
O
1
CO
$- 3 '
O -M
-(-> CQ
0
fO «>
u. o
c ->*
O CsJ
r- O
CO Z
§
^:
co
I" 1
O)
cn 3
co +->
CO CQ
CVl
D O
U.
3
U.
(U
O.
>»
r
cn
c
r-
*
*r"
^~
(U
CL
H-
C
(U
E
O.
3
cr
Ul
CO
^
o
o
CO
ra
CD
c
o
'^J
r
C
*r~
NX
i-
03
Q.
CO
CO
CD
VO
^
UD
CO
O
0.
JC
0
o
LO
A
(U
co
(U
r-
O
cn
^j.
CsJ
CD
5
CO
o
CO
cr>
CO
n
o
^
O
CD
&
C\j
O
r-I
LO
3
O
o.
.C
o
0
LO
V
r_
OJ
CO
0)
r-
Q
c
*->
o
O CO
t-
-------
TABLE A2-16. ANNUAL FUEL CONSUMPTION BY INTERNAL
COMBUSTION ENGINES 1012 (Btu)
Fuel
on
Gas
Source
Shell
1971
(Reference 2-29)
477
1497
MSST
1972
(Reference 2-4)
519
1627
Current
1974
540
1617
TABLE A2-17. ANNUAL NOX EMISSIONS FROM INTERNAL COMBUSTION ENGINES (106 tons)
Equipment
Reciprocating
Engines
Turbl nes
Fuel
011 and
Dual
Gas
on
Gas
Total
Source
Shell
1971
(Reference 2-29)
0.40
1.74
0.03
0.08
2.25
MSST
1972
(Reference 2-4)
0.316
1.871
0.119
0.172
2.48
Current
1974
0.44
2.22
0.12
0.14
2.92
A-15
-------
o
o.
.*j **- t/i
£ "> 0
i- ^{ **
3 2«
o e o
UJ
1
u.
o
, t
IO
u
'o.
o.
^^
^*
r\
Ifl
1
o
1 1
§ §
o o
tfMB^^^^^^^^k
eC £ o
V£) I ID
O O O
0 O O
^ IO ^
O CD O
i.
0* 4->
C 0 >,
-i- 10 f
4-> C 10
f C 4->
i: il! o
0)
«/) O
01 U
^"* 4^
C9 O.
1
1
CM
VD
O
tfi
IO
0
(/)
^
0>
IO
OJ
Process Hi
M^
O
o
fn
o
o
o
5
c
J*
o>
1
i
o>
o
^
o^
CM
o
o
1
vo
8
o
I/)
5
O)
^
5:
L
QJ
>
0
1
O
V
o
o
c
o
l_
TJ
C
i
4^
4/1
o ro vo
cy>
§CM
CM
CM . O
0 I 0
o o
r*. CM CT> oo
O CM CM O
o o o o
o o o o
l/> irt
« ,*s
CD CD
0 0
U)
O
C7)
10 C
cr> aj -r-
C £ I- 0)
f- OJ C
4J C -> T
10 o> c t-
O) O. -^- 3
3: o co o
in
.*:
u
L>
CO
1
cy>
o
o
o
U)
3
Flares
«/>
3
o
O)
c
IO
^""
u
t"
so
o
*IO
o
A-16
-------
LU
Z
ii
O
o
a:
oo
z
o
oo
oo
X
o
o
>-
00
I
CVJ
to
o
to
to
r- tO
E C
LU O
-4 >
X
CD *o
z o
o
I
c
o
r
-M
fO
o
r-
'QL
0.
^
to
3
o
» 1
*i" cr>
OJ r-
o o
.
0 0
rO r
r (O
L. a.
-»-> -r-
to u
7 *r"*
T3 E
H-. SL
O
05
L.
(U
e
o
cr
CO
^^
0
o
1
+J
o
1
A-17
-------
TABLE A2-20. SUMMARY OF ANNUAL EMISSIONS FOR
NONCOMBUSTION SOURCES
Industry
Acid
Explosive
Total
Application
Sulfuric
Nitric
Adi pic
NO , 106 tons
^
0.012
0.140
0.016
0.056
0.224
TABLE A2-21. ESTIMATE OF ANNUAL NOX EMISSIONS
FROM OTHER SOURCES
Source
Solid waste disposal
Forest wildfires
Prescribed burning
Agriculture burning
Coal refuse fires
Structural fires
Misc. (welding, grain silos, etc.,
Total
NOY 103 tons
/\
165
152
33
14
58
7
50
479
A-18
-------
00
o
<
o
CU
1 *
It)
r~
3
U
k.
o
*->
u
CU
oo
.0
r
-M
O
-
1o
«->
o
M
<4-
O
^j
CU
(J
00
CU
<0
c
CU
i-
CU
o.
+J
v.
c <«_
2°
O CU
u
CU
CL
r_
>r_
o
IXJ
o
t^
to
CJD
cr>
5
oT
5
* ^
00
CM
i^O
in
CO
VO
"*-*
CM
00
o
. *
^
00
CM
0
00
co
o"
r>!
* *
vo
*^
°I
CD*
t/l
J_
01
o
CO
>,
ZD
r-» r^
f co
VO VO
CO O
CTi CM
^~^
0 0
VO vo
VO CM
cj o
CO CT>
CO CD
0 0
** CM
CD CD
1 1
1 1
In" ^
VO r^-
1 ^^
^"^
0 0
CM *J-
CM i
CM CD
0>
c in
r- CU
M C
(O -i-
tO O J3
CU O i-
C C 3
i- Q. t-
C O CO
UJ CU IO
CC CD
U
cn
CO
CM
oo'
^^
ej-
CM
CO
VO
'
co
CO
CD
^^
CO
LT5
CO
VO
CD
CM
VO
»
o
CM
CO
CD
to
L.
CU
r
0
m
lo
i/i
a
c
CTi
CD
CTi
O
i
CM
O
CM
cvT
in
o
o
^ s
^r
CD
CO
«^-
tn
0
CD
CO
co
^ '
vo
ffH^
-------
cr»
CD
<
to
<
Lo_
O
1
CU
3
1o
o
h-
^^
(0
o
*->
I)
o
CU
en
to
4J
c
CU
o
s-
Q)
CL
S-
o
4_>
O
CU
GO
1
CD
r
(X3
O
0
to
to
CD
in
in
CO
CO
CO
m
^j-
in
r-..
r
VO
r
^J-
00
CVJ
-
^J"
o
CVJ
**^r
f*+.
1
^f
cr>
in .
cr>
vo
v_^
0
o
CVJ
CO
to
S-
CU
r
p-
O
CO
>j
4_>
p
r
M
=
^j-
f**«
CVJ
'
CVJ
VO
CVJ
^~
in
tn
CD
«^-
in
CVJ
1
CTi
»
CVJ
^_^
CO
0
0
r^
en
c
r-
4J
pj
'
«;f
cn
CO
CVJ
vo
0
vo
CO
CVJ
1
CVJ
CO
,_!
x_^
CO
0
vo
to
0)
c
r-
JD
t-
^
^
H-
p
CO
CD
CO
in
vo
CT»
CO
vo
vo
CO
CT>
00
cr>
CO
oo"
r^.
vo
^_^
CO
CVJ
r
CO
f^.
00
^jl
r
V_*"
CO
in
CO
vo
to
i_
CU
r
p-
0
CO
,
OJ
p-
^.
4_)
to
^
^
-a
c
i i
,5-j-
CT>
CVJ
"~"
r
VO
O^
in
in
^
r^l
p
CO
*^-
CO
00^
in
0
«_^
r>x
vo
CVJ
,
CT»
^
^_^-
CO
vo
CVI
CVJ
to
S-
O)
p
p
0
CO
^_
fC
p-
0
i-
cu
^
o
o
CVJ
in
OC
CO
CO
in
00
,
o
f^
00
CVJ
CVJ
CO
1
,
in
,_!
r
x_^>
in
o
CO
in
cn
c
p
M
(O
O)
^r
^_
re
r-
^_)
C
CU
-o
to
CU
ct:
0
CD
o
r
1 '
CO
^*.
o
vo
in
*±
CD
CO
O">
CVJ
o
p^
o"
00
f»^
CVJ
>_«>
rv.
0
CO
CVJ
p
in
r^-
,_!
^~
v_^
f^»
CO
CVJ
cr»
i
h-
o
h-
CU
to
to
CU
o
to
CU
X
LU
to
A-20
-------
oo
o
oo
oo
X
o
O
oo
o
oo
II
a:
«=c
a.
CM
I
CM
CO
£ en
o*
cc ^
t- c\T
oo r^
to en
oo^T
g2
o*
1 Cn
D. r
ooo*
oo vo
oo en
LiJ i
s-
O
0
Q)
OO
00
CM
VO
in
O
ID
vo
vo
ID
00
CO
ID
_
*
00
CO
CO
^B
0>
*
o
CO
$
=D
0 0
vo vo
VO CM
CM O
vo «a-
^~ f^
00 O^ O*i
^t r CM
CM CM O
,
-o -a
o
03
CM
CD
C CO
r CU
M C
CO U X>
a> o t.
.c *" 3
c G co
LU
^
o
CO
r
T!
CO
3
o
C
1
CO
CO
0
^^
^^^
o
vo
CO
o
vo
CO
ID
CD
CO
CM
O
0
0
-
to
o
0
o
o
CM
O
CM
CO
O
£
o
vo
CO
ID
CD
ID
O
O
O
-'
'to
r
+*
C
O
D
r-
CO
01
cm
vo
«3-
0
1
1
en
CO
o
to
0
CM
CD
to
CD
^
4_>
to
0)
1-
ocess
t.
o.
CM
CM
O
i
1
cr>
o
o
CM
CD
1
CM
CD
C
O
4-J
CO
3
XI
o
c
o
CO
0
o
o
o
o
o
0
CD
CO
0
CD
10
C
o
1
to
ai
c
o
c
en
p»^
^
0
o
u
a
t.
o>
0
vo
r
^"
CO
r
00
-
vo
,_
*""
en
vo
CD
CM
CO
CD
en
en
O
_j
Q.
CO
3 CO
D tO
C C CL) C
r to ~£3 *t
3
o tu -a
O
C *r- O C
i Q_ ^ t i
O ^i U T3
-------
CO
o
CO
o
o
CO
o
o
LU
LO
CM
I
CM
CO
3
M
CO
in
*O
|R
& ^^
3
o"""
«r ^
S '"*"
^ i
**""*
^_^
1 CM
co r**
co cr>
2: r-
LO- »
f~~ O^
r- VO
O_ r
»
S^ O">
oc.
S-
O
u
o
cu
CO
5:
LO
r
0
00
*j!
*"~
CO
**
«5j"
r
CM
*~~
^*-
o
^
(/)
t.
cu
^~
r
O
CO
^-
*J
r-
r
r-
4_)
°
CM
r
O
CM
VO
CM
r
1
1
1
a>
c
r
M
(O
(0 O
C 2
i- a.
c o
LU CU
t^s
0
*~*
00
o
CO
,_I
cr>
00
CD
1
1
I
cu
c
f
n
i.
3
|_
CO
(O
CD
hx VO CO
C7> O"> LO
CO LO CO
"""
l*s. LO CM
CM ^- CTi
co ^t* r^»
cr> co ^~
fS^ f f^^
r S- CU
CO CU ^3
3 E *r-
"O E CO
C O CU
i o a.
CO
o
vo
^
a>
00
CO
vo
LO
CM
CM
CO
a*>
CO
o
CO
_J
-o
3
13
«3
A-22
-------
TABLE A2-26. ANNUAL NATIONWIDE NOY EMISSIONS PROJECTED TO 2000
(Reference 2-37) A
Source Category
Stationary Fuel Combustion
Electric Generation
Industrial
Commercial -Institutional
Residential
Industrial Process Losses*5
Solid Waste Disposal
Miscellaneous
TOTAL
NOX Emissions (106 tons)
1972
12.27
5.94
5.39
0.65
0.29
0.70
0.18
0.59
13.74
1980
15.96
(17.12)a
8.16
(9.32)
6.73
0.76
0.31
0.95
0.22
0.74
17.87
(19.03)
1985
16.82
(21.43)
8.20
(12.81)
7.46
0.84
0.32
1.14
0.25
0.87
19.08
(23.69)
1990
18.46
(27.14)
8.88
(17.56)
8.31
0.93
0.34
1.38
0.28
1.02
21.14
(29.82)
2000
21.74
(44.46)
10.24
(32.96)
10.01
1.11
0.38
1.85
0.34
1.32
25.25
(47.97)
N0x emissions for no new nuclear power plants after 1975 are given in parentheses.
Industrial process losses corrected for 1972 reporting error in Reference 2-36.
A-23
-------
TABLE A2-27. ESTIMATED FUTURE NSPS CONTROLS
(Reference 2-38)
NOX Source
Utility and Large
Industrial Boilers
(<73 MW)a Coal
Oil
Gas
Large Packaged Boilers
(<7.3 MW)a Coal
Oil
Gas
Small Packaged Boilers
(>7.3 MW)a Coal
Oil
Gas
Small Commercial and
Residential Units
Oil
Gas
Gas Turbines
1C Engines Dist Oil
Nat Gas
Gasoline
Process Combustion
Date Implemented
1971
1977
1981
1985
1988
1971
1971
1979
1985
1990
1979
1979
1979
1979
1979
1983
1983
1977
1983
1972
1985
1979
1985
1979
1985
1981
1990
Standard (lb/106 Btu)
0.7
0.6
0.5
0.4
0.3
0.3
0.2
0.6
0.5
0.4
0.2
0.3
50% reduction
0.2
0.3
0.07
0.04
0.3
0.2
3.2
2.4
2.9
2.1
2.2
1.6
20% reduction
40/K reduction
aThermal Input
A-24
-------
o
o
o
CM
CO
CO
X
o
O CM
h- cu
>*t r**. LO r^ r ^- LO
^3- «3- vo LO co oo LO
vo o *tf" CM co o co
VO CM r O O O O
oo CM LO oo r**. oo «*
r*« LO cr> co oo LO LD
o CM r- CM CM o CM
r%. CM i o o o o
O LO LO r>» r ^J- LO
^J- r VO VD CO 00 LO
i 00 ^J" CM CO O CO
r- r- C 0 0 O
1
o i vo oo r^ oo "^
i»>. r^ cr> oo oo LO vo
i LO r*. CM CM O CM
00 CM r O O O O
r^ LO CM «a- vo «a- co
ro oo co LO r-» «d- i
I LO LO CO ^t O CM
LO CM CM O O O O
c a
O O co
I ! CO
-P co -P en cu
ro S- CO C O
S- CU =3 -r- O
cu i .0 -P S- c:
C T- E t3 D. C O
cu o o cu o -
CD co o n: i T- -»J
ro -M CO
O "O i i_ -r- ro ZJ
i- CU ro t- co C O
O J* CU E =3 «i- O
CU U -P i- "O O C
r ro C ro C c: O
1 1 1 r> i_i *g i i H- « 2Z
CO
CTt
1
r
CO
CM
0
CM
r~~
LO
^
LO
p~"
CO
CO
^~
CO
CM
r
_J
^c
h-
o
1
A-25
-------
t/>
o
CO "
CM
UJ
LU to
H- O)
=3
0
Q.
C
o
C
o
JO
s.
to
o
3
CO
40
O
JD
^,
C
0
111
kU
^J
<0
U
fc.
IO
o_
X
o
1
0)
C
"
"
JJJ
to
00
X
o
§
_J
C
r-
I/)
in vo
*
in o
i t
CO
co in
CM O
^t CO
9 CO
*fr CO
1 |
CM O\
i in
co in
CO CT\
CO CO
1 1
co m
o vo
co r^
^
3
C *O
O 0)
fsl to
*- O
fc- CX
0 CX
3:0
CO
co m
CO r-
*~
1 1
CO GO
^ 0
'""
CO CO
in
CM r
1 1
CM
CM O^
^ O
CM
r>» in
in co
in vo
i i
en r^.
«ss- co
in CM
r CM
in «<3-
i i
0 0
o in
CM CM
10
C
C7>
C
to
1-
A-26
-------
i
u
i
o
o
E
00
< 01
^^
tt£
UJ (U
Of
CM
9
**
00 .
O
in
vt
UJ
O
e
ox
1*
~~"v
£3
j
i
«r
L
i
u
8
i **
s£
g
i i
*
Decrease
A/F
Su.
c
e
^ o
t ^
>
V
s
: 2
i
o cS
1
0
o
CM
1 CM
' 0
f*.
o
0 *l
r^.
c§
o
I s
0 CM
1 I
0$ eo
o »
0 O
r«. to
o o
1 1
CM tO
S 2
0 0
ro
O 00
O I-*
i i
tr> tr>
to f»
0 0
w u
>s X
i i
CM «
1ft
S
c
2
^
s 7
o
00
CM
to ~
o
en
o
f*.
1
1
1 ^
tO CM
r- ^-
1 1
CM S
m to
0 0
CM
§ 0
CM
0 '
CM
O
f»» CM
O v-
I I
Sin
to
»- CM
0 0
u u
>, X
CM «
*»
*
o
00
1 1
"".
$ r*>
d ^
So^
00
tf> tO
d d
I o
CM 00
to «r
O CM
1 1
00
to oo
IT) IT)
O r-
00 CM
to r-
tr> to
d d
to
to
S *~
«r> i
O to
to
o
u u
CM «
u.
i
A-27
-------
TABLE A6-4. CAPITAL AND OPERATING COSTS FOR DIFFERENT NOx ABATEMENT SYS-
TEMS IN A 300 TPD NITRIC ACID PLANT (Reference 6-6 and 6-26)
Capital Investment, ($)
Royalty
Operating Labor, (hr/yr)
($/yr)
Maintenance Labor,
($/yr)
Labor Overhead (incl. fringe
benefits & supervision, $/yr)
Catalyst or Molecular Sieve
Cooling Water, (gpm)
($/yr)
Steam, (Ib/yr)
($/yr credit)
Electricity, (kW)
($/yr)
Boiler Feed Water, (gpm)
($/yr)
Fuel, (106 Btu/hr)
($/yr)
Nitric Acid, (tpd)
($/yr)
Urea, tpd
($/yr)
Ammonium Nitrate, (tpd)
($/yr)
Depreciation (11-yr life)
Return on Investment (@ 20%)
Taxes & Insurance, (@ 2%)
Total Annual Cost, ($/yr)
Annual Cost, ($/ton)
Catalyst
Reduction
1,384,000
360
2,200
315
2,200
4,400
77,800
~ " . p
(15,833)
(387,590)
- 128
20,890
35
12,850
28.5
465,120
125,900
276,800
27,700
628,270
6.16
Molecular
Sieve
1,200,000
360
2,200
315
2,200
4,400
45,600
500
7,330
250
6,120
322
52,550
2.0
32,640
(6.6)
(112,200)
109,090
240,000
24,000
413,930
4.06
Grande
Paroisse
1,000,000
included
360
2,200
315
2,200
4,400
300
4,420
90
14,690
--
(6.0)
(102,000)
90,910
200,000
20,000
236,780
2.32
CDL/
Vitok
575,000
none
360
2,200
315
2,200
4,400
1,020
14,980
715
17,500
265
43,250
(6.0)
(102,000)
52,300
115,000
11,500
161,330
1.58
Masar
663,000
fee
360
2,200
540
3,775
5,975
--
1,310
32,070
20
3,260
(5.28)
(89,760)
1.37b
74,528
1.25
(42,500)
60,300
132,600
13,260
195,708
1.92
Goodpasture
425,000
51 ,000
360
2,200
315
2,200
4,400
30
440
--
45
7340
--
--
(13.0)
(422,000)
38,640
85,000
8,500
( 42,290)
(0.42)
Investment estimates exclude interest during construction, owners expenses, and land costs.
Includes credit for 0.0017 tons of urea/ton or nitric acid produced present in the spent
solution (D.SITPD).
'Parenthesis indicate credit taken.
A-28
-------
TABLE A6-5. ANNUAL ENERGY REQUIREMENTS (109 Btu) FOR NOX ABATEMENT SYS-
TEMS FOR A 300 TPD NITRIC ACID PLANT (Reference 6-6 and 6-26)
Steam (Credit)
Electrical
Natural Gas
Oil
Basic Nitric
Acid Plant
(71.4)
-
163.2
-
91.8
Catalytic
Reduction
(129.20)
10.97
232.56
-
114.33
Molecular
Sieve
2.04
27.59
-
16.32
45.95
Grande
Paroisse
-
7.71
-
-
7.71
CDL/
Vitok
5.83
22.71
-
-
28.54
Masar
10.69
1.71
-
-
12.40
Goodpasture
-
1.38
-
-
1»38
TABLE A6-6. BASIS FOR TABLES A6-4 AND A6-5 (Reference 6-6)
(Plant Capacity @ 300 tpd and 102,000 tons/yr)
(March 1975 Dollars, ENR Index = 2.126)
1. Operating Labor
2. Maintenance Labor
3. Overhead
4. Cooling Water
5. Boiler Feedwater
6. Natural Gas
7. Oil
8. Depreciation
9. Return on Investment
10. Taxes and Insurance
11. Nitric Acid
12. Urea
13. Ammonium Nitrate
14. 1 kWh = 10,500 Btu
15. Electricity
16. Ammonia
@ $6.1/hr
(3 $7.0/hr
@ 100% of labor (including fringe benefits
and supervision)
(3 $0.03 1,000 gal
(3 $0.75 1,000 gal
(3 $2.00/1O6 Btu
(3 $2.00/1O6 Btu
@ 11 yr straight line
@ 20% of capital cost
@ 2% of capital cost
0 $90/ton
(3 $160/ton
0 $TOO/ton
(3 $0.02/kWh
@ $157/ton
A-29
-------
-------
APPENDIX B
PREFIXES FOR SI UNITS
The names of multiples and submultiples of SI units may be formed by application of these
prefixes:
1 UV- \*\J i uj «M i ** *
Unit is Multiplied
1018
1015
1012
109
106
103
102
10
lo-1
10'2
io-3
io-6
io-9
io-12
io-15
io-18
Prefix
exa
peta
tera
giga
mega
kilo
hecto
deka
deci
centi
mi 1 1 i
micro
nano
pi co
femto
atto
Symbol
E
P
T
G
M
k
h
da
d
c
m
y
n
P
f
a
B-l
-------
-------
APPENDIX C
GLOSSARY
Biased Firing - An off-stoichiometric combustion technique in which the burners of a wall-fired
utility boiler are operated either fuel- or air-rich in a staggered configuration.
Boiler Efficiency - Heat Output x 100.
Heat Input
The overall figure reflects combustion efficiency, radiation and convection losses from the boiler,
and heat lost in exhaust gases.
Burners Out Of Service (BOOS) - An off-stoichiometric combustion technique in which some burners
are operated on air only.
Combustion Modification -An alteration of the normal burner/firebox configuration or operation
employed for the purpose of reducing the formation of nitrogen oxides.
Derating Reducing the heat input and power or steam output of a boiler below the level for which
it was designed.
Excess Air - Any increment of air greater than the stoichiometric fuel requirement. With gas-, oil-
and coal-fired boilers, some excess air is used to assure optimum combustion.
Field-Erected Boiler - All components of a boiler are delivered to the site and assembled in the
field. Mainly pertains to utility and large industrial boilers.
Firetube Boiler - Steam or hot water generator with heat transfer surface consisting of steel tubes
surrounded by water and carrying hot combustion gases.
Flue Gas Recirculation (FGR) -A combustion modification in which a portion of the boiler exhaust
gases are recirculated to the burners to inhibit NO formation.
Flue Gas Treatment - A process which treats tail gases chemically to remove N0x before release to
the atmosphere.
Fuel Nitrogen - Nitrogen that is chemically bound in the fuel.
Heat Input - The product of the fuel feedrate and the higher heating value, e.g., 10 Mg per hour
of coal with a higher heating value of 29 MJ/kg provides a heat input of 80.5 MW (290GJ/h).
C-l
-------
Heat Release Rate - The rate of combustion per unit volume of firebox, typically 1n terms of MH/m3.
Higher or Gross Heating Value (HHV) -The heat generated by complete combustion of a fuel, always
referenced to baseline temperature, e.g., 16°C. Heat available at the reference temperature 1s
Included 1n the higher heating value even 1f 1t 1s not practically available, I.e.* heat of con-*
denslng water vapor.
Low Excess A1r - A combustion modification 1n which NOX formation Is Inhibited by reducing the excess
air to less than normal ratios.
Lower or Net Heating Value (LHV) -The heat that 1s practically available from a fuel to generate
steam or otherwise raise the temperature of the media receiving energy. The net heating value assumes
complete combustion. It differs from the higher heating value 1n that heat of vaporizing water of
combustion 1s considered a recoverable loss.
Off-Stoich1ometr1c Combustion (OSC) -A combustion modification technique 1n which burner stolchl-
ometry 1s altered to Inhibit NOX formation. Types of OSC Include biased firing, burners out of
service, and two-stage combustion.
Packaged Boilers - These are usually boilers that are smaller and more economically assembled at
the plant, then shipped to the boiler site as one integral unit ready for operation after connection
to water, stream, and power.
Polycyclic Organic Matter (POM) - Organic compounds which exists in condensed phase at ambient tem-
perature and are emitted as either "carbon on particulate" or condensed onto emitted partlculate.
Polynuclear Aromatic Hydrocarbons (PNA) - Same as POM.
Stoichiometric Air - That quantity of air which supplies only enough oxygen to react with the com-
bustible portion of the fuel.
Two-Stage Combustion - A type of off-stoichiometr1c combustion 1n which the burners are operated
fuel-rich and the remainder of the required combustion air is Introduced through separate ducts
located above the burner. This is also called "overfire air" or "NOX port operation.
Matertube Boiler - A steam generator with heat transfer surface consisting of steel tubes carrying
water that are exposed to hot combustion gases.
C-2
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-450/1-78-001
3. RECIPIENT'S ACCESSI Of* NO.
4. TITLE AND SUBTITLE
Control Techniques for Nitrogen Oxides Emissions from
Stationary Sources - Second Edition
5. REPORT DATE
January, 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
R.M. Evans, R.J. Schreiber, H.B. Mason, W.M. Toy
L.R. Waterland, and C. Castaldini
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
Acurex Corporation/Energy & Environmental Division
485 Clyde Avenue
Mountain View, California 94042
11. CONTRACT/GRANT NO.
Contract No. 68-02-2611
Task No. 12
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
United States Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
Final
14. SPONSORING AGENCY CODE
200/04
15. SUPPLEMENTARY NOTES
Acurex Project Engineer: Michael Evans
EPA Project Officers: Gilbert Wood and Michael Davenport
16. ABSTRACT
This second edition of Control Techniques for Nitrogen Oxides Emissions
from Stationary Sources (AP-67) presents recent developments of nitrogen oxides
(NOx) control techniques which have become available since preparation of the
first edition (published, March 1970). As required by Section 108 of the Clean
Air Act, this second edition compiles the best available information on NOx
emissions; achievable control levels and alternative methods of prevention and
control of NOx emissions; alternative fuels, processes, and operating methods
which reduce NOx emissions; cost of NOx control methods, installation, and
operation; and the energy requirements and environmental impacts of the NOx
emission control technology.
Each stationary source of NOx emissions is discussed along with the various
control techniques and process modifications available to reduce NOx emissions.
Various combinations of equipment process conditions and fuel types are
identified and evaluated for NOx emission control.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Nitroqen Oxides Emissions
Control Techniques
Fossil Fuel Combustion
Nitric Acid Manufacturing
Costs
Photochemical Oxidants
13. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
4QQ
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
C-3
-------
------- |