EPA-450/3-83-002
Control Techniques for Nitrogen  Oxides
  Emissions from Stationary Sources-
          Revised Second Edition
             Emission Standards and Engineering Division
             U.S. ENVIRONMENTAL PROTECTION AGENCY
                Office of Air, Noise, and Radiation
              Office of Air Quality Planning and Standards
             Research Triangle Park, North Carolina 27711

                     January 1983

-------
This report has been  reviewed by the Emission Standards and Engineering Division of the
Office of Air Quality Planning and Standards, EPA, and approved for publication. Mention of
trade names  or commercial  products  is not  intended to constitute  endorsement  or
recommendation for use. Copies of this report are available through the Library Services Office
(MD-35), U.S. Environmental Protection Agency, Research Triangle Park, N.C. 27711, or from
National Technical Information Services, 5285 Port Royal Road, Springfield, Virginia 22161.

-------
                                          TABLE OF CONTENTS


Section                                                                                        Page

          LIST OF FIGURES	,	     viii

          LIST OF TABLES	•	.  . .  .	     xi

          SUMMARY	     S-l

   1      INTRODUCTION.	. .	>	     1-1

   2      CHARACTERIZATION OF M0v EMISSIONS	.	     2-1
                                X

          2.1  Definitions and Formation Theory	     2-1
          2.2  Sampling and Analysis Methods	.  .     2-2
          2.3  Equipment Descriptions, Emissions Estimates, and Emiss'ion Factors
               by Application Sector	   2-3

          2.3.1  Stationary Fuel Combustiori Sources	     2-7

          2.3.1.1  Utility Boilers	     2-7
          2.3.1.2  Industrial Boilers 	 . 	     2-15
          2.3.1.3  Commercial and Residential  Space Heating	.  .  . .	     2-18
          2.3.1.4  Internal Combustion Sources.	     2-18

          2.3.1.4.1  Stationary Reciprocating Internal Combustion
                       Engines	     2-18
          2.3.1.4.2  Gas Turbines	  .     2-19

          2.3.2  Industrial Process Heating	     2-22

          2.3.2.1  Iron and Steel Industry	     2-26
          2.3.2.2  Glass Industry	     2-29
          2.3.2.3  Cement Manufacturing Industry	     2-29
          2.3.2.4  Petroleum Refinina Industry	  .     2-31
          2.3.2.5  Brick and Ceramic Kilns	  .     2-31
          2.3.2.6  Nonccmbustion and Otner Minor Industrial Sources 	     2-34

          2.3.3  Solid Waste Disposal Sources	.....*..	     2-34
          2.3.4  Other Miscellaneous NO,, Sources	     2-45
                                       /\

          REFERENCES FOR SECTION 2	     2-49

   3      CONTROL TECHNIQUES	     3-1

          3.1  Combustion Modifications i	     3-1

          3.1.1  General Concepts on NO  Formation and Control	    -3-1

          3.1.1.1  Thermal NOV.	     3-2
          3.1.1.2  Fuel NOX ,	     3-4
          3.1.1.3  Summary of Process Modification Concepts 	     3-12

          3.1.2  Modification of Operating Conditions .....  	     3-14

          3.1.2.1  Low Excess Air Combustion (LEA)	     3-14
          3.1.2.2  Off-Stoichiometric or Staged Combustion(OSC)  	     3-15
          3.1.2.3  Flue Gas Recirculation(FGR)	     3-19
          3.1.2.4  Reduced Air Preheat Operation(RAP) 	     3-20
          3.1.2.5  Reduced Firing Rate	     3-21
          3.1.2.6  Steam and Water Injection(HI). ,v	     3-22
          3.1.2.7  Combinations of Techniques . . 	 ...     3-22
                                                iii

-------
                                    TABLE OF CONTENTS (Continued)

Section                                                                                        Page
          3.1.3  Equipment Design Modification  	     3-24
          3.1.3.1  Burner Configuration 	     3-24
          3.1.3.2  Burner Spacing 	     3-28
          3.1.3.3  Advanced Burner/Furnace Designs  	  	  .  .  .     3-28
          3.1.4  Fuel Modification.  .	     3-30
          3.1.4.1  Fuel Switching	i  .  .  .  .     3-31
          3.1.4.2  Fuel Additives	     3-34
          3.1.4.3  Fuel Denitrification 	     3-34
          3.1.5  Alternate Processes	'.	     3-35
          3.1.5.1  Fluidized Bed Combustion 	     3-35
          3.1.5.2  Catalytic Combustion 	 ,  	     3-36
          3.1.5.3  Repowering	     3-37
          3.1.5.4  Combined Cycles	.'.....     3-33
          3.2  Combustion Flue Gas Treatment	     3-39
          3.2.1  Dry Flue Gas Treatment	     3-39
          3.2.1.1  Selective Catalytic Reduction(SCR) 	     3-40
          3.2.1.2  Selective Noncatalytic Reduction (SNR)	     3-42
          3.2.1.3  Other Dry FGT Processes	     3-43
          3.2.2  Wet Flue Gas Treatment	     3-44
          3.3  Noncombustion Gas Cleaning	     3-45
          3.3.1  Plant Design for NO  Pollution Abatement at New Nitric Acid Plants  ....     3-45
          3.3.1.1  Absorption Column Pressure Control	."	'	     3-46
          3.3.1.2  Strong Acid^Processes	     3-46
          3.3.2  Retrofit Desian for NOV Pollution Abatement at New or Existing
                 Nitric Acid Plants .	     3-47
          3.3.2.1  Chilled Absorption 	  ...     3-47
          3.3.2.2  Extended Absorption. . ;	1  ....     3-48
          3.3.2.3  Wet Chemical Scrubbing 	  ....     3-48
          3.3.2.4  Catalytic Reduction	'	     3-50
          3.3.2.5  Molecular Sieve Adsorption 	     3-53
          REFERENCES FOR SECTION 3	     3-54
   4      LARGE FOSSIL FUEL COMBUSTION PROCESSES	     4-1
          4.1  Utility Boilers	     4-1
          4.1.1  Control Techniques 	     4-3
          4.1.1.1  Combustion Modifications	     4-3
          4.1.1.2  Flue Gas Treatment	,	     4-25
          4.1.2  Costs	     4-33
          4.1.2.1  Combustion Modification	,	     4-33
          4.1.2.2  Flue Gas Treatment 	   •  4-38
                                                 iv

-------
                                    TABLE OF CONTENTS (Continued)

Section                                                     ,                             .      Page
          4.1.3  Energy and Environmental Impact	    4-44
          4.1.3.1  Energy Impact  	    4-44
          4.1.3.2  Environmental Impact 	 	    4-48
          4.2  Industrial Boilers	    4-55
          4.2.1  Control Techniques	    4-56
          4.2.1.1  Combustion Modification	    4-59
          4.2.1.2  Post Combustion Techniques	    4-76
          4.2.2' Cost Impact	  •  4-78
          4.2.2.1  'Combustion Modification Techniques 	    4-78
          4.2.2.2  Post Combustion Techniques . »  	    4-83
          4.2.3  Energy and Environmental Impact   	    4-84
          4.2.3.1  Energy Impact	    4-84
          4.2.3.2  Environmental Impact	 	    4-87
          4.3  Prime Movers ...... 	    4-88
          4.3.1  Reciprocating  Internal Combustion Engines. ... 		  .    4-88
          4.3.1.1  Control Techniques  ,	    4-88
          4.3.1.2  Costs	,	    4-96
          4.3.1.3  Energy and Environmental Impact,		    4-102
          4,3.2  Gas Turbines	    4-109
          4.3.2.1  Control Techniques	    4-114
          4.3.2.2  Costs	    4-118
          4.3.2.3  Energy and Environmental Impact	    4-122
          4.4  Summary	*	    4-124
          REFERENCES FOR SECTION 4	    4-127
   5      OTHER COMBUSTION PRCCESSES	    5-1
          5.1 " Space Keating	    5-1
          5.1.1  Emissions. -	    5-2
          5.1.2  Control Techniques ......'	  '5-4
          5.1.2.1  Tuning	".  . .	    5-4
          5.1.2.2  Equipment Replacement	    5-8
          5.1.3  Costs...	    5-14
          5.1.4  Energy  and Environmental Impact	    5-16
          5.1.4.1  Energy Impact. 	    5-16
          £.1.4.2  Environmental Impact 	 ..... 	    5-16
          5.2  Incineratjon and Open Burning	.,	    5-18
          5.2.1  Municipal and  Industrial Incineration.	    5-18
          5.2.1.1  Emissions	,	    5-19
          5.2.1.2  Control Techniques  	 	    5-19
          5.2.1.3  Costs	    5-22

-------
                                    TABLE OF CONTENTS (Continued)

Section                                                                                        Page
          5.2.2  Open Burning	     5-22
          5.2.2.1  Emissions	     5-22
          5.2.2.2  Control Techniques 	     5-23
          5.3  Industrial Process Heating 	     5-26
          5.3.1  Petroleum, Chemicals, and Natural Gas	     5-27
          5.3.1.1  Fired Heaters	     5-27
          5.3.1.1.1  Process Description	     5-27
          5.3.1.1.2  Emission and Control Techniques	     5-40
          5.3.1.2  Catalytic Crackers and CO Boilers	 .     5-68
          5.3.1.2.1  Process Description	     5-68
          5.3.1.2.2  Emissions and Control Technology 	     5-68
          5.3.2  Metallurigical Processes	     5-70
          5.3.2.1  Process Description and Control Techniques 	     5-70
          5.3.2.2  Emissions	     5-77
          5.3.3  Glass Manufacture	.,	     5-79
          5.3.3.1  Process Description	     5-79
          5.3.3.2  Emissions	     5-83
          5.3.3.3  Control Techniques 	 	     5-85
          5.3.4  Cenient Manufacture	.'...„	     5-85
          5.3.4.1  Process Description. ......... 	     5-85
          5.3.4.2  Emissions	     5-87
          5.3.4.3  Control Techniques 	     5-38
          5.3.5  Coal Preparation Plants	'.	     5-89
          REFERENCES FOR SECTION 5	,	     5-91
   6      NONCOMBUSTION PROCESSES 	     6-1
          6.1  Nitric Acid Manufacture	     6-2
          6.1.1  Dilute Nitric Acid Manufacturing Processes	•	     6-?.
          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	     5-10
          6.1.3  Control Techniques for NOV Emissions from Nitric Acid Plants 	     6-11
                                          X
          6.1.3.1  Chilled Absorption 	     5-14
          6.1.3.2  Extended Absorption	I ....     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

-------
                                    TABLE OF CONTENTS (Continued)

Section                                                                                        Page
          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	     5-50
          6.2.4.3  Controls	     6-52
          6.2.4.4  Costs.	......;..     6-52
          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-l
          APPENDIX C - GLOSSARY	     C-l
                                                vii

-------
LIST OF FIGURES
Figure
2-1
2-2
3-1
3-2
3-3

3-4
3-5

3-6

3-7
3-8
4-1
4-2

4-3

4-4
4-5

4-6
4-7
4-8
4-9
4-10
4-11
4-12
4-13
4-14
4-15
4-16


Stationary Sources Of NO Emissions 	
Load Reduction Coefficient As A Function Of Boiler Load 	
Nitrogen And Sulfur Content Of U.S. Coal Reserves 	 ,
Conversion Of Fuel N In Practical Combusters 	 • 	
Possible Fate of Fuel Njtrogen Contained in Coal Particles During
Combustion 	
Conversion Of Nitrogen In Coal To NO 	 , .
Typical Arrangements For (B) Overfire Air, (A) Burners Out Of Service, And
CC) Flue Gas Recirculation 	
Correlation Of NO Emissions With Water Injection Rate For Natural Gas
Fired Gas Turbifie (Houston L&P Wharton No. 43 Unit) 	
Flow Diagram For Typical NO -Only SCR Process 	 ; 	
Extended Absorption System On Existing Nitric Acid Plant 	
Baseline NO Emissions - Coal-Fired Utility Boilers 	
Effect Of Combustion Modification Methods On Total Nitrogen Oxides .
Emissions And Boiler Efficiency 	 , 	 	
Effects Of Flue Gas Oxygen On The Formation Of NO Emissions From
Gas- And Oil-Fired Boilers 	 x 	 	
FGR Test Results on 5.1 MW (17.5 x 106 Btu/hr) Packaged Watertube Boiler ....
Effects Of Air Preheat Temperature On NO Emissions From Gas- And
Oil-Fired Boilers 	 x 	
Effects Of Excess Air On NO,, Emissions From Stoker Coal-Fired Boilers 	
/\
Effects Of NOX Emissions Level On Fuel Penalty For Lightduty Trucks 	
Effect Of Derating On 1C Engine HC Emissions 	
Effect of Retarding Ignition Timing On 1C Engine KC Emissions 	
Effect Of Air-To-Fuel Ratio On 1C Enoine HC Emissions 	
Effect Of Decreased Manifold Air Temperature On 1C Engine HC Emissions 	
Effect Of Water Injection On 1C Engine HC Emissions 	
Smoke Levels Versus NO Level For Large Bore Diesel Engines 	 ....
NO Emissions From Large Gas Turbines Without NO Controls 	
X X
NO Emissions From Small Gas Turbines Without NO Controls 	
NOV Emissions From Gas Turbines Having NOV Controls And Operating
5n Liquid Fuels 	 x 	
_Page
2-4
2-14
3-6
3-8

3-9
3-11

3-17

3-23
3-41
3-49
4-2

4-60

4-64
4-68

4-70
4-73
4-103
4-106
4-106
4-107
4-108
4-108
4-110
4-112
4-113

4-117
     viii

-------
LIST OF FIGURES (Continued)
Figure
5-1

5-2

5-3

5-4
5-5
5-6

5-7

5-rS
5-9
5-10

5-11

5-12

5-13
5-14
5-15
5-16
5-17

5-18
5-19
6-1
6-2
6-3
6-4
6-5
6-6
6-7
6-8

General Trend Of Smoke, Gaseous Emissions, And Efficiency Versus
Stoichiometric Ratio For Residential Heaters. . 	 	
Effect of Excess Air On NO Emissions From A 45.3 Mg (50 Ton) Per Day
Batch-Feed Incinerator. . 	
Effect Of Underfire Air On HO Emissions From A 227 Mg (250 Ton) Per Oay
Continuous Feed Incinerator 	 . 	
Examples Of Fired Heater Designs 	 	 	
Typical Burners By Type Of Fuel Burned. . 	
MO Emissions From Standard Premix Burners 	
X
NO Emissions From A Refinery Gas-Fired Reforming Heater With Low Excess
Hir Burners 	 	 	 	
NOV Emissions From A Single Burner, Forced Draft Debutanizer Bottoms Reboiler . .
A
Forced Draft Crude Heater With Staaed Combustion Burners 	 	 	 '
NO Emissions From A Natural Draft Naptha Reformer With Staged Combustion
Air Burners 	 '....... 	 	
NOX Emissions From A Balanced Draft Crude Preheater With Staged Combustion
Air Burners, Combustion Air Preheater 485 To 500K ......; 	
NO Emissions As A Function Of Stack Oxygen For A Distillate Oil -Fired,
Natural Draft Process Heater. .... 	
Residual Oil -Fired Horizontal Heater With Staged Combustion Air Burners 	
Residual Oil-Fired Natural Draft Cylindrical Heater 	 , 	
Residual Oil/Refinery Gas-Fired Natural Draft Heater 	 	 .
NO Emissions From A Natural Draft Atmosphere Crude Heater With Staged
Combustion Air Burners, Oil And Gas Fuel Combustion 	 	
NO Emissions From A Natural Draft Vacuum Distillation Column Heater
with Staged Combustion Air Burners, Oil And Gas Firing 	
NO Emissions As A Function Of Time For An Open Hearth Furnace 	
The Effect Of Cement Kiln Temperature- On NOX Emissions 	 	 .
Single Pressure Nitric Acid Manufacturing Process 	 	
Dual Pressure Nitric Acid Plant Flow Diaarem 	 ,
Nitric Acid Concentrating Unit 	
Process Flow Diagram For Direct Production Of Highly Concentrated Nitric Acid . .
Schematic Diagram Of The CDL/VITOK NO Removal Process. . . 	
TVA Chilled Absorption Process 	
Grande Pariosse Extended Absorption Process For NO Treatment 	
Flow Diagram Of The MASAR Process 	 	 	 	
Page

5-3

5-20

5-21
5-34
5-39
5-42


5-46
547
. 5-49

5-51

5-52

5-54
5-55
5-56
5-58
5-60

5-61
5-81
5-90
6-5
6-6
6-8
6-9
6-15
6-17
6-18
6-20

-------
                           LIST OF FIGURES (Continued)
Process Flow Diagram For The Gcodpasture Process	    6-23
Nonselective Catalytic Reduction System ....  	  .  	    6-26
Molecular S*eve System	    6-30
Batch Process For The Manufacture Of Nitroglycerln  (N6)  	 	    6-42
Schmld-Heissner Continuous-Nitration Plant	    6-43
Biazzi Continuous-Nitration Plant 	    6-44
Recovery Of Spent Add	    6-45
Trinitrotoluene (Batch Process) Manufacturing Diagram  	    6-51

-------
LIST OF TABLES
Table,
S-l
S-2
S-3
S-4
S-5
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


Nationwide 1980 Nitrogen Oxide Emissions From Stationary Sources 	
Summary of NO Control Techniques 	 .............
Summary of Combustion Process Modification Concepts 	
Overall Evaluation of NOV Control Techniques for Combustion Sources 	
x s
NO Abatement Methods on New or Existing Nitric Acid Plants 	
National Estimates of Nitrogen Oxide Emissions 	 .......
i960 Nitrcgenoxide Emissions From Fuel Combustion In Stationary Sources
By Category for Major Fuels 	 	 	 ......
Nitrogen Oxide Emission Factors for Bituminous and Sub-Bituminous Coal
Combustion 	 	 	
Nitrogen Oxide Emission Factors for Anthracite Coal Combustion Without
Control Equipment 	 	 	 	
Nitrogen Oxide Emission Factors for Lignite Combustion Without Control
Equipment . . '. 	 	 	 , . .
Nitrogen Oxide Emission Factors for Fuel Oil Combustion Without Control
• Equipment 	 	 	 	 	
Nitrogen Oxide Emission Factors for Natural Gas Combustion Without
Control Equipment . : 	 	 	
Nitrogen Oxide Emission Factors for LPG Combustion Without Control
Equipment 	 	 	
Nitrogen Oxide Emission Factors for Wood/Bark Waste Combustion Without
Control Equipment 	
Nitrogen Oxide Emission Factors for Bagasse Combustion Without Control
Equipment 	 	 ..........
Nitrogen Oxide Emission Factors for Residential Fireplaces 	
Nitrogen Oxide Emission Factors for Heavy-Duty, Natural Gas-Fired Pipeline
Compressor Engines Without Control Equipment 	 	
Nitrogen Oxide Emission Factors for Stationary Large Bore Diesel and

Composite Nitrogen' Oxide Emission Factors for the 1971 Population of
Electric Utility Turbines Without Control Equipment 	 . 	
Nitrogen Oxide Emission Factors for Keavy-Duty, Natural Gas-Fired Pipeline
Compressor Gas Turbine Engines Without Control Equipment. , 	
Nationwide Nitrogen Oxide Emissions Estimates From Major Industrial Process
Sources, 1970 - 1980 	 	
1
Nitrogen Oxide Emission Factors for Coke Manufacture Without Control
Equipment 	 . 	 	
Nitrogen Oxide Emission Factors for Gray Iron Furnaces Without Control
Equipment 	 	 	 	 	 	
Page
. . S-2
. . S-4
. . S-5
. . S-8
. . S-13
'. . 2-5

. . 2-6

. . 2-10

. . 2-11

. . 2-11

. . 2-12

. . 2-13

. . 2-16

. . 2-16

. . 2-17
. . 2-17

. . 2-20

. . 2-21

. . 2-23

. . 2-24

. . 2-25


, , 2-27

2-27
     xi

-------
LIST OF TABLES (Continued)
Table
2-19

2-20

2-21

2-22

2-23
2-24

2-25
2-26
2-27
2-28
2-29

2-30
2-31
2-32

2-33

2-34
2-35
3-1
3-2
3-3
4-1
4-2
4-3
4-4

4-5
4-6
4-7

Nitrogen Oxide Emission Factors for Steel Foundries Without Control
Equipment 	
Nitrogen Oxide Emission Factors for Glass Manufacturing Melt Furnaces
Without Control Equipment 	
Nitrogen Oxide Emission Factors for Glass Fiber Manufacturing Without
Control Equioment 	 	 	
Nitroaen Oxide Emission ^Factors for Cement Manufacturing Without Control
Equipment 	 	
Nitrogen Oxide Emission Factors for Petroleum Refineries. . . . r 	
Nitrogen Oxide Emission Factors for Brick Manufacturing Without Control
Equipment 	 	
Nitrogen Oxide Emission Factors for Nitric Acid Production 	
Nitrogen Oxide Emission Factors for Adiptic Acid Manufacture 	
Nitnjgen Oxide Emission Factors for Explosive Manufacturing 	
Nitrogen Oxide Emission Factors for Miscellaneous Industrial Processes 	
Nitrogen Oxide Emission Factors for Refuse Incinerators Without Control
Equipment 	
Nitrogen Oxide Emission Factors for Sewage Sludae Incinerators 	
Nitrogen Oxide Emission Factors for Auto Body Incinerators. . . . 	
Nitrogen Oxide Emission Factors for Waste Incineration in Conical Burners
Without Control Equipment 	
Nitrogen Oxide Emission Factors for Open Burning of Nonagri cultural
Materials 	
Nitrogen Oxide Emission Factors for Forest Wildfires 	 	
Nitrogen Oxide Emission Factors for the Dentonation of Explosives 	
Factors Controlling the Formation of Thermal NO 	
Summary of Combustion Process Modification Concepts 	
NO Formation Potential of Some Alternate Fuels 	
Average NO Reduction With Low Excess Air Firing. ... 	
Average NO Reduction With Burner Out of Service 	
X
Average NQX Reduction With Overfire Air 	
Average NOX Reduction With Flue Gas Recirculation 	 	 . .

Average NOX Reduction With Reduced Firing Rate 	 * 	
Average NO Reduction With Off Stoichicmetric Combustion and Flue Gas
Secirculatlon ..... 	 	
Average NO Reduction With Reduced Firing Rate and Off Stoichiometric
Combustion 	
Page

2-28

2-28

2-30

2-32
2-33

2-35
2-36
2-37
2-38
2-39

2-43
2-44
2-44

2-46

2-46
2-47
2-48
3-5
3-13
3-32
4-5
4-6
4-7
4-8

4-9
4-10
4-11
           xii

-------
LIST OF TABLES (Continued)
Table.
4-8

4-9
4-10
4-11
4-12
4-13
4-14
4-15
4-16
4-17
4-18



4-20

4-21-

4-22

4-23

4-24

4-25
4-26
4-27
4-28

4-29

4-30
4-31
4-32
4-33

4-34

Average NO Reduction With Load Reduction, Off Stoichiometric Combustion
and Flue Gas Recirculation. . 	
Effect of Low NO Operation on Coal -Fired Boilers 	
Effect of Low NO Operation on Oil-Fired Boilers. 	 	
Effect of Low NO Operation on Gas-Fired Boilers 	 	
X
Summary of Utility Boiler Combustion Modification Control Costs 	
Projected Retrofit Control Requirements for Alternate NO Emission Levels ....
NO FGT Processes Selected for Evaluation By TVA in 1980 Study 	
Capital Investment Developed By TVA for Alternative FGT Systems . . . 	 	
Annual Revenue Requirements Developed by TVA for Alternative FGT Systems. ....
Summary of Capital Investments Developed in 1981 TVA Study 	
Sunmarv of Annual Revenue Reciri rements Developed in 1981 TVA Study. .......
a. Estimated Total Capital Requirement of Post Combustion NO Y Control Processes

b. Estimated Levelized Cost Reauirement of Post Combustion N0» Control Processes
Partial Summary of EPA/Acurex Combustion Modification Environmental Assessment
Field Testing 	 	 	 	
Measured Change in the Concentrations of Selected Pollutants Across the"
EPA Pilot Plant SCR Reactors 	
Installed Capacity of U.S. Watertube Industrial Boilers By Unit Size and
Fuel Type (in 1977) 	
Installed Capacity of U.S. Industrial Fire-Tube Boilers By Unit Size and
Fuel Type (1977) 	 	
NO Emission Control Techniques for Gas-Fired and Oil-Fired Industrial
Boilers 	 	 	
NO Emission Control Techniques for Coal-Fired Stoker Industrial Boilers. ....
Safe Operating Levels for LEA 	 . 	
Estimated Cost of Low Excess Air Operation for New Boilers (.1978 Dollars) ....
Estimated Cost of Staged Combustion Operation for New Boilers (1978
Dollars) 	 	 	
Estimated Cost of Flue Gas Recirculation Operation for New Boilers
(1978 Dollars^ 	 	 	 . .
Annual Cost of NO Control Systems Applied to Coal -Fired Boilers 	
Annual Cost of NO Control Systems Applied to Oil-Fired Boilers . . . 	 	
Annual Cost of NO Control Systems Applied to Natural Gas-Fired Boilers ......
A
Summary of NOV Emission Control Techniques for Reciprocating Internal
Combustion Engines 	
Effect of NO Controls on Large-Bore Internal Combustion Engines 	
Page.

4-12
4-14
4-20
4-23
4-35
4-36
4-40
4-42
4-42
4-43
4-43




4-49

4-54

4-57

4-58

4-61
4-62
4-65
4-80

4-81

4-82
4-85
4-85
4-85

4-89
4-91
        xiii

-------
LIST OF TABLES (Continued)
Table
4-35

4-36
4-37
4-38
4-39
4-40
4-41

4-42

4-43

4-44
4-45
4-46
4-47
4-48
4.49

4-50
5-1
5-2
5-3
5-4

5-5

5-6
5-7

5-8
5-9
5-10
5-11

Control Techniques for Truck Size Diesel Engines [^375 kW (500 HP)] to Meet
1975 California 13.4 G/KWHR (10 6/HP-HR) Combined NOX and HC' Level 	
1975 Vehicle Emission Limits 	
Emission Control Techniques for Automotive Gasoline Engines 	
Emission Control Systems for Conventional Gasoline Internal Combustion Engines. .
Cost Impacts of NO Controls for Large-Bore Engines 	
Typical 1974 Baseline Costs for (Large>75 kW/cylinder) Engines • • • 	
Typical Control 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 NO Control on CO Emissions from Internal Combustion
Engines 	 	 	
Relationship Between Smoke, EGR, and Retard 	
Gas Turbine - Summary of Existing Technology - Combustion Modifications .....
Impact of NO Emission Control on the Installed Capital Cost of Gas Turbines . .
Water Injection Costs, Mills/kWh . . 	 	
Representative Effects of NO Controls on CO Emissions from Gas Turbines ....
Summary of the Effects of NO Controls on Vapor Phase Hydrocarbon Emissions from
Gas Turbines 	 	
Summary of NO Controls Technology 	 ...
X
Emission Factors for Residential and Commercial Heating Systems .........
Uncontrolled NOX Emissions from Residential Space Heating Systems 	
Comparison of Hean Emissions for Cyclic Runs on Residential Oil -Fired Units . . .
Performance Summary of Low-NO Control Equipment for Natural Gas-Fired
Residential Heaters 	 	
Performance Summary of Low-NO Control Equipment for Distillate Oil-Fired
Residential Heaters 	 	
Cost Impact of NO Control Alternatives 	
Effects of Low-NO Operation on Incremental Emissions and System Performance
for Residential Wlrm Air Furnaces 	
Annual Emissions of Nitrogen Oxides from Open Burning 	
Major Refinery Processes Requiring a Fired Heater 	
Typical Fired Heater Applications in the Chemical Industry 	
Basic Fired Heater Applications 	
Page

4-93
4-94
4-94
4-95
4-98
4-98

4-99

4-100

4-104
4-111
4-115
4-119
4-120
4-123

4-123
4-125
5-5
5-6
5-7

5-9

5-10
5-15

5-17
5-23
5-28
5-31
5-32
                    xiv

-------
                                     LIST OF TABLES (Continued)

Table                                                                                          Page
 5-12     Summary of NO  Test Data on a Standard Duel Fuel Burner	    5-44
 5-13     Retrofit Costs for Low-NO  Burners	    5-66
 5-14     NO  Emissions from Petroleum Refinery CO Boilers. 	    5-69
 5-15     Estimated HOX Emissions from Steel Hill Processes and Equipment 	    5-78
 5-16     Effects of NO  Controls on Steel Industry NO  Emissions 	    5-80
 5-17     NO  Emissions from Glass Melting Furnaces 	    5-84
 5-18     Recommended Programs for Reducing Emissions and Energy Consumption in the
          Glass Industry	    5-86
 6-1      NO,. Abatement Methods on Nevf or Existing Nitric Acid Plants	    6-12
            A
 6-2      Performance of Hercules Purasiv N Unit During Three Day Pun 	    6-31
 6-3   •   Performance of U-. S. Army-Holston Purasiv N Unit During Three-Day Run	    6-32
 6-4      Capital and Operating Costs for Different NO  Abatement Systems in a 270 Mg/d
          Nitric Acid Plant		    6-34
 6-5      Annual Energy Requirements (TJ) for NO  Abatement Systems for a 270 Mg/d
          Nitric Acid Plant . .		    6-35
 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 in 1970*	    6-48
 6-9      Emission Factors for Manufacture of Explosives	    6-53
 A6-4     Capital and Operating Costs for Different NO  Abatement Systems in a 300 TPD
          Nitric Acid Plant	    A-2
                                        9
 A.6-5     Annual Energy Requirements (10  Btu) for NO  Abatement Systems for a 300 TPD
          Nitric Acid Plant	x	    A-3
                                                 xv

-------

-------
                                               SUMMARY
CHARACTERIZATION OF NOV EMISSIONS
                      A
       Manmade oxides of nitrogen are currently emitted at a rate of about 20 Tg (22 million
tons/yr) in the United States.  Stationary sources account for approximately 60 percent of these
emissions, of which 97 percent are due to combustion sources.  Combustion generated NOX 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 njtrogen compounds.  This can be the dominant mechani-sm
with the firing of coal and high nitrogen residual oils.  The rate of formation of both thermal NO
                                                                                                  X
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 NOV are sharply increased by increased local
                                                       A                    ,
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 classification of stationary sources according to factors known to influence NOX formation.
Over 100 combinations of equipment type and fuel type are identified as having significantly
different potential for NO  emissions and/or NO  control.  The emission compilation for these
                          X                    X
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.

       The total 1980 nationwide NO  emissions from stationary sources, grouped according to appli-
cation sector, are shown in Table S-l,  On an uncontrolled basis, utility boilers accounted for
about 58 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 about 51 percent of
total NOX, the firing of oil yielded about 8 percent, and the firing of natural gas yielded
30 percent.  Wood, kerosene, LPG, and coke generated the majority of the remaining NO  emissions.
                                                 S-l

-------
                   TABLE S-l.  NATIONWIDE 1980 NITROGEN OXIDE EMISSIONS
                               FROM STATIONARY SOURCES
Stationary Source Category
Fuel Combustion
Electric Utilities
Industrial
Commercial -Institutional
Residential
Fuel Combustion Total
NO Emissions,
Tgx(10° tons)

, 6.4 (7.0)
3.0 (3.3)
0.3 (0.33) , '
0.3 (0.33)
10.0
Percent of Total
Emissions

58
27
2.7
2.7
(11)


91
Industrial Processes                              0.7 (0.77)                6.3

Solid Waste Disposal

     Incineration                    0.0 (0)       .              0
     Open Burning                    0.1 (0.11)                   0.9
     Solid Waste Total                            0.1 (0.11) .               0.9

Miscellaneous

     Forest Fires                    0.2 (0.22)                   1.8
     Other Burning                   0.0 (0)                      0
     Hisc. Organic Solvent           0.0 (0)                      0_
     Miscellaneous Total0.2 (0.22)                1.8

Total of All Categories                          11.0 (12.1)              100
                                 S-2

-------
CONTROL TECHNIQUES
      • Current and advanced methods for stationary source NOV control  operate either through
                                                            A
suppression of NOV formation in the process or through physical  or chemical  removal  of NO  from the
                 X                                                                       X
stack gases.  Both processes are effective with combustion sources, though most domestic experience
has involved suppression of NO  formation.  Candidate NO  suppression  approaches include combustion
process modification through alteration of operating conditions  on existing systems  or alternate
design of new units; fuel modification through fuel switching, fuel denitrification, or fuel
additives; and use of alternate combustion concepts such as catalytic  combustion and fluidized bed
combustion.  Stack gas NO  removal systems, principally, using selective catalytic or noncatalytic
reduction, have been extensively applied to combustion systems in Japan.  However, experience in the
United States has primarily been limited to a few pilot scale demonstrations of catalytic systems
and some industrial applications of noncatalytic systems.  Removal of  NO  from stack gases is also
effective with noncombustion sources of NO , chiefly chemical manufacturing.  Candidate approaches
for this application include catalytic reduction, wet chemical scrubbing, extended and chilled
absorption, and adsorption with molecular sieves.  A summary of general stationary source NO
control techniques is given on Table S-2.

       Combustion process modifications have been extensively implemented on existing coal-, oil-,
and gas-fired boilers to comply with local emission standards.  Combinations of external control
techniques such as low excess air firing, flue gas recirculation and staged combustion have yielded
emission reductions of up to 60 percent compared to the uncontrolled,  baseline emissions of units
designed prior to the 1970s.  Staged combustion techniques include biased burner firing, burners out
of service, and overfire air injection.  A summary of combustion modification concepts is given in
Table S-3.

       Currently, the most commonly applied low NOX technique for coal-fired utility boilers is
staged combustion through the introduction of overfire air.  This technique has been used in both
new and retrofit applications, achieving NOX reductions of 30 to 50 percent compared to older
baseline levels, or controlled emissions of 210 to 300 ng/J (0.5 to 0.7 lb/106 Btu).  More recently,
first generation low NO  burners have been installed on some units and found to be at least as
effective as overfire air.  In fact, new wall-fired units subject to the 1979 new source performance
standards will generally rely on low NOX burners, enlarged furnace designs, and overfire air
combustion.  Corner fired units will rely on overfire and low excess air.  The combination of
overfire air with low NOX burners has resulted in 40 to 60 percent NO   reductions compared to older
                                                 S-3

-------
TABLE S-2.  SUMMARY OF HOX CONTROL TECHNIQUES
Ttchntqut
Ccwfcustlon
Modification
Flut 6«s/
Noncombustlon
Tail Gas
Treatment
Fuel
Switching
Fuel
Additives
Fuel
Oenltriflcitlon
Catalytic
Combustion ';
Fluldtied Bed
Cotfcustlon
Principle of Operation
Suppress thermal NO. through re-
duced flame temperature, reduced
02 level ; suppress fuel HOX
through delaying fuel/air nix-
ing or reduced 0; level In pri-
mary Dane zone
Additional absorption of NOX to
HK03J conversion of NO. to
NtyNOj; reduction of NOX to N?
by catalytic treatment
Simultaneous SOX and NO, con-
trol by conversion to clean
fuels; synthetic gas or oil
from coal; SRC; methanol;
hydrogen
Reduce or suppress NO by
catalytic action of fuel
additives
Rasovil of fuel nitrogen com-
pounds by pretreatment
Heterogeneously catalyzed
reactions yield low combus-
tion temperature, ION ther-
nal NOX
Coal cMbusttofl 1n solid bed
yields low teoperature, loa
»«
Status of Development
Operational for point
sources; pilot-scale and
full scale studies on com-
bined Modifications, opera-
tional problems and ad-
vanced design concepts for
area sources
Operational for existing
and new nitric acid plants
meeting NSPS; pilot Stale
feasibility studies for
conventional combustion
systems in U.S., but
operational in Japan
Synthetic fuel plants In
pilot-scale stage; cora-
nercial plants due by
mid-1 980 's
Inactive; preliminary
screening studies indi-
cated poor effective-
ness
Oii desulfurization
yields partial deni-
trification
Pilot-scale test beds for
catalyst screening,
feasibility studies
Pilot-scale study of at-
mospheric and pressurized
systems; focus on sulfur
retention devices


Degree of control
United by opera-
tional problems
Catalytic processes
gaining experience
with high parti cu-
late applications;
tests with additional
coal types needed.
Fuel cost differentia]
nuy exceed NOX, SOX,
control costs with
coal
Large make-up rate of
additive for signifi-
cant effect; presence
of additive as pollu-
tant
Effectiveness for coal
doubtful ; no effect on
thermal NOX
Halted retrofit appli-
cations; requires clean
fuels
Fuel nitrogen conversion
nay require control
(staging); may require
large nuke-up of lime-
stone sulfur absorbent
Applications
Near-Tern
Retrofit utility.
ndiistrial toilers,
gas turbines; im-
iroved designs;
low utility
Killers
Noncombustlon
sources (nitric
acid plants)
Negligible use
Negligible use
Negligible use
Seal' space
heaters
Negligible use
Long-Term
Optimized design trei.
point sources
Possible supplement to
combustion modifications;
simultaneous SOX/NOX
removal
New point sources,
(combined cycle)
Convert area sources
(residential)
Not pi-wising
Supplement to cottustion
modification
Possible use for resi-
dential heating, snail
boilers, gas turbines
Utility. Industrial boil-
ers beginning 1960's;
possible coablneil cycle.
waste fuel application

-------
                               TABLE S-3.  SUMMARY OF COMBUSTION PKOCESS MODIFICATION  CONCEPTS
Combustion
Conditions
Decrease
primary
flame zone
02 level
Decrease
peak
flame
temperature
Control
Concept
Decrease overall
0; level
Delayed mixing
of fuel and air
Increased fuel/
air mixing
Primary fuel-
rlch flame
zone
Decrease
ad 1 alia tic flame
temperature
Decrease com-
bustion
Intensity
Increased flame
zone cooling/
reduce resi-
dence time
Applicable
Equipment
Boilers, furnaces
Boiler, furnaces
Gas turbines
Boilers.
furnaces, 1C
Boilers, fur-
naces. 1C,
gas turbines
Boilers, furnaces
Boilers, furnaces
Effect on
Thermal NOX
Reduces 02-rich,
Mgh-IHL pockets
In the flame
- Flame cooling am!
dilution during
delayed mix re-
. duccs peak temp.
Reduces local hot
stokliiotnetric
regions in over-
all fuel lean
combustion
Flame cooling In
low-02, low-temp.
primary zone re-
duces peak temp.
Direct supres-
slon of thermal
NOx mechanism
increased flame
zone cooling
yields lower
peak temp.
Increased flame
zone cooling
yields lower
peak temp.
Effort nn
Fuel HOX
Reduces exposure
of fuel nitrogen
Intermediaries '
to 02
Volatile fuel N
reduces to llj in
the absence of
oxygen
Increases
Volatile fuel H
reduces to Hj in
the absence of
oxygen
Ineffective
Minor direct
effect! indirect
effect on mixing
Ineffective
Primary Applicable Controls
Operational
Adjustments
Low excess air
firing
Curner
adjustments

Burners out of
service; biased
burner firi.tg
Reduced air
preheat
Load reduction
Durner tilt
Hardware
Modification
Flue gas reclrcu-
lation (FGR)
Low HO x burners

Overflre air
ports, stratified
charge
Hater injection.
FGR


Major
Redesign

Optimum burner/
firebox design
Hew can design;
premlx, prevap.
Burner/firebox
design for two-
stage combustion

Enlarged firebox
increased burner
spacing
Redesign heat
transfer sur-
faces, firebox
aerodynamics
I
01

-------
baseline levels, or controlled emissions of 170 to 260 ng/J (0.4 to 0.6 lb/10  Btu).   Similar
control levels should also be achievable on large pulverized coal  industrial  boilers.
       Current testing activities are attempting to identify and quantify potential  operating
problems with combustion modifications, such as increased water-wall tube corrosion under reducing
conditions.  Such problems are expected to be most prevalent for retrofit applications where the
boiler was not designed for low'NOv operation.  However, the results of most  short term corrosion
                                  X
tests to date have not indicated this to be a major problem.

       Retrofit combustion process modifications have also been extensively applied to gas turbines.
Water injection has been sucessfully implemented to achieve emission levels of 75 ppm at 15 percent
excess oxygen.  Current activity is focusing on development of dry controls using premixing,
prevaporizatio"n and controlled mixing for application to new combustor can designs.

       There has been only limited field implementation of combustion process modifications for
other stationary combustion equipment e.g., small 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 NO  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 NOX 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 NO  control, e.g., desulfurization cost tradeoffs, system
efficiency.

       Fuel oil denitrification, usually as an adjunct to oil desulfurization, shows  promise for
reducing fuel NOV.  This concept may be effective for augmenting combustion modifications for NO,,
                A                                                                               X
control with the firing of residual oil but is expensive when applied for NO   reduction alone.  Fuel
additives are not directly effective for suppressing NOV emissions.  Their use to suppress fouling
                                                       X
and smoke emissions, however, may permit more extensive use of combustion control methods than would
otherwise be practical.
                                                 S-6

-------
       Alternate combustion concepts under development include catalytic combustion and fluidized
bed combustion (FBC).  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 turbines and space heating systems.  Fluidized bed combustion-pilot plants have demonstrated NO
                                                                                                   X
emissions of the same order as conventional coal-fired power plants using process modifications for
NO  control (170 ng/J, or, 0.4 Ib N02/10  Btu).  The potential for replacement of conventional
utility and industrial boilers by FBC depends on a number of other factors such as SO  control cost
tradeoffs and operational flexibility, e.g., load following.

       Stack gas treatment for NO  removal has been implemented in the U.S. primarily on noncombus-
tion sources.  Here, an additional incentive is the recovery of NO- 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 HO-  as a
potential feedstock is causing interest to shift to alternate processes such as molecular sieve
absorption 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
accelerated, however, as a result of increased emphasis on stationary source NO  controls in  the
                                                                               X
national NOV abatement program.  Flue gas treatment could be effective in the 1980's to augment
           X
combustion process modifications on large sources if stringent emission control is required,  for
example, to comply with a potential short-term NOj ai'r quality standard.  Current developmental
activity includes transferring FGT technology from Japan where stringent NO  controls are enforced
                                                                           X
and demonstrating this technology on pilot and full scale systems in the U.S.  The most advanced
processes include selective catalytic reduction and selective noncatalytic reduction.  Other
techniques under development include electron beam irradiation and wet scrubbing.  However, the dry
techniques currently appear to be the most cost effective, even when used in combination with wet
flue gas desulfuriration systems.
       A summary evaluation of NO  control techniques for combustion sources is given in Table S-4.

LARGE FOSSIL FUEL COMBUSTION PROCESSES

       The three largest stationary emitters of NO  are electric power plant boilers (58 percent of
the total), industrial boilers (10 percent) and prime movers, such as gas turbines and I.C. engines
(17 percent).  The most widely applied NOX reduction technique is modification of operating
                                                 S-7

-------
                                  TABLE S-4.   OVEltALL EVALUATION Of  NOX  CONTI-OL TECHNIQUES FU'< CWil'STUih SCtPr.fS
Control
Technique
Low excess air
(LEA)
Flue gas reclrcu-
latlon (FGR)
Off-stokhlometrlc
combustion (OSC)
Incl. OFA, BOOS.
B8F
load reduction
Burner
modifications
Existing
Applications
Retrofit and new
utility boilers;
iome use In Indus-
trial boilers
Retrofit use on many
gas- and oil-fired
utility boilers;
demonstrated on some
types of industrial
boilers
New and retrofit use
on many utility
boilers; demonstrated
on some types of
industrial boilers
Some retrofit use
on gas and oil
utility boilers;
enlarged fireboxes
on new coal units
New and retrofit
use on utility
boilers; demon-
strated on resi-
dential furnaces
Effectiveness
10X to 30X for
thermal and fuel
N0x .
20X to SOX for
thermal ML; no
effect on fuel
N0x
20% to SOX for
thermal and fuel
NOX
Of to 40X for
thermal N0x
30X to 60X for
thermal and fuel
N0x
Operational
ImpJCt
Increase In effi-
ciency; amount lim-
ited by smoke or
CO at very low EA
Possible flame In-
stability; In-
creased vibration
No major impact
with new design;
potential for flame
instability, effi-
ciency decrease,
increased corrosion
(coal-fired) with
retrofit
Decrease in effi-
ciency and power
output; limited by
spare capacity and
smoke formation
No major impact
with new design;
retrofit use con-
strained by firebox
characteristics
Projected
Applications
Widespread use for
efficiency In-
crease; Incorpor-
ate into advanced
designs all sources
Possible use in new
Industrial boiler
designs
Widespread use in
large boilers; In-
corporate Into ad-
vanced designs
Enlarged fireboxes
used In new unit
design; limited use
for retrofit
Incorporate Into
advanced designs
utility, Industrial
boilers, residen-
tial , process fur-
naces, GT; combine
with OSC
Control
Evaluation
Primary emphasis near-
term and far-tern appli-
cations (all sources);
combined with OSC & bur-
ner mods for far-ten.i appl.
Primary emphasis near-terra
applications larg^ boilers;
possible far-term Industrial
boiler application
Primary titphasls near-
ter.n i.nd far-t£rM appli-
cations all sources
Secondary emphasis near-
term applications (boilers);
combined with OSC or burner
mods for far-term appl.
Primary emphasis near- and
far- term applications all
sources
LO
00

-------
                               TABLE S-4.
OVERALL EVALUATION OF NOX CONTROL TECHNIQUES FOR COMBUSTION SOURCES
Control
Technique
Water, s learn
Injection
Reducrd air
preheat (RAP)
Aramnia injection
Fuel
denltriffcatlon
Fuel additives
Alternate and
mined fuels
Existing
Appl ic.it ions
Widely used for gas
turbines
Widespread use ir.
large turbocharged
1C engines
Demonstrated on
oil* and gas-fired
Industrial boilers
Oil denUrificatlcm
accompanies rfesul-
furlwtlon for some
large boilers
Fuel additives for
HOX not used
Combustion of low
nitrogen alternate
fuels being
demonstrated
Effectiveness
301 tr. 90S for
thermal NO^
101 to 401 for
thermal N0x
40S to 70S for
thermal and fuel
N0x
10X to 40f for
fuel N0x
Generally In-
effective for dir-
ect HOX reduction
Varies
Operational
Impact
Slight decrease in
efficiency; limited
by CO formation;
power output
increases
S11(|lit decrease In
efficiency, In-
crease power output
Retrofit use lim-
ited; possible ad-
verse environmental
impact
No adverse effects
Byproduct emissions
formed
Varies
Projected
Applications
Use in new gas tur-
bines; possible use
in process furnaces
Continued use in 1C
engines, applicable
to some industrial
arid utility boilers .
Use In large
boilers In some
areas (19BO's)
Use of oil de-
nitrification in
large boilers AS
supplement^ CM
tech.
Additives for cor-
rosion, fouling,
participate, smoke,
etc. can provide
increased fl«»1-
billty with CM tech.
on large boilers
Combined cycles and
residential ami
ccnmerctal heating
sys terns
Control
Evaluation
Primary emphasis near-term
applications, gas turbines;
possible far-term indus-
trial process application
Secondary emphasis
»
Primary emphasis far-
term application to large
boilers; evaluate impact
with coal firing
Secondary emphasis; eval-
uate as alternate fuel
Secondary emphasis; con-
sider impact of additives
Secondary emphasis far- term
application; evaluate dif-
ferential impact of fuel
switching; transfer results
of other E/A's .
in
vo

-------
                              TABLE S-4.
OVERALL EVALUATION OF N0x CONTROL TECHNIQUES  FOR COMBUSTION  SOURCES  (Continued)
Control
Technique
Catalytic
combustion
Fluidized bed
coffimstion
Flue gas
treatment (FGT)
Existing
Applications -
Only tested In
experimental
combustors
Tested in pilot/
prototype
combustors
Used in Japan on
large boilers
Effectiveness
>90X for thermal
N0x
20X to 50X for
fuel NOX (pres-
surized FBC)
40X to >90t for
fuel and
thermal NO
X
Operational
Impact
Requires clean
fuel; combustors
1 imited by cata-
lyst bed temp.
capability
Requires sulfur
acceptor
Requires temp, con-
trols, catalyst,
scrubbing soTn. ,
or oxidizing agent;
possible adverse
envlronn^riUl Impact
Projected
Applications
Gas turbines and
residential and
commercial heating
sys terns
Combined cycle,
utility boilers,
industrial boilers
(1980's)
Possible supple-
ment to CM for
utility and large
industrial boilers
(1980's)
Control
Evaluation
Primary emphasis far-term
applications; compare im-
pact to burner mods, al-
ternate fuels
Transfer results from
FBC E/A; compare impact
to combustion modifications,
conventional combustion
Secondary emphasis; trans-
fer results of other
studies to compare impact
to combustion mods
I
o

-------
conditions.  For utility boilers and large industrial  boilers, techniques such as lowering excess
air, off-stoichiometric or staged combustion, and, for gas- and oil-fired units, flue gas recircula-
tion have resulted in NOV reductions of up to 60 percent making it possible for them to meet •
                        A
emissions regulations at costs typically ranging from $0.60 to $6 per kW (electric output, 1978$).
The incremental costs of new burner and enlarged furnace designs installed on new units also fall
within this range.  Ongoing performance tests are investigating potential side effects of the
modifications, such as increased corrosion and particulate emissions with coal firing.

       Although less well developed for most industrial boilers, some combustion modifications for
these 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 combus-
tion, and flue gas recirculation.

       The energy impacts of applying combustion modification NO  controls to utility and industrial
                                                                A
boilers occur largely through the effects on unit fuel-to-steam efficiency.  This is usually
expressed as an increase or decrease in fuel consumption for a constant output.  Generally, low
excess air, flue gas, recirculation and off-stoichiometric combustion have very little effect on
efficiency.  Low excess air controls actually improve fuel efficiency in many cases.  In some cases,
taking burners out of service may result in reduced capacity.  Reduced air preheat has 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 NOX 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 NO
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 NOX 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
NOX up to 90 percent at costs of 0.4 to 14 mills/kWh (1975 costs), 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.
                                              S-ll

-------
       The energy impacts of applying NOV control to internal  combustion engines and gas turbines
                                        X
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.

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 5 percent of the
nation's stationary NO  emissions.  Emissions of CO and particulates from the major equipment types,
residential and commercial warm air furnaces, can be controlled by burner maintenance and tuning.
These techniques are not very effective for NOX reduction, however.   The most promising prospect for
HO  control in space heating systems is for new equipment applications.  New low NO  systems are
  A                                                                                A
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.

       There has been negligible application of combustion modification to incineration and open
burning.

NONCOMBUSTION PROCESSES

       Noncombustion-generated NO , 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
noncombustlon sources are mainly those that use nitric acid as a feedstock.  Control methods are
similar to those used for nitric acid manufacturing,  fable S-5 gives a summary of tail gas
abatement processes and applications.
                                                S-12

-------
                                        TABLE S-5.   NOX ABATEMlNT METHODS ON NEW OR EXISTING NITRIC ACID PLANTS
Process
Chilled
Absorption
Extended
Absorption
«
Wet Chemical
Scrubbing
Method
Increased solu-
bility of NOX In
chilled water
Increased absorp-
tion of NOX by
additional ab-
sorption equip-
ment
Scrubbing tall
gases with urea
solution or
ammonia to
recover NO
Comments
Usually cannot meet
NSPS without other add-
on technology or low-
ered acid product
concentration
Inlet pressure of 760
kPa required (addi-
tional compressors
may be required)
Requires additional
compressor

Performs better at high
pressure but operable at
lower pressures. Recovers
ah.-noniuni nitrate and urea
solution. Requires re-
frigeration.
May require an evaporator
to produce a concentrated
ammonium nitrate by prod-
uct. No refrigeration
required.
Licensors
COL-VITOK
TVA
J. F. Prltchard
(Grande Paroisse)
P. M. Weatherly
Chemlco
Uhde
C&l Glrdler
CoFAZ
MASAR (urea
scrubbing)
Norsk Hydro
(urea scrubbing)
Goodpasture
(ammonia
scrubbing)
Examples
2-318 Mg/d (350 tons/day) (with Gulf cata-
lytic reduction add-on). Nltrant, Tampa. Fla.
2-50 Mg/d (55 tons/day) plants at Muscle
Shoals, (1972)
327 Mg/d (360 tons/day) plant, Miss.
Chemicals, Yazoo City, Miss. 1973.
272 Mg/d (300 tons/day) Holston Army
Ammunition Plant, Klngsport, Tenn.
Comlnco Plant, Beatrice Neb.
Kaiser. Tampa. Fia. pnq Bainbridae. Ohio
9 U.S. plants, 1 Japan plant (employs
chilled absorption process)
900 Mg/d (1000 tons/day) Monsanto,
Pensacola, Fla. 1977.
250 Mg/d (275 tons/day) plant, Allied
Chemical, Oman, Neb. 1975
None built to date
111. Nitrogen PH., Marsalles, 111.
Air Products & Chem. , Pace, Fla.
Norsk Hydro, Proggunn, Norway
90 Mg/d (100 tons/day) Goodpasture pit.,
1974. Dlnmitt, Texas
Chevron 011 Co., Richmond, Calif. 1976
C.F. Industries, Fremont, N.D.
2 scrubbers for 7 plants totalling 544
mg/d (000 tons/day),. Cyanamid, Welland, Ont.
I
_4

oj

-------
                                   TABLE S-5.   HOX ABATEMENT METHODS W HEH OR EXISTING NITRIC ACID PLANTS (Concluded)
Process
Catalytic
Nonselectlve
Catalytic
Selective
Heterogeneous
Catalysis
Chemical
Absorption
Holecul-ar
Sieve
Method
Burns NOX and 02
with Cfy or Kg to
form N2. H20, C02
Burns NO* with
ammonia to form
Ng and H20; 02
not affected '
Oxidation of NO +
N02 catalyzed by
heterogeneous ca-
talysis before
absorption
Oxidation with
KMnO^ (KMn04
electrolytlcally
reclaimed
Absorption by
molecular sieve,
regeneration of
the sieve by
thermal cycling
Conrnents
• Consumes natural gas,
uneconomical If high
NOX or 02 content
(also reacts with 02)
• May be used In con-
junction with extended
absorption
• Energy recovery
possible
• Works at low or high
pressure
• Uses ammonia, can be
expensive to operate
• Often used with ex-
tended absorption
• Works at low or high
pressure
a Energy recovery
usually not possible
* Can achieve very low
emission If desired
Limited success
Uneconomical not pres-
ently offered
o High eneryy and capi-
tal dem.tn.ls
a Hard to fit cycl Ing
of sieve Into con-
tinuous plant opera-
tion
Licensors
CM Glrdler
D. M. Weatherly
Chemlco
Gulf
Uhde (BASF cata-
lysts)
Mitsubishi
CDL/VITOK
Carus Chemical
Purasiv N
(Union Carbide)
Examples
Ol1n» Lake Charles, La. (also, Heatherby
plants)
IMC Corp., Strellngton, La. (1976) (with
extended absorption). 817 Mg/d (900 tons/
day), 1977. Columbia Nitrogen, Augusta,
Ga.
Location not available
Nltram plants 1n Tampa, Fla., Installed
after CDL/VITOK process.
10 plants 1n U.S.
Plants 1n Europe and Jap'an
Under development
2 plants 1n Jaoan, not currently offered
1n U.S.
50 mg/d (55 tons/day) Hercules, Inc.
Berseimrr, Ala. 1971
50 mg/d (6'j tons/day) U.S. Army, Hols ton, '
Kingston, Tenn. (Inoperable, disiiuntleil)
i
	 	 	 	 	 	 1
I
.£.

-------
                                              SECTION 1
                                            INTRODUCTION


     The first edition of "Control Techniques for Nitrogen Oxide from Stationary Sources" (AP-67)
was published in March 1970 as provided by the Air Quality Act of 1967.  This first edition was used
to support the National Ambient Air Quality Standards (NAAQS) for Nitrogen Dioxide as required .by
Sections 108 and 109 of the Clean Air Amendments of 1970.  The first control techniques document was
updated in January 1978 to include later developments in regulatory control of NOX and emission
control techniques for stationary sources.  Since the second edition, still further developments
have occurred.  These developments have been largely in the field of control of NO  emissions from
combustion.  Consequently, the current revisions have been limited to the combustion sections
leaving the other parts of the text unchanged except for the addition of text to show the dates of
the various cost data.
       The NO  control technology development to support the implementation of the 1971 NAAQS
             X
standards has shown widespread advancement since the publication of the original 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.  Work is continuing in these areas.
       Combustion process modification is the commonly used method for control of stationary
combustion sources,accounting for 98 percent of stationary source NO  control.  Process
modifications have been extensively 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 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 heating systems.
       Stack gas treatment is the commonly used method for control of NO  emissions from stationary
noncombustion 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
control techniques, including extended absorption, catalytic reduction, wet scrubbing, and molecular
                                              1-1

-------
sieve absorption, have been developed and implemented on existing and new equipment.   Reductions  in



NOV in excess of 95 percent have been demonstrated.
  X




       The purpose of this report is to update and revise the control techniques document issued  in



1978 by incorporating improved emissions estimates and NO  control  technology developments since



that time.  Emphasis is placed on identifying the significant stationary sources of NO  emissions,



based on the most recent EPA emissions data (Section 2); summarizing the developmental  status  of



candidate NO  control techniques (Section 3); and reviewing the effectiveness, cost and user



experience with the implementation of NOV controls on large combustion sources (Section 4), other
                                        A


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 Clean Air Act as amended in 1977.
       This report is concerned only with the quantifying and controlling stationary source NOV
                                                                                              A


emissions.  The effects upon health and welfare of nitrogen oxides and their secondary atmospheric



reaction products are considered in two related documents, "Air Quality Criteria for Ozone and Other



Photochemical Oxidants," (EPA-600/8-78-004) and AP-84, "Air Quality Criteria for Nitrogen Oxides."



The foregoing documents are being revised with publication scheduled within the next few months.
                                         1-2

-------
                                              SECTION 2
                                  CHARACTERIZATION OF NOX EMISSIONS

       This section presents a nationwide inventory of emissions of oxides of nitrogen.  Section 2.1
defines NO  and summarizes the basis of its occurrence in stationary source combustion.  Section 2.2
describes the standard EPA methosd for analysis, of source and ambient NO  concentrations.
                                                                       X
Section 2.3 describes specific stationary source equipment types and presents nationwide NO
emissions estimates and NO  emissions factors for each source equipment type.
                          X

2.1    DEFINITIONS AND FORMATION THEORY

       Seven oxides of nitrogen are known to occur: NO, NOg, NO,, NgO, NpO,, N-O. and NgOg.  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,
"NOX" refers to either or both of these two gaseous oxides of nitrogen.  Nitrogen dioxide is
deleterious to human respiratory functions and is a key participant in the formation of photo-
chemical smog.  Nitric oxide, taken alone, is relatively less harmful but is important as the main
precursor to NO- formation in the atmosphere.
       Approximately 95 percent of oxides of nitrogen from stationary combustion sources are emitted
as nitric oxide.  Two separate mechanisms, thermal NOX formation and fuel NO  formation, have
been identified as generating NO  during fossil fuel combustion.
       Thermal NOX results from the thermal fixation of molecular nitrogen and oxygen in the combus-
tion air. ' Its rate of formation is extremely sensitive to local flame temperature and somewhat less
so to local oxygen concentrations.  Virtually all thermal NO  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 fay the thermal quenching of the combustion gases.  The
flue gas NOX concentrations are therefore between the equilibrium level characteristic of the peak
flame temperature and the equilibrium 'level at the flue gas temperature.  This kinetically
controlled behavior means that thermal NO  emissions are dominated by local combustion conditions.
                                                 2-1

-------
       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 the local oxygen concentration in particular.  The flue gas NO  concen-
tration due to fuel nitrogen is typically only a fraction (e.g.i 20 to 60 percent) of the level
which would result from complete oxidation of all nitrogen in the fuel.  Thus, fuel NO  formation,
like thermal NO  formation, is dominated by the local combustion conditions.  Additionally, fuel  NO
               X                                                                                   A
emissions are dependent on the nitrogen content of the fuel.  The NO  emissions characterization
detailed in this section, therefore, takes account of variations in equipment operating conditions
and 1n fuel type which influences the emissions as well as the potential for control.  Additional
discussion on thermal and fuel NO  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 NO-.  The NO- concentration in the process vents is
typically 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
                                              X
less sensitive to minor process modifications than are combustion generated NO  emissions.

2.2    SAMPLING AND ANALYSIS METHODS

       The standard EPA method for compliance testing of NO  from stationary sources is the phenol-
disulfonic acid (PDS) method.  This method was developed for the measurement of nitrate i-n 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 (NgO).  The sample is then processed following the procedures of Reference 2-2.
The absorbence of 420 nm wavelength light by the treated samples is then measured.  A calibrated
relationship between absorbence and NO, concentration is used to relate the measurement to the
sample NOg 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
disadvantages are the long time elapsed between samples, a possible interference from halides, and
the Inherent problems with grab sampling.
                                                  2-2

-------
       Continuous type instrument methods are also used to measure NOX concentrations.  The most
common type of instrument method is the chemiluminescence method.  This method is described in
Reference 2-3.  Performance specifications and specification test procedures for monitors of NO
                                                                                               A
emissions are given in Performance Specification 2, Appendix B, Part 60, Title 40, Code of Federal
Regulations.

2.3    EQUIPMENT DESCRIPTIONS, EMISSIONS ESTIMATES, AND EMISSION FACTORS BY APPLICATION SECTOR

       An overview of stationary sources of NOV emissions is provided in Figure 2-1.  The first
                                              A
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 the petroleum and related products industries have
generally 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, primarily in urban areas.  Noncombustion sources are largely within the area of
chemical manufacture, more specifically nitric and adipic acids and explosives.  The final descrip-
tion level in Figure 2-1 gives the important equipment types.  Although these equipment categories
do not include all the possible variations or hybrid units, the bulk of the equipment is included in
the breakdown.                      ,
       Table 2-1 summarizes annual nationwide emissions of nitrogen oxides from all sources for the  •
period 1970 to 1980 (Reference 2-4).  As shown, transportation (or mobile sources), stationary
source fuel combustion, and industrial processes are the major NO  emission sources.  In the
following sections, equipment descriptions, nationwide NO  emissions estimates, and NOV emission
                                                         A                            A
factors will be presented for sources in the stationary fuel combustion and industrial processes
categories.  Transportation NOX sources are not addressed in this document.                        .„

       Particular emphasis is placed in this report on stationary fuel combustion sources because of
their large contribution to stationary source NO  emissions.  Table 2-2 presents a breakdown of
                                                «
estimated 1980 nationwide NO  emissions from stationary fuel combustion sources by fuel type and
consuming sector (Reference 2-4).  The coal-fired electric utility sector and the gas-fired
                                               2-3

-------
                   APPLICATION
          SECTOR
              EQUIPMENT TYPES
               -PRIME HOVER
STATIONARY
SOURCES OF
NO-
                  ,   STEAM
                   GENERATION
                                         ELEC POWER
                                         GENERATION
                                         INDUSTRIAL
                                       PROCESS STEAM
                    SPACE
                  'HEATING
 ELEC POWER GEN.
 OIL AND GAS
 PIPE LINE PUMPING
 HAT GAS PROCESSING
                                    —  RESIDENTIAL
                                   1—  COMMERCIAL

                                      -  MUNICIPAL
               — INCINERATION
L_  INDUSTRIAL.

     PETROLEUM
  "*  REFINING
                     PROCESS
                     HEATING
   METALLURGICAL
               I—NONCOMBUSTIQN
                                         CERAMICS
                                    I	  CHEMICALS
                                           GLASS
                                          BRICKS
                                          CEMENT
     CHEMICAL
   MANUFACTURING
                             FIELD ERECTED
                           WATEHTU8E BOILERS
                                              r FIS-B ERECTED
                           WATERTU8E BOILERS H
                                              L PACKAGED
                                                                  PACKAGED
                                                              FIRETUBE BOILERS
                                                              RECIPROCATING
                                                                1C ENGINES
                                                              •GAS TURBINES

                                                              . FURNACES
                         .CAST IRON BOILERS

                        ,—   BOILERS



                        —  FURNACES
                          -HATERTUBf

                            FIRETUBE

                          L.CA5T IRON
-r
PROCESS HEATERS

FLUID CATALYTIC CRACKERS
    -KEATING AND ANNEALING OVENS

    —   COKE OVEN UNOERFIRE

    —   OPEN HEARTH FURNACE

    _    S UTTERING OVENS
      .  FURTMCES

    r— MITKIC ACID

      • AOIPIC ACID

    L. EXPLOSIVES
         Figure  2-1.   Stationary sources  of NO  emissions.
                                                       *\
                                             2-4

-------
              TABLE 2-1.   NATIONAL ESTIMATES OF NITROGEN OXIDE EMISSIONS, Tg (106 Tons)/yr (Reference 2-4)
Source Category 1970
Transportation
High Vehicles 5.5(6.1)
Aircraft 0.1 0.11
Railroads 0.6 0.66
Vessels 0.1(0.11
Other Off-Highway 0.8(0.88
Vehicles
Transportation Total 7.1(7.8)
1971 1972

5.9 6.5) 6.5(7.2)
0.1 0.11) 0.1(0.11
0.6 0.66) 0.7(0.77
i 0.1 0.11) 0.1(0.11
0.8 0.88) 0.9(0.99

7.5(8.3) 8.3(9.1)
1973

6.8 7.5)
0.1 0.11
0.7 0.77
0.1 0.11
0.90.99

8.6(9.5)
1974 1975 1976 1977

6.5(7.2) 6.6 7.3) 6.9 7.6) 7.0 7.7)
0.1(0.11) 0.1 0.11) 0.1 0.11) 0.1 0.11
0.7(0.77) 0.7 0.77) 0.7 0.77) 0.7 0.77
0.1(0.11) 0.1 0.11) 0.1 0.11) 0.1 0.11
0.9(0.99) 0.9 0.99) 1.0 1.1) 0.1 1.1)

8.3(9.1) 8.4(9.2) 8.8(9.7) 8.9(9.8)
1978

7.1
0.1
0.7
0.2
1.1


7.8)
0.11
0.77
0.22
1.2)

1979 1980

6.8(7.5) 6.7
0.1(0.11) 0.1
0.8(0.88) 0.8

7.4)
0.11)
'0.88)
0.2 0.22) 0.1(0.11)
1.1(1.2) 1.0(1.1)


9.2(10.1) 9.0(9.9) 8.7(9.6)
Stationary Source Fuel Combustion
Electric Utilities 4.5(5.0)
Industrial 3.9 4.3)
4.7(5.2) 5.0(5.5)
3.8 4.2) 3.9(4.3)
5.3(5.6)
3.9(4.3)
5.3(5.8) 5.2 5.7) 5.6(6.2) 6.0 6.6)
3.7(4.1) 3.4 3.7) 3.7 4.1) 3.7 4,1)
5.9(6.5}
3.7(4.1)
6.2(6.8) 6.4
3.5(3.9) 3.0
Comnercial- 0.3 0.33) 0.3 0.33) 0.3(0.33) 0.3(0.33) 0.3(0.33) 0.3 0.33) 0.3 0.33) 0.3 0.33) 0.3(0.33) 0.3(0.33) 0.3
Institutional

Residential 0.4(0.44) 0.4(0.44) 0.4(0.44
Fuel Combustion Total 9.1(10)
9.2(10.1) 9.6(10.6





7.0)
3.3)
0.33)

0.4(0.44) 0.4(0.44) 0.4(0.44) 0.4(0.44) 0.4(0.44) 0.4(0.44) 0.4(0.44) 0.3(0.33)
' 9.9(10.9) 9.7(10.7) 9.3(10.2) 10.4(11.4) 10.0(11 .0)
10.3(11.3) 10.4(11.4) 10.0(11)
Industrial Processes 0.7(0.77) 0.7(0.77) 0.7(0.77) 0.7(0.77) 0.7(0.77) 0.7(0.77) 0.7(0.77) 0.7(0.77) 0.7(0.77) 0.7(0.77) 0.7(0.77)
Solid Haste Disposal
Incineration 0.1 0.11
Open Burning 0.3 0.33
Solid Haste Total 0.4 0.44
liscellaneous
Forest Fires 0.2 0.22
Other Burning 0.1 0.11
Misc. Organic 0.0 0)
Solvent

0.1(0.11) 0.1(0.11
0.20.22) 0.1(0.11
0.3 0.33) 0.2(0.22


0.0(0)
0.1 0.11
0.1 0.11


0.0(0) 0.0(0) 0.0 0) 0.0 0)
0.1(0.11) 0.1 0.11) 0.1 0.11) 0.1 0.11
0.1(0.11) 0.1 0.11) 0.1 0.11) 0.1 0.11



0.0(0)


0.0(0) 0.0(0)
0.1(0.11) 0.1(0.11) 0.1(0.11)
0.1(0.11) 0.1(0.11) 0.1(0.11)




0.2(0.22) 0.2(0.22) 0.1(0.11) 0.2(0.22) 0.1(0.11) 0.2(0.22) 0.2(0.22) 0.2(0.22) 0.2(0.22) 0.2(0.22)
0.1(0.11) 0.1(0.11) 0.0 0)
0.0 0) 0.0(0) 0.0 0)


0.0(0) 0.0(0) 0.0(0) 0.0 0)
0.0(0) 0.0(0) 0.0(0) 0.0 0)

0.0(0)
0.0(0)


0.0(0) 0.0(0)
0.0(0) 0.0(0)


Miscellaneous Total 0.3(0.33) 0.3(0.33) 0.3(0.33) 0.1(0.11) 0.2(0.22) 0.1(0.11) 0.2(0.22) 0.2(0.22) 0.2(0.22) 0.0(0.22) 0.2(0.22)
Total of All 17.6(19.4) 18.0(19.8) 19.1(21.0) 19.4(21.3) 19.0(20.9) 18.6(20.5) 19.8(21.8) 20.3(22.3) 20.5(22.6) 20.4(22.4) 19.7(21.7)
Categories







ro
i

-------
               TABLE 2-2.  1980 NITROGENOXIDE EMISSIONS  FROM FUEL COMBUSTION IN STATIONARY SOURCES
                                BY CATEGORY FOR MAJOR FUELS, 103 tons  ) (Reference 2-4)
Fuel Type

Anthracite Coal
Bituminous Coal
Residual Oil
Distillate Oil
Natural Gas
Wood
Kerosene
LPG
Coke
Coke-Oven Gas
Bagasse
TOTAL
Electric
Utility

8.6 (9.5)
5149.4 (5664)
405.5 (446)
38.2 (42)
775.3 (853)
0.0 (0)
0.0 (0)
0.0 (0)
0.0 (0)
0.0 (0)
0.0 (0)
6377.0 (7015)
Industrial

4.8 (5.3),
398.6 (438)
162.3 (179)
57.4 (63)
2239.7 (2464 )a
35.1 (39)
11*4 (12.5)
32.6 (36)
16.5 (18.2)
5.0 (5.5)
5.2 (5.7)
2968.6 (3265)
Commercial-
Institutional

2.2 (2.4)
14.8 (16.3)
103.9 (114)
37.1 (41)
118.4 (130)
0.0 (0)
0.0 (0)
0,0 (0)
O.Q (0)
0.0 (0)
0.0 (0)
276.4 (304)
Residential

1.5 (1.7)
3.0 (3.3)
0.0 (0)
77.0 (85)
215.5 (237)
6.8 (7.5)
9.6 (10.6)
17.5 (19.3)
Q.O (0)
0.0 (0)
0.0 (0)
330.9 (364)
Total

17.1 (18.8)
5565.8 (6122)
671.7 (739)
209.7 (231)
3348.9 (3684)
41.9 (46)
21.0 (23)
50.1 (55)
16.5 (18.2)
5.0 (5.5)
5.2 (5.7)
9952.9 (10,948)
..                  33
 Includes 1884 x 10  Mg (2072 xv!0  tons) from  stationary internal  combustion engines and gas turbines.

-------
industrial sector are the major emission sources in the fuel  combustion category.   In Section 2.2.1
the operations of the boiler sources in the fuel combustion category are described and NO  emission
factors are presented for the various equipment types within each use sector.

       In Section 2.3.2 the operations of the sources that constitute the industrial  processes
category are briefly discussed.  Nationwide NO  emissions estimates for 1980 and NO  emission
factors are presented for each major Industrial source.  Section 2.3.3 presents nationwide emissions
estimates and emission factors where available for the minor industrial process sources of nitrogen
oxides.  The nationwide NOX emissions estimates are. derived from unpublished data collected by the
Monitoring and Data Analysis Division (MDAD) of EPA.  The NO  emission factors were obtained from
the EPA publication AP-42 (Reference 2-5) and its six supplements (References 2-6 to 2-11).

2.3.1  Stationary Fuel Combustion Sources

2.3.1.1  Utility Boilers

     Most of the nation's electricity is generated in large fossil fueled central  station power
plants, which primarily consist of high-pressure field-erected watertube boilers in the 100 to
1300 MW  range serving turbine generators.  On a thermal input basis, these boilers range in
capacity from about 300 to 3700 MW (1,000 to 13,000 x 10  Btu/hr).  Industrial electric generating
boilers will generally be smaller than the sizes identified in this range as indicated in the next
subsection.  Field-erected watertube boilers operate at steam temperatures up to 570°C (1050°F) and
steam pressures up to 26 MPa (3800 psi).  Depending upon manufacturer, units greater than about
2250 MW (7700 x 10  8tu/hr) thermal input operate at supercritical steam pressures above 24 MPa
(3500 psi) (Reference 2-12).  In general, utility boiler thermal efficiencies range up to 90 percent
of the heat liberated during fuel combustion.  Approximately half of this heat energy is absorbed by
radiant heat transfer to the furnace walls.  However, because of the various thermodynamic cycle and
mechanical losses, total power plant fuel-to- electric efficiencies are considerably lower, around
34 to 38 percent.

       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
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.  The major firing modes are: single- or
opposed-wall-fired, tangentially-fired, turbo-fired, and cyclone-fired.  Vertically-fired units and
                                                   2-7

-------
stoker units are used to a small extent in older steam generating stations.   All  of the major firing
types can be designed to burn fossil fuels - gas, oil and coal, either singly or in combination.
However, the cyclone unit is primarily designed to fire coal as the principal fuel.  All of the
coal-fired units use pulverized coal except for the cyclone units which use crushed coal and the
stokers which accept' lumps of coal.

       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
temperatures below the ash-fusion temperature, and ash is removed as a solid.  For wet-bottom
furnaces the ash is removed as a molten slag through a bottom tap.  Although wet bottom units were
once used extensively in burning low ash-fusion temperature coals, they are less frequently used due
to operational problems with low sulfur coals and because their high combustion temperatures promote
HOV formation.
  A
       In single-wall firing (front-wall) burners are mounted normal to a single furnace wall.
Furnace wall area generally limits the capacity of these units to about 1200 MW (4100 x 10  Btu/hr)
thermal input.  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 thermal input (Reference 2-12).  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
introduced 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 sufficient turbulence for
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 NOX formation
accompany the high temperatures of slagging operation.  Due to the inability of this design to
readily adapt to low NO  operation, this type of furnace is no longer being  constructed.
                                                2-8

-------
       Vertical-firing furnaces were developed for pulverized fuels prior to the advent of water-:
walled 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 relatively
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
stationary grate through which ash falls or a moving grate which dumps the ash into a hopper.   The
main types of stokers are overfeed and underfeed designs.  Spreader stokers are overfeed designs and
distribute the fuel by pneumatically or mechanically projecting the fuel into the furnace where it
falls evenly over the fuel bed.  Other overfeed stokers generally deposit fuel on a continuously
moving grate.  Underfeed designs introduce fuel beneath the fuel bed.  Ash is pushed aside by  the
newly introduced fuel.

       Tangential firing, single-wall and horizontally opposed-wall firing, and turbofurnace firing
accounted for about 40 and 36 and 14 percent of the fuel consumed by utility boilers in the
mid-1970s (Reference 2-13). t In terms of numbers of units, their distribution was estimated at 19,
59 and 8 percent, respectively.  Cyclone, vertical and stoker designs make up the remainder of
utility units.

       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.
Historically, the trend was toward increasing unit capacities.  However, this appears to have  slowed
in recent years with many utilities electing to install two small boilers rather than a single
larger unit (Reference 2-12).

       Nationwide NOX emissions estimates for 1980 from electric utility boilers are shown in
Table 2-2 according to the type of fuel consumed (Reference 2-4).  Tables 2-3 to 2-7 present NO
                                          •                                                     X
emission factors that are applicable to utility boilers burning coal, oil, and gas
(Reference 2-11).  For utility boilers Figure 2-2 should be used in conjunction with Table 2-7 to
determine NOX emissions from natural gas combustion at reduced boiler loads.
                                               2-9

-------
      TABLE 2-3.  NITROGEN OXIDE EMISSION FACTORS FOR BITUMINOUS AND
                  SUB-BITUMINOUS COAL COMBUSTION3, (Reference 2-11)
     Firing Configuration
        NOX Emission Factors
                                        kg/Mg of
                                      coal burned
                       (Ib/ton of
                        coal  burned)
Pulverized Coal Fired

     Dry bottom
     Wet bottom
Cyclone Furnace6
Spreader Stoker

    "Uncontrolled
     With flyash reinjection
     No flyash reinjection


Overfeed Stdkerd'f

     Uncontrolled
     After multiple cyclone


Underfeed Stoker9

     Uncontrolled
     After multiple cyclone


Handfired Units
 10.5 [7.5]c
 17
18.5
  7
  7
  7
  3.25
  3.25
  4.75
  4.75
  1.5
 21)  ([15])
181
 37
(14)
(14)
(14)
(7.5)
(7.5)
(9.5)
(9.5)
(3)
 Factors represent uncontrolled emissions unless otherwise specified.

 Total nitrogen oxides expressed as N02.  To express these factors as NO,
 multiply by a factor of 0.66.  All factors represent emissions at 60 to
 110% load conditions.

cBracketed value is for tangentially fired units.

 Includes traveling grate, vibrating grate, and chain grate stokers.

eUsed primarily in utility and large industrial applications.

 Used primarily in large commercial and general industrial applications.

9Used primarily in commercial and domestic applications.
                                     2-10

-------
         TABLE 2-4.   NITROGEN OXIDE EMISSION FACTORS FOR ANTHRACITE
                     COAL COMBUSTION WITHOUT CONTROL EQUIPMENT,
                     (Reference 2-11)
Type of Boiler
NOV Emission
A
kg/Mg of
coal burned
Factors
(Ib/ton of
coal burned)

Pulverized Coal Fired
Traveling Grate Stoker
Handfired Units
9
5
1.5
(18)
(10)
(3)
          TABLE 2-5.   NITROGEN OXIDE EMISSION FACTORS FOR LIGNITE
                      COMBUSTION WITHOUT CONTROL EQUIPMENT,
                      (Reference 2-11)
Type of Boiler


NOV Emission
A
kg/Mg of
fuel burned
Factors9
(Ib/ton of
fuel burned).

Pulverized Coal Fired
Dry Bottom
Front Wall or Horizontally
Opposed Wall
Tangential
Cyclone Furnace
Spreader Stoker
Other Stokers
6b

7
~
(4)
8.5
3
3
(12)b

(14)

(8)
(17)
( 6)
(6)
 Total  nitrogen oxides expressed as N09.
h
 Nitrogen oxide emissions may be reduced  by 20 to 40 percent with low
 excess air firing and/or staged combustion in. front fired and opposed  wall
 fired units and cyclones.
                                    2-n

-------
         TABLE 2-6.  NITROGEN OXIDE EMISSION FACTORS FOR FUEL OIL
                     COMBUSTION WITHOUT CONTROL EQUIPMENT,
                     (Reference 2-11)
          Boiler Type
   NO.. Emission Factors
                                        kg/10-3 liters
                                        of oil burned
                  (lb/103 gal of
                   oil burned)
UTILITY BOILERS - Residual Oil1

     Tangentially fired
     Vertical fired
     All others

INDUSTRIAL BOILERS

     Residual Oil
     Distillate Oil
                                          5
                                         12.6
                                          8
                                          6.61
                                          2.4
                     (42)
                    (105)
                     (67)
                     (55)c
                     (20)
COMMERCIAL BOILERS

     Residual Oil
     Distillate Oil
6.6
2.4
                                                                (55)
                                                                (20)
RESIDENTIAL FURNACES

     Distillate Oil
2.2
                                                               (18)
aTotal nitrogen oxides expressed as N02.
bFactors are for boilers at full load and normal ( 15%) excess air.  Several
 combustion mod.ifications can be employed for NO  reduction: (1) limited
 excess air can reduce NO  emissions 5-20%, (2) Staged combustion 20-40%,
 (3) using low NO  burners 20-50%, and  (4) ammonia injection can reduce NO
 emissions 40-70% but may increase emissions of ammonia.  Combinations of
 these modifications have been employed for further reductions in certain
 boilers.
GNitrogen oxides emissions from residual oil combustion in industrial and
 commercial boilers are strongly related to fuel nitrogen content, estimated
 more accurately by the empirical relationship:   ,                  ?
     kg N0?/103 liters = 2.75 + 50 (Mr [lb NO?/10J gal = 22 + 400(NT]
     where N is the weight % of nitrogen in the oil.  For residual oils
     having high ( Q.5 weight %) nitrogen content, use 15 kg N02/10  liter
     (120 Ib N02/10d gal) as an emission factor.
                                    2-12

-------
          TABLE 2-7.   NITROGEN OXIDE EMISSION FACTORS FOR NATURAL
                      GAS COMBUSTION WITHOUT CONTROL EQUIPMENT,
                      (Reference 2-11)
Furnace Type
NO Emission
kg/106 m3 of
gas burned
Factors3
(lb/106 ft3 of
gas burned)

Utility Boilers
Industrial Boilers
Domestic and Commercial Boilers
8800b
2240
1600
(550)b
(140)
(100)
 Total nitrogen oxides expressed as N02.  Test results indicate that 95 wt.
 percent of NO  is NO.

bUse 4400 kg/106 m3 (275 lb/106 ft3) for tangentially fired units.  At
 reduced loads, multiply this factor by the load reduction coefficient given
 in Figure 2-2.
                                    2-13

-------
   u
   1.0
UJ

g

u.
UJ
o
u
0.8






0.6





0.4






0.2
5
§
o
     40
               60
        80


LOAD, percent
100
110
    Figure 2-2.   Load reduction coefficient as  a

                  function  of boiler  load.

                  (Reference 2-11).
                             2-14

-------
2.3.1.2  Industrial Boilers

       This equipment category is comprised of industrial boilers ranging in capacity up to 250 MM
thermal input (850 x 10  Btu/hr).  Industrial boilers are either field-erected or packaged units.
The field-erected units are typically the units larger than 45 MW (150 x 10  Btu/hr) and are quite
similar in design and size to small utility boilers.  Field erected boilers are typically of the
watertube design.  Packaged boilers, which are equipped and shipped from the factory complete with
fuel burning equipment, are mainly watertube and firetube designs.  Other designs such as cast iron,
and shell type are also used in packaged designs.  Each of these designs has a fairly distinct
capacity range.  Packaged boilers far out-number field- erected units, but their combined fuel
consumption is 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 10  Btu/hr) are watertube
boilers.

     Population statistics for 1977 indicate that 74 percent of the industrial boiler capacity was
composed of watertube boilers and the remaining 26 percent were predominantly firetube boilers.  Of
the watertube boilers burning fossil fuels, 43 percent were predominately natural gas-fired,
32 percent were predominately oil-fired, 15 percent were stoker coal-fired, and 10 percent were
pulverized coal-fired.  Less than 5 percent of the industrial boiler capacity is supplied by
non-fossil fuel-fired boilers.
       In firetube boilers hot gases are directed from the combustion chamber through tubes which
are submerged in water.  Firetube boilers generally burn fuel oil and natural gas because the design
is particularly sensitive to fouling with ash-containing fuels.  Natural gas and distillate oil are
the main fuels for the smaller watertube units.  All fossil fuels are represented in the large
watertube industrial boiler category.,  Recent sales statistics indicate that the firetube has
diminished in sales in the past few years (Reference 2-14).
                                                                                      /
       Nationwide NO  emissions estimates for 1980 from industrial boilers are shown in Table 2-2
according to the type of fuel consumed (Reference 2-4).  Tables 2-3, 2-4, and 2-5 present NO
emissions factors that are applicable to industrial boilers burning coal.  Tables 2-6, 2-7, and 2-8
present NO  emissions factors for industrial boilers burning oil, gas, and liquefied petroleum gas
(LPG), respectively.  Emission factors for NO  from industrial boilers burning wood waste and
bagasse are presented in Tables 2-9 and 2-10 (References 2-5 and 2-11).
                                                2-15

-------
            TABLE 2-8.  NITROGEN OXIDE EMISSION FACTORS FOR LPG
                        COMBUSTION WITHOUT CONTROL EQUIPMENT*,
                        (Reference 2-11)
          Type of Source                     NOX Emission Factors

           and Fuel                     kg/103 liter of     (lb/103 gal of
                                        LPG burned           LPG burned)
Industrial Process Furnaces

     Butane                             '    1.58                (13.2)
     Propane                                1.49                (12.4)


Domestic and Commercial Furnaces

     Butane                 >                1.13                ( 9.4)
     Propane                                1.05                ( 8.8)
aLPG emission factors calculated assuming emissions are the same, on a heat
 input basis, as for natural gas combustion.

 Nitrogen oxides expressed as NOg.


         TABLE 2-9.  NITROGEN OXIDE EMISSION FACTORS FOR WOOD/BARK
                     WASTE COMBUSTION WITHOUT CONTROL EQUIPMENT,
                     (Reference 2-11)
Type of Source NOX Emission
kg/Mg of
fuel burned
Factors3
(Ib/ton of
fuel burned)

Industrial Boilers 0.7
(1.5)
aNitrogen oxides expressed as N02-
                                     2-76

-------
   TABLE 2-10.  NITROGEN OXIDE  EMISSION  FACTORS  FOR  BAGASSE  COMBUSTION
                WITHOUT CONTROL EQUIPMENT,  (Reference  2-11)
         Type of Source                     NOV Emission Factors
                                                  ,3
                                       g/kg (Ib/KHb) of  kg/Mg  (Ib/ton).
                                       steam generated     of fuel  burned
         Bagasse Boiler                  0.3 (0.3)           0.6  (1.2)
Emission factors are expressed in terms of the amount of steam produced.
These factors should be applied only to that fraction of steam resulting
from bagasse combustion.
Emissions are expressed in terms of wet bagasse, containing approximately
50 percent moisture, by weight.  About 2 kg (4.4 Ib) of steam are produced
from 1 kg (2.2 Ib) of wet bagasse.
                TABLE 2-11.  NITROGEN OXIDE EMISSION FACTORS
                             FOR RESIDENTIAL FIREPLACES,
                             (Reference 2-5)
               Fuel Type                      NO   Emission  Factors

                                        kg/Mg of            (Ib/ton of
                                        fuel  burned           fuel  burned)
              Wood                        0.5                    (1)


              Coal                        1.5                    (3)
                                    2-17

-------
2.3.1.3  Commercial and Residential Space Heating
       This category is made up of commercial and residential warm air furnaces and boilers.   Warm
air furnaces are subdivided into space heaters, where the unit is located in the room which it
heats, and 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 warm air heater equipment sector.  Combustion products pass through flue gas
passages of the heat exchanger and exit through a flue to the atmosphere.  Boilers used for
residential space heating are generally cast iron desigps.  Residential warm air furnaces and cast
iron boilers are available in sizes up to 0.12 MW (4 x 10  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 gas- and 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 10  Btu/hr) to
3.6 MW (12.5 x 106 Btu/hr).
       Fuels burned for residential and cnmmercial space heating are primarily natural gas and
distillate oil and much smaller amounts of coal, residual oil, LPG, and wood.  Nationwide NO
emissions estimates for 1980 from commercial and residential combustion sources are given in
Table 2-2 according to the type of fuel burned (Reference 2-4).  Emission factor data for NO
emissions from commercial and residential sources are presented in Tables 2-3, 2-4, 2-6 to 2-8, and
2-11.
2.3.1.4  Internal Combustion Sources
                                                                       *
       This stationary fuel combustion subcategory consists of stationary reciprocating internal
combustion engines and gas turbines.  Nationwide NOV emissions for these two sources in 1980 were
                                                   i\
                         "3              3
estimated to be 1884 x 10  Mg (2072 x 10  tons) as shown in Table 2-2.  Equipment descriptions and
N(L emission factors for the internal combustion engine and gas turbine sources are presented in  the
following sections.

       2.3.1.4.1  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.
                                               2-18

-------
       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 the 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.  Either 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
adjacent 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
installation time and the fact that they can be operated remotely.   Applications  range from shaft
power forjarge electrical generators to small air compressors and welders.

       Table 2-12 presents NOV emission factors for a heavy-duty, stationary reciprocating internal
                             /\
combustion engine that is firing natural gas.  These types of engines are generally used to power
pipeline compressors.  Emission factors are presented in terms of gas flow and  in terms of energy
produced (Reference  2-5).  NOX emission factors for large bore diesel and dual  fuel  engines are
given in Table 2-13  (Reference 2-11).

       2.3.1.4.2  Gas turbines.  Gas turbines are rotary internal combustion engines  fueled by
natural gas, diesel  or distillate fuel oils, and occasionally residual or crude oils.  These units
range in capacity from 30 kW (40 tip) 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 1090°C (2000°F).  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 (430 to 590°C, 800 to 1100°F)  to  preheat the inlet  air
between the compressor and the combustor.  This makes it passible 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.  The combined-cycle turbine is basically a simple-  cycle  unit which
                                               2-19

-------
        TABLE 2-12.  NITROGEN OXIDE EMISSION FACTORS FOR HEAVY-DUTY,
                     NATURAL GAS-FIRED PIPELINE COMPRESSOR ENGINES
                     WITHOUT CONTROL EQUIPMENT, (Reference 2-5)
          Source                        NO  Emission Factors3

                              g/kWh (lb/10* hph)     kg/106 sera (lb/106 scf)c
Reciprocating Internal
Combustion Engine                  Ib (24)               55,400 (3,400)
aTotal nitrogen oxides expressed as NO?.
u                                     £
 These factors are for compressor engines operated at rated load.  In
 aeneral, NO  emissions will increase with  increasing load and intake
 (manifold) air temperature and decrease with increasing air-fuel ratios
 (excess air rates) and absolute humidity.

cThese factors calculated from the energy-based factors for reciprocating
 engines assuming a hea.ting value of 41 MJ/scm  (1100 Btu/scf) for natural
 gas and an average fuel consumption 10.6 MJ/kWh (7500 Btu/hph).
                                      2-20

-------
            TABLE 2-13.  NITROGEN OXIDE EMISSION FACTORS FOR
                         STATIONARY LARGE BORE DIESEL AND DUAL
                         FUEL ENGINES WITHOUT CONTROL EQUIPMENT,
                         (Reference 2-11)
          Engine Type
                         NO  Emission Factors3
                           A
Diesel

g/hph
(lb/103 hph)
(lb/103 gal)h

Dual Fuel

g/hph
(lb/103 hph)
                                             11

                                            (24)

                                             60

                                           (500)




                                              8

                                            (18)
 Total  nitrogen oxides expressed as N02.   Factors are for engines
 operated at rated load and speed.

JThese  factors  calculated from the  above  factors assuming a heating
 value  of 40 MJ/liter (145,000 Btu/gal)  for oil, 41 MJ/scm (1100 Btu/scf)
 for natural  gas,  and an average fuel  consumption of 9.9 MJ/kWh
 (7000  Btu/hph).
                                    2-21

-------
exhausts to a waste heat boiler to recover thermal energy from the exhaust gases.   In some cases,
the 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 (Idw
buildings, short stacks, little visible emissions).  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 caseload), pumping,
gas compression, standby electricity generation, and miscellaneous Industrial uses.

       National NOV emissions estimates for gas turbines and internal  combustion engines combined
                  X                    -                 .
are presented in Table 2-2.-  NOX emission factors for two uses of gas  turbines are presented in
Tables 2-14 and 2-15.  Table 2-14 presents NO  emission factors for electric utility turbines, while
                                             X
Table 2-15 gives factors for NO  emissions from gas turbines used to power pipeline compressors
(Reference 2-6).

2.3.2  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 refining,
chemical manufacturing, cement 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, painting curing ovens, and metal ore smelting, to name only  a few.  Brief process
descriptions for some of the more significant NO  emission sources are given in the following
paragraphs.  Where available, national NO  emissions estimates and NOV emission factors are
                                         X                          X
presented for the process heating categories.  NO  emissions from the  industrial process equipment
sector are the most difficult to quantify of all stationary sources.  This is largely due to the
extreme diversity of equipment types currently in use.  Nationwide NO   emissions estimates for the
                                                                    X
more significant industrial processes over the period 1970 to 1980 are presented in Table 2-16
(Reference 2-4).
                                               2-22

-------
        TABLE 2-14.  COMPOSITE NITROGEN OXIDE  EMISSION  FACTORS  FOR
                     THE  1971 POPULATION OF  ELECTRIC UTILITY TUR-    .
                     BINES  WITHOUT  CONTROL EQUIPMENT,  (Reference  2-6)
Emission Factor Basis                       -NO  Emission Factor
Time Basis
     Entire population
       kg (lb)/hr rated loada                   4.01  (8.84)

     Gas-fired only
       .kg (lb)/hr rated load*                   3.54  (7.81)

     Oil-fired only
       kg (lb)/hr rated load*                   4.35  (9.60)
Fuel Basis
     Gas-fired only

       kg/105 m3 (lb/106 ft3) gas               6615  (413)

     Oil-fired only

       kg/103 liter (lb/103 gal) oil            8.13 (67.8)
aRated load expressed in megawatts.
                                      2-23

-------
       TABLE 2-15.  NITROGEN OXIDE  EMISSION  FACTORS FOR HEAVY-DUTY,
                    NATURAL GAS-FIRED PIPELINE COMPRESSOR GAS
                    TURBINE ENGINES WITHOUT  CONTROL EQUIPMENT,
                    (Reference 2-6)
                                   NO  Emission Factors3
Source                      g/fcWh (lb/103 hph)    kg/106 son (lb/106 scf)c
Gas Turbines                  1.7   (2.9)            4,700     (300)
aTotal nitrogen oxides expressed as N09.
L                                     f-
 These factors are for compressor engines operated at rated load.  In >
 general, NO  emissions will increase with increasing load and intake
 (mantifold) air temperature and decrease with increasing air-fuel ratios
 (excess air rates) and absolute humidity.

°These factors are calculated from the energy-based factors for gas
 turbines by assuming a heating value of 41 MJ/scm (1100 Btu/scf) of
 natural gas and an average fuel consumption of 14.1 MJ/kWh (10,000 Btu/hph)
                                    2-24

-------
                      TABLE 2-16.   NATIONWIDE NITROGEN OXIDE EMISSIONS ESTIMATES  FROM MAJOR

                                   INDUSTRIAL PROCESS SOURCES, 1970-1980, (Reference 2-4)
Source Category

Pulp Mills
Organic
Chemicals
Manufacturing
Ammonia
Production
Nitric Acid
Production
Petroleum
Refining
Glass
Manufacturing
Cement
Manufacturing
Iron & Steel
TOTAL
NOV Emissions, Gg (103 tons)/yr
1970 1972 1974 1975 x 1976 1977 1978 1979 1980

20(22)
60(66)
40(44)
150(165)
220(242)
40(44) '
90(99)
70(77)
690(759)

30(33)
60(66)
40(44)
140(154
230(253)
50(55)
100(110)
70(77)
720(792)

30(33)
60(66)
40(44)
130(143)
240(264)
50(55)
100(110)
80(88)
730(803)

20(22)
60(66)
40(44)
110(121)
240(264)
50(55)
80(88)
70(77)
670(737)

30(33)
50(55)
40(44)
110(121)
240(264)
50(55)
90(99)
70(77)
680(748)

30(33)
50(55)
50(55)
100(110)
260(286)
60(66)
100(110)
80(88)
720(792)

30(33)
50(55)
40(44)
100(110)
260(286)
60(66)
100(110)
80(88)
720(792)

30(33)
60(66)
50(55)
100(110)
250(275)
60(66)
100(110)
70(77).

30(33)
40(44)
50(55)
100(110)
240(264)
50(55)
90(99)
60(66)
1
720(792) 1 660(726)
ro
i
IVJ
01

-------
2.3.2.1  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
  A
furnaces, soaking pits and reheat furnaces.  The remaining combustion-related processes
(palletizing, heat treating, and finishing) are less important because they use relatively small"
amounts of fuel (Reference 2-15).

       Sintering machines are used to agglomerate ore fines, flue dust, and coke breeze for charging
of a blast furnace.  The use of this operation is presently declining 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,
thereby producing 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
                               X
are nonetheless substantial because of the very large quantity of fuel consumed in this process.

       Open hearth furnaces are now being replaced in the U.S. steel industry by the basic oxygen
furnace, however, they are still an important source of NO  emissions because of the very high
combustion air preheat temperature, the high operating temperatures, and the practice  of oxygen
lancing.
       Soaking pits and reheat furnaces are used to .heat steel billets and ingots to correct working
temperatures prior to forming.  Current trends are toward continuous casting of molten metal, and
the need for the soaking and reheat units is being reduced.  At present, however, soaking pits and
reheating furnaces still consume more fuel than any other single process in the steel  industry.  In
spite of the fact that soaking pits and reheat furnaces are being phased out, consumption of process
fuel continues to increase in the iron and steel industry as a whole.

       Tables 2-17 to 2-19 present NOV emission factors for the various emission sources within the
                                     A
Iron and steel industry.  Table 2-17 presents emission factors for the coking process.  Table 2-18
gives NO  emission factors that are applicable to gray iron furnaces. NOV emission factors for steel
        A                                                               A
foundry operations are presented in Table 2-19.  All NO  emission factors for iron and steel
processes were obtained from Reference 2-9.
                                                  2-2fr

-------
      TABLE  2-17.   NITROGEN OXIDE EMISSION FACTORS FOR COKE MANUFACTURE
                   WITHOUT CONTROL EQUIPMENT, (Reference 2-9)
     Type of Operation                 'NO  Emission Factors3
                                   kg/Mg of            (1b/ton of
                                   coal charged         coal charged)
     Coal Charging                    0.015               (0.03)


     Door Leaks                       0.005               (0.01)
aTotal nitrogen oxides expressed as N07.
u
 No data available for NO  emissions from other coke manufacturing sources
 including coal preheaters, coke pushing, quenching, and combustion stacks.
        TABLE 2-18.   NITROGEN OXIDE EMISSION FACTORS FOR GRAY IRON
                   ... FURNACES WITHOUT CONTROL 'EQUIPMENT, (Reference 2-9)
          Furnace Type3                      NO  Emission Factors
                                               A
                                        kg/Mg of gray       (Ib/ton of gray
                                        iron produced        iron produced)
          Electric Arc                   0.02 - 0.3         (0.04 - 0.6)
3No data available for NO  emissions from other types of furnaces including
 cupola, electric induction, and reverberatory.
                                    2-27

-------
           TABLE  2-19.   NITROGEN  OXIDE EMISSION  FACTORS  FOR  STEEL
                        FOUNDRIES WITHOUT CONTROL  EQUIPMENT3,
                        (Reference 2-9)
      Furnace Type                        NO  Emission  Factors
                                          /\
                                    kg/Mg of metal            (Ib/ton of metal
                                    processed                 processed)
      Electric Arc                       0.1                       (0.2)


      Open Hearth                        0.005                     (0.01)
a
 The  streams measured  for  NO  emissions  had  controls  for  particulates,
 however,  these would  not  hatfe  any  effect  on reducing NO  .
h
 No NO  emissions  data were available  for  open  hearth oxygen  lanced  and
 electric  induction  furnaces.
   TABLE 2-20.  NITROGEN OXIDE EMISSION FACTORS FOR GLASS MANUFACTURING
                MELT FURNACES WITHOUT CONTROL EQUIPMENT, (References 2-6)
                                                            a
          Type of Glass                 NOX Emission Factors
                                   kg/Mg of            (Ib/ton of
                                   glass produced       glass produced)
          Container                  1.6 - 4.5,        (3.3 - 9.1,
                                     avg. 3.1           avg. 6.2)

          Flat                       2.8 - 5.2,        (5.6 - 10.5,
                                     avg. 4             avg. 8)

          Pressed and Blown          0.4 - 10.0,       (0.8 - 20.0,
                                     avg. 4.3           avg. 8.5)
a
 NO  emissions from forming and finishing in all three glass categories
 an* negligible.
                                     2-28

-------
2.3.2.2  Glass Industry
       In the glass industry, melting furnaces and annealing lehrs are the two fuel  combustion
processes 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
occurring.  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 electrically 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.

      'As shown in Table 2-16, nationwide NO  emissions from the glass industry were estimated to be
50 Gg (55 x 10  tons) in 1980.  Nitrogen oxide emission factors for basic glass manufacturing are
presented in Table 2-20.  Table 2-21 presents NO  emission factors specifically for glass fiber
manufacturing, a subcategory of the glass manufacturing industry.  Emission factors for both tables
were obtained from Reference 2-6.

2.3.2.3  Cement Manufacturing 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 1480°C (2700°F).

       Coal, fuel oil, and natural gas are the main fuels used in cement kilns.  As of 1977 natural
gas accounted 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 for the past 20 years show that the industry has grown at an average rate of
1.9 percent annually (Reference 2-16).
                                                   2-29

-------
       TABLE 2-21.  NITROGEN OXIDE EMISSION FACTORS FOR GLASS FIBER
                    MANUFACTURING WITHOUT CONTROL EQUIPMENT,
                    (Reference 2-6)
          Source
          NO  Emission Factors
kg/Mg of material        . (Ib/ton of material
  processed                 processed)
Glass Furnace - Wood
     Electric
     Gas-Regenerative
     Gas-Recuperative
     Gas-Unit Melter
Glass Furnace - Textile
     Regenerative
     Recuperative
     Unit Melter
Oven Curing - Wool
     Rotary Spun
     Flame Attenuation
Rotary Spun Cooling - Wool
    0.14
    2.5
    0.85
    0.15
   10
   10
   10
    0.55
    1.0
    0.15
Oven Curing and Curing - Textile  1.3
 (0.27)
     ,
  1.7)
 (0.3)
(20)
(20.)
(20)
 (1.1)
 (2)
 (0.3)


 (2.6)
                                    2-30

-------
       Nationwide NO  emissions estimates from cement manufacturing are given in Table 2-16.
Table 2-22 gives NO  emission factors for both wet and dry process cement manufacturing.  All
                   X
emission factors were obtained from Reference 2-5.
2.3.2.4  Petroleum Refining Industry

       A wide variety of process combustion takes place in the petroleum refining industry,
including 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 570 to 650°C (1050 to 1200°F), but fuel requirements are on the order of
829 kJ/ (125,000 Btu/Bbl) feedstock.  Future growth of catalytic cracking will depend on the
national energy and environmental policies, and particularly on the demand for low sulfur fuel oil.

       Catalytic reforming is a process where paraffinic hydrocarbons are converted into aromatic
compounds.  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 impurities 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 2.9 percent annual
increase in process heating to 1985 (Reference 2-16).  The fuel mix expected 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 question of natural
gas price regulations is settled.
       Nationwide NO  emissions estimates from petroleum refineries, are presented in Table 2-16
(Reference 2-4).  References 2-4 and 2-5 were used to obtain NOX emission factors for petroleum
refinery sources.  Table 2-23 presents the petroleum refinery emission factor data.
2.3.2.5  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
                                                2-31

-------
          TABLE 2-22.  NITROGEN OXIDE EMISSION FACTORS FOR CEMENT
                       MANUFACTURING WITHOUT CONTROL EQUIPMENT,
                       (Reference 2-5)
     Process Source                NO  Emission Factors
                              kg/Mg of             (Ib/ton of
                              cement produced      cement produced)
     Dry Process Kilns             1.3                   (2.6)


     Wet Process Kilns             1.3                   (2.6)
a
 These emission factors include emissions from fuel combustion, which should
 not be calculated separately.
                                    2-32

-------
TABLE  2-23.    NITROGEN  OXIDE  EMISSION  FACTORS  FOR PETROLEUM  REFINERIES,
                   (References 2-4,  2-5)
                    Refinery Process8
                                      NOX Emission  Factor
FLUID CATALYTIC CRACKING UNITS

     Uncontrol lad
          kg/10J,liters fresh feed
          (1&71Q  bbl fresh feed)

     ESP and CO,Boilerc
         ' kg/10J..liters fresh feed
          (lb/10J bbl fresh feed)


MOVING-BED CATALYTIC CRACKING UNITS

     kg/103,1iters fresh feed
     (lb/10° bbl fresh feed)


COMPRESSOR ENGINES

     Reciprocating-Engines
          kg/10%nr gas burned
          Mk/inJ ft3 gas tupped)
     Gas
                        Turbines ,
                         kg/10 ,nT  gas burned
                         (lb/103
                SLOWDOWN SYSTEMS

                    Uncontrolled
                         kg/10 ..liters  refinery feed
                         (lb/10J bbl  refinery feed)

                    Vapor Recovery System and Flaring
                         kg/10J,Hters  refinery feed
                                 bbl  refinery feed)
                VACUUM DISTILLATION AND COLUMN
                CONDENSERS

                     Uncontrolled
                         kg/lQ3,liters vacuum feed
                         (lb/10J bbl  vacuum feed)
                     Controlled
                FLARES
                PROCESS HEATERS

                     Oil
                     Gas
                                                      0.107 - 0.416, avg.  » 0.204
                                                      (37.1 - 145, avg. -  71)
                                                      0.107 - 0.416, avg. « 0.204°
                                                      (37.1 - 145, avg. * 71)
                                                               0.014
                                                                (5)
                                                             55.4
                                                             (3.4)
                                               4.7
                                              (0.3)
                                             Negligible
                                            (Negligible)
                                               0.054
                                             (18.9)
                                             Negligible
                                            (Negligible)

                                             Negligible


                                      0.054 kg/103  liters (0.01916 lb/bbl)e
                                      6.57 kg/103  liters (2.30 lb/bbl)e
                                      2.2 kg/nT {0.14 Ib/ff3)6
                 NOX emission factor data fof fluid coking  units were not available.
                bTotal nitrogen oxides  expressed as N02-
                °ESP » electrostatic precipitator, CO « carbon monoxide.
                 Hay be higher due to the combustion of ammonia.
                eTotal nitrogen oxides  and not N02-
                                           2-33

-------
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 1090°C (2000°F).  Combustion products are ducted  from  the kiln to  a
drfer, 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.  NO
emission factors for brick manufacturing were obtained from Reference 2-5 and  are  presented  in
Table 2-24.

2.3.2.6  Noncombustion and Other Minor Industrial Sources
       Noncombustion-based NO  emissions from the chemical industry dominate this  category.   Oxides
of nitrogen are chemically released  during the production of nitric acid, adipic acid,  terephthalic
acid, acrylonitrile, adiponitrile, and explosives.  Tables 2-25 to 2-27 present NO emission factors
for nitric acid, adipic acid, and explosives manufacturing, respectively.  Uncontrolled NO
emissions from a terephthalic acid reactor are 6.5 kg/Mg (13 Ib/ton) of• acid produced (Reference
2-5).  Uncontrolled NOX emissions from acrylonltrile and adiponitrile production are  4.9  kg/Mg
(9.8 Ib/ton) and 70.6 kg/Mg (140 Ib/ton), respectively (Reference 2-4).
       Other Industrial processes such as primary copper smelting, coffee roasting, lime  production,
nitrate fertilizer production, and ammonia production also generate oxides of  nitrogen  in lesser
amounts.  NO  emission factors for these minor industrial sources are presented in Table  2-28
(References 2-5 to 2-10).  Nationwide 1980 NO  emissions estimates for some of the minor categories
including ammonia, explosives, organic chemicals, and pulp mills are presented in  Table 2-16.
2.3.3  Solid Waste Disposal Sources

       This category of NO  emissions includes such sources as refuse and sewage sludge incinera-
                          X
tors, auto body incinerators, conical burners, and open burning.  The 1980 nationwide HO  emissions
from solid waste disposal are presented in Table 2-1.  Each of the solid waste disposal NO  sources
are briefly described in the following paragraphs.  NO  emission factors are presented for each
category of incineration.

       The most common type of refuse incincerator consists of a refractory-lined chamber with a
grate upon which refuse is burned.  Combustion products are formed by heating and burning of refuse
on the grate.  Specific operating conditions, refuse composition, and basic incineration design have
                                             2-34

-------
           TABLE 2-24.   NITROGEN OXIDE EMISSION FACTORS FOR BRICK
                        MANUFACTURING WITHOUT CONTROL EQUIPMENT,
                        (Reference 2-5)
     PROCESS SOURCE
     N0x EMISSION FACTOR
                              kg/Mg of brick
                                 produced
                   /lb/ton of brick]
                        produced   /
CURING AND FIRING
     Tunnel  Kilns
          Gas-fired
          Oil-fired
          Coal-fired

     Periodic Kilns
          Gas-fired
          Oil-fired
          Coal-fired
0.08
0.55
0.45
0.21
0.85
0.70
(0.15)
(1.1)
(0.9)
(0.42)
(1.7)
(1.4)
                                     •2-35

-------
              TABLE 2-25.  NITROGEN OXIDE EMISSION FACTORS FOR
                           NITRIC ACID PRODUCTION3, (Reference 2-9)
          SOURCE
        NOV EMISSION FACTOR
          X
                                   kg/Mg acid
                                    produced
                      Mb/ton acid\
                      y  produced /
WEAK ACID PLANT TAIL GAS
     Uncontrolled
     Catalytic Reduction
       Natural gas
       Hydrogen
       Natural gas/hydrogen
        (25%/75%)c
     Extended Absorption
HIGH STRENGTH ACID PLANF
7-43, avg. 22

0.03-0.6, avg. 0.2
0-0.8, avg. 0.4
0.4-0.6, avg. 0.5

0.4-1.4, avg. 0.9
(14-86,  avg.  43)

(0.05-1.2, avg.  0.4)
(0-1.5,  avg.  0.8)
(0.8-1.1, avg.  1.0)

(0.8-2.7, avg.  1.8)

       (10)
  Based on 100% acid.  Production rates are in terms of total  weight of
  product (water and acid).  A plant producing 454 Mg (500 tons)/day
  of 55 wt. % nitric acid is calculated as producing 250 Mg (275 tons)/day
  of 100% acid.
  Based on data from two plants with these process conditions:  production
  rate, 118 Mg (130 tons)/day at 100% rated capacity; absorber exit
  temperature, 32°C ,(90°F); absorber exit pressure, 600 kPa (87 psig);
  acid strength, 57%.
c Based on data from two plants with these process conditions:  production
  rate, 188 Mg (208 tons)/day at 100% rated capacity; abosrber exit
  temperature, 32°C (90°F); absorber exit pressure, 550 kPa (80 psig);
  acid strength, 57%.
d Based on a unit that produces 6615 kg/hr (3000 Ib/hr) at 100% rated
  capacity, of 98% nitric acid.
                                   2-36

-------
   TABLE  2-26.
NITROGEN OXIDE  EMISSION  FACTORS FOR
ADIPIC  ACID  MANUFACTURE,  (Reference 2-5)
PROCESS SOURCE

RAW MATERIAL STORAGE ,
Uncontrolled
CYCLOHEXANE OXIDATION
Uncontrolled
w/boiler
w/thermal incinerator
w/f Taring
w/carbon adsorber
w/scrubber plus boiler
NITRIC ACID REACTION
Uncontrolled0
w/scrubber
w/thermal reduction6
w/flaring or combustion*
ADIPIC ACID REFINING
Uncontrolled
ADIPIC ACID DRYING, LOADING,
AND STORAGE
Uncontrolled
NOX EMISSION FACTORS3
kg/Mg of acid
produced


0
0
0
0
0
0
0

27
8
0.5
8

0.3

0
/lb/ton of acid\
\ produced j


(0)
(0)
(0)
(0)
(0)
(0)
(0)

(53)
(16)
(1)
(16)

(0.6)

(0)
  NO*  is  in the form of NO and NCfc.  Although large quantities of N20 are
  also produced, N?0 is not considered a criteria pollutant and is not,
  therefore, included in these factors.

  Uncontrolled emission factors are after scrubber processing since
  hydrocarbon recovery using scrubbers is an integral part of adipic acid
  manufacturing.

c Uncontrolled emission factors are after NOX absorber since nitric acid
  recovery is an integral part of adipic acid manufacturing.

  Based on estimated 70% control.

e Based on estimated 97.5% control.
                                    2-37

-------
   TABLE  2-27.
NITROGEN OXIDE  EMISSION FACTORS FOR
EXPLOSIVE MANUFACTURING,  (Reference 2-5)
     PROCESS SOURCE
                        NO  EMISSION  FACTORS
                                   kg/Mg produced
                                     (Ib/ton produced)
TNT - BATCH PROCESS-
     Nitration Reactors
          Fume recovery
          Add recovery
     Nitric Acid  Concentrators
     Sulfuric Acid  Concentrators
          ESP exit
          ESP w/scrubbera
     Red Water Incinerator
          Uncontrolled
          Wet scrubber
TNT - CONTINUOUS PROCESS
     Nitration Reactors
          Fume recovery
          Acid recovery
     Red Water Incinerator


NITROCELLULOSE
     Nitration Reactorsd  .
     Nitric Acid Concentrator
     Boiling Tubs
               3-19, avg. 12.5
               0.5-68, avg.  27.5
               8-36, avg. 18.5

               1-40, avg. 20
               1-40,. avg. 20

               0.75-50, avg. 13
               .  2.5
               3.35-5.0, avg.  4
               0.5-2.25, avg.  1.5
               3-4.2, avg. 3.5
               1.85-17, avg. 7
               5-9, avg. 7
                       1
(6-38, .avg. 25)
(1-136,  avg. 55)
(16-72,  avg. 37)

(2-80, avg. 40)
(2-80, avg. 40)

(1.5-101,  avg. 26)
(5.0)
(6.7-10,  avg. 8)
(1-4.5,  avg. 3)
(6.1-8.4, avg. 7)
(3.7-34,  avg.  14)
(10-18, avg.  14)
       (2)
a No data available  for NOX emissions after the scrubber.  It is assumed
  that NO  emissions are unaffected by the scrubber.
  Use low end of range for modern, efficient units  and high end for older,
  less efficient units.
  Apparent reductions in NOX after control may not  be significant because
  these values are based on only one test result.
  For product with low nitrogen content (12 percent), use high end of range.
  For products with  higher nitrogen content, use lower end of range.
                                      2-38

-------
                                   TABLE 2-28.   NITROGEN  OXIDE  EMISSION  FACTORS  FOR
                                                MISCELLANEOUS  INDUSTRIAL-PROCESSES,
                                                (References  2-5 to  2-10)
            INDUSTRY
     SOURCE
          NOV EMISSION FACTOR0
            A *
I
to
u>
       AMMONIA PRODUCTION
       CARBON BLACK
        PRODUCTION
       COFFEE ROASTING
       GRANULAR NITRATE
        FERTILIZER MFG.
Primary Reformer
  Natural gas
  Fuel oil
Oil Furnace Process
  Main process vent
  Flare
  CO boiler &
   i nci nerator
  Dryer vent
    bag filter

    scrubber
  Solid waste
   incinerator
Thermal Process
Roaster
  Di rect-fi red
  indirect-fired
Grandulator
Dryers & Coolers
2.9 kg/Mg (5.8 Ib/ton) of ammonia produced
2.7 kg/Mg (5.4 Ib/ton) of ammonia produced

0.28 kg/Mg (0.56 Ib/ton) of carbon black produced
Not Available

4.65 kg/Mg (9.3 Ib/ton) of carbon black produced
                                                         0.12-0.61  kg/Mg,  avg.  0.36
                                                           (0.24-1.22  Ib/ton,  avg. 0.73)
                                                         1.1  kg/Mg  (2.2  Ib/ton)  of carbon  black  produced

                                                         0.04 kg/Mg (0.08  Ib/ton) of  carbon  black  produced
                                                         Unknown0
0.05 kg/Mg (0.10 Ib/ton) of coffee produced
0.05 kg/Mg (0.10 Ib/ton) of coffee produced
0.45 kg/Mg (0.9 Ib/ton) of fertilizer produced
1.5 kg/Mg (3.0 Ib/ton) of fertilizer produced

-------
TABLE 2-28.  NITROGEN OXIDE EMISSION FACTORS FOR MISCELLANEOUS INDUSTRIAL PROCESSES (CONTINUED)
ORCHARD HEATERS
ASPHALTIC CONCRETE
 PRODUCTION
PRIMARY COPPER
 SMELTING6
LIME PRODUCTION9
 GAS  SWEETENING PLANTS1
 PULP MILLS
Type of Heater
  Pipeline
  Lazy flame
  Return stack
  Cone
  •Solid fuel
Rotary Drum

Reverb Furnace
 Followed by Converters
  Reverb furnace

  Converters
Crushers, Screens,
 Conveyors, Storage
 Piles
                         Rotary Kilns
                         Vertical Kilns
                                         .h
Calcimatic Kilns
Fluidized-Bed Kilns
Product 'Coolers
Hydrators
Amine
Kraft Pulp Process
Negligible0
Negligible
Negligible
Negligible
Negligible
18 g/Mg (0.036 Ib/ton) of concrete produced6'
                                                  0.045 kg/Mg (0.09 Ib/ton) of ore
                                                   concentrate processed
                                                  0.025 kg/Mg (0.05 Ib/ton) of ore
                                                   concentrate processed
Negligible
1.5 kg/Mg (3 Ib/ton) of lime produced
Data Not Available
0.1 kg/Mg (0.2 Ib/ton) of lime produced
Data Not Available
Negligible
Negligible
Negligible
0.90 kg/Mg (1.79 Ib/ton) of pulp produced

-------
FOOTNOTES FOR TABLE 2-28.
a All emission factors are for uncontrolled emissions unless otherwise specified.
  Average values are based on the results  of six sampling runs conducted*at a representative
  plant with the industry mean production  rate of 5.1 x 10  Mg (5.6 x 10  tons)/yr.
c Emissions data are not available, but all  emissions are believed to be negligible.
  Little nitrogen oxide is formed because  of the relatively low combustion temperatures.
e Total nitrogen oxides expressed as N02 .
  Based on limited test data from a single asphaltic concrete plant.
^ All emission factors for kilns and coolers are per unit of lime produced.  Divide  by two to
  obtain factors per unit of limestone feed to the kiln.   Factors for hydrators are  per unit
  of hydrated lime produced.  Mul.tiply by  1.25 to obtain factors per unit of lime  feed to
  the hydrator.
  Calcimatic kilns generally employ stone  preheaters.  All  factors represent emissions after
  the kiln exhaust passes through a preheater.
1 Emission factors are presented only for  smokeless flares and tail gas incinerators  on the
  amine gas sweetening process.   These factors represent emissions after smokeless flares
  (with fuel gas and steam injection) or tail  gas incinerators.

-------
a pronounced effect on emissions.  Nitrogen oxide emissions increase'with an  increase  in  the
temperature of the combustion zone, an increase in the residence time iri the  combustion zone  before
quenching, and an increase in the excess air rates to the point where dilution cooling overcomes  the
effect of increased oxygen concentration (References 2-5 to 2-6).  Refuse incinerators have been
labeled as municipal, industrial/commercial, trench, domestic,  flue-fed, pathological, and
controlled-air depending on the specific use sector or use function.   NO  emission  factors for
refuse incinerators are given in Table 2-29 (References 2-5 to  2-6).
       Incineration is being used to dispose of sewage sludge because it destroys the  organic matter
present in sludge, leaving only an odorless, sterile ash, and also reduces the solid mass by  about
90 percent.  The most prevalent types of sewage sludge incinerators are multiple hearth and
fluldized bed units.  In multiple hearth units the sludge enters the top of the furnace where it  is
first dried by contact with the hot, rising-, combustion gases,  and then burned as it moves slowly
down through the lower hearths.  At the bottom hearth any residual ash is then removed.   In
fluidized bed reactors, the combustion takes place in a hot,, suspended bed of sand  with much  of the
ash residue being swept out with the flue gas.  Temperatures in a multiple hearth furnace are 320°C
(600°F) in the lower, ash cooling hearth; 760 to IIOO'C (1400 to 2000°F) in the central combustion
                          *
hearths, and 540 to 650°C (1000 to 1200°F) in the upper, drying hearths.  Temperatures in a
fluidized bed reactor are fairly uniform, from 680 to 820°C (1250 to 1500°F).  In both types  of
furnace an auxiliary fuel may be required either during startup or when the moisture content  of the
sludge is too high to support combustion.
       NO  emission factors for sewage sludge incinerators were obtained from Reference 2-5 and are
         X
presented in Table 2-30.

       Auto incinerators consist of a single primary combustion chamber in which one or several
partially stripped cars are burned. Approximately 30 to 40 minutes is required to burn two bodies
simultaneously.  As many as 50 cars per day can be burned in this batch-type operation, depending on
the capacity of the incinerator.  Continuous operations in which cars are placed on a  conveyor  belt
and passed through a tunnel-type incinerator have capacities of more than 50 cars per  8-hour  day.
Both the degree of combustion as determined by the incinerator  design and the amount of combustible
material left on the car greatly affect emissions.  Temperatures on the order of 650°C (1200°F) are
reached during auto body incineration.  Some auto incinerators  are designed for two stage combus-
tion, using afterburner control devices;  Afterburners result in a reduction  of both NO   emissions
and the emissions of the other criteria pollutants.
                                                 . 2-42"

-------
           TABLE 2-29.  NITROGEN OXIDE EMISSION FACTORS FOR REFUSE
                        INCINERATORS WITHOUT CONTROL EQUIPMENT,
                        (References 2-5, 2-6)
     INCINERATOR TYPE
  NO  EMISSION FACTORS0
                                   kg/Mg of refuse
                                       burned
               Ib/ton of refuse\
                    burned
MUNICIPAL
     Multiple Chamber
INDUSTRIAL/COMMERCIAL
     Multiple Chamber
     Single Chamber
     Trench
          Wood
          Rubber tires
          Municipal refuse
     Control!ed-air
FLUE-FED SINGLE CHAMBER
FLUE-FED (MODIFIED)5
DOMESTIC SINGLE CHAMBER
     w/o Primary Burner
     w/Primary Burner
PATHOLOGICAL
.1.5

1.5
1
(3)

(3)
(2)
2                     (4)
   Data not.avail able
   Data not available
5                    (10)
1.5                   (3)
5                    (10.)
0.5
1
1.5
(1)
(2)-
(3)
  Total nitrogen oxides expressed as N02«
  W.ith afterburners and draft controls.
                                    2-43

-------
                 TABLE  2-30.   NITROGEN OXIDE EMISSION FACTORS
                              FOR SEWAGE SLUDGE INCINERATORS,
                              (Reference 2-5)
CONTROL STATUS



Uncontrolled
After Scrubber Control
NO EMISSION
X
kg/Mg
FACTORS9 fb

' (Ib/ton)

3
Z.b
(6)
(5)
a Unit weights in terms of dried sludge
  Total nitrogen oxides expressed as
                  TABLE  2-31.   NITROGEN OXIDE  EMISSION  FACTOR
                               FOR AUTO BODY  INCINERATORS,
                               (Reference 2-5)
CONTROL STATUS

Uncontrolled
With Afterburner
N0x
kg/car
EMISSION FACTORSa'b
(Ib/car)

0.05
0.01
(0.1)
(0.02)
  Based on 113 kg (250 Ib) of combustible material on stripped car body.
  Total nitrogen oxides expressed as N02.
                                     2-44

-------
       NO  emission factors for auto body Incinerators are  presented In Table 2-31 (Reference 2-5).

       Conical  burners are generally a truncated metal cone with a  screened top vent.   The charge 1s
placed on a raised grate by either conveyor or bulldozer;  however,  the use of a conveyor results 1n
more efficient burning.  No supplemental fuel  1s used, but  combustion air 1s often supplemented by
underflre air blown Into the chamber below the grate and by overflre air Introduced through
peripheral openings 1n the shell.  The quantities and types of pollutants released from conical
burners are dependent on the composition and moisture content of the charged material, control of
combustion air, type of charging system used, and the condition In  which the Incinerator 1s
maintained.  The most critical of these factors seems to be the level of maintenance on t!,e
Incinerators.  It 1s not uncommon for conical burners to have missing doors and numerous holes 1n
the shell, resulting 1n excessive combustion air, low temperatures, and therefore, Mgh emission
rates of combustible pollutants.  NO  emission factors for conical  burners handling municipal and
wood refuse are presented 1n Table 2-32 (Reference 2-5).

       Open burning can be done In open drums or baskets, 1n fields and yards, and 1n large open
dumps or pits.  Materials commonly disposed of In this manner are municipal waste, auto body
components, landscape refuse, agricultural field refuse, wood refuse, bulky Industrial refuse, and
leaves.  Ground-level open burning 1s affected by many variables Including wind, ambient tempera-
ture, composition and moisture content of the refuse  burned, and compactness of the pile.  In
general, the relatively low temperatures associated with open burning suppress the emission of
nitrogen oxides.  Estimates of nationwide NO  emissions from open burn1 ig are given in Table 2-1.
NO  emission factors for the open burning of nonagr1cultural materials are presented 1n Table 2-33
(Reference 2-6).
 2.3.4   Other Miscellaneous  NO  Sources
      Forest  fires,  structural  fires, and explosives detonation are some of the NO  sources
 Included  1n  this  category.   Nationwide  NO  emissions estimates for miscellaneous sources such as
 th?s*j are given  in  Table  2-1.   Available NOX emission factors for forest fires and explosives
 detonation are presented  in  Tables  2-34 and 2-35  (References 2-5 and 2-8).
                                                2-45

-------
     TABLE 2-32.  NITROGEN OXIDE EMISSION FACTORS FOR WASTE INCINERATION
                  IN CONICAL BURNERS WITHOUT CONTROL EQUIPMENT,
                  (Reference 2-5)
TYPE OF WASTE

Municipal Refuse
Wood Refuse3
NOV EMISSION
A
kg/Mg of refuse
burned
FACTORS
Mb/ton of refuse]
\ burned /

2.5
0.5
(5)
t1)
  Moisture content as fired is approximately 50 percent for wood waste.
            TABLE 2-33.  NITROGEN OXIDE EMISSION FACTORS FOR OPEN
                         BURNING OF NONAGRICULTURAL MATERIALS,
                         (Reference 2-6)
     TYPE OF WASTE
NOV EMISSION FACTORS
  J\
                              kg/Mg of refuse
                                  burned
              Mb/ton of refuse\
              \     burned     /
Municipal Refuse
Automobile Components'
                      (6)
                      (4)
  Upholstery, belts, hoses, and tires burned in common.
                                     2-46

-------
            TABLE 2-34.  NITROGEN OXIDE EMISSION FACTORS
                         FOR FOREST WILDFIRES, (Reference 2-5)
GEOGRAPHIC AREAb

Rocky Mountain Group
Northern - Region 1
Rocky Mountain - Region 2
Southwestern - Region 3
Intermountain - Region 4
Pacific Group
California - Region 5
Pacific Northwest - Region 6
Alaska - Region 10
Southern Group
Southern - .Region 8
North Central Group
Eastern - Region 9
Eastern Group (with Region 9)
TOTAL United States
NOV EMISSION
X
kg/hectare of
forest burned
FACTORS
/ Ib/acre of \
^forest burned 1

166
269
135
45
36
85
81
269 ,
72
40
40
49
49
49
76
(148)
(239)
(120)
(40)
(32)
(76)
(72)
(239)
(64)
(35.6)
. (35.6)
(43.6)
(43.6)
(43.6)
(67.6)
Areas consumed by fire and NO* emissions are for 1971.
Geographic areas are specifically defined in Reference 2-5.
                                  2-47

-------
                TABLE 2-35.  NITROGEN OXIDE EMISSION FACTORS
                             FOR THE DETONATION OF EXPLOSIVES,
                             (Reference 2-8)
EXPLOSIVE3
    COMPOSITION
      NOV EMISSION FACTOR
        J\            	
                                               kg/Mg
                                               (Ib/ton)
Dynamite,
  Gelatin
ANFO
20-100% nitroglycerine
ammonium nitrate with
  5.8-8% fuel oil
4-59, avg. 26   (8-119, avg. 53)
      8
(17)
  Emission factors for black powder, smokeless powder, dynamite-strai'ght,
  dynamite-ammonia, TNT, RDX, and PETN were not available.
                                   2-48

-------
                                    2.4  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:  Hamil,  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 1974.
2-3  Federal Register, Volume 41, No. 232.  December 1, 1976.
2-4  Unpublished Data of the Monitoring and Data Analysis Division of the Office of Air Quality
     Planning and Standards, U.S. Environmental Protection Agency.  Research Triangle Park,
     North Carolina.  July 1982.\
2-5  "Compilation of Air Pollution Emission Factors (Third Edition)," Publication No. AP-42.  U.S.
     Environmental Protection Agency, Research Triangle Park, North Carolina.  August 1977.
2-6  Supplement No. 8 of Reference 2-5, May 1978.
2-7  Supplement No. 9 of Reference 2-5, July 1979.
2-8  Supplement No. 10 of Reference 2-5, February 1980.
2-9  Supplement No. 11 of Reference 2-5, October 1980.
2-10 Supplement No. 12 of Reference 2-5, April 1981.
2-11 Draft Supplement No. 13 of Reference 2-5, August 1982.
2-12 Personal communication with H.J. Melosh III, Foster Wheeler Corporation.
2-13 Surprenant, N.F. et. a!.  "Preliminary Emissions Assessment of Conventional Stationary  •
     Combustion SystemsT"" GCA Corporation, EPA Report No. 600/2-76-046b. March 1976.
2-14 Devitte, T., et. al., "The Population and Characteristics of Industrial/Commercial Boilers,"
     EPA-600/7-79-T78a, August 1979.
2-15 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, North Carolina.
     October 1974.
2-16 Foley, 6.  "Industrial Growth Forecasts," Stanford Research Institute, Contract No. 68-02-1320.
     September 1974.
                                                 2-49

-------

-------
                                              SECTION 3

                                         CONTROL TECHNIQUES

       This section presents a survey of the general principles and developmental status of
potential techniques for NO ' control for stationary sources.  It is intended to provide a broad
perspective 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 status of development, 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 general concepts on NO  formation and
control during combustion.  Much of the material in Section 3.1 was taken directly from
Reference 3-1 and the reader is referred back to this document for further discussion of combustion
modifications, especially as applied to utility boilers.
3.1.1  General Concepts on NO.. Formation and Control
       """' '   -• "-™1™""-111"—"••"•' ------   -  — X ••— --•-	—: --.-.-.-.-
       Oxides of nitrogen formed in combustion' processes are due either to the thermal fixation of
atmospheric nitrogen in the combustion  air, which produces "thermal NO  ," or to the conversion of
chemically bound nitrogen in the fuel, which produces "fuel NOV."  For natural gas and light
                                                              A                            . '
distillate oil firing, nearly all NO  emissions result from thermal fixation.  With residual oil,
                                    X
crude oil, and coal, the contribution from fuel bound nitrogen can be significant and, in certain
cases, predominant.
      '                                          \
       A third potential mechanism of NOX 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
       During combustion, nitrogen oxides are formed by the high temperature,  thermal  fixation  of
N«.  Nitric oxide (NO) is the major product, even though N02 is-thermodynamically favored at lower
temperatures.  The residence time in most stationary combustion processes is too short for signifi-
cant NO to be oxidized to NCL.

     The detailed chemical mechanism for thermal NO  formation is not fully understood.   However, it
is widely accepted that thermal fixation in the postcomfaustion zone occurs according to the extended
form of the Zeldovich chain mechanism (Reference 3-2):
                                        N2 + 0  * NO  + N                          '       (3-1)
                                        N  + 02 * NO  + 0 '                                (3-2)
                                        N  + OH * NO  + H                                 (3-3)
assuming that the combustion reactions have reached equilibrium.   Reaction (3-1)  has  a  large
activation energy (317 kJ/mol) and is generally believed to be rate determining.   Oxygen  atom
concentrations are assumed to have reached equilibrium according  to:

                                        02+M*0+0+M                            (3-4)

where H denotes any third substance (usually N2).

       In the flame zone itself, the Zeldovich mechanism with the equilibrium oxygen  assumption  is
not adequate to account for experimentally observed NO formation  rates.   Several  investigators have
observed the production of significant amounts of "prompt" NO, which  is  formed very rapidly in the
flame front (References 3-3 through 3-11), but there is not general agreement on  how  it is produced.
Prompt NO is believed to stem from the existence of "superequilibrium" radical concentrations
(References 3-11, 3-12, and 3-13) within the flame zone which result  from hydrocarbon chemistry
and/or nitrogen specie reactions, such as suggested by Fenimore (Reference 3-14),  To date, prompt
NO has only been explicitly measured in carefully controlled laminar  flames, but  the  mechanism
almost certainly exists in typical combustion flames as well.  Of course, in an actual  combustor,
both the hydrocarbon and NO  kinetics are directly coupled to turbulent  mixing in the flame zone.

       Recent experiments at atmospheric pressure indicate that under certain conditions  the  amount
of NO formed in heated N2, 02, and Ar mixtures can be expressed as (Reference 3-15):
                                               3-2

-------
                                   [NO] = kj exp(-k2/T)[N2][02J4t                         (3-5)
     where     [  ]    = mole fraction
               k., k2  = constants
               T       = temperature
               t       = time

Although this expression certainly will not adequately describe NO formation in a turbulent flame,
it does point out several features of thermal NO  formation.  It reflects the strong dependence of
NO formation on temperature.  It also shows that NO formation is directly proportional  to NX concen-
tration and to residence time, and proportional to the square root of oxygen concentration.

       Based on the above relations, thermal NO  can theoretically be reduced using four tactics:
     j                    "                 '     A          '

               1)  Reduce local nitrogen concentrations at peak temperature;
               2)  Reduce local oxygen concentrations at peak temperature;
               3)  Reduce the residence time at peak temperature; and
               4)  Reduce peak temperature.

Since reducing N2 levels is quite difficult, efforts in the field have focused on reducing oxygen
levels, peak temperatures,  and time of exposure in the NO  producing regions of the combustion
chamber.
       On a macroscopic scale, techniques such as lowered excess air and off stoichiometric (or
staged) combustion have been used to lower local 02 concentrations in boilers.  Also, staged combus-
tion 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 .
                                                                     X
       Flue gas recirculation and reduced air preheat have been used in boilers to control thermal
NO  by lowering peak flame temperatures.  Flue gas recirculation also reduces combustion gas
residence time, but its primary effect as a thermal NO  control is through temperature reduction.
                                                      X
Analogously, exhaust gas recirculation (EGR), reduced manifold air temperature (1C engines) and
reduced air preheat (regenerative gas turbines) have been 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 for prime movers which specifically reduce exposure time at
                                           3-3

-------
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 NOX reduction
alter combustion conditions on a rnacroscopic scale.  Although these macroscopic techniques have all
been relatively successful in reducing thermal NO  , local microscopic combustion conditions
ultimately determine the amount of thermal NO  formed.  These conditions are in turn intimately
related to such variables as local combustion intensity, heat removal rates, and internal mixing
effects.  Modifying these secondary combustion variables at microscopic levels requires fundamental
changes in combustion equipment design..

       For example, recent studies on the formation of thermal NOV 1n gaseous flames have confirmed
                                                                 X
that internal mixing can have large effects on the total amount of NO formed (References 3-16,
3-17).  Burner swirl, combustion air velocity, fuel injection angle and velocity, quarl angle, and
                  *
confinement ratio all affect the mixing between fuel, combustion air, and recirculated products.
Mixing, in turn, alters the local temperatures and specie concentrations which control the rate of
NO  formation.

       Unfortunately, generalizing these effects is difficult, because the 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 NOX 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 NO  emissions from combustion sources has been
recognized since 1968 (Reference 3-18).  Although the relative contribution of fuel  and thermal  NO
to total NOX emissions from sources firing nitrogen containing fuels has not been definitively
established, recent estimates indicate that fuel NOV is significant and may even predominate.  In
                                                   A
                                             3-4

-------
one laboratory study (Reference 3-19), residual  oil  and pulverized coal  were burned in an
argon/oxygen mixture to eliminate thermal N0v effects.  Results show that fuel  NOV can account for
                                            A                                    A
over 50 percent of total NO  production from residual oil firing and approximately 80 percent of
                           X
                                                                    coal-fired  utility boiler,
confirm this prediction (Reference 3-20).  Flue gas  recirculation, which controls primarily thermal
total NOX from coal firing.  Tests on a full scale system, a 560
NO , was a relatively ineffective NO  control measure for the coal-fired boiler tested.
  X                         .        X
            TABLE 3-1.  FACTORS CONTROLLING THE FORMATION OF THERMAL NOX (Reference 3-1)
          Primary Equipment
         and Fuel Parameters
                                             Secondary
                                      Combustion Parameters
   Fundamental
   Parameters
          Inlet temperature,
          velocity
          Firebox design
          Fuel composition
          Injection pattern
          of fuel and air
          Size of droplets
          or particles
          Burner swirl
          External mass
          addition
                                      Combustion intensity
                                      Heat removal rate
                                      Mixing of combustion
                                      products into flame
                                      Local fuel/air ratio
                                      Turbulent distortion
                                      of flame zone
Oxygen level
Peak temperature
Exposure time at
peak temperature
Thermal
  NO.,
       Fuel bound nitrogen occurs in coal and petroleum fuels.  However, the nitrogen containing
compounds in petroleum tend to concentrate in the heavy resin and asphalt fractions upon
distillation (Reference 3-21).  Therefore, fuel NO  is of importance primarily in residual oil and
coal firing.  The nitrogen compounds found in petroleum include pyrroles, indoles, isoquinolines,
acridines, and porphyrins.  Although the structure of coal has not been defined with certainty, it
is believed that coal-bound nitrogen also occurs in aromatic ring structures such as pyridine,
picoline, quinoline, and nicotine (Reference 3-21).

       The nitrogen content of most residual oils varies from 0.1 to 0.5 percent.  Nitrogen content
of most U.S. coals lies in the 0.5 to 2 percent range (Reference 3-22); anthracite coals contain the
least and bituminous coals the most nitrogen.  Figure 3-1 illustrates the nitrogen content of
various U.S. coals, expressed as ng N02 produced per joule for 100 percent conversion of the fuel •
nitrogen (Reference 3-23).  The figure clearly shows that if all coal bound nitrogen .were converted
                                             3-5

-------
CO
I
en
3000






c
o
« 2000
Ol
o

i«
O
O

t
•-3
•«v.
0°^ 1000
2J
en
c






•*



6


c 5
o
•r"
- 10
01
r-
8 4
*^.a
§
o
*~
3
3
4->
CQ
iO
° 2
*~,
rvi
»>l
o
.0
" 1


0


0*t» From:
U5.B.M. Ttch. fltpu. | JB22-1MW
Averui(1»491 ' • MtifidH
~" Vtndwttr UW5). p, 45 O Bilunwnoui
M»rk. •! tl. (1B64) £ Sgb-Bitumlnout
D**Fk> lt&C.*kl *
rwry t«w3) Hi" •
Pwki&0'Donnfl||185fl) 9 tt9n(W
•*»

Di


A / t5» c4 RtMraM H*w t»MH
/ MbSCMMMi
n ° 0 /
:r 	 ^u 	 1
£ n ttft.l.
A"0i
' 3> CW
/*A^1
*f ** J
HFk
* .
— *£$
2y\
i *4
£» '
— %



EPA 1971 Star
|
1
I
1!
0 1

1
0
*£! On
.^ ^n ,
£!.,.,» ^ 	 "1
i f
Or
^







ida



*° Jl
i <«'/%
j (j^ 9
/« 0
<* A °
A <





d


1
75./. "°C
0 b
0 0 |
o 0 o
^^
•








1 1 1

Q
o
0 0










1


0°
0 a

X









I 1 1
234 56 789 10
Ib S02/10° Btu 100% conversion
1 i 1 i
1000 2000 3000 4000
                                                  ng S02/J    100% conversion
                    Figure 3-1.   Nitrogen and sulfur content of U.S.  coal  reserves  (Reference 3-23).

-------
to NO , emissions for all coals would exceed even the 1971 Standards of Performance for Large Steam
     X
Generators (NSPS).  Fortunately, only a fraction of the fuel nitrogen is converted to NO  for both
oil and coal firing, as shown in Figure 3-2 (Reference 3-24).  Furthermore, the figure indicates
that fuel nitrogen conversion decreases as nitrogen content increases.  Thus, although fuel  NO
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 3-25).  From observations such as these, the effectiveness of partial fuel
denitrification as a NO  control method seems doubtful.

       Although the precise mechanism by which fuel nitrogen is converted to NO  is not understood,
certain aspects are clear, particularly for coal combustion.  In a large pulverized coal-fired
utility boiler, the coal particles are conveyed by an airstream into the hot combustion chamber,
                                               n                                               •
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 (<300 ms).

       Figure 3-3 summarizes what may happen to fuel nitrogen during this process (Reference 3-26).
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 3-27).  Both total
mass volatilized and total nitrogen volatilized increase with higher pyrolysis temperature;  the
nitrogen volatilization increases more rapidly than that of the total mass.  Total mass volatilized
appears to be a stronger function of coal composition than total nitrogen volatilized
(Reference 3-28).  This supports the relatively small dependence of fuel NO  on coal composition
observed in small scale testing (References 3-19 and 3-29).

       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 evolved react
to form other reduced species before being oxidized.  Prior to oxidation, the devolatilized nitrogen
may be converted to a small number of common, reduced intermediates, such as HCN and NH,, in the
fuel-rich regions of the flames.  The existence of a set of common 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 3-21, 3-30).  More recent experiments suggest that HCN is the
predominant reduced intermediate (Reference 3-31).  The reduced intermediates are then either
oxidized to NO, or converted to Np in the postcombustion zone.  Although the mechanism for these
conversions is not presently known, one proposed mechanism postulates a role for NCO
(Reference 3-32).

                                            3-7

-------
00

X
o
z
° 100-
c
o
•r—
I/)
1 80-
c
o
o
u 60-
£
o
M 40'

20
?2 - 5 | Calif, residual fuel oil i
r, •-. II innito ™al 1 1 B1 tUIIll nOUS
« •• • ^ ifuel oil. 1 Lignite coai | *
I/ rn ISuh-hi tuminnus coal j

i antnracice #]-? fupl
Coal Data
0 Pereira, et al. (1974)
*Q ^ Pershing, et al. (1973)
g A McCann (1970)
^ Q <•> • Jonke (1970)
\& H Bituminous M.I. T. (1975)
^<£0^ ^ ^ Lignite M.I.T. (1975)
O^D ^

  Pershing, et al. (1973)

-------
                      Volatile fractions
                     (Hydrocarbons. RN etc.)
                                                             Ash
                                                          virtually
                                                          nitrogen
                                                            free
Figure  3-3.
Possible  fate of fuel  nitrogen  contained  in coal
particles during combustion  (Reference 3-26).
                                 3-9

-------
       Nitrogen retained 1n the char may also be oxidized to NO,  or reduced to N. through
heterogeneous reactions occurring 1n the postctmbustlon zone.  However, it is clear that the
conversion of char nitrogen to NO proceeds mucd more slowly than  the conversion of devolatllized
nitrogen.  In fact, based on a combination of experimental and empirical  modeling studies, It 1   now
believed that 60 to 80 percent of the fuel NOX results from volatile nitrogen oxidation (References
3-27, i-33).  Conversion of the char nitrogen to NO 1s in general lower,  tv faccors of two to three,
than conversion of total coal nitrogen (Reference 3-30).
       Regardless of the precise mechanism of fuel NO  formation, several general trends are
evident, particularly for coal combustion.  As expected, fuel nitrogen conversion to NO is highly
dependent on the fuel/air ratio for the range existing 1n typical combustion equipment, as shown in
Figure 3-4.  Oxidation of the char nitrogen is relatively insensitive to fuel/air changes, but
volatile NO  formation Is strongly affected by fuel/air ratio r.Manges.

       In contrast to thermal NO. fuel NO  production is relatively Insensitive to small changes  in
                                K         X
combustion zone temperature  (Reference 3-30).  Char nitrogen oxidation appears to be a very weak
function of  temperature, and although the amount of nitrogen volatiles appears to Increase as
temperature  increases, this  1s believed to be partially offset by a decrease ii. percentage
conversion.  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  a'r during coal devolatilization will
dramatically increase volatile nitrogen conversion and  increase  fuel NO  .  In contrast, char NO
formation  is only weakly dependent on initial nixing.

        From  the above modifications,  it appears  that,  in  principle, the  best strategy for fuel NO
abatement  combines low excess  air  (LtA) firing,  optimum burner design, and two stage  ombustion.
Assuming suitable  stage  separation,  low excess  air may  have little  effect on fuel NU  , but  it
 Increases  syster,  efficiency.   Before  using LEA  firing,  the  need  to  gt t good  caibon  burnout  and  low
CO emissions rrust  be considered.
        Optimum burner design for coal ensures  locally  fuel-Hch  conditions Juring devolati 1 ization,
which  promotes reduction of devolatllized nitrogen  to  N,.   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.  Higfi  secondary  <".ir  preheat may  also be desirable,  because  it  promotas  more
                                              3-10

-------
I/)
QJ
X

O
c
a
c

o
+J

c
o

>
V.
01

c
o
(J
ijali temp 1500°K
Flame temp 1600"K


A Lignite 75-90 Min
        te 38-45 ,>m
                                                                 Bituminous 38-45
                                    234

                                         Fuel  equivalence ratio

                                  (Inverse of stoichiometric  ratio)
     Figure  3-4.   Conversion  of nitrogen In coal to NO  (Reference 3-25).
                                                       J\

-------
complete nitroyen uevo'aLlMiotlor. ir. the fuel  rich initial  cofbustlon stage.   This leaves  less char
nitrogen to be subsequently oxidized in the fuel-le.in second stage.   Unfortunately, it also tend: to
favor thermal NO formation, and at present there is .10 general  agreemtnt 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 Mnetically or
aerodynamically limited in that their emission rates are far below the levels which would prevail at
equilibrium.  Thus, the rate of formation of both thermal and fuel NO  is dominated by combustion
conditions and is amenable to  suppression througi combustion process modifications.  Although the
mechanisms are different, both thermal and fuel NO  are promoted by rapid mixing of oxyren with the
fuel.  Additionally, thermal NO   is  greatly increased by lono 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:

           (1)  Decrease primary flame zone 0*  level oy
               --  Decreased overall 0,,  levsl
               --  Controlled  mixing of  fuel dnd air
               --  Use of  fuel-rich  primary fleme  zone
           (2)  Decrease time of exposure  at nigh temperature by
               --  Decreased peak temperature:
                   --  Decreased  idiabatic flame temperature through  dilution
                   --  Decreased  combustion  intensity
                        Increased  flame cooling
                   --  Controlled mixing  of fuel and  air or use  of  fuel-rich primary  flame icne
               --  Decreased primary flame zone  residence  time

        Table 3-2  relates  these cont.ol  concepts  to applicable combustion  proce1:: modifications and
 equipment types.   The  process  modifications  are  categorized according to  their  role  1t>  the control
 development  sequence:   operational  adjustments,  hardware modifications  of existing eouipment  or
 through factory  installed controls,  and  major  redesigns of  new  equipment.   The  controls  for
 decreased Oj are also  generally effective for  peak temperature  reduction  but have  not  been repeated.
 The following subsections briefly review the  status of each of  the  applicable  control  techniques.
                                              3-12

-------
TABLE 3-2.  SUMMARY OF COMBUSTION PROCESS MODIFICATION CONCEPTS (Reference 3-1)
Combustion
Conditions
Decrease
primary
**ame zone
02 level
Decrease
peak
flame
temperature
Control
Concept
Decrease overall
02 level
Delayed nixing
of fupl and air
Increased fuel/
air mixing
Urinary fuel-
rich flame
zone
Decrease
adlabatlc flaw
temperature
Decrease com-
bust Ion
intensity
Increased flame
zone cooling/
reduce resi-
dence time
Applicable
Equipment
Boilers, furnaces
Boiler, furnaces
L-is turbines
Boilers,
furnaces, JC
Boilers, fur-
naces. 1C.
ga$ turbines
Boilers, furnaces
Boilers, furnaces
Effect on
Thermal HO
Rsduc>s O^-rich,
h1gh-NOs pockets
In the flame
Flame cooling and
dilution during
'leleyed mix re-
duces peak temp.
Reduces local hot
itctchlometrlc
regions IP over-
all fuel lean
combustion
Plane cooling In
low-02, I CM- temp.
primary zone re-
duces peak temp.
Direct sup.-es-
slon of thermal
NOx mechanism
Increased flame
zone cooling
yields lowe*
peak temp.
Increased flame
zone cooling
yields lower
peak tewp.
Effect on
Fuel N0x
Reduces exposure
of fuel nitrogen
Intermediaries
to 02
Volatile fuel N
reduces to N; In
tne absence <>f
oxygen
Increases
Vo'.atllo fuel K
reduces to «2 In
the absence of
oxygen
Ineffective
HI nor direct
effect; Indirect
effect on mixing
Ineffective
Primary Applicable Controls
Operational
Adjustments
low excess air
firing
Burner
adjustments

Burners out of
service; biased
burner firing
Reduced air
preheat
Lead reduction
Burner tilt
Hardware
Modification
Flue gas reclrcu-
latlon (FGR)
Low NOX burners

Over fire air
ports, stratified
charge
Mater Injection,
FGR


Major
Redesign

Optimum burner/
firebox design
New can design;
premix, prev*p.
Burner/firebox
design for two-
stage combustion

Enlarged firebox
Increased burner
spec Ing
Redes t 99 heat
transfer sur-
faces, firebox
aerodynamics

-------
3.1.2  Modification of Operating Conditions

       The modification techniques described in this subsection include low excess  air,  off
stoichiometric or staged combustion, flue gas recirculation, reduced air preheat,  reduced firing
rate, and steam or water injection.

3.1.2.1  Low Excess Air Combustion (LEA)

       Reducing the total amount of excess air supplied for combustion is an effective demonstrated
method for reducing NO  emissions from utility and industrial boilers, residential  and commercial  .
                      A
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
                                                                          A
is furthermore easy to implement and increases boiler efficiency.   It is, therefore,  used
extensively in both new and retrofit applications, either singly or in combination  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 usually the first NO  control technique applied to  utility boilers.
It may be used with virtually all fuels and firing methods.  It was initially implemented to
increase thermal efficiency and reduce stack gas opacity due to acid mist, and it  is  now often
considered more of an energy conservation measure than a NO  control technique. A number of studies
have shown LEA firing to be effective in reducing NOX emissions without significantly increasing CO
or smoke levels (References 3-34 through 3-39).  As shown in Section 4, Table 4-1,  numerous tests of
low excess air firing on utility boilers have indicated NO  emission reductions averaging between 16
                                                          X
and 21 percent for coal, oil, and natural gas firing compared to earlier baseline  levels.

       The minimum practical level of excess air which can be achieved i,n existing  boilers and
process heaters, 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.
                                      *
                                             3-14

-------
       Other factors such as secondary air register settings and steam temperature control
flexibility 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 distribution 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.

       As discussed in Section 4.2, LEA firing  is also a very effective method for controlling  NO
in industrial boilers.  For residential and commercial furnaces, however, while LEA is a potentially
feasible NOX control technique, the trend in NO  control for these  sources has been in improved
burner design in order to obtain low NO  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,
feasible, and effective technique for utility and industrial boilers but  is less  applicable for
other stationary sources of combustion.  For certain applications such as utility boilers,  LEA
firing is presently considered a routine operating procedure and is incorporated  in all  new units.
Also, more and more industrial boilers are incorporating this techniques  as well.  Since it is  often
efficient and easy to implement, LEA firing may see increasingly widespread use in other
applications.  Most sources will require additional control methods, in conjunction with LEA, to
bring NOV emissions within statutory limits.  In such cases, the extent to which  excess  air can be
        A
lowered will depend upon the other control techniques employed.  However, virtually all
developmental programs for advanced NO  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 or Staged Combustion (OSC)

       Off stoichiometric or staged combustion  seeks to control NO   by carrying out initial
                                                                 X
combustion in a primary, fuel-rich combustion zone, then completing combustion at lower  temperatures
in a second, fuel lean zone.  In practice OSC is implemented through biased burner firing (BBF),
burners out of service (BOOS), or overfire air  injection (OFA).
                                          3-15

-------
Biased Burner Firing (BBF). Burners Out of Service (BOOS)

       Biased burner firing consists of firing the lower rows of burners more fuel-rich than  the
upper rows of burners.  This may be accomplished by maintaining normal  air distribution to  the
burners while adjusting fuel flow so that a greater amount of fuel  enters the furnace  through the
lower rows of burners than through the upper rows of burners.  Additional air required for  complete
combustion enters through the upper rows of burners which are firing air rich.

       In the burners out of service mode, individual burners, or rows  of burners,  admit air  only.
Correspondingly the total fuel demand is supplied through the remaining fuel  admitting or active
burners.  Thus the active burners are firing more fuel-rich than normal, with the remaining air
required for combustion being admitted through the inactive burners.

       These methods reduce NOV emissions by reducing the excess air available  in the  firing  zone.
                              A
This reduces fuel and thermal NOX formation.  These techniques are  applicable to  all fuels  and  are
particularly attractive as control methods for existing units since few, if any,  equipment  modifica-
tions are required (References 3-40 and 3-41).  In some cases, however, derating  of the unit  may be
required if there is too limited extra firing capability with the active burners.  This is  most
likely to be a problem with pulverized coal units without spare pulverizer cetpacity.

       Monitoring flue gas composition, especially 0« and CO concentrations,  is very important  when
employing these combustion modifications for NO  control.  Local reducing atmospheres  may cause
                                               X
Increased furnace slagging when burning coal because of the lower ash fusion  temperature associated
with reducing atmospheres (References 3-42 and 3-43).  In addition, it  is important to closely
monitor flue gas, excess air, and CO to avoid reducing boiler efficiency through  flue  gas heat  and
unburned combustible losses, and to prevent unsafe operating conditions caused  by incomplete
combustion.  For these reasons, accurate flue gas monitoring equipment  and increased operator
monitoring of furnace conditions are required with these combustion modifications.
       As shown in Section 4.1, Table 4-2, emission tests of burners out of service firing  on
utility boilers have indicated average NOV reductions of 31 to 37 percent for coal, oil, and  natural
                                         A
gas firing compared to earlier baseline levels.  A typical burners  out  of service pattern is  shown
in Figure 3-5(a).
                                               3-16

-------
                                                                            OVIRfllt All
                                                                            NOZZlii
   O Active  burners

   )SC Burners admitting air only

 a.  Typical  burners out of
     service  arrangement
     opposed  fired unit
                                             WIMSSOX
                                     MCONOAIV All CAMPUS
                                          UCOMOAtY AH
                                        BAMNI MlVf UNIT
                                                                           — COAIMOZZUS
                                                                            OIL OWN
    b.  Typical  overfire air system for
        tangential  fired unit (Reference 4-21)
Burners
                 Apportioning
                 dampers
                                                                        Air
                                                           Forced draft fan
Flue gas recirculating
fan
                  c.  Typical  flue  gas  recirculation system for  NO   control
                                                                   A
    Figure 3-5.  Typical arrangements  for (b)  overfire air,, (a) burners  out
                 of service, and  (c) flue gas  recirculation (Reproduced  from
                 Reference 3-1,'p.  4-26).
                                      3-17

-------
Overfire Air (OFA)

       The overfire air technique for NO  control  involves firing the burners  more  fuel  rich  than
normal while admitting the remaining combustion air through overfire air ports or an  idle  top row of
burners.

       Overfire air is very effective for NO  reduction and may be used with all  fuels.  However,
there is an increased potential for furnace tube wasteage due to local  reducing conditions when
firing coal or high sulfur oil.  There is also a greater tendency for slag accumulation  in the
furnace when firing coal (References 3-23, 3-41, 3-43, and 3-44).  In addition, with  reduced  airflow
to the burners, there may be reduced mixing of the fuel and air.  Thus, additional  excess  air may be
required to ensure complete combustion.  This may  result in a decrease in efficiency
(References 3-41 and 3-44).

       Overfire air is more attractive in original designs than in retrofit applications because of
cost considerations.  Additional duct work, furnace penetrations, and extra fan capacity may  be
required.  There may be physical obstructions outside of the boiler setting making  installation more
costly.  There may also be insufficient height between the top row of burners  and the furnace exit
to permit the installation of overfire air ports or to allow sufficient residence time for the
completion of combustion (Reference 3-41).

       Also, overfire air is more easily implemented without large efficiency  or cost penalties on
large units than on small ones.  As unit size is decreased, furnace volume decreases  faster than the
available wall surface.  Hence, furnace residence  times available for fuel  combustion tend to be
shorter in small units.  Staged combustion techniques such as overfire air serve to dalay  or  prolong
the combustion process.  Thus, on small units larger proportional increases in furnace size and cost
may be required to assure complete fuel combustion with the application of these techniques.   Or
alternatively, increased excess air rates through  the overfire air ports may be required thus
leading to decreased unit efficiency.  For these reasons, staged combustion techniques are commonly
applied to utility size boilers; but their application to smaller units is more limited.
       As shown in Section 4.1, Table 4-3, some emission tests of overfire air on utility boilers
have indicated average NO  reductions of about 25 to 60 percent for coal,  oil, and natural  gas
firing compared to earlier baseline levels.  A typical  overfire air system is  shown in
Figure 3-5(b).
                                             3-18

-------
       An off stoichiometric or staged combustion technique similar to overfire air is also
applicable to control of NOV from natural  draft process heaters (Reference 3-45).   The staged
                           A
combustion system that has been demonstrated in process heaters uses a number of air lances arranged
around each burner.  The burner is operated under fuel  rich conditions with about  65 percent of the
air required for combustion entering through the burner air registers.  The remainder of the air
required for complete combustion is injected into the flame zone some four feet above the burner.
This technique, in conjunction with low excess air operation, reduced NO  emissions by 21 percent
below a baseline of 66 ng/J for a unit firing refinery gas.  Fuel  consumption was  reduced by nearly
5 percent.                                                                           •  -          ,
                                          *
3.1.2.3  nue£a? Recirelation (FGR)

       Flue gas recirculation for NO  control consists of extracting a portion of  the flue gas,
                                    A
usually from the economizer outlet with utility boilers, and returning it to the furnace.  The flue
gas may be admitted through the furnace hopper or through the burner windbox or both.  Flue gas
recirculation lowers the bulk furnace gas  temperature and reduces the oxygen concentration in the
combustion zone (References 3-41 and 3-44).

       Flue gas recirculation through the  furnace hopper and near the furnace exit has long been
used for steam temperature control.  Flue  gas recirculation through the windbox and, to a lesser
degree, through the furnace hopper is very effective for NOX control on gas- and oil-fired units
(References 3-40 and 3-44).  However, it has been shown to be relatively ineffective on coal-fired
units (Reference 3-20).

       Flue gas recirculation for NO  control is more attractive for new designs than as a retrofit
                                    A                   .         »
application.  Retrofit installation of flue gas recirculation can be quite costly.  The fan, flues,
dampers, and controls as well as possibly  having to increase existing fan capacity due to increased
draft loss, can represent a large investment.  In addition, the flue gas recirculation system itself
may require a substantial maintenance program due to the high temperature environment and potential
erosion from entrained ash.  Thus the cost-effectiveness of this method of NO  control has to be
examined carefully when comparing it to other control techniques.

       As a new design feature, the furnace and convective surfaces can be sized for the increase  in
mass flow and change in the furnace temperatures.  In contrast, in retrofit-applications the
increased mass flow increases turbulence and mixing in the burner zone, and alters the boiler heat
absorption profile.  Erosion and vibration problems may result (References 3-44 and 3-46).  Flame
                                             3-19

-------
detection can also be difficult with flue gas recirculation through the windbox.   In  addition,
controls must be employed to regulate the proportion of flue gas to air so that a sufficient concen-
tration of oxygen is available for combustion (Reference 3-47).
       On utility boilers, flue gas recirculation has most often been used in combination  with  other
low NO  combustion techniques.  Test data for these types of applications are discussed in
Section 4.1.  A typical flue gas recirculation system in a utility boiler application is shown  in
Figure 3-5(c).
       Flue gas recirculation has also been applied to a few process heaters.  However, no perfor-
mance data are available for these units.  The technique may not be applicable to all types of
heaters because it lowers flame temperature and can cause problems with flame stability
(Reference 3-48).  Flue gas recirculation is therefore unlikely  to be used in high temperature
applications such as ethylene pyrolysis heaters.

3.1.2.4  Reduced Air Preheat Operation (RAP)

       Reducing the amount of combustion air preheat lowers the  primary combustion zone peak
temperature, generally lowering thermal NO  production as a result.  Because of the energy penalty
associated with this technique, it has been used only sparingly  in utility and industrial  applica-
tions.  It is applicable to utility steam generators and large industrial boilers which employ  heat
exchangers to impart about 280°K (500°F) incremental heat to the combustion air.
       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 NOV control if means other than air preheat were developed to recover heat from 420°K to
          X
700°K (300°F to 800°F) gases.  For example, in new industrial boilers it is often practical to
replace the air preheater with an economizer thereby reducing or eliminating the  energy penalty
associated with this technique.  However, this technique appears relatively ineffective in
suppressing fuel nitrogen conversion (References 3-49, 3-50).
       This technique is also applicable to turbocharged internal  combustion engines  and regenera-
tive gas turbines.  The turbocharged 1C engines normally have 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.  However, regenerative gas turbines recover some
of the thermal energy in the exhaust gas where temperatures range from 700°K to 870°K (800°F to
                                            3-20

-------
1100°F) to preheat the combustion air.  Any reduction in air preheat would cause severe fuel
penalties unless other means of recovering the heat in the exhaust could be implemented.

3.1.2.5  Reduced Firing Rate
       Thermal NO  formation generally increases as the volumetric heat release rate or combustion
intensity increases.  Thus, NO  can be controlled by reducing combustion intensity through load
reduction (or derating) in existing units, and by enlarging the firebox in new units.   The reduced
heat release rate
(Reference 3-51).
heat release rate lowers the bulk gas temperature which in turn reduces thermal  NOX  formation
       The heat release rate per unit volume is generally independent of unit rated power output.
However, the ratio of primary flame zone heat release to heat removal increases as the unit capacity
is increased.  This causes NO  emissions for large units to be generally greater than for small
units of similar design, firing characteristics, and fuel.

       The increase in NO  emissions with increased capacity is especially evident for gas-fired
boilers, since total NOV emissions are due to thermal NOV.  However,  for coal-fired and oil-fired
                       A                                A
units the effects of increased capacity are less noticeable, since the conversion of fuel nitrogen
to NO  for these fuels represent a major component of total NOV formation.  Still, a reduction in
     A                                                        A
firing rate will affect firebox aerodynamics which may, consequently, affect fuel NO  emissions.
But such effects on fuel NOX production are less significant.

       Reduced firing rate often leads to several operating problems.  Aside from the limiting of
capacity, low load operation usually requires higher levels of excess air to maintain steam
temperature and to control smoke and CO emissions.  The steam temperature control range is also
reduced substantially.  This will reduce the operating flexibility of the unit and its response to
changes in load.  The combined results are reduced operating efficiency due to higher excess-air and
reduced load following capability due to a reduction in control range.

       When the unit is designed for a reduced heat release rate, the problems associated with
derating are largely avoided.  The use of an enlarged firebox produces NO  reductions similar to
load reduction on existing units.
                                           3-21

-------
3.1.2.6  Steam and Water Injection (MI)
       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 NOV emissions from internal  combustion engines  and gas turbines;
                                   A
Figure 3-6 shows NOV emission reductions from a gas turbine as high as 80 percent (Reference 3-52).
                   A
       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
Injection.  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
decreased thermal efficiency 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 has therefore  not
gained much acceptance as a NOX reduction technique for stationary combustion equipment except for
gas turbines (References 3-49 and 3-50).  Gas turbines, in addition to having the lowest efficiency
losses with water injection, also showed no major operational problems or reduced equipment life
with this technique.  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  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.
                                                                          X
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
                                             3-22

-------
   §
                                                   .Ol
                                                   ,P
                                                    CM
                                                           O

                                                           CO

                                                           03
                                                           2
                                                           O
                                                          .o
                                                           o
                                                           UJ
                                                           I
                                                                       CM
                                                                       LO
                                                                        t
                                                                     (0  O)
                                                                     10 S-
                                                                     S- CD
                                                                     o o:
                                                                    tt  XIM ^
                                                                     o
                                    O O
                                    CD -Z.
                                                                    •r- O
                                    i. S-
                                    O) (O
                                                                     00 -»->
                                                                     £=  01
                                                                     O  =
                                                                    ir^  O

                                                                     V 3^
                                                                      XQ
                                                                     O  OO
                                    s: en
                                    o
                                    a s-
                                                                     s_  en
                                                                     O  n3
                                                                     cj  cr>
                                                                      i
                                                                     CO
o
CM
(SISV8 ABO **O %Sl) -ON  Wdd
                                  3-23

-------
combustion temperatures, while off-stolchiometric operation reduces the amount of fuel  burned at
peak temperature under oxidizing conditions.  For the most part, combining control  techniques has
been shown to be complementary but not additive for NOX reduction (Reference 3-49).

3.1.3  equipment Oe-.lgn Modification

3.1.3.1  Burner Com 1gurat1 on
       Bur-ner or combustor modification for NOX control 1s applicable to all stationary combustion
equipment categories.  The specific design and configuration of a burner has an Important bearing on
the amount of NOX formed.  Certain design types have been found to give greater omissions than
others.  For exarcple, the spud-type gas burner appears to give a higher emission rate than the
radial spud type., which, 1n turn, produces more NO  than the ring type.
       During the early 1970's  specially designed "low-NOx" burnsrs were produced for thermal N0x
control.  For the most part,  they were designed for utility and Industrial boilers and employ
inflame  LEA, OSC, or  FSR principles.  The aim  1s to strike a balance between "inimum NOX formation
and acceptable  combustion of  carbon and hydrogen in the fuel.

       There are currently several commercial  low-NOx gas and ..11 burner resigns for borers and
process  heaters in  operation  and under development  (References  3-53 through 3-57).  Full scale test
results  for boilers in Japan  show reduction: a).  Some of the more innovative
methods  for oil  burners  include:  flame splitting distributor tips which cause a flower petal flamr
arrangenqnt, and atomizers with fuel  injection holes of different diameters which create fuel-rich
and fuel-lean combustion  zones  (References  3-53, 3-56, 3-58).   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-NOx  burner (offered  by  at least one  company in  the
 U.S.) for oil-  and  gas-f1rad package  boilers.   This burner uses shaped fuel  injection  ports  »nd
 controlled  air-fuel mixing to create  a  thin stuhby  ring-shaped  flame  (References 3-53,  3-55).  With
 this modification,  reductions 1n  NO  from 20 to 50  percent are  claimed.   The most  extensive  air-fuel

                                                  3-24

-------
modifications Involve tne self-red rculating and staged combustion chamber type of burners,  used in
Industrial process furnaces.  These burners are equipped with a prevaporizatlon or a precombustior
chamber 1n the wlndbox.  In the chamber the fuel 1s vaporized and premixcd with part of the
combustion air, or is allowed to undergo partial combustion under oxygen def^ciert conditions  before
being discharged Into the furnace.  NO  reductions of about 55 percent are typical for these
devices.

       Low N0x '.turners are the most common typo of NO,  emission control technique applied to process
heaters.  The most common type of low NO  butner tr  oe best described as & two-stage combustion
burner which 1s fired fuel rich In the first stagt.  The burner 1s designed to Inject tertiary air
after sufficient time has elapsed 1n the '•educing zone of the fla.ne.  Controlled introduction of the
tertiary  air provides reduction of NOX emissions In the reducing zone of the flame without
significant  changes  1n flame pattern and burner operation.

       A  second klrd of  staged combustion burner Is the staged fuel burner.  This technique Involves
combustion of a fuel with high excess air.  The remainder of the fuel 1s injected In the second
stage of'  the reaction and combustion is completed at low excess air.  The high excess air permits
the  first stage of the combustion to occur at a low temperature.  Depending en the amount of excess
air, the  theoretical temperature may be as low  as 1030°C (2000°F).  As the combustion reaction goes
to completion  in the first  zone, the additional fuel 1s Injected.  The second  reaction begins with a
reduced partial pressure of oxygen which tends  to limit the formation of NO .  Other vendors offer
low  NO  burners based  on flue gas  recirculitlon and two stage  combustion.  The self-reelrcu^ating
gasification  (SRG) burner has been designed to  employ  flue gas rec1rcui?tion,  two-staged combustion,
gasification reactions,  and low excess air.  The key feature  1s the creation of an exceptionally
strong  recirculation eddy  in the  burner throat  tile.   This draws combustion reaction products from
the  furnace  to gasify  the  fuel  stream.  The primary air flow  1s between  ten and thirty percent
stlochiometric depending upon  the  design.  The  result  1s  that  the gases  leaving the burner  throat
are  very  rich  in  hL  and  CO.

        Several  utility boiler  manufacturers nave also  beer active  in  the development of new burners
 designed  to  reduce NO   emissions  from coal-fired units.   Most  low NO   burners  designed for  utilitv
 boilers control  NOX  by reducing flame turbulence,  delayirg fueVair mixing, and establishing  fuel-
 rich zones where  combustion initially takes  place.   This  represents a  departure from  the  usual
 burner design procedures which prcr.jte high tL-bulence,  high intensity,  rapid  combustion  flames.
 The longer,  less  intense flames produced  with  low  NO  burners  result  1n  lower  flame temperatures
                                              3-25

-------
which reduce thermal NO  generation.  Moreover, the reduced availability of oxygen 1n the Initial
combustion zone Inhibits fuel NOV conversion.  Thus, both thermal and fuel NO, are controlled by the
                                A                                            A
low NOX burners.

       The Babcock and Mil cox Company Is currently Installing the dual register pulverized coal-
fire^ burner 1n all Its not utility bo'lle»-s  in order to meet current NSPS (Refsrences 3-46 and
3-59).  TV. limited turbulence, control1:*! di^efus1on flame burner Is designed to minimize *uel and
air mixing at the burner to that retired to obtafn Ignition and sustain iv.able combustion of the
coal.  A VentuH mixing device, located in the coal nozzle, provides a uniform coal/primary air
mixture at the burner.  Sscondary air 1s Introduced through two concentric zone* surround!r-g the
coal nozzle, each of which 1s Independently controlled by Inner and outer air zone registers.

       At least seven dual register burner-equipped utility boilers have been tested for NO
emissions (Reference 3-53),  Tests on four bituminous coal-fired units showed NO  emissions ranging
fro*! 190 to 260 ng/J (0,45 to 0.6 lb/106 Btu, 320 to 420 ppm).  Tests on three subbltuminous coal-
f1r«d unit:, showed  NOX emissions In the range of 130 to 150 ng/J (0.3 to 0.35 lb/106 Btu. 210 to
250 ppm).  Comparisons with NO  emissions from similar unite equipped with tne high turbulence older
burners show reductions in "iOx  levels from 40 to 60 percent due to the new burner design.

       In another recently reported test of  the dual register burner on a bituminous coal unit,
 EPA collected  approximately 68  days of  continuous monitoring  data  (Reference 3-60).  Dally average
 emissions were consistently below  200 rvyj  (0.47 lb/10   Btu}  and 30 day  rolling average emissions
 ranged froai 160 to 170  ng/J (0.37-0.39  lb/10G Btu).

        BftW  claims  that  NOX control  through  Its dual  register  burners  1s  superior  to  staging as  It
 maintains *he  furnace '  Injected above the burner  zone.

        Although the du«l  register  burners were developed for  use in new  boilers,  they  can also  be
 retrofitted to older units,   However, the new boilers are  also designed  to provide airflow control
 en a per pulverizer basis.  This may not  be possible in some  of the older units,  or  the cost
 involved In retrofitting a  compartmented  windbox and making the necessary changes 1n pulverizer
 burner piping  may  be high.   If careful  control of  fuel  and air tc  each burner  is  not feasible,  the
                                                3-26

-------
burners will not be as effective in reducing NOX emissions.   Nevertheless,  the  new  burners  should
reduce NO  levels below those obtained with the older high turbulence burners.   They  may still  be
considered for retrofit arplication, perhaps 1n conjunction with other NOX  control  techniques,  but
much development work remains.
       Foster Wheeler Energy Corporation has developed a dual register rosl burner  for Installation
In Its new boilers (References 3-43 and 3-61).  The new burner reduces turbulence as  compared to the
older designs and causes controlled, gradual mixing of fuel  and air at the  burner.  This 1s achieved
using a dual throat with two registers which splits the secondary air Into  two concentric streams
with Independently variable swirl.  The mixing rate between the primary and secondary air streams
and the rate of entrapment of furnace gases can thus be varied.
       Test results for the new Foster Wheeler burners are reported 1n References 3-43 and 3-61.
Reductions  in NO  emissions of about 40 percent were observed on a four-burner ste^m generator when
operated at full load with the new burners.  Three utility steam generators, two 265  MW opposed
fired  units and one 75 MU front wall unit, have been retrofitted with the new burners and tested for
NOX emissions.  Controlled NOX emissions were  In the 170 to 220 ng/J  (0.4 to 0.5 lb/10S Btu, 280 to
350 ppm) range, about 40 to 50 percent lower than similar units with older design burners.
        In addition to NOX control  in new units, the Foster Hhaeler dual register burner 1s
technically we'll »uited for retrofit application.  The airflow to the new burners is controlled
individually at each burner by means of a  perforated hood.  !!ence, precise air/fuel control at each
burner is possible without Incurring major hardware changes besides burner replacement.

        RHey Stoker Corporation 1s  currently modifying the burners used 1n Its turbo furnace to
lower  NO  emissions  (Reference  3-62).  The new burners are designed to be more flexible and to
control fuel/air mixing to reduce  thermal  and  fuel NOX>  Uith  the new burners and changes  1n furnace
design Riley Stoker  expects  to  meet current  NSPS requirements  without Increased carbon or  unburned
hydrocarbon losses.   The new  burners  can be  used with  coal,  oil, and  gas fuels but are not being
considered  for retrofit application.   No tes.*,  data are available on the performance of the new
burners at  present.
        In  summary,  low NOX burners appear  very attractive, with  potential  NOX  reductions of  the
 order of  50 percent.   Data  from long  term, full  scale  demonstrations  a;-e imminent, and  commercial
 application 1s well  underway.
                                            3-27

-------
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
NO  emissions with tighter spacing and a decreased ability to  radiate to  cooling surfaces.   This
effect 1s Illustrated by the higher NOX emissions from  larger  boilers with greater  multiples of
burners and tighter spacing.

       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  bunt to
allow sufficient time for complete c  ^ustlon with slower and  more controlled heat  release rates,
such as that associated with the oft-st    -iometric operating  mode.  Furthermore, furnace  plan areas
have been Increased to allow for larger heat transfer to the cooling  walls.  This increase in the
burner zone dimensions creates more wall area thus increasing  the  distance between  evenly  spaced
burners.
       Horizontal burner spacing is largest for tahydntially 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 know spiral configuration.  As  a result the flames radiate  widely  to the
surrounding  cooling surfaces before Interacting with one another.   Also,  the tangential firing
configuration  results  in s'ow mixing of fuel with the combustion air.  For these reasons,
tangent1ally-f1red boilers  show baseline, uncontrolled emissions beloi; those for other utility
boilers firing configurations.  It has been observed, however, that for many tangentially-fired
units, reductions 1n  NOX balow the naturally low uncontrolled levels is more difficult than reduci^
NO  on units with higher uncontrolled emissions  (References 3-39 and 3-49).

3.1.3.3   Advanced Burner/Furnace Designs

       A  number of advanced burner designs  are being developed and tested  to reduce NO  emissions
from  coal-  and o1l-f1red utility and  Industrial  boilers.  Advanced burners, as  compared to  low NO
burners,  are defined  as those  devices  still  under  experimental or  pilot  scale development for
 lowering  NOX emission.  Burner modification has  the potential of lowering  N0y emissions well below
 levels attainable by  conventional  combustion modification techniques.  Burner modification  also has
 the advantage of  requiring  minimal changes  1n  current  boiler  design  and  operation  and  is  suitable
 for retrofit application.   A  few  of the  techniques under  development  are  discussed below.
                                              3-28

-------
       Some manufacturers of oil-firing equipment ait ',r. +*<• process of developing burners capable
of operating at very low levels of excess air.   The low excess air requirements Increase boiler
efficiency and reduce fan power" corsumption while decreasing NO  emissions.   The low excess air nay
also reduce SO* conversion.  The Peabody Engineering Company has designed the Air Pressure Recovery
(APR) burner designed to operate at excess oxygen levels down to 1/2 percent without Increase 1n
partlculate and unbumed hydrocarbon emissions.  The Coen Company 1s developing the LEA burner which
uses a tip swlrler to operate down to 0.1 percent excess oxygen (Reference 3-63).  Both burners are
currently undergoing testing and no data on NOX emissions are available.

       For coal-fired utility boilers, Foster Wheeler 1s currently testing an advanced dual register
split flame burner design.  A device id-led at the burner no.z1e splits the primary air-coal flow
Into several distinct streams.  Coal particles become concentrated within each stream and, hence
diffuse more slowly Into the secondary air.  This further Inhibits NOX formation by extending the
slow-burning characteristics of the dual register burner.  Results from an Industrial size test
boiler are promising with a NOX level of approximately 130 ng/J (0.3 lb/10  Btu) for subbltumlnous
coal (Reference 3-61).  However, the burner tested on a 375 MU electrical output boiler produced
approximately 215 ng/J  (0.5 lb/10  Btu).  A further modification of this burner with a variable
velocity split flame nozzle will be Installed, and a NOX level of 15C to 170 ng/J (0.35 to
0.4  lb/106 Btu) 1s expected (Reference 3-64).

       Babcock & W11cox and Energy and Environmental Research, under EPA sponsorship, are developing
an advanced low NOX coal burner,  the distributed fuel/air mixing burner, for field testing
 (Reference 3-65).  The  burner  if designed to control both thermal and fuel NOX.  Coal and prlmiry
air  are  Injected from the  center of the burner with  a moderate axial component.  This stream 1s
 surrounded by a divided secondary alrstream with a swirl component  for  stabilization.  Tertiary air
 for  burnout 1s added axlally around the periphery of the burner.  The arrangement results  In t. hot,
 rich redrculatlon zone at the center  or  the 'flame.  Time 1n  the rich zone helps maximize ev./lutlon
 of nitrogen from the char  and  reduce most of the fuel NO  that may  be formed.   Also, axial addition
 of the tertiary air  leads  to  a large flame zone.  Heat  extraction prior to completion of  burnout,
 along with dilution  of  the tertiary air  by combustion products,  lowers  the peak flame temperature,
 thus reducing thermal NO .  Although experimental  prototypes  have achieved NO   emissions  below
 86 ng/J (0.2  lb/10   Btu),  actual  field testing has  not  yet  been  conducted.

        Advanced burners are also  under development  for  tangentlally-fired  systems  (Reference  3-67).
 Combustion Engineering  and Acurex Corporation, the  latter under sponsorship  of the  EPA,  are
                                                 3-29

-------
developing a burner called the "fuel-rich fireball."  .This burner achieves a fuel-rich, relatively
low temperature flame by diverting some of the normal air flow along the furnace walls.  The air
rejoins the combustion zone higher up in the furnace for complete fuel burnout.  Demonstration
testing is underway.  Another advanced burner for tangentially-fired systems is under development by
Mitsubishi Heavy Industries of Japan.  The test program for this burner has also involved Japan's
Electric Power Development Company and Combustion Engineering of the United States.   Pilot scale
testing has yielded emissions as'low as 86 ng/J (0.2 lb/10  Btu).
       Babcock & Wilcox Company is developing a primary combustion furnace concept for coal-fired
utility boilers in a program sponsored by the Electric Power Research Institute (Reference 3-66).
The fundamental NO  control process in this furnace is conversion of fuel  nitrogen to N2 through
fuel-rich combustion.  Pulverized coal is introduced into an extended combustor with substoichio-
roetric air, so that combustion occurs under fuel-rich conditions isolated from the rest of the
furnace.  The length of the combustor is sufficient to provide the necessary residence time to
partially oxidize the coal and permit the desirable N2 producing reactions to occur.  Heat is
removed along the combustion chamber to prevent slagging.  Secondary air is added at the exit of the
primary combustion furnace to bring the combustion products to oxidizing conditions  before they
enter the secondary furnace.  Pilot scale testing of a 1 MW (4 x 10  Btu/hr) heat input prototype
has achieved the targeted NO  level of below 86 ng/J (0.2 lb/10  Btu).  Commercial offering of a
full scale furnace is not expected until at least 1983 (Reference 3-66).

       In summary, several promising advanced burner/furnace concepts are under development and may
       available in the next few years.  These techniqt
lower than current combustion modification techniques.
become available in the next few years.  These techniques may yield NO  emissions  substantially
3.1.4  Fuel Modification

       While not necessarily thought of as NC'X control  techniques, some additional  methods  either
are, or potentially will be, available for reducing NOV emissions in unique situations.   These
                              ,                        A
include various fuel modification techniques.  Three candidate techniques are fuel  switching, fuel
additives, and fuel denitrification.
                                            3-30

-------
3.1.4.1  Fuel  Switching
       In the past, the predominant reasons for fuel  switching were related to fuel  cost  or SCL
control.  However, conversion to a fuel  with a reduced nitrogen content or to one that  burns at  a
lower temperature may also result in a reduction in NOX emissions.   NOX emission  factors  for various
types of fuels were presented in Section 2.

       Natural gas firing is an attractive NO  control strategy because of the absence  of fuel NO
                                             X                                                  X
in addition to the flexibility it provides for the implementation of combustion modification
techniques.  However, for large fuel burning applications, fuel choice decisions  tend to  depend  on
fuel costs and other regulatory constraints, such as  the Powerplant and Industrial  Fuel Use Act  of
1978.  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.

       Western coals constitute one abundant alternate source of potentially low-NO  fuels.  The
direct combustion of western subbituminous coals in large steam generators may produce  lower NO
emissions than with combustion of eastern bituminous coals.  Three mechanisms could result in lower
NO  emissions:  first, western coals in general contain less bound nitrogen than  eastern  coals on a
unit heating value basis; second, the excess Op 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.  However, some studies have indicated that these factors may  be offset by the
higher fuel 02 levels in western coals.  These higher levels may lead to increased conversion of
fuel bound nitrogen to NO  (Reference 3-68).
                                                                                            \
       Some potential problems associated with burning low sulfur, high moisture  content  coals in
combustion equipment designed for higher quality coals are listed below (Reference 3-69):
                 -  Poor i-gnition;
                 -  Reduced boiler load capacity;
                 -  Increased carbon loss;
                 -  Boiler slagging/fouling; .
                 -  High superheat steam temperature;
                 -  Flame instability;
                 -  Increased boiler maintenance;
                 -  Reduced boiler efficiency; and
                 -  Reduced collection efficiency of electrostatic precipitator (ESP).
                                            3-31

-------
However, most of these operational problems can be solved with current boilers specifically  designed
to bum 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 gets scrubbing for S02 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 Amendments
                                .'
require that NSP5 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 reduced
or eliminated the advantage of fuel switching in many applications.

       A potential long-range option is the use of clean synthetic fuels derived from coal.
Candidate fuels include low to high Btu gas (3.7 to 30 MJ/Nm3, or 100 to 800 Btu/scf) and  synthetic
liquids and solids.  Two alternatives for utilizing low- and intermediate-Btu 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.  The NO  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 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 expected to be high in
fuel nitrogen content even though some denitrification may occur in the desulfurization process.
This high nitrogen content, carried over from the parent coal, would promote high  NOX emissions.
Other/potential alternate fuels that might be considered and their potential for fuel or thermal NOX
are listed in Table 3-3.
                     TABLE 3-3.  NOX FORMATION POTENTIAL OF SOME ALTERNATE FUELS
FUEL
Shale Oil
Coal -Oil Mixture
Coal -Liquid Mixtures
Methanol
Water-Oil Emulsion
Hydrogen
THERMAL NOX
Moderate
Moderate
Low
Low
Low
High
FUEL NOX
High
Moderate
Unchanged
Low
Unchanged
Low
               aFuel NO  is probably unchanged unless a significant amount of low
                nitrogen oil or methanol replaces part of the coal  on a heating
                value basis.
                                                3-32

-------
       Besides coal, shale oil  is another abundant source of nonpetroleum fossil  fuel  in the United
States.  However, current shale oil  production is very limited due to economics.   The  combustion of
shale oil will cause higher levels of fuel NO  because this fuel  generally contains bound nitrogen
                                             X
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 reduce
fuel costs for existing oil-fired boilers.  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.

       Other types of coal liquid mixtures are also under development, including coal-water, coal-
oil-water, and coal-alcohol.  The primary incentive for development of these  fuels is to obtain a
relatively cheap coal-based fuel for .combustion in existing oil-fired boilers.   However, some boiler
modifications are expected to be required to burn these fuels.  These fuels should burn at a lower
temperature than the parent coal thereby reducing thermal NOV formation.
                                                            X
       Methanol is currently produced from the synthesis of methane from natural  gas.   Some future
production may also come from synthetic gas generated from coal and biomass.   Baseline NO  emissions
                                                                                         X
from the combustion of methanol in an experimental 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 NO
emissions from methanol combustion were reduced to 10 ppm, or 15 percent of the baseline level
(Reference 3-70).
       In gas turbines 74 percent less NO  was produced using methanol, compared to distillate oil.
The hot wall experimental furnace showed a 20 percent increase in stack heat loss compared to a loss
of 14  percent for distillate oil (based on 115 percent theoretical air at a 473°K (390°F) stack
temperature).  For natural gas, turbine efficiency levels increased by 6 percent due to higher inlet
temperatures.
       Since water-oil emulsions affect only thermal NO  these alternate fuels have a definite NO
                                                       X                                         X
reduction potential when distillate oil is used (Reference 3-71).  NO  emission levels from
emulsions with approximately 50 mass percent water in distillate oil  approached the, levels, obtained
from methanol combustion (Reference 3-72).
                                               3-33

-------
       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
systems.  Thermal NO  levels are, however,.high when hydrogen meets with oxygen in the presence  of
atmospheric nitrogen.
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 NO  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 emissions
but some additives containing nitrogen increased NO formation (Reference 3-73).  Fuel  additives
reduced NOX emissions from gas turbines by an average of 15 to 30 percent but are not  attractive due
to added cost, serious operational difficulties and the presence  of the additives,  as  pollutants, in
the exhaust gas (Reference 3-74).  Average NO  reductions of 15 to 18 percent have  been recorded
using fuel additives in diesel engines (Reference 3-75).
3.1.4,3  Fuel Denitrification

       Fuel denitrification of coal or heavy oils could in principle be used to control  the
components of NO  emission due to conversion of fuel  bound nitrogen.  The most likely  use  of this
concept would be to supplement combustion modifications implemented for thermal NO  control.
                                                                                  X
Current technology for denitrification is limited to  the side benefits of fuel pretreatment to
remove other pollutants.  There is preliminary data to indicate that marginal  reductions in fuel
nitrogen result from oil desulfurization (Reference 3-76) and from chemical  cleaning or  solvent
refining of coal for ash and sulfur removal  (Reference 3-77).  The low denitrification efficiency  of
these processes does not make them attractive solely  on the basis of NO  control.  They  may prove
cost effective, however, on the basis of total environmental  impact.
                                              3-34

-------
3^1.5  Alternate Processes
       In the future, some alternate processes may be available which result in lower NO  emissions
than conventional stationary combustion technologies.  Some candidate processes include fluidized
bed combustion, catalytic combustion, repowering, and combined cycles.
3.1.5.1  Fluidized Bed Combustion
       In a fluidized bed combustor (FBC) combustion occurs in an air-supported bed of relatively
large (~3.2mm) coal ash and sand or limestone particles.  The temperature- in the bed is generally in
the range of 1,070°K to 1,270°K (1,500°F to 1,900°F) which makes the combustion process self-
sustaining.  The combustor may be at atmospheric pressure (A.FBC) or it may be pressurized (PFBC).

       A 30 MW AFBC pilot plant began operation in late 1976 (Reference 3-78).  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, DOE, and EPRI FBC programs can be found in
Reference 3-78.
       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, (3) lower combustion temperature resulting in less
fouling and corrosion and reduced NO  formation, (4) potentially efficient sulfur oxides control by
                                    A
direct contact of coal with an SOg 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 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 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.
       Nitrogen oxides 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 3-79 and 3-80).
                                                                       X            j
Experiments with nitrogen-free fuels resulted in NO  concentrations in agreement with equilibrium

                                           3-35

-------
values at the bed temperature.  However, coal-fired experiments resulted in NO  concentrations  in
                                                                              A
excess of the equilibrium values.  Furthermore, experiments using nitrogen-free gases  with  coal
yield substantially similar NOX levels as combustion in air (Reference 3-79).
       NOV emissions have been found to be slightly dependent on coal  particle size, the  type and
         X
amount of sulfur acceptor, the amount of excess air and the design of  the combustor itself.
Emission 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, NO  emissions from a PFBC are  well below the
current NSPS.  Results of 160 ng/J (0.37 lb/10  Btu) have been reported (Reference  3-78).

       In general, NO  control in F5C is a matter of good management of the normal  process
variables.  If more stringent standards are enacted, conventional NO  controls, such as flue gas
                                                                    X                          '
recirculation and off-stoichiometric combustion, may be used.  Exploratory results  indicate that
two-stage combustion could be advantageous for both NO  and SO  control.
                                                      A       X

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 formation of
NO .  Moreover, the catalyst itself serves to sustain the overall combustion process,  thereby
  A
minimizing the stability problems (References 3-81 and 3-82).  However, the overall  success of  a
catalytic combustion system in reducing, CO and unburned hydrocarbons (UHC) to  low levels  is a
function of both heterogeneous and gas phase reactions; surface reactions alone appear to be unable
to achieve the desired low levels.

       Emissions from catalytic experiments have typically been:  NO  <2 ppm,  UHC e4 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 3-82).
       At high temperatures, above 1,270°K (1,830°F), 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.
                                             3-36

-------
       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
low emissions (2 to 3 ppm NO , 20 to 30 ppm CO) were achieved for both fuels, and unburned hydro-
                            A                                                                   •
carbons were less than 1 ppm (Reference 3-83).   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
NOV, CO, and UHC, but at present, catalytic combustors are limited by the catalyst bed temperature
  X
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 in the future catalytic combustion concepts may
be incorporated into new gas turbine and residential, commericial, and industrial  heating designs.
                                                                              »•
3.1.5.3   Repowering

       Repowering adds a combustion turbine to  an existing steam plant,  providing additional
capacity 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
recovery 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,
requiring 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 3-84 and 3-85).
       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;
                                            3-37

-------
            -  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; and
            -  Gas turbines may be operated independently as peaking units, which provides
               greater plant flexibility.

       References 3-84 and 3-85 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 MM and reduced power cost to the consumer by 8 percent (Reference 3-86).   Under contract  from the
Electric Power Research Institute, Westinghouse Electric Corporation is evaluating repowering
conventional steam power plants without replacing the boiler.  Earlier pilot  scale work for  EPRI by
KVB Inc. shows a low NOV potential for repowering.   The boiler is fired fuel-rich using approxi-
                       A
mately 85 percent of the NOX bearing gas turbine exhaust as the combustion air.   The  remaining 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 percent 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 percent.
              X

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-fueled steam stations (Reference 3-87).  The combined gas and steam turbine system could
consist of a gas turbine firing a coal-derived fuel., which exhausts into an unfired waste-heat
recovery boiler.  In this system, a portion of the power would be generated by the gas turbine and a
portion by the steam boiler system.  Combined cycle efficiency improves significantly as the gas
turbine inlet temperature is increased.  At turbine inlet temperatures of  1,480°K (2,200°F), an
efficiency improvement of 2 percentage points per 55°K (100°F) increase in turbine inlet temperature
is found.
                                           3-38

-------
3.2    COMBUSTION FLUE GAS TREATMENT

       Historicany4 the major NOX,control  emphasis in the United States has been on combustion or
process modification.  However, in Japan where NO  emission standards are more stringent,  flue gas
                                                 X
treatment (FGT) technologies have undergone extensive development and implementation.  Recently, in
the U.S. several pilot and demonstration scale units have been built and operated.

       Flue gas treatment consists of any of several technologies designed to, remove or eliminate
NO  in the flue gas downstream of the combustion zone.  Several  FGT processes are potentially
capable of very high (greater than 80%) NO  removal efficiencies.  And since FGT processes are
located downstream of the combustion zone,  this NO  control can occur in addition to any control
                                                  X
already achieved with combustion modifications.  The application of FGT will usually involve
considerably more construction and expense  than the comparatively simple combustion  modifications.
But FGT may prove to be a viable control alternative for situations where very high  levels of
control are needed.

       These postcombustion processes can be divided into dry or wet types.   Some dry processes are
designed to control NO  alone while others  are designed to control SOV and NOV.  The wet processes
                      A                                              A       X
employ a wet scrubber to control both pollutants.  This subsection briefly describes some  of the
major FGT processes under development and how they work to control NO .  Most of the development of
FGT processes has been undertaken with application to utility boilers in mind.  Consequently, more
extensive discussions of FGT technology development status and process impacts are reserved to the
section on NOX control for utility boilers, Section .4.1.  Much of the material on FGT was  taken from
References 3-88 and 3-89; the reader is referred to these and other documents (e.g.  References 3-90,
3-91, and 3-92) for additional information.
3.2.1  Dry Flue Gas Treatment
       The dry processes can be categorized into four subdivisions:  catalytic reduction,
noncatalytic reduction, adsorption, and irradiation.  The majority of the dry processes are of the
reduction type.  These catalytic and noncatalytic reduction processes can also be classified as
selective or nonselective processes based on the type of reducing agent used. •  The majority are
selective and usually use NH3 as the reducing agent.  If the NH, is injected after the boiler
economizer, where'temperature of the flue gas is about 370°C to 450°C (700°F to 800°F), a catalyst
is necessary.  These processes are described as selective catalytic reduction (SCR) processes.  If
NH3 is injected into the secondary superheater region of the boiler, where temperature of the flue
gas approaches 980°C (1,800°F), a catalyst is not necessary.  These processes are described as
                                            3-39

-------
selective noncatalytic reduction (SNR) processes.   This subsection briefly describes  SCR,  SNR,  and
other dry F6T processes.

3.2.1.1  Selective Catalytic Reduction (SCR)
       The SCR method is the most advanced FGT method, and the one on which the overwhelming
majority of existing FGT units are based.  As with the majority of all FGT processes, most of the
SCR processes were developed in Japan.
       SCR systems use ammonia to selectively reduce nitrogen oxides.  The chemical mechanisms  can
best be summarized by the following gas phase reactions.

               4NO  + 4NH3 + 02  *   4N2 + BHgO                                           (3-6)
               2N02 + 4NH3 + 02  *   3N2 + 6H20                                           (3-7)

The first reaction predominates since flue gas NOX consists primarily of NO.  Oxygen  is in large
excess in the flue gas and does not limit the extent of reaction.

       A process flow diagram is shown for a NO  only SCR process in Figure 3-7.  Flue gas is taken
from the boiler between the economizer ^and air preheater.  Ammonia., taken from a liquid storage tank
and vaporized, is injected and mixed with the flue gas prior to the reactor.  The flue gas passes
through the catalyst bed where NO,, is reduced to N9.  The flue gas then exits the reactor and is
                                 A                ^
sent to the air preheater and, if necessary, additional pollutant control devices (e.g. FGD system,
ESP).
       With SCR systems it is desirable to treat flue gas exiting the economizer at 300 to 400°C,
prior to any air preheater, since it is at this temperature range that the catalysts  show the
optimum range of reactivity and selectivity (Reference 3-88).  Research and development on catalyst
formulations and shapes during recent years has resulted in some standardization among the catalyst
types offered.  A catalyst formula consisting primarily of oxides of titanium and vandium appears to
be universally used (Reference 3-92).  This formulation has proven to be resistant to poisoning.by
sulfur compounds in the flue gas.
       Reactor designs tend to vary depending on the application.  Catalyst pellets in a fixed bed
are commonly used for gas-fired applications.  For oil- or coal-fired applications where the flue
gas contains particulate matter, reactor designs usually incorporate honeycomb, pipe, or parallel
                                          3-40

-------
 Boiler
                      Economi zei
                          i
                                NH.
                        Air
                     Preheater
                   Air in
Reactor
       Gas to stack
       or additional
       pollution control
Figure 3-7.  Flow diagram for typical NO -only SCR process.
                                        X
                     3-41

-------
plate shaped catalysts which allow the flue gas to pass in parallel  along the catalyst surface.
Another design uses a moving bed arrangement.

       NOV emissions from boilers using SCR processes in Japan are generally reduced by 80 percent.
         X
Higher reductions are possible but costs are greater for these units (Reference 3-92).  As discussed
1n Section 4.1, there are presently over 60 full scale SCR units operating on gas- or oil-fired
boilers in Japan.  Also, two commercial units are operating on coal-fired boilers.  Construction  is
scheduled to be completed during 1981-1984 on at least 14 additional SCR units on coal-fired boilers
in that country.  In the U.S., SCR applications have been limited to a few pilot scale units on
coal-fired boilers and a demonstration scale unit under construction on an oil-fired utility boiler.

3.2.1.2  Selective Noncatalytic Reduction (SNR)

       Exxon Research and Engineering Corporation developed the SNR process in which NH, is injected
into the boiler where proper flue gas temperatures (about 900 to 1010°C, 1650 to 1850°F) allow the
reduction of NOV by reaction with NH, to proceed without a catalyst.  Generally, 40 to 60 percent
               A                    «J
NO., reduction is achieved with NH,:N()  molar ratios of 1:1 to 2:1.  SNR may be more attractive than
  X                              OX
SCR in cases where only 40 to 60 percent NO  is needed since SNR is simple and does hot require
expensive catalysts.

       The general disadvantage of SNR is the limited NO  control achievable, especially with larger
                                                        />
boilers.  This limited control generally results from the difficulty of achieving rapid uniform
mixing of NHj with the flue gas and from the variations of flue gas temperature and composition
usually present within the boiler region where SNR occurs.  NH, consumption and unreacted NH, levels
can also be high.

       There are several large SNR units installed in Japan, between 30- and 100-MW  capacity,
                                     *                                                      »
mostly supplied by Tonen Technology (a subsidiary of Toa JJenryo) which has a license from Exxon.
These units are operated on gas- and oil-fired boilers or furnaces.   Practically all are only for
emergency use during a photochemical smog alert or when total plant emissions exceed the regulation.

       There are presently two commercial SNR plants operating in the United States.  One is on a
glass melting furnace and the other a petroleum refinery, both located in California.  The construc-
tion of five other industrial-scale units is planned.  The SNR process is also being installed by
Exxon at the No.4 oil-fired unit of the Haynes Station of the Los Angeles Department of Water and
Power.
                                             3-42

-------
3.2.1.3  Other Dry FGT Processes





       In addition to SCR and SNR, other dry NOV control processes are being developed which will
                                   i            A


allow simultaneous control of SO .  These include:



          (1)  Activated carbon processes where NH, reduces NO,, to N,;
                                                  «j           X     &


          (2)  Copper oxide processes where NH. reduces NO  to N^; and



          (3)  Electron beam irradiation processes in which NH, is added



               to produce ammonium sulfate and nitrate.
     The optimum temperature range for simultaneous SO  and NO  control with activated carbon
                                                      X       A


processes is 220°C to 230°C (430°F to 445°F).  Although NOV may be adsorbed below 100°C (212°F), for
                             *                             A


treating large quantities of flue gas above 100°C the carbon is mainly useful as an NO  reduction



catalyst.  Therefore, while NO  is converted to N2 by reaction with NH3 in the presence of the



activated carbon catalyst, S02 is simultaneously adsorbed by the carbon to form H2SO,.  The H^SO,



may also compete for NH, in forming ammonium sulfate or bisulfate.  The formation of these ammonium



salts increases NH, consumption and also lowers catalyst activity.  The carbon must be regenerated,



either by washing or thermal regeneration.  Washing produces a dilute solution.  Concentration of



the solution to produce a fertilizer requires much energy.  Therefore, thermal regeneration seems to



be preferred.  A concentrated S02 gas is recovered, which can be used for sufuric acid or elemental



sulfur production.





       The major drawback of the activated carbon processes is the enormous consumption of activated



carbon, which is more expensive than ordinary carbon used only for SO  removal.  Since carbon and
                                                                     A


ammonia consumption increase with the S02 content of the flue gas, the process is best suited for



flue gases relatively low in S02«  In Japan' Sumitomo Heavy Industries and Unitika Company have



operated activated carbon pilot plants of 0.6 MW and 1.5 MW capacity respectively.





     The Shell Flue Gas Treatment process may simultaneously remove SO  and NOX.  SOX reacts with



the copper oxide acceptor to form copper sulfate.  The copper sulfate and copper oxide are SCR



catalysts for the NOV reduction by NH,.  Regeneration of the multiple catalyst beds by a reducing
                    X                «•'


gas, such as H-, yields a S02-rich stream that can be used to produce liquid S02> elemental sulfur,



or sulfuric acid.  By eliminating NH3 injection, the process is strictly an FGD process, whereas,



eliminating regeneration of the catalyst beds allows the process to be used for only NO  control.



The major disadvantages are the large consumption of fuel for making hydrogen and the catalyst



expense.
                                              3-43

-------
     In addition to the EPA-sponsored pilot plant mentioned earlier,  the process  has  been  installed
in Japan on a 40-MW oil-fired boiler.  The unit has demonstrated 90%  SOX removal  and  70% NOX
reduction.
     Another process for simultaneous SO  and NO  control  is the electron beam process  developed  by
Ebara Manufacturing Company in Japan.  NH, is added to the flue gas,  after which  the  gas stream is
irradiated with an electron beam in a reactor, promoting the conversion  of SO . NOV,  and ,NH,  to
                                                                             A    A         O
ammonium sulfate and ammonium nffrate.  The ammonium sulfate and aranonium nitrate may be collected
downstream in an ESP or baghouse and potentially sold as a fertilizer.   The most  economically
practical removal efficiency range appears to be 80% to 90% for each  of  NO  and SO ,  though higher
removals can be achieved with much greater electron beam energy input.   The optimum temperature
range is 70°C to 90°C (160°F to 195°F).

     Ebara has worked on the process since 1971.  It has been tested  at  a 0.3 MW  and  3  MW  scale in
Japan.  Avco Corporation in the United States has also examined this  technique and has  a
cross-licensing agreement with Ebara in sharing of technology and in  marketing of the process.
Although the process appears attractive because of simplicity, simultaneous SOV and NO   control,  and
                                           '                                   A      A
byproduct formation, there are still many questions concerning costs  and byproduct quality which
must be determined.

     Development of an alternate electron beam scrubbing process was  begun in 1979 by ,
Research-Cottrell under contract to the Department of Energy (DOE).   With this process  a lime spray
dryer is located upstream of the reactor.  Calcium sulfate and calcium nitrate are produced in the
reactor and caught in a downstream baghouse.   Some bench scale testing has been done  with  this
process.  DOE plans proof-of-cohcept scale testing of both the ammonia injection  and  lime  slurry
Injection electron beam processes on real coal-fired slip streams (Reference 3-93).

3.2.2  Wet Flue Gas Treatment

     The wet F6T processes normally involve simultaneous removal of SO  and NO .   The major problem
                                                                      X        A
associated with wet NO  control processes is the absorption of NO  by the scrubbing solution. NO
in the flue gas is predominantly NO, which is much less soluble than  NOp, whereas, N0~  is  even less
soluble than S0£.  The two common methods of removing the NO  in flue gas by wet  processes are:
(1) direct absorption of the NO  in the absorbing solution or (2) gas-phase oxidation to convert  the
                                           3-44

-------
relatively insoluble NO to NOp, followed by absorption of NOg.   Presently,  development of the wet
NO  FGT processes has practically ceased because of the complexity and unfavorable  economics  of
these processes in comparison with the dry processes (Reference 3-94).
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 NO  emissions  from nitric
acid manufacturing are estimated for the year 1980 at 100  Gg (110,000  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 NO .  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 NO  in  several states.  These  standards
                                                              X
were established in consideration of the then available technology, which was catalytic  reduction of
NO  to N£ 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 manufacture
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 NO  emissions without  add-on control  equipment. 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  reduction,
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 NOX Pollution Abatement at New Nitric Acid Plants

       Nitric acid is manufactured in the United States by the catalytic oxidation of ammonia over a
platinum catalyst with the subsequent absorption of the product gases, primarily N02 and NO, by
                                            3-45

-------
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 M02 to HNO, allow for smaller equipment for both
the main process plus any tail gas treatment required to meet emission standards.  Currently 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 absorber 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
          X
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 conversion of NO, to nitric acid and minimize NO  emissions.  However,
                                            fc                               A
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
liquid phases.  With higher 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
distillation and then liquified as NO.
                                              3-46

-------
       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
(760 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 3-95).  Tail gas emissions from this process are within the 1.5 g/kg (3 Ib/ton) NO
                                                                                             X
regulation.  This occurs primarily by ensuring oxidation to NO- 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 Reactivi-
cation) process.  Ammonia combustion occurs at near atmospheric pressure and at 1,120°K (1,560°F)
with the usual waste-heat boiler, tail gas preheater, 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 product is
produced by vacuum distillation.
3.3.2  Retrofit Design for NOX Pollution Abatement at New or Existing Nitric Acid Plants

       Most existing nitric acid plants were not designed with the present NO  emission standards in
                                                                             A
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 NOX 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 NO^/kg of acid product (3 Ib/ton) emission standard.

3.3.2.1  Chilled Absorption

       The basic principle involved is that the amount of NO  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 NOV reduction has only provided marginal  results
                                                   A
and has had problems in continuously meeting the NSPS, especially in warm weather.  Refrigeration
requirements can prove costly, both in equipment and energy use.
                                                3-47

-------
3.3.2.2  Extended Absorption

       One of the most commonly used retrofit processes, which has been used effectively  to  meet  the
HSPS, 1s extended absorption.  Figure 3-8 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  compression
                         *
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 (100 psia) is preferred to make the economics of this method attractive.

3.3.2.3  Wet Chemical Scrubbing
       Wet chemical scrubbing removes NO  from nitric acid plant tail gases  by chemical reaction.
Liquids such as alkali hydroxide solutions, ammonia, urea, and potassium permanganate convert NO*  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
1s a valuable byproduct which can be reclaimed and sold as fertilizer.

Caustlc Scrubbi ng

       In this process, NO  in the tall gas reacts with sodium hydroxide, sodium  carbonate,  or
ammonium hydroxide to form nitrite 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.

Urea Scrubbing
       Urea can be used to treat all gases for NO  control since it reacts rapidly with nitrous
acid.  Nitrogen dioxide, N02 reacts with water to form both nitric acid (HN03) and nitrous acid
(HN02) in equal proportions.  Nitrous acid will rapidly decompose to form NO and  N02-  Urea
(CO (NHpJo) when contacted with the tail gas will absorb N0« indirectly as nitrous acid to form
ammonium nitrate, NH.NOg and free nitrogen, N^.  By depleting .the liquid phase of nitrous acid the
equilibrium conversion of nitric oxide, NO, to nitrogen dioxide occurs to remove  NO also. The
                                                3-48

-------
            wroonla
                 compressor
                            converter
*»
ID
                                             absorber
j-^i»
                                 heat      condenser
                                 recovery
              A

               I
               I
               I


        extended P~
 X" ~ ~ T absorber I

A   |      J   '
                    tall gas.^
               I
               I

 ^-jEyip;--^

   product Lfll
    pump
                     power
                     recovery
                                                                      product
                                                                      •acid

             Figure 3-8.  Extended absorption system on existing nitric acid plant.

-------
 result  is conversion of N02 to either free nitrogen which is vented to the atmosphere or ammonium
 nitrate which is sold as a 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 solution
'(NH.NO-) which can be recovered and sold as fertilizer.  This process can be applied to tail gas
   *r  O
 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 process
 uses  both physical absorption and  stripping and chemical oxidation absorption.  The process uses
 only  water and nitric acid and converts nitrogen oxides  in the tail gas to nitric acid at concentra-
 tions which can be commercially utilized (Reference 3-96).

 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
 nitri'te, and potassium nitrite.  The potassium permanganate  is regenerated by oxidizing the
 potassium sodium manganate electrolytically (References  3-97 and 3-98).

 3.3.2.4 Catalytic Reduction
        There are three types of catalytic reduction processes used for NO  control:  nonselective
 reduction, which removes both NOV  and oxygen; selective  reduction, which removes only NOV; and
                                X                                                       A
 heterogeneous catalysis used in conjunction with wet scrubbing.  Each of these will be discussed in
 the  following paragraphs.

 Nonselective Catalytic Reduction
        The nonselective reduction  process reacts NO  with Hp or CH^ to yield Ng, COp and HpO.  The
 process is called nonselective because  the reactants first deplete all the oxygen present in the
                                            3-50

-------
tail gas, and then remove the NO .  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 NO  abatement.  Decolorization and power recovery units reduce NOx and 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                                         (3-8)

                    CH4 + 202  + C02 + 2H20                                               (3-9)

                    CH4 + 4NO  - 2N2 + C02 + 2H20              .                           (3-10)

       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
reduction reaction (3-8) is faster than reduction reaction (3-9), but abatement reaction (3-10) is
much slower than reaction (3-9).  Thus, decolorization can be accomplished by adding just enough
fuel for partial oxygen burnout.  If NO  abatement is required, however, sufficient fuel must be
                                       x               N
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
                                      A                                £
areas near the plant is reduced substantially.

       NOX 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.
                                           3-51

-------
       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 NO  control.   The availability and
cost of natural gas, increasing catalyst cost and poor performance have led  to a decline  in  interest
1n this process.

Selective Catalytic Reduction
       In selective catalytic reduction, ammonia is reacted with the NO  to  form N2- The oxygen in
the tall 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 483°K to 544°K (410°F to 520°F).  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 HgO                                       (3-11)

                         4NH3 + 6NO  * 5N2 + 6H20                                         (3-12)

Above 544°K, ammonia may oxidize to form NOX; below 483°K, it may form ammonia nitrate.
       Selective oxidation with ammonia has several advantages over nonselective reduction:
                 -  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 20°K to 30°K (36°F to
                    54°F) so that energy recovery equipment, such as a waste heat boiler  or
                    high temperature gas turbine, is not required; and
                 -  Lower raw material costs since the amount of ammonia required is
                    approximately equal to the molar equivalent amount of NOV abated.
                                                                            X

Heterogeneous Catalysis
       One wet scrubber process uses heterogeneous catalysis in a packed column to  oxidize NO to N02
(References 3-99 and 3-100).  This system is currently in  the development stage.
                                           3-52

-------
3.3.2.5  Molecular Sieve Adsorption
       One method of NO  control  involves the adsorption of NO  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  N02-
The required regenerating gas is  obtained by using a portion of the treated tail  gas stream to
desorb the adsorbed N02 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  process
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.
                                            3-53

-------
                                      REFERENCES FOR SECTION 3


3-1       Lira, K.J., et al., "Environmental Assessment of Utility Boiler Combustion  Modification NO
          Controls:  Wlume 1.  Technical Results," EPA-600/7-80-075a, April  1980.   pp.  4-1  to 4-56.

3-2       Zeldovich, J., "The Oxidation of Nitrogen in Combustion and Expolsions," Acta  Physiochem
          URSS, (Moscow), Vol. 21, p. 4, 1946 (as cited in Reference 3-1, Section 4~T

3-3       Bowman, C.T. and Seery, D.J., "Investigation of NO Formation Kinetics  in Combustion
          Processes:  The Methane-Oxygen-Nitrogen Reaction," In Emissions from Continuous  Combustion
          Systems, Cornelius, W. and Agnew, W.6., eds., Plenum, 1972 (as cited in Reference  3-1,
          Section 4).

3-4       Bartok, W., et al., "Basic Kinetic Studies and Modeling of NO Formation in Combustion
          Processes," ftlCnT Symposium Series No. 126, Vol. 68, 1972 (as cited in Reference 3-1,
          Section 4).

3-5       Halstead, C.J. and Munro, A.J.E., "the Sampling, Analysis, and Study of the  Nitrogen
          Oxides Formed in Natural Gas/Air Flames," Company Report, Shell Research,  Egham, Surrey,
          U.K., 1971, (as cited in Reference 3-1, Section 4).

3-6       Thompson, D., et al., "The Formation of Oxides of Nitrogen in a Combustion System,"
          presented at th~e TUth National AIChE Meeting, Atlantic City, 1971,  (as cited in
          Reference 3-1, Section 4).

3-7       Lange, H.B., "NO  Formation in Premixed Combustion:  A Kinetics Model  and  Experimental
          Data," presented at the 64th Annual AIChE Meeting, San Francisco,  1971 (as cited in
          Reference 3-1, Section 4).

3-8       Sarofim, A.F. and Phol, J.H., "Kinetics of Nitric Oxide Formation  in Premixed  Laminar
          Flames," 14th Symposium (International) on Combustion, The Combustion  Institute,
          Pittsburg, 1973 (as cited in Reference 3-1, Section 4).

3-9       Iverach, D., et al., "Formations of Nitric Oxide in Fuel-Lean and  Fuel-Rich  Flames,"
          ibid., 1973 (as cfted in Reference 3-1, Section 4).

3-10      Wendt, J.O.L. and Ekmann, J.M., "Effect of Fuel Sulfur Species on  Nitrogen Oxide Emissions
          from Premixed Flames," Comb. Flame, Vol. 25, 1975 (as cited in Reference 3-1,  Section 4). .

3-11      Malte, P.C. and Pratt, D.T., "Measurement of Atomic Oxygen and Nitrogen Oxides in  Jet-
          Stirred Combustion," 15th Symposium (International) on Combustion,  The Combustion
          Institute, Pittsburgh, 1975 (as cited in Reference 3-1, Section 4).

3-12      Mitchell, R.E. and Sarofim, A.F., "Nitrogen Oxide Formation in Laminar Methane Air
          Diffusion Flames," presented at the Fall Meeting, Western States Section,  The  Combustion
          Institute, Palo Alto, California, 1975 (as cited in Reference 3-1,  Section 4).

3-13      Bowman, C.T., "Non-Equilibrium Radical Concentrations in Shock Initiated Methane
          Oxidation," 15th Symposium (International) on Combustion, The Combustion Institute,
          Pittsburg, 1975 (as cited in Reference 3-1, Section 4).

3-14      Fenimore, C.P., "Formation of Nitric Oxide in Premixed Hydrocarbon  Flames,"  13th Symposium
          (International) on Combustion, The Combustion Institute, Pittsburgh, 1971  (as  cited in
          Reference 3-1, Section 4).

3-15      MacKinnon, D.J., "Nitric Oxide Formation at High Temperatures," Journal of the Air
          Pollution Control Association, Vol. 24, No. 3, pp. 237 to 239, March 1974  (as  cited in
          Reference 3-1, Section 4).

3-16      Heap, M.P., e_t al_., "Burner Criteria for NO  Control; Volume I ~  Influence  of Burner
          Variables on NO  in Pulverized Coal Flames,8 EPA-600/2-76-061a, NTIS-PB-259-911/AS,
          March 1976, (as cited in Reference 3-1, Section 4).

3-17      Bowman, C.T., ejt al_., "Effects of Interaction Between Fluid Dynamics on Chemistry  or
          Pollutant Formation in Combustion," In:  Proceedings of the Stationary Source Combustion
          Symposium; Volume I — Fundamental Research, EPA-600/2-76-152a. NTIS-PB-256-320/AS.
          June 1976, (as cited in Reference 3-1, Section 4).
                                                 3-54

-------
3-18      Shaw, J.T. and Thomas, A.C., "Oxides of Nitrogen in Relation to the Combustion of Coal,"
          presented at the 7th International  Conference on Coal  Science,  Prague,  June 1968 (as  cited
          in Reference 3-1, Section 4).

3-19      Pershing, D.W., et al., "Influence  of Design Variables on the Production of Thermal and
          Fuel MO from Residual Oil and Coal  Combustion,"  AIChE  Symposium Series, No. 148, Vol.  71,
          pp. 19 to 29, 1975 (as cited in Reference 3-1, Section 4).

3-20      Thompson, R.E. and McElroy, M.W., "Effectiveness of Gas Recirculation and Staged
          Combustion in Reducing NO  in a 560-MW Coal-Fired Boiler,"  EPRI FP-257, NTIS-PB-260-582,
          September 1976 (as cited Sn Reference 3-1, Section 4).

3-21      Sorofim, A.F., et al., "Mechanisms  and Kinetics  of NO   Formation:   Recent Developments,"
          presented at the 6lTfh Annual AIChE  Meeting, Chicago, November 1976  (as  cited in
          Reference 3-1, Section 4).

3-22      Martin, 6.B. and Berkau, E.E., "An  Investigation of the Conversion  of Various Fuel
          Nitrogen Compounds to Nitrogen Oxides in Oil Combustion," presented at the 70th National
          AIChE Meeting, Atlantic City, August 1971 (as cited in Reference 3-1, Section 4).

3-23      Hablet, W.W. and Howe11, B.M., "Control of NO  Formation in Tangentially Coal-Fired Steam
          Generators," In:  Proceedings of the NO Control  Technology Seminar, EPRI SR-39,
          NTIS-PB-253-661, February 1976 (as  cited in Reference  3-1,  Section  4).

3-24      "Air Quality and Stationary Source  Emission Control,"  U.S.  Senate,  Committee on Public
          Works, Serial No. 94-4, March 1975  (as cited in  Reference 3-1,  Section  4).

3-25      Pohl, J.H. and Sarofim, A.F., "Fate of Coal Nitrogen During Pyrolysis and Oxidation,"  In:
          Proceedings of the Stationary Source Combustion  Symposium;  Volume I --  Fundamental
          Research, EPA-600/2-76-152a, NTI5-PB-256-320/AS. June  1976  (as  cited in Reference 3-1',
          Section 4).

3-26      Heap, M.P., et jil_., "The Optimization of Burner Design Parameters to Control NO  Formation
          in PulverizecTCoal and Heavy Oil Flames," In:  Proceedings  of the Stationary Source
          Combustion Symposium; Volume II —  Fuels and Process Research and Development,
          EPA-600/2-76-152b, NTIS-PB 256 321/AS, June 1976 Reference 3-1, Section 4).

3-27      Pohl, J.H. and Sarofim, A.F., "Devolatilization  and Oxidation of Coal Nitrogen," presented
          at the 16th Sympsoium (international) on Combustion, Cambridge, Massachusetts, August 1976
          (as cited in Reference 3-1, Section 4).

3-28      Blair, D.W., et^ al_., "Devolatilization and Pyrolysis of Fuel Nitrogen from Single Coal
          Particle Combustion," 16th Symposium (International) on Combustion, Cambridge,
          Massachusetts, August 1976 (as cited in Reference 3-1, Section  4).

3-29      Brown, R.A., et al., "Investigation of Staging Parameters for NO  Control in Both Wall and
          Tangentially UoaT^fired Boilers," In:  Proceedings of  the Second Stationary Source
          Combustion Symposium: Volume III. New Orleans, EPA-600/7-77-073C, NTIS-PB-271-757/AS,
          July 1977 (as cited in Reference 3-1, Section 4).

3-30      Pershing, D.W., "Nitrogen Oxide Formation in Pulverized Coal Flames," Ph.D, Dissertation,
          University of Arizona, 1976 (as cited in Reference 3-1, Section 4).

3-31      Axworthy, A.E., Jr., "Chemistry and Kinetics of Fuel Nitrogen Conversion to Nitric Oxide,"
          AIChE Symposium Series. No. 148, Vol. 71, pp. 43 to 50, 1975 (as cited in Reference 3-1,
          Section 4).

3-32      Axworthy, A.E., e_t jil_., "Chemical Reactions in the Conversion of Fuel Nitrogen to NO  ,"
          In:  Proceedings of the Stationary  Source Combustion Symposium, Volume I,          x
          EPA-600/2-76-152a, NTIS-PB-256-320/AS, June 1976 (as cited in Reference 3-1, Section  4).

3-33      Pershing, D.W. and Wendt, J.O.L., "The Effect of Coal  Combustion on Thermal and Fuel  NO
          Production from Pulverized Coal Combustion," presented at Central  States Section,  The
          Combustion Institute, Columbus, Ohio, April 1976 (as cited in Reference 3-1, Section  4).
                                               3-55

-------
3-34      Habelt, W.W. and B.H. Howell, "Control of NO Formation in Tangentially Coal-Fired Steam
          Generators," In:  Proceedings of the NO  Control Technology Seminar, EPRI SR-39,
          February 1976.                      ;*'

3-35      Barr, W.H., and D.E. James, "Nitric Oxide Control - A Program of Significant Accomplish-
          ments," ASME 72-WA/Pwr-13.

3-36      Barr, W.H., et^l_., "Retrofit of Large Utility Boilers for Nitric Oxide Emissions
          Reduction - "Experience and Status Report."

3-37      Crawford, A.R., et a±., "Field Testing:  Application of Combustion Modifications  to
          Control NO  Emissions from Utility Boilers-," Exxon Research and Engineering Co.,
          EPA-650/2-74-066, June 1974.

3-38      Crawford, A.R., et £l_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.

3-39      Blakeslee, C.E., and H.E. Burbach, "Controlling NO  Emissions from Steam Generators,"
          C.E. Inc., APCA 72-75, 65th Annual Meeting of the Air Pollution Control Association,
          June 18-22, 1972.

3-40      Norton, D.M., et al., "Status of Oil-Fired NO  Control Technology," In:  Proceedings of
          the NO  ControTTic'hnology Seminar, EPRI SR-3$, NTIS-PB 253 661, February 1976  (as cited
          1n Reference 3-1, Section 4).

3-41      Campobenedetto, E.J., Babcock & Wilcock Co., letter to Acurex Corp., November 15, 1977 (as
          cited in Reference 3-1, Section 4).

3-42      Durrant, O.W., "Pulverized Coal — New Requirements and Challenges," presented  to ISA
          Power Instrumentation Symposium, Houston, Texas, May 1975 (as cited in Reference  3-1,
          Section 4).

3-43      Vatsky, J., "Attaining Low NO  Emissions by Combining Low Emission Burners and  Off-
          Stoichiometric Firing," presented at the 70th Annual AIChE Meeting, New York,
          November 1977 (as cited in Reference 3-1, Section 4).

3-44      Rawdon, A.H. and Johnson, S.A., "Control of NO  Emissions from Power Boilers,"  presented
          at Annual Meeting of the Institute of Fuel, Adelaide, Australia, November 1974  (as cited
          in Reference 3-1, Section 4).

3-45      Tidona, R.J., W.A. Carter, and H.J. Buening, (KVB) "Application of Advanced Combustion
          Modifications to Industrial Process Equipment — Process Heater Tests" (Preliminary draft
          report prepared for Office of Research and Development, U.S. Environmental Protection
          Agency.)  Research Triangle Park, North Carolina.  EPA Contract No. 68-02-2645.
          November 1981.

3-46      Campobenedetto, E.J., "The Dual Register Pulverized Coal Burner — Field Test Results,"
          presented to Engineering Foundation Conference on Clean Combustion of Coal, New Hampshire,
          August 1977 (as cited in Reference 3-1, Section 4).

3-47      Barr, W.H., et aj.., "Modifying Large Boilers to Reduce Nitric Oxide Emissions," Chemical
          Engineering "Progress, Vol. 73, pp. 59 to 68, July 1977 (as cited in Reference 3-1,
          Section 4).

3-48      U.S. Environmental Protection Agency.  Control Techniques for Nitrogen Oxides Emissions
          from Stationary Sources — Second Edition.  (Prepared for* U.S. Environmental  Protection
          Agency.)  Research Triangle Park, N.C.  Publication No. EPA-450/1-78-001.  January 1978.
          pp. 3-19 (as cited in Reference 3-1, Section 4).

3-49      Brown, R.A., H.B. Mason, and R.J. Schrfeber, "Systems Analysis Requirements for Nitrogen
          Oxide Control of Stationary Sources," Environmental Protection Technology Series
          EPA-650/2-74-091, September 1974.
                                            3-56

-------
3-50      U.S. Environmental  Protection Agency,  "Draft - NO  SIP Preparation  Manual  Volume  II  -
          Support Sections,"  Office of Air Quality Planning and Standards,  Research  Triangle
          Park, N.C., April,  1976.

3-51      Bell, A.W., et a\_., "Nitric Oxide Reduction by Controlled Combustion  Processes,"
          KVB, Inc., Western  States Section/Combustion Institute, April  20-21,  1970  (as  cited  in
          Reference 3-1, Section 4).

3-52      Crawford, A.R., E.H. Manny and U. 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."

3-53      Ando, J. and T. Heiichiro, "NO  Abatement for Stationary Sources  in Japan," Environmental
          Protection Technology Series, £PA-600/2-76-013b, January 1976.

3-54"    • Shoffstall, D.R., "Burner Design Criteria for Control of Pollutant  Emissions from Natural
          Gas Flames," Institute of Gas Technology, EPA-600/2-76-152b, June 1976.

3-55      Koppang, R.R., "A Status Report on the Commercialization and Recent Development History of
          the TRW Low NOV Burner," TRW Energy Systems Group.
                        X

3-56      Tsuji, S., et al.,  "Control Technique for Nitric Oxide - Development  of New Combustion
          Methods," IHI Engineering Review, Vol. 6, No. 2.

3-57      Ando, J., et al., "NOV Abatement for Stationary Sources in Japan,"  August  1976
          (Preliminary Draft). x
3-58
Heap, M.P., et al., "The Optimization of Burner Design Parameters to Control  NO  Formation
in PulverizecTCoal and Heavy Oil  Flames," In:   Proceedings of the Stationary  Source
Combustion Symposium, EPA-600/2~76-152b, June  1976.
3-59      Barsin, J.A., "Pulverized Coal Firing NO  Control," In:   Proceedings:   Second NO  Control
          Technology Seminar, Electric Power Research Institute, Report No.  FP-1109-SR, Pafo Alto,
          California, July 1979 (as cited in Reference 3-1, Section 4).

3-60      Peduto, E.F., R.R. Hall, and G. Tucker, "Characterization of the NO  and S02  Control
          Performances:  Southern Indiana Gas and Electric Co., A.B.  Brown Unit No.  1,  Volume 1  ~
          Program Results," Prepared for U.S. Environmental Ptoection Agency, Washington, D.C.,  EPA
          Contract No. 68-02-3168, March 1982.

3-61      Vatsky, J., "Experience in Reducing NO  Emissions on Operating Steam Generators,"  In:
          Proceedings:  Second NO  Control  Technology Seminar, Electric Power Research  Institute,
          Report Nol FP-1109-SR, Palo Alto, CA, July 1979 (as cited in Reference 3-1, Section 4).

3-62      "NO  Control Review," Vol. 2, No. 4, EPA Industrial Environmental  Research Laboratory,
          RTP, North Carolina, Fall 1977 (as cited in Reference 3-1,  Section 4).

3-63      Stavern, D.V., "The Coen Low Excess Air Burner," presented at the NO  Control Technology
          Workshop, Pacific Grove, California, October 1977 (as cited in Reference 3-1, Section  4).

3-64      Vatsky, J., "Larger Burners and Low NO ," Heat Engineering, Vol. 49, No. 2, pp. 17-25,
          April-June 1979 (as cited in Reference 3-1, Section 4).

3-65      Martin, G.B., "Field Evaluation of Low NO  Coal Burners on Industrial  and Utility
          Boilers," In:  Proceedings of the Third Stationary Source Combustion Symposium, Volume I,
          EPA-600/7-79-050a, February 1979 (as cited in Reference 3-1, Section 4).

3-66      Johnson, S.A., et al., "The Primary Combustion Furnace System -- An Advanced  Low-NO
          Concept for Pulverized Coal Combustion," In:  Proceedings:   Second NO  Control Technology
          Seminar, Electric Power Research Institute, Report No. FP-1109-SR, PaTo Alto, California,
          July 1979.

3-67      Whitaker, R., "Trade-offs in N0x Control," EPRI Journal, January-February 1982, pp. 18-25.

3-68      Habelt, W.W.  The Influence of the Coal Oxygen to Coal Nitrogen Ratio on NO  Formation.
          Presented at the 70th Annual AIChE Meeting.  New York.  November 13-17, 1977.
                                              3-57

-------
3-69      Ctvrtnicek, I.E., et aT_., "Evoluation of Low-Sulfur Western Coal  Characteristics,
          Utilization and Combustion Experience," Monsanto Research Corp.,  EPA-650/2-75-046,
          Hay 1975.

3-70      Martin, 6.B., "Evaluation of NO  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.

3-71      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.

3-72      Martin, 6.B., "Environmental Considerations in the Use of Alternate  Clean Fuels  in
          Stationary Combustion Processes."

3-73      Martin, 6.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.

3-74      Shaw, H., "Reduction of Nitrogen Oxide Emissions from a Gas Turbine  Combustor  by Fuel
          Modifications," ASME Transactions, Journal of Engineering for Power,  95,  4, October  1973.

3-75      Altwicker, E.R., et al., "Pollutants from Fuel Oil Combustion and the Effects  of
          Additives," Paper No7"71-14, 64th Annual APCA Meeting, Atlantic City, N.J., June 1971.

3-76      Barrett, R.E., et al., "Field Investigation of Emissions from Combustion  Equipment for
          Space Heating," EP/PR2-73-084a, June 1973.

3-77      Frey, D.J., "De-Ashed Coal Combustion Study,"  Combustion Engineering, Inc., October  1964.

3-78      Energy Research and Development Agency, "Proceedings of the Fourth International
          Conference on Fluidized Bed Combustion," McLean, VA., December 1975.

3-79      Jonke, A.A., et a\_., "Pollution Control Capabilities, of Fluidized Bed Combustion," AIChE
          Symposium Sertes No. 126, Vol. 68, 1972.

3-80      Chronowski, R.A., and B. Molayem, "NOV Emissions from Atmopsheric Fluidized-Bed  Boilers,"
          ASME-75-PWR-4, October 1975.         x

3-81      Pfefferle, W.C., e£ al., "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.

3-82      Kesselring, J.P., et al., "Catalytic Oxidation of Fuels for NOV Control for Area Sources,"
          EPA Report, EPA-60^72^76-037, February 1976.                  x

3-83      DeCorso, S.M., et al., "Catalysts for Gas Turbine Combustors - Experimental Test Results,"
          paper presentecTaY~A"SME Gas Turbine Conference and Products Show, Mew Orleans, March 1976,
          ASME Paper I76-GT-4.

3-84      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.

3-85      Ahuja, A., "Repowering Pays Off for Utility and Industrial Plants,"  Power Engineering,
          pp. 50-54, July 1976.

3-86      Stambler, I., "Repowering Gives Glendale Extra 75 MW and Lower Rates," Gas  Turbine World,
          September 1977.

3-87      Robson, F.L., and A.J. Giramonti, "The Use of Combined-Cycle Power Systems  in  Nonpolluting
          Central Stations," JAPCA, Vol. 22, pp. 177-180, 1972.

3-88      Jones, G.D. and K.L. Johnson, "Technology Assessment Report for Industrial  Boiler
          Applications:  NOX Flue Gas Treatment," EPA-600/7-79-178g, December  1979.

3-89      Maxwell, J.D. and L.R. Humphries, "Evaluation of the Advanced Low-NO  Burner,  Exxon, and
          Hitachi Zosen DeNOx Processes," EPA-600/7-81-120, TVA/OP/EDT-81-28,  3u1y  1981.
                                              3-58

-------
3-90      Ando, J., "NOX Abatement for Stationary Sources  in  Japan,"  EPA-600/7-79-205, August 1979.

3-91      Faucett, H.L., J.D.  Maxwell, ,and T.A.  Burnett,  "Technical Assessment of  NO  Removal
          Processes for Utility Application,"  TVA Bulletin Y-120,  EPA-600/7-77-127  (NTIS
          PB-276-637/6WP), EPRI FP-1253,  1977.

3-92      Jones, G.D. "Selective Catalytic Reduction and  NO  Control  in  Japan, A Status Report,"
          EPA-600/7-81-030, January 1981.                  x

3-93      Trexler, E.C. "DOE's Electron Bream  Irradiation  Developmental  Program,"  prepared for
          Seventh Symposium on Flue Gas Desulfurization, May  17-20, 1982,  p.  11.

3-94      Reference 3-89, p. 11.

3-95      "Nitric Acid from Ammonia,"  Hoechst-Uhde Corp.  brochure  (FWC 11  619),
          Englewood Cliffs, N.J.

3-96      Mayland, B.J., "The CDL/VITOK Nitrogen Oxides Abatement  Process," Chenoweth Development
          Laboratory, Louisville, KY.

3-97      "New System Knocks NOX Out of Nitric," Chemical  Week,  September  3,  1975,  pp. 37-38.

3-98      "NO  Removal System Now Available,"  Wet Scrubber Newsletter, September 30, 1973, pp. 3-4.

3-99      Mayland, B.J., "Application of the CDL/VITOK Nitrogen  Oxide Abatement Process," presented
          at Sulfur and Nitrogen Symposium, Salford, Lancashire, U.K., April  1976.

3-100     Mayland, B.J., and R.C. Heinze, "Continuous Catalytic  Absorption for NO   Emission
          Control," Chemical Engineering Process, Vol. 6,  May 1973, pp.  75-76.
                                                     3-59

-------

-------
                                              SECTION 4
                               LARGE FOSSIL FUEL COMBUSTION PROCESSES
       Fossil  fuel  combustion  by utilities  and  industry  accounted  for about  85 percent of  NO
emissions from stationary sources in 1980 (see Section 2).  This  combustion  occurred principally in
                               s
boilers and  large internal combustion  (1C)  engines.   The large boiler category encompasses applica-
Jion to utility and  industrial  power generation and industrial process steam  generation.   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 experience,  and energy and environmental  impacts of  the
implementation of NO  controls on these equipment categories.

4.1  UTILITY BOILERS
       Most  of  the  nation's  electricity  is generated  in large fossil-fueled  central  station  power
plants, which  primarily consist  of  high-pressure watertube  boilers  in  the 100 to  1300  MW*  range
serving turbine-generators.   Firing  capacities  of individual   burners in utility  boilers commonly
have thermal inputs as high as 29-58 MW* (100-200 x 106 Btu/hr).  A 1000 MW* opposed wall-fired unit
may require  as many as 60 separate burners.

       Although there are differences among utility  boiler  designs  in  furnace volume,  operating
pressure,  and  configuration  of  internal  heating transfer  surface,  the  principal  distinction is
firing  mode.  Firing  mode is  characterized  by  the type of  firing  equipment,  the  fuel  handling
system, and  the placement of the burners on the furnace walls (see Section 2.3.1).

       The total NOX emitted in 1980 by the electric utility industry was 6.4 Tg (7.1 x 106 tons) or
58 percent  of  the  total  stationary  source  emissions  of NO  .  Coal-fired  boilers  accounted  for
approximately  81  percent of the  total utility  emissions.   A  more  detailed  emission breakdown is
presented  in Section 2.  Ranges  of .uncontrolled NO  emissions for  three principal  types  of  coal-
                                                    X
fired boilers are presented  in Figure  4-1.  These emission rates essentially represent baseline
*Utility boilers are commonly described in terms of electrical output rating rather than in terms of
 thermal input.  Burners are commonly described in terms of thermal input to the boiler as noted by
 the subscript "t".  This convention will be used throughout Section 4.1 unless otherwise noted in
 the text.
                                                4-1

-------
    1600
    1400
     1200
6   1000
Cu
a.
 CM
o
en
O
z
     800
     600
     400
     200
        0
                          I        I

                          Wet bottom
Wall
                          I
                200     .400      600


                             MW
                                800
1000
    Figure 4-1.  Baseline HO  emissions - coal-fired  utility
                 boilers (Reference 4-1).
                              4-2

-------
emissions from units designed prior to the early 1970s, without the combustion modification controls
often  designed into  later  units.   As  shown,  cyclone-fired  boilers  typically  have  the  highest
uncontrolled  emission  rates,  tangentially-fired  units  the  lowest,  with  one wall-  and  opposed
wall-fired units at an  intermediate  level.   The old wet bottom (slag tap) units are  exceptions  and
have very  high NO  emissions.   All  of  these  boilers fire  pulverized  coal  except for  the cyclone
units which fire crushed coal.  Very few stokers are used in utility applications.

4.1.1  Control Techniques
       Techniques  for  controlling  NOX  emissions   from   utility   boilers   fall   into   two  broad
classes; combustion modifications  and  flue gas treatment (F6T).  Considerable  experience exists in
the U.S. with  various combustion modification controls.  Flue gas treatment techniques, on the other
hand,  have  primarily  been developed and applied  in Japan,  with some testing performed  in the U.S.
This section  discusses  the  status of  application of  these  techniques  to utility  boilers,  the  NO
control performance achieved, and the process impacts associated with their application.

4.1.1.1  Combustion Modifications
       The  general concept  of  combustion  modification  as  potential  NO   control  techniques  for
stationary  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 NO  sources and their control  flexibility,  are
the  most extensively modified  stationary  equipment  type.   The  selection  and  implementation  of
effective NOV  controls  for a specific utility boiler  are generally  dependent on several  variables.
            X
These  include the furnace  characteristics  (i.e.,  geometry and operational  flexibility), fuel/air
handling systems and  automatic controls, and 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.
       Combustion modifications which have been applied or proposed for utility boilers include:
                 -  Low Excess Air Firing (LEA);
                 -  Off Stoichiometric Combustion (OSC) or Staged Combustion (SC),
                    including Biased Burner  Firing (BBF), Burners Out of Service (BOOS),
                    and Overfire Air (OFA);
                 -  Flue Gas Recirculation  (FGR);
                 -  First Generation Low NOX Burners (LNB);
                 -  Enlarged Furnace Design/Reduced Firing Rate;

                                                  4-3

-------
                 -  Reduced Air Preheat (RAP);
                 -  Water Injection (WI); and
                 -  Advanced Burner/Furnace Designs.
Some of these techniques, such  as  enlarged furnace designs, first generation low NOV burners,  OFA,
                                                                                    X r-
and LEA, have become  a  standard part of new unit design.  Techniques such as LEA,  F6R,  and  various
methods of  OSC  have  also been  extensively applied  to  existing units.   Other  techniques have  not
gained wide  acceptance  because of  adverse impacts on unit  efficiency  or capacity.  These  include
reduced firing rate, RAP, and WI.  Also, various OSC techniques may result in reduced unit capacity
1n some retrofit  applications.   Advanced burner and furnace designs represent techniques which  are
stm under development but  may become available in the  next few years.
       Retrofit NO  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  BBF or BOOS  are  implemented.
If further reductions  of NO   are necessary, these minor  modifications are followed by the more major
                           X
retrofits, Including OFA ports, FGR systems, and new burners.

       The feasibility, effectiveness, and method of applying 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 NOV,  so this technique is usually not cost-effective for  NOV  control
                                 A                                                         X
on coal-fired units.  Also,  such techniques as BOOS or OFA are implemented differently on wall-fired
and  tangentially-fired  units  due   to  burner configuration  and  hardware differences.    Tables  4-1
through 4-8 summarize available  test data  by  fuel  and firing type for  various combustion modifica-
tions applied to  utility boilers.    These  data  were obtained from  Reference  4-2  and the reader  is
referred to this reference for further information about these tests.

       The  practical  limits  on  combustion modifications  are  based initially on  three  subjective
criteria:   emission of other  pollutants (i.e.,  CO, smoke, 'and carbon in  fly ash), onset  of slagging
or fouling, and incipience ot flanie instability at the burner.
       The remainder  of this  section describes  recent combustion modification experience on  coal-,
oil-, and gas-fired boilers.  The material  in this section was taken from Reference  4-3.   The reader
1s  referred to  this  document  for  additional  information  and  for the original   sources  of  this
material.
                                                4-4

-------
                            TABLE 4-1.   AVERAGE  NO   REDUCTION  WITH LOW  EXCESS  AIR  FIRING (LEA)a
                                                        x     (Reference  4-2)
Equipment
Type

Tangential
Opposed Hall
Single
Hall
All Boilers*
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas*
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
All
Fuels

Number
of
Boilers
Tested
11
~
1
5
4
6
7
(2)
4
3
(1)
23
8
10
41
Baseline
Stoichionetry
to Active Burners
(percent)
124
-'
117
126
120
115
123 u
(134)"
120
117
(124)
. 124
120
116
120'
NOX Emissions
(ppm dry Q 3X 02)
459
-- -
340
746
357
717
624
(1338)
409
418
(992)
609
383
492
495
Low Excess Air (LEA)
Stoichiometry
to Active Burners
(percent)
116
-
113
US
113
110
114
(118)
112
108
(112)
116
115
110
114
NO Emissions
[ppm dry 9 38 03)
373
" --
245
660
290
600
522
(1325)
315
356
(931)
522
302
400
408
Average
NO. Reduction
(percent)
19
—
..,- 28
12
19
16
16
(1)
23
15
(6)
16
21
20
19
Maximum NOX
Reduction
Reported
(percent)
42 .
-
28
23
30
33
25
(3)
26
15
(6)
30
28
25
28
I
en
               jjBoiler load  at or above 80 percent HCR^  For individual  tests, corresponding baseline and controlled loads were nearly identical.
               "Numbers in parentheses refer to boilers originally designed for coal firing with wet bottom furnaces.

-------
                                      TABLE 4-2.   AVERAGE NO   REDUCTION  WITH  BURNER OUT OF SERVICE  (BOOS)3
                                                                            (Reference 4-2)
en
E
16
<24)<=
16
(12-24)l»
K .
(12-16)>»
28
(16-56)0
20 h
(12-24)"
16
(6-36)0
20
(a-ffijb
StotchlOMtry
to Active
Burners
(percent)
121
-
112
122
107
US
123
(134)<=
119
11?
122
113
115
117
NO,
Emissions
(pp* dry
»M02)
462
--
146
670
442
674
610
(1196JC
425
A\Q
S83
433
412
4/6
lurners Out of Service (MIOS)
Percent
Burners
on Air
Only
17
--
HA
16
33
28
19
<33)C
ia
22
17
25
25
n
Stolchlowtry
to Active
Burners
(percent)
98
--
06
102
73
84
97
(a9)<=
95
as
99
84
86
90
HOX
Caisslons
(ppmdnr
»3X02)
293
--
146
522
292
290
412
(577 )C
256
214
409
274
217
300
Average NO.
Reduction
(percent)
37
--
0
22
34
57
33
{52)£
40
4$
31
37
35
34
Maxima HO.
Reduction
Reported
(percent)
56
--
0
46
34
61
48
(52)C
48
S3
SO
41
43
45
                             'Boiler load at or above 80 percent HCR.  For individual tests, turrespondlnu, baseline and controlled loads were nearly identical.
                             "Range in number of burners firing
                              Numbers in parentheses refer to boilers originally designed for coal firms with *et bottom furnaces.

-------
                   TABLE 4-3.   AVERAGE NO -  REDUCTION WITH OVERFIRE  AIR  (OFA)a
                                          .    (Reference 4-2)
Equipment
Type
Tangential
Opposed Wall
Single
Wall
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Number
of
Boilers
Tested
6
—
—
~
* 5
2
-
« ,
~
Baseline
Stoichiometry
to Active
Burners
(percent)
129
~
--
~
118
114
—
..
-
N°x
Emissions
(ppra dry
0 3X 02)
454
—
~
-
376
928
--
-
-
Over fire Air (OFA)
Stoichiometry
to Active
Burners
(percent)
105
---
— •
~
96
99
--
-
—
Furnace
Stoichiometry
(percent)
122
—
—
—
118
112
—
—
--
NOX Emissions
(ppm dry
3 3X 02)
311
"
—

287
378
—

-
Average NOX
Reduction
(percent)
31
~
~
--
24
59
--
--
--
Maximum NOX
Reduction
Reported
(percent)
41
•
-
--
30
66
--
-
-
aBoi1er load at or above 80 percent NCR.  For individual tests, corresponding baseline and controlled loads were nearly identical.

-------
                               TABLE 4-4.  AVERAGE NO  REDUCTION WITH FLUE GAS RECIRCULATION  (FGR)a
                                                                (Reference  4-2)
I
00
Equipment
Type
Tangential
Opposed Wall
Single
Hall
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Number
of
Boilers
Tested
—
—
1
* 1
1
—
—
«
1
Baseline
Stoichionetry
to Active
Burners
(percent)
—
-
117
128
122
—
~
~
106
>»x
Emissions
(ppa dry
@ 3* 02)
-
—
340
855
304
—
-
—
470
Overfire Air (Of A)
Stoichionetry
to Active
Burners
(percent)
-
—
115
127
126
—
—
—
107
FGH
(percent)
--
—
23
15
11
--
—
—
11
NOX Emissions
(ppra dry
• 3X 02)
—
—
135
735
263
. ~
—
—
307
Average NOX
Reduction
(percent)
..
—
60
17
13
—
—
--
35
Maximum I(0X
Reduction
Reported
(percent)
— •
—
60
17
13
—
- ~
—
35
                afloi)er load at or above SO percent HCR.  For individual tests, corresponding baseline and controlled loads were nearly identical.

-------
                 TABLE 4-5.   AVERAGE  NO   REDUCTION  WITH  REDUCED FIRING  RATE0
                                           (Reference 4-2)
Equipment
Type
Tangential
Opposed Wall
Single
Hall
All
Boilers
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Coal
on
Nat
Gas
Coal
011
Nat
Gas
All
Fuels
Number
of
Boilers
Vested
7
—
1
4
4
5
2
(2)
3
2
(1)
13
7
8
28
Baseline (BOX HCR or Above)
Firing
Rate
(percent
HCR)
93
—
100
93
98
98
92
(90)
98
97
(98)
93
98
99
97
Stolchiometry
to Active
Burners
(percent)
112
—
117
131
118
115
125
(133)»
119
118
(115)
126
119
117
120
NOX
Emissions
(ppra dry
9 3* QZ)
462
—
340
825
362
651
651
(1338)
425
442
(992)
646
393
478
506
Reduced Load
Firing
Rate
(percent
HCR)
64
—
75
70
61
57
67
(54)
53
35
(59)
67
57
55
60
Stolchlometry
to Active
Burners
(percent)
127
—
135
136
121
US
130
(138)
119
117
(131)
131
120
122
124
NO),
Emissions
(ppm dry
9 3t 02)
408
..
332
758
249
269
496
(990)
296
125
(522)
554
272
242
356
Average NQX
Reduction
(percent)
12
—
2
8
31
59
24
(26)
30
72
(47)
14
31
44
30
Maximum NOX
Reduction
Reported
(percent)
25
-
32
18
48
64
25
(33)
45
82
(47)
23
47
59
43
'Numbers 1n parentheses refer to boilers originally designed for coal firing with wet bottom furnaces.

-------
                TABLE 4-6
                      AND
.   AVERAGE N0₯ REDUCTION WITH OFF STOICHIOMETRIC COMBUSTION
 FLUE  GAS RECALCULATION (OSC AND  FGR)a  (Reference 4-2)
Equipment
Type
Tangential
Opposed Hall
Single
Wall
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Number
of
Boilers
Tested
—
—
1
1
1
-
—
2
i
Baseline
Stolchlometry
to Active
Burners
(percent)
—
-
117
128
122
-
-
.118
106
HOX
Emissions
(ppm dry
9 3X Oz)
.-
—
340
781
304
—
-
355
470
OSC and FGft
Type of
OSC
—
—
BOOS
• BOOS
OFA
-
—
BBF
BOOS
BOOS
Stoichiowetry
to Active
Burners
(percent)
~
-
75
99
97
~
~
91
75
FOR
(percent)
—
..
21
19
11
—
-
14
12
KOX
Emissions
(ppm dry
9 3X 02)
—
-
105
453
247
--
—
154
115
Average HOX
Reduction
(percent)
~
-
69
42
19
--
--
57
76
Maximum NOX .
Reduction
Reported
(percent)
--
—
69
42
19
—
—
59
76
toiler load at or above 80 percent MCR.  For individual tests, corresponding baseline and controlled loads were nearly identical.

-------
                  TABLE 4-7.   AVERAGE NO   REDUCTION WITH  REDUCED FIRING RATE AND
                            OFF STOICHIOMETRIC COMBUSTION  (Reference 4-2)
Equipment
Type

Tangential


Horizontally
Opposed Wall

Single
Wall
All
Boilers
Fuels
Fuel
Coal
011
Nat
Gas
Coal
Oil
Nat
Gas
Coal
011
Nat
Gas
Coal
011
Nat
Gas
All

Number
of
Boilers
Tested
8
—
—
3
4
6
4
(2)
3
2
(1)
15
7
8
30
Baseline
Firing
Rate
(percent
NCR)1
93
-
-
93
99
100
90
98
97
(98)
92
99
99
97
Stoichlometry
to Active
Burners
(percent)
122
~
--
129
118
115
124
(133)4
120
118
(125)
125
119
117
120
NOX
Emissions
(ppro dry
» 3% 02)
453
—
-
820
362
717
663
(1338)
426
442
(992)
645
.394
579
b39
Low Load and OSC
Firing
Rate
(percent
NCR)
61
—
~
73
64
58
73
(59)'
56
35
(71)
69
60
31
bJ
Type of
OSC
BOOS
OFA
—
--
BOOS
BOOS
OFA
BOOS
OFA
BOOS
BBF
BOOS
BBF
BOOS
BOOS
OFA
BBF
BOOS
OFA
BOOS
OFA
DBF
noos
OFA
Stolen iometry
to Active
Burners
(percent)
95
—
-
102
117
88
99
(91)
97
93
(102)
99
107
91
99
1
NO.
Emissions
(ppra dry
9 3X02)
248
~
—
634
177
148
381
(887)
228
78
(641)
421
202
- 113
245
Average HOX
Reduction
(percent)
45
~
-
23
51
79
43
(34)
46
82
(35)
37
49
80
S5
Maximum
NOX
Reduction
Reported
(percent)
62
-
--
32
67
89
50
(55)
59
87
(351. '
48
63
8S
66
'Numbers in parentheses refer to boilers originally desigmn) for coal firing with wot .but turn furnaces.

-------
TABLE 4-8.  AVERAGE NO  REDUCTION WITH LOAD REDUCTION,  OFF STOICHIOMETRIC
          COMBUSTION AND FLUE GAS RECIRCULATION (Reference 4-2)
Equipment
Type

Tangential


Opposed Hall

Single
Hall
All
Boilers
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Sas
Coal
Oil
Nat
Gas
All
Fuels

Hunter
of
Boilers
Tested
-
--
--
-
3
2
--
2
2
—
5
4
9
Baseline
Firing
Rate
(percent
NCR)
—
...
—
~
99
100
--
98
100
—
99
100
100
Stoichlonetry
to Active
Burners
(percent)
«
—
—
—
118
113 .
--
118
110
—
118
112
115
HOX
Emissions
(pp* dry
9 3* 02)
—
--
—
~
398
945
--
355
421
~
376
683
530
Controlled/Low Load and OSC and FGR
Firing
Rate
(percent
HCR)
—
~
—
—
46
43
—
62
65
—
54
54
54
Type of
OSC
--
—
—
-
BOOS
OFA
BOOS
OFA
—
BOOS
BOOS
-
BOOS
OFA
BOOS
OFA
BOOS
OFA
Stoichioroetry
to Active
Burners
(percent)
--
—
—
—
87
90
—
92
81
—
90
86
88
FGR
(percent)
--
-
-
--
39
27
—
30
20
-•
35
23
29
NOX
Emissions
(pp* dry
9 3% 02)
—
—
.. .
--
194
130
--
152
171
—
173
ISO
162
Average HOX
Reduction
(percent)
—
—
—
-
56
87
—
57
59
—
57
73
66
Maximum HOX
Reduction
Reported
(percent)
—
-
—
--
59
90
-
57
83
—
58
87
73

-------
Application to Coal-Fired Boilers                                  .                .  _
       Currently the most commonly applied low NO  technique for coal-fired boilers is off stoichio-
                                                 X
metric or staged combustion through the introduction of oyerfire air (OFA).  This technique has been
used in both  retrofit  and new unit applications.  Application of burners  out  of service (BOOS), an
alternate staging technique with retrofit application, is limited because it is often accompanied by
a 10 to 25 percent load reduction.  Average NO  reductions of 30 to 50 percent from the levels shown
in Figure 4-1 (controlled emissions of 210 to  300  ng/J,  0.5  to  0.7  lb/10   Btu)  can be expected with
either technique in dry  bottom  boilers.   Flue  gas  recirculation (FGR)  has  been, tested but was found
to be a relatively ineffective control for coal firing, giving only about 15 percent NO  reductions.
                                                                                       A
More recently, new low NO   burners  (LNB)  have  been installed on some  units and found to be at least
as effective as OFA.  In fact,- new wall-fired units currently being brought online generally rely on
LNB to meet  the 1971 New Source  Performance  Standard (NSPS).  The combination  of OFA  with LNB has
resulted  in  40  to 60 percent  NO '  reductions  (controlled  emissions of  170  to 260  ng/J,  0.4 to
                                  A
0.6 lb/106 Btu).

       There has been a steady improvement in combustion modification control technology over recent
years.   Current demonstrated  technology is  capable of  40 to 60 percent  NO   reductions  from the
                                                                             X
levels shown  in  Figure 4-1, easily meeting the  1971  NSPS of 300 ng/J (0.7  lb/106 Btu)  and in most
cases the  1979  revised standard of 210 to 260 ng/J  (0.5  to 0.6 lb/10  Btu).  Current R&D programs,
such as  the  development  and demonstration of  the  EPA advanced  low NO  burner  and the  EPRI primary
combustion  furnace,  should  result  in  combustion  modification  techniques  capable  of  meeting
NOX emission levels of 86 to 130 ng/J (0.2 to 0.3 lb/106 Btu).

       The effects  of low  NO  operation on  coal-fired boilers  are  summarized  in  Table 4-9.   The
                     »                                                 •
table describes  experience  obtained through field test  programs  on the  several  units  listed.   The
major  concerns  regarding low  NO  operation on  coal-fired  boilers  have  been  the possible adverse
effects on boiler  efficiency,  load capacity, water wall  tube corrosion  and slagging,  carbon loss,
heat absorption profile, and convective section tube and steam temperatures.

       In  most  past  experience  with  staged  combustion,  optimal   excess  air  levels  have  been
comparable to  levels  used under baseline conditions.   In these cases the efficiency of the boiler
remains unaffected  if  unburned  carbon losses  do not  increase appreciably.  However, in  some cases
when, due  to  nonuniform  fuel  air distribution  or other causes,  the excess  air requirement*increases.
under staged  firing, a  significant decrease in efficiency occurs.  From Table  4-9,  it  is seen that
efficiency decreases of up to 1 percent have been experienced.  It is also seen that the same boiler
                                                4-13

-------
TABLE 4-9.  EFFECT OF LOW NO  OPERATION ON COAL-FIRED  BORERS*
                        (Reference  4-3)
Boiler
Tanqentlai
Barry No. 2

Columbia
No. 1
Huntington
Canyon No. 2
Barry No. 4
Navajo No. 2
Ccaanche No. 1
Kingston No. £

Opposed Uall
Harllce Branch
No. 3
i
Four Corners
He «
Low NOX
Technique
BOOS
OFA
OFA
OFA
LEA. BOOS
LEA. BOOS. OFA
OFA
BBF
BOOS
LEA, BOOS
LEA. BOOS
Boiler
Efficiency
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
0.5X average
Increase
0.6X average
decrease
0.6X increase
Corrosion
Measured 75X
increase, but
within nora. )
range
Measured 701
increase, but
within normal
range
No change
Measured 25X
decrease, but
within noraal
range
No significant
change
No significant
change
No significant
change
—
~
Slight increase
No significant
change
Load
Capacity
20* derate
Unaffected
'Jnaffected
Unaffected
20X AT more
derate with
BOOS
Unaffected
Uni^fected
Unaffected
202 derate
Up to 17X
derate
with BOOS
Up to 25X
derate
with BOOS
Carbon Loss
in Fly*sh
Slight increase
Slight increase
Slight increase
Slight increase
~50JC average
decrease
No change
~3W average
decrease
Unaffected
Unaffected
~130;i average
~50X average
decrease
Oust Loading
~100X increase
— 100* ircrease


~SOX average
increase
~40t average
increase
~20* average
decrease
Unaffected
Unaffected
~1W average
increase
~15X average
decrease
Part. Sue
Distribution
--



--
No change
NO significant
change
Unaffected
Unaffected
—
"
Other Effects.
C omen Is
Minor changes in heat
absorption profile
SH atteaperatton
increased by 7M
Minor charges in heat
absorption prof lie
SH atteaperation
increased over 20QX
Xinor changes in heat
absorption profile
SH *ttMper«ttoii
increased hy 70t
Minor changes in heat
Absorption profile
No SH atteaperation
required








-------
                                                   TABLE  4-9.   (Continued)

Boiler
Kstfteld No. 3









F..C. Gaston
Ko. 1




Crist No. 7
Slnole Hall
Widows Creek
Ho. 5 (TVA
Test)
Widows Creek
No. 5 (Exxon
test)
Widows Creek
No. 6
Mercer Station
No. 1 (wet
bOttOM)

Crist Station
No. 6

Low NOX
Technique
BOOS

FGR







LNB. LEA. COOS





BOOS

BOOS


LEA, BOOS


LEA, BOOS

LEA. Biased
firing


LEA. BOOS


Boiler
Efficiency
O.jl decrease

0.4t decrease
in bailer
efficiency.
Sow decrease
In cycle
efficiency
due to RH.
atteaperatlon.
0.31 decrease
on average
(LNB baseline)



Unaffected

It decrease


It average
Increase

Unaffected

Unaffected



0.4t decrease



Corrosion
	

_.







So significant
Increase




Increase

Results of
tests
Inconclusive
No significant
increase

—

No significant
increase


	


| Load
Capacity
10X derate

Unaffected







Up to 30t
derate
(LNB with
BOOS)


Unaffected

UnaffecteJ


Unaffected


Unaffected

U(vaff(>cted



Up to 15*
derate
1
Carbon Loss
in Flyash
~30X average
increase
~120t average
ircrease






~13(Jt average
ifK.ftS~t (L"*
baseline)



Sllghi increase

30t increase


30t average
decrease

70S average
Increase
80t average
Increase


SOU increase



Oust Loading •
Unaffected

Unaffected







~15t average
increase (LNB
base line}



Unaffected

No significant
increase

}*)t average
dec ease

20t average
decrease
lOt average
increase


tflff 4«u"i»«kACtt


Part. Size
Dlstr .lion
	

--







Shift towards
smaller par-
ticles (LNB.
with or with-
out BOOS)

Unaffected

_ _


_.


	

No significant
change


	


Other Effects,
Cements
No slagging cr fouitno.
No sitwiiflcart Increase
In tube temperatures.
Stable flatus and
uniform coobusvi;^.
Increase In RH
atteaperatton. No
significant Increase
in tube temperatures.


Urit retrofitted
wi^h low NO. burners.
Baseline. LEA and BOOS
tests with LNB compared
to baseline tests on
sister boiler with no













1



— denotes tni: pameter MIS not lavestlgat-d.

-------
(Widows Creek No. 5) tested at-a  different  time with BOOS showed an average  increase  in  efficiency
of 1 percent.
       Many new boilers now come factory equipped with OFA ports.  Older  boilers  can be  retrofitted
with OFA  ports,  or can  operate with minimal  hardware changes under  BOOS  firing.   BOOS firing  is
normally accomplished  by shutting  off one  or  more pulverizers supplying coal  to the upper  burner
levels.  If  the  other pulverizers  cannot  handle the extra fuel  to  maintain a constant  total  fuel
flow, boiler  derating  will  be  required.  From Table 4-9  it is seen that boiler derating of  10  to
25 percent is not uncommon with BOOS firing.

       The  possibility of  increased corrosion  has  been  a major  cause  for  concern  with  staged
operation.  Furnaces fired with certain  Eastern  U.S.  bituminous coals  with  high sulfur contents may
be susceptible to  corrosion attack under reducing  atmospheres.   Local reducing  atmosphere  pockets
may  exist  under staged  combustion  operation  even  when  burner stoichiometry  is  slightly  over
100 percent.  The problem may be further aggravated by slagging as slag generally fuses at  lower
temperatures  under  reducing conditions.  The  sulfides  in the molten  slag  may then readily  attack
tube walls.  Still, as noted  in Table 4-9,  experience with most short-term  tests  has generally been
that no  significant acceleration in  corrosion rates occurs under  staged firing.  Recent  12-month
corrosion measurements by Exxon during low NO  firing in a 500  MW boiler showed significantly higher
corrosion  rates  with  low  NO   operation   (Reference 4-4).   Nevertheless,  the  issue  cannot  be
considered  resolved  until  further results  are  available  from long-term  corrosion  tests  with
measurements  on  units  equipped  with  factory  equipped  NO   controls.    Insofar  as  slagging  is
concerned, short-term tests performed to date generally indicate no  significant increase  in  slagging
or fouling of tubes under staged combustion.

       Increased carbon  loss  in fly  ash  may occur  with  staged firing if  complete burnout of the
carbon particles does  not occur in the  furnace.  'High  carbon  loss  will  result in  decreased  boiler
efficiency and may also  cause electrostatic precipitator  (ESP)  operating  problems.  From Table 4-9,
it  is  seen  that increases in  carbon los%  vary over a  wide  range and  can  be  as high as  70  to
130 percent of baseline values  in some cases.   However, -increased carbon  loss levels have generally
been considered acceptable and not a major problem associated with staged  combustion.

       Extension  of the  combustion  region to  higher  elevations  in  the  furnace may  result  in
potential  problems  with excessive  steam and tube temperatures in retrofit  applications.   However,
among  the  numerous  short  term combustion  staging tests  conducted,  no  such problems  have  been
reported.   In some  tests  where  furnace and, convective  section tube temperatures  were  measured
                                                4-16

-------
directly, no significant increase was found.  Changes in heat absorption profiles were also found to
be  minor.    Superheater attemperator  spray  flowrates  tripled  in  one  case  under  retrofit  OFA
operation,  but  in  all  cases  they were well  within spray flow  capacities  of the units.   Reheater
attemperator  spray flowrates  did  not  show any  increase  due  to  staged  operation,  thus  cycle
efficiencies were not affected.

       Many new wall-fired coal boilers are,being fitted with low NOX burners  (LNB).   These burners
are designed  to  reduce NOV levels either  alone,  or in some  cases,  in  combination with OFA ports.
                          X
Using the new burner designs has the advantage of eliminating or decreasing  the presence of reducing
or  near reducing  conditions  near furnace  walls  that  could  cause  corrosion.   Although  low  NO
                                                                                                   A
burner  flames are  generally less turbulent  and  hence longer than flames  from normal  burners,  the
combustion  zone will   probably  not  extend  any  farther  up the  furnace  than with  overfire  air
operation.   Potential  changes  in  heat absorption  profile and excessive steam  and  tube temperatures
are, therefore, less likely to occur.
       As  fuel  and  air flows  are  controlled  more  closely in  LNB equipped systems,  nonuniform
distribution  of fuel/air ratios  leading  to excessive CO generation or  high  excess air requirements
should  be  reduced.  Boiler efficiencies  should,  therefore, not be affected, especially  in new unit
designs where adequate volume for carbon  burnout  is available.  However, Table 4-9  shows that the
efficiency of one  boiler  decreased  slightly (0.3  percent) when  retrofitted with low  NO   burners.
                                                                                   *
The decrease  in efficiency was mainly due to the large  increase  in  unburned carbon  loss.   However,
such problems noted in retrofit  applications can  be avoided in units specifically designed for low
NOX  burners.   Corrosion rates  inferred from tests with  corrosion  coupons  showed  no  significant
increase with the new burners. . Some BOOS tests were also carried out on the LNB equipped boiler.  A
substantial  decrease  in NOX emissions resulted, below those already achieved with the  new burners
alone.  However, the  boiler was  derated  by up to 30 percent.  Other potential  problems  noted above
as being associated with staged combustion could also arise with this type of firing.

        It  should be emphasized that  the  operational  effects  of NO  control, in many  cases, will be
critically dependent  on boiler operating conditions.   Still,  with proper design of retrofit systems
and adequate maintenance programs, low NO  operation should not  result  in a  substantial  increase in
operational  problems  over  normal  boiler operation.  Moreover,  when  NO   controls  are  designed into
new units, potential problems can be anticipated and largely avoided.
                                                4-17

-------
       Other advanced NO   control  techniques  under development, including the EPA advanced  low  NO
burner and EPRI primary combustion  furnace, are designed to further minimize some of  the  potential
problems  with conventional' combustion  modifications,  such  as  losses  in  boiler  efficiency  and
Increases in corrosion rates.

Application to Oil-Fired Boilers
       For cost reasons, most  oil-fired  utility boilers fire residual oil.  The most  commonly used
low NOX  techniques  for these boilers are  staged  combustion and flue gas  recirculation  (FGR), both
employed in combination with low excess  air firing.  These techniques have generally  been  employed
on a retrofit basis since  few  new oil-fired units  are  being installed.   Other techniques which have
been tested  are  water  injection (WI) and  reduced air  preheat (RAP).   However,  these have  found
little application due to associated efficiency losses.

       Staged combustion  has  been  applied through the  use of  overfire air  ports  (OFA)  and   by
removing burners  from service  (BOOS).  Typical NO  reductions  using OFA are 20 to  30 percent from
baseline conditions  (controlled emissions of  150  to  170 ng/J, 0.35  to 0.4 lb/106 Btu).  BOOS  has
been slightly more  effective, giving 20 to  40  percent reductions  (controlled  levels of 130  to
170 ng/J, 0.3 to 0.4 lb/10  Btu).  A load reduction may accompany the use of BOOS  on  some units.

       Flue gas  recirculation  also typically  gives 20  to  30  percent NOV reductions,  but  requires
                                                                         A
more hardware modifications.  The combinations of  BOOS  or OFA with  FGR have  been  most effective,
resulting  in  30  to  60 percent  reductions   (controlled  emission  of  86  to   170 ng/J,   0.2   to
0.4 lb/106 Btu).   With FGR,  OFA  is  preferred  over BOOS because flame stability  seems  to be  more of
a problem with the combination of FGR + BOOS.
                                                                                      t
       There  have  been  some R&D  efforts  by EPA  and  private industry  to  develop low  NO   emission
burners  for  oil  firing.  Although  no detailed test reports have been  released  yet,  the Babcock &
WUcox dual register low NO  oil-fired burner has been successfully  retrofitted  on a  utility boiler,
achieving  emissions  below 130 ng/J (0.3  lb/10  Btu)  (Reference 4-5).   The combination  of  overfire
air  and low  NO   burners  may   potentially  achieve emissions  below  86  ng/J  (0.2 lb/10  Btu).   An
oil-fired  low NO   burner  is  also under development  by Southern  California  Edison  to  meet  the
stringent NOV regulations for utility boilers  in the Los Angeles area.
            X
                                               4-18

-------
       The effects  of low  NO   operation on  oil-fired  boilers are  summarized in Table 4-10.   The
major concerns regarding  low  NO  operation on oil-fired  boilers  are effects  on  boiler  efficiency,
load capacity, vibration and flame instability, and steam and tube temperatures.

       Staged combustion  operation  generally  increases  the minimum  excess  air requirements of  the
boiler, which  may result  in  a  loss  in boiler  efficiency.    In  extreme cases when  the boiler  is
operating close to the limits of its fan capacity, boiler derating may be required.  Derating  of as
much  as  15 percent has  been required  in  some  cases due  to the lack  of  capability  to meet  the
increased airflow  requirements  at  full  load.   In addition, under BOOS  firing the fuel flow to  the
active burners  must  be  increased  if  load  is  to  remain constant.   In many cases,  it  has been
necessary to enlarge the burner tips in order to accommodate these increased flows.

       Other potential  problems  associated with applying staged  combustion  on a retrofit basis  in
oil-fired boilers  have  concerned flame  instabilities,  boiler vibrations,  and excessive  convective
section .tube  temperatures.   However,  in  past  experience,  none   of these  problems   has  been
significant.  Staged  operation  can  result  in hazy flames and obscure flame zones.   Thus new  flame
scanners and  detectors  are often  required in retrofit  applications.  In  addition, because staged
combustion  produces  an  extended   flame  zone,  flame  carryover  to  the   convective  section  may
occasionally occur.  However, in one case  where  intermittent  flame carryover occurred,  no  excessive
tube temperatures were recorded.

       Similarly,  there  are  a  number  of  potential  problems  which  can   occur in  retrofit  F6R
applications.  The most common problems, such as F6R fan and duct vibrations, can usually be  avoided
by good design.  Other problems such as flame instability, which can  lead to furnace  vibrations,  are
caused by the  increased gas velocity at the  burner  throats.   Modifications to the burner  geometry
and  design  such  as enlarging  the  throat,  altering  the burner tips,  or adding diffuser plates  or
flame retainers, may then be required.

       Another potential  problem sometimes  associated with FGR is high  tube and  steam temperatures
in the convective  section.  The increased  mass velocities which occur with  FGR  reduce  furnace heat
transfer, but increase  convective heat  transfer  in  the  convection section.   The  result is  increased
convection section flue  gas,  steam  and tube temperatures.  Increased mass flowrates in  the  furnace
may also cause furnace pressures to increase beyond safe limits.

     The  combination  of  staged  combustion and  FGR is  very  effective  in  reducing  NO   emissions.
However, the  potential  problems associated with each technique are  also combined.  Tube  and  steam
                                                 4-19

-------
TABLE 4-10.  EFFECT OF LOW NO  OPERATION ON OIL-FIRED BOILERS2
                        (Reference  4-3)
Boiler
Tangential
South Bay No. 4


Pittsburg Ho. 7
SCE tangential
boilers
Opposed Wall
Moss Landing
Nos. 6 and 7
(Early
experience)
Moss Landing
Ho. 6 (NO, EA
tests)

Ormond Reach
Nos. I and ?

Low MOX
Technique
LEA
BOOS
RAP
OFA and FGR
BOOS and FGR
BOOS and FGR
FGR
BOOS and FGR
BOOS and FGR
Water Injection
Efficiency
5< Increase
«
Decrease in efficiency
compared to LEA due to
Increased excess air
requirements
Unaffected due to
special preheater
design
"
—
Increased excess air
requirements resulting
In decreased efficiency
Unaffected
Unaffected
Increased excess air
requirements resulting
in decreased efficiency
Increased sensible and
latent stack losses
Load
Capacity
—


Slower startups
and load changes
—
—
Unaffected
Unaffected
10 to 15X derate
due to maxed FD
fan capacity

Vibration and
Flame Instability
--

--
FGR fan vibration
problems
—
FGR fan and duct
vibration, furnace
vibration problems.
Associated flame
instability.
None
None
Flame instability
and associated
furnace vibration

Stew ind Tuhe
Te«per«tures
--

-•
High water wall tube
temperatures
--
—
No attemper at Ion
changes required
No attemperation
changes required
•""

Other Effects, Comments
No adverse effects reported.
Fan power consumption
reduced.
No other adverse effects
reported
Limited tests. NOX
control effectiveness not
demonstrated.

No adverse effects reported
High furnace pressures.
Increased FGR and forced
draft fan power consumption.

Unit currently operating
under low NOX continuously
Flame detection problems
due to change In flame
characteristics
Limited tests carried out
with WI at oartial loads.
Excess a'*- requirements
increased.

-------
                                                       TABLE  4-10.   (Continued)

Boiler
SCE 6&W Units



Sewaren Station
No. 5



Single Mall
Enctna Hos. 1,
2 and 3








Turbo
South Bay No. 3









Potrero No. 3-1







Low NOX
Technique
BOOS and FGR



LEA. BOOS





LEA and BOOS
(2 burners
.on air only)


BOOS
(3 burners on
air only)



Airflow
adjustments

Water injection


Reduced air
preheat


OFA and FGR








Efficiency
FGR reduced minimum
excess air requirements
increasing unit
efficiency
„





Increased unit effi-
ciency. Some adverse
effect on cycle eff i-
ciency. due to lower
steam temperatures.
increased excess air
requirements resulting
In reduced efficiency



Slight reduction in
EA resulting in slight
increase in efficiency
65! decrease at full
load

Reduction in effi-
ciency greater than
that with water
injection
Higher excess air re-
quirements, hut addi-
tion of economizer
surface expected to
improve efficiency


•
Load
Capacity




„





_ M




53 derate due to
maxed 10 fan
capacity



—-.


^ »


_ —



51 derate due to
excessive tube
temperatures





Vibration and
Flame Instability
Boiler vibration
problems


__





_—




In most tests no
flame instability
or blowoff noted



_.


No flame Instability
noted even at high
rates of MI
— ^



Side to side
wind box oxygen
cycling





Steam and Tube
Temperatures




....





Decrease In SH & RH
steam temperature

*£

Intermittent flame
carryover to SH
inlet but tube
temperature limits
not exceeded

~» -


•»«>


_.



Tube and steam tem-
perature limits ap-
proached. Increased
SH tube failures.





Other Effects, Comments,
Flame detection problems
due to change in flame
characteristics

Tests carried out at partial
loads. No adverse effects
reported. Particulate load-
ing and size distribution
unaffected.

No other adverse effects
reported



"No abnormal tube fouling.
corrosion or erosion noted.
Increased tendency to smoke
and obscure flame zone.


No adverse effects reported


No other adverse effects
reported

Limited tests



Increased tendency to smoke
required higher minimum
excess Op levels., RH
surface ranoved to avoid
excessive RH steam attem-
peratlon. Larger economizer
installed to compensate for
RH surface removal.
I
ro
        — denotes not investigated.

-------
temperature problems in the upper furnace are an area .of concern, as both combustion staging and  F6R
tend  to  increase  upper  furnace  temperatures  and  convective  section  heat  transfer  rates.    In
addition,  boiler efficiencies  usually decline  slightly with  combined staged  combustion  and  FGR
firing due-to higher excess air requirements and greater fan  power  consumption.   However, potential
adverse effects for retrofit NO  control systems can be minimized by proper design and installation.
Many of the problems experienced in the past can now be avoided.

Application to Gas-Fired Boilers

       The most  commonly applied  NO  control  technique for  gas-fired  boilers,  as with oil-fired
                                     X
boilers,  is  staged combustion  through the use  of OFA  or  BOOS.  These  prove  more effective when
combined with  FGR;  however,  flame stability may be  of greater  concern  when  BOOS is combined with
FGR.  Typical  NO  reductions  under  either OFA, BOOS, or  FGR  are  30  to  60 percent from  baseline
conditions (controlled  emissions  of 86 to  150  ng/J, 0.2 to  0.35 lb/10  Btu).   The combination  of
staged  combustion  and FGR is  capable  of  50  to 80  percent reduction (controlled  levels of 43  to
110 ng/0, 0.1 to 0.25 lb/106'Btu).

       There  are no  major efforts  toward developing a low NO   burner  or other new  combustion
                                                                 A
modification techniques for  gas firing because NO   emissions under current control  techniques  are
                                                  X
already relatively low, and no new gas-fired utility boilers are currently being  sold.
       The effects  of low NO   operation on  gas-fired  boilers are  summarized  in Table 4-11.   The
effects of  low NO  firing  on  gas-fired boilers  are very.similar  to those for  oil-fired  boilers.
                  X
Usually, there is no  distinction between oil- and gas-fired  boilers as they are  designed to  switch
from  one  fuel to the other according  to  availability.  Since boiler  design  details, NOX  control
methods, and the effects of low NOX operation  are  similar  for  gas-  and  oil-fired  units, most  of  the
above discussion of applicable NOX control measures  for oil-fired boilers and potential resulting
problems applies here as well.   Some effects specific to gas-fired boilers alone are  treated briefly
below.
       NO   emissions  are  often difficult to control   after  switching  from  oil  to  gas   firing.
Residual oil  firing  tends  to  foul  the furnace  due to  the  oil  ash  content.   Thus,  NO   control
                                                                                          X
measures which have been tested on a clean furnace with  gas may be found inadequate  after  oil  firing
due to the changed furnace conditions.
                                                4-22

-------
                                TABLE 4-11.   EFFECT OF LOW N0y  OPERATION ON  GAS-FIRED BOILERS0
                                                         (Reference 4-3)
I
M
CO
Bo Her
Tangential .
Smith Bay No. 4
Pittshiirg Ko. 7
Horizontally
Opposed
Hoss Landing
Not. 6 and 7
Pittsburg
Has 5 and 6
Contra Costa
Has. 9 and 10
Single Hall
Incina Nos. 1,
2 and 3
LowHOx
Technique

BOOS
OFA and FGR


OFA and FGR
OFA and FGR
OFA and FGR

BOOS
(? and 3
burners out
of service)
Efficiency*

Slight decrease in
efficiency due to
increased excess air
requirements



O.BJC decrease in cycle
efficiency due to BH
steam atlemperation
"


low EA levels were
possible even with
BOOS, resulting in
increased efficiency
Lnada
Capacity


25X derate 
-------
                                                    Table 4-11. Concluded.
Boiler
Turbo
South Bay No. 3
Potrero Ho. 3-1
Low NOX
Technique

Water Injection
OFA and FGR
Efficiency

10% decrease at full
load
Installation of larger
economizer expected to .
improve efficiency
Load
Capacity

--
5X iterate due to
prohlens with high
terrperatures
Vibration .ind
Flame Instability

No flame instability
noted even at high
rates of HI
Side to side
trindbox oxygen
cycling
Steam and Tune
Temperatures

—
Tube metal and steam
temperature limits
reached at high
loads
Other Effects, Coaments

No other adverse effects
"reported
Hardware modifications
included partial RH surface
removal to avoid excessive
RH steam attemperation.
Larger economizer fion
installed to compensate for
smaller RH surface.
I
rv>
•£»

-------
        Boilers  tired with gas usually have higher gas temperatures at the furnace outlet than those
 fired with  oil.  The upper furnace and convective section inlet surfaces are thus subject to higher
 temperatures  with gas firing.  These  temperatures may  increase further  under staged firing or F6R.
 Upper furnace and convective  section  tube failures and  excessive  steam temperatures are therefore
 more likely to  occur with staged firing and FGR applied to gas-fired boilers.  The situation may be
 aggravated  further if switching  to gas fuel occurs after  an oil  burn, as  fouling will further reduce
 furnace absorption and,  hence, increase gas temperatures.  Excessive tube temperatures will usually
 require derating of  the  system.   These problems  could be  minimized  on new units but no new gas-tired
 utility boilers are  being built.

 4.1.1.2  Flue Gas Treatment

        Historically, the major NOX control emphasis in the United  States has been on combustion or
 process modification.  However,  in Japan where  NO  emission  standards  are  more stringent, flue gas
 treatment (FGT) technologies have undergone extensive development and implementation.  Recently, in
 the U.S.  several pilot and demonstration  scale units have been  built and  operated.

        As discussed  in  Section 3.2,  flue gas  treatment consists of  any  of  several  technologies
 designed to remove or eliminate  NO  in the flue gas downstream of the  combustion zone.   Since FGT
                                    X
 technologies  are  distinct from  the combustion  process,  their  performance  capabilities'are usually
 described in terms of NO  percentage  reduction  in the flue  gas  rather  than in terms of achievable
 emission levels.  NO  reductions  with I-GT can  occur over and  above  the reductions attributable to
 combustion  modifications.

        These  postcombustion  processes can be divided into dry  or wet types.  The dry processes can
 be  further  categorized  into  four  subdivisions:   catalytic   reduction,   noncatalytic  reduction,
 adsorption, and  irradiation.  Ihe majority of the dry processes are of the reduction type.  These
 catalytic and noncatalytic  reduction  processes  can also  be  classified  as selective or nonselective
                                                                                       *
 processes based on the  type of reducing agent used.  The  majority  are selective and usually use NH,
 as the reducing agent.  If the NH, is  injected after the  boiler economizer,  where  temperature of the
 flue gas is  about 370°C to 430°C  (700°F  to  800°F),  a catalyst is necessary.   These processes are
 described as selective  catalytic  reduction (SCR)  processes.  If NH3  is  injected into the secondary
 superheater region  of the boiler, where temperature of  the  flue gas  approaches 980°C (1,800°F),  a
, catalyst is not necessary.  These  processes are described as selective noncatalytic reduction (SNR)
 processes.
                                                 4-25

-------
     The remainder of  this  section  describes  the development status of flue das treatment technolo-
gies and  experience  in applying  them to boiler  flue  gases.  The  discussion  is organized  by  fuel
type.  Much  of the material  was obtained  from Reference 4-6; the  reader  is  referred to  this  and
other documents (e.g. References 4~7, 4-8, and 4-9) for additional information.

Selective Catalytic Reduction (SCR)

       The SCR method  is  the  most advanced  FGT method,  and the  one on  which the  overwhelming
majority of existing NOX FGT units are based.  As with the majority of all  types of NOX FGT, most of
the SCR processes were developed  in Japan.   The Japanese have found that with the  optimum reaction
temperature, usually 300°C to 450°C (570°F to 840°F), an NH3:NO molar ratio of 1:1 typically reduces
NOV emissions by 90% with residual NH, concentrations of 10 to 20 pprn or higher.  It should be noted
  A                                  O
that the Japanese seem to  prefer 80% NOX removal in which NH3:NO molar  ratios  range  from 0.81:1 to
0.9:1.  Under  these  conditions the  unreacted NH3 concentration  is  usually  less than 5  ppm.   This
reduces capital and  operating costs as well  as effects on downstream equipment  from ammonium  salt
deposition.
       Presently,  there are over 60 full scale SCR units successfully operating on gas- or oil-fired
boilers in Japan.   Over 10 percent of these units are larger  than 330 MW.  Two commercial  SCR units
began operating in 1980 on  coal-fired boilers in Japan.  Construction is scheduled  to be completed
during 1981-1984 on at least 14 additional SCR units for coal-fired boilers  ranging  in capacity from
75 to  700  MW (References  4-10, 4-11,  and 4-12).  The  results  of the Japanese experience  on these
units must be tempered by the dissimilarities with  U.S.  facilities  including maintenance  practices,
load cycling, and overhaul  practices.
       In the United States, EPA and the Electric Power Research Institute  (EPRI) are evaluating SCR
on coal-fired pilot scale units.  EPA sponsored two 0.5 MW size tests.   The  Shell Flue Gas Treatment
process for  simultaneous  NOV and  SOV control  was  evaluated at Tarnpa  Electric Company's  Big  Bend
             .              *        x
Station.  This  process controls NO  by  selective  catalytic reduction although it uses a different
                                   X
catalyst than other SCR processes.   The Hitachi Zosen  SCR process was tested  at  the  Plant Mitchell
Station of Georgia Power Company.   Test results showed that  both processes  are technically capable
of achieving significant NO  reductions from coal-fired boilers in the U.S.

       One conclusion  of the  testing was that SCR test work  is needed when  considering SCR process
applications for untested coals.  Another was that a prototype scale test on a  10 to 100 MW facility
would be  useful for demonstrating  the  technology for coal-fired  sources  in the U.S.  For further
details on the results of these tests, the reader is referred to Reference 4-13.
                                                4-26

-------
       In other U.S. testing,  EPRI  has  operated a 2.5 MW pilot  unit  on a coal-fired boiler at  the
Arapahoe Station of Public Service  Company  of  Colorado  using  the Kawasaki  Heavy  Industries  process.
Also, testing of  the  first large scale  SCR demonstration unit  in  the  U.S.  is planned by  Southern
California Edison Company  at  the Huntington Beach Station.   The unit will be 107.5 MW in  capacity
and applied to a oil- and gas-fired boiler.

       Since the SCR reactor is located downstream of the boiler economizer,  its  process impacts  are
also largely limited to this region.  These include potential problems with the  SCR  reactor itself,
the air preheater, and downstream emission control systems.

       Dr. Oumpei  Ando  (Reference  4-14)  reports  that  early ,in its  development SCR  had  several
serious problems.   These included:

          (1)  Catalyst poisoning by SO  in the flue  gas;
          (2)  Plugging of the catalyst by dust;
          (3)  Deposition of ammonium bisulfate on the catalyst at reduced
               boiler loads;
          (4)  Deposition of ammonium bisulfate on the air preheater;
          (5)  Promotion of the oxidation of SOg to S03  in the flue  gas; and
          (6)  Erosion of the catalyst by fly ash from the coal.

These problems have been largely resolved as follows:

          Problem (1) -  Development of SOX resistant catalysts based  on TiOg
                         rather than Al20g or Fe203;
          Problem (2) -  Using parallel-flow catalysts such  as honeycomb,
                         plate, and tube shapes, parallel passage reactors, and
                         sootblowing when needed;
          Problem (3) -  Keeping the reactor temperature above 300°C by  using
                         economizer bypass gas when needed;
          Problem (4) -  Keeping NH, leakage at the reactor  outlet at  a  low
                         level  (e.g. below 5 ppm);
          Problem (5) -  Development of low oxidation catalysts which  also
                         helps to solve problems 3 and 4; and
                                                4-27

-------
          Problem (6) -  Using moderate gas velocities, hard catalysts,  and
                         a device (e.g. a dummy spacer) tor erosion prevention.

       Jones  (Reference 4-15)  reports  that a  major  area of  research  and  development  involves
minimizing the  impacts  of SCR systems on  downstream  equipment such as  air  preheaters,  particulate
collection devices and  SCL  removal  equipment.   Problems with the air preheater occur when  ammonium
blsulfate  (NH.HSO.)   deposits  plug  and  corrode  the  elements.   NH.HSO.  is  the  product  of  a
                               \
condensation reaction between NHj,  SO, and H-0, which can  occur  when  the flue  gas  temperature  drops
below about 210°C.  Japanese pilot unit tests have shown that the plugging problem  is most severe in
units which fire  coal or high sulfur  oil and also  remove fly ash upstream of the NOX reactor.   When
fly ash 1s removed downstream, plugging problems are significantly reduced.  It is  felt  that the fly
ash produces a  sandblasting effect  that  cleans the air  preheater elements  and also  that  some of the
NH^HSO^ condenses on the fly ash particles  rather than  the  elements.  Plugging problems  are reduced
or eliminated by  installing soot blowers on both sides  of  the  air preheater  and increasing  both the
frequency and pressure of the soot  blowing  operation.   In some cases, special air  preheater designs
will  be  used  in  which the intermediate  and  low  temperature zones  are  manufactured  as  a single
element.   These have been tested on pilot unit  equipment.   A full scale  installation is  scheduled at
the Electric Power Development Company's  Takehara  Power Station.
       NHg from an SCR  reactor  apparently  does  not impair F6D system performance although,  in  some
cases, the wastewater must  be treated to remove nitrogen compounds.   It  is not known if  SCR systems
will  affect  dry S02 removal  systems  (e.g., spray drying)  since these  techniques  are  not used  in
Japan.  The one apparent adverse  impact  that may  occur is  NH.HSO. affecting the performance of the
downstream baghouse.  The effect of an SCR .system on baghouses  is under investigation.   Pilot  unit
tests are underway.

       Several of the coal-fired SCR  applications that  are  under construction utilize hot-side  ESPs
tor  upstream  particulate  removal, and  there  are  a  variety of  reasons  for  selecting  hot-side
particulate removal.   These reasons include:

               (1) To eliminate fly ash from entering the NO  system and
                   potentially causing plugging or erosion  problems;
               (2) To obtain the capability to  remove particulates from a
                   wide range of coals with varying characteristics;  and
                                                 4-28

-------
               (3) To avoid ammonia compounds in the ash that can result
                   when a cold-side ESP is used.

However, cold-side ESPs also have unique advantages such as:

               (1) Lower capital and operating costs; and
               (2) Allowing the fly ash to reduce or eliminate NH4HS04
                   deposits on air preheaters.

      Jones further discusses the fact that .there has been concern that the catalyst and reactor may
plug with ash when applied to coal-fired boilers.  Pilot unit  tests  have  indicated  that plugging is
not a  problem when honeycomb' or pipe shape  catalyst  is  used in a  vertical,  downflow  arrangement.
However, soot  blowers  will be  installed  in  the  reactors  of current  full  scale applications  as  a
conservative design measure.   Another concern  in  the  U.S.  has been catalyst poisoning  by  flue gas
components.  While it is true that certain alkali  metals, such as  sodium  and  potassium, will  slowly
poison the catalyst, the concentrations  are  low enough that catalyst  activity will  not be  affected
during the guarantee  period.   Catalyst life  guarantees  are usually one year  for coal, one  to two
years for oil, and two  to  three years  for  gas although the  experience on  gas- and oil-fired boilers
has been that actual  catalyst life exceeds the guarantee.

       The  labor  requirements   of  the  operating,  full-scale  systems  are  small.   No  additional
operating  personnel  are required and  maintenance  labor  consists*  primarily  of  NH,  and  catalyst
loading and cleaning the air  preheater during the annual  outage.   Since there have been no  catalyst
changes to date,  the labor estimates  for  this work vary widely.   Operators  indicate that the SCR
processes themselves are very reliable, essentially 100%.  However, in some cases, a boiler  shutdown
has been necessary where air preheater plugging has  occurred.  In most cases  steps  have been  under-
taken to reduce the plugging  rate to  the extent that cleaning is only required during  normal  boiler
outages.
      Some additional process impacts have been reported for retrofit applications (Reference  4-16).
hor example,  in  many cases all  equipment  and  ducting downstream  of the economizer,  including the
stack, will  have  to  be relocated  to make  room  for the reactor  and  the  ammonia flue gas  mixer.
Existing structures, equipment, and other site constraints may interfere with  the required expansion
of  the  back  end of  the boiler, thus  requiring major site rearrangements*. This  requirement could
cause long construction periods and extended unit outages.  Also, larger units may require a booster
                                               4-29

-------
fan to overcome the  added  system pressure drop, causing a conversion from forced draft  to  balanced
draft operation.  This may result in somewhat greater risk of boiler implosion.

Selective Noncatalytic Reduction (SNR)

       Exxon Research and Engineering Corporation developed the SNR process  in which  NHg  is  injected
Into the boiler where  proper flue gas temperatures allow the reduction of NO  by reaction  with  NH,
                              *                                               x                     3
to proceed without a catalyst.   Generally,  40%  to 60% NOV reduction is achieved with  NH,:NOV  molar
                                      •                   X                               OX
ratios of  1:1 to  2:1.   SNR  may be more attractive  than SCR  in  cases where only  40%  to 60%  NO
control 1s needed since SNR is simple and does not require expensive catalysts.

       The general disadvantage of SNR is the limited NO  control  achievable,  especially  with  larger
boilers.   This limited control results from the difficulty of achieving rapid uniform  mixing of  NH,
with the flue  gas and from the  variations  of flue gas temperature and composition  usually present
within the boiler region  where the SNR  is  operated.   NH3 consumption  and unreacted NH3 levels  can
also be high because  of  the high NH,:NO  molar ratios needed with  this process.  Also,  many of  the
same problems of ammonium salt formation, previously discussed  for  SCR  processes, occur  with SNR as
well.
       There are several  large SNR units installed in Japan, between 30- and  100-MW  capacity, mostly
supplied by Tonen Technology (a  subsidiary of Toa Nenryo) which  has  a license from  Exxon.   These
units  are  operated  on gas-  and  oil-fired  boilers  or  furnaces.   Practically  all  are  only  for
emergency use during a photochemical smog alert or when total plant emissions  exceed  the  regulation.

       There are  presently  two commercial  SNR plants operating in  the  United States.   One is on  a
glass  melting  furnace and  the  other  a  petroleum  refinery, both located  in  California.    The
construction  of five  other  industrial-scale units  is  planned.    The  SNR  process  is  also   being
Installed by Exxon at the No.4 oil-fired unit of the Haynes Station of  the Los Angeles  Department of
Water and Power.

Other Flue Gas Treatment Techniques

       In addition to SCR and SNR, dry processes which are being developed for simultaneous SO  and
NO  control include:

               (1)  Activated carbon processes where NH, reduces NO  to  N-;
                                                 4-30

-------
               (2)  Copper oxide processes where NH3 reduces NO  to N2; and
               (3)  Electron beam irradiation processes in which NH, is added
                    to produce ammonium sulfate and nitrate.
Also, work has been conducted on various wet processes for simultaneous SO  and NO  control.
                                                                          X       X
       The optimum  temperature range  for simultaneous  SO   and NO   control  with  activated  carbon
processes is 220°C to 230°C (430°F to 445°F).  Although NOV may be adsorbed below 100°C (212°F), for
                                                          A
treating large quantities  of  flue gas above 100°C  the carbon is mainly useful  as  an  NO  reduction
catalyst.  Therefore,  while NO   is  converted  to N9  by. reaction with NH, in  the presence  of the
                               X                    £                      -3
activated carbon  catalyst,  S02 is simultaneously adsorbed  by the carbon to form  H2SO..   The H-SO.
may also compete for NH., in forming ammonium sulfate  or  bisulfate.   The  formation  of these ammonium
salts increases NH3 consumption and  also  lowers  catalyst activity.   The  carbon must be regenerated,
either by washing or thermal   regeneration.  Washing  produces a dilute  solution.   Concentration  of
the solution to produce  a  fertilizer requires  much energy.   Thus,  thermal  regeneration  seems to  be
preferred.  A concentrated  SO- gas is recovered, which  can be used for sulfuric  acid or elemental
sulfur production.

       The major drawback of the activated carbon processes is the enormous consumption of activated
carbon, which is  more  expensive than ordinary carbon used  only for SOV removal.   Since carbon and
                                                                       X
ammonia consumption increase with the S02 content  of  the flue gas, the process is  best suited for
flue  gases  relatively low  in SO,,.   In Japan  Sumitomo  Heavy  Industries  and  Unitika Company have
operated activated carbon pilot plants of 0.6 MW and 1.5 MW capacity respectively.

       Ihe Shell  Flue  Gas Treatment  process  may  simultaneously remove SO  and  NO .  SO  reacts with
the  copper  oxide acceptor  to  form  copper sulfate.  The copper sulfate and  copper oxide are SCR
catalysts for the NO   reduction by NHg.   Regeneration  of the multiple catalyst beds  by a reducing
gas, such as H2, yields  a S02-rich stream  that can  be used  to produce liquid S02,  elemental sulfur,
or sulfuric  acid.   By eliminating NH3  injection, the process is strictly  an  F6D  process, whereas,
eliminating  regeneration of the catalyst beds allows  the process  to be used  for  only NO  control.
The  major disadvantages are  the  large  consumption of  fuel  for making  hydrogen  and the catalyst
expense.
       In  addition  to  the  EPA-sponsored  pilot plant  mentioned  earlier, the  process  has  been
installed in Japan, on .a  40-MW  oil-tired boiler.   The unit has demonstrated 90%  SO  removal and 7056
NO  reduction.
                                                4-31

-------
       Another process for simultaneous  SO   and NO  control  is the electron beam  process  developed
                                          X        A
by Ebara Manufacturing Company  in  Japan.  NH3 is added to the flue gas, after which the gas  stream
is irradiated with an electron  beam  in  a reactor,  promoting  the conversion of SOV, NOV, and  NH,  to
                                                                                 XX         O
ammonium sulfate and ammonium nitrate.   The  ammonium sulfate and ammonium nitrate may  be  collected
downstream  in  an ESP  or baghouse  and  potentially  sold as  a fertilizer.   The most  economically
practical removal efficiency range appears to be 80% to 90% for each of NOV and SOV> though  higher
                                                                           X       X
removals can  be achieved with  much  greater  electron  beam energy  input.   The optimum  temperature
range is 70°C to 90°C (160°F to' 195°F).
       Ebara has worked on the process since  1971.   It has been tested at a 0.3 MW  and 3 MW scale  in
Japan.   Avco  Corporation  in   the  United  States  has  also  examined   this  technique  and  has   a
cross-licensing  agreement with  Ebara  in sharing  of technology and  in marketing of  the  process.
Although the process appears attractive because of simplicity, simultaneous SOV and NOV  control, and
                                                                              X      X
byproduct formation, there  are  still many questions concerning costs  and byproduct quality  which
must be determined.

       Development  of  an  alternate  electron  beam  scrubbing  process  was  begun   in   1979  by
Research-Cottrell under contract to the  Department  of  Energy  (DOE).   With  this process  a lime spray
drier is located upstream of the reactor.  Calcium sulfate and calcium nitrate are produced  in the
reactor  and  caught  in a  downstream baghouse.   Some bench scale  testing  has been  done with  this
process.  DOE  plans  proof-of -concept scale testing  of both  the ammonia  injection and  lime  slurry
injection electron beam processes on real coal-fired slip streams  (Reference 4-17).
       The wet F6T processes  normally  involve simultaneous  removal  of  SO   and NO .   The  major
problem associated with wet NO  control  processes is the absorption of NO  by  the scrubbing solution
in which it  can be  concentrated and converted  into other forms.   NO   in  the flue gas  is  predomi-
nantly NO, which is much less soluble than NOg, whereas, NO- is even less soluble than  SOg.   The  two
common methods of removing the NO   in  flue gas by wet processes are:   (1) direct absorption  of  the
NO  in the absorbing  solution  or (2) gas-phase oxidation to convert the  relatively  insoluble NO to
NOp.  followed  by  absorption of  NOp.   Presently,  development of  the wet  NO   F6T processes  has
practically  ceased because  of  the complexity  and  unfavorable  economics  of  these  processes  in
comparison with the dry processes.
                                                4-32

-------
4.1.2  Costs    •                                               .

       In this section  some  available  estimates of the capital and annualized  costs  of alternative
NO   control  techniques are  presented.   The  section is  divided  into  two  major  subsections  -
combustion modifications and flue gas treatment.  The  costs and discussion for  combustion modifica-
tions were  largely  taken from two reports  prepared  for EPA. by Acurex  Corporation  (References  4*18
and 4-19).  Much  of the costs and discussion for flue gas treatment techniques were  taken  from a
report prepared for the EPA  by  TVA (Reference 4-20).   However,  as  noted in  the  section,  several
other references were also used.

       The  reader should be  careful  in attempting to  compare the costs presented  in this section.
Some differences  in methodologies and bases  were used  in  developing these costs.    For detailed
discussions of how these costs were developed, the reader is referred to the original  references.

4.1.2.1  Combustion Modification

       Recently, as discussed in the above references, Acurex Corporation conducted an environmental
assessment  for EPA of various utility boiler  combustion modification  NO controls.  As part of this
study costs were developed for a  number of cases representing retrofit and new boiler applications.
The following discussion is largely taken from this study.

       The  use of accepted estimation  procedures for costing  NOX control  implementation in current
dollars was employed  in this study, with  heavy reliance on discussions with boiler  manufacturers,
equipment vendors,  and  utilities.  For the case of retrofit  control  costs, preliminary design work
was  performed to  allow estimation  of  hardware and  installation  needs,  as  well  as  engineering
requirements.  The  analysis  was  applied to a number of  cases to give a  range of  retrofit control
costs.   For the cost  of NOX controls  in  new boilers,  the  services of  two major  suppliers,  the
Babcock & Mil cox Company and the Foster WheeleY Energy Corporation,  were enlisted.

       For  the analysis of the cost of  controls,  regulated public utility  economics were adopted.
Based on  a  revenue  requirement approach,  an annualized cost methodology was developed, adapted from
that used by  the Tennessee Valley Authority in evaluating the cost of  power plant  projects for EPA
(Reference  4-21)  and EPRI  (Reference  4-22).   This  procedure has  been generally  accepted  in  the
industry  (References 4-23 through 4-25).
                                                 4-33

-------
       For the  present application, the additional  revenue  requirement represents the  incremental
cost of operating a boiler under controlled conditions over and above the cost of operating the  same
boiler  uncontrolled.    In  other  words, the  revenue  requirement  takes  into account  the  initial-
investment, the annual capital  charges resulting from  that investment, and  all direct  operating
costs such as operation and maintenance.  This methodology is described in detail  in  Reference 4-18.

Retrofit Control Costs
       Based on  this  cost analysis methodology, typical  retrofit  control  costs (1978 dollars)  are
summarized in Table 4-12.  The costs shown in the table should be  considered  only  representative of
retrofit costs.  They apply to retrofitting  relatively  new boilers,  approximately  5  to 10 years  old
with at least 25 years of service  remaining.   With  the exception of BOOS for coal-fired units,  and
FGR/OFA for  oil-  and  gas-fired units,  annualized  control  costs  generally  fall  in  the  $0.50  to
0.70/kW-yr, based  on  a 7000-hour  operating  year.   For  comparison,  the cost of operating  a  power
plant is approximately $175/kW-yr.
       BOOS was not treated in the  cost analysis as  a  recommended control  technique for coal firing.
Rather  it was  included   to  show  the  extremely high  cost  of  derating.   This  high cost  was  due
principally to the need to purchase make up power and  to account  for  the lost  capacity of the  system
through a capital charge.
       Tables 4-12  and 4-13  present projected  retrofit  control  requirements  for alternative  NO
emission'levels.  Based on favorable process analysis  results, it  is evident  from  an examination of
these tables that OFA and LNB  are  the  preferred, cost-effective  NO  controls for  coal firing.   For
high levels  of NO  control for  coal-fired units (170 ng/J), both OFA  and  LNB  may be appropriate.
For more  moderate  levels of control,  LNB are  less expensive and more  cost-effective  than OFA  in
reducing  NO   in wall-fired  units.   However,  retrofit   of low  NO   burners   may  not  be  widely
            X                                                        X
applicable.
       As far as moderate control  for  oil- and gas-fired units,  staged combustion via BOOS appears
to be the  preferred route, as indicated  in  Tables 4-12  and 4-13.   Initial  investment is minimized
since there are  no  associated  major hardware requirements, only  engineering and startup costs.   To
reach the next level of  NOX control  (86 ng/J for oil,  43 ng/J for  gas), FGR with OFA would seem to
be in  order.   However, this  alternative results in a  cost  increase from  $0.52/kW-yr  for BOOS  to
$3/kW-yr for FGR and OFA.
                                                4-34

-------
                         TABLE 4-12.  SUMMARY OF UTILITY BOILER COMBUSTION MODIFICATION
                                  CONTROL COSTS (1978 DOLLARS) (Reference 4-19)
I
c*>
U1
Boiler/Fuel Type
Tangential /Coal -Fired
OFA
Opposed Wall /Coal -Fired
OFA
LNB
BOOS
Single Wall/Oil- and Gas-Fired
BOOS
FGR/OFA
MCRa
(MW)

225

540
540
540

90
90
Initial
Investment
($/kW)

0.96

0.66
2.17
0.09

0.32
6.09
Annual ized
Indirect
Operating
Cost
($/kW-yr)

0.22

0.17
0.36
5.70

0.05
1.22 .
Annual ized
Direct
Operating
Cost
($/kW-yr)

0.34

0,55
0.06
26.40

0.47
2.04
Total Cost To
Control .
($/kW-yr)D

0.57

0.73
0.42
32.10

0.52
3.26
        aMaximum continuous rating in MW of electrical  output.

         Based on 7000-hour operating year.  Typical  costs only.

-------
I
CO

-------
Control Costs tor New Boilers





       Estimating the incremental costs of NOX controls for new boilers is in some  respects  an  even



more difficult task than costing retrofits.   Certain modifications on new units, though effective in



reducing NO  emissions, were originally  incorporated due to operational  considerations  rather  than
           X


from a  control  viewpoint.   For example, the  furnace of a  typical  new  unit has  been  enlarged'to



reduce slagging potential and allow the burning of poorer quality  fuels.  But this  improvement  also



reduces NO   due  to the lowered  heat  release rate.   Thus, since  the design  change would  have  been



implemented even without the anticipated NO  reduction, the cost  of  that  design modification should
                                           **


not be attributed to NO  control.
                       X




      Babcock & Wilcox  has  estimated  the incremental costs of  NO  controls  on  a  coal-fired boiler
                                                                  X


designed  to  meet  19/1  NSPS (Reference 4-26).   Units designed prior  to 1971  did not  include  NOX



controls.  NO  emissions from coal-fired units  designed  at  that time were on the  order of 430  ng/J



(1.0 lb/106 Btu).    The  1971    NSPS   required   that  these  emissions   be   limited   to   300  ng/J



(0.7 lb/10  Btu).  The  two  units used in the comparison  by  Babcock  & Wilcox were  identical except



for NO  controls on the NSPS unit which included:
      X                                                                               ,


            -  Replacing the high turbulence, rapid-mixing cell  burner



               with the limited turbulence dual register (low NO ) burner;
                                                                X


            -  Increasing the burner zone by spreading the burners



               vertically to include 22 percent more furnace surface; and



            -  Metering and controlling the airflow to each row of



               burners using a compartmented windbox.
To  provide  these' changes  for NO   control,  the price  increase  was about  $1.87 to  $2.67/kW  (1978


dollars).  If these costs are annualized, they translate to $0.30 to 0.43/kW-yr.
       In addition,  hoster Wheeler has performed a -detailed  design study aimed  at  identifying the



       =ntal costs of NO  control to meet
                       X


unit designs with the following results:
incremental  costs of NO  control  to  meet 1971  NSPS  (Reference  4-27).   Foster Wheeler  looked at three
                       X                    •                   ,                       "
                              Boiler Design                      Relative Cost



                    Unit 1:  Pre-NSPS base design                     100



                    Unit 2:  Enlarged furnace, no                     114



                             active NOV control
                                      A
                                                  4-37

-------
                    Unit 3:  NSPS design; enlarged        .            115.5
                             furnace, new burner design,
                             perforated hood, overfire
                             air, boundary air

       For  a pre-NSPS  coal-fired  boiler  costing  about $192/kW  In 1978  construction costs,  the
Incremental cost of active NOX controls (LNB plus OFA) is $2.97/kW, or about $0.47/kW-yr annualized.
The  Foster Wheeler  estimate,  which  includes  both  LNB  and OFA,  thus  agrees  quite  well  with  the
Babcock & Mil cox estimate, which includes only LNB and associated equipment.

       Comparing these costs with the retrofit costs (0.40 to 0.70 S/kW-yr for LNB or OFA)  presented
in Table 4-13 and  considering  the better NO  control  anticipated  with  NSPS units, it  is  certainly
                                            X
more cost effective to implement controls on new units.  Furthermore, fewer operational  problems  are
expected with units specifically designed for these controls.

       Advanced combustion modification concepts under development, such as the EPA advanced  low  NOX
burner  (Reference  4-28)  and  EPRI  primary combustion  furnace  (Reference 4-29),  are  targeted   to
achieve NOX  emission levels around 86  ng/J  (0.2 lb/106 Btu) on a  commercial  basis in the  1980's.
Projected cost  for the EPRI  furnace  is $4/kW  or $0.80/kW-yr  (Reference  4-30).   The  EPA advanced
burner costs  should fall  in  the same  range.   These  developing advanced  combustion  modifications
should eventually  prove much  more  cost-effective  than  the developing  post combustion techniques.
discussed next.   However, the latter techniques are currently closer to  commercialization.

       In conclusion,  conventional  combustion  modifications are indeed a cost-effective  means  of
control for  NOX,  raising the cost  of electricity  less than 1 percent in  most  cases.   Furthermore,
the initial capital investment required should also only be of the order of 1 percent  or less of  the
installed cost  of a  boiler.   Advanced  techniques  such  as  advanced low  NO  burners  and advanced
boiler/furnace concepts have projected costs of the  same  order  as  conventional  combustion  modifica-
tions.   Therefore,  preferred  current  and projected  combustion  modification techniques  are  not
expected to have a substantial  adverse economic impact.

4.1.2.2  Flue Gas Treatment

       As discussed in Reference 4-20, TVA, under contract to EPA,  conducted a  preliminary economic
analysis in 1980 to compare several  flue gas treatment (F6T) processes.   The following discussion is
largely  taken from  this  reference.   Some  additional  cost information  and  references  are also
presented.

                                               4-38

-------
1980 TVA Study

       The TVA study developed preliminary economics, comprising total capital investment and annual
revenue requirement estimates, for  seven  NO   F6T processes.   The economics were calculated based on
a consistent set gf design  and economic premises that have formed the basis  for many previous flue
gas  desulfurization  (F6D) studies  done by TVA.   The reader should  be  cautioned that  at  the time
these  costs  were  developed, most  of  these systems  were  at  an  early  stage  of  development  for
coal.-fired applications.   Thus,  actual  systems could vary  significantly from  the  costs presented
here.

       The FGT  processes  evaluated are shown'in Table 4-14.   These  include one dry  and three wet
processes  for  simultaneously controlling  SOV  and NOV and three  dry processes  for  controlling NOV
                                            A      '  A                                             X
alone.  The dry processes all employ selective catalytic reduction (SCR) for controlling NO .  Among
the  dry processes,  the Hitachi  Zosen process employed a fixed bed reactor  with "honeycomb" shaped
catalyst  cells  through  which the  flue  gas passes  in parallel  flow;   the  Kurabo  Knorca  process
employed a moving bed  reactor with  spherically  shaped catalyst  located downstream of a hot ESP; and
the  Shell  Flue  Gas Treatment process  employed a  fixed  bed, parallel passage reactor in which the
catalyst  is  contained in  unit, cells and  the flue gas  is  forced across  the face  of the  catalyst
layer.

       The power plant assumed as  a basis for this study was a new,  500 MW coal-fired boiler.  The
coal had a heating  value  of 10,500 Btu/lb and contained 3.5  percent sulfur and 16 percent ash.  The
plant efficiency was 9000 Btu/kWh and the boiler on-stream time was 7000 hr/yr.

       In addition to  the design premises for the.NO  FGT process itself,  the design premises for an
overall FGT  system,  including PM  removal  and FGD, were developed  to allow for  comparisons  of the
various  dry  and wet   processes.   The   design  premises for  the FGT  system included  the following
removal efficiencies:
                 -  particulate:  99.5%;
                 -  SOX:  90%; and                   ;
                 -  NOX:  90% from a baseline of 600 ppm.

       The  capital  investment  estimates  were  based  on mid-1979  construction  costs.  The  annual
revenue  requirements  were based on mid-1980 operating  costs  using average capital  charges  with a
'7000 hr/yr on-stream  time.   Additional  details on the design and economic premises  are provided in
Reference 4-20.
                                               4-39

-------
     TABLE 4-14.  NOY FGT PROCESSES SELECTED FOR EVALUATION
              BY TVAXIN 1980 STUDY (Reference 4-20)
          Process
             Type
Dry NOV -only
      A
     Hitachi Zosen
     Kurabo Knorca
     UOP Shell Flue Gas
      Treatment (SFGT-N)
Selective catalytic reduction
Selective catalytic reduction
Selective catalytic reduction
Dry SOX -
     UOP Shell Flue Gas
      Treatment (SFGT-SN)
Sorption of SO  and selective
 catalytic reduction of NO
                          A
Wet SOY -NOY
      A   > A
     Asahi
     Ishikawajima-Harima
      Heavy Industries (IHI)
     Moretana Calcium
Absorption-reduction
Oxidation-absorption-reduction

Oxidation-absorption-reduction
                               4-40

-------
       The capital investment and annual revenue  requirement  estimates  by  TVA for these systems >are
presented in Tables 4-15 and 4-16,  respectively.   The capital investments for the various  NO -only
FGT processes  ranged from  $38/kW to  $48/kW.   For  the  combined SCR-FGD-ESP systems, the  capital
investments ranged from $165/kW  to  $175/kW.   For the wet SO  -NOV processes  the  capital  investments
                                                            A   X
ranged from $205/kW to $482/kW.   The dry. SO -NO -ESP was $169/kW.
                                           A   A
       The annualized costs of a dry SCR system ranged from 2.1 to  3.6  mills/kWh.   The total  system
costs of SCR combined with  FGD  and ESP ranged from 7.1 to 8.6 nrills/kWh.   In comparison,  the costs
of a dry SOX-NOX-ESP system were 7.5 mills/kWh and the costs  of the wet SOX-NOX  systems ranged from
12 to 20 mills/kWh.

       The results of  this  study can  be  summarized as follows.  The wet  SO -NO  processes  do not
                                                                                *
appear economically  attractive  for new power  plant  applications  when compared with either  the dry
SOX-NOX process  or the SCR-FGD  systems.   Comparisons between the  dry  SFGT-SN process and  SCR-FGD
systems were close enough to  be inconclusive considering the state of  development of  these systems
at that time.  Comparisons within the  SCR-FGD  systems, i.e.,  moving versus parallel  flow,  fixed bed
reactors were  also inconclusive.  However,  recent  trends in Japan indicate the  fixed bed  reactor
systems are preferred.
0_thjir; Stud1 es
       Some other studies have also presented cost information for various FGT systems as summarized
below.                     '
       A 1981  TVA study  (Reference  4-6) made  a  preliminary economic  evaluation of  three  control
methods for obtaining  50  percent  NO   reduction and three methods for obtaining  90%  reduction.  The
base case power plant was a new 500 MW  coal-fired unit  emitting  0.6  Ib  NO-XIO Btu in the flue gas.
Capital  investment estimates  were based  on projected  mid-1982 construction  costs.   The  revenue
requirements were based on projected 1984 costs.
       The three  50  percent NO   reduction  processes  evaluated  were the  EPA  sponsored advanced low
NOX  burner (ALNB),  the  Exxon Thermal  DeNOx  process,  and  the Hitachi  Zosen  SCR  process.   For
90 percent NO  reduction the ALNB was combined with the Hitachi  Zosen process, the Exxon process was
combined with  the Hitachi  Zosen  process, and  the  Hitachi  Zosen process was used  alone.   Capital
                                                                                          *
investments and  annual revenue requirements  for these  processes  are presented  in  Tables 4-17 and
4-18, respectively.
                                            4-41

-------
            TABLE 4-15.   CAPITAL  INVESTMENT DEVELOPED BY TVA FOR
                  ALTERNATIVE FGT SYSTEMS (Reference 4-20)
Total Capital Investment

Process

FGT
M$
FGD ESP

Total
$/kW
Total
$/aft3/nri n
Total

Dry NO -only
x\
UOP SFGT-N
Kurabo Knorca
Hitachi Zosen
Dry SOX-NOX
UOP SFGT-SN3
Wet SO -NOV
„ A /\
Moretana Calcium
Asahi
IHI

18.4
21.2
23.3

67.2

88.0
104.9
203.6

50.4 10.8
50.4 12.1
50.4 10.8

14.6

7.2
_ -
— -

79.6
83.7
84.5

81.8

95.2
104.9
203.6

165
174
175

169

205
233
482

37.1
39.0
39.4

38.1

44.4
48.9
94.9
aBased on hydrogen production from naphtha and H9SOA byproduct from S09.
       TABLE 4-16.  ANNUAL  REVENUE REQUIREMENTS  DEVELOPED  BY  TVA  FOR
                 ALTERNATIVE  FGT  SYSTEMS  (Reference  4-20)
Annual Revenue
Requirements,

Process

FGT
M$
FGD

ESP

Total
Equivalent Unit
Revenue Requirements,
mills/kWh
Total

Dry NO -only
A
UOP SFGT-N
Kurabo Knorca
Hitachi Zosen
Dry SOX-NOX
UOP SFGT-SNa
Wet SO -NO
Mofetafia Calcium
Asahi
IHI

7.2
9.3
12.2

22.5

38.1
39.8
58.6

14.7
14.7
14.7

-

_
-
—

2.2
2.7
2.2

3.0

1.5
-
-

24.1
26.7
29.1

25.5

39.6 .
39.8
58.6

7.13
7.91
8.60

7.53

12.20
12.63
19.82
Based on hydrogen production from naphtha and
                                                     byproduct from
                                    4-42

-------
      TABLE 4-17.   SUMMARY OF  CAPITAL  INVESTMENTS DEVELOPED
                 IN  1981 TVA STUDY  (Reference 4-6)
     Process
                                        Capital  Investment
                                      (projected nrid-1982$)
 M$
 $/kW
50% NO  Reduction
      y\

     ALNB
     Exxon
     Hitachi Zosen

90% NOV Reduction
      A

     ALNB/Hitachi Zosen
     Exxon/Hitachi Zosen
     Hitachi Zosen
 2.4
 9.9
15.7
25.9
32.1
25.5
 4.8
19.7
31.4
51.8
64.2
50.9
   TABLE  4-18.   SUMMARY  OF  ANNUAL  REVENUE  REQUIREMENTS  DEVELOPED
                 IN  1981 TVA  STUDY (Reference  4-6)
Annual Revenue Requirements
(projected 1984$)

Process

M$
First Year
Mills/kWh

Mi
Level ized
Mills/kWh

50% NOV Reduction
X
ALNB
Exxon
Hitachi Zosen
90% NOV Reduction
X
ALNB/Hitachi
Zosen
Exxon/Hitachi
Zosen
Hitachi Zosen


0.45
3.4
8.0


11.5

14.2

13.3


0.17
1.2
2.9


4.2

5.2

4.9


0.54
5.1
13.0


18.4

22.6

21.9


0.20
1.9
4.7


6.7

8.2

7.9
                               4-43

-------
       As expected, the ALNB, a combustion modification,  is  projected  to  be  the  least  expensive  NO
control method  among  those studied.   Also,  the costs  for obtaining  high levels  of NO   reduction
(90 percent) are significantly greater than for more moderate levels (50 percent).

       In November  1982  the Electric  Power  Research Institute completed  a  detailed technical  and
economic  review of  four  FGT processes  (Reference  4-31).   The  economic  analysis was  based  on
preliminary designs for a 500 MW  (net)  capacity power plant  located in the midwest  and burning  low
sulfur (0.5%) Powder River Coal.   The results of the economic analysis  are presented  in Table 4-19.

       As found in previous  studies,  non-catalytic processes are less  expensive  than  catalytic
processes from both a net cost and  a  cost  per ton  of pollutant  removed basis.  Also, costs increase
significantly  for  higher  levels  of  control  and  for, treating  flue  gases  with  higher  NOX
concentrations.

       In a recent paper  (Reference 4-32), Ando reports that most coal-fired boilers in Japan will
use a  flue  gas cleaning  system consisting of SCR  using a honeycomb or plate catalyst, an  ESP,  and
FGD by the  limestone-gypsum process using  a  prescrubber with a separate liquor loop.  The  costs  of
generating power by coal, including, gas cleaning and wastewater treatment, is less than that by  low
sulfur  oil  without gas  cleaning.   SCR is  much  less  costly  than FGD  but  is more  costly than
combustion modification.  Simultaneous  SOV/NOV removal  processes  and  fluidized bed  combustion have
                                         A   A
not yet proved to be better alternatives to combined cleaning.

4.1.3  Energy and Environmental  Impact

       In addition to  affecting  the cost  of operating  electrical  generating combustion equipment,
Implementing NO  control  techniques can also impact overall plant  efficiency  and  emissions  levels  of
pollutants other than NO .  These energy and environmental impacts  are  discussed  below.

4.1.3.1  Energy Impact
       This section  discusses the  energy impacts resulting or  expected with  various combustion
modification and flue gas treatment NO  control techniques.
                                                4-44

-------
TABLE 4-19a    ESTIMATED TOTAL CAPITAL REQUIREMENT OF POST COMBUSTION' MOX CONTROL  PROCESSES
                                            (S/fcW, 1979)  ...
          NON-CATALYTIC
inlet NOx, PPm
NOx Reduction, %
Outlet NOx. PPm
Exxon • Dual Grid
Exxcn - With «2

33
200
11
12
300
50
150
11
13

55
135
12
13
600
55
270
14
17
          CATALYTIC
Inlet NOx. ppm
t4Qx Seduction, %
Outlet NOx. opm
Kawasaki H!
Hitachi Zosen
Shell FQT*
300
67
100
41
58
161
S3
50
42
85
169
90
30
46
68
163
600
§5
30
56
73
17S
  TABLE 4-19b  ESTIMATED LEVELIZED COST REQUIREMENT OF POST COMBUSTION  NOX CONTROL PROCESSES
                                             (MUls/kWh 1979)
           NON-CATALYTIC
Inlet NOx. PPm
NOx Reduction, %
Outlet NOx, ppm
Exxon - Dual Grid
Exxon - With »2
300
33
200
0.96
0.97
50
150
1.12
1.18
55
135
1.19
1.24
600
55
270
1.68
1.85
            CATALYTIC
Inlet NOx, pom
NOx Reduction, %
Outlet NO», ppm
Kawasaki HI
Hitachi Zosen
Shall FGT*
300
67
100
5.20
5.05
11.00
83
50
6.56
7.30
12.00
90
30
7.12
8.00
12.10
600
95
30
9.09
9.75
12.90
            Accuracy; +30% - »0%
            "Include* 20% Removal of 500 PPm SO2
            (Reference 4-31)
                                             4-4S

-------
Combustion Modifications
       The following  discussion  of combustion  modification energy  Impacts  nas  taken  largely  from
Reference 4-33.  The  largest potential  energy  Impact of  combustion :ioJi fixations  is  their  effect
upon boiler  thermal  efficiency.   Another significant source of  energy  "impact 1s the change  1n  fan
power requirements caused by these '.ontrols.  Boiler con'rol sy:tems Installed for low NO  operation
also Increase electricity and instrument air requirements, but the energy Impact Is usually minimal.
Some discussion of the  energy  Impacts  ot applied  NOX controls  was already  presented on  a boiler-by-
boner  bas^s earlU.r In  Section 4.   As  noted  thore, with  proper  engineering  and  Implementation,
there should be  no  major adverse energy  impacts with  preferred  combustion  modifications.   A review
of  that analysis follows.

       Applying  low  excess  air (LEA) firing not  only  results  In a  small decrease  1n  NO  emissions
but also  results in  an increase 1n boiler efficiency through  reduced  sensible heat  loss  out  the
stack.   For  this  reason the  technique has  gained acceptance  and  das become  more of  a standard
operating procedure  than a specific NO   control method 1n  both old and new units.

       The  other  commonly applied combustion modifications, OSC and FGR, often  lead to decreases in
bci'ler  efficiency  when  implemented  on  a  retrofit basis.    OSC  usually   increases  excess  air
requirements res'iltinq  In decreases  in  efficiency of up to 0.5 parcent.  Unbu^ned fuel  losses either
due to OSr.  or FGR  may cause a decrease  in  efficiency   of  up  to  O.b percent.  If  a substantial
increase  in  reheat  steam attemperation is required due to OSC or FGR, cyclt; efficiency losses of up
to   1 percent  may  occur.   Increased  tan  power requirements  due to  OSC  or  FGR will   also i.npact
efficiency,   resulting  1n  losses  of  up  to 0.?  to 0.3  percent.  No  significant energy  impact is
expected  with LNB,  either retrofit or  new Installation.

        Other combustion modification  techniques,  WI  and  RAP, ciir  Impose  quite significant energy
 penalties on boiler operation, with  decreases  in  efficiency from b  to  10 percent.  As a consequence,
 these techniques have found  little acceptance.
        These decreases  in aoiler etticiency (increases  1n  energy  consumption)  discussed above  for
 the  preferred NOX  control   techniques  (OSC,  FGR,  and   LNB)  are  mo  t  likely  to  occur  when  the
 techniques are applied 01.  a  retrofit basis.  These  same combustion  modifications are not  expected  to
 adversely affect  unit  efficiency when included  1n  the   design  of  a  new  unit.   Thus, with  proper
 engineering  and development, combustion modification NO,   controls can be Incorporated  into  new  unit
 designs with no significant adverse energy Impacts.
                                                 4-46

-------
       A related problem that may occur witr  retrofit  application  of some  techniques  1s  derating of
the unit.  Loss 1n boiler  load  capacity due tc limited coal pulverizer capacity will  occur 1n nwny
coal-fired boilers operated  with burners-out-of-serv1ce  (BOOS).   Derating of 10 to  25  percent may
occur   For o1l-f1ret boilers on  OSC,  higher excess air  requirements nay  causa  fan  capacity limits
to be  reached  1n some cases.  Although derating  due to fan  capacity  is  not common,  reductions of up
+o 15  percent  have been reported.  With  OSC and flue  gas  redrculation   (FGR), excessive  tube anu
steam  temperatures may  lead to  derating,  especially for gas  fired  boilers,  and 1n  some  cases for
oil-fired boilers.

Flue Gas Treatment

        Estimates of the energy  Impacts of various dry NOX control systems  i»i:d of dry and wet SOX/NOX
control  systems have been  presented  in  several sources (see References  4-31 and 4-34 through 4-38).
The  following  discussion  focuses on  the  direct  energy  Impacts of  FGT systems.  However,  many of
these   systems  utilize  chem.cals  such  as  ammonia  which  are  energy   intensive  to  manufacture
themse'ves.

        The energy  requirements  of dry SCR  NOX control  systems  result primarily from the e'ectrical
energy required to overcome  the reactor pressure drop  and the compressed air and steam usijd  for  soot
blowing.  These requirements range from about  0.2 percent 01  the holler capacity for gas-fired units
to about 0.3 percent for coal-fired units.

        When  5CR systems  are used  1n combination  with wet  FGO systems  and  ESPs,   total  flue gas
cleaning energy rsquireiKnts for  coal-fired boilers range from about 3.5  to 4 percent of the boiler
capacity.   By  comparison, recent estimates  of  the  energy  requirements  of  the Shell  FGT process
achieving  similar  control  levels  (90 percent control  of SOX and N0x)  are about 5  percent of the
boiler capacity (Reference 4-36).   Most of  this energy  consumption  is  in  tne fo>m of fuel  reqiired
by  the process.
        Fnergy  requirement  estimates of various wet  SOX/NO  control processes are considerably  higher
 - about 8  percent  of the boiler capacity  for the Moretana Calcium process, 11 percent for  the  Asahi
 process, a  d 19 percent for the  IHI  process  (Reference 4-35).   For  the latter process, most  of the
anetgy Is  consumed  in  gt-neratlng  ozone.

        Enc "gy   requirements  for  the   £xxon  Thsrmal  OeNOx  process   have   been  estimated  at  about
 0.4 percent  of tne boiler capacity  with most of the energy consumed  by the large air  compressors  in
 the ammonia  storage and  injection  section  (Reference  4-34).   However,   these  estimates  are for  a
 lower level  of NOX  control (about 50  percent) than  the processes described above.
                                                 4-47

-------
       Estimated energy requirements for achieving large reductions of SOX and NOX with the electron
beam process  range from about  3 to 4  percent  of the  boiler  capacity (References 4-32  and  4-371.
However, this process 1s still relatively far from commercial application.

4.1.3.2  Environmental Impact
       This  section  discusses  the  environmental  Impacts  resulting  or  expected  with  various
combustion modification and flue gas treatment NOX control techniques,.

Combustion Modifications
       The Environmental Protection  Agency  recently  sponsored  a throe year evaluation of combustion
modification  controls  for  emissions  of  NOX  and other pollutants from  stationary combustion sources.
Some  of  the results of  this  program Included  field  tests  of  gaseous, liquid,  and solid effluents
from  seven  stationary  combustion sources and estimates of environmental effects of using combustion
modification   control.   Detailed  results   are   presented   1n  Reference 4-38.   A  few  highlights
pertaining  to  utility  and  Industrial boilers  are  presented below.

        Environmental  assessment  field  testing was conducted on five boilers.  Table 4-20 summarizes
the  key aspects  of these  field  tests.  Test  results  were  evaluated  by  comparing effluent stream
pollutant  concentrations  to discharge  stream compositions desirable  to preclude adverse effects  to
human  health.   In addition to NO  and  SO^, potentially hazardous flue  gas streair pollutants  include
vapor  phase  SO, and condensed sulfate.  organic acids, and trace elements such is As, Be, Cd, and  V.
Coil-fired  sources are generally more  hazardous due to these pollutants.  Potentially hazardous ash
stream pollutants from coal-fired  sources  are the trace elements  Fe, Nn, Cr,  N1,  Be,  Ba, Pb, and
occasionally  As,  Se,  Tl, and  Sn.

        Conclusions evident from the  field testing and  analysis  program were as  follows.
                (1)  For the sources  tested,  the flue  gas  stream presents  the
                     greateit  potential  environmental  hazard.
                (2)  NOX and SCU are  the potentiaMy  most  hazardous  flue gas,
                     pollutants.
                                                  4-48

-------
                           TABLE 4-20.  PARTIAL SUMMARY OF EPA/ACUREX COMBUSTION MODIFICATION
                                 ENVIRONMENTAL ASSESSMENT FIELD TESTING (Reference 4-38)
Source Category
Coal-fired
Utility Boiler
Description
Kingston #6; 180 MW
tangential; twin
furnace, 12 burners/
Test Points
(Ur.it Operation)
Baseline
Biased Firing (2)
BOOS (2)
Sampling Protocol
Continuous NO , SO,,
CO, CO, 0 * i
Inlet to^lst ESP:
lest
Collaborator
TVA
                       furnace,  3  elevations;
                       cyclone,  2  ESP's  for
                       particulate control
— SASS
— Method 5
— Method 8
-- Gas grab (C.-C,
                                                                                           HC)
                                                                                      .-,
                                                                        Outlet ot  1st ESP0
                                                                        -- SASS
                                                                        -- Method  5
                                                                        - Method  8
                                                                        — Gas grab  (C.-C, HC)
                                                                        Button ash   *
                                                                        Hopper ash (1st ESP,
                                                                          cyclone)
                                                                        Fuel
                                                                        Operating  data
Coal -fired
Utility Boiler















Oil-fired
Utility Boiler






Crist 17; 500 MW Baseline
opposed wall-fired; DOOS (2)
24 burners, 3 eleva-
tions; ESP for
partlculate control












Moss Landing 16 aselfne
740 MW Opposeo wall- .UR
fired; 48 burners, FGfi * OF A
6 elevations





Continuous NO., CO, Exxon
co2, o x
ESP tnlet
— SASS
— Method 5
- Method 8
— Gas grab (C.-C, HC)
ESP Outlet l °
— SASS
— Method 5
— Method 8
— Gas grab (C,-C HC)
Botton ash l
FSP hopper ash
Fuel
Operating data
Bioass^y
Continuous NO , CO, ' None
co2. o, x
- SASS
— Method b
— Method 8
—Gas grab (C,-CK HC)
Fuel '- 5
Operating Data
Bioassay
Key to ?crony»s: (see end of table).

-------
                                                      (ABLE 4-20.   (Continued)
         Source Category
    Description
   Test Points
(Unit Operation)
Sampling Protocol
    Test
collaborator
         Coal-fired
         Industrial Boiler
Traveling grate
spreader stoker,
38 kg/s (300,000 Ib
steam/hr)
Baseline
LEA + high OFA
Continuous HU , CO,
  C0?, 0?    x
Boilfer exit:
— SASS
~ Method 5
— Shell-Emeryville
-- Gas grab (C,-C,. HC)
ESP outlet    l  °
— SASS
~ Method 5
— Shell-Emeryville
— Gas grab (C,-Cfi HCJ
Bottom ash    x  °
Cyclone hopper ash
Fuel
Operating Data
     KVB
         Coal-fired
         Industrial Boiler
01
o
Traveling grate
spreader stoker,
25 kg/s (200,000 Ib
steam/fir), ESP for
particulate control
Baseline
LEA
Continuous NO , CO,
  co2, o2    x
Boiler exit:
~ SASS
— Method 5
— Shell-Emeryville
— Gas grab (C.-C,- HC)
ESP outlet    *  °
~ SASS
~ Method 5
-- Shell-Emeryville
— Gas grab (C^C,- HC)
Bottom ash    x  °
ESP hopper ash
Fuel
Operating Data
Bioassay
     KVB
         Key to acronyms:

         BOOS:    btaged combustion technique with burners out ot service.
         ESP:     Electrostatic precipitator.
         FGR:     Flue gas recirculation through burner windbox.
         LEA:     Low excess air firing.
         OFA:     Staged combustion with overfire air injection.
         SASS:    Source assessment sampling system for organic and inorganic  emission  collection  and measurement.

-------
               (3)   The total  flue gas hazard is  decreased or,  at worst,  does
                    not increase with applying the combustion modifications
                    tested.   Changes in emissions due to day-to-day fuel
                    composition changes are often of greater magnitude  than
                    those attributable to NO,, control.
                                            X
               (4)   The effluent streams from the sources tested are not
                    mutagenic, and, in general, elicit nondetectable toxicity
                    in bioassay testing.
               (5)   The combustion modifications  tested:
                    —  Have no effect, or increase only slightly, emissions
                        of CO and vapor phase hydrocarbon;
                    ~  Have no effect on particulate mass emissions;
                    —  Have no effect, or tend to increase slightly, emitted
                        particle size distribution;
                    ~  Have no measurable effect on inorganic  trace element>
                        emissions or on trace element partitioning tendencies;
                    —  Have no effect, or decrease slightly S03 and particulate
                        sulfate emissions;
                    —  Have little effect on total higher molecular weight
                        organic emissions; and
                    —  Marginally increase POM emissions; however, emission
                        levels remained on the order of the detection levels  .
                        of the instrument.
               (6)   Emissions of many organic priority pollutants were  below the
                    detection limit for the sources tested.

       It was  further  concluded  in  the  program that  the NO   control  methods  investigated are
environmentally  sound  since  the  potential  adverse  impact  of  the  controlled  source  is  either
decreased or unchanged with NOV control applications.  However, it  should be pointed out that these
                              A
conclusions are based on  short  term  tests  conducted  under steady  operating conditions, and  that the
controls tested were the relatively straight forward, current technology  combustion modifications.
                                                4-51

-------
Flue Gas Treatment
       Since F6T processes have been  less  extensively applied than combustion modifications,  their
environmental impacts  have  been less extensively  tested.   However,  the principal  impacts  expected
can be qualitatively assessed.  The following discussion presents some  assessments  of  these impacts
and the  results  of some pilot unit testing  sponsored  by  EPA.  The discussion focuses  primarily  on
SCR processes since these are the most extensively developed and applied F6T processes.

       The  principal   impacts  expected  to  result  from the  application of  SCR processes are  as
follows.  NOV emissions in the flue gas are reduced considerably, but some  emissions of ammonia and
            A
ammonium compounds will result and some  increases  in  SO, or other  compounds are  possible.   Also,  in
some  cases wastewater treatment  and disposal  may  be complicated  by  the addition  of  nitrogen
containing  compounds.   An  additional  impact  with catalytic  systems  is   the  need  for  catalyst
disposal.
       By way of comparison, SNR  processes  are less efficient at removing NO   from the flue  gas.
                                                                              X
And the  increased ammonia injection levels will result  in greater emissions of ammonia  and  ammonium
compounds  than   the  SCR  processes,   potentially  leading to greater  wastewater impacts  as  well.
However, these processes will  not likely increase flue gas  SO,  levels  and  they are not faced  with
the need for catalyst disposal.
       In  a recent  report  covering  a  survey of SCR  systems in  Japan  (Reference  4-39),  Jones
evaluates  potential  SCR environmental  impacts as  follows.  NO   emissions from boilers  utilizing
catalytic  de-NO   are  generally reduced  by  80 percent.  Higher  reductions  are  possible, but  costs
will be  greater  for these units.   NH, emissions for oil-  and gas-fired  applications are reported  to
be  about 3 to 10  ppm.   Emissions from  full-scale,  coal-fired  facilities  will  not be known  until
after the  start-up of  several  units  currently under construction.  In  addition to concern  about NH,
emissions, there is some concern that other compounds, such  as cyanides, nitrosoamines  and  nitrates,
may also be emitted.    SCR  system vendors  and  operators, however, were not aware  of  any  instances
where compounds  such  as these were  emitted.   The possibility  of  a  visible ammonium  sulfite  plume
resulting from NH, emissions  entering a  downstream, wet FGD system may  be a problem if the  slip NH,
is  high  (>50 ppm).   Plumes  of  ammonium  sulfite are  known to  occur  during  certain  atmospheric
conditions  when  NH_   based  FGD  systems  are  used.   However,  the  Japanese have  experience  with
situations  in which  10 ppm of NH3 enters  the scrubber and,  based on  this  experience, do  not  feel
that SCR systems will   cause visible plume formation.
                                                 4-52

-------
       Jones goes on to  discuss  other potential environmental impacts  including  nitrogen  compounds
in the wastewater and  catalyst disposal.  NH,  can  enter the wastewater  through  the F6D  system  or
from air preheater  washwater.   In locations where  discharge  of  this water will  cause  problems,  an
activated sludge  technique can  be used  to  treat  the  wastewater.  In  other locations  it may  be
possible to  blend wastewater  containing NH, with  other water  discharges.   The catalyst disposal
issue has not been  fully addressed.   While the process  vendors  indicate that they will dispose  of
spent catalyst, the specific method  of disposal had not been identified  at the time  of the survey.
This is  partially due to  the fact  that none  of  the full  scale systems had  required a catalyst
change.
       As discussed earlier in  Section 4, EPA recently  sponsored  pilot plant  tests  of two  SCR
processes applied to coal-fired  utility boilers.   These  included the Hitachi Zosen  and Shell  Flue
Gas  Treatment  processes.   As  part  of  the  pilot  plant  tests,   EPA conducted a sampling  program
designed  to quantify  process  emissions  of  pollutants  other  than NO   and  S0«.   The  test  and
evaluation  program  is  discussed  in  Reference 4-13.  Parts  of that discussion,  pertaining to  the
environmental impacts of these processes, are presented below.

       The objective of the sampling program was to determine  if any adverse  flue gas concentration
impacts  can result from  application  of SCR  technology to a  coal-fired   source.   Samples  were
collected at the inlet and outlet  of the pilot  plant  reactors  and analyzed for selected pollutants.
Inlet and outlet  samples were collected simultaneously,  so  that measured differences  in  pollutant
concentration were due to changes across the reactors.

       Table 4-21  identifies  the  specific pollutants  measured  and presents  the  results of  the
sampling program.  As shown, the concentrations of  some pollutants changed across the reactor  while
others did not.  And in some cases, the concentrations of pollutants were below the detection limits
of the analytical techniques employed during the sampling program.

       As illustrated  in Table 4-21,  concentrations  of hydrogen  cyanide and nitrosoamines at  the
outlet of the  pilot plant  reactors  were  below the detection limits of the analytical  techniques
employed.  For hydrogen cyanide, the detection limit is equivalent to 10 ppb  and for nitrosoamines a
maximum of 2 ppb (the actual concentration level depends on the nitrosoamine  compound(s) present).
                                                 4-53

-------
                      TABLE 4-21.   MEASURED CHANGE IN THE CONCENTRATIONS OF SELECTED POLLUTANTS

                                       ACROSS THE EPA PILOT PLANT SCR REACTORS (Reference 4-13)
I
in
Pollutant
Hydrogen Cyanide
Nitrosoaminesa
Carbon Monoxide
Hydrocarbons (C, - Cg)
Sulfur Tri oxide
Ammonia
Hitachi Zosen
Reactor Inlet
Concentration
<10 ug/m
3
<5 yg/m
<0.017*
<1.0 ppmv
8.4 ppmv
0
Pilot Plant
Reactor Outlet
Concentration
<10 yg/m
<5 ug/m
<0.017%
<1.0 ppmv
20.7 ppmv
54.8 ppmv
Shell
Treatment
Reactor Inlet
Concentration
<10 yg/m3
<5 yg/m
0.13%
28.5 ppmv
11.4 ppmv
0
Flue Gas
Pilot Plant
Reactor Outlet
Concentration
<10 yg/m
<5 yg/m
<0.017«
21.0 ppmv
0.1 ppmv
15.3 ppmv
         Gas  volume taken  at  i5.5°C,  1 atm, dry basis.

-------
     At the Hitachi Zosen plant, the concentrations "f both CO and hydrocarbons were also below the
detection limits of the analytical techniques.  But at the Shell FGT pilot plant, the concentrations
of these compounds were found to decrease across the reactor.  This decrease represents an environ-
mental benefit for the Shell process, and 1t 1s believed to result from oxidation of these compounds
1n the reactor.  It should be noted that the Hitachi Zosen reactor may also oxidize CO and hydro-
carbons, but no conclusions can ue drawn since those compounds were present 1n such low
concentrations.

     One of the most significant results of the emissions sampling program was the effect of the
Hitachi loicn and Shell processes on the concentration of SO* In the flue gas.  As shown In
Table 4-21, SO, was produced In the Hitachi Zosen pilot plant reactor while the Shell FGT process
removed SO, from the gas stream.  These results are significant because of the effects SO- can have
on equipment 'located downstream of an SCR reactor, especially the air preheater.

     Another significant result of the er  Mons sampling program is the measured NH~ emissions from
the  processes.  As shown in Table 4-21, NH. emissions from the Hitachi Zosen process were over three
times greater  than the emissions from the Shell FGT process even though the NH./NO  Injection ratio
was  higher at  the  Shell pilot plant.  The higher NH, emissions are due to the  fact that  the catalyst
used 1n the Shell  process  promotes NH., oxidation while essentially i.b NH, oxidation occurred 1n    •
Hitachi Zosen  process.

4.2     INDUSTRIAL  BOILERS

        Industrial  boilers  are a  /ery common piece  of equipment  in  industrial plants.  These boilers
typically  range 1n size  from 3  to 250 MW  thermal  inpi.'t  (10 to 850  x  10  Btu/hr)  am<  Include a wide
variety  of  firing  types  and fuels.   In  1980 industrial  boilers  represented  the second largest
stationary  source  of  NOX emissions,  preceded  only  by electrical  utility boilers  (See Chapter 2).
Nationwide  NOX emissions from  Industrial  boilers  were  estimated to be 3 Tg  (3.3  x  10  tons)
annually.

        Industrial  boilers  are  topically classified by  the type  of firing  mechanism employed, the
 heat transfer mechanism, and  Che type  of fuel  fired.   Firing mechanisms  Include either  burners,
 spreader-fed, or mass-fed.  With burners, the fuel  is  Injected  Into  the boiler through  a nozzle  anO
 burns while suspended within the boiler combustion chamber.   Mass-fed arid spreader-fed  boilers are
                                                 4-55

-------
used for most solid fuel industrial boilers.  They combust the fuel  on a grate in the boiler.

       Watertube is the most common mechanism used for heat transfer in industrial  boilers.   In
watertube boilers the water for steam generation is contained in banks of tubes suspended in  the
boiler combustion chamber and flue.  Firetube boilers invert this configuration and pass hot  flue
gases through tubes suspended in a water drum.  Firetube boilers are seldom sized larger than 7.3 MW
(25 x 10  Btu/hr) thermal input (Reference 4-40).

       Industrial boilers,are fired with a wide variety of fossil and nonfossil fuels.  Most  common
among the fossil fuels are natural gas, accounting for 43 percent of the fossil fuel-fired
Industrial boiler capacity, followed by fuel oil and coal comprising 32 percent and 25 percent,
respectively.  Nonfossil fuels fired in industrial boilers include wood, bark, agricultural waste,
municipal waste, and industrial waste; the most common being wood and bark.  Nonfossil fuels  account
for less than 5 percent of the industrial boiler capacity.

       The following discussion on NO  emission control techniques for industrial boilers will  focus
on fossil fuel-fired boilers below 73 MW (250 x 10  Btu/hr) in size.  Those boilers greater than
approximately 73 MW are essentially identical to utility boilers and are able to apply the same NO
emission control technologies that are addressed in Section 4.1.  Additionally, non-fossil fuels
generally exhibit relatively low NOV emissions with respect to solid fuels resulting in a lack of
                                   A
demonstrated NO  emission control technologies for non-fossil fuels  (Reference 4-41).  The
population distribution of U.S. fossil-fired watertube boilers by size range is presented in
Table 4-22.  The corresponding material for U.S. firetube boilers is presented in Table 4-23.

4.2.1  Control Techniques
       Currently, the most promising NOV control options for industrial boilers include combustion
                                       X
modification and post combustion techniques.  The former has been the most successful and widely
used option, and is described below for gas-, oil- and coal-fired units.  The post combustion
techniques are discussed after combustion modification and are currently being demonstrated for some
industrial boiler applications.
                                                4-56

-------
                                               TABLE 4-22. INSTALLED CAPACITY OF U.S. HATERTUBE  INDUSTRIAL BOILERS

                                                             BY UNIT SIZE AND FUEL TYPE  (IN  1977)


                                                                (MW Thermal  Input (106 Btu/hr))
Fuel

(0 to 2.9
(0 to 10)

2.9 to 14.7
(10 to 50)
Capacity by
14.7 to 29.3
(50 to 100)
unit size
29.3 to 73.3
(100 to 250)

>73.3
(>250)

Totals

•ulverized coal
Spreader-stoker coal
Inderfeed-stoker coal
Dverfeed-stoker coal
Fotal Coal
Residual oil
Distillate oil
Total Oil
Natural gas
Total all fuels
0
70
(240)
680
(2,300)
85
(290)
835
(2,830)
3,960
(13,500)
2,560
(8,700)
6,520
(22,200)
4,475
(15,300)
11,830
(40,330)
0
(0)
4,650
(15,900)
14,105
(48,000)
3,470
(11,800)
22,225
(75,700)
48,190
(164,000)
8,280
(28,200)
56,470
(192,200)
57,900
(197,500)
136,595
(465,560)
0
(0)
6,175
(21,060)
17,265
(58,900)
4,455
(15,200)
27,895
(95,160)
36,640
(122,000)
4,295
(14,600)
39,935
(136,600)
53,585
(182,800)
121 ,415
(414,560)
19,895
(67,800)
20,295
(69,000)
7,000
(24,200)
3,555
(12,100)
50,825
(173,100)
44,790
(153,000)
6,370
(21,700)
51,160
(174,700)
63,320
(216,000)
165,305
(563,800)
40,180
(,137,000)
11,010
(37,600)
5,230
(17,800)
3,510
(12,000)
59,930
(204,400)
43,570
(148,600)
4,085
(13,900)
47,655
(162,500)
95,935
(327,200)
203,520
(694,100)
60,075
(204,800)
42,200
(143,800)
44,360
(151,200)
15,075
(51,390
. 161,710
(551,190)
176,150
(601,100)
25,590
(87,100)
201,740
(688,200)
275,215
(938,800)
638,665
(2,178,190)
I
in
    Reference 4-40.

-------
                       TABLE 4-23.  INSTALLED CAPACITY OF U.S.  INDUSTRIAL FIRE-TUBE BOILERS

                                        BY UNIT SIZE AND FUEL TYPE (1977)


                                         (MM Thermal Input (106 Btu/hr))
Fuel
Coal
tesidual Oil
)i still ate Oil
latural Gas
Total All Fuels
Unit
0.1 to 0.4
(0.4 to 1.5)
1,690
(5,700)
8,960
(30,600)
4,160
(14,200)
15,420
(52,600)
30,230
(103,100)
Capacity, MW
0.4 to 2.9
(1.5 to 10)
3,960
(13,500)
26,320
(89,800)
13,610
(46,400)
43,700
(149,100)
87,590
(298,800)
Thermal (106
Btu/h)
2.9 to 7.3 7.3 to 14.7
(10 to 25) (25 to 50)
4,950
(16,900)
19,280
(65,800)
11,760
(40,200)
37,270
(127,200)
73,260
(250,100)
2,830
(9,600)
6,580
(22,500)
4,010
(13,700)
9,230
(50,700)
28,280
(96,500)

Totals
13,430
(45,700)
61,140
(208,700)
33,540
(114,500)
111,250 .
(379,600)
219,360
(748,500)
I
in
00
      Reference 4-40.

-------
4.2.1.1   Combustion Modification

       Combustion modification control techniques reduce the formation of NOX emissions by altering
the combustion conditions present in the combustion chamber.  These techniques include modifications
to the fuel and combustion air feed systems, and modifications to the combustion chamber design.

     One of the most extensive sets of data on combustion modification technique performance was
derived in an EPA-sponsored study involving the testing of 65 boilers.  Ten different combustion
modification techniques were implemented resulting in a total of 116 test runs.  The effects of
these techniques on NO  emissions and boiler efficiency are summarized in Figure 4-2 for 73 separate
boiler tests (Reference 4-42).

     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
                          A         I
combination of five different methods.  These methods were:  excess air reduction, burner out of
service, flue gas recirculation, overfire air addition, and reduced air preheat.  In the first three
methods boiler efficiency was generally unimpaired.

     Since the original combustion modification study, the EPA has continued to study all of the
above techniques and several newly developed techniques.  The most promissing of the recently
developed NO  combustion modification techniques is low !M  burners for oil- and gas-fired boilers.
            "                                             X

     Tables 4-24 and 4-25 summarize the results of EPA's studies on the performance of combustion
modification techniques on gas-, oiU, and coal-fired boilers.  The remainder of this section is
devoted to the discussion of combustion modification experience on these boilers, including control
efficiency, operational problems, and applicability.

Gas- and Oil-Fired Boilers
       Combustion modification controls have been most successful in the reduction of NOV emissions
                                                                                        X
from gas- and oil-fired industrial boilers.  In large part, this success can be attributed to the
greater flexibility of fluid fuels with respect to alterations in fuel firing conditions.
                                                4-59

-------
   +75
                 WORST QUADRANT
   +50
 ~ +25
 x
 O
 a
 o
 oc
2
o
tu
o
•4
   -25
   •so
•75
 -1C
Figure  4-2.
                                COMBUSTION MODIFICATION METHOD
                                      O FLUE GAS RECIRC.
                                      O AIR REGISTER ADJ.
                                      A OIL VISCOSITY
                                      O BURNERTUNEUP
                                      V ATOMIZATION PRESSURE —
                                      • ATOMIZATION METHOD
                                      • REDUCED EXCESS AIR
                                      A OVERRRE AIR
                                      • REDUCED AIR PREHEAT
                                      T BURNI-R-OUT-OF-SERVICE
                                     a
                                                       BEST QUADRANT
                                                                          +10
         -50+5
                CHANGE IN EFFICIENCY, percent
Effect  of combustion modification methods  on total nitrogen
oxides  emissions and boiler efficiency  (Reference 4-42.)
                                       4-60

-------
                              TABLE  4-x/l. N0x EMISSION CONTROL TECHNIQUES FOR GAS-FIRED AND OIL-FIRED INDUSTRIAL BOILERS
                         Effectiveness
                        X N0x Reduction)
                       Operational Impact
                                        Applicability
                                   Commercial Availability/R & D Status
Low Excess Air (LEA)
 (Gas)     0-30
 (d.  oil)  0-30
 (r.  oil)  0-20
Increased boiler efficiency
Applicable to all gas and oil
industrial boilers.  Generally
stack 0« can be reduced to 1-2%
for gas, ?% for dist. oil, and
yt> for resid. oil.  Developing
burners will allow lower stack
V
Method well demonstrated and control
equipment  commercially available for
all boiler types.
     Over-Fire Air
     Ports (OFA)
 (Gas)     25-45
 (d.  oil)  20-40
 (r.  oil)  20-50
Slight decrease in boiler  •
efficiency of 0 to 3%.   The
decrease can be mitigated
in part with the combined
use of LEA controls.
Applicable to all gas and oil
industrial boilers.  Generally
70%-90% burner stoichiometries
can be used with proper instal-
lation of secondary air ports.
Best  implemented on  new units.  Not
commercially available for all design
types especially fire tubes.  Retrofit
not feasible for most units, especially
packaged units.
     Burners Out of
     Service (BOOS)
 (Gas)    25-45
 (d.  oil) N/A
 (r.  oil) 10-40
Perhaps slight decrease in
boiler efficiency.   The
decrease can be mitigated
with the combined use of
LEA controls. • May require
derating unless fuel
delivery system is
modified.
Applicable only to multi-burner
boilers.  Best suited to square
burner pattern.
Commercially available.  Retrofit
application only.  Not demonstrated for
distillate oil.
Low-N0x Burners (LNB)
 20-50
May potentially require
larger fire box area.
Retrofit application may
require derating due to
larger flame pattern.
New burners described generally
applicable to all boilers.
More specific information
needed.
Commercially offered for burner  sizes
up  to  150  x 10   Btu/hr.  Only demon-
strated  for a limited  range of boiler
types  at this time.
Flue Gas Recirculation
(FGR)
  Gas)    45-75
  d.  oil) 40-70
  r.  oil) 15-20
Possible flame instability
and fan errosion problems
which can be reduced
with proper engineering.
Applicable to all design types
except gas ring burners.
Commercially available for most boiler
types but best suited for new boilers
because retrofit would result in possi-
bly extensive modifications.
Deduced Air Preheat
(RAP)
 (Gas)  up  to 55%
 (d.  oil)  up to
  453;
 (r.  oil)  UD to
 20%
Depending on
amount of pre-
heat.
Significant loss in boiler
efficiency unless compen-
sated for by use of feed
water economizer.
Applicable to all design types.
 Comraerically  available  but best  suited
 to  new boilers  where  designs  can be
 modified  to include feed water
 economizers.
Hmmonia Injection
 40-70%
Possible implementation
problems:  fouling and
corrosion problems with
high sulfur oils.   Close
operator attention required.
Appears most applicable to gas
and low sulfur oils.
 Commercially  available  but  very  limited
 demonstration (in Japan only.) Best
 suited to base load boilers.
selective Catalytic
deduction (SCR)
 70-90
No impact on efficiency.
Fouling and corrosion
problems with high sulfur
oils due to ammonium
sulfate salts.   Close
operator attention required.
Appears applicable to all gas-
and oil-fired boilers, although
high -sulfur oils may pose
greater operational problems.
 Commercially  available  but
 limited  demonstration (in  Japan  only.)

-------
TABLE 4-25. NOX EMISSION CONTROL  TECHNIQUES FPR  COAL-FIRED STOKER  INDUSTRIAL BOILERS

Low Excess Air (LEA)
Staged Combustloi Air
(SCA)
Reduced Air Preheat
(RAP)
Aaino.ila Injection
Selective Catalytic
Reduction (SCR)
Effectiveness
(t Nt>x Reduction)
5-25
not well
defined
8
40-60
70-90
Operational Impact
Increased boiler efficiency.
Close operator attention
required to prevent grate
overheating and clinker
formation.
Possible SOM; decrease 1n
efficiency. Close operator
attention required to prevent
grate overheat and cUnker
formation.
Significant loss in boiler
efficiency unless compen-
sated for by use of feed-
Mater ecoiunizers.
Possible i^jlementaMon
difficulties, fouling
problems with high sulfur
fuels, ioac restrictions.
Close operator attention
required.
No Impact on efficiency.
Fouling and corrosion
problems with high sulfur
coals due to iinnonlum sulfate
salts and fly ssh. Close
operator attention required.
Applicability
Applicable to all stokers.
Generally stack 0. can be
reduced to 4-61. L
Most stokers already equipped
with OFA ports.
Applicable to boilers with
combustion air preheat ers.
Appears most applicable to
low sulfur fuels.
Appears applicable to coal-
fired boilers.
Commercial Avallablllty/R 1 0 Status
Commercially available and performance
well demonstrated. Minimum 0, level
•ay not > Identified for some boiler
types.
Commercially available but further
research required to identify optimum
Of A port position and performance
levels.
Commercially available but not signifi-
cantly effective. Best suited for new
boilers where design can be modified to
Include feedwater economizers.
CoMerclally available but not demon-
strated on stoker boilers.
Commercially available, but very
limited demonstration in Japan only.
Significant questions persist about
operational problems on coal -ft red
boilers.

-------
     Low Excess Air - Low excess air (LEA) operation has proven to be ex'i-emely effective 1n
lowering NO  emissions from gas and distillate oil-fired Industrial boilers.  As discussed in
Chapter 3, LEA controls are most effective 1n reducing thermal-NO  emissions, which compose the
major fraction of the NO  emissions from these two fuels.
       An EPA study of 213 gas-fired, 60 distillate-fired, and 148 residual-fired boilers revealed
      DX emissions from these sources could t
the following correlations (Reference 4-431:
that NO  emissions from these sources could be related to the quantity of excess air present using
       Eg  -  0.079 «0.20T1.!1,0.17
       EdQ  -  0.32 H°-46 T°'31 A0'29
       Ero  .  24.2 T°'34 A0'24 0.055 N1'06
 ,1?  E   »  total NO  emissions for gas cnnbustlon adjusted to 3 percent
     9     02 and dry basis (ppm)
    E.   »  total NO  emissions for distillate oil combustion adjusted
            to 3 percent 02 and dry basis (ppm)
    E    *  total NOV emissions from residual oil combustion adjusted
            to 3 percent 02 and dr/ basis (ppm)
    H  *  combustion zone heat release rate (10" Btu/hr - ft")
    T  *  combustion air preheat temperatura °R
    A  «  flue gas oxygen concentration (mole fraction)
    N  *  fuel nitrogen content (Ib N/10" Btu)
 These correlations are presented in Figure 4-3 as a plot of flue gaj 02-vs-NOx for natural gas,
 distillate oil, and a 0.3 wt  percent nitrogen residual fuel oil.
        Based on  these  correlations,  reducing excess air from a  typical  flue gas 02  level of
 5  percent  down to  1  percent  1n  a  gas-fired boiler will result in  a  NO   emission reduction of
 24 percent.   Correspondingly, a reduction 1n excess air from a  flue gas  0- level of 5 percent down
 to 2 percent in  a  distillate oil-fired  boiler will reduce NO  emissions  by 23  percent.  Residual
 oil-fired  boilers  with their characteristically  high  fuel NO  emissions  exhibit a much  lower
 reduction  with LEA controls. Reducing  excess air from a typical  flue  gas 0,, level  of 6 percent  down
 to 3 percent 1n  a  residual  oil-fired boiler combusting a 0.3 wt percent, nitrogen oil will result in
 a  N0x emission reduction of only 6  percent.  Table 4-26 presents  both  the typical and the minimum
 flue gas 02 levels applicable to gas- and oil-fired boilers.
                                                 4-63

-------
    0.4
    0.3
    0.1
                                                           RESIDUAL OIL
                                                           0.3 wt * N
                                                           No A1r Preheat
                                                           DISTILLATE OIL
                                                           No Air Preheat
                                                           Full Load
                                                           NATURAL GAS
                                                           No Air Preheat
                                                           Full Load
                          Flue Gas Oxygen (;)

Figure  4-3.   Effects of flue  gas  oxygen  on the  formation of NO
               emissions  from gas-  and  oil-fired  boilers  (Reference 4-43)
                                       4-54

-------
                     TABLE 4-26.  SAFE OPERATING LEVELS  FOR LEA  (Reference 4-45)
MINIMUM FLUE GAS 0,
FUEL/FIRING TYPE (percent) z
Natural Gas 0.5-3
Distillate and 2-4
Residual 011
TYPICAL FLUE GAS 0«
(percent)
4 - 8
4 - 8
       EPA studies of LEA controls concluded that these controls can be applied  to  some  degree  on
all gas- and oil-fined boilers.   However, the lowest excess air levels can be  achieved with  newer
LEA burners which Incorporate design feature*     •'emit complete fuel  combustion at very low air
levels.  These studies also recommend the use of v.   gen trim systems as an Integral  part of  all  LEA
control system to maintain a minimum but safe excess air level.  If air levels are  allowed to drop
too low. Incomplete combustion can occur, resulting 1n Increased emissions of  hydrocarbons,  carbon
monoxide, and smoke (Preference 4-44).

     Staged Combustion A1r (SCA) - These controls reduce NO  emissions by selectively staging tha
Introduction of combustion air Into the combustion zone.  On gas- and oil-fired  boilers  staged
combustion air  (SCA) controls can be employed by two techniques, burners-out-of-service  vBOOS)  and
overfire air ports (OFA).  These techniques are described in Chapter 3.

       EPA conducted a study of five watsrtube gas-fired boilers, four of which  were equipped with
combustion air  preheat.  Without SJA controls the NO  emissions from these boilers  ranged from
103 to 142 ng/J (24 to 33 lb/106 3tu) and averaged 120 ng/J (28 lb/106 Btu).  With  the application
of SCA controls,  the NOX emissions from these boilers ranged from 60 to 129 ng/J (14 to
30 lb/10  Btu)  and averaged 82 ng/J  (19 lb/10  Btu). These data demonstrate an average emission
reduction of 33 percent  (Reference 4-46).
       An EPA study of a firetube boiler combusting natural gas documented a 25 percent reduction in
 NOX  emissions using SCA controls.  This reduction was achieved with a burner air stoichicmtetry of
 90 percet.t  ano  a  flue gas oxygen concentrecion of 2.9 percent (Reference 4-47).
                                                 4-65

-------
       SCA controls demonstrate their greatest NO  emission reduction efficiencies on residual
                                                 X
oil-fired boilers because of their high effectiveness on fuel-NO  emissions.   EPA studies  on  two
packaged watertube boilers firing residual oil documented NOX emission reductions of 40-45 percent
using OFA ports to achieve SCA controls (Reference 4-48).  Numerous tests on  field-erected watertube
boilers firing residual oil documented NO  emission reductions of 25-40 percent (References 4-42  and
4-49).  This latter set of boilers used BOOS techniques to achieve SCA controls.
       Very little data is available on the performance of SCA controls on distillate oil-fired
boilers, or on residual oil-fired firetube boilers.

       Operational Impacts of SCA controls include possible flame stability problems  and boiler
derating for retrofit applications.  Flame stability problems can be mitigated by the proper
location of OFA ports and by the use of commercially available air flow controls which maintain  the
required staged air injection and burner combustion air flowrates throughout the boiler load range
(Reference 4-50).

       With retrofit applications where BOOS controls are being applied, some boiler  derating may
occur due to the size limit of the burners which remain in service.   This problem can generally  be
eliminated by the installation of larger burners and by modification of the air registers to allow
greater air supply to the burners remaining in service (Reference 4-44).
     Flue Gas Recirculation - A third technique for NOX control  by combustion modification is  flue
gas redrculation (FGR).  This technique involves extracting a portion of the flue gas and returning
it to the furnace through the burner windbox.  FGR suppresses NO  formation by diluting the 0, level
                                                                X                           £*
in the combustion zone and by reducing peak flame temperatures with the heat absorptive capacity of
the recirculated gas.  FGR primarily reduces-thermal NO  and is, consequently, most effective  when
applied to gas- and distillate-fired boilers.

       EPA sponsored FGR tests on two gas-fired watertube industrial  boilers.  On a 73 MW (250 x 10
Btu/hr) boiler NO  emissions were reduced an average of 70 percent at a flue gas recirculation rate
of 45 percent (Reference 4-47).  The second test was conducted on a 5 MW (17 x 10  Btu/hr) boiler
and achieved an average 75 percent NO  emission reduction with a 20 percent flue gas recirculation .
rate (Reference 4-48).
                                                4-66

-------
       F6R controls demonstrated similar NO  emission reduction capabilities on two long-term tests
conducted by EPA, reaching a maximum control  efficiency of 70 percent (Reference 4-48).

       The results of a multi-fuel test on a 5.1 MW (17.5 x 10  Btu/hr) packaged watertube boiler
are presented in Figure 4-4.  This series of tests demonstrates both the relative performance of F6R
controls on various fuels as well as the impact of varying the flue gas recirculaton rate.  The
greater impact of FGR on NOV emissions from gas and distillate oil combustion is clearly
                           A
demonstrated in these results.  More specifically, these test data showed that when firing natural
gas, FGR reduced NO  emissions from 28 ng/J (0.07 lb/10  Btu) at normal operating conditions to an
                   X
average of 13 ng/J (0.03 lb/10  Btu).  These data for gas combustion represent a NO  emission
reduction of 53 percent.  When the same boiler was tested combusting a 0.14 percent nitrogen
residual fuel oil, FGR only reduced NO  emissions from a LEA emission level of 95 ng/J
(0.22 lb/106 Btu) to a FGR emission level of 78 ng/J (0.18 lb/106 Btu).  (References 4-42 and 4-46).
       Ther.e are two operational problems associated with the application of FGR controls; flame
stability problems and errosion of the recirculation fan blades..  Flame stability problems will
necessitate the use of different burner configurations and the use of flame sensors to detect the
onset of instability problems.  Proper fan design will reduce fan'problems.  It has also been
observed that FGR controls may not be easily retrofitted to small industrial boilers due to space
limitations on the extra ducting requirements (Reference 4-44).

     Reduced Air Preheat - By reducing the amount of preheating applied to combustion air, a
reduction of the peak flame temperature can be achieved in the combustion zone.  This reduction  in
turn lowers thermal NO  production.  Most industrial watertube boilers with design heat input
capacities greater than 15 MW (50 x 10  Btu/hr) recover some flue gas heat in combustion air
preheaters or feedwater economizers to maximize thermal efficiency (Reference 4-42).  The
installation of an economizer to replace or reduce the use of a combustion air preheater will result
in lower peak temperatures while still allowing for effective flue gas heat recovery
(Reference 4-51).  Lowering peak temperature is primarily effective for reducing thermal NO , but
has little effect on fuel NO .  Hence, the technique of reduced combustion air preheat will result
in higher percent reductions for low nitrogen fuels — distillate oil and natural gas
(Reference 4-52).                   .
                                               .4-67

-------
  100
  80
 
-------
       The impact of combustion air temperature on a typical  gas-, distillate oil-, and residual
oil-fired boiler is presented in Figure 4-5.  These correlations were developed from the study of
over 400 short-term tests on industrial watertube boilers.  The results of this-study conclude that
a"222°C (400°F) reduction in combustion air preheat would reduce NO  emissions from natural  gas
combustion by 44 percent, from distillate oil combustion by 33 percent, and from residual  oil
combustion by 7 percent.  Within the data base used in this study, combustion air preheat
temperatures typically ranged from 149°C (300°F) up to 360°C (680°F) (Reference 4-43).

       Although RAP is applicable to all gas- and oil-fired boiler types it may'pose a significant
energy penalty for certain applications.  New boiler applications can generally recover waste flue
gas heat by use of feed water economizers.  However, situations do exist where feed water
temperatures are returned to the boiler at temperatures too high to make the use of feed water
economizers practical.  In many retrofit applications, feed water economizers are not already in
place, and the retrofit of feed water economizers is not practical for most industrial boilers
(Reference 4-43).

     Low NO  Burners (LNB) - New burner designs are being developed for industrial boilers which
alter the mixing of air, fuel, and combustion products within the burner flame zone to reduce NO
emissions.  At this time LNB controls have seen very limited application, and are commercially
available for only a limited range of industrial boiler types.
       In a 30-day test of a 30 MW gas-fired boiler, LNB controls reduced NO  emissions from
113 ng/J (0.26 lb/106 Btu) down to 33 ng/J (0.08 lb/106 Btu) under full load conditions.  When
tested at partial load conditions, LNB controls reduced NO  emissions from 95 ng/J (0.22 lb/10  Btu)
down to 44 ng/J (0.10 lb/10  Btu).  These results demonstrate a 70 percent NOV emission reduction at
                                                                             X
full load and a 54 percent NOX emission reduction at partial loads (Reference 4-53).

       Preliminary vendor test results indicate that in typical boiler applications LNB controls
will achieve a 30-50 percent reduction in NOV emissions from distillate oil- and gas-fired boilers
                                            A
and a 17-23 percent reduction in NOX emissions from heavy residual oil combustion (.3 wt % N)
(Reference 4-54).
                                                4-69

-------
        0.4
        0.3
   3
   ±J
   ca
                                                               RESIDUAL OIL
                                                               0.3 wt % N
                                                                 3% 00
  1
        0.2
        0.1
                                                               NATURAL GAS
                                                 Full Load

                                                 DISTILLATE OIL
                                                   2% 02

                                                 Full Load
                           600              800
                              Air Preheat Temnerature  ("P.)
                                             1000
Figure  4-5.
Effects of air preheat temperature on  NO  emissions  from
gas-  and oil-fired boilers  (Reference  4-43).
                                       4-70

-------
       Operational problems with LNB controls have not yet been well  identified.   However,
preliminary information indicates that some LNB controls increase either the length and/or the width
of the burner flame.  In retrofit applications where adequate firebox space is not available some
degree of derating may be required (Reference 4-55).

Coal-Fired Boilers
       The baseline NO  emissions from coal-fired boilers are generally higher than those from gas
and oil-fired units.  Emissions range from 100 to.550 ng/J (0.23 to 1.3 1b/10  Btu).  Although the
fuel nitrogen contents of the test coals are high, ranging from 0.8 to 1.5 percent by weight, field
studies indicate no strong dependence of NOX emissions upon fuel nitrogen content.  Other factors
are apparently more important in,determining NOX production, such as furnace geometry, excess air,
firing rate, burner type, and other fuel properties (Reference 4-56).

       Coal-fired boilers are generally classified by their coal feeding mechanism:  pulverized
coal, spreader stoker, and mass fe'd stoker.  Studies have found that pulverized coal-fired boilers
produce the highest uncontrolled NOX emissions -- approximately 328 ng/J (0.76 lb/10  Btu).  In
these units finely pulverized coal is blown into the boiler through burners.  The pulverized coal
burns relatively rapidly in suspension, resulting in a very high combustion intensity.

       Spreader stoker boilers, in which the coal is thrown onto the grate from above, exhibited
intermediate NO  emission rates ~ approximately 274 ng/J (0-.63 lb/10  Btu) for uncontrolled
full-load firing.  In these units some of the fuel is burned in suspension with air supplied by over
fire air ports, and the remainder is combusted on the grate with under fire air.  The resulting
combustion is therefore partially staged.  Due to design characteristics and the partial  staging
characteristics, the combustion intensities of stoker boilers are less than of pulverized coal-fired
boilers, contributing to a decrease in NO  emissions.

       Mass fed stokers had the lowest emissions ~ approximately 145 ng/J (0.34 lb/10  Btu).  In
these units the combustion air fed up through the grating is insufficient for complete oxidation, so
additional air is introduced above the grating through over fire air ports.  Combustion is,
therefore, effectively staged, and the NOV emissions are quite low.
                                         A
                                               4-71

-------
       The following discussion of combustion modification controls for coal-fired boilers  will
focus primarily on spreader stoker boilers.  Pulverized coal-fired boilers are predominantly
constructed in sizes greater than 73 MW (250 x 10  Btu/hr).  Large industrial  boilers  that  are
greater than this size are very similar to utility boilers which are discussed in Section 4.1.   Mass
fed stokers will generally not be discussed because there is very little data  available on  the
control of NO  emissions from these sources.  This lack of data can be attributed to the smaller
size of these units ^generally less than 29 MW (100 x 10  Btu/hr)] and the naturally lower  NOX
emissions from this source (Reference 4-57).
     Low Excess Air - Low excess air (LEA) controls are the most effective of the combustion
modification techniques for the control of NO  emissions from stoker-fired boilers.   At full load,
stoker boilers typically exhibit a flue gas oxygen concentration .of 6 to 11 percent.   Studies
conducted by the American Boiler Manufacturers Association have demonstrated that excess air rates
for spreader stoker boilers can be safely reduced to a flue gas oxygen concentration  of 4 to
6 percent at full load conditions (Reference 4-56).  At full load conditions, the heat release  rate
is the highest and consequently, so is the potential for NO  emissions.
               %
       The EPA conducted a series of tests on 17 spreader stoker boilers to determine the
performance characteristics of LEA controls.   Under normal operating conditions, these stokers
exhibited a flue gas oxygen concentration of 9 percent.  When LEA controls were applied the  average
flue gas oxygen level was reduced to 6.4 percent with an accompanying 26 percent reduction in NO
emissions (References 4-42, 4-46, 4-49, 4-58,'4-59).
       A more indepth study was conducted on four of these boilers to determine the specific
      onship between excess air levels and NO  emissions.
                                             X
in Figure 4-6 for full load conditions (Reference 4-60).
relationship between excess air levels and NO  emissions.   The results  of  this  study  are  presented
                                             X
       The EPA also conducted 30-day studies on two spreader stoker boilers.   On  a  36  MW (125  x
10  Btu/hr) boiler, NOV emissions under LEA controls averaged 170 ng/J (0.4 lb/106  Btu)
                      /\
(Reference 4-61).  For a 55 MW (190 x 106 Btu/hr) boiler the NOV emissions under  LEA controls
                                                               A
averaged 208 ng/J (0.48 lb/106 Btu) (Reference 4-62).
                                                4-72

-------
    0.3
 CM

O
   0.7
    0.6
    0.5
S3
o •
,!   0.4
 X

o
    0.3
    0.2
20        40
                                        60         80



                                    Percent Excess Air
                                                            11-1
100        120
 Figure 4-6.   Effects of excess air on NO   emissions from stoker

               coal-fired boilers.   (Reference 4-60).
                                     4-73

-------
       The American Boiler Manufacturers Association (ABMA) has also recently completed a  series  of
tests on stoker boilers.  The ABMA study found that for five out of six spreader stokers studied,
the HO  emissions decreased from 9.0 to 15.5 ng/J (0.021 to 0.036 lb/106 Btu) for each  10  percent
decrease in excess air (approximately 1 percent decrease in flue gas 02).  The sixth spreader stoker
boiler exhibited a 28.8 ng/0 (0.067 lb/10  Btu) decrease in NO  for the same decrease in excess air.
For these spreader stokers, LEA controls achieved an average NO  emission reduction of  24  percent
over high excess air operation at full load (Reference 4-56).
       The ABMA results from testing several underfed stokers exhibited NO  emission reductions  of
6.9 to 11.6 ng/J (0.016 to 0.027 lb/10  Btu) for each 10 percent reduction in excess air
(approximately equivalent to 1 percent reduction in flue gas Og).   The average NO  reduction
achieved by LEA controls on these boilers was 20 percent over high excess air operation
(Reference 4-56).
       LEA controls are applicable to all types of stoker boilers.   The major potential  problem with
LEA controls is insufficient combustion air.  If adequate combustion air is not supplied to the
grate, there is an increase in carbon monoxide, VOC, and unburned carbon emissions,  in addition to
the formation of clinkers on the grate.  Oxygen trim systems are commercially available which
monitor 0« and/or CO concentrations in the stack and adjust combustion air flow appropriately to
insure good combustion (Reference 4-45).

     Staged Combustion Air - Staged combustion air (SCA) has not proved to be as effective in
reducing NO  emissions from stoker boilers as it was for other types of boilers.  This
ineffectiveness has been attributed to the observation that stoker boilers generally achieve some
degree of staged combustion by their inherent design.  Fuel is burned relatively slowly on a grate
supplied by overfire air (OFA) and undergrate air.  Using OFA tends to reduce undergrate air,
creating a locally oxygen difficient zone at the fuel bed (Reference 4-44).

       The ABMA conducted a study of the impact of further reducing undergrate air and compensating
with an increase in OFA on 11 stoker boilers.  The conclusion from this study was that SCA controls
had an insignificant impact on NOX emissions from both spreader- and mass-fed stoker boilers
(Reference 4-56).
                                                 4-74

-------
       Similar results were also obtained from a SCA study of a 41 MW (140 x 10  Btu/hr)  spreader
stoker boiler,  a 22 MW (75 x 106 Btu/hr) mass-fed stoker boiler, and a 85 MW (290 x 106  Btu/hr)
mass-fed boiler (Reference 4-46).
       The major operational problem associated with the implementation of SCA controls is
insufficient undergrate air.  This problem results in increased emissions of VOC, carbon monoxide
and particulate emissions, in addition to grate slagging and corrosion problems.   These operational
problems can be avoided by the use of commercially .available oxygen trim systems  which monitor stack
02 and CO concentrations, and make combustion air adjustments appropriately (Reference 4-45).

     Flue Gas Recirculation - The partial recirculation of flue gas (F6R) to the  combustion chamber
has not been demonstrated on stoker boilers.  However F6R has been tested on other high nitrogen
fuels such as residual oil-fired boilers and pulverized coal-fired boilers.  The  results of these
tests have shown that recirculation rates of up to 15 percent decreased NO  emissions by only
17 percent, whereas similar recirculation rates decreased NOX by as much as 50 percent for gas- and
distillate oil-fired boilers (Reference 4-47, 4-63,  4-64).

     Reduced Air Preheat - The,technique of reduced  air preheat (RAP) attempts to reduce NO
emission formation by reducing the temperature of preheated combustion air.  This technique has been
studied on only a very limited scale on stoker boilers.  A comparison of the typical  emissions from
six spreader stoker boilers equipped both with and without combustion air preheaters  showed no
significant difference in NOX emissions which can be associated with combustion air preheat
(Reference 4-60).  Some researchers claim that the coal bed preheats the combustion air before the
combustion occurs and thus defeats the purpose of the method.

       This technique is of course limited to stokers equipped with combustion air preheaters.  Only
larger stokers, greater than 29 MW (100 x 10  Btu/hr) tend to have air preheaters.  In addition,
significant losses in boiler efficiency will occur if flue gas temperatures leaving the stack  are
increased as a concequence of bypassing the preheater.  Economizers can be added  to avoid these
efficiency losses.
                                                4-75

-------
4.2.1.2  Post Combustion Techniques

       Post combustion techniques promise to be more effective than combustion modification
techniques for control of NOX emissions from industrial boilers.  However, the demonstration of post
combustion techniques currently lags a significant distance behind combustion modification.

       The two most developed post combustion techniques are ammonia injection and selective
catalytic reduction.  Both of these techniques are based on the reaction of ammonia with nitrogen
oxide to form elemental nitrogen and water.  These two techniques will be discussed in this section.
Other post combustion techniques such as wet scrubbing and electron beam irradiation are still  in
the research and development stage, and will not be discussed in this section.  Accurate information
is not yet available on these emerging technologies with respect to their application, performance,
cost, and operating problems for industrial boilers.

Awnonia Injection
       Ammonia injection is the best demonstrated of the post combustion NO  control techniques.   In
                                                                           X
this process, ammonia (NHg) is injected into the flue gas downstream of the firebox where it reacts
in a gas phase reaction with NO to produce N2 and H20.  One advantage of ammonia injection is that
it can be used alone, or else in conjunction with combustion modification to achieve an additive  NO
control effect.

       In Japan, ammonia injection was applied-on four industrial boilers; one oil-fired an.d three
gas-fired ranging in size from 16 to 79 MW (55 - 270 x 106 Btu/hr) heat input.  The Japanese tests
demonstrated a 40 to 65 percent reduction in NOX emissions.  The most important variable in
determining performance was the flue gas temperature at the point of ammonia injection
(Reference 4-65).
       In another test conducted on an oil field steam generator in California,  ammonia injection
       d a 50 to 1
(Reference 4-66).
achieved a 50 to 70 percent reduction in NOX emissions.   This unit was burning a heavy crude oil
       The performance of ammonia injection has not been tested on firetube boilers, residual
oil-fired boilers or stoker boilers.  However the performance of ammonia injection is primarily
dependent on the flue gas temperature at the point of injection and is relatively independent  of
fuels and combustion conditions.  Therefore, the performance of this technique should be very
similar for all common types of fossil fuel-fired boilers.
                                                4-76

-------
       There are two operational problems of concern with the application of ammonia injection;
maintaining the optimum injection location and ammonium sulfate deposits.  As the boiler load
fluctuates, there is a corresponding fluctuation in the temperature profile in the flue gas ducts.
Because of the sensitivity of the ammonia - no reaction to temperature - load fluctuations also
result in a fluctuation of the optimum injection location.  For this reason ammonia injection is
best suited for constant load boilers.  Alternatively, ammonia injection systems may be accompanied
with multiple injection points and as'sociated controls so that ammonia can be injected at the proper
point for the corresponding boiler load (Reference 4-66).

       Ammonium sulfate problems are associated with burning high sulfur fuels.  Unreacted ammonia
reacts with sulfur oxide in the flue gas to form ammonium sulfate salts.  These salts create
plugging and corrosion problems for down stream preheaters and boiler parts.  Ammonium sulfate  ;
problems are best mitigated by applying ammonia injection only to very low sulfur fuels such as
natural gas and clean fuel oils and by the use of ammonia analyzers/controllers to detect and
control the presence of excess ammonia  (Reference 4-66).

     Selective Catalytic Reduction - Selective catalytic reduction (SCR) is a technique involving
removal of the flue gas NO  by reacting the NO  with ammonia in a catalytic reactor to form
                          A                   X
elemental nitrogen.  With the exception of the use of a catalyst, it is similar to the ammonia
injection NO  control technique just discussed.

       Although not demonstrated in the U.S., SCR is thought to be applicable to all types of fossil
fuel-fired industrial boilers.  Greater'than 90 percent NOV reduction is achieved at ammonia to NO
                                                          X                                      X
ratios of 1:1 on commercial systems applied to industrial boilers in Japan.  These systems have been
applied to a variety of gas- and oil-fired boilers in Japan, and appear to be viable, techniques of
attaining up to 90 percent NOV control on all industrial boilers (Reference 4-67).
                             A

       Two operational problems associated with SCR are the formation of ammonium sulfate salts «and
the loss of unreacted ammonia.  The ammonium sulfate salts pose plugging and corrosion problems for
down stream equipment and the loss of unreacted 'ammonia poses an emission problem.  Both of these
problems are the focus of ongoing EPA and Electric Power Research Institute studies
(Reference 4-67).
                                                 4-77

-------
4.2.2  Cost Iiapact

       NOX emission control costs are a very Important concern to the owners and operators of
Industrial boilers.  These costs can vary greatly with control technique, degree of control, and
boiler size and type.  This section discusses the capital and operating cost of various NOX control
technologies, and projects the Impact of these controls on the cost of producing steam.

       The discussion on combustion modification and ammonia Injection controls was summarized from
a single  study conducted for the Industrial Environmental Research Laboratory (IERL) of EPA 1n July
1981  (Reference 4-54). This study provides a comprehensive analysis of NO, control costs an a common
cost  basis.  The reader 1s referred to this report for detailed Information on the assumptions and
basis applied 1n developing these costs.

       The discussion on SCR costs was summarized from a study conducted for the IERL on N()x flue
gas  treatment technologies, completed December 1979.  The reader 1s referred to this study,
Reference 4-67, for further Information  on the assumptions and basis applied In developing these
costs.

       Section 4.2.2.1 discusses the costs associated with the application of combust'on
modification  controls for  NOX  emissions.  And Section 4.2.2.2 discusses  the costs associated with
the  application of post combustion controls  for  NOX emissions.

4.2.2.1   Combustion Modification Techniques
 Basis
        Two cost components are studied 
-------
       Detailed Information on the cost basis  applied to jenerate the  following  combustion  modifi-
cation costs 1s presented 1n Reference 4-54.   In genera1 the equipment,  labor, fuel,  and  utility
costs are based on 1978 prices and are presented 1n 19'8 dollars.  A load factor of 45  percent  was
assumed for boilers smaller than 7000 kg steam/hr (15,000 Ib/hr)  and a load factor of 60  percent  was
assumed for all boilers larger than that size.  A capital recovery factor of 0.2 was  applied to
convert capital costs to annual 1zed capital charge'*.

       In Che following analysis, the costs to irodlfy ths boiler are given as a  percentage  of the
boilers Installed cost.  The boiler's Installed cost does not Include costs for auxllllary support
equipment such as water treatment and fuel hardllng.  However, the Impacts on steam costs do
consider the to*ii cost to produce steam, 1nOud1ng the costs for auxiliaries.

       Information on retrofit costs are not available  in the same detail.  However the costs for
applying combustion modification controls to existing boilers are estimated to be twice the cost of
appHc tlon to new boilers (Reference 4-68).
 Results
        The  estiiwted capital and operating costs associated with applying LEA, SCA, and FGR controls
 on  industrial boilers are presentee1 1n Tables 4-27, 4-28, and 4-29 respectively.

        For  the  reduced  air  preheat  (RAP) control option, the cost: are negligible for new boilers
 when a feed water economizer is substituted for the combustion fir preheater.  Feed water econo-
 mizers are  estimated to recover the equivalent amount of waste heat for the equivalent cost of a
 combustion  air  preheater.   The retrofit of RAP controls on an existing boiler where c; feedwater
 economizer  is not feasible  can sasily result  in a boiler efficiency loss of up to 3 percent and a
 stea.n price increase of as  much as  2 percent  (Reference 4-69, 4-70, 4-71).  Efficiency drops of
 twice this  amount could result from significant preheat reductions.

        Since low NOX burners are  still  in  the developmental  stage, costs ''re noc yet firmly
 established.  For new  boilers, costs would be effected by  the incremental cost difference between a
 low NOX burner  and a  standard  burner plus  the cost  of an oxygen  trim  system, if needed.  Windbox
 modifications may also be needed.   Finally,  1t  1s expected that  LNB will allow LEA operation which
 would keep the  economic impact of LNB  at  a minimum,  thus,  it seems likely that LNB operation wll1
 cost no more than SCA  operation  and,  in  fact, may have an  even  smaller effect on steam costs than
 SCA (Reference 4-72 and 4-73).
                                                 4-79

-------
                                           I ABIE 4-27.   ESTIMATED COST OF LOU EXCESS AIR OPERATION  FCft NEW BOILERS (1976 Dollars)
00
o
AHNUALIZED COSTS
TYPE
Pulverized Coa1
Spreader Stoxer
Chain Grate
Stoker
Underfeed
Stoker
Residual Oil
WatertL-be
Residual Oil
Flretube
Distillate Oil
Uatertube
Distillate Oil
Flretube
Natural Gas
Vatertube
Natural Gas
Fire lube
»EAT,IHPUT
MU (10° Btu/hr)
59 (200)
44 (IbO)
22 (75)
9 (30)
4« (ISO)
4.4 (15)
29 (100)
4.4 (15)
29 (?00)
4.4 (Ib)
PERCENT NO
REDUCTION"
15
15
15
15
20
15
10
10
5
5
CAPITAL
COST
1000 $
27
22
17
14
17
9
14
9
14
9
HXEO
, Kills/
10J kg/steam
1J 16)
16 (7)
26 (12)
54 (24)
12 (5)
91 (41)
16 (7)
91 (41}
16 (7)
91 (41)
OPERATING*
, Bills/
10J kg steam
-28 (-13)5
-31 (-14)
-28 (-13)
-23 (-10)
-75 (-34)
-59 (-27)
-41 (-19)
-26 (-12)
-23 (-11)
-11 (-5)
TOTAL
, Bills/
10J kg/stea*
-15 (-7)
-15 (-7)
-2 (-1)
21 (M)
-63 (-29)
32 (14)
-25 (-11)
65 (29?
-9 (-4)
80 (36)
PERCENT CHANGE IN COST OF
BOILER
0.4
0.5
1.2
1.5
1.9
9.0
3.4
9.0
4.0
9.0
STEAM
-D.I
-0.1
0
0.1
-0.5
0.1
-0.2
0.2
-0.2
0.3
          dNuirt>ers In parentheses are  in units of rail Is/103  Ib stream

          ''Negative values represent cost savings

          cReference 4-54.

-------
TABLE 4-?K.   ESTIMATED COST OF STM&D COMBUSTION OPERATION FOR NEW BOILERS (1978 Dollars)
ANNUAL I ZED COSTS1
TYPE HEAT,INPUT PERCENT MO CAPITAL FIXED OPERATING
MU (10b Btu/hr) REDUCTION" CUM









Pulverized loai M>
Spreader Stoker 44
Chain Grate 22
Stoker
Underfeed 9
Stoker
Residual Oil 44
Uatertube
Residual Oil 4.
Flretube
Distillate Oil 29
Uatertube
Natural Gas 29
Uatertube
1000 t
, Bills/ , Bills/
Mr kg/steaa 10J kg stCM
(200) 25 47 . 24 (11) 
-------
                                     TABLE 4-29.  ESTIMATED COST OF aUE GAS RECIRCUtATlON OPERATION FOR HEM BOILERS (1978 Dollars)
oo
PO
AHNUALIZED COSTS3
TYPE
Distillate Oil
Water-tube
Distillate Oil
Flretube
Natural Gas
Hatertube
Natural Gas
Flretube
HEAT,INPUT PERCENI NOV
HW (10b Btu/hr) REDUCTION55
zy (i(10) 40
4.4 (15) 40
29 (1UO) 40
4.4 (15) 40
CAPITAL
COST
1000 $
26
19
26
19
FIXED
, mills/
10J kg/steam
29 (13)
192 (87)
29 (13)
192 (87)
OPERA 1 ING
, mills/
10 kg steam
135 (blj
175 (80)
120 (54)
159 (72)
TOTAL
, mills/
10J kg/steam
164 (74)
367 (167)
150 (67)
351 (159)
PtRCENT CHANGE IN COS 1 Oh
BOILER STEAM
7.0 1.7
21.0 1.2
7.0 2.4
21.0 1.7
        a   Numbers  In parentheses are in units of mills/10  Ib stream.

        Reference 4-54.

-------
              Based on the limited data available on combustion modification costs for industrial  boilers,
       some tentative conclusions can be made.

              -  Low excess air operation, in many cases, will actually lower steam costs due to the
                 increase in thermal efficiency.  In general, LEA operation is recommended for use with
                 other control techniques to lessen their cost impact and to give higher NO  reductions.
                                                                                           X       i
              -  Staged combustion causes an estimated small increase in steam cost but with careful design
                 and operation this estimated cost increase can probably be reduced.
              -  Flue gas recirculation, though cost'ly, is the most effective for the low nitrogen fuels,
                 distillate oil and natural gas.  Again, optimal design and operation will probably lower
                 the cost.
              -  Low NO  burners hold the promise of being the most cost-effective technique for oil and
                 gas boilers.  However, they are still under development.
              -  Reduced air preheat is recommended for those boilers where an economizer may be installed
                 in place of an air preheater.

       4.2.2.2  Post Combustion Techniques

              This section presents the costs for installing and operating the two most advanced post
       combustion techniques for industrial boiler NOV control:  ammonia injection and selective catalytic .
f                                                    A
       reduction.                                                     •

              Ammonia injection is not a demonstrated control technology for industrial boilers within the
       U.S.  The few sources available on ammonia injection costs are based on large utility boilers.
       Extrapolating down to an industrial-size boiler, the capital cost for installing ammonia injection
       on a 59 MW (200 x 10  Btu/hr) pulverized coal boiler is estimated to be about $236,000.  This cost
       represents about a 4 percent increase over the cost of an uncontrolled boiler.  The incremental
       steam cost is estimated to be 220 mills/10  kg steam (100 mills/10  Ib steam) or a 2.5 percent
       increase in steam price.  Because of the large error in extrapolating from a utility- size  boiler
       down to an industrial-size pulverized coal boiler or to other boiler types, the above price
       estimates have very large uncertainties  (References 4-74, 4-75, 4-76).
                                                       4-83

-------
       Selective catalytic reduction (SCR) Is another post combustion technique which has not been
applied to industrial boilers in the U.S.  The most extensive cost estimates for SCR controls is
available from a detailed engineering cost analysis of proposed plant designs (Reference 4-77).   The
results of this analysis are presented in Tables 4-30, 4-31, and 4-32.  These cost estimates are
based on mid-1978 costs.  Annualized costs are calculated from capital costs using a capital
recovery factor of 0.13, reflecting a 10 percent interest rate and a 15 year recovery period.  The
load factors for coal, residual oil, distillate oil, and natural gas were 60 percent, 55 percent,
45 percent, and 45 percent respectively.  Finally, the most stringent control level  reflects
90 percent NOV control and the moderate control level reflects 70 percent N0y control.
             J\                                                              A

4.2.3  Energy and Environmental Impact

4.2.3.1  Energy Impact

     The energy impacts of applying NO  emission controls are very important to the owner/operator
of industrial boilers.  By far, the major cost of operating an industrial boiler is the fuel cost.
Combustion inefficiencies created by NOV emission control techniques translate directly into higher
                                       A
steam costs.  This section discusses the energy impacts associated with applying NO  emission
control technologies.  These energy impacts are a combination of changes in boiler thermal
       "\
efficiency, and direct energy consumption by the control technology.  The energy impacts presented
here were summarized in Tables 4-24 and 4-25.

Low Excess Air

     Low excess air (LEA) is the simplest of all NOX control techniques, and one which saves fuel.
Virtually all boilers tested show an increase of about 0.5 percent in efficiency for each 1 percent
decrease in flue gas oxygen.  However, in a few isolated cases, LEA did not increase efficiency for
natural-gas-fired units.  The average energy savings was approximately 1 percent of the thermal
Input (Reference 4-42, 4-46, and 4-49).

Staged Combustion Air

     Staged combustion air (SCA) can be achieved by use of BOOS and by use of OFA.  In the limited
number of studies conducted on BOOS controls, very little impact on thermal efficiency has been
observed.  The thermal efficiency has ranged from an 0.5 percent increase to an 0.5 percent decrease
(Reference 4-42, 4-49, and 4-68).
                                                4-84

-------
   TABLE 4-30.   ANNUAL COST OF NO  CONTROL SYSTEMS
    APPLIED TO COAL-FIRED BOILERSX(Reference 4-67)

Boiler
Underfeed Stoker
Chaingrate
Spreader Stoker
Pulverized Coal
Size,
MBtu/hr
30
75
150
200
Annual Cost, $1000/yr
Control System
Parallel Flow SCR
Parallel Flow SCR
Parallel Flow SCR
Parallel Flow SCR
Moderate
Control
108
153
221
254
Stringent
Control
130
197
291
351
   TABLE 4-31.  ANNUAL COST OF NOV CONTROL SYSTEMS
    APPLIED TO OIL-FIRED BOILERS (Reference4-67)

Boiler
Distillate Oil
Distillate Oil
Residual Oil
Residual Oil
Residual Oil
Residual Oil
Size,
MBtu/hr
15
150
30
30
150
150
Annual Cost, $1000/yr
Control System
Fixed Packed Bed SCR
Fixed Packed Bed SCR
Parallel Flow SCR
. Moving Bed SCR
Parallel Flow SCR
Moving Bed SCR
Moderate
Control
64
137
96
120
181
168
Stringent
Control
67
176
108
130
223
204
   TABLE 4-32.  ANNUAL COST OF NOV CONTROL SYSTEMS
APPLIED TO NATURAL GAS-FIRED BOILERS (Reference 4-67)

Boiler
Package,
Package,

Fi retube
Watertube
Size,
MBtu/hr
15
150
Annual Cost, $1000/yr
Control System
Fixed Packed Bed SCR
Fixed Packed Bed SCR
Moderate
Control
64.4
129
Stringent
Control
67.6
' 175
                         4-85

-------
     Tests on OFA controls have shown a wider thermal  efficiency impact,  ranging from a  1  percent
efficiency gain to a 3 percent efficiency loss.   With  OFA controls there  is  an  additional  energy
loss associated with increased fan power requirements.  This additional fan  power is  required  to
overcome the pressure drop in the air ducts leading to the OFA ports,,  and amounts to  less  than
0.1 percent of the thermal input (Reference 4-42 and 4-72).

     Both SCA controls are sensitive to burner stoichiornetry and to location of air injection. The
use of oxygen trim systems will help offset possible boiler efficiency losses associated with  either
SCA control.

F1ue Gas Reelrculati on

     In tests run to date flue gas recirculation (F6R) had only a small effect  on boiler thermal
efficiency.  Thermal efficiency impacts ranged from a  0.5 percent increase to a 1.0 percent
decrease, while most impacts were less than a 0.5 percent decrease (Reference 4-42, 4-48,  4-49,
4-68, 4-77).  According to utility boiler data,  an increase of 0.25 percent  of  boiler heat input
could be required to power the F6R fan (Reference 4-72).
Low HO  Burners

     At present, there is very little data available on the energy impacts  of low NO  burners  (LNB).
LNB are not expected to have a significant impact on thermal efficiency.   In fact some improvement
1n efficiency may be possible due to their expected use of lower excess air (Reference 4-78).
Reduced A1r Preheat

     The use of reduced air preheat (RAP) has the potential  for significantly impacting thermal
efficiency.  Tests have demonstrated thermal efficiency decreases of up to 3 percent with  the  use  of
RAP.  Even greater energy penalties are incurred with the full  reduction of air preheat.   This
efficiency decrease is associated with the loss of valuable  waste heat when the preheater  is
bypassed.  However, on new units where feedwater economizers can be substituted for preheaters in
the boiler design, this thermal efficiency loss can be entirely mitigated (References 4-42, 4-46,  •
and 4-49).
                                                4-86

-------
Amnonfa Injection

     Very little information is available on the energy impacts associated with ammonia injection
controls.  However, since ammonia injection is a post combustion control technology, its primary.
energy impact will be associated with the energy usage of ammonia handling and injection equipment.
For SCR controls, .the ammonia handling and injection equipment consumed steam and electricity
totaling less than 0.1 percent of the thermal energy input  (Reference 4-67).

Selective Catalytic Reduction

     Selective catalytic reduction (SCR) is a post combustion control technology and therefore is
not expected to have an impact on the direct thermal efficiency of the boiler.  However, SCR
controls do consume electricity and steam.  The electrical demand is required to overcome the
pressure drop across the catalytic reactor and to transfer ammonia.  Steam is consumed in vaporizing
and diluting the ammonia.  The combined electrical and steam demand required to operate SCR is
consistently less than Q.64 percent of the boiler thermal input and generally less than 0.3 percent
(Reference 4-67);

4.2.3.2  Environmental Impact

     Very little research has been conducted on the environmental impacts associated with the use of
NOX emission controls on industrial boilers.  Much more extensive research has been conducted on
utility size boilers, the results of which are presented in Section 4.1.3.2.  Because of the
similarities between industrial and utility boilers with respect to emission characteristics, the
reader is referred to the above section for information on the environmental impacts of applying NO
                                                                                                   X
controls to industrial boilers.

      In general, combustion modification controls were found to generate very slight or no increase
in carbon monoxide, VOC, particulates, or trace elements.  However there was a slight increase in
POM  emissions.

   ,  Post combustion techniques including ammonia injection and SCR are expected to increase
emissions of ammonia, amonia salts, and S03.  In some cases, wastewater treatment and disposal may
be complicated  by  the addition of nitrogen compounds.  Finally SCR systems will require waste
catalyst disposal.  The  full extent of these environmental Impacts is not known and is the subject
of ongoing  EPA  research.
                                                4-87

-------
4.3    PRIME HOVERS

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 10s 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 4-79) provides a good overview of emissions from station-
ary engines, particularly the large bore engines used 1n the oil and gas industry and for electric
power generation.  An EPA-sponsored Standards Support and Environmental Impact Study (SSEIS) for
these engines (Reference 4-80)  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
             A
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
NOX control techniques for these engines often involve HC and CO control since these emissions fre-
quently increase as NOX 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-
  X                                                     ,
fueled engines have relatively small HC and CO emissions (less than 4 g/kWh*).
       The following paragraphs will discuss NOX control techniques in general followed by a tabula-
tion of specific HOX 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 4-33 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, 1n 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
                                                  4-88

-------
              Table 4-33.
SUMMARY  OF NOX EMISSION CONTROL TECHNIQUES  FOR
RECIPROCATING INTERNAL COMBUSTION  ENGINES
CONTROL
RETARD
injection (CD*
ignition (SI)b
CHANGE AIH-TO-FUa (A/F)
RATIO
DERATE
INCREASE SPEED
DECREASE INLET HANIFOU
AIR TEMPERATURE
EXHAUST GAS RECIRCUUTION
(€SR)
External
Internal
valve overlap
or rttird
exhaust back
pressure
CHAJQER MODIFICATION
Preeombustlon (CD
Stratified charce (SI)

HATER INDUCTION
CATALYTIC CONVERSION
PRINCIPLE OF REDUCTION
Reduces peak temperature
by delaying start of
combustion during the
combustion stroke.
Peak conbustlon tenpera-
ture 1i reduced by off-
stolchlomtHc operation.
Reduces cylinder pres-
sures and temperatures.
Decreases residence tine
of gases at elevated
temperature and pressure.
Reduces peak temperature. '
Dilution of 1ncon1ng com-
bustion charge with Inert
gases. Reduce excess
oxygen and lower peak
coabustlon temperature.
Cooling by increased
scavaglng, richer
trapped air-to-fuel
ratio.
Richer trapped air-to-
fuel ratio.
Combustion In ante-
chamber permits lean
combustion fn main cham-
ber (cylinder) with less
available oxygen.
Reduces peak combustion
temperature.
Catalytic reduction of
NO to «2.
APPLICATION
An operational adjustment. Delay
can or Injection punp tilting (CI);
delay Ignition spark (SI).
An operational adjustment. In-
crease or decrease to operate on
off-»to1chiometr1c mixture. Reset
throttle or Increase air rate.
An operational adjustment. Units
maximal bmepc (governor setting).
Operational adjustnent or design
change.
Hardware addition to Increase
aftercoiHng or add aftercoollng
(larger heat exchanger, coolant
punp).
Hardware addition; plumbing to shunt
exhaust to Intake; cooling nay be
require! to be effective;' controls
to vary rate with load.
Operational hardware modification:
adjustamt of valve can timing.
Throttling exhaust flow.
Hardwam modification; requires
different cylinder head.
Hardware! addition: Inject water Into
Inlet onnifold or cylinder directly;
effective at water-to-fuel ratio •
1 (kg H;>0/kg fuel).
Hardware) addition: catalytic con-
verter Installed In exhaust plumbing
or reducing agent (e.g. amonla)
Injected Into exhaust stream.
BSFCd
' INCREASE
Yes
Yes
Yes
Yes
No
No If EGR
rates not
excessive
Yes
Yes
Yes
No
HO
COWENTS -- LIMITATIONS
1
•
Particularly effective with moderate amount i
of retard; further retard causes high exhaust
temperature with possible valve damage and
substantial 3SFC 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 8SFC, 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 1s lowered for given power requirements.
Compressor applications constrained by vibra-
tion considerations. Not a feasible tech-
nique for existing and oast new facilities.
Ambient temperatures limit maximum reduction.
Raw water supply may be unavailable.
Substantial fouling of heat exchanger and f1»
passages; anticipate increased maintenance.
Hay cause fouling in turfiocharged, aftercooleal
engine. Substantial Increases in CO and smote
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.
Llnlted for turbocharged engines due to
choking of turboconpressor.
5 to 10 percent Increase in 8SFC over open-
chamber designs. Higher heat loss implies
greater cooling capacity. Major design
development.
Deposit buildup (requiring denlnerallzation);
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
ebnep — brake mean- effective pressure .
dSSFC — brake specific fuel consumption
                                               4-89

-------
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 NOX, 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 4-80, 4-81)..
       Hater 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 4-80).
       Combustion chamber modifications such as precombustion and stratified chambers have demon-
strated large NOV reductions, but also produce substantial fuel consumption increases (5 to 8 per-
                A
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 4-80).
       Table 4-34 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 4-33 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 NO  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.
                                                     4-90

-------
               Table 4-34.   EFFECT  OF  NOX  CONTROLS ON  LARGE-BORE
                            INTERNAL COMBUSTION  ENGINES
                            a.   Normalized  percent  reductions  of NO
Fuel
Number Cylinders

Baseline3
Retard
Alr-to-Fuel
Derate
MAT
Gas
2
BSb
20
2.5
0.19
6.2
0.9
TC
17
3.1
4.5
2.6
1.3
4«
NA
24-29
1.5
1.8
0.25-1.3
— •
TC
17-30
4.1-0.6
3.3
0.34-1.9
0.4-0.9
Dual Fuel
2
TC
12
9.1
1.7
—
1.3
4
TC
10-17
1.5-6-3
2.4-2.5
0.01-0.94
0.6-0.8
Diesel
2
BS
18-26
6.9
—
0.84-0.92
0
TC
14-19
5.3-5.7
—
—
0.2-0.4
4
TC
13-15
2.7-4.4

0.17
0.1-0.3
                            b.  Percent  increase  in  brake  specific
                               fuel  consumption
Retard
A1r-to-Fue1
Derate
MAT
5.2
2.0
2.6
1.3
4.3
1.5
6.1
0.5
3.6
1.0
8.2C
—
1.2
2.3
c
1.1
0
3.4
.2.6
7.0C
0.4
1.0C
1.9
+0.5
—
C
3.4
—
3.3C
«.
1.6
2.2C
9.6
0
 Baseline data in gm/kWh shaft output,  all  other data in percent NOX reduction/unit control.  Unit
 control is 1   retard,  1 percent air flow increase,  1  percent derating,  or 1.8K (l.OF)  air
.temperature decrease.
|!BS - blower scavenged, TC - turbo-charged,  NA - naturally  aspirated.
 Average value

-------
       Numerous investigators have studied control techniques to reduce NOX in diesel-fueled auto-
motive truck applications.  Many of these studies are summarized in Reference 4-81.   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 1b)  emission
limits for diesel-fueled engines.
       Table 4-35 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 NOX 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 MOX 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,
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 4-36 lists Federal and State emission limits, and Table 4-37 lists the various controls that
are used in several combinations by manufactures to meet these limits.  Table 4-38 gives specific
examples of control techniques recently applied to meet Federal light duty vehicle emission limits.

        Based on the preceding discussion, potential NO  emissions reductions for stationary recipro-
 cating engines can be  summarized as follows:.
        •   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.
        t   Exhaust gas recirculation, watar 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
                                                   4-92

-------
Table 4-35.
CONTROL. TECHNIQUES FOR TRUCK SIZE
DIESEL ENGINES [<375 kW (500 HP)]
TO MEET 1975 CALIFORNIA  13.4 G/KWHR  .
(TO G/HP-HR) COMBINED N0₯ AND HC LEVEL'
                        /\
          Control
                    Percent
                 bsfc" Increase
  Retard, modify fuel
  system and turbocharger

  Retard, modify fuel
  system and turbocharger,
  add aftercooler

  Add turbocharger and
  aftercoolerc

  Retard0 (naturally
  aspirated version)

  Precombustion chamber
                       3
                       3
                     5-8
   Based 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.
                       4-93

-------
                   TABLE 4-36.   1975  VEHICLE  EMISSION LIMITS

Passenger Car,
g/Wh (g/m1)a
California
Federal
Light duty truck,
g/kWn (g/ral)
California
Federal
Heavy duty vehicles,
g/kHh
California
Federal
MX

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
 lun/hr requiring 8195 W (11 bhp) for the '/-mode composite cycle.
  TABLE 4-37,   EMISSION CONTROL TECHNIQUES FOR AUTOMOTIVE GASOLINF ENGINES
CoMtrol
NOX:
Rich or lean A/F ratic,
Ignition timing retard
Exhaust gas redrculatlon
(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
de
-------
                              TABLE 4-33.  EMISSION CONTROL SYSTEMS FOR CONVENTIONAL GASOLINE INTERNAL COMBUSTION ENGINES
                                           (ADAPTED FROM Reference 4-81 i
Number
—
0
1
2
3
Year
1972
1973 Federal
1975 Federal
1975 Calif.
Syste»
EM-1
EM* + El + FC * AI + E6R
EM' + El * 1C + QHI + AI + EGR
«• + El + 1C + QKi + EfiR * AI + OC
Fuel Penalty v
_
7*3
5 ± 2
8 t 2
Reduction Factors1*
HCC
1 ± 0.375
1.35 i 0.30
0.65 i 0.15
0.18 ± 0.05
COC
1 t 0.375
1.0 t 0.23
0.55 t 0.1S
0.15 t 0,03
NO/
1 i 20
0.6 t 0.10
0.06 t 0.10
0.06 t 0.10
Syste.
Oc erloratlon
L
L
L
M ;HC. co)
L (NOX)
vo
in
        aH972 baseline engine:  andiflcatlons Included fn tht baseline engine configuration are retard,  lera air-to-fuel, and reduced
         compression ratio.

        Cc^ponent Identification

        EM   -  Engine Modifications; retard, air-to-fuel, compression ratio
        El   -  Electronic ignition
        FC   -  Fast choke
        QHI  -  Quick heat intake
        AI   -  Exhaust manifold air injection
        EGR  -  Exhaust gas reclrculatlon
        1C   -  Improved carbureticn
        OC   -  Oxidizing catalyst.


        Deduction factor defined as:  "S^baseffneHgssloBS  basea on LA~4 dr1vi"9 c>cle
        CA11 enissions dcta taken using or corrected to 1975 CVS-CH test procedure

        Deterioration of present systems; L - 10X.  M = 10 - 301,  '  =  301

-------
        •   NO  control technology for automotive applications can be adapted to stationary engines;
                                                   »
            however, NQX reductions and attendant fuel penalties for automotive applications are
            closely related to the load cycle, which in some cases nay differ from stationary
            applications
        •   Viable control  techniques may involve an operational adjustment, hardware addition, or
            a combination of both
        •   Additional  research is necessary to
            —   Establish controlled levels for gaseous-fueled engines (<75 kW/cylinder, or
                100 hp/cylinder)
            —   Establish controlled levels for roediurn-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 4-82 and information
 supplied by Reference  4-80 (1974 costs).   These costs, however, relate to emission reduction
 achieved by engines tested in laboratories rather than to field installations.  Reference 4-83'
 indicates, nevertheless, that these data are representative (1972 costs).
        In contrast to the large stationary engines, more published cost data exists  for smaller
 (<37S 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
                                                  4-96

-------
load) can differ from stationary-duty  cycle's  (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 4-39 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 4-40 (1974 costs).  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 operating adjustment with the exception of  derating and manifold  air cooling, which would require
hardware additions.   Derating is not  a  viable  technique for  existing installations un'iess 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.
                                                    4-97

-------
    TABLE 4-39.  COST IMPACTS OF NOX CONTROLS FOR LARGE-BORE ENGINES
        Control
            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 4-40.  TYPICAL 1974 BASELINE COSTS FOR
                          LARGE (>75  KW CYLINDER) ENGINES3
Costs
1. Initial,5 $/kW
2. Maintenance,
$/kWh
3. Fuel and lube,
$/kWh
Total Operating,
2 + 3
Gas
174
4 x 10'3
10 x 10-3
14 x 10-3
Dual Fuel
174
4 x 10"3
10 x 10-3
14 x 10-3
Diesel
174
4 x 10-3
23 x 10"3
27 x lO'3
            Based on Reference 4-82 and information supplied to
            Reference 4-80 by manufacturers.
           'includes basic engine and cooling system.
                                     4-98

-------
 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 4-41   lists  typical costs for techniques implemented for 1975 diesel-fueled  truck
 engines  (1974  costs).  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 4-42  gives  control hardware costs to meet gasoline-fueled passenger vehicle emission
 limits through 1976 (1973 costs).  Note that cost increases correspond to increasingly more complex
 controls to meet  more stringent emission limits.
            TABLE 4-41.  TYPICAL CONTROL COSTS FOR DIESEL-FUELED  ENGINES  USED  IN  HEAVY-DUTY
                         VEHICLES (>2700 kg OR 3 tons)
                              engine
                              cooling system
                       turbocharger
                       aftercooler
                       EGR
$40-$67/kW ($30-$50/hp)
8%-14% engine
$4/kW ($3/hp)
6%-10% engine
$3-$4/kW ($2-$3/hp)
              Operating
                   Fuel:          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).
                   Maintenance:   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.
     aBased on information  supplied  to Reference 4-80 by manufacturers (1974 costs).
                                                   4-99

-------
TABLE 4-42.  ESTIMATES OF STICKER PRICES FOR EMISSIONS HARDWARE FROM
             .966 UNCONTROLLED VEHICLES TO 1976 DUAL-CATALYST SYSTEMS
             (Reference 4-81, Costs Taken From A 1973 Report)
Model
Year
1966
1968
1970


t

1971-
1972




1973






Configuration
PCV-Crank Case
Fuel Evaporation
System
Carburetor Air/Fuel Ratio
Compression Ratio
Ignition Timing
Transmission Control
System
Total 1970
Ant1-D1ese11ng
Solenoid
Thermo A1r 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 - 14S)
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
                                     4-100

-------
TABLE 4-42.   ESTIMATES OF STICKER PRICES FOR EMISSIONS HARDWARE FROM 1966 UNCONTROLLED
             VEHICLES 70 1976 DUAL-CATALYST SYSTEMS (Reference 4-81) (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
$2/1b)
Cooling System Changes
Underhood Temperature
Materials
Body Revisions
Welding Presses
Assembly Line Changes
End of Line Test So/No--Go
Quality Emission Test
Total 1975
2 NOX Catalytic Converters*
Electronic Control3
Sensors3
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
                                            4-101

-------
        Figure 4-7 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 (1972 costs). 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 NOX  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 NOX 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 SOX and  trace  ruetals 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)  ratiols,  reducing manifold air
temperatures  (MAT), and water Injection.   The effects of each of these NOX  controls on engine CO
emission levels are summarized in  Table 4-43.
       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-
ctnt, while derating four-cycle engines actually gives a 60 to 100  percent  decrease in CO levels.
                                                 4-102

-------
  1.5
  1.0
  0.5
                                                VARYING EGR AND

                                              SECONDARY AIR RATES
                                                                      3.0
                                                                      2.5
                                                                      2.0 §
                                                                         a
                                                                         LU

                                                                      1.5"

                                                                         cc
                                                                         a.
                                                                         CO
                                                                          X

                                                                      1.0?
                                                                      0.5
                          10                    20

              INCREASE IN BSFC, purcent (OVER UNCONTROLLED VEHICLE)
30
                                GENERAL CORRELATION              	

                                ESTIMATED FOR ADDITION OF NOX CATALYST

                                BED AT 75 PERCENT EFFICIENCY

                                VARYING DRIVING CYCLES

                                AND CONTROL TECHNIQUES            	
                                    "I—	1
                5          10         15         20         25

              INCREASE IN BSFC, percent (OVER UNCONTROLLED VEHICLE)
30
                                                                         a
                                                                         ai
                                                                         o
                                                                         o
                                                                         cc
                                                                        CO

                                                                        o

                                                                         X
                                                                        a
Figure 4-7. Effect of NOx emissions level on fuel penalty for light duty trucks

 (Reference 4-83).
                                4-103

-------
1^
o
                         Table 4-43.  REPRESENTATIVE EFFECTS OF NOX CONTROLS ON CO EMISSIONS FROM INTERNAL COMfftfSTIGN ENGINES3
                                      (Reference 4~80).
Fuel
Natural Gas
Diesel
Dual Fuel
Engine Type
2-cycle
4-cycle
2-cycle
4-cycle
2-cycle
4-cycle
Baseline
Emissions
(ng/J)
15 - 40
75 - 3350
72 - 325
114 - 546 '
165
200 - 670
NOX Control CO Emissions (ng/J)
Derate
40 - 94
54 - 150
89
100 - 180
244
289
Retard
Ignition
35 - 45
80 - 1000
140 - 628
260 - 654
244 - 267
679 - 1070
Increase
A/F
29 - 31
__h
439
288
Decrease
A/F
117
675
244
296
Reduce
MAT
29 - 45
131
71
142 - 550
67
632
Mater
Injection
194
464
460 - 606
503 - 507
                This table Is Included 1n Appendix A 1n English-units.
                Denotes no data reported.

-------
       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 NO  combustion controls on 1C engines can also have significant effects on HC
emissions, with different NOX reduction techniques eliciting different effects.
       As shown in  Figure 4-8,, 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  4-9  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 4-10,. 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 4-1H 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 4-12.  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
                                                 4-105

-------
  100
 O 2 CYCLE, BLOWER SCAVENGED

 Q 2 CYCLE, TUREIGCHARGED

 A 4 CYCLE, NATURALLY ASPIRATED
 04 CYCLE, TURBOCHARGED

G  NATURAL GAS

OF DUAL FUEL              	
0  DIESEL
§
                                   20              30

                                   POWER DERATE, percent
                                                                  40
                                                                                  50
     Figure 4-8.   Effect of derating on 1C engine  HC emissions  (Reference 4-80),
 1000
                                        O 2 CYCLE, BLOWER SCAVENGED
                                        Q 2 CYCLE. TURBOCHARGED
                                          4 CYCLE. NATURALLY ASPIRATED
                                        O4 CYCLE, TURBOCHARGED
                                       G  NATURAL GAS
                                       OF  DUAL FUEL
                                       D  DIESEL
                                                                  8
                                   4              6

                                  TIMING RETARD, degrees

Figure 4-9. Effect of retarding ignition timing on 1C engine HC emissions (Reference 4-80)
                                                                                 10
                                       4-106

-------
    1000
  CO
  o
  53
  S2
     100
    " 20
      •30
•20
                                                        O 2 CYCLE, BLOWER SCAVENGED
                                                        Q 2CYCLE.TUR80CHARGED
                                                        O 4 CYCLE, TURBOCHARGEO
                                                       G  NATURAL GAS
                                                       OF  DUAL FUEL
                                                       D  DIESEL
•10    .        0           10
     CHANGE IN A/F RATIO, percent
20
30
Figure 4-10.  Effect  of air-to-fuel  ratio on  1C engine HG  emissions  (Reference  4-80)
                                          4-107

-------
1000
 O 2 CYCLE, BLOWER SCAVENGED
 Q 2 CYCLE,TUR80CHARGEO
 O4CYCLE TUR80CHARGED
G  NATURAL GAS
OF DUAL FUEL            	
D  DIESEL
                                   10              15
                                     MAT DECREASE, °K
          20
                          25
   Figure 4-11.  Effect of decreased manifold air temperature (MAT)  on  1C engine
                 HC emissions (Reference 4-80).
1000
                        >DF
                                                       O 2 CYCLE, BLOWER SCAVENGED
                                                       O4CYCLE.TURBOCHARGED
                                                      G  NATURAL GAS
                                                      OF DUAL FUEL
                                                      0  DIESEL
                                  0.4              0.6
                                        W/F RATIO
          0.8
1.0
   Figure 4-12.  Effect of water injection in 1C engine  HC  emissions  (Reference 4-80),
                                         4-108

-------
diesel engines (Reference  4-84  and 4-85).   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 4-13 shows the relationship between smoke emissions and NOX reduction as a function of NOX
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 (E6R) cause the most significant increases
•fn smoke level.
       Since NOX controls which caused smoke levels to exceed 10 percent opacity were considered
unacceptable in the tests summarized in figure 4-13  none of the data points for controlled engines
are above this value.  However, the effect qf progressively applying ignition retard and ESR on
smoke emissions is best demonstrated by data v/hich include higher smoke levels.  Such data are pre-
sented in Table 4-44 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 increase!! 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 NOV emission contribution from these sources.  In response
                                           A -
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 NO  concentrations primarily due to higher
combustion temperatures and to increased residence time at high temperatures.  Oil-fired turbines
generally have higher NO  concentrations than gas-fired units.  Typical uncontrolled NO, emissions
                        A                         .                                     X
from gas turbines are illustrated in  Figures  4-14 and 4-15 for large and small units, respectively.
                                                 4-109

-------
                                       NOX LEVEL, j/kWh
                                    12            16
                                                    20
                                        24
a
         A
         C
         0
         E
         H
u
s
                             CONTROL CODE
AIR-TO-FUEL RATIO      I
REDUCE COMP. RATIO    M
DERATE               R
EXTERNAL EGR
H20 INDUCTION         S
INTERNALEGR
MANIFOLD AIR TEMP.
RETARD IGNITION
TIMING
INCREASE SPEED
    R
~"\. FUEL
TYPE^^s^
2 STROKE
BLOWER
SCAVENGED
2 STROKE
TURBO-
CHARGED
4 STROKE
NORMALLY
ASPIRATED
4 STROKE
TURBO-
CHARGED
DIESEL
o
a
A
•
DUAL
FUEL



0
                                   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 WERf MEASURED
                                   WITH A 30SCH METER.
                                   FINAL SMOKE LEVEL IS AY END OF LINE HAVING
                                   CONTROL CODE.)         I         I
  10
<
         #34


           D
                                         10       12
                                       NOX LEVEL, j/hp-hr
                                               14
                                16
18
20
               Figure 4-13.  Smoke levels versus NOX  levels  *or large bore diesel engines.
                                              4-110

-------
            TABLE 4-44.   RELATIONSHIP BETWEEN SMKE,
                         EGK, AND RETARD
                         (Reference  4»80).
Engine Type
2-cycle, Blower
Scavenged Diesel






2-cycle,
Turbocharged Diesel



Control a
None
10% EGR
20% EGR
39% EGR
4° advance
None
48 retard
None
4.9% EGR
8.4% EGR
12. U 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  1s  not  a  control; data included to show trend,

-------
   250
   200
o
£  150
   100
u
 M
    SO
                               O GAS-FIRED UNITS
                               D OIL-FIRED UNITS
                          NOTE:   DATA NOT ADJUSTED FOR
                                  GASTURBINE EFFICIENCY
                      a
       PROPOSED NSPS
                                    a
                                    o
                                                                                   a
                   to
15             20

        TURBINE SIZF, MW
25
30
                                                                                          35
       Figure 4-H.   NOX emissions from large gas turbines without NOX controls  (Reference  4-86).
                                             4-112

-------
   200
I
   ISO
    O GAS-FIRED UNITS


    DOIL-FiRED UNITS


NOTE: DATA NOT ADJUSTED FOR

      GAS TURfclNE EFFICIENCY
a:
u<
ci
   100
o
u
    50
       PROPOSED NSPS
                                                                              a
                                                                              a
                   o.s
   1.0
2.5
                                           1.5            2.0


                                     TURBINE SIZE, MW


Figure  4-15.  NOX emissions from small gas turbines  without HOX controls  (Reference 4-86).
3.0
                                             4-113

-------
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 NO
                                                                                               A
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 NOX contra! in gas turbines is to eliminate the high temperature
stolchioroetric regions through water injection, premixing, improved'primary zone mixing, and  down-
Streamed dilution.
       Combustion modifications for gas 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 NQX at 15 percent oxygen.  Presently available wet and dry methods  for NO   reduc-
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 4-45.
       Wet techniques are the irost 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
developmant 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:
       •   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
                                                  4-114

-------
TABLE 4-45.  GAS TURBINE -  SUMMARY OF  EXISTING TECHNOLOGY  -  COMBUSTION MODIFICATIONS
Modification
Wet Controls
Water Injection
Stead Injection
Dry Controls
Lean Out Prlnary
Zone
Increase Mass
FloMrate
Earlier Quench
with Secondary
Air
Air Blast or
Air Assist
Atomlzation
Reduce Inlet
Preheat
(Regenerative)
Other Minor
Combustor Modi-
fications and
Retrofit
Exhaust Gas
Recirculatlon
Catalytic
Combustion
Approach to NOX Reduction
Control Potential
Lower peak flame temp 50-90+ t
by utilization of
heat capacity and
heat of vaporization;
most effective when
Injected Into primary
flame zone
Same as above 50-90+ %
Lower peak flame temp 10-201
Reduce residence time To 159E
at peak temperatures
Reduce residence tine To 1SX
Reduce peak flame temp To -90%
by Increasing mixing
thereby reducing local
A/F ratio
Reduce peak flame temp Not Available
Reduce peak flame temp To 60S
through premlxlng, Combined
secondary air Injec-
tion, primary zone
flow reclrculatlon
Reduce peak flame To 38!
temperatures
Complete combustion To 98%
at lower peak
temperatures
Hear Term
To date, most effective
measure, reasonable cost
To date, most effective
measure after water
Injection
Attractive option, re-
quires additional con-
trols to meet standards
Attractive option If
feasible
Minor combustor modi-
fication used pre-
sently with wet
controls
Limited application
for retrofit
Not attractive due to
thermal efficiency
reduction
Attractive near tern
as an Interim solu-
tion
Option has seen use in
minor combustor modi-
fications
Technology not avail-
able
Far Term
Not as attractive as
dry controls, but
Is adequate If dry
not developed
Same as above
•»
Generally seen as an
option to be incor-
porated Into new low
NOX designs
Not an attractive long
term option due to In-
flexibility
An attractive concept
to be employed in
advanced combustors
Promising method to be
Incorporated Into new
low NOX design
Not attractive for long
term solution
Unknown at this
time
An attractive option
for future design
with Internal com-
bustors
Attractive method for
new combustor designs
Additional Comments
Reduces efficiency. Increases capital
costs up to 10%. Operating costs as
low as It depending on usage. Hin-
dered by requirement for "clean"
water supply. Ineffective In reducing
fuel NOX.
Increases overall efficiency by In-
creasing flowrate. Installation and
operating costs same as water Injec-
tion. Requires high pressure steam.
Ineffective In reducing fuel NOX.
Possible decrease In power output,
less control over flame stabilization
Possible Increase In shaft speed
constant torque
An attractive option both for near
tern minor combustor modifications
and for Incorporation Into new de-
signs. Limited by flowrates and
Incomplete combustion.
Generally considered a new design
concept
Reduces efficiency.
With proper design, efficiency
unimpaired
Reduced efficiency requires additional
controls.
Current research aimed at reducing
reliability, maintainability, life-
time and start-up
Refs.
4-81
4-86
4-87
4-88

4-81
4-86
4-23
4-86
4-81
4-86
4-89
4-86
4-86
4-86
4-86
4-86

-------
       •   Uncertainty regarding long-term control effects on turbine components.
       Although few combinations of presently available dry controls have the NOX 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 NO   levels up to 20 percent by Towering 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.
       t   Increase mass flowrates — Possible NOV reductions up to 18 percent by reducing
                                                J\                        '
           residence time at peak flame temperature.  This control essentially increases
           the turbine speed at constant torque and is not feasible in many applications.
       •   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.
       •   A1r 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 NOX 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 NOX reduction  dif-
ferently than do wet controls, their effects are complementary and, consequently,  can be used toge-
ther.  Figure 4-16 illustrates the effect of dry and wet controls used separately and in combination
                                                  4-116

-------
IIU



100
s
a.

-------
for liquid fuels (Reference 4-86).  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 NOX at 15 percent oxygen,.  A possible long term option
1s 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 4-81  and
4-86).
4.3.2.2   Costs
      The most  recent cost study of NOX controls for gas  turbines  has been performed by the EPA
 .(Reference 4-86).  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 efficienty, their complete  development and application to large produc-
 tion  turbines is still several years away.   Cost considerations for dry methods are, therefore,  not
 discussed.  All costs in Reference 4-86 are stated  in  terms of 1975 dollars.
       Table 4-46, derived from Reference 4-86, 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  4-47 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  4-86).   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.
                                                  4-118

-------
                        TABLE 4-46.  IMPACT OF NOX EMISSION CONTROL ON THE INSTALLED CAPITAL COST

                                     OF GAS TURBINES (Reference 4-86, 1975 Costs)
I
•-J


to
Application
A. Standby
1 . 260 kwa (350 hp)
2. 820 kWa (1100 hp)
B. Industrial
1. 3 MWa (4000 hp) -typical
2. 3 MWa - offshore
C. Utility
1 . 66 MWb
Installed Cost (1000$)
Uncontrolled

55.6
177.9

352.8
352.8

9900.0
Controlled

58.0
179.3

366.8
379.8

10070.0
% Increase

2.4
0.8

4.0
7.1

1.7
               ashaft output



                electrical output

-------
                                              Table 4-47.   WATER INJECTION  COSTS, MILLS/kWh (Reference  4-86, 1975 Costs)
Item
Unit Size
Hours of Operation
Per Year
Annual 1 zed Fixed
Costs
Operating Cost of
Hater Treatment
Hater for Injection
Energy Penalty
Hater Transport
Costs
Output Enhancement
Total
Baseline Costs
Percent Impact
Standby
S-l
260 kH
80
13.65
0.46
0.01
O.S1
-
-
14.63
611.29
2.39
S-2
260 kU
200
5.46
0.46
0.01
0.51
-
-
6.44
264.79
2.43
S-3
820 kU
80
4.32
0.37
0.01
0.51
- «
-
5.21
611.29
0.85
S-4
820 kU
200
1.73
0.37
0.01
0.51
-
-
2.62
264.79
0.99
Industrial
1-1
3HU
2000
0.48
0.10
0.01
0.43
-
-
1.02
43.77
2.32
1-2
3HU
2000
0.12
0.10
0.01
0.43
-
0.09
0.57
32.53
1..75
1-3
3 KU
3000
0.12
0.10
0.01
0.43
0.64
0.09
1.21
35.53
3.71
Utility
U-l U-Z U-3 U-4
66 KU
200
2.58
0.11
0.01
0.34
-
-
3.04
180.00
1.69
66 MM
500
1.03
0.11
0.01
0.34
-
-
1.49
85.45
!.7S
66 HU
2000
0.26
0.11
0.01
0.34
-
-
0.72
38.20
1.88
6HH
8000
0.06
0.11
0.01
0.34
-
0.11
0.41
26.39
1.5S
Offshore
Drilling
Platform
11 GJ/hr
8000
0.23
0.35
-.
0.43
-
0.09
0.92
32.53
2.83
f\J
o
               9Ut111tv turbine size (is electrical output, others «s shaft output

-------
       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 2i 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 abso?ute 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.
                                                  4-121

-------
       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
1n operating costs.
4.3.2.3   Energy and Environmental Impact
       As was the case for reciprocating 1C engines, the energy impacts of applying NOX 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 HgO/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 NOX 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 SO , trace metal, and, to some extent, higher molecular weight organic emissions are
insignificant.
       The effects of some comnonly applied NOX control techniques on CO emissions from gas turbines
are summarized in Table 4-48.   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 4-48..
       The very limited data on incremental hydrocarbon emissions due to NOX combustion controls
applied to stationary gas turbines are summarized in Table 4-49.   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
                                                 4-122

-------
TABLE 4-48.    REPRESENTATIVE  EFFECTS  OF NOX  CONTROLS'ON CO EMISSIONS FROM GAS TURBINES (Reference  4-86).
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
Basel i ne
102
102
53
195
969
53
147
252
99
135
93
NOX Control
51
96
99
59
no
36
1134
1512
144
162
30
                   02> dry basis.  Emission:; levels at full load.
            TABLE 4-49    SUMMARY OF THE EFFECTS OF NOX CONTROLS ON VAPOR PHASE HYDROCARBON
                          EMISSIONS FROM GAS TURBINES (Reference 4.36).
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
NOX Control
41
11
9-12
12
7
12
372
246
27
12
Comment
Idle
Full load
( Full load
/ W/F = 0.5
               a3% 02, dry basis.
                                                  4-123

-------
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 NO
                                                                                                  X
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 4-86).  Some tests show smoke level reduction of 1.5 to 1.75 smoke sopt numbers when
water injection is used.  Other tests, however, indicate particulate emissions with water injection
at peak load.

4.4    SUMMARY
       Table 4-50 summarizes current and emerging NOX control  technology for the major source
categories (1974 costs).  As shown, most of the- current NO  control  technologies rely on combustion
process modification.  Emerging technologies include various combustion modification and flue gas
treatment processes.  However, combustion process modifications will  likely predominate in  the U.S.
except for situations where very stringent emission levels are required.
       Historically, utility boilers have been the most extensively regulated source category, and
accordingly, NO  control technologies for utility boilers are  the most advanced.  Available combus-
tion modification technology ranges from simple operational adjustments such as low excess  air,
biased burner firing, and burners out of service to application of overfire air ports, flue gas
recirculation, low NO  burners, and enlarged furnace designs.   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
boilers, the emphasis in emerging technology is on development of controls applicable to new unit
design.  Advanced low NO  burners and/or advanced off-stoichiometric combustion techniques  are the
                        X
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.

                                                4-124

-------
                                       TABLE 4-50.   SUMMARY OF NOV CONTROL TECHNOLOGY
                                                                 A
Equipment/
Fuel Category
Existing coal -fired
utility boilers
New coal -fired
utility boilers
Existing oil-fired
utility boilers
Existing gas-fired
utility boilers
Oil-fired industrial
watertube boilers
Stoker-fired,
industrial water-
tube boilers
Gas-fired industrial
watertube boilers
Current Technology
Available
Control
Technique
LEA + OSC
(OFA, BOOS,
BBF); new
burners
LEA + OFA;
new burners;
enlarged
furnace
design
LEA + OSC •*•
FGR;
LEA + OSC
+ FGR;
LEA + OSC
(OFA, BOOS,
BBF); FGR;
new burners
LEA + OFA
LEA + OSC
(OFA, BOOS,
BBF); FGR;
new burners
Achievable
NO Emission
l.8vel,ng/J
(lb/10° Btu)
215 - 300
(0.5 - 0.7)
170 - 260
(0.4 - 0.6)
86 - 170
(0.2 - 0.4)
43 - 100
(0.1 - 0.25)
85 - 130
(0.2 - 0.3)
150 - 190
(0.35-0.45)
86 - 130
(0.2 - 0.3)
Estimated
Differential
Capital Cost
$0.7 - 2.2/kVJ
1978$ e
$2 - 3/kW
1978$ e
$6/kH
1978$e
$6/kW
1978$e
$1 - 1.5/kW.
1978$ *
$1 - 3/kW.
1978$ l
$1.5/kU.
1978$ fc
Operational Impact
Possible increase
in corrosion &
slagging & carbon
in fly ash
No major problems
Possible flame
instability;
boiler vibration
Possible flame
instability;
boiler vibration
Possible IX
increase in fuel
consumption; flame
instability;
boiler vibration
(retrofit)
Possible 1%
increase in fuel
consumption; .
corrosion; slagging
of grate (retrofit)
Possible 1%
increase in fuel
consumption; flame
instability;
boiler vibration
(retrofit)
Emerging
Technology
Advanced low-NO
burners; SCR; SRR
Advanced low-NO
burners; advanced
furnace designs;
SCR; SNR;
fluidized bed
combustion
Low-NO burners;
oil defiitrifica-
tion; SCR; SNR
Low-NO burners;
SCR; SNR
Low-NO burners;
OFA in new unit
designs, optimized
burner/ firebox
designs; oil
denitrifi cation;
SCR; SNR
Inclusion of OFA
in new unit design;
fluidized bed
combustion; SNR
Low-NO burners;
OFA in new unit
design; optimized
burner/firebox
designs; SNR
Comments
Flue gas treatment techniques are
potential supplements to combustion
modifications if needed
Flue gas treatment techniques are
potential supplements to combustion
modifications if needed
No new units; emission levels are
limit of current technology
No new units; emission levels are
limit of current technology
Current technology still undergoing
development
Current technology still undergoing
development
Current technology still under-
going development
ro
tn

-------
                                 TABLE 4-50.  SUMMARY OF NOV CONTROL TECHNOLOGY (Continued)
                                                           3\
Equipment/
Fuel Category
Industrial flretube
boilers
Gas turbines
1C engines
Current Technology
Available
Control
Technique
LEA + FGR;
LEA + OSC
Water, steam
injection
Fine tuning;
changing A/F
Achievable
NOX Emission
Level fing/J
(lb/100 Btti)
65 - 110
(0.15 - 0.25)
110 - 150
(0.25 - 0.35)
1,070 - 1,290
(2.5 - 3.0)
Estimated
Differential
Capital Cost
$6/kVL
1978$c
$1 - 2/kW
1975$
$0.70 - 2.00/kW
($0.50 - 1.50/BHP)
1975$
Operational Impact
Possible IS
Increase 1n fuel
consumption; flame
instability;
(retrofit
Possible 1%
increase in fuel
consumption;
affects only
thermal NOX
5 - 1Q% increase
in fuel consump-
tion; misfiring;
poor load response
Emerging
Technology
Low-NO burners;
OFA or FGR in new
unit design;
optimized burner/
firebox design
Advanced combustor
designs for dry
NO controls;
catalytic combus-
tion; advanced
can designs
Include moderate
control in new
unit design;
advanced head
designs
Comments
Development continuing on current
technology
Current technology widely used
Technology still being tested
I
t-»


01

-------
                                      REFERENCES FOR SECTION 4

4-1       Maxwell, J.D. and L.R. Humphries, "Evaluation of the Advanced Low-N0v  Burner,  Exxon,  and
          Hitachi Zosen DeNOx Processes," EPA-600/7-81-120, TVA/OP/EDT-81/28,  July 1981.   p.  5.
4-2       Lim,, K.J., e_t aQ_., "Environmental Assessment of Utility Boiler Combustion Modification  NO
          Controls; Volume 1. Technical Results," EPA-600/7-80-075a,  April  1980.   pp.  4-1  to  4-56.
4-3       Waterland, L.R., et al., "Environmental Assessment of Stationary  Source NO  Control
          Technologies — FTnaTReport," EPA-600/7-82-034, May 1982.   pp.  133-196.  x
4-4       Manny, E.H., and P.S. Natanson, "Fireside Corrosion and NO   Emission Tests on  Coal-Fired
          Utility Boilers," In:  Proceedings of the Joint Symposium on Stationary Combustion  NO
          Control, Volume I, Utility Boiler NO  Control by Combustion Modification. IERL - RTP-I083,
          October 1980 (as cited in Reference 4-3).
4-5       Barsin, J.A., "Pulverized Coal Firing NO  Control," In:  Proceedings Second  NO  Control
          Technology Seminar, EPRI FP-1109-SR, Jul$ 1979 (as cited in Reference  4-3).*
4-6       Reference 4-1, 111 pp.
4-7       Ando, J., "NOX Abatement for Stationary Sources in Japan,"  EPA-600/7-79-205, August 1979.
4-8       Faucett, H.L., J.D. Maxwell, and T.A. Burnett, "Technical Assessment of NO  Removal
          Processes for Utility Application," TVA Bulletin Y-120, EPA-600/7-77-127 (fiTIS PB 276
          637/6WP), EPRI FP-1253, 1977 (as cited in Reference 4-6).
4-9       Jones, G.D., "Selective Catalytic Reduction and NO  Control in Japan,  A Status Report,"
          EPA-600/7-81-030, January 1981.                 ' x
4-10      Reference 4-9, p. 10.
4-11      Reference 4-1, p. 9.
4-12      Ando, J., "SO, and NO  Removal for Coal-Fired Boilers in Japan,"  Prepared for  the Seventh
          Symposium on Flue Gas Desulfurization, May 17-20, 1982.  p. 5.
4-13      Mobley, J.D. and J.M. Burke, "EPA's Pilot Plant Evaluations of NO  and NO /SO   Flue Gas
          Treatment Technology," prepared for the Seventh Symposium on Flue Gas  Desulfurization,
          May 17-20, 1982.
4-14      Reference 4-12, p. 10.
4-15      Reference 4-9, pp. 1-5.
4-16      Johnson, L.W., C.L.W. Overduin, D.A. Fellows, "Status of SCR Retrofit  at Southern
          California Edison Huntington Beach Generating Station Unit  2."  In:  Proceedings of the
          Joint Symposium on Stationary Combustion NO  Control, Volume II Utiliby Boiler NO  Control
          by Flue Gas Treatment, IERL-RTP-1084. Octobir 1980.pp. 32, 33.                  *
4-17      Trexler, E.G., "OOE's Electron Beam Irradiation Developmental Program," Prepared for the
          Seventh Symposium on Flue Gas Desulfurization, May 17-20, 1982.  p.  11.
4-18      Reference 4-2, pp. 7-1 to 7-49.
4-19      Reference 4-3, pp. 149 to 154.
4-20      Maxwell, J.D., T.A. Burnett, H.L. Faucett, "Preliminary Economic Analysis of NO  Flue Gas
          Treatment Processes," EPA-600/7-80-021, EPRI FP-1253, TVA ECDP B-6,  February 1980.
4-21      McGlamery, G.G.,-et al., "Detailed Cost Estimates for Advanced Effluent Desulfurization
          Processes," EPA-60072^75-006, January 1975 (as cited in Reference 4-18).
4-22      'Waitzman, D.A., et al., "Evaluation of Fixed-Bed Low-Btu Coal Gasification Systems  for
          Retrofitting Power Plants," EPRI Report 203-1, February 1975 (as  cited in Reference 4-18);
                                                4-127

-------
4-23      Ponder, W.H., R.D. Stern, and G.6. McGlamery, "SO  Control  Methods  Compared,"  The  Oil  and
          Gas Journal, December 1976.  pp. 60-66 (as cited tn Reference 4-18).
4-24      Engdahl, R.B., "The Status of Flue Gas Desulfurization," ASME Air Pollution  Control
          Division News, April 1977 (as cited in Reference 4-18).
4-25      Princiotta, F.T., "Advances in SO  Stack Gas Scrubbing," Chemical Engineering  Progress,
          February 1978.  pp. 58-64 (as cited in Reference 4-18).
4-26      Campobenedetto, E.J., Babcock & Wilcox Co., Barberton, OH,  Letter to  K.J.  Lim, Acurex
          Corporation, Mountain View, CA, November 15, 1977 (as cited in Reference 4-19).
4-27      Vatsky, J., "Effectiveness of NO  Emission Controls on Utility Steam  Generators,"  Foster
          Wheeler Report to Acurex Corporation, FW Contract 2-43-3245, Livingston, NJ, November  1978
          (as cited in Reference 4-19).
4-28      Martin, G.B., "Field Evaluation of Low NO  Coal  Burners  on  Industrial  and  Utility
          Boilers," In:  Proceedings of the Third Stationary Source Symposium;  Volume  I,
          EPA-600/7-79-050a, February 1979 (as cited in Reference  4-19).
4-29      Johnson, S.A., et al., "The Primary Combustion Furnace System — An Advanced Low-NO
          Concept for Pulverized Coal Combustion," In:  Proceedings:   Second  NO  Control Technology
          Seminar. EPRI FP-1109-SR, July 1979 (as cited in Reference  4-19).*
4-30      Teixeira, D., "NO  Control Technology."  EPRI Journal.  V3(9):37, November 1978  (as  cited
          1n Reference 4-19J.
                                                                              m
4-31      "Technical and Economic Feasibility of Ammonia-Based Post Combustion  NO Control",
          EPRI CS-2713, Electric Power Research Institute, Palo Alto, California, November 1982.
4-32      Reference 4-12, p. 19.
4-33      Reference 4-2, pp. 8-39 to 8-41.
4-34      Reference 4-6, pp. 81-85.
4-35      Reference 4-20, pp. xxv to xxviii.
4-36      Reference 4-13, p. 8.
4-37      Reference 4-17, pp. 5,6.
4-38      Reference 4-3, 448 pp.
4-39      Reference 4-9, pp. 6,7.
4-40      Devltt, T., et al., "The Population and Characteristics  of  Industrial/Commercial Boilers,"
          EPA-600/7-79^I75a", August 1979.
4-41      "Non-fossH-fuel Industrial Boilers, Background  Information," EPA-450/3-82-007,
          Environmental Protection Agency, Research Triangle Park, NC, March  1982.
4-42      Cato, G. A., jet al_., "Field Testing:  Application of Combustion Modification to  Control
          Pollutant Emissions from Industrial Boilers—Phase II,"  EPA-600/2-76-086a, April 1976.
4-43      Burklin, C. E., and W. D. Kwapil, "Regressions for NO Emissions from Oil- and Gas-Fired
          Industrial Boilers," EPA Contract No. 68-02-3058, Radtan Corporation,  Durham,  NC,
          May 27, 1982.
4-44      "Fossil Fuel-Fired Industrial Boilers, Background Information, Volume  I:   Chapters 1-9,"
          EPA-450/3-82-OQ6a, Environmental Protection Agency, Research Triangle  Park, NC,
          March 1982.
4-45      McElroy, M. W., and D. E. Shore, "Guidelines for Industrial  Boiler  Performance
          Improvement," EPA-600/8-77-003a, January 1977.
                                             4-128

-------
4-46     'Hunter, S. C.,  and H.  J.  Buening, "Field Testing:   Application of Combustion  Modifications
          to Control Pollutant Emissions from Industrial  Boilers,  Phase I and II,  Data  Supplement,
          "EPA 600/2-77-122, June 1977.

4-47      Chicanowica, J. E., et jil_.,  "pollutant Control  Technique for Packaged Boilers.   Phase  I.
          Hardware Modifications and  Alternate Fuels,"  Ultrasystems 'Draft Report,  EPA Contract
          No. 68-02-1498, Ultrasystems Corporation, Irvine,  CA,  November 1976.

4-48      Carter, W. A.,  et al., "Emissions Reduction on  Two Industrial Boilers with  Major
          Combustion ModiTTcations,"  EPA-600/7-78-099a, June 1978.

4-49      Cato, G. A., et al., "Field Testing:  Application  of Combustion Modifications to Control
          Pollutant Emission's from Industrial Boilers—Phase I," EPA-600/2-74-078a, October 1974.

4-50      Lim, K. J., et  al., "Environmental Assessment of Utility Boiler Combustion  Modification
          NO  Control sT1" ETA Contract No. 68-02-2160, Acurex Corporation, Research Triangle
          Park, NC, April 1978.

4-51      Palazzolo, M. A., "Air Preheat vs. Economizers," Technical  Note, Radian  Corporation,
          Durham, NC, September 2,  1981.

4-52.      Mason, H. 8., et al., "Preliminary Environmental Assessment of Combustion. Modification
          Techniques:  VoTuie II, Technical Reports," EPA-600/7-77-119b, October 1977.

4-53      Langsjoen, P. L., ejb al., "Field Tests of Industrial Stoker Coal-Fired Boilers for
          Emissions Control and~E~fficiency Improvement—Site K," EPA-600/7-80-138a, May 1980.

4-54      Lim, K. J., C.  Castaldini,  and H. I. Lips, "Industrial Boiler Combustion Modification  NO
          Control:  Volume  \, Environmental Assessment,"  EPA-600/7-81-126a, July 1981.

4-55      Burklin, C. E., "'Application of Staging Burners to Industrial Boilers,""Draft Technical
          Note, EPA Contract No. 68-02-3058, May 1982.

4-56      American Boiler Manufacturers Association, "Emissions  and Efficiency Performances of
          Industrial Coal Stoker-Fired Boilers, Volume 1," DOE/ET/10386-T1 (Vol.lj, Arlington, VA,
          August  1981.

4-57      "Environmental  Assessment of Stationary Source  NO  Control  Technologies: Final Report,"
          EPA-600/7-82-034, Environmental Protection Agency, Research Triangle Park,  NC, May 1982.

4-58      Maloney, K. L., et al., "Low Sulfur Western Coal Use in Existing Small and  Intermediate
          Size Boilers,"  EPA"-600/7-78-l53a, July 1978.

4-59      Gabrielson, J.  E., et  al., "Field Test of Industrial Stoker Coal-Fired Boilers for
          Emissions Control and "ETficiency Improvement -  Site A," EPA-600/7-78-136a,  July 1978.

4-60      Keller, L. E.,  and M.  S. Jennings, and W. D. Kwapil, "Regressions for NO Emissions  from
          Coal-Fired Spreader Stoker Industrial Boilers," EPA Contract No. 68-02-3058,  Radian
          Corporation, Durham, NC, July 22, 1982.

4-61      Carter, W. A.,  et al., "Thirty-day Field Tests  of Industrial Boilers:  Site 1—Coal-Fired
          Spreader Stoker"?1' EPA-600/7-80-085a, April 1980.

4-62      Carter, W. A. and J. R. Hart, "Thirty-day Field Tests  of Industrial Boilers:  Site 4  —
          Coal-fired Spreader Stoker," EPA-600/7-80-085d, April  1980.  ,

4-63      Carr, R. C., "Effectiveness of Gas Recirculation and Staged Combustion in Reducing NO  on
          a  560-MW Coal-Fired Boiler," Electric Power Research Institute, NTIS No. PB-260-282, x
          September 1976.

4-64   '   Carterm W. A., et. al_., "Emission Reduction on Two Industrial Boilers with Major Combustion
          Modifications," EPA-600/7-78-099a, June 1978.

4-65      Exxon Research and Engineering Company, "Exxon  Thermal DeNo  Process," New  Jersey, Exxon
          Technology, April 1978.
                                                 4-129

-------
4-66      "Exxon Corp. Stationary NO  Emissions Significantly Reduced at Plant,"  Air and Water
          Pollution Report, February 20, 1978.
4-67      Jones, 6. D., and K. L. Johnson, "Technology Assessment Report for Industrial  Boiler
          Applications:  NO  Flue Gas Treatment, Final Report," EPA-600/7-79-178g,  December 1979.
4.-6S      Heap, M. P., et al., "Reduction of Nitrogen Oxide Emissions from Field  Operating Package
          Boilers, Phase~~lTT," EPA-600/2-77-025 , January 1977.
4-69      Schwieger, R. , "Industrial Boilers - What's Happening Today," Power,  February  1977.
4-70      Schwieger, R., "Industrial Boilers - What's Happening Today, Part II,"  Power,
          February 1978.
4-71      Personal communication, K.J. Lim of Acurex Corporation with J. Lindsay, Zurn Industries,
          San Francisco, CA, August 17, 1978.
4-72      Lira, K. J., et al,, "Environmental Assessment of Utility Boiler Combustion Modification
          NOV Control s7rlPA-600/7-80-075a and b, April 1980.
            A
4-73      Personal communication, K.J. Lim of Acurex Corporation with B. Morton,  E.  Keeler, Co.,
          Williamsport, PA, August 8, 1978.
4-74      Lyon, R. K. , and J. P. Longwell, "Selective, Non-Catalytic Reduction  of NOV by NH,,"
          Proceedings of the NO  Control Technology Seminar, EPRI SR-39, NTIS-PB  2532661,  J
          February 197FT

4-75      Wong-Woo, H., and A. G. Goodley, "Observation of Flue Gas Desulfurization and Denitrifica-
          tion Systems in Japan," California Air Resources Board, SS-78-OOK,  March  7,  1978.
4-76      Vargo, G. M. , Jr., et al . , "Applicability of the Thermal  DeNO  Process  to Coal-Fired
          Utility Boilers," EPA-600/7-79-079 , March 1979.              x
4-77      Cato, G. A., et al., "Reference Guideline for Industrial  Boiler Manufacturers to Control
          Pollution witFCoinbustion Modification," EPA-600/8-77-003b, November 1977.
4-78      Koppang, R. R. , "A Status Report on the Commercialization and Recent Development History  ,
          of the TRW Low NO  Burner," TRW Internal Report, TRW, Inc., Redondo Beach, CA,
          January 1976.
4-79      McGowin, C.R., "Stationary Internal Combustion Engines in the United States,"
          EPA-R2-73-210, April 1973.
4-80      "Standard Support and Environmental Impact Statement - Stationary Reciprocating  Internal
          Combustion Engines," (Draft Report).  Acurex Corp. /Aero therm Division,  Mountain  View,
          California, Project 7152, March 1976.
4-81      Aerospace Corporation, "Assessment of the Applicability of Automotive Emission Control
          Technology to Stationary Engines," EPA-650/2-74-051, July 1974.
4-82      The American Society of Mechanical Engineers (ASME), "Power Costs,  1974 Report on  Diesel
          and Gas Engines," March 1974.
4-83      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.'
4-84      Bascom, R.C., et al., "Design Factors that Affect Diesel  Emissions," SAE  Paper 710484, .
          July 1971.
4-85      Hills, F.J., et al., "CRC Correlation of Diesel Smokemeter Measurements," SAE
          Paper 690493 ,~Hay~1969.
4-86      "Standards Support and Environmental Impact Statement, Volume I:  Proposed Standards of
          Performance for Stationary Gas Turbines," EPA-450/2-77-017a, September  1977.
                                               4-130

-------
4-87      Shaw, H., "The Effects of Water,  Pressure and Equivalence Ratio  on  Nitric  Oxide  Production
          in Gas Turbines," ASME Paper 73-WA/GT-l.

4-88      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.

4-89      Stern, R.D., "The EPA Development Program for NO  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.
                                                4-131

-------

-------
                                         SECTION 5

                                OTHER COMBUSTION PROCESSES

       Significant amounts of the total  fuels burned and NO  emissions released in the United States
                                                           A
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 localized,
residential5 areas.  Control techniques,  costs, and energy and environmental impacts are discussed
for those systems where data are available.

5.1    SPACE HEATING

       Emissions from stationary source  fuel combustion in 1980 were present in Section 2.  As shown
in Table 2-2, commercial and residential combustion sources contributed about 6 percent to the
national NO  emissions.  A study of 1977 NO, emissions showed that 56 percent of the total emissions
           A                               A       •
from residential and commercial combustion sources were from residential space heating systems and
28 percent were from commercial space heating systems (Reference 5-1).••

       Natural gas and distillate oil are the primary fuels burned in residential heating systems.
The combination of gas and oil accounted for 90 percent of the fuel burned for domestic heating in
1976.  Gas and oil are also the primary fuels consumed in commercial heating systems.  The
predominance of gas and oil is also reflected in the distribution of heating equipment types in the
residential sector.  Electrical heaters  accounted for nearly 14 percent of residential heating
equipment in 1976; LPG, coal, and wood-fired units accounted for 8 percent; and the remaining
equipment was gas- and oil-fired (Reference 5-1).

       Residential.heading units are characterized by thermostatically controlled heating cycles
(on/off cycles).  Natural gas-fired residential heating equipment generally employs single port
upshot or tubular multiport burners.  A standing pilot flame is often used to ignite the burner;
however, there is a rapid trend toward interrupted pilots which are turned off in the standby mode.
Oil-fired residential heating systems usually use high pressure atomizing gun-type burners.  Nearly
all new oil-fired systems use the flame retention burner because of its high efficiency.  Residen-
tial coal-fired furnaces are generally stoker fed and wood-fired heaters are usually hand fed
(Reference 5-1).
                                               5-1

-------
       Commercial heating systems can be divided Into three general categories:  space heaters, warm
air furnaces, and hot water or ot^am systems.  Warm air furnaces may be direct or indirect-fired.
Qiriic".-fired neoters use clean gaseous fuels and exhaust combustion products directly into the
heated soace.  Indirect fired heaters vent to the outdoors a>id an? simitar to residential warm air
furnace designs.  Commercial steam and hot water units include wtitertubes, firetubes, and cast iron
Doi'ieis.  Gas- and uii-fired boilers normally employ single powe" burner designs.  Some atmospneric
gas burners are used for snail units.  Coal-fired steam and hot water units are underfed stoker
units (Reference 5-1).

5.1.1  Emissions
       Hall, e_t  a_K  (Reference 5-2)  studied  the factors that affect emission  levels from residential
heaters.   This project, which concentrated or. an oil-fired warm air furnace,  showed thet excess  air,
resiij,.ce  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 purticulate  matter pass  through a
ninimum  as  excess air  is  increased from stoichiometric conditions.  By  contrast, both therm?1
efficiency  and NO '"-missions pass through maximum points as excess  air  is  increased.  The experi-
Tiental  results showed  that  increased residence  time  of the combustion  products  reduces  emissions of
CO   gaseous HC,  and  smoke but has no  ffeu  on  N0x emissions.  Combustion chamber material was found
to  affect  i':1  emissions.  Fu-riaces wltt steel-linecl  chambers requirsd  higher  excess air levels to
rearfi  optimum  emission levels, thus  reducing efficiency.  The  shape of  the  combustion chamber  had
 little  effect  on pollutant  generation.  A  specially  designed flame retention  device designed to
decrease participate emission* was found to  increase NOX  emissions, but such  a  dtvice also  increased
furn?i;e  efficiency.   Poorly maintained  f^rnac^s also yielded higher NOX emissions.
        In  another study rf  spa-e heating equipment  (Reference  6-3), emission  levels 'vere  found to be
dependent  upon boiler size, design,  b'"-ner tyoe, burner  age, and  operating  conditions.   The  type of
 fuel  used  i-i the combustion equipment for  space heating  al^o affect;  NOX emissions  because  of  fuel
 •••trogen convt-sion (Reference b-1).  Ot  tha fuels  typically used for space heating,  coal  and
 residual oil <.:antain significant quantities  of  nitroyen  which  may be  converted to  NO  emission,
 Conversion rates lor coal  rarje  from 20 to 60  percent and conversion  rates  in residential  and
 coirtr.e.'cia 1 heatiincj systems  . •«  on tne hiqh sicie of  the range  (Reference S-l),  Because  natural  gas
 any nial oi'i i:"9 ve/' low in  fuel nitrocfcii,  N0x ^reduction  from natural gas ana distillate oil

-------
.2.0
                         OPTIMUM SETTING FOR MINIMUM
                           EMISSIONS AND MAXIMUM
                     SMOKE      EFFICIENCY
                     (10TH
                                                                              8
                                                                              I
                                                                              X
                                                                              >-
                                                                              a:
                                                                              sc
                                                                              a
                          1.4       1.6      1.8      2.0

                              STOICHIOMETRIC RATIO
2.4
2.6
    Figure  5-1.   Stneral trend  of smoke, gaseous emissions,  and effi-
                  ciency versus  stoichiometric  ratio for  residential
                  heaters (Reference 5-4).
                                      5-3

-------
systems is due primarily to tnerml  NO  which  is  promoted by high temperatures  and  long  residence
times (Reference 5-1).
       NO  emission factors for domestic heatinc  reported in AP-42 (Reference 5-5)  are summarized  in
Table 5-1.  Some additional emission rates reported in the literature are summarized in  Table 5-?..

       Allen (Reference 5-6) found that although  NO  emissions from wood stoves have not been
recognized as serious they can be significant.  He found that even though the fuel  nitrogen content
in wood is low, conversion can be significant.  He further found fiat temperature:  within wood
stoves are not expected to reach the level where  atmospheric nitrogen fixation would occur.

5.1.2  Control Techniques

       Currently available emission reduction techniques for space haatir.j units are:  (1) tuning:
the  bust  adjustment in  terms of the smoke-COj relationship that can be achieved by normal cleanup,
nozzle replacement, simple scaling and adjustment with tha benefit of Vielo instruments, (2) equip-
ment replacement:  installation of a new, advanced low-NOx uriit, and installation of a new
low-emission burner.
 5.1.2.1  Tuning

        Reference  5-3  indicates that tuning has a beneficial effect on all pollutants with the
 exception  of  NO.   In  the field program, oil-fired un'ts considered in "poor" condition were
 replaced and  all  others were tuned, resulting in reductions in smoke, CO, HC, and filterable
 particulate matter  by  59, 81, 90, and 24 percent respectively, with no significant change in NO
 levels.  Table 5-3  shows mean emission  levels prior  to and after  replacement or tuning.  Although
 tuning  or  replacement  '?as been snown 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.

        As  an  aid  to controlling emission levels  from residential  and  commercial space heating
 systems,  EPA has  made available guidelines  *   oil  and gas burner adjustments  (References 5-7,  5-2
 5-9).   These guidelines  are  intended  for the  jse of  skilled technicians  and  for training service
 personnel.  The reroronended  adjustment  guidelines  provide  for efficient  fuel utilization and
 rimmize air pollution with  reliable  automatic operation.
                                                5-4

-------
        TABLE 5-1.  EMISSION FACTORS FOR RESIDENTIAL AND COMMERCIAL
                    HEATING SYSTEMS*
          FUEL
  EMISSION FACTOR
     Gas
     Distillate  Oil
     Anthracite
     Bituminous
     Lignite
1280 - 1920 kg/106 m3
(80 - 120 lb/105 ft3)b
2.3 kg/10J liter
18 lb/103 gal)

1.5 - 9 kg/Mg
(3 - 18 Ib/ton)
3 kg/Mg
(6 Ib/ton)

3 kg/Mg
(6'lb/ton)
a AP-42 (Referenrj 5-5).
  Lower value for residential  systems.   Higher value for commercial  systems,
                                     5-5

-------
                    TABLE 5-2.  UNCONTROLLED N0v EMISSIONS FROM RESIDENTIAL SPACE HEATING SYSTEMS
                                               A
                  FUEL
 N0x EMISSION RATE (as N02)
        SOURCE
         Natural gas (blue flame)
         Natural gas (yellow flame)
.42 ng/J (0.098 Ib m/10b Btu)
38 ng/J (0.088 Ib m/106 Btu)
(Reference 5-10)
Laboratory tests
in
I
Oi
         Natural gas (mostly boilers)
         Natural gas (mostly.hot air
           furnaces)
42 ng/J (0.098 Ib m/10b Btu)
49 ng/J (0.114 Ib m/106 Btu)
(Reference 5-11)
Field tests:  measurements
of emissions in chimneys of
natural gas heated homes.
         Fuel oil
fri.5 ng/J "(0.143 lb/10b Btu)
1.8 g/kg fuel
(Reference 5-3)
Field tests
         Anthracite
         Bituminous
3.9 g/kg fuel
0.9 g/kg fuel
(Reference 5-12)
Composites of many emission
rate measurements.

-------
                         TABLE  5-3.   COMPARISON  OF MEAN  EMISSIONS  FOR CYCLIC RUNS ON RESIDENTIAL
                                            OIL-FIRED UNITS  (REFERENCE  5-3)
Units

All Units
All units,
except those
in need of
replacement
Condition

As Found
Tuned
As Found
Tuned
Units
In
Sample

32
33
29
30
Mean
Smoke
No.

-
-
3.2
1.3
Mean.,Emission Factors
kg/rrf (lb/1000 gal)
CO

>2.65
(>22.1)
>1.96
(>16.4)
0.93
(7.8)
0.52
(4.3)
HC

0.68
(5.7)
0.36
(3.0)
0.09
(0.72)
0.07
(0.57)
NOX

2.32
.C19.4)
2.34
(19.5)
2.35
(19.6
2.34
(19.5)
Filterable
Parti cul ate

0.35
(2.9)
0.28
(2.3)
0.29
(2.4)
0.26
(2.2)
Ul
I
•-4

-------
5.1.2.2  Equipment Replacement
       Control equipment options for controlling NO  from gas- and oil-fired residential  furnaces
are listed in Tables 5-4 and 5-5.  Each of the tables also shows a conventional  unit for purposes  of
comparison.

Gas-Fired Equipment

       The American Gas Association Laboratories developed radiant screens capable of reducing  NO
emissions from 36 to 76 percent with an average reduction of 58 percent.   Incandescent radiant
screens in a natural gas flame radiate heat to the surroundings and cool  the flame.   Test results
showed that performance was more significant for multiport burners than for single port burners.
The screens were also found to increase steady-state furnace efficiency slightly because of
increased flame and burner radiation (Reference 5-1).  The Gas Appliance  Manufacturer Association
has reported installation and performance problems with the screens.  Some of the potential  problems
cited were performance sensitivity to screen location, effect on CO emissions, and deterioration due
to thermal shock (Reference 5-1).
       Secondary air baffles control secondary air flow into the flame front.  Decreasing the
concentration of excess oxygen at peak temperatures with reductions in secondary air was found  to
decrease NOX emissions from 10 to 40 percent (Reference 5-10).  However,  NQX reductions without
increases in CO emissions were generally limited to about 15 percent; (Reference 5-1).  The Gas
Appliance Manufacturer Association expressed concern over reliability and performance of secondary
air baffles.  Furthermore, it is not clear that secondary air baffles can be applied to all  types  of
residential furnaces (Reference 5-1).
       A surface combustion burner employs surface combustion or premixed natural gas and air on a
refractory material.  The burner radiates heat to an air-cooled firebox, and the combustion zone is
maintained below about 1,250 K (1,790°F).  NOX emissions of about 7 ng/J  (0.016 lb/106 Btu)  have
been reported with a prototype furnace using the surface combustor (Reference 5-13).  The surface
combustor is similar to some larger commercial systems which employ surface combustors.
       Another manufacturer's design incorporates a perforated burner. Natural  gas mixes with  air
through the perforations and combustion occurs at the burner surface.  Heat is transferred to a
glycol solution in small tubes imbedded in a fin arrangement surrounding  the burner.  The
                                                 5-8

-------
                                            TABLE 5-4.  PERFORMANCE SUMMARY OF LOW-NOV CONTROL EQUIPMENT FOR
                                                        NATURAL GAS-FIRED RESIDENTIALxHEATERSa
CONTROL
Conventional
Units
Radiant
Screens
Secondary Air
Baffles
Surface Combus-
tion Burner
Perforated
Burner
Modulating
Furnace
Pulse
Combustor
Catalytic
Combustor
AVERAGE
OPERATING
EXCESS AIR
(percent)
40-120
40-120
60-80
10
HA
NA
NA
HA
CYCLIC POLLUTANT EMISSIONS
ng/J HEAT INPUT
N0xb
28-45
15-18
22
7.5
7.7
25
10-20
< 5
CO
8.6-25
6.4
14
5.5-9.6
26
NA
NA
NA
UHCC
3.3-33
NA
NA
NA
, Nft
NA
NA
NA
STEADY STATE
EFFICIENCY
(percent)
70
75
NA
NA
85
75
95
90
CYCLE
EFFICIENCY
(percent)
60-65
70
NA
NA
80
70
.95 .
85
1978
INSTALLED
CONTROL
COST
d
NA
NA
$100-$200
$100-$300
over
conventional
furnace
$50-$250
over con-
ventional
furnace
$300-$600
$100-$250
COMMENTS
Costs include installation.
Emissions of CO and HC can
increase significantly if
screen is not placed properly
or deforms.
Requires careful installation.
Suited for single port upshot
burners.
IF
Not commercially available.
Still under development.
Commercially available design.
Spark ignited thus requires no
pilot.
Furnace 1s essentially derated.
It requires longer operation
to deliver a given heat load.
Currently being Investigated
by AGAL.
Still at the R&D stage.
  Reference 5-1.
b Sum of NO + N02  reported as N02.
c Unburned hydrocarbons calculated as methane  (CH4).
  Typical costs of uncontrolled unit $500-$800.
  NA = not available.

-------
TABLE 5
                                                          -5.  PERFORMANCE SUHHARY OF LOW-NOr COHTROL EQUIPMENT FOR
                                                               DISTILLATE OIL-FIRED RESIDENTIAL HEATERS3




CONTROL
Conventional
Units




Flame Reten-
tion Burner
Head

Controlled
Mixing
Burner Head



Integrated
Furnace Sys-
tem




"Blue flaw"
Burner/Furnace
System

Internal
Redrculatlon



AVERAGE
OPERATING
EXCESS AIR
(percent)
50-85




20-40



10-50

fc



20-30






20



10-15

"

CYCLIC POLLUTANT EMISSIONS
ng/0 HEAT INPUT


"*
37-85




26-88



34





19






10



10-25




CO
15-30




11-22



13





20






4.5-7.5



<30




UHCC
3.0-9.0




0.2-1.8



0.7-1.0





1.2






1.5-2.5



NA



Smoke
ttaufcer
3.2




2.0



<1.0





<1.0






zero



<1.0




Participate
7.6-30




NA



NA





NA






NA



NA





STEADY STATE
EFFICIENCY
(percent)
75




80-83
also depends
on heat
exchanger
. 80
also depends
on heat
exchanger


84






84



85





CYCLE
EFFICIENCY
(percent)
65-70




NA



NA





74






74



NA




1978
INSTALLED
COHTROL
COST
d




$52e



$43e





$250 over
conven-
tional
furnace



$100 over
conven-
tional
furnace
NA







COMMENTS
Range In NO emissions Is
for residential systems
not equipped with flame
retention burners.
Emissions for other
pollutants are averages.
If a new burner 1s needed
as well as a burner head,
the total cost would be
$385.
Cost of mass produced
burner head only about
$1.50. Combustible
emissions are relatively
low because hot firebox
was used.
Uses optimized burner
head. For new furnace
only. Combustible
"emissions are higher than
with burner head because
of quenching in air
cooled firebox.
New installation only.
Furnace is commercially
available.

Both for retrofit or new
Installations. Not yet
commercially available in
U.S.
(X
I

o
         Reference 5-1.

         Sum of NO and N02 reported as NOj.

       c Unburned hydrocarbons  calculated as methane.
         Typical  costs of uncontrolled unit $650-$!,000.

       e Original costs  reported  for years other than  1978 were-corrected for  Inflation using
         Gross National  Product (GNP) Implicit price inflators (Reference 1-9).
         NA - Not available

-------
glycol solution then transfers the heat to room air.  Ignition occurs by means of a spark rather
than by means of a standing pilot; therefore, seasonal fuel consumption is reduced compared with
pilot ignition furnaces.  Reported NO  emissions averaged 7.7 ng/J (0.018 lb/10  Btu)
(Reference 5-10).

       The modulating furnace is commercially available and differs from conventional units in that
the firing rate responds to heating load demand instead of cycling on and off.  The reported
emission rate for this furnace design is 25 ng/J (0.058 lb/10  Btu).  The American Gas Association
Laboratories attributes the lower emissions to the fact that the furnace is .essentially derated,
yielding a cooler flame and thus low NO  emissions.  However, the Gas Appliance Manufacturer
Association attributes the decreased emission rate to the single-port inshot burner that was used
(Reference 5-10).
       Pulse combustion involves combustion in a chamber fitted at one end with flapper valves and
at the other end with an open exhaust pipe.  Fuel and air entering through the flapper valves are
ignited with a spark.  The pressure resulting from combustion forces the flapper valves closed and
forces the product gases out the exhaust pipe.  The valves open again as the exhaust gases leaving
the chamber create a negative pressure.  Preliminary measurements of NO  emissions from pulse
                                                                       X
combustion of natural gas are about 20 ng/J (0.047 lb/10  Btu) (Reference 5-14).  There are no
marketed pulse combustion systems in use today.

       Catalytic residential combustors, although not available commercially, offer the potential
for very low NO  emissions with good combustion efficiency in residential furnaces.  The catalyst
promotes combustion at low temperatures so that thermal NO  formation is significantly reduced.
Catalytic combustors require large amounts of combustion air.  The use of a condensing residential
system in which the latent heat of vaporization is recovered is being considered for minimizing heat
losses due to high excess air levels (Reference 5-2).
Oil-Fired Equipment

       Essentially all new residential and commercial sized oil-fired furnaces and boilers are
equipped with flame retention burners.  Flame retention devices are generally desireable for all
conventional high pressure atomizing gun burners because they allow for operation at low excess air
levels and stay tuned longer.  However, laboratory experiments have shown that most flame retention
burners increase NOV emissions (References 5-15, 5-3).  In a test of ten commercially available high
                   A                 *
pressure atomizing oil burners, only one burner was found to reduce NO  emissions while  .
                                                  5-11

-------
also reducing smoke emissions.  NO  emissions were reduced from 37 to 26 ng/J (0.087 to
0.061 lb/106 Btu) (Reference 5-15).
       An advanced residential warm air oil furnace has been developed in an EPA-funded program
(References 5-16 and 5-17).  The integrated furnace system 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  NOX
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 commercial burners.
       In a study related to the development of the integrated furnace system, a controlled  mixing
burner head for retrofit application to residential oil heating equipment was developed
(Reference 5-16).  Laboratory testing indicated that the burner design is feasible  for
commercialization and can be retrofitted on existing residential space heating equipment.  The
burners operated successfully with long life potential.  Retrofit of the burner into standard
o1l-f1red furnaces would result in an estimated 20 percent decrease in NOV emissions with an
                                                                         A
accompanying increase in thermal efficiency of up to 5 percent (Reference 5-17).
       Another advanced burner/furnace design consists of a "blue flame" oil burner integrated  with
the firebox of a warm air furnace package (Reference 5-18).  Two sizes are currently available:
0.63 cm  oil/sec (0.6 gph), and 0.79 cm3 oil/sec (0.75 gph).  The efficiency of the burner is
reported to be about 84 percent and the NOV emission level is about 20 ppm.  This is a  significant
                                          X  •
Improvement over conventional systems for which typical efficiencies are 75-80 percent, and  NOV
                                                                                             X
emissions range from 70 to 90 ppm.  In the blue flame system combustion air and gases are
redrculated throughout the combustion chamber; the recirculation zone is designed  such that
blue-flame burnout of CO and organics results.  These systems are available as a single unit
(burner/furnace combinations) for new installations.  Retrofits to existing burners are not
practical since the blue flame burner must be matched to the firebox geometry and heat  transfer
characteristics (Reference 5-19).  The blue flame furnace system is the only commercially available
low NOX system in the U.S. (Reference 5-1).
                                                5-12

-------
       Another oil burner recirculates combustion gases in a manner similar to the blue flame*
system.  However, this burner recirculates combustion gases Internally; whereas, the blue flame
system uses external gas recirculatlon and requires an air tight combustion chamber to prevent air
leaks.  This Internal recirculatlon feature imiy permit retrofit Installation on oil-fired furnaces.
Estimates of NOX emissions from this burner are 10-25 ng/J (0.059 lb/10  Btu) (Reference 5-2).

Coal and Wood-Fired Equipment

       NO  emission control techniques for residential coal- and wood-fired heaters and boilers have
         A
not been widely investigated, primarily because of the declining use of this type of equipment in
the past.  Also, small coal-fired equipment is; not amenable to extensive modification for
controlling NO  emissions.  Excess air and ov«rf1re air Injection in some units are the only
feasible control alternatives which have some impact on overall NO  emissions.  However, excess air
                                                                  A
reduction in residential coal-fired equipment is very limited due to Increases in carbonaceous
emissions.  Overfire air injection, only available with larger commercial stokers, 1s only
moderately effective in reducing NOV (Reference 5-1).
                                   A

Commercial Equipment

       Application of control technology to commercial heating equipment has been very limited.
However,'the potential for applying some of the control techniques applied to residential systems
exists.  Compared to residential gas-fired equipment, a greater percentage of commercial warm air
heaters or duct heaters utilize power burners instead of naturally aspirated burners.  Power burners
generally have more flexibility for excess air control while maintaining low CO and VOC emissions
(Reference 5-1).  Furthermore, theoretical considerations indicate that the flame quenching and
surface combustor concepts of gas-fired residential burners could be Implemented for commercial
systems.  Application of control techniques similar to those for residential oil-burners may also be
possible for commercial oil-fired furnaces.  The EPA controlled mixing burner head design was found
to minimize NOX emissions from burners with oil flow capacity up to 13 ml/s (12 gph)
(Reference 5-1).
       Although NOX control techniques for small firetube boilers have not been widely investigated,
the similarity in equipment design between small firetube boilers and larger Industrial firetube
boilers may allow application of similar combustion modification control techniques.  NOX control
                                                5-13

-------
techniques investigated for industrial stoker coal-fired boilers could also be potentially
applicable to commercial size stokers (Reference 5-1).

5.1.3  Costs                                             »

       Table 5-6 summarizes estimated cost data for the most effective NOV control  alternatives  for
                                                                         A
residential heating systems.  The use of advanced low-NO  burner-furnace units for new sales  appears
to be the most attractive option for NOX control in space heating equipment.  The perforated
burner is presently considered the best available control technology for gas-fired residential and
commercial heaters (Reference 5-1).

       The blue flame furnace and the integrated furnace system are options for new oil-fired
furnace installations.  The blue flame unit has been commercially available since 1974 and has been
widely tested in field installations.  The integrated furnace system (Reference 5-16)  is undergoing
field demonstration preparatory to certification and potential commercialization.  With proper
maintenance, both units offer a NOV reduction potential of 50 percent or greater compared to
                                  A
conventional units.  Fuel savings of 5 percent or more, relative to standard units, are achievable
with these units.  Use of these low-NOv units in new houses and for replacement of obsolete
                                      A
conventional units in existing installations would yield a nationwide decrease in residential NO
emissions which would more than offset the potential emissions increase due to population growth for
several decades.

       For long term application to NOX control in new residential units, there is the additional
possibility of utilizing the alternate design concept of catalytic combustion.  This concept,
discussed in Section 3.5, offers the potential for extremely low levels of NO  (1-10 ppm) when
firing natural gas or distillate oils.  Catalytic combustion is still in the exploratory stage of
development and no reliable cost estimates are available for residential heating systems.
       As indicated in Table 5-6, retrofit of the controlled mixing burner head (EPA/Rocketdyne) for
existing residential oil-fired furnaces primises to be the most cost-effective approach for
achieving lower NO  emission levels.  However, these burners are not currently commercially
available (Reference 5-1).
       Furnace tuning and, if required, burner head replacement (conventional burner head) are
strongly recommended for reduction of carbon monoxide and smoke and for improving unit efficiency.
The impact on NOX is negligible, however.  Furnace tuning (cleaning, leak detection, sealing  and
burner adjustment) costs a minimum of $40 for the average residential unit.  Burner head retrofit
                                                5-14

-------
                                                         TABlf 5-6.  COST 1HPACT OF HO, COHTitOl ALTERNATIVES*
V
«*l*
(ft
CONTROL
Perforated 8um«r
Htidtriattng Furnace
Surface cw&ugtlon
Burner
pulse Combustion
Rumen *
Catalytic
Couinistion
fl«ri>er
flame Retention
Burner Head
Claud Retention
Burner
Controlled Mixing
Burrurf (lead
Integrated Furnaca
System
Blue Plane
FUEL
«»tural Gas
Natural Gas
Natural Gas
Natural GM
datura! Cas
Distillate Oil
distillate 01 1
Distillate OH
IHstltlate Oil
Distillate Oil
ACHIEVABLE HO LEVEL
ng/J USEfW. HEAf •
12
(339 ng/w3 fuel)
i 36
(920 ng/«3 fuel)
12
(3WI tig/a? g«S
21
(603 ng/»3 UBS)
Estimate S
(153 ng/M3 g*$)
50
(l.B g/Sg fuel)
SO
(1.8 k/kg fualj
45
(l.« 9/H fuel)
29
(0.7 B/kg/fuel)
SO
(0.7 a/fca fual)
1978
IHCflEtttHTM.
IHtfESnUNT CAST
$100-J300 ovur cost of
conventional furnace
$BO-$Z5U over cost of
conventional funitca
$100-$200 itver cost of
conventional furnace/
heater
$30fl-$600 aver cost of
conventional furnace/
heater
fl6Q-$2SO over vost of
conventional furnace/
heater
«2 -- retrofit
including instillation
|385 — retrofit of
reduced capacity burner
t« - retrofit •
including installation
9250 over cost of •
conventional furnace
$100 over cost of
conventional furnace
COST EFFECTIVENESS
l/Bfl/J1'
(Ib/UT Btu)
1.1 - 6.2
1.4 - /.O
1.7 - 3.4
6.1 - li.Z
2.3 - 3.9
2.6
12.8
1.3
4.2
1.7
ttHf&tCK PERIOD
BASED ON ANNUAL
FUEL BILL OF $500
1-3 years
1 ' 3.9 years
3.S - 8.U
1.7 - 3.5
1.4 - 2.3
Less than 1 .year
3.S years
Less titan 1 yettr
2.S years
1 year
OEVELOFff.HI
STATUS
Counerclally
available
CoflMrciolly
«vi liable
Hot coranercially
available
Not Rocwercialty
available
Hot coHiwrclally
available
Cowiercltflly
available
Cwnnerclally
available
Hot cowerc tally
available
Mot commercially
available
Ciwffierc tally
availaftle
        (Reference 5-1}.
      b Based oa uncontrolled Mix emissions of 'IS ng/J host output for natural ass-fired heaters and flO ng/.) heat output
        for distillate oil-fired heaters.  tost-e?fectlv«n«ss Is based on the differential investment cost of the control.
      c Based on installation of » condensing system whtsr» se&swta! efficiencies can be as high as 55 percent.
        Only one fla«e retention burner tested lowered fiOx emissions.

-------
replacement costs an additional $25 less installation.  These control  measures are usually cost
effective in view of the fuel savings and increased safety derived from the maintenance.

5.1.4  Energy and Environmental Impact

5.1.4.1  Energy Impact

       Both of the NOX emission reduction techniques (tuning, equipment) 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 equipment were replaced with new
designs.

5.1.4.2  Environmental Impact

       The effect of lower excess air on CO, VOC, and parti oil ate 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-NO  combustion conditions without increasing adverse
emissions of other species (Reference 5-17).  Table 5-7 shows a comparison of the emissions from
typical uncontrolled units and from a prototype unit with an optimized burner/firebox.  Incremental
emissions were held constant or reduced when using the low-NO  furnace.  Table 5-7 also shows
incremental emissions with a commercially available oil emulsifier burner.  Again, low-NO  operation
was achieved with no adverse effects on incremental emissions (Reference 5-20).

       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 furnaces, sulfates, trace metals and especially POM's could cause severe localized environ-
mental problems.  However, except for fuel switching, it is doubtful that NOV controls will be
                                                                            X
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 pSr-ticulate matter (smoke)
                                                5-16

-------
                       TABLE 5-7.   EFFECT OF LOW-NO  OPERATION  ON INCREMENTAL EMISSIONS

                                   AND SYSTEM PERFORMANCE  FOR RESIDENTIAL WARM AIR FURNACES
,
Typical uncontrolled
field units
(References 5-2,5-3)
Optimum low-NO unit
J\
(Reference 5-17)
Water/distillate oil
emulsifier burner:
(Reference 5-20)
Excess
Air
90%
15%
32%
Thermal
Efficiency
(Steady-State)
70*
802
802
NO
g/kg fuel
1.1 - 2.7
0.6
0.85
CO
g/kg fuel
1.05
1.0
0.3
HC
g/kg fuel
0.1
0.1
iM«l
Smoke
Bacharach
3.2
1
-1
01
I

-------
(References 5-2 and 5-3).  The effect of NOV controls on transient emissions has-not been widely
                                           A
studied.  Incremental steady-state emissions must eventually be weighed against the  transient
emissions for this significance to be shown.

       Comparative 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 NOX controls.   Additionally, it
should be emphasized that the incremental emissions data shown in Table 5-7 are for  well-maintained
laboratory operation.  Data are needed on long-term field operation with NO  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 5-21),  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.R kg (10 pounds)  per day
(Reference 5-21).

       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 underflre air and  waste on  the
grates in the primary chamber.  Additional air is admitted above the burning waste to promote
burnout of the primary combustion products.

       Incinerators are used in a variety of applications.  The main ones are municipal and
industrial 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).
                                                5-18

-------
5.2.1.1  Emissions
       Nationwide NO  emissions from incineration in 1974 amounted to 39 Gg per year (43,400 tons
per year) which is 0.3 percent of the total NOX 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 5-22).  AP-42 (Reference 5-23) specifies the
same value for multichamber industrial and municipal incinerators.  For single chamber industrial
incinerators, a lower factor.of 1 g NOg/ kg refuse (2 Ib/ton) is.specified.
       Stenberg, et al., conducted field tests to study the effects of excess combustion air on NO
                  '" •  --"                                                                            X
                                *
emissions from municipal incinerators (Reference 5-24).  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 NOX 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 .
                                                                                          A
Activity in pollution abatement for incinerators to date has focused on particulate control rather
than NOX.

5.2.1.2  Control Techniques
       The use of waste disposal methods other than combustion may be the most likely means for
reducing NO  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
                                              5-19

-------
   0.6
   0.5
   0.4
•I

ec*
o
x
01
LU
O
cc
11J
cu

g
o
111

si! 0.3

cc
a
 x-
a
st
   0.2
   0.1
                                                O FURNACE - BEFORE SCRUBBER

                                                    NOX = 0.081 + 0.00144 (PERCENT EXCESS)

                                                O STACK   - 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 5-24).
                                        5-20

-------
                                NOX = 0.365 • 0.00183 (PERCENT UNDERF1RE AIR)
                            40             60

                           UNOERFIRE AIR, percent
Figure 5-3.
Effect of  underfire air on NO  emissions from  a
227 Mg (250  ton) per day continuous feed incinerator
(Reference 5-24).
                                 5-21

-------
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 1s obviously very Inexpensive, but extremely objectionable and offensive in and near popu-
lated areas.
       Sanitary landfills nay be alternatives,  to the extent that land usable for this  purpose
Is available.  Approximately 1233 m3 (1 acre-foot} of land 1s required per 1000 persons per year
of operation for a waste production of 2 kg  (4.5 pounds) per day per capita (Reference 5-25).   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
                        i
       At present, gaseous emission controls are not applied to Incinerators.  As described earlier,
only particulate emission controls are employed.  Reference 5-26 presents estimated construction
costs in 1966 and operating costs for particulate pollution control.
5.2.2  OpenBurning

       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
druas or boskets, 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 fronr open burning vary from year to year, and the data for the various
sources are not entirely consistent.  Table 5-a shows the estimated NOX emissions from open burning
                                                5-22

-------
sources for 1971 as reported in Reference 5-27.  More recent estimates from the 1976 NEDS data file
                                                                                                  +•
and Reference 5-28 are also given in Table 5-8.  Increasing awareness of air pollution problems
has contributed to a general decline in the quantity burned (and thus the emissions) from tnose
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 So (32,000 tons) in
1969 to 13 Sg (14,300 tons) for 1973 {Reference 5-23).


                  TABLE 5-8.  ANNUAL EMISSIQHS OF NITROGEN OXIDES FROM OPEH BURNING
Source
Solid Waste Disposal
Forest Wildfires
Prescribed Burning
Agricultural Surning
Coal Refuse Fires
Structural Fires
NOX Emissions
1971, .
Reference 5-27
So.
ISO
138
19
29
31
6
1C3 Tans
16S
152
21
32
34
7
1S7S NEDS
Ss
55
48
30
13a
S3
5
10s Tons
105
53
33
14a
58
6
                       a1973 estimate from  Reference  5-28.
5.2.2.2  Control Techniques
Solid Waste Disposal
*                                       •                         •          '               •
       Frora the standpoint of air pollution, sanitary landfills are alternatives ta open burn-
Ing.  In addition i» the land necessary for sanitary landfill. cover material approximating 20
percent by volume of tfte compact waste is required.  The availability of caver material nay 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 5-29 reports that in a pilot project the refuse Is shredded and baled for
loading on rail cars for shipment to abandoned strip urine landfill sites.
       Other noncoabustion alternatives may have application fn some localities.  Composting is
now being tested on a practical scale (Reference 5-30).  Hog feeding has been used for disposal
of cartage.  Dumping at sea has been practiced by some seaesast cities, but is now extensively regulated.
                                                 S-23

-------
       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 5-31).
Forest Wildfires
                                                                    12 '^
       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.
       The U.S. Forest Service estimated that 2.06 x TO10 m2 (5.11 million acres) of land were burned
1n 1976 (the World Almanac, 1978).  A similar estimate for 1971 (Reference 5-27) was 1.73 x TO10 m2
(4.28 minion 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 5-32):
         •  Slash disposal resulting from forest harvesting operations
         •  Forest management operations for forest floor fuel reduction, seedbed preparation, pest
            control, forest thinning and undergrowth control
         »  Public works construction operations to clear reservoir and dam-sites, utility and high-
            way rights-of-way and building and structure site areas
                                                 5-24

-------
         «   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 NO  and other emissions  such  as particulate
 matter,  organics, odorous compounds, and carbon monoxide; or  it may be  possible  to develop equipment
 that cart 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
 up6n 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  5-28):
       r   Removal and disposal  of  residue at low cost
       •   Preparation of farmlands for cultivation
       «   Clearing  to facilitate harvest
       •   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.
                                                 5-25

-------
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
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 5-27).  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 two1 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.  NOX emissions and control techniques for. the latter equipment types have been de-
scribed in previous sections of this document.  The former equipment types are described 1n 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 5-33).
       There is currently very little application of NOX control to industrial process heating equip-
ment.  Consequently there are very few data on MOX 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 NOX control in a
"diversity of process furnaces, ovens, kilns, and heaters.  Partial results from that study are given
in Reference 5-34 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.
                                                5-26

-------
5.3.1  Petroleum, Chemicals, and Natural  Gas;
     Petroleum refining is the process of converting crude oil  into salable products.   This con-
version into salable products is accomplished in various operations which require the feedstock to
be at elevated temperatures.  Typically,  feedstocks are raised  to the required temperature in a
fired heater.  Many chemical manufacturing plants, natural gas  plants, and pipeline stations have
fired heaters using the same general design principles.  Methods used to control  NOX emissions from
fired heaters used in these industry groups are similar and therefore will be discussed together.
Another source of NO  emissions at petroleum refineries that will be discussed in this section is
catalytic crackers and CO boilers.

5.3.1.1  Fired Heaters                                               \

5.3.1.1.1  .Process Description
     Petroleum Refining and Chemical Manufacturing - Process emissions of nitrogen oxides from the
petroleum refining and chemical, manufacturing industries are produced primarily by fired heaters.
Fired heaters transfer heat, that is liberated by the' cumbustion of fossil fuels, to fluid contained
in tubular coils.  Industrial processes usually use fired heaters when fluid temperature require-
ments are above 204°C (400°F).  The fluids include any gas or liquid, with the exception of liquid
water when it is used for the generation of hot water or steam.  Industry also refers to fired
heaters as process heaters, process furnaces, and direct fired heaters.

     Typical applications of fired heaters include heating of oil or other heat transfer fluid,
steam superheating, distillation, thermal cracking, coking, pyrolysis, and reforming.  Although
there are some applications of fired heaters in other industries, the dominant use is in the
petroleum refining and chemical manufacturing industries.

     Petroleum refineries (SIC 2911) process crude oil to form products like transportation fuels,
heating fuels, lubricating oils, and chemical feedstocks.  In 1981, there were 303 refineries
                                                          9                     6
operating in the U.S. with a capacity to process 2.93 x 10  liters/d (18.45 x 10  barrels/day) of
crude oil (Reference 5-35).  These refineries range in size from 6.36 x 10* liters/day (400 b/d) to
         8
1.02 x 10 . liters/day (640,000 b/d)(Reference 5-35).  Forty-one states had at least one refinery.
Approximately 56 percent of the industry capacity is located in Texas, California, and Louisiana.
                                                 5-27

-------
               TABLE 5-9.   MAJOR REFINERY  PROCESSES REQUIRING A FIRED HEATER
Process
Distillation
Atmospheric
Vacuum
Themal Processes
Theresa! Cracking
Cokl ng
₯i streaking
Catalytic Cracking
Fluldlred
Catalytic
Cracking
Catalytic
Hydrocracfcing
Hydroprocessing
Hydrodesulfur-
Izatlon
Hydrotr eating
Hydroconverslon
Alkylation
Catalytic
Reforming
Process Description

Separate light hydrocarbons
from crude in a distillation
column under atmospheric
conditions.
Separates heavy gas oils
from atmospheric distillation
bottoms under vacuum.

Thermal decomposition of
large molecules into lighter,
more valuable products.
Cracking reactions allowed to
go to completion. Lighter
products and coke produced.
Mild cracking of residuals
to Improve their viscosity
and produce lighter gas oils.
Cracking of heavy petroleum
products. A catalyst is used
to aid the reaction.
Cracking heavy feedstocks
to produce lighter products
in the presence of hydrogen
and a catalyst.

Remove contaminating metals,
sulfur, and nitrogen from
the feedstock. Hydrogen is
added and reacted over a
catalyst.
Less severe than hydrodesulfur-
Izatlon. Removes metals,
nitrogen, and sulfur from
lighter feedstocks. Hydrogen
is added and reacted over a
catalyst.
Combination of two hydrocarbons
to produce a higher molecular
weight hydrocarbon. Heater
used on the fractionator.
Low octane napthas are
converted to high octane,
aromatic napthas. Feedstock
Is contacted with hydrogen
over a catalyst.
Process Heatd
Heaters • Requirements
Used kj/liter (103 3tu/b) feed

Preheater, 511 (77)
Reboiler
Preheater, 418 (63)
Reboiler

Fired 4,648 (700)
Reactor
Preheater 1,460 (220)
Fired 1,328 (200)
Reactor
Preheater 385 (58)
Preheater 1,262 (190)

Preheater 465 (70)a
Preheater 239-498 (36-75)
Reboiler 1,992-7,304 (300-1, 100)*
Preheater 2,258 (340)
Feedstock Temperature
Outlet of Heater
°C (°F)

371 (700)
399-443 (750-830)

454-538 (850-1000)
482-524 (900-975)
454-510 (850-950)
e, f
316-474 (600-385)
204-454 (400-850)

199-427 (390-800)
316-427 (600-800)
-204 (-400)
454-538 (850-1000)
'heavy gas oils and middle distillates
b!1ght distillate
           of total alfcylate
'erence 5-36
dRefi
Reference 5-39
 Reference 5-40
                                                    5-28

-------
      In general,  each process in a refinery requires at least one fired heater.   All  refineries have
 atmospheric and vacuum distillation units to separate the lighter products from crude oil.   The
 number and type of downstream processes depend on the type of crude processed and the products
 desired.  For example, a complex refinery that takes heavy crude oil  and produces unleaded  gasoline
 and lighter oils  will require processes such as cracking, hydrotreating, reforming, and alkylation.
 A refinery of this type can require as many as 100 heaters (Reference 5-36).   A small, simple
 refinery (topping refinery) whose main operation is to separate the crude oil into its major
 fractions may have 10 heaters.  Table 5-9 presents the major refining processes along with  a brief
 description of each process, the types of heaters it requires, and its fired  heat and temperature
 requirements.

      The heaters  for the refinery processes are usually either preheaters, fired reactors,  or
. reboilers.  Hot oil furnaces can be used as preheaters or reboilers.  Circulating heaters that heat
 the heavy crude oils to decrease the viscosity and improve flow through pipes are also used.  In
 refinery fired heaters, the feedstock usually flows through the radiant tube coils only once.  The
 heaters require an even heat distribution to prevent coking wi-thin the tubes  and to control feed-
 stock temperature.

      The total refining industry fired heater energy requirements can be estimated by assuming 75
 percent of the total fuel consumed by the refining industry is used by fired heaters (Reference
 5-37).  For 1979, the fired heater energy requirements are estimated to be 2300 PJ/yr (2.18 x 1015
 Btu/yr).  This calculated value compares well with other reports which show a range of fired heater
 energy requirements of 1910 to 2330 PJ/yr (1.131 - 2.21 x 10   Btu/yr)(Reference 5-37).  In  a report
 published by the American Petroleum Institute, the estimated number of fired heaters used in the
 petroleum refining industry in 1977 was 3240(Reference 5-38).

      Chemical Manufacturing Industry - In the chemical manufacturing industry (SIC 28), tubular
 fired heaters are used to drive endothermic reactions such as natural gas reforming and thermal
 cracking.  They are also used as preheaters to raise the feedstock temperature to a certain
 temperature to control a reaction and as reboilers in some distillation processes.  When a  narrow
 temperature range is required, hot oil furnaces are preferred because they allow better temperature
 control.
                                                   5-29

-------
     The chemical industry fired heater applications are similar to those used in the refining
industry.  Fired heaters are used when the feedstock temperature requirements are above 177-204°C
(350-400°F).  The predominant heater application is as fired reactors.

     Ten chemicals have been identified in the chemical manufacturing industry as major users of
fired heaters (Reference 5-37).  Table 5-10 presents these chemicals along with their manufacturing
process, temperature requirements, and the type of heaters they use.  Of these ten chemicals, all
but one, ammonia, are organic chemicals.  In 1980, about 38.5 Gg (85,000 x 103 Ibs)  of these nine
organic chemicals were produced.  Seven of these nine chemicals are among the top 50 produced
chemicals in the U.S.  Also in 1980, 17.3 Gg (38,100 x 10  Ibs) of ammonia were produced.   Ammonia
1s the second largest volume chemical in the chemical industry (Reference 5-41).   Other smaller
volume organic chemicals may use tubular fired heaters.  However, the ten chemicals  presented
represent a large portion of the chemical industry fired heater energy requirements.  In addition,
all of the basic heater applications expected to be used in the chemical industry are represented by
these ten chemicals.  Therefore, it is believed that these ten chemicals are representative of the
fired heater population for the chemical industry.  As in refineries, a major portion of the
production capacity is located in Texas and Louisiana.

     An estimate of the total chemical industry heater population can be based on process  heat
requirements.  Thirty percent of the total fuels consumed by the chemical industry is required for
process heating (Reference 5-37).  From this the process heat energy requirements are estimated as
approximately 779 PJ/yr (7.38 x 10   Btu/yr).  However, this estimate of process  heat requirements
may include other types of process heat besides tubular fired heaters such as dryers, kilns, and
roasters.  Based on the available data, these other types of process heat account for approximately
14 percent of the total process heat energy requirements (Reference 5-42).  Therefore, 86  percent of
the total process heat energy requirements is assumed to be met by fired heaters, or 670 PJ/yr (6.35
x 1014 Btu/yr).

     Fired Heater Design - The design of fired heaters can vary depending on the  heater application
and client preference.  Industry uses fired heaters in a variety of applications. Table 5-11
describes the common applications of fired heaters used in the chemical  manufacturing and  refining
industries.  The following design classifications distinguish the various heater  designs from each
other:
                                                 5-30

-------
                                             TABLE  5-10.  TYPICAL FIRED HEATER APPLICATIONS  IN  THE CHEMICAL INDUSTRY
01
I

Chemical
Ethyl ene/.
Propyl ene
Butadiene
Ethanolf
Benzene
Dimethyl
Styrene6
b e
Ammonia '
Methanol9
Process
Naptha Cracking
Dehydrogenation
Ethyl ene Hydra -
tion
Extraction of
Reformate
Formed from
D-Xylene and
Methanol
Dehydrogenation
of Ethylbenzene
Natural Gas'
Ref ormi ng
Hydrocarbon
Reforming
Heater Type
Fired Reactor-
Thermal Cracking
Preheater and
Reboiler
Preheater
Reboiler
Preheater-llot
Oil Furnace
Fired Reactor-
Steam Hydro-
carbon Reformer
Fired Reactor-
Steam Hydro-
carbon Reformer
Fired Reactor-
Steam Hydro-
carbon Reformer
Feedstock
Temperature
°C (°F)
800-900 (1,472-1,652)
425 (800)
750 (1,382)
375 (707)
250-280 (482-536)
630-710 (1,166-1,310)
800 (1,472)
840-900 (1,544-1,652)
Heat Requirements8 >c
PJ/yr 10" Btu/yr
(205.16)
(3.58)
(1.76)
. (47.30)
(4.11)
(20.93)
(146.65)
(17.54)
194.46
3.39
1,67
44,83
3.90
19.84
139.0
16.63
                              b
 All  1971 data, except ammonia
 1975 data
Reference 5-42
^Reference 5-43
 Reference 5-44
                               Reference 5-45
                              ^Reference 5-46.

-------
                                                                  TABLE  5-11.   BASIC FIRED HEATER APPLICATIONS
en
i
to
CO
                           Type
                      Fired Reactor
                             Description
 Outlet Feedstock
 Temperature *C (*F)
                                                                                                              Firebox Temperature6
                      Beboller            He»ts liquid charge stock from
                                          »  fractionating column before
                                          returning the liquid to the
                                          column.
 Fractionating-     Heats  charge stock before it
  colum feed       enters a distillation colum.
  preheater

 Reactor feed       Raises the  temperature of the
  preheater         feedstock to control a reaction
                    taking place in an adjacent
                    process.

Circulation         Heats  feedstocks to lower their
 Heaterviscosity thus improving pumping
                    and flow through' pipes.

                    Prevents condensation when gases
                    are transferred.
                       Steam/hydrocarbon Heats  the  feedstock along with
                        reformer heater  steam  to drive an endotheraic
                                         reaction.   The reaction usually
                                         takes  place over a catalyst.

                       Pyrolysls heater  Keats  the  feedstock for thermal
                                         decomposition within the heater
                                         tubes.
                                          ileats a recirculating medium which
                                          transfers heat to a feedstock.
                                          Used when a narrow temperature range
                                          is  required.
204-288   (400-550)







371-443   (700-830)* '



  538      (1000)U

  399      (750)b


< 260     (<500)



< 260     (<500)



                     d
788-899 (1.450-1,650)




815-899 (l,500-l,650)d







177-399   (350-750)e
                                                                                         538-982   (1000-1800)   Charge stock from ataospherlc or
                                                                                                                 vacuum distillation columns  In a
                                                                                                                 refinery or a fractionating  colum In
                                                                                                                 organic chemical Manufacturing.
                                                                                                             871-1,093   (1600-2000)   Heats feedstock  before  it enters
                                                                                                                                       ataospheric or vacuum distillation
                                                                                                                                       columns.
                                                                                                           871-1,093   (1600-2000)
                          Refinery catalytic reforming

                          Ethylene hydration to fom ethanol.
                                                                                                            538-982    (1000-1800)     Crude oil  heating  in  a  refinery to
                                                                                                                                       improve flows.
538-982    (1000-1800)      Natural gas processing plants.
                                                                                      1,093-1,427   (2000-2600)   Natural  gas reforming  to yield
                                                                                                                 hydrogen for ammonia synthesis.
                                                                                      1.093-1.427   (2000-2600)   Thermal cracking In refineries  to
                                                                                                                 produce lighter petroleum products
                                                                                                                 and cracking of natural  gas  liquids
                                                                                                                 and petroleum feedstocks in  organic
                                                                                                                 chemical manufacturing to produce
                                                                                                                 olefins.

                                                                                       538-982     (1000-1800)   Used as a reboiler In benzene
                                                                                                                 extraction from refinery feedstocks.
                       Reference 5-39

                      Reference 5-46

                      cReference 5-48
                         Reference  5-49

                        "Reference  5-50

-------
               Radiant and convective tube coil orientation
               Draft type
               Use of preheated combustion air
               Feedstock temperature
               Fuel types used
               Burner location
               Burner type

     Many designs of fired heaters are available.  Some examples of the variety of designs available
 are presented  in Figure 5-4.  All fired heaters have a radiant  section and the majority have a
 convection  section.  The radiant section  is  located within the  firebox and contains the burners and
 a  single  row of tubular coils.  The primary  heating of the feedstocks occurs within the radiant
.section.  As the name implies,  radiation  is  the primary method  of  heat transfer.

     The  tube  coil in the rad.iant section consists of a number  of  tubes connected in  series by 180
 degree  return  bends (Reference  5-47).  Each  set of consecutive  tubes is considered a  "pass" or
 parallel  stream.' The inlet feedstock stream can make one pass  or  can be  separated into a number of
 passes. Tube diameter can vary, but an average diameter is around  10.2 cm (4 inches)(References 5-52,
 5-53).  The spacing between the tubes and the distance of the tube coil from the refractory walls
 depends on  the tube diameter.   Spacing between the tubes usually ranges from 1.5 to 3 diameters
 (Reference  5-53).  Increasing the tube to wall clearance improves  heat flux to the tube until the,
 distance  reaches about two tube diameters (Reference 5-54)., The walls are lined with an insulated
 material  such  as insulated firebrick, castable refractory, or ceramic fiber.  The insulation
 protects  the steel structure from overheating and flue gas corrosion.  In addition, the insulation
 minimizes heat loss and keeps the firebox at.a high temperature by reradiating heat to the tube
 coils  (Reference 5-47).

     The  convection section is  located after the radiant section and also contains a  set of tubes.
 The  convection section recovers the residual  heat of the flue gas  before  it goes to the stack.  The
 temperature of the flue gas leaving the  radiant section usually ranges from 816- 982°C (1,500-
 1,800°F)(Reference 5-52).  The  first  few  rows of tubes, called  shield tubes, are .subject to some
 radiant heat transfer.   In most heaters,  the feedstock flows through the  convection section to
 preheat it  before flowing to the  radiant  section (Reference 5-52).  Some  convection sections are
 also used to generate steam.  Convection  sections can improve heater efficiency particularly if
                                             5-33

-------
                              Canwedon coil
                               llint coil'
                                                                         Convection coil-.
                                                                        n«dtant ceil
        Artor or wickwtyp*
V«ro'«aJ-tub«, linglarow, douWe-fir»d
Vertical-cylindrical with eranflow-
     convection saction
        . Sw*.ffr»d box                     Cabin                               Tw-xail box


Figure 5-4.   Examples  of  fired heater designs  (Reference 5-51).
                                                 5-34

-------
they have extended surface area.   Many older heaters have bare convection section tubes and operate
at only 65 to 70 percent efficiency (References 5-37, 5-55).  New extended surface tubes can improve
efficiency by 10 percent and reduce flue gas temperature by 149°C (300°F).  Fins or studs are
usually used to extend the tube surface area.  Newer heaters can operate at around 90 percent
efficiency.  The lower limit of exit flue gas temperatures is around 149-177°C (300-350°F)
(Reference 5-55).  Because many flue gases contain SO,, temperatures below this lower limit will
cause corrosion problems due to sulfuric add condensation.  Typical exit temperatures are approxi-
mately 371°C (700°F).

     Two basic draft types are available to supply combustion air and to remove flue gas.  These are
natural-draft and mechanical-draft.  Approximately 90 percent of all gas-fired heaters and 76
percent of all oil-fired heaters have natural-draft (Reference 5-51).  Natural-draft heaters rely on
the natural stack effect to remove flue gas and induce the flow of. combustion air into the firebox.
The natural stack effect, maintains a negative pressure within the fired heater.  The negative
pressure prevents leaking of flue gases and overheating of the heater structure.

                                  •
     In a mechanical-draft heater, a fan supplies the combustion air and removes flue gas.  A
mechanical-draft heater can use either an induced- draft, forced-draft, or induced-draft/
forced-draft (balanced draft) design.  An induced-draft heater uses an induced-draft fan located
above the convection section and before the atack to induce the flow of combustion air and remove
flue gas.  The fan also maintains a negative pressure in the fired heater.  A forced-draft heater
uses a forcedrdraft fan to supply combustion air under positive pressure.  Although the combustion
air is under positive pressure, the firebox still remains under negative pressure.  The negative
pressure occurs because the flue gas is removed, as in a natural-draft heater, by the stack effect.
A balanced draft heater uses a forced-draft fan to supply combustion air and an induced- draft fan
to maintain a negative pressure within the heater and remove flue gas.

     Typical heaters have a negative pressure of 0.007 - 0.01 kPa (0.03 - 0.05 in H20)(References
5-52, 5-54).  Typical stack flue gas velocities and mass flowrates range from 7.6 to 12.2 m/sec
(25-40 ft/sec) and 3.7 to 4.9 kg/s m2 (0.75 -1.0'lb/s ft2),  respectively (References 5-52, 5-54,
5-56).                                                                                     .

     Combustion air preheaters are often-used to improve the efficiency of a fired heater.  The
maximum thermal efficiency obtainable with current air preheat equipment is 92 percent
                                                  5-35

-------
(LHV)(Reference 5-55).  In the preheater, heat is transferred from the flue gas to the combustion
air.  Therefore, less heat is required to heat the combustion air which allows  a greater proportion
of the total heat released to be absorbed in the radiant section.  And less fuel is required to
reach the required combustion temperature.  In addition, the preheater raises the adiabatic flame
temperature above that of ambient air heaters.  The trend  in  the 1980's will be to apply air
preheaters to larger sources because of improved fuel  efficiency (Reference 5-57).  Because of  the
lower density of the heated combustion air, all heaters using air preheaters will  also require  a
fan.

     The temperature requirement of the feedstock is an important factor in heater application  and
design.  Typical feedstock temperature requirements for refinery and chemical manufacturing
processes are included in Tables 5-9 and 5-10.  Feedstock temperature can  influence the number  and
spacing of tubes and the mass velocity of the feedstock within the tube.   The mass velocity of  the
feedstock determines the tube'size and number of passes.  Mass velocities  usually range from 222  to
1,972 kg/s m2 (45 - 400 Ib/s ft2)(Reference 5-52)  In addition, heat transfer rate and firebox
temperature will be determined by the feedstock temperature requirement.   Heat transfer rates range
from 85 to 160 MJ/hr m2 (7,500 - 14,000 Btu/hr ft2)(Reference 5-52).   Firebox temperatures are
included in Table 5-11 and range from 538 to 1427°C (1000-2600°F)..
     The lowest temperature fired heaters, less than 260°C (500°F) feedstock temperature,  are used
in natural gas processing plants to prevent condensation.  These heaters average only about 4.4 MW
(15 x 106 Btu/hr) fuel input and can be as low as 0.15 MW (0.51  x 106 Btu/hr)(Reference 5-38). The
lowest temperature heaters in the refining and chemical manufacturing industries are the circulation
heaters, while the highest are fired reactors.

     Fired heaters can use a variety of fuels.  In general, the  chemical manufacturing and refining
industries use oil or gas.  These heaters also burn a wide variety of waste liquids or gases that
are not usually considered as fuels.  Other fuels such as coal and petroleum coke can be used, but
they are not expected to make a significant contribution to the  fuels used during the 1980's.  Both
industries use more off gases and natural gas than oil.  However, the refining industry burns more
off gases than the chemical industry.  Off gases are a by-product from manufacturing processes.
They can be made up of a variety of components whose relative composition can vary considerably over
time.  Heating values of refinery off gases can range from under 100 Btu per cubic foot to over 3000
Btu per cubic foot (Reference 5-58).   It is expected that in the future, refineries will  use more
                                               5-36

-------
refinery off gas and heavier fuels because of the higher market demands for lighter fuels.   The
chemical industry will  probably continue to use primarily natural  gas.

     The burners in a fired heater can be arranged to fire from the top, bottom, or sides of the
heater.  Most heaters fire from the bottom or sides because of the design simplicity and efficiency.
Some high temperature specialty units such as pyrolysis heaters and steam-hydrocarbon reformer
heaters are designed to use many small radiant-wall burners to heat the refractory surface.   Other
designs of these specialty units include the use of bottom fired burners near the wall  to heat the
refractory, a combination of bottom fired and radiant-wall burners to heat the refractory,  and
bottom fired burners midway between the refractory and the process tubes.

     Many different types of burners are available for fired heaters.  Burners can be differentiated
by their flame shape, method of mixing of fuel and air, atomization type, and draft type.  The
primary objective of a burner is to mix the fuel and oxygen before and during ignition.

     The application of the heater and the temperature requirements affect the type of burner
selected.  For example, some processes will require a more even heat distribution or more intense
heat.  One of the important operating features of a burner that is considered in the heater design
is the flame type.  The type of flame will determine heat intensity and heat distribution.
Important flame properties are shape and heat flux rate.  A burner must be able to maintain a
flame-stable operation during a wide range of operating variables and provide a reasonable  flame
shape when fuel.and air input varies (Reference 5-58).. Flame shape is affected by the manner in
which the fuel and air are introduced.  The length of the flame is affected by the type and amount
of fuel, combustion air temperature, register draft loss, and excess air rate (Reference 5-59).  The
flame shape must also comply with the mechanical configuration of the radiant section so that the
flame or hot gasses do not impinge on the tubes.  It should also provide an even heat distribution.
Typical flame shapes are flat and conical.  Flat flame burners can produce short, wide flame
patterns or longer patterns with less spread.  Conical shaped flames can have short flames  or
extremely long flames.  Short flames usually are high intensity while long flames provide a uniform
heat flux rate throughout the radiant section.  The heat flux rate is usually determined by the
process and heater design (Reference 5-58).

     Burners can be designed to burn gas, oil, or a combination of gas and oil.  Gas-fired  burners
are simpler in design and operation than oil-fired burners.  The basic gas burner classifications
                                                 5-37

-------
are prenrfx insoirating and raw gas burning.  In the prenrix burner (Figure S-5a), primary air 1s
mixed with the fuel gas prior to ignition at the burner tip.  Induction and mixing of primary air
occurs due to the kinetic energy of the fuel gas as it expands through orifices in the burner.
Secondary air is usually required to allow for swings in fuel calorific values and to complete
combustion.  If fuel calorific value is very constant, linearity of excess air levels can be
maintained from 33 percent to 100 percent of firing capacity by inspiratlng all primary air
(Referencs 5-56).  As swings in calorific value of the fuel increase, the use of secondary air must
be increased to permit excess air linearity.  However, as secondary *ir use increases, the range of
firing rates where excess air levels can ba maintained without burner register adjustments
decreases.  When the firing rate is below the range the burner is capable of maintaining excess air
linearity, excess air levels Increase significantly as firing rate is decreased.  Generally, anout
50 to 60 percent of the combustion air is mixed as primary air in the burner.  Premix burners
rsquire a fuel gas pressure of greater than 170 kPa (10 psig) to maintain combustion air mixing.  An
advantage of prenrix burners is that they can operate with high or low excess air rates
(Reference 5-47).

     Raw gas burners (Figure 5-5b) receive fuel gas without any prenrlxing of combustion air.  Mixing
occurs in the combustion rone at the burner tip.  The tip has a series of small ports to aid mixing.
The raw gas burner can handle large turndown ratios for a given combustion condition; however,
combustion air adjustments must be made over the full burner operating range (Reference 5-47).

     Oil-fired burners  (Figure 5-5) are classified by the type of fuel atomization.  Oil  is
atomized to improve the mixing of fuel and combustion afr.   Atomization methods commonly  used are
Htechanfcal, air, or steam.  Steam atomization is usually preferred because it is mor« economical,
controls the flame better,' and can handle larger turndown ratios.  Typical steam requirements are
0.07 - 0.16 kg steam/kg of oil (0.07 - 0.16 Ib steam/'lb of oil).   The main requirements for the
steam are that it must be dry and  available at a constant  pressure,  approximately 791 kPa  (100 psig).
As in premix burners, cil-Fired  burners  utilize primary  and secondary air to  improve  heater efficiency
(Reference 5-47).

     Combination burners (Figure S-5d) are designed to bum all  gas, all  oil  or any combination  of
oil and gas.  A typical arrangement is to have a single oil gun in the center of an array of raw gas
nozzles.  The air for oil or gas firing can be controlled separately in these units.  Therefore,
                                              5-38

-------
                           .Burner Tip
                      Primary Air
                      Register
                                          fcfractnry
                                                                                           Register-
                                                                                           Assembly
                                                                                      Inspiriting
                                                                                      Qas Pilot
                                                                               Full Gas
                                                                               :ntet
a.   Premix Burner
b.   Raw Gas Burner
                                     HUfft*
                                     Block
                                  Air Shutter
                                            Sas Burner
                                            Ajieably
                             •Oil ami Stea Inlet   'Primary Air
                             InUt              Control
                                            Fuel 011
                                            Burner
                                                                                             .^luffl
                                                                                                   t Block
                                                                                             Secondary Air
                                                                                            Control
                                                                                       .Rax tnlet
                            . 1 And Steam
                            Inlet
  c.   Oil Burner                                 d.   Combination Oil and  Gas  Burner

       Figure 5-5.   Typical  burners  by  type  of  fuel  burned.
                                    5-39

-------
these burners can control primary air for oil combustion and secondary air for fuel-gas combustion
(Reference 5-47).

     High intensity burners are generally used when a heavy fuel,or when a fuel that has a large
quantity of inerts, is to be burned.  High intensity burners usually have a large, cyclindrical  .
refractory lined combustion chamber.  The mixing of fuel and combustion gases is very intense in
this chamber.  Combustion is fully established in the chamber but not completed.  Because of cir-
culation patterns present in the chamber, flames with stable and controlled size and shape are
                     f
produced at relatively low excess air.  The fuel gases from the high intensity combustion are
expelled at a high velocity and temperature.  This produces a uniform temperature profile in the
firebox.  High intensity burners require forced-draft air supplied at relatively high pressures.
 *
These air pressures can range from 1.5 to 5 kPa (5-20 in HgOHReference 5-47).

     The burner type and design are affected by the type of draft.  In the past few years the trend
has been towards forced-draft burners because they are more fuel efficient.  Forced-draft burners
can control airflow, therefore, they can be operated at lower excess air, control flame shape
better, control fuel/air ratio, and be used with air preheaters (Reference 5-54). They can also
cleanly combust fuels with low or varying heat values at low excess air rates.  Having the
capability to control or use each of these features allows the heater to be operated more
efficiently.  In addition forced-draft burners have a larger capacity per burner; therefore, less.
burners are required which reduces cleaning, maintenance costs, and makes them more amenable to
automatic control (Reference 5-60).

5.3.1.1.2  Emissions and Control Techniques

     Presently there are a wide range of control techniques that have been studied or demonstrated
on fired heaters.  These techniques include, operational changes, combustion modification,, flue gas
treatment, or a combination of these techniques since they are usually more effective  than  flue
gas treatment alone.  Combustion modification techniques have been the most widely used methods  to
control  NO  emissions.   Because an uneven distribution  of  heat to  the  process  tubes  could cause
          X
coking of the feedstock, combustion modification techniques  have been  limited  more than  in other
combustion processes.
     Since combustion modifications are the predominant technologies being used their descriptions
and capabilities will dominate the following discussion.  Prior to'discussing the various combustion
                                                 5-40

-------
modification techniques, emissions from uncontrolled  furnaces,  and  operational  control  methods and
their emissions will be discussed.  Flue gas  treatment will  be  discussed last.

     Operational Control Techniques - Reducing emissions  of  NOX from standard burners used In fired
heaters 1s sometimes possible using methods available to  the heater operator.  These methods Include
monitoring stack oxygen and combustibles, and then  reducing  average excess  air  levels,  reducing the-
amount of oil combusted, and adjusting burner air registers  for minimum NO   emissions levels.
Typically, heater operators adjust -the stack  damper and burner  air  registers for a compact, well
defined and stable flame.  Additionally, the  stack  damper 1s adjusted to assure the operator that
sufficient combustion air is available during all operating  conditions.   At these settings the stack
oxygen level usually averages 5 to 9 percent.  Recently,  because of energy  conservation programs
within some plants, stack oxygen levels are being monitored  and controlled  to lower levels.
      Emissions from a patrolewa  refinery  reformer with  70 preraix  burners  were analyzed using over
 2000  hourly measurements  (Reference  5-61).   Using an  auto regress ion technique,  NQX  emissions as  a
 function of stack oxygen level  were predicted.  Figure 5-6 shows the results  and  the  95  percent
 confidence limits of this predictive  procedure for firing rates that were typical during the test
 period.  The maximum design firing rate for the heater was 20.7 HW  (70 x 10  Btu/hr).  The actual
 operating conditions for this heater  was as follows:

      Fuel Firing Rate                     Stack 0,         ,          NOV Emissions
      KM (106 Btu/hr)                         %  £                  ng/Q (Ib/lpS  Btu)
        14.3  (48/9)                      3.15                       39  (0.09)
         3.8  (13)                        0,,1                         0  (0.0)
        19.0  (65)                        17.0                       120  (0.28)
                                                5-47

-------
01
i
fvj
  170


  160


  150

  140


  130

  120


55 110

n 100
5
jj  90


]  80
 x

:  70

   60


   50


   40

   30


   20


   10
                                Naptha Reformer: Prenrix
                                Burner, 70 percent load
                                                                  Btu/hr)
                                            3          4          56
                                                Stack Oxygen Level, Percent
8
                                                                                                                0.4
                                                                                                                0.35
                                                                                                                0.3
                                                                                                                0.25
                 V)

             0.2  §
                 in
                                                                                                                     cr
                                                                                                                     _j
                                                                                                                     o
                                                                                                                0.15
                                                                                                                     CO
                                                                                                                     <•*
                                                                                                                0.1
                                                                                                                0.05
                                 Figure 5-6.  NOW emissions from standard premix burners.

-------
     This heater was closely monitored and adjusted to maintain NO  emissions within  permit  limits.
The NO  emissions are controlled by manual adjustment of the stack damper to control  excess  air
levels.  The plant estimates that NOV monitoring and control require 4 manhours/day for six  heaters.
                                    A
Although it appears that NO  emissions per unit of heat input can be reduced by about 20 percent by
reducing stack oxygen levels from 6 percent to 3 percent, in actuality average NO  emissions may be
                                                                                 /*
reduced less.  The reason is that NO  emissions per unit of heat input tend to maximize at some
                                    A
firebox oxygen level above 6 pe'rcent to 8 percent oxygen due to the cooling effect of the-air on the
flames.  By operating for some periods of time above this point of maximum emissions, the average
emissions would be reduced from the emissions produced at the average stack oxygen level. It should
be noted, however, that because of the increased usage of fuel, the NOX emissions per unit of time
may continue to increase above this stack oxygen level.

     Less comprehensive data are available on a standard combination raw gas and oil  burner
(Reference 5-62).  The data were collected on a 108 cubic meter per hour (16,250 barrels per day)
natural draft, vertical cylindrical, crude oil heater with six 2.7 MW (9.2 x 10  Btu/hr) maximum
capacity burners.  Although 89 one hour NOX emission tests were conducted, there were six variables
that could be controlled by the operator that may have affected emission levels.  These variables
included the process rate, heat input rate, fuels combusted, secondary air register setting, steam
injection rate and stack oxygen level.  Due to the large 'number of variables, the effect of excess
air levels on NO  emissions could not be quantified with precision.  However, a trend can be
observed when data are arranged in an order of increasing stack oxygen concentrations in groups
where  the other variables are similar.  .Table 5-12 contains selected data from these 89 tests where
the other variables were similar and there were at least three different values of stack oxygen
level.  A definite trend of increasing NOX emissions with increasing stack oxygen levels is apparent
within each of the groups.
     Combustion Modification Techniques - As stated previously, combustion modification techniques
are the most common and widely used method to control NO  emissions.  Techniques that have been used
on burners designated by the vendor as 1ow-NO  burners include low excess air, high swirl burners;
staged addition of combustion air; combustion gas self recirculation; and staged addition of fuel.
These techniques are effective on a number of raw gas and dual fuel burner types.  Because of design
                                                5-43

-------
       TABLE 5-12.
SUMMARY OF NOX TEST  DATA ON  A STANDARD DUAL  FUEL
BURNER
Fuel* Process Rate
MG-FG/AG/16 (m3/hr)
100/0/0 55



100/0/0 74.9



100/0/0 94.1



0/100/0 76.2







56/0/44 55



37/0/63 76.2


54/0/46 94.1


63/0/37 95.4



Heat Input
(MM)
7.4
7.9
7.9
8.6
10.9
10.8
11.3
11.8
15.4
15.8 •
' 15.9
16.3
11.5
11.0
11.1
11.8
11.1
11.8
11.5
11.6
8.2



11.4


14.1


14.3



Secondary Air Stack Op
Register /„,%
(% open) (%>
502 2.3
4.2
5.6
8./
bO% 0.9
2.0
3.9
5.6
50% 1.9
3.3
4.2
4.5
50% 1.3
1.7
2.0
3.6
3.9
4.0
4.1
4.1
50% 1.7
2.9
4.8
6.7
50% 1.25
2.2
3.8
50% 1.9
2.6
3.4
50% 1.2
1.6
2.8
2.9
NO Emissions
ng/J (1b/106 Btu)
29
44
48
47
' 32
44
46
52
35
41
41
42
42
47
51
54
56
50
54
56
75
88
90
63
64
82
101
73
71
87
64
73
79
85
(.Ob8)
(.102)
(.111)
(.110)
(.07)
1.102)
(.107)
(.121)
(.081)
(.096)
(.096)
(.098)
(.100)
(.110)
(.120)
(.130)
(.130)
(.120)
(.130)
(.130)
(.175)
(.205)
(.210)
(.147)
(.149)
(.191)
(.235)
(,170)
(.165)
(.203)
(.149)
(.170)
(.184)
(.198)
Percentage ot natural gas-fuel gas/adsorber gas/number 6 fuel oil.
                                        5-44

-------
requirements, premix burners incorporate tha intimate mixing of fuel and air and can combine this
with combustion air staging and low excess ciir firing.  This may explain why the heater with
standard premix burners had significantly lower NOV emissions than the standard dual fuel  burner
                                                  A                            «
discussed previously.  Therefore, the discussions that follow will be limited to raw gas or liquid
fuel burners that incorporate one or more of these technologies.

     Low Excess Air, High-Swirl Burners - Burners that depend on low excess air high-swirl
conditions to reduce NO  emissions have been used on a number of petroleum refinery process
heaters.  Two of these heaters were tested to determine emissions of NO. CO, and SO, and to
                                        '                               X            t
determine the effects of heater operation on emissions (Reference 5-63).  The first test was of a
                                                                                             3
four cell, balanced draft refinery gas fired reformer heater with a maximum capacity of 134 M /h
(20,000 bbl/day) of reformate.  Each cell was separated from adjacent cells with a common wall and
had four burners mounted in each end wall.  The combustion air for all burners was preheated to
approximately 544 K (520°F).  NOX emissions for this heater ranged from 27.4 ng/J (0.06 lb/10
Btu) to 77.4 ng/J (0.18 lb/10  Btu).  NO  emissions exhibited a strong dependence on excess air
levels and a moderate dependence on heat input rate.  Figure 5-7 is a presentation of the multiple
regression of NOX emissions as a function of heat input rate and stack oxygen concentrations.  For
simplicity the five curves were plotted at the average heat input rate five groups of four process
conditions of the four cells tested.
     The second heater tested was a vertical cylindrical debutanizer bottoms reboiler.  A single
floor mounted burner was used with ambient temperature combustion air supplied by forced draft.  The
                                            o
process rate for the second heater was 187 M /hr (28.3 bbls/day) with the burner firing at a heat
input rate of 7.62 MW (26.1 x 10  Btu/hr).  This was about 95 percent of the process capacity of
the heater.  Compared to the first heater at near maximum load, NO  emissions from the second
heater were less.  Figure 5-8 presents the NO  emissions during the four tests of the second
heater.                                                                                      '    •   •
     Staged Combustion Air Burners - Burners that stage combustion air to reduce NOX emissions
have been the predominant design used in U.S. petroleum refinery heaters.  Most of the burners
appear similar to standard dual fuel burners except that a third air register exists to provide
additional control.  Other burner designs incorporate the gas and oil gun within the primary tile
case.  Air to the primary tile case for initial combustion of the fuel is regulated by a single air
register with secondary (staged) combustion air controlled by a second air register and admitted
between the primary and secondary tile case.  Non-vendor published data are available only on
burners of the three air register design.
                                                 5-45

-------
    (0.2
      80
      70




   (0.15)


      60
3
CD
      50
tn

o
•r*
I/I
in
O
as
40
      30
   (0.05)
      20
      10
               Heat Input Rate

                   10.2 MW  (35 x  106  Btu/hr)	


                    9.5 MW  (32 x  106  Btu/hr)	


                 8.27 MW  (28.3 x  106  Btu/hr)	

                 6.45 MW  (22.1 x  106  Btu/hr)	
                 4.56 MW  (15.6  x  10G  Btu/hr)
               Heater Operating Conditions

                 Preheat Air Temperature: 544 K
                                          (520°F)
                    1           23            456

                                       Stack Oxygen Level, Percent


               Figure 5-7. NO  emissions from a refinery gas-fired reforming heater
                           with low excess air burners.
                                              5-46

-------
 (0.15)
    60
3
4->
CO
    50
g(O.l)
01
VI
    30
 (0.05)
    20
    10
                                             345
                                       5itack  Oxygen Level,  Percent
                Figure 5-8. NOX emissions from  a  single  burner,  forced draft
                            debutanizer bottoms reboiler.
                                                5-47

-------
     NO  emission data for the three air register design burner type firing gaseous  fuels  include
one set of short terra emission tests of a forced draft heater and two sets of long term data  from  a
natural draft heater and a balanced draft heater with preheat.

     The forced draft heater was a vertical cylindrical heater rated for 92.7 M /h (14,000 bbl/day)
crude oil throughput (Reference 5-63).   Three burners were floor mounted in the heater and fired  a
combination of natural gas and refinery gas.  During the test the process rate varied from 57
percent to 80 percent of rated capacity and the firing rate varied from 3.48 MW (11.9 x 10 Btu/hr)
to 4.85 MW (16.5 x 10  Btu/hr).  NOX emissions for this heater were as low as 24 ng/J (.056 lb/10
Btu) to as high as 44 ng/J (0.10 lb/10  Btu).  Although the air register settings were adjusted to a
number of different settings, 13 tests were conducted with air registers set to approximately the
same openings that existed prior to the start of the tests.  Of the other air register settings
evaluated during the test, a setting of primary and tertiary registers 100 percent open and
secondary registers closed was designed as optimum for low NOV operation.  Figure 5-9 presents the
                                                             X
results of the 13 tests and three tests conducted at the approximate setting designated as optimum.
Additionally, the two tests conducted with the air registers set at openings that may be considered
as simulating a standard burner are also included.  Although not conclusive, it appears that  the air
register setting could increase or decrease NO  emissions from 3 to 20 percent from  the levels
produced with the original register settings.  Also it appears that excess air levels have a  greater
impact on NO  emissions than either air register settings or load.

      The two sets of long term data for gas fired heaters were collected by the plants as a  result
of local permit requirements (Reference 5-61).   The first was a naphtha reformer that consisted of
four separate vertical cylindrical heaters with a common'convective section.  The combined capacity
of the four heaters is 25.4 MW (87 x 10  Btu/hr) with four burners in three of .the heaters and three
burners in the remaining heater.  The outlet temperature of the naphtha from the heaters was  approx-
imately 770 K (930°F).  Over 1000 hourly data points of stack oxygen level, fuel firing rate, and
NO  emissions were collected.  The plant monitors oxygen level and adjusts the stack damper for
energy conservation considerations.  Adjustments are not performed to the burners or stack damper  to
control NOV emissions because the plant is well within its permit limits.  Fuel firing rate for
          A
these heaters averaged 20.9 MW (71.3 x 106 Btu/hr) and ranged from 1.37 to 24.5. MW (4.7 to 83.8 x
10  Btu/hr).  Stack oxygen levels averaged 5.5 percent and ranged from 2.7 to 8.4 percent  for this
same period.  The resulting NOX emissions for these heaters during this period averaged 42.9  ng/J
(0.10 lb/106 Btu) and ranged from 21.5 ng/J (0.05 lb/106 Btu) to 68.7 ng/J (0.16 lb/106 Btu).
                                               5-48

-------
   (.12
     50
01
-fc.
lO
 CO
VD
 O
   (.10)
 C(-08)

      30
 c?(.06)
 I   20
 « (.04)
 i
 O
      10
    (.02)
                  Trend Line
                             3.SWW (12  x  KT Btu/hr)
                         	 4.7 MW (16 x 106 Btu/hr)
                            OProcess Rate, 53 m3/hr (8000 bbl/day),  Fuel Firing Rate 3.48-4.16  MW
                          £ AProcess Rate, 72.87-73.13 m^/hr (11,000-11,040 bbl/day), Fuel  Firing  Rate 4.48-4.85
                          "SO   Optimum Low-N0x Register Adjustment  (High Process Rate)
                          D.   1-2
                          
-------
      Because there was  a  significant  amount  of missing data in the middle of the data, both parts of
 the data were analyzed  to determine the  effects  of  load and stack excess air levels on NOV
                                                                                         A
 emissions.   Load  effects  on  the  NOV emissions could not be determined to be of significance with
                                   A
 either data  set analyzed.  However, stack oxygen level did have a significant effect on NO
 emissions.   Figure 5-10 shows  the  predicted  average NO  emissions as a function of stack oxygen and
 the 95 percent confidence limits for  these values.  Only the data set with the highest predicted NO
 emission rate is  shown.  The other data  set  predicted approximately 20 percent less NO  emissions.
                                                                                      A

      The second set  of  data  collected was on a balanced draft vertical cylindrical crude heater that
 used preheated combustion air  (Reference 5-61).   The heater was retrofitted with nine staged air
 burners with a total  heat input  rating of 25 MW  (86.3 x 10  Btu/hr).  The preheat temperature was
 between 485  K (414°F) and 50C  K  (550°F).  Almost 3,000 hourly data points of stack oxygen level,
•fuel firing  rate,  and NO   emissions were collected.  The plant monitors and controls NO  emissions
 because of local  permit limits.  Adjustments to  NO  emissions are accomplished by adjusting the
 stack damper to control excess air levels.   Burner  registers are adjusted to control flame impinge-
 ment on the  process  tubes.   Fuel firing  rate for this heater averaged 16.4 MW (56 x 10  Btu/hr) and
 ranged from  2 to  20.3 MW  (6.9  to 69.3 x  10   Btu/hr).  Stack oxygen levels averaged 2.11 percent and
 ranged from  0.1 to 19.0 percent.   NOX emissions  ranged from 3.4 ng/J (.008 lb/10  Btu) to 68.8 ng/J
 (0.16 lb/106 Btu)  and averaged 34.4 ng/J (0.08 lb/106 Btu).
     The data were  analyzed  statistically to determine the effects of load and stack excess air
 levels  on NO emissions.   Increases  in both load and stack excess air levels increased emissions of
 NO .  Figure 5-11 shows the  predicted NO  emissions as a function of stack oxygen level for two
   A                                    A
 different loads.  The  95  percent  confidence limits for these values are also indicated.
      Limited NO  emission  data  from staged  combustion air burners firing 100 percent liquid fuels
 are available.   As  a  result  of  a  study  to determine emission levels from process heaters with staged
 combustion air  burners  three emission tests were  conducted  (Reference 5-63).  The first test was on
 a natural  draft vertical cylindrical crude  heater which fired a 0.15 percent nitrogen distillate
 oil.   A single  burner of 4.4 MW (15 x 10  Btu/hr) maximum rated firing capacity was installed in the
 floor of the heater.  Actual  firing rate varied from about  4.03 MW (13.8 x 10  Btu/hr) to 4.49 MW
 (15.3 x 10  Btu/hr).  Stack  oxygen  levels varied  every few  seconds resulting in a variation of from
 5 to 8 percent  02 in  the "as found" condition.  This variation was attributed to the low overall
 pressure drop of the  heater  (which  did  not  have a convective section) and local variation in wind
                                                 5-50

-------
               70
en
i
               60
               50
            ?40
            VI

            c
            to

            VI
              30
              20
              10
                                                                                                                   0.15
                                                                                                                   0.10
                                                                                                                        CO
                                                                                                                   0.05
                           1          234          5          6          7          8

                                                    Stack Oxygen Level, Percent



                          Figure 5-10. NOX emissions from a natural draft Naptha reformer with staged combustion

                                      air burners.

-------
         60
         50
at
i
in
       cn
       c


       w"40
       c
       o

         20
         10
                                                                                                      0.15
                                                     18.8
                                       0.10
                                             o
                                             Ol

(43
                                       0.05
                                                      I
                      1          2345678

                                             Stack Oxygen Level, Percent


                 Figure'5-11. NOX emissions from a balanced draft crude preheater with staged

                             combustion air burners, combustion air preheater to 485 to 500 K.

-------
conditions at the site.  Stack oxygen levels and burner air register settings were adjusted to
determine their effects on NO  emissions.   Reducing stack oxygen levels decreased NO  emissions
                             A                                                      X
significantly as shown in Figure 5-12.  Minor adjustments in.burner air register settings did not
seem to affect NO  emissions significantly.  However, as the two tests indicated in Figure 5-12  by
the hexagonal points, closing the tertiary air register to about 15 percent open and the primary air
register to 50 percent open increased emissions moderately. The complete closure of the tertiary
air register would simulate a standard burner.

     The second test was on a natural draft horizontal heater  firing 0.81 percent nitrogen residual
oil (Reference 5-63).  Twelve 0.97 MW (3.3 x 10  Btu/hr) staged air burners were mounted on the  long
walls of the heater and fired toward the center.  The heater was fired at about 9.8 MW (33 x 10
Btu/hr) during the test.  The stack damper on this heater was  stuck, therefore, the burner air
registers were adjusted to attempt to reduce excess air levels.  Figure 5-13 shows the results of
the six tests on this heater.  The data points, connected by lines represent tests where air register
settings were the same.  Although the tests with more open air register settings had higher
emissions, the differences are probably not significant.

     The third test was on a natural draft vertical cylindrical heater firing 0.81 percent nitrogen
oil (Reference 5-63).  Three 2.24 MW (7.64 x 10  Btu/hr) staged combustion air burners were mounted
in the floor of the heater.  Although most of the tests were conducted at about 6 MW (20 x 10
Btu/hr) heat input, one test was conducted at about 6.5 MW (22 x 10  Btu/hr).  The "as found" stack
oxygen level was between 8.1 and 11.4 percent.  This was reduced to as low as 6.9 percent by
partially closing one of the two stack dampers and by minor adjustments of the burner air registers.
No clear indication of the effect of stack oxygen on NOX emissions can be discerned from the data.
Possible reasons for this are that too few tests were conducted and the stack oxygen levels may  not
have been reduced past the level of maximum emissions above which flame cooling by the combustion
air begins to reduce emissions.  The results of the seven tests are shown on Figure 5-14.

     The simultaneous firing of both liquid and gaseous fuels  in the same heater is a more typical
occurrence in the chemical and petroleum industries.  This is  accomplished either by firing only
liquid fuels in some burners and only gaseous fuels in the remainder of the burners or by base loadint
all of the burners on liquid fuel and supplementing the heat requirements with gaseous fuels. Three
tests are available on combined firing of liquid and gaseous fuel in staged combustion air burners.
                                                  5-53

-------
Ul
I
01
130


120


no

100


 90

 80

 70


 60


 50

 40


 30


 20

 10
                                                  0
                                                             O
                                            Burner Type:  Staged Combustion Air
                                           s Fuel:  Distillate Oil
                                            Load:  33.12 m3/h (5000 bbl/d)
                                        Q All registers 100X Open (Baseline)

                                            Reduced Tertiary Air and Primary Air

                                            Reduced Primary Air Only

                                            Primary and Secondary Air Registers 502 Open

                                            Primary Register 50% Open,
                                            Secondary Register 70% Open
                                            I           I          1          I          I
                                                      4          5
                                                Stack Oxygen, % Dry
8
                                                                                                            0.30
                                                                                                            0.25
            0.20
                  o
                                                                                                                  CO
                                                                                                                  rt-
                                                                                                                  C
                                                                                                            0.15
            0.10
            0.05
           Figure  5-12.  NO emissions as a  function of  stack oxygen for a distillate  oil-fired,
                          natural  draft process heater.

-------
   250
  (Q.S8)
   Z40
  (0.56)
03
 ,  230
 UQ.53)
 . 220
2(0.5.1}
   210
  (0.49)
   200
  (0.47)
                                             Firing Rate  9.8  MW (33 x 10b Btu/hr)
                                                  35/86/81
                                                                 Mr Register Setting:
                                                                 Primary/secondary/tertiary
                                           Stack Oxygen, % cry
                      Figure 5-13. Residual oil-fired  horizontal  heater with
                                   staged  combustion air burners.
                                               5-55

-------
         290

       (0.6$
      CO
     IO
      O
   280

 '(0.65;

-------
     The first of these tests consisted of 14 short term emissions evaluations (Reference 5-63).
The heater was a vertical rectangular natural draft heater with 10 staged combustion air burners
installed.  Each burner is rated at 1.2 MW (4 x 10fi Btu/hr).-  The firing rate for the heater during
the tests varied from 8.79 MW {30 x'106 Btu/hr) to 11.98 MW (40.9 x 106 Btu/hr)'.  The liquid fuel
fired was 0.81 percent nitrogen oil.  All but three of the tests were conducted with all burner air
registers completely open.  A curve for the tests with over 80 percent oil firing and all burner air
registers open shows a definite lowering of NOX emissions with reduced excess air levels.  The other
three tests were conducted with only minor adjustments to the burner air registers.  Considering the
small number of tests and variations in firing rate and oil-to-gas ratios, a significant change in
NOX emissions with changes in burner air register settings is not noticeable.  Figure 5-15 shows the
results of these tests.  Tests with different burner air register settings are indicated with
different shape data point markers.  The percentage of oil fired is also indicated  in Figure 0.
Based upon the two tests with lower rates of oil firing it appears that the percentage or amount of
fuel oil fired affects th« emissions of NO^.

     The other two tests of staged combustion air burners are from long term continuous emission
data on heaters firing a combination of refinery gas and residual  oil  (average nitrogen content of
about 0.65 percent)(Reference 5-61).  Both heaters were operated by the same refinery.  The refinery
continuously monitored NOX emissions to show compliance with a permit  requirement that they easily
achieved.  Therefore, heater adjustments were not usually made to reduce NO  emissions.  The plant
did, however, monitor stack oxygen level and adjustments were made when stack oxygen levels were
outside control limits.

     The first heater was a natural draft vertical cylindrical atmospheric crude heater
(Reference 5-61).  Eight floor fired burners with a total firing capacity of 377.5  MW (110.6 x  106
Btu/hr) are installed 1n the heater.   The crude temperature to the atmospheric distillation column
was between 636 K and 639 K  (685°F and 690"F).  The oil fraction averaged 31 percent and varied from
20 to 45  percent of the  total heat  requirement during the almost 1000  hours of data collection.  The
total fuel firing rate averaged 416.7  MW (122.1 x 10° Btu/hr) and ranged between 67.6 and 466.6 MW
(19.8 x 10  and 136.7 x  10   3tu/hr).   The stack oxygen levels were maintained between 1.0 and 9.0
percent and averaged 3.5 percent.  NOX emissions for this heater averaged 51.5 ng/J (.12 Ib/lQ6 Btu)
and  ranged from a low of 30.0 ng/J  (.07  lb/105 Btu) to as high as 68.7 ng/J (.16 lb/10S Btu).
Mathematical  modeling of the data could not detect a relationship between oil-to-gas ratio or heater
load and  NO   emissions.  However, stack oxygen levels did correlate with emissions  of KQX>  Figure

                                                5-37

-------
*->
CO
••3


C
O
      (0.5)

       210
       200



     (0.45)j—

       190
       180
(0.4)
 170
       160
     (0.35)
       150
       140
       130
    . (0.3)
           Burner Air Register Openings

               O « 100/100/100 (Curve  drawn for  this data)
               D * 65/72/100
               A - 60/59/76
               (Oil firing  rate in  MW/percent  of
                      heat input by oil)
                                                            (10.0/85%J
                                                                O
                                      (8.9/82%)
                                                       (9.4/84%)
                                                            O/  D(9.5/85%)
                                                              0(9.4/83%)
                                                             0(9.1/82%)
                                                            O(l 0.0/84%)
                                    (8.2/81%)
                                       O
                                         39.4/84%)
                                           '(9.4/843!)
                                                 A(9.4/84%)'
                    (7.9/81%)
                         O
                                               (5.9/61%)

                                                   O
                              (5.0/57%)
                                   O
                     I
                          I
                                           I
                                                                8
               4567
                           Stack Oxygen Level, Percent
             Figure 5"-15.Residual  oil/refinery gas-fired  natural  draft  heater.
                                             5-58

-------
5-16 shows the NO  emissions as a function  of stack  oxygen and also the 95 percent confidence limits
on the NO  emissions.

     The second heater was a natural  draft  vertical  cylindrical  vacuum distillation  column  heater
(Reference 5-61).  Four floor fired burners can provide a total  heat  input  capacity  of 126.3  MW
(37.0 x 106 Btu/hr).   The process fluid is  heated to approximately 650 K  (710°F).  The oil  fraction
fired in this heater varied more (0 to 40 percent) but averaged  less  (6 percent)  than  the previous
heater.  The total firing rate averaged 93.2 MW (27.3 106 Btu/hr)  varying from 67.6  to 151.9  MW
(19.8 x 10  to 44.5 x 10  Btu/hr).  The stack oxygen levels  were maintained  between  0.9 and 11.4
percent and averaged 3.9 percent.  NOX emissions during these conditions  ranged from 17.2 to  51.5
ng/J (0.04 to 0.12 lb/106 Btu) and averaged 30.0 ng/J (.07  lb/106  Btu).   Mathematical  modeling of
the data indicated that emissions increased with increasing oil  fraction  and decreasing load. One
possible explanation for this is that the heater was base loaded on oil.  Therefore, as load  was
reduced the oil fraction increased, however, emissions rates were  not reduced or reduced only
slightly, resulting in an increased emission factor.  Figure 5-17  includes  the NO emissions  with
95 percent confidence limits for two different loads as stack oxygen  level  varies.   As with the other
heaters NO  emissions decrease with decreasing stack oxygen  level.

     All of the above data on staged combustion air burners were collected  on cylindrical flame
burner designs where the oil gun was in the center of the recirculating primary oil  tile, the gas
tips were between the primary and secondary tile cases and a tertiary air port surrounds the
secondary tile.  Data on other burner designs and on other burner types that stage the combustion
air are available.  At least one manufacturer claims that their burner is designed to  incorporate
flue gas self-recirculation in addition to  staged combustion and the  ability  to operate at  low excess
air levels (Reference 5-64).  Emissions reported by the vendor for this burner installed in a 6000
barrels/day hydro desulfurization unit show reductions in ,NOV emissions while firing gaseous  fuels
                                                            A
of from 60 to 80 percent compared to a burner considered by this manufacturer to be  standard. Data
collected by a refinery on a heater with these burners installed show about a 50 percent reduction
of NOV emissions (Reference 5-65). Test furnace data on another manufacturers burner of a somewhat
     A
similar appearance (gas tip'and oil gun within primary tile and and secondary air port around
primary tile) show NO  emissions as a function of primary air register opening at. similar excess air
levels from 40 to 70 percent less than a burner considered to be standard by this manufacturer
(Reference 5-58).  Data for both of these manufacturers' burners show reductions in  NO  emissions
during liquid fuel firing of about 30 to 40 percent compared to a burner  considered  by each to be
                                                  5-59

-------
in



k
                 70f—
                 60
                 50
              '-3



              O1
              VI

              c
              o
                 30
                 20
                 10
                                                                                                              0.15
                                                            345

                                                       Stack Oxygen Level,  Percent
                                                                                                                   o
                                                                                                                    o>
                                                                                                              0.10
                                                                                                              0.05
                     Figure 5-16. NOX emissions from a natural draft atmosphere crude  heater  with staged

                                combustion air burners, oil and gas fuel combustion.

-------
               60
                                                                                                                   0.15
01
I
en
               50
V)
c
o
•r-
V)
l/>
   30
             x
            o
               20
                                                                                                                   0.10
                                                                                                                        cr
CO
H-
                                                                                                                   0.05
               10
                                                              I
                             1           2          3          4          56.7          8
                                                      Stack Oxygen Level, Percent

                        Figure 5-17.NOX emissions from a natural draft vacuum distillation column heater with
                                   staged combustion air burners, oil and gas firing.

-------
standard.  Data for the two different burners  cannot easily be compared because of  differences  in
test conditions.
     At least one manufacturer has  two burner designs incorporating staged  combustion air that can
be substituted for standard high intensity burners and standard flat flame  burners  (Reference 5-58).
Although these burners have been successfully used in full scale turnovers, emission  data from full
scale furnaces are not available.   Test furnace data for the high Intensity burners show reductions of
NOX emissions during gas or oil firing of about 60 percent from a burner the manufacturer considers
standard.  Data presented by the manufacturer show that as fuel nitrogen increases  the amount of
reduction of NOV emissions increases.  Test furnace data for a low NO., flat flame burner firing next
               A                                                    X
to a radiant wall shows a reduction of between 40 and 50 percent by changing the staged air port from
the side away from the radiant wall to the side of the burner nearest the radiant wall.  According to
the Manufacturer, emissions for both of these flat flame burners are less than the standard flat
flame radiant burner.
     Staged Air Lances - As a  result of a research project to develop  a method to reduce  NOX
emissions from refinery process heaters, an air lance system  was  developed, installed, and  tasted  on
an operating natural draft vertical cylindrical crude heater  (Reference 5-62).  The design  consists
of four 3.18 CM (1.25 1n) stainless steel lance tubes inserted through the furnace  floor  around each
of the six 2.68 HW (9.14 x 10S Btu/hr) dual fuel burners.  A  45°  elbow is welded  on the end of  each
lance and a fan was used to supply air to the tubes.  An optimum  staging  height of  0.3 m  (1 ft)
above the furnace floor was established for gas firing and 1.22 m (4 ft)  was established  for  oil/gas
firing.  Since burner rile tops were about 0.23 m (.75 ft) above  the furnace floor, staging heights
of less than this impinged on  the  burner tile.  An optimum' burner equivalence ratio  (the ratio of
air drawn through the burner to the stoichiometric requirement) was-not determined, however,
emissions declined with decreased  equivalence ratio (increased staging).   A leveling off  of
emissions was not apparent at  the  maximum air injection  rate.' During  short term  tests, when  fan
capacity limited the air injection rate, an NO  emission reduction of  35  to 45 percent was
demonstrated during gas firing and a reduction of about 35 percent was demonstrated during
combination oil/gas firing.  Short term tests that were conducted after increasing  the fan  capacity,
thereby, allowing increased staging through the air lances, demonstrated  lower NOV  emissions  during
                                                                          i       X
combination oil/gas firing.  Additionally, by combining staged air injection with low overall excess
air reduction, a further reduction of NO  emissions could be  achieved.

     A long term continuous emission test was conducted on this heater.   During the test, the crude
throughput was maintained at 71.2  m /hr (10,717 bfals/day) or  66 percent of capacity.  The fuel  fired
                                                                               3             "?
airing the test was refinery gas with an average calorific value  of 61,030  kj/m   (1638 Btu/ft ).
Fifteen minute average values  for  stack oxygen, carbon dioxide, carbon monoxide, and nitrogen oxide

                                               -5-62

-------
were continuously recorded for a 34-day period.  The stack oxygen level  averaged 2.8 percent during
the test with a range from 0.1 percent to 10.8 percent.  The first 15 days of the test the heater
was operated at an average of 1.9 percent stack oxygen.  The average NO  emissions for this period
were 18.5 ng/J  (0.04 lb/10  Btu).  The last 16 days of the test the heater was  operated at an
average of 3.4 percent stack oxygen.  The average NO  emissions for this period were 25.5 ng/J (0.06
lb/10  Btu).  Minimum NOV emissions were 2 ng/J, however, CO emissions of over 400 ng/J (1 lb/10
                        X
Btu) occurred at the same time.  Minimum NO,, emissions without an increase of CO emissions were 12
                                           A
ng/J (.03 lb/10  Btu) when stack oxygen levels were 1.0 percent.  Maximum NO  emissions were 59 ng/J
(0.14 lb/10  Btu) where stack oxygen levels were at about 7.3 percent.  Although baseline emissions
(at 4 percent stack oxygen level) of 66 ng/J (0.15 lb/10  Btu) were established on this heater prior
                                         »
to the long term test,' other tests of the standard burners without air lances being used  (see Table
5-12) at about the same load documented emissions as low as 54 ng/J (0.13 lb/10  Btu).
     Staged Fuel Burners - A burner that is being installed in at least two high temperature steam
hydrocarbon reformers, one at a U.S. methanol plant, and one at a Canadian methanol plant, offers
substantial reductions of NO  emissions compared to presently used standard burners and possibly
many of the low NOV burners.  The burner requires mechanical draft because of the high pressure drop
                  A
required across the burner (Reference 5-58).  At the present time, only gaseous fuels have been used
in the burner, however, the manufacturer believes that with proper nozzle design liquid fuels may be
able to be fired.  The burner can be made in round or conical flame and flat flame designs. .Test
furnace data by the vendor show a 60 to 70  percent reduction in NOV emissions from a baseline of 90
                                                                   X
ppm at 3 percent Og (approximately 46 ng/J or 0.11 lb/10  Btu)(Reference 5-58).  Actual emissions
data from an operational furnace would probably be higher because of higher furnace temperatures and
the use of a high degree of preheat.
     Flue Gas Treatment - Two methods of treating process heater flue gases to reduce NOX emissions
are available.  Both methods use ammonia as a reactant with NO to form nitrogen gas and steam
(Reference 5-65).  The difference in the methods is the use of a catalyst to reduce the reaction
temperature of 1250 to 1370  K (1800 to 2000°F) required in the non-catalytic method.  The two
methods are called selective catalytic reduction (SCR) and selective non-catalytic reduction (SNCR)
or "Thermal  DeNO ."  The  "Thermal  DeNOx method,  patented by Exxon, can incorporate the use of hydrogen
to lower the reaction temperature  to approximately 980 K  (1300°F).
                                               5-63

-------
     There are at least 24 refinery heaters that have had an SNCR system installed (Reference 5-65).
Emissions reductions of 35 to 50 percent  have been achieved in new and retrofit situations.  It is
expected that with Improved ammonia injection methods, increased reductions can be achieved.  Since
one of the disadvantages of SNCR is that  the fraction of NOX reduced decreases as the concentration of
NO., decreases, this technology may not be as effective when used with low NO  burners firing gaseous
  •rC                                                                   ,     A
fuels.  However, some refineries use this as an interim technique to meet permit limits  during times
of high NOX emissions such as during oil  firing or during high production periods.

     At least one U.S. refiner has  retrofitted an SCR system on a balanced  draft crude topping
heater.  The heater has been operating since May 1979 without difficulty  (Reference 5-65).  The
catalyst activity has not deteriorated nor the pressure drop increased  in spite of  seven emergency
shutdowns unrelated to the heater and  several occasions when the furnace  operated under a temporary
upset condition.  Emissions have been  reduced by 90 percent from an  average of 100  ppm to an average
of 10 ppm.
     Costs - The cost of controlling NOX emissions from refinery heaters has been  reported  in  at
least two studies.  The most comprehensive study of costs was recently accomplished by the  South
Coast A1r Quality Management District and the Stationary Source Control Division of the California
Air Resources Board.  This study, which will be referred to as the CARB reoort, was developed  as
part of the background information for a public meeting considering NOX emissions  control from
boilers and process heaters at refineries.  The major technologies for which cost  data were
developed were the retrofit Installation of low NOX burners, selective non-catalytic reduction and
selective catalytic reduction.  A less comprehensive study evaluated the cost of retrofitting  a
staged air lance system and automatic oxygen trim control on a process heater.  These two studies
art the source of cost Information presented in this section (References 5-65, 5-66).  The  costs of
NO  control will be presented in the following order:  automatic oxygen trim control, low NO
burners, selective non- catalytic reduction and selective catalytic reduction.
     Automatic Oxygen Trim Control - Although automatic oxygen trim systems are common on large
chtsrical industry heaters, roost refineries that control stack oxygen levels do this manually by
using plant operators to adjust stack dampers.  The daily time required to monitor and adjust stack
oxygen level varies with the--level of oxygen control desired and the variabilities in process heat
requirements and fuel composition.  One plant has estimated that about two hours per day are
required to adjust stack oxygen levels on six heaters in order to control NO  emissions
                                             5-64

-------
(Reference 5-61).  Much of this time could be reduced by installing an automatic oxygen  trim system.
Additional benefits that may be gained include tighter control  of the excess  air levels  and
decreased potential of upset conditions occurring.   The installed cost of retrofitting an  automatic
draft controller to a natural draft heater of about 16 MW (55 x 10  Btu/hr) thermal  input  capacity
Was expected to be about $40,000 (Reference 5-62).   It was estimated that the initial cost of an
automatic draft controller would be proportional  to the heat input rate to the 2/3 power.  Annual
maintenance costs for the automatic systems were  estimated to be about $200.   Because of the tighter
control achievable with an automatic draft controller, additional fuel savings could be  achieved
resulting in an overall cost savings.  These fuel savings were estimated at about $32,000  per year
for a 16 MW (55 x 106 Btu/hr) heater; $59,000 per year for a 29.3 MW (100 x 106 Btu/hr)  heater; and
            •                           c
$296,000 per year for a 147 MV1 (500 x 10  Btu/hr) heater (Reference 5-62). The annualized savings   •
for these heaters were estimated to be about $17,000, $35,000 and $230,000 respectively
(Reference 5-62).
     Low NO  Burners - Low NO  burners are capable of being installed in most process heaters.
However, because of the differences in installation requirements and heat distribution requirements,
the cost to retrofit low NO  burners in two different heaters of the same size may be substantially
different.  Data collected from eight refineries on the cost of retrofitting low NO  burners in nine
separate heaters is presented in Table 5-13 (Reference 5-65).  As can be seen in this table," a wide
variation in burner costs and retrofit costs exists, and the variation would not be estimated by
adjusting for the size of the furnace.   In the document where these data are presented,  it  is  stated
that the annual cost of retrofitting low NO  burners can be estimated using a scale up factor of 0.6
                                                          n
from the cost of $11,795 for retrofitting a 9.4 MW (32 xlO  Btu/hr) furnace.  A contingency factor
of 35 percent was added to this estimation in a later supplement to the report.  Therefore, the cost
of retrofitting low NO  burners using this procedure could be estimated using the formula I = 4160
(Q)0'6, where Q is in MW (or I = 1990 (Q) ' , where Q is in 10  Btu/hr).  However, by regression
analysis of the retrofit cost and heater size presented in the table, a scale up factor of 0.96 is
calculated with a correlation coefficient of 0.85.  Using this analysis with a contingency factor of
         *                                 •                                    0 96
35 percent, the retrofit cost could be estimated using the formula I = 1500 (Q) .*   where Q is in MW
(or I = 456 (Q)0'95 where Q is in 106 Btu/hr).  It should be realized that because of the vari-
abilities in the cost of retrofitting low NOX burners, either one of these estimation methods may.be
off by a factor of two or three.
                                             5-65

-------
TABLE 5-13.  RETROFIT COSTS FOR LOW NOX BURNERS
Refinery
Beacon
Chevron
Coastal
Petroleum
ECO Petroleum
Golden Eagle
Kern County
Newhall
Newhall
Texaco
Average

Size of
Furnace
(106 Btu)
31.29
90
21.7
27
25.7
.20.6
5
5
40
.286.29
9
= 31.8
Cost of
Burners-
Dollars
(Year)
10,400
(1978)
120,000
(1981)
10,000
(1979)
4,360
(1977)
4,800
(1979)
5,800
(1979)
4,800
(1979)
2,220
(1979)
8,640

*
Retrofit
Cost-
Dollars
(Year)
4,000
(1978)
80,000
. (1981)
11,000
(1979) .
17,660
(1977)
10,000
(1979)
20,000
(1979)
9,320
(1979)
4,660
(1979)
13,820


Total
Cost-
Dollars
(Year)
14,400
(1978)
200,000
(1981)
21 ,000
(1979)
22,020
(1977)
14,800
(1979)
25,800
(1979)
14,120
(1979)
6,880
(1979)
22,460


Total
Cost
(1981
Dollars)
19,645
200,000
26,261
32,204
18,508
32,263
17,657
8,604
32,848


Total
Annual 1 zed
Cost (1981
Dollars)
5,375
54,720
7,185
8,811
5,064
8,827
4,831
2,354
8,987
106,154
9
= 11,795

-------
     Selective Non-catalytic Reduction - The use of selective non-catalytic reduction requires  the
installation of the system and controls to inject the ammonia at the proper location  in  the  heater
or the ammonia and hydrogen if the temperatures in the heater are lower than required, and a
hydrogen and/or an ammonia storage tank.  The cost of one refinery to retrofit the system, controls
and a large ammonia storage tank to control NO  emissions from a 9.4 MM (32 xlO  Btu/hr) heater were
$83,000 in 1980.  Of this $83,000 installation cost, $20,000 was attributed to installing an ammonia
tank.  To estimate the retrofit costs for other facilities, the authors of the CARS report made the
assumption that an ammonia storage tank would not be required at all facilities and that the system
and controls can be estimated using a scale-up factor of 0.6.  The resulting formula  to  estimate-the
capital cost of a SNCR system is I = 16420 (Q)0-6, where Q is in MW (or I  = 7860 (Q)0'6, where  Q  is
in 10  Btu/hr.)  The annual cost of controlling NOX with SNCR was made by  assuming a  capital
recovery factor of 0.274; maintenance, labor and spare parts cost of 3 percent of capital cost; and
a plant overhead cost of 25 percent of maintenance and labor.  Therefore,  excluding annual costs  of
ammonia, steam and hydrogen, the annual costs would be about 0.31 of the total capital cost. The
annual cost of ammonia, steam and hydrogen were estimated using the following additional
assumptions: (!) an ammonia to NOX mole ratio of 1.5 to 1.0 at a retail cost.of ammonia  of $0.25  per
pound (wholesale costs of $0.125 per pound could be used'if 2 large storage tanks were installed),
(2) a steam requirement of 12 pounds per 10  Btu of fuel used at a cost of $3.50 per 1000 pounds  of
steam, and (3) a hydrogen to ammonia mole ratio of 1.0 to 1.0 at a cost of $1.10 per pound of
hydrogen.
     Selective Catalytic Reduction - Estimating the costs of retrofitting an SCR system to reduce
NO  emissions is more difficult than predicting the costs of SNCR or low NO  burner retrofits.   Some
of the reasons for these difficulties include greater site characteristic dependencies; whether
reheat before the catalyst bed is required; and other contingencies such as furnace down time
requirements, frequency, type, and amount of liquid fuel fired in the furnace, start up costs,  and
vendor guarantee costs.                             . •

     One refinery installed a selective catalytic reduction sytem using in-house labor on a new 14.2
MW (48.5 x 106 Btu/hr) boiler.  The capital cost of this installation was $333,000 in 1980.  By
using a factor of 1.25 to account for contracting the installation, adding a 35 percent contingency
factor, and including a site factor of 2.2 when reheat is required or 1.5 when reheat is not
necessary, a formula to estimate the capital cost of an SCR system for heaters of different sizes
was developed.  The formuVa for a unit requiring reheat is: I = 255,400 (Q)    + 3170 Q, where  Q is
                                                   5-67

-------
the rated heat Input 1n MW (or I = 122,000(Q)°'6 + 9300 Q,  where Q is  in  106 Btu/hr).  The formula for
a unit that does not require reheat is:  I - 174,000(Q)°'6  + 31700 Q,  where  Q is  the  rated heat input
1n MW  (or I « 83,000(Q)0>6 + 9300 Q, where Q is in 106 B.tu/hr).   The second  term  in these equations is
the'cost of the catalyst, which is estimated to last about two years,   Annual operating  costs can be
estimated by multiplying the capital cost of the system by the factor specified before,  which covers
the capital recovery factor; maintenance, labor and spare parts requirements; and plant  over-head
requirements.  In addition, the cost of ammonia which must be added at a NO,, mole ratio  of 1  to 1,
                                >                                          x
catalyst replacement every two years, steam costs at a rate of 0.5 Ib per 10  Btu fired, and
electrical costs at 0.25 KM per 106 Btu fired must be included.
5.3.1.2  Catalytic Crackers and CO Boilers

5.3.1.2.1  Process Description

     A fluid-bed catalytic-cracking unit is often an integral part 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.

     The regenerator flue gas contains from 6 to 12 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 and to allow variable firing rates to meet
the refinery steam demands.

5.3.1.2.2  Emissions and Control Technology

     KOX 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
5-14, has shown results ranging from 100 ppm to 230 ppm of NO .  Hunter (Reference 5-34) reported
                                                  5-68

-------
   TABLE 5-14.  NOX EMISSIONS FROM PETROLEUM
               REFINERY CO BOILERS (REFERENCE  5-33)
          Investigator
       NOX
(ppm as measured)
Schulz, et. al., (Reference 5-67)
Schulz, et. al., (Reference 5-68)
Shea (Reference 5-71)
Shea (Reference 5-69)
Cowherd (Reference 5-70)
     104-116
  (average 106)

      70-89
   (average  78)

      96-233
  (average 163)

     101-159
  (average 135)

     108-162
  (average 129)
                         5-69

-------
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 has negligible effect on NO  emissions.  CO emissions,  however, were very sensitive
to excess air and increased rapidly below about 2 percent excess oxygen,  the lack of response of
N(L to combustion modifications is attributed to NOV that is formed  from ammonia in the CO gas feed
  x                                                x
acting similarly to fuel nitrogen*in oil or coal.

     The average emission factor for NO  from fluid catalytic cracking units is estimated in
Reference 5-23 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 was estimated  to be  45 Gg (50,000 tons) in 1974.
If the regenerator exhaust is burned in a CO boiler , the resulting  NOX emissions can presumably be
controlled by the classical methods discussed in Section 4.2 of this  report.

5.3.2  Metallurgical  Processes

5.3.2.1  Process Description  and  Control  Techniques
       The iron and steel  industry  is  the predominant source of NO  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  NOX.   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 NOX 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
section are taken from a 1976  I6T study (Reference  5-33)  which  uses 1971 steel industry data as a
source for fuel consumption and NO  emissions estimates.
Palletizing
       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 palletized 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 palletizing furnaces will reach about  7.65 Gg (8,500 tons) by 1985.  The steel industry

                                                  5-70

-------
and equipment builders are considering coal firing the palletizing 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 5-33).
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.
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 th« 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 106  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 I6T estimates (Reference 5-33)  show that NO  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 5-72).
                                                  5-71

-------
       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/Nm1, or, 92 Btu/ft* (Reference 5-72).
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 bums 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 5-73).  Much of the remaining heat in the partially spent combustion
gases 1s 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 NOX 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

                                               5-72

-------
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.

       Although it may be reasonable  to assume  that substitution of form coke  may result  in a  sub-
stantial reduction in NO  production, the general -opinion  in  the steel  industry  is  that form coke
will not be a significant factor  in 1985 (Reference 5-33).
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/Nm3, 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 NOX
stack-gas concentration, the NOX  emission.in 1985 will be  17.7 Gg/yr (19,600 tons/yr).  Because of
the low estimated NO  concentration and the presence of inerts in  the fuel gas, equivalent to  flue-
gas recirculation, the potential  for NO  reduction  is probably small (Reference 5-33).

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.
 Basic Oxygen Furnace                           5-73

-------
       The optn hearth furnace 1s both reverberator? «nd regenerative,  like  tht  glass  melting  fur-
naces   It 1s reverberator? In that the charge is wlted 1n «  shallow hearth by  heat fro* a  flaw
passing over the charye and by rac'latlon frcw the hefted dome.  It Is regenerative 1n that the
remaining heat 1n the partially spent combustion gases from the reverberatory chamber 1s accumulated
In a brick filled chamber, or "checker", and released to preheat the Incoming com&ustlon air when
the cycle 1s reversed.  Fuel of low calorific value such as blast furnace g«s as well  as the com-
twitlor. 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 11me are the usual  materials
charged to an or>en hearth furnace.  These are heated over a period averaging 10 hours, at a tempera-
ture as high as the refractories will permit.  Fuel of! Is the preferred fuel and 1s burned with
excess *1r to provide an oxidizing Influence on the charge.
       NOX emissions from open hearth furnaces ere very high because of the high combustion air- pre-
heat temperature, high operating temperature, and the use of oxygen Unces to Increase production
rates.  The data avullable  Indicate that NOX concentrations will be In tie  1000 to  2000-ppm range.
Although many open hearths  ere 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 1n 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 1t 1s anticipated that about 13.5 Pa. (15
million tons) will still be trade by the open hearth process in 1985.  Fuel consumption has been
decreasing and may reach 2.9 MJ/kg (2.5 million Btu/ton) 1n 1985.  This will require a fuel con-
sumption of 40 PJ (37.5 trillion Btu) for open hearth steel production and result in an NO  emis-
sion level of 14 Gg  (15,750 tons)  (Reference 5-33).
 Basic Oxygen  Furnace
        In the basic oxygen furnace  (BOF), oxygen  1s blown downward through a water-cooled lance Into
 a bath  containing  scrap and hot metal.  Heat produced by oxlditlon of carbon, silicon, manganese, and
 phosphorous  Is sufficient to  bring  the metal to pouring temperature and auxiliary fuel 1s not  required.
 The furnace  1s an  open top, tillable, refractory-lined vessel  shaped somewhat like the cld-fashloned
 glass  nrilk bottle.  Furnace capacities range up to 309 Mg  (340 tors).  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
                                                5-74

-------
mount of BOF capacity 1n an Integrated $te«l  plant 1s, therefor*, closely associated with hot  Mtal
availability.  Acdltlonol 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 1t can be used In conjunction with BOF shops to
provide the required flexibility to accommodate variations 1n hot metal-scrap ratio.  A combination
of BOF shops and eiectrlc furnace shop't provides the maximum 1n flexibility and nay re^esent the
makeup of future steelmakliig facilities.
       Excluding fuel use for scrap preheating, other uses are for refractory dryout and to keep
the BOF vessel from cooling between heats.  Their us&s amount to about 232 kg  per kg (200.000 Btu
per ton) of steel produced.
       DecarbuHzatlon of the Iron charged to the BOF produces tbout 467 kJ of carbon monlxlde per
kilogram of steel (400,000 Btu/ton).  The off-gases also contain large amounts of participates,
wMch must be removed before discharge Into the atmosphere.  Typical American practice 1s tc bum
the combustible gases 1n w«ter-cooled hoods mounted above the BOF vessel, cool with excess air or
water sprays, and pass the cooled gases through h
-------
       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 NO- emission level.  At the present time, only experimental information is
available concerning the effect of these parameters on NO  levels.
       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 NOV output is the trend toward use of continuous casting to replace some ingot casting.  In
         A                                                             ,       '    ' •         '
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 I6T projection,
conventional steel processing will account for 144 Tg (160 x 10s tons) in 1985.   At present fuel
consumption of 5.4 MJ/kg (4.7 x 10s  Btu/ton), the total fuel consumed for soaking-pits and reheat
furnaces in 1985 will be 795 PO (750 x 1012 Btu).  This fuel consumption will result in estimated
HOV emissions of 143 Eg (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 101!  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 NOV emission in 1971 for this category will  be about 7.6 Gg (8,400
                               A
tons).  Assuming that production of cold-rolled products remains at about 25 percent of total  steel
production, the 1985 NOX 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 5-33).
                                                 5-76

-------
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 production transfers the NOX emissions to the
utility plant where the problem is easier to control.  Electric furnaces are, in any case, a very
minor source of NO  from the steel industry (Reference 5-33).

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 SOg, 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 NO  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 NOX concentrations in stack gases (Reference
5-33).  This relationship is affected by other variables, such as combustion air preheat temperature
and oxygen enrichment of combustion air.
       Table  5-15shows the estimated NO  concentrations for the major energy-intensive processes
                                       or
 energy consumption data  (Reference 5-33).
and the resulting total annual combustion-related NOX production based on 1971  steel  production
                                                 5-77

-------
                         TABLE  5-15.  ESTIMATED NOX EMISSIONS FROM STEEL HILL PROCESSES AMD
                                      EQUIPMENT (Reference 5-28 except where noted)
Process
or
Equipment
Pelletizlng
Sintering
Blast Furnace
Coke Oven
Blast Furnace Stove
Open Hearth Furnace
Basic Oxygen Furnace
Soaking Pit and
Reheat Furnaces
Heat Treating and Finishing
Electric Furnaces
Annual Fuel Consumption
PJ
29
98
; ndb
225
i 212
135
; nd
541
64
; nd
10" Btu
27
93
nd
212
200
127
nd
510
60
nd
NO Emission Factors3
A
ppm
(as measured)
300
500
' (230)e
negc
200
(10-485)6
100
600
nd
300
(92)e
200
6-25
ng/J
180
300
neg
120
60
360
nd
180
120
nd
Ib/lO* Btu
0.42
0.70
neg
0.28
0.14
0.84
nd
0.42
0.28
nd
Annual NO^ Emissions3
Gg tons
5.1 5670
29.6 32550
neg neg
26.9 24680
12.6 14000
48.3 53340
nd nd
96.4 107100
7.6 8400
.02-. 09 26-110
Notes:    expressed as NO,
        l                L-
         nd = no data
         neg « negligible  emissions
         Reference 5-75
        American Iron and Steel  Institute  test data provided by Dr. Walter Jackson (U.S. Steel), (Reference 5-74)

-------
       Other test results provided by the American Iron and Steel  Institute (Reference 5-74}  indi-
cate different emission factors as shown in parentheses in Table 5-15-. The emission levels for the
coke ovens are the result of three separate tests (10, 186, and 485 ppm).

        Results of recent tests reported by Hunter, e_t a_l_. (Reference 5-34) are summarized in Table
 5-16. The open hearth furnace was tested while operating on natural ;as and Number 6 fuel oil
 (50/40).  The wide fluctuations in NO  and CO observed as various operations were performed are
                                      A
 shown in Figure'5-18.  Large changes in excess air occurred as the operators opened doors to look at
 the steel and to add material or adjust fue'l  flow to change heating rate.  NOX 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-  ParticuTate 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
                                                                                                 *v
 was 1094 ng/J (1250  ppm), a reduction of about 40 percent.  During baseline tests, N0x frequently
 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 NO  reduction, and burners out
 of service which produced a 43 percent NO  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 MHBtu/hr) through a single burner.  Baseline NOX emissions at 2 MW were 52 ng/J (101 ppm).
 Reduction of excess  air reduced NOX fay 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:
                                                5-79

-------
                                TABLE 5-16.   EFFECTS OF NOX  CONTROLS ON STEEL INDUSTRY

                                             NOX  EMISSIONS  (Reference 5-34)
in

&>
O
Device
Type
Steel Open Hearth
Furnace.
Steel Reheat
Furnace
Steel Soaking Pit
Fuel
Nat. Gas +
No, 6 Oil
Nat. 'Gas
Nat. Gas
Average Baseline
N0x
ng/J
1094
56
52
ppm
@ 3$ 02
2070
no
101
Max.
Ctf
/o
Reduction
40
43
69
Combustion
Modification
Low 00-
3/27 BOOS
Low 02

-------
  4000
  3500
  3000
  2500
cs
9
 »
CM
a


&
Ui
u
cc
vu
a.
2000
  1500
  1000
            BASELINE TEST
         N0.6QILAND GASFliEL
   SOO	
    1200
             130G
1400
1500       1600       1700

     TIME OF DAY, hr
                                                                 1800
                                                                         1SQO
2000
           Figure 5-18.   NOX emissions as  a function of time for an
                           op£n hearth furnace (Reference 5-34}.
                                            s-ai

-------
       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 NOV control.
                                                                           A
                            *
       Continuous reverberatory furnaces have a hoi ding'capacity of up to 1.27 Gg (1,'400 tons) and a
daily output of as much as 270 Hg (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 c'heckerbricks, 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 nir and exhaust gases passing through the reversal  valve, and
                                                5-82

-------
combustion 1s then reestablished in what was previously the exhaust port.  Both types of me Hers 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-type 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 5-33).
       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 NOX 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 1s the major source of NO  emission 1n 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, NO  emissions were measured during a complete firing cycle of a
glass meltef.  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 5-17 summarizes the emissions from several glass melters as measured by a number of
investigators.
                                                 5-83

-------
CO
CO
s:
o
a: co
u. LU

CO et
z: z
o a:
HH Z3
co u.
LU l-H


 xti
O LU
 t
LA
LU


CQ







C/)
4«^
C

g
Is
o
0








'•a"
cu
s_
3
CO
. (O
xai
f^ S
*s^
CO
re
g
Q.
o.







i-
o

(O
Cft
•1—
•M
CO
CU

C
1— 1









&-
re

CO
CO
a;
o
X
cu

^s>
oo

««i
•o
cu
s.
*^»
«*-

CO
re
en


o
o
'"•
0
4-»

0
CT>

•










*—+
o>

1
LO

(U
a
c:
£
.«
•fr™

•a
02





j™
3
O
c

C
3

S-

'Jo

CO
co
cu
u
X


LO




•x_
?
LO
0}
^~+ U
vo c
r- a>
1 y^
LO  CM

o
re
ai
>
— '




^^.
oo
r*.
i
LO

CU
u
c
CD
S-

lf"»
CU
oc

c
o
CO

*cu

c
(O
o
                             5-84

-------
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 to 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, SOX, NOX, and particulates are the primary air pollutants from the glass manu-
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 tit
the combustion process is  NO  emissions.

        While the formation of NOV in the combustion process is not entirely understood, it is clear
                                 A
that the goals of reducing NOY emissions and reducing energy consumption are seemingly at odds.  NO
                             A                                                                     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 NO  emissions.  Analysis of the process modifications under
                        ..                 *
consideration in the glass industry shows that there is a possibility of increasing NO  emissions.'
If the implementation is carried out properly, however,'this need not occur.
        Six recommended modification programs are listed in Table 5-18.  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 5-33).   Cost data for these programs are not available
at this time.  Two of the six recommendations are currently, being pursued by EPA/1ERL-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:  port!and 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
                                                 5-85

-------
                                TABLE 5-18. RECOMMENDED PROGRAMS FOR REDUCING EMISSIONS AND ENERGY
                                            CONSUMPTION IN THE GLASS INDUSTRY (REFERENCE 5-33)
                         Program
                                        Expected Improvements in
                                         Energy Consumption, %
                             Expected Improvements in
                             Air Pollutant Emissions
oo
.en
            1.  Develop current emission data
            2.  Raw batch pretreatment — i.e.,
                preheating and agglomeration
            3.  Oxygen enrichment
4.  Augmentation of heat transfer
    from flames — e.g.,  burner
    positioning
            5.  Use of low-temperature heat
                for driving compressors
            6.  Development of submerged
                combustion process
25-50



 5-15


10-20
                                                50
                                                                      25% to 50% potential NOX reduction,
                                                                      may reduce particulate in form of
                                                                      batch carry-over

                                                                      No effect on NOX, SOX, or particulates
Proportional NOX reduction
                       Will substantially reduce NOX
                       emissions

-------
cement manufactured in the United States, with the remaining  5  percent split among the other
types.
       Raw materials used in the manufacture of portland 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, sil'icon, 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 5-33).

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
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 MQX emissions from cement plants in 1972 are 42.7 Gg
(4.7 x 10* 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.
                                               5-87

-------
Conversely, adequate removal of the additional heat resulting from the applicable process modifica-
tions should maintain HO  emissions at their current level.
                        J\                                                   :
       Of the process modifications deemed to be near-term, only the use of oxygen enrichment has
any great potential of increasing air pollutant emissions, primarily NOX-  In some applications in
other industries, for example, glass melting, oxygen enrichment can be used without increasing NO
emissions.  However, due to the different type of load in the cement industry and the different
patterns of heat transfer, it is suspected that NO  would increase with the implementation of oxygen
enrichment (Reference 5-33).

5.3.4.3  Control Techniques
     /
       There is very little information in the literature regarding commercial installation of equip-
ment for removing NOX from kiln waste gas or of modifications to kiln operations to reduce NOX
production.  Water scrubbing is sometimes used for particulate removal from waste gas from lime
kilns.   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 recirculation, which is used
                          Ct              "                                  i
to control temperature in some lime kilns, should reduce NO  emissions by lowering flame temperature.
       Reference 5-34 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 kiln 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.
                                                                            I
       Lower NOX emissions on solid and liquid fuels compared to gas are attributed to the highly
adiabatic nature of the process.  Many cement kilns are currently being converted from gas to solid
fuels.   This conversion wil-1 be beneficial in reducing NO  and could be pursued as an NO  control
                                                         A                              A
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 ppm).  Combustion modifications Investigated included variation of
combustion air inlet temperature and excess oxygen.  Increase of combustion air temperature from
644K (700F) to 767K (920F) increased NO emissions to 1518 ng/J, and 15 percent increase.   Reduction
of excess oxygen at baseline air temperature reduced NO  to 846 ng/J, a 36 percent reduction.   The
                                                                            I
Independent reductions of either excess air or air temperature caused unacceptable reduction of
kiln temperature that can result in a process upset.  The NO emissions were found to  be a  strong
                                                 5-88

-------
function of kiln temperature, as shown in Figure 5-19.  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 using 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 NOX 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 po-llutant 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 5-81).  These emission rates are below the
NSPS of 300 ng/J (0.7 lb/108 Btu) for large steam generators.  In any case, no NOX standards have yet
been proposed since the thermal dryer capacities are generally less than the smallest power plants
required to control NOX emissions:  73.2 HW, .or 250 x 10s Btu/hr (Reference 5-81>.
                                                 5-89

-------
    1422
  30QO,—
1500
KILN TEMPERATURE, °K

  1600               1700
1800
1867
  2500
^2000
ec
CM
O
  1500
u
ec

  1000
   500
    0
    2100
        SHADED AREA SHOWS EFFECT OF INLET AIR TEMPERATURE VARIATIONS

        NUMBERS REPRESENT EXIT 02 CONCENTRATIONS IN PERCENT       ~
        ROTARY CEMENTKILN, WET PROCESS
     2300
         2500

    KILN TEMPERATURE, °F
                                             2700
           2900
          Figure 5-19.   The effect of cement kiln temperature  on NOX
                         emissions .(Reference 5-34).
                                          5-90

-------
                                REFERENCES FOR SECTION 5


5-1       Castaldini, C., e£al., "Combustion Modification Controls for-Residential  and Commercial
          Heating Systems:  VFTume I.  Environmental  Assessment," EPA-600/7-81-123a, July 1981.

5-2       Hall, R. E., 0. H. Wasser, and E. E. Berkau, "A Study of Air Pollutant Emissions from
          Residential Heating Systems," EPA-650/2-74-003, January 19.74.

5-3       Barrett, E. R., S. E. Miller, and 0. 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., e£ aj_., "Status of EPA's Combustion Research Program for Residential  Heating
          Equipment," presented at the 67th APCA Annual Meeting, June 1974.

5-5       U.S. Environmental Protection Agency, "Compilation of Air Pollutant Emission. Factors,"
          AP-42, February 1980.

5-6       Allen, J. M., "Control of'Emissions from Residential  Wood Combustion by Combustion Modifi-
          cation" Proceedings of the Joint Symposium on Stationary Combustion NO  Control. Vol.
          III.. IERL-RTP^1085, October 1980.                                    x

5-7       Locklin, 0. W. and R. E. Barrett, "Guidelines for Residential  Oil Burner Adjustments,"
          EPA-600/2-75-069a, October 1975.

5-8       Locklin, D. W. and R. E. Barrett, "Guidelines for Burner Adjustments of Commercial
          Oil-Fired Boilers," EPA-600/2-76-088, March 1976.

5-9       DeWerth, D. W., R. L. Himmel, and 0. W. Locklin, "Guidelines for Adjustment of Atmospheric
          Gas Burners for Residential and Commercial  Space Heating and Water Heating,"
          EPA-600/8-79-005, February 1979.

5-10      Thrasher, W. H. and D. W. DeWerth, "Evaluation of the Pollutant Emissions from Gas- Fired
          Air Furnaces," Research Report No. 1503, American Gas Association, Cleveland Laboratories,
          Cleveland, Ohio, May 1975 as cited in Okuda, A. S. and L. P. Combs, "Design Optimization
          and Field Verification of an Integrated Residential Furnace - Phase 1," Rockwell Inter-
          national, Rocketdyne Division, EPA-600/7-79-037a, February 1979.

5-11      Kalika, P. W., G. T. Brookman, and J. E. Yocum, "A Study on Measuring the Environmental
          Impact of Domestic Gas-Fired Heating Systems," Fipal  Report, The Research Corporation of
          New England, June 1974 as cited in Okuda, A. S. and L. P. Combs, "Design Optimization and
          Field Verification of an Integrated Residential Furnace - Phase 1," Rockwell
          International, Rocketdyne Division, EPA-600/7-79-037a, February 1979.

5-12      De Angelia, D. G. and R. B. Reznik, "Source Assessment:  Residential Combustion of Coal,"
          Monsanto Research Corporation, EPA-600/2-79-019a, January 1979.

5-13      Martin, G. B., "Evaluation of a Prototype Surface Combustion Furnace," Proceedings of the
          Second Stationary Source Combustion Symposium, Volume III, EPA-600/7-77-073c,
          NTIS-PB 271 757, July 1977.
                                                                             f
5-14      Personal communication with Chuck Mueller, Amana, Inc., Amana,  10, April 14, 1979 as cited
          in Reference 5-2.

5-15      Hall, R. E., et al., "Study of Air Pollutant Emissions from Residential Heating Systems,"
          EPA-650/2-74-WTNTIS-PB 229697, January 1974.

5-16      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.

5-17      Comb, L. P. and A. S. Okuda, ".Residential Oil Furnace System Optimization - Phase I,"
          Rocketdyne Division, Rockwell International, EPA-600/2-76-038,  February 1976.

5-18      National Petroleum News, January 1975, pp.  34-35.
                                              5-91

-------
5-19      Personal Communication, Lenney, R.  J., Blueray Systems,  Inc., Weston,  Massachusetts,
          September 1975.

5-20      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.

5-21      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
          Administration, Rockville,  Maryland,  1968.

5-22      Niessen, W. R., et jil_., Systems Study of Air Pollution from Municipal  Incineration, Report
          to NAPCA under contract CPA 22-69-23, Arthur D. Little,  Inc., Cambridge,  Mass., 1970.

5-23      McGraw, J. J. and R. L. Duprey, Compilation of Air Pollutant Emission  Factors (Revised),
          AP-42, EPA, February W72.

5-24      Stenburg, R. L., et cil_., "Field Evaluation of Combustion Air Effects on Atomspheric
          Emissions from Municipal Incinerators,"  J. Air Pollution Control Assoc.,  Vol. 12,
          pp. 83-89, February 1962.

5-25      Kirsh, J. B., "Sanitary Landfill,"  In:  Elements  of Solid Waste Management Training Course
          Manual Public Health Service, Cincinnati, Ohio, 1968, p, 1-4.

5-26      Fife, J. A., and R. H. Boyer, Jr.,  "What Price Incineration Air Pollution Control?,"
          Proceedings of 1966 National Incinerator Conference, American Society  of  Mechanical
          Engineers, flew York, 1966.

5-27      OAQPS Data File of Nationwide Emission,  1971.  National  Air Data Branch Monitoring and
          Data Analysis Division, May 1973.

5-28      Chi, C. T., and D. L. Zanders, "Source Assessment:   Agricultural Open  Burning,
          State-of-the-Art," EPA-600/2-77-107a, July 1977.

5-29      Air Pollution Problems from Refuse  Disposal Operations in the Delaware Valley,  Department
          of Public Health, Air Management Services, Philadelphia, Pa., February 1969.

5-30      Wiley, J. S. et al_., "Composting Developments in  the U.S.," Combust. Sci. 6(2)-.5-9, 1965.

5-31      Kurkey, C., "Reducing Emissions from  Refuse Disposal," J. Air Pollution Control Assoc.,
          19: 69-72, February 1969.
           *
5-32      Personal Communication, Mr. Peter L.  Cook, Office of Federal Activities,  U. S.
          Environmental Protection Agency, November 1977.

5-33      Ketels, P. A., J. D. Nesbitt, and R.  D.  Oberle, "A Survey of Emissions Control  and
          Combustion Equipment Data in Industrial  Process Heating," Final Report by Institute of Gas
          Technology for EPA, IGT Project No. 8949, June 1976.

5-34      Hunter, S. C., e£ al., "Application of Combustion Modifications to Industrial Combustion
          Equipment, "ProceedTngs of  the Second Stationary  Source  Combustion Symposium. Vol. Ill,
          EPA-600/7^77-0736, July Wf!

5-35      Cantrell, A.  Annual Refining Survey.  Oil and Gas Journal.  79(13):  110-153.  March 30,
          1981.                                                      ~

5-36      Hunter, S. C. and S. C. Cherry.  (KVB) NO  Emissions from Petroleum  Industry Operations.
          (Prepared for the American  Petroleum  Institute.)   Washington, D. C.  API  Publication
          No. 4311.  October 1979.

5-37      Radian Corporation, unpublished data.  Source Category Survey Report on. NO  Emissions From
          Fired Heaters.  (Prepared for Emission Standards  and Engineering Division, U. S.
          Environmental Protection Agency, Research Triangle Park, North Carolina.)
          DCN 81-231-372-19-04, EPA Contract  No. 68-02-3058.   July 1981.
                                               5-92

-------
5-38      Cherry, S. C. and S. C. Hunter.  (KVB) Cost and Cost Effectiveness of NO  Control  in
          Petroleum Industry Operations.   (Prepared for the American Petroleum Institute.)
          Washington, D. C.  API Publication No. 4331.  October 1980.

5-39      Radian Corporation.  Assessment of Atmospheric Emissions From Petroleum Refining:
          Volume 5, Appendix F.  (Prepared for the U. S: Environmental  Protection Agency.)
          Publication No. EPA-600/2-80-075E.  Research Triangle Park,  N-.C.   July 1980.

5-40      Murcia, A. A., et £]_., Add Flexibility to FCC's.  Hydrocarbon Processing.   58_:134.
          September 1979.

5-41      Production by the U. S. Chemical Industry.  Chemical and Engineering News.  59(23):  33-35.
          June 8, 1981.

5-42      Hoover, J. R., J. R. Blacksmith, and P. W. Spaite.  (Radian  Corporation) Energy Use
          Patterns and Environmental Implications of Direct-Fired Industrial Processes.   (Prepared
          for U. S. Environmental Protection Agency.)  Cincinnati, Ohio.  EPA Contract No.
          68-01-4136.  August 1980.

5-43      Parsons, T. B., C. M. Thompson, and G. E. Wilkins.  (Radian  Corporation) Industrial
          Process Profiles for Environmental Use:  Chapter 5.  Basic Petrochemicals  Industry.
          (Prepared for U. S. Environmental Protection Agency.)  Washington, 0. C.
          EPA-600/2-77-023e.  January 1977.

5-44      Kent, J. A. (ed.) RiegeVs Handbook of Industrial Chemistry.   Seventh edition.  New York.
          Van Nostrand Reinhold Company, 1974.

5-45      1979 Petrochemical Handbook Issue.  Hydrocarbon Processing.   5_9_:154.  November 1979.

5-46      Considine, D. M. (ed.).  Chemical and Process Technology Encyclopedia.  New York,
          McGraw-Hill Book Company.  1974.

5-47      Berman, H. L.  Fired Heaters - II:  Construction materials,  mechanics! features,
          performance monitoring.  Chemical Engineering.  85_(17):89.  July 31, 1978.

5-48      Smith, T. M.  Applying Ceramic Fib«r Furnace Linings.  Hydrocarbon Processing.
          60:169-172.  April 1981.

5-49      Berman, H. L.  Fired Heaters - I:  Finding the Basic Design for Your Application.
          Chemical Engineering.  85_(14):99.  June 19, 1978.

5-50      Safety in High-Temp Heat Transfer Ruid Systems.  Chemical Engineering.  88(9): 14-15.
          May 4, 1981.                                                            ~~

5-51      Ayraud, S. L;, and R; J. Schreiber.  (Aerotherm/Acurex) Refinery Heater Screening  Study.
          (Prepared'for U. S. Environmental Protection Agency.)  Research Triangle Park, North
          Carolina.  Publication No. EPA-450/3-79-007.  March 1979.

5-52      Berman, H. L.  Fired Heater III:  How Combustion Conditions  Influence Design and
          Operation.  Chemical Engineering.  85_(18): 138.

5-53      Evans, F. L.  Equipment Design Handbook for Refineries and Chemical Plants:  Volume 2.
          Houston, Gulf Publishing Company, 1980. p. 9.

5-54      Goyal, 0. P.  Guidelines Help Combustion Engineers.  Hydrocarbon Processing.   59:  209.
          November 1980.

5-55      Berman, H. L.  Fired Heaters - IV:  How to reduce your fuel  bill.  Chemical Engineering.
          85_(20): 166-7.  September 11, 1978.

5-56      Reed, R. D.   Furnace Operations.  Houston, Gulf Publishing Company, 1976.

5-57      Waterland, L. R., et al_., (Acurex/Aerotherm)  Environmental  Assessment of Stationary
          Source NO  Control~Technologies:  Second Annual Report.  (Prepared for U.  S.  Environmental
          Protection Agency.)  Research Triangle Park, North Carolina.   Publication No.
          EPA-600/7-79-147.
                                              5-93

-------
5-58      Martin, R. R. Burner Design Parameters  for Flue Gas  NO  Control.  John Zink Technical
          Publication No. 4010.  Tulsa, Oklahoma, John Zink Company,  1981.

5-59      Coe, W. W.  How Burners Influence Combustion.   Hydrocarbon  Processing.  60: 181.  May
                                                                                 ~~
5-60      Cox, N. D., and W.  Paul  Jensen.   Improving Fired Heaters  Saves  Fuel.  Oil and Gas Journal.
          77 (49): 73.  December 3, 1979.

5-61      Radian Corporation, unpublished  data.   Interim Data  Collection  and Analysis for
          Development of New Source Performance  Standards for  NO Emissions from  Fired Heaters.
          (Prepared for Emission Standards and Engineering Division,  U. S. Environmental Protection
          Agency.)  Research  Triangle Park, North Carolina.  DCN 82-231-372-19-15.  Contract
          No. 68-02-3058.  May 1982.

5-62      Tidona, R. J., W. A. Carter, and H.  J.  Buening.  (KVB) Refinery Process Heater NO  Control
          Using Staged Combustion Lances.   (Prepared for U.  S.  Environmental Protection Agency.)
          Research Triangle Park, North Carolina.  EPA Contract No. 68-02-2645.   November 1981.

5-63      Tidona, R. J., H. J. Buening, and J. R. Hart.   (IVB)  Emissions  From Refinery Process
          Heaters Equipped with Low-NO  Burners.   Publication  No. EPA-600/7-81-169.  Research
          Triangle Park, North Carol in!.  October 1981.

5-64      Chapman, Kirk S. (Coen Company,  Inc.)  NO  Reduction  on Process  Heaters  With a Low NO
          Burner.  Presented  at the 72nd Annual  Meeting of the  Air  Pollution Control Association.
          Cincinnati, Ohio.  June 24-29, 1979.

5-65      State of California Air Resources Board.  Public Meeting  to Consider a  Suggested Control
          Measure for the Control  of Emissions of Oxides of Nitrogen  from Boilers and Process
          Heaters in Refineries.  (Prepared by the staff of the South Coast Air Quality Management
          District and Stationary Source Control  Division Air  Resources Board.)   October 1981.

5-66      State of California Air Resources Board.  A Suggested Control Measure for Emissions of
          Oxides of Nitrogen  from Boilers  and  Process Heaters  in Refineries.  (Prepared by the staff
          of the South Coast Air Quality Management District and Stationary Source Control Division
          Air Resources Board.)  March 1982.

5-67      Schultz, E. J., L.  J. Hellenbrand, and R.  B. Engdahl, "Source Sampling  of Fluid Catalytic
          Cracking Plant of Standard Oil of California,  Richmond, California,"  Batelle-Columbus
          Labs, July 1972.

5-68      Schultz, E. J., L.  J. Hellenbrand, 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.

5-69      Shea, E. P., "Source Testing, Standard Oil Company,  Richmond, California," Midwest
          Research Institute, Kansas City, Missouri, 1972.

5-70      Gowherd, C., "Source Testing, Standard Oil of California  Company, El Segundo, California,"
          Midwest Research Institute, Kansas City, Missouri, 1972.

5-71      Shea, E. P., "Source Testing, Atlantic Richfield Company, Wilmington, California," Midwest
          Research Institute, Kansas City, Missouri, January 1972.

5-72      McGannon, H. E., The Making, Shaping and Treating of  Steel, 8th ed. , Pittsburgh, United
          States Steel Co.,~T567T                          ~~

5-73      Russel, C. C., "Carbonization."   In:  Kirk-Othmer Encyclopedia  of Chemical Technology.
          Standen, A. (ed.),  Vol.  4,  2d ed., New York, Interscience PublilFers, Inc., 1964, p.
          400-423.

5-74      Personal communication, Dr. Walter Jackson, U.  S.  Steel, November 1977.

5-75      Personal Communication, Mr. Andrew Trenholm of the Office of Air Quality Planning and
          Standards, U. S. Environmental Protection Agency,  Durham, North Carolina, May 1976.
                                              5-94

-------
5-76     , Stockham, J.  D., "The Composition of Glass  Furnace  Emission,"  presented  at the 63rd Annual
          meeting of the Air Pollution Control  Association, St.  Louis, June  1970.

5-77      Mills, J. L., e£ a]_., "Emissions  of Oxides  of Nitrogen from Stationary Sources in
          Los Angeles County;  Oxides of Nitrogen Emitted by Medium and Large Sources," Joint
          District, Federal, State, and Industry Project, Los Angeles County Air Pollution Control
          District, Los Angeles, Calif., Report Number 3, April  1962.

5-78      Air Pollution Engineering Manual, Daniel son, J. A.  (ed.).   National  Center for Air
          "FoTlution Control, Cincinnati, Ohio, PHS  Publication No.  999-AP-40.

5-79      Nesbitt, J. D., D. H. Larson, and M.  Fejer, "Improving Natural  Gas Utilization in a
          Continuous End Port Glass-Melting Furnace," In: Proceedings of the Second Conference on
          Natural Gas Research and Technology, Session IV, Paper 9,  Chicago, 1972.

5-80      Ryder, R. J., and J.  J. McMackin, "Some Factors Affecting Stack Emissions from a Glass
          Container Furnace,"  Glass Ind., 50, June-July 1969.

5-81      "Background Information for Standards of  Performance:   Coal Preparation  Plants; Volume 1:
          Proposed Standards,"  EPA 450/2-74-021 a, October 1974.
                                                      5-95

-------

-------
                                              SECTION 6
                                       NONCOMtJUSTION PROCESSES

       The problem of NO  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 NOX control in the manufacture and
uses of nitric acid.
       The NO  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 N02 and NO present.  The N0x-containing chemical gas is commonly a process stream which
must be recycled with maximum NO  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 NOV,
                                                                                                 A
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 NOX, or 10,000 ppm.
       The total amount of NO  emitted from all chemical manufacturing is about 1.7 percent
(203 Gg or 2.2 x 10s 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 NO^/NO 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 NOX at a 1:10 ratio of N02: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 N0« at 10,000 ppm may be visible but less hazardous than
many tiroes- as much NO  emitted from a large stack at a lower concentration.  The total  amount in  a
                     A
short, concentrated emission may be too small to have a detectable effect on NOV levels in ambient
                                                                        f       X
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 add 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
                                      (AH   = -906 J/mole)
The stream 1s cooled to 311K (100F) or less, and the NO then reacts with oxygen to form nitrogen
dioxide and Its liquid dlmer, nitrogen tetroxide.
                              2ND  +  02   *  2 N02 *  N204
                                      (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 N02 (or N204) reacts with  water to form nitric acid and NO,  as
follows:
                                                 6-2

-------
                              3N02  +  H20  -*•  2HN03  +  NO
                                                                                            (6-3)
                                      (AH   3  -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 irrconcentrations of 55 to 65 percent.
The air entering the bottom of the tower serve;; to strip N02 from the product and to supply oxygen
for reoxldizlng the NO formed In making nitric add (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 11(5 psia) or at high pressure (1000 to 1200 kPa or 145 to
174 psia).  Both operations may be at the same pressure or at different pressures.
       Before corrosion-resistant materials were developed, the ammonia 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 th« 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  stm 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 oxidatjon reaction shortens catalyst's lifetime  (1 to 2
months) through increased erosion and lowers the ammonia 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 NO  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 1s 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 NO  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


AMMONIA
(ANHYD
ROUS)
COMPRESSOR

EVAPORATOR




F
•W


WEAK k ^
ACID ] *"
iiTcn AIR au

kTERj
I
ABSORBER
                                                             NITRIC ACID

Figure 6-1. Single pressure nitric acid manufacturing process (Reference 6-3).
                            6-5

-------
1. AMMONIA EVAPORATOR             9.
2. AMMONIA PREHEATER            ; 10.
3. AMMONIA FILTER                j 11.
4. AIR FILTER                     i 12.
5. AIR COMPRESSOR                 13.
6. AMMONIA-AIR MIXER              14.
7.  BURNER WITH LA MONTE BOILER     15.
8.  TAIL GAS PREHEATER I
                                      GAS COOLER
                                      TAIL GAS PREHEATER II
                                      ABSORPTION COLUMN
                                      DEGASSING TOWER
                                      MIXING JET
                                      NO COMPRESSOR
                                      TAIL GAS TURBINE
Figure 6-2.  Dual pressure nitric acid plant flow diagram (Reference 6-2).

-------
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 NO- 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 N0« or N-O. 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 NO.
                                                 6-7

-------
                                                                   TAIL GAS TO
                                                              ATMOSPHERE (VOLUME %)
                                         COUNTERCURRENT
                                            CONDENSER
 VAPOR
98XHN03
FEED
HN03
      DEHYDRATING
        COLUMN
                                     VAPOR
                           i
             1
CONDENSATE
                        BLEACHER
                          NCIN-
                      CONDEIUSABLE
                         GASES
                           T
                       95-99% HN03
                  TO COOLER AND STORAGE
                   VAPOR
           ILIQUID
     STEAM-
      COIL •
                                       74.3 N2
                                       20.4 02
                                        1.0NO + N02
                                        4.2 H20
                                           \
ABSORPTION
 COLUMN
                         BOILER
                70% H2S04
                TO COOLER
    I
                                                                                .AIR
                                                                    55% HN03
                          Figure 6-3. Nitric acid concentrating unit.
                                           6-8

-------
oc
3
                 1.   AIR FILTER
                 2.   AMMONIA-AIR MIXER
                 3.   AMMONIA GAS FILTER
                 4.   BURNER WITH WASTE-HEAT BOILER
                 5.   GAS COOLER I
                 6.   BLOWER
                 7.   GAS COOLER II
                 8.   OXIDATION COLUMN
 9.   ACID COOLER             17.
ID.   CIRCULATING ACID TANK   18.
11.   ACID COOLER             19.
12.   FINAL OXIDATOR          20.
13.   BRINE GAS COOLER        20.
14.   ABSORPTION COLUMN      21.
15.   TAIL GAS SCRUBBER       22.
RAW ACID COOLER
LIQUEFIER
AGITATOR VESSEL FOR
RAW MIXTURE
REACTOR VESSEL
BLEACHING COLUMN
FINAL ACID COOLER
16.   HEAD TANK FOR RAW ACID  23.  MIXING JET
            Figure 6-4. Process flow diagram for direct production of highly concentrated nitric !acid (Reference 6-2).

-------
The overhead vapors are compressed to a pressure that allows the equilibrium di-nltrogen  tetroxide
                                                                   'ii	 :,  i"              '         .
(N20.) to be liquefied with the use of cooling water alone.   The liquid  NgO^  is  converted  to  nitric
add of about 95 to 99 percent by reacting the N204 with oxygen at a pressure of  5000 kPa (50  atm).
The conversion reaction is:  2N20.  +  2H20  •*•  4HNO,.  Tail gases from  the absorption column  are
scrubbed with water and condensed N^O* in a tall-gas scrubber before being released.   The liquid
from the tail-gas scrubber is mixed with the concentrated add from the  absorption column,  which
has been bleached and liquefied.  The combined product liquid (containing  N20. as well as HNO,) 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
Ooliet, 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
1n operation at present, although, both were designed to meet the New Source Performance  Standards.
       In the Davy Powergas SA6AR process (Reference 6-4), like the Uhde process, ammonia and  air
art 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 add.  As in the Uhde process, N02 1s then produced from the product gases  and  absorbed
1n concentrated nitric acid.  However, whereas Uhde forms N204 from this liquid,  and  reacts the N20.
                         »
with oxygen, the SABAR process takes the concentrated HN03 to a vacuum rectification  column, where
concentrated HNO, 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
1n 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-
                                              4                          '
cent nitric add, 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 NOX 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 NOX 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 N0y Emissions from Nitric Acid Plants

       Nitric acid plants can be designed for low NOX 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  NOX 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 riecessary.                                         >
       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 NOX are described.  These  techniques may also be
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

-------
                                             TABLE  6-1.  HOX ABATEMENT METHODS OH NEH OR EXISTING HITRIC ACID PLANTS
Process
Chilled
Absorption
Extended
Absorption
yet Chemical
Scrubbing
Method
Increased solu-
bility of NOX 1n
chilled water
Increased absorp-
tion of NOX by
additional ab-
sorption equip-
ment
Scrubbing tail
gases with urea
solution or
ammonia to
recover NOX
*
Comments
Usually cannot meet
NSPS without other add-
on technology or low-
ered acid product
concentration
Inlet pressure of 760
kPa required (addi-
tional compressors
may be required)
Requires additional
compressor

Performs better at high
pressure but operable at
lower pressures. Recovers
ammonium nitrate and urea
solution. Requires re-
.friqeration.
May require an evaporator
to produce a concentrated
ammonium nitrate by prod-
uct. No refrigeration
required.
Licensors
COL-VITOK
TVA
J. F. Pritchard
(Grande Paroisse)
P. M. Weatherly
Chemico
Uhde
C&I Girdler
CoFAZ
MASAR {urea
scrubbing)
Norsk Hydro
(urea scrubbing)
Goodpasture
(ammonia
scrubbing)
Examples
2-318 Mg/d (350 tons/day) (with Gulf cata-
lytic reduction add-on). Nitram, Taiupa, Fla.
2-50 Mg/d (55 tons/day) plants at Muscle
Shoals, (1972)
327 Mg/d (360 tons/day) plant, Miss.
Chemicals, Yazoo City, Miss. 1973.
272 Mg/d (300 tons/day) Holston Army
Ammunition Plant, Kingsport, Tenn.
Cominco Plant, Beatrice Neb.
Kaiser. Tampa. Fla. and Bainbridae, Ohio
9 U.S. plants, 1 Japan plant (employs
chilled absorption process)
908 Mg/d (1000 tons/day) Monsanto,
Pensacola, Fla. 1977.
250 Mg/d (275 tons/day) plant, Allied
Chemical, Oman, Neb. 1975
None built to date
111. Nitrogen Pit., Marsalles, 111.
Air Products & Chem. , Pace, Fla.
Norsk Hydro, Proggunn, Norway
90 Mg/d (100 tons/day) Goodpasture pit.,
1974. Dimmitt, Texas
Chevron Oil Co., Richmond, Calif. 1976
C.F. Industries, Fremont, N.D.
2 scrubbers for 7 plants totalling 544
mg/d (600 tons/day).. Cyanamid, Wei land, Ont.
I
_J

ro

-------
TABLE 6-1.  NOX ABATEMENT METHODS ON NEW OR EXISTING NITRIC ACID PLANTS (Concluded)
Process
Catalytic
Nonselectlve
Catalytic'
Selective
Heterogeneous
Catalysis
Chemical
Absorption
Molecular
Sieve
• Method
Burns NOX and 02
with CH4 or H2 to
form N2, HzO, C02
Burns NOx with
ammonia to form
Hy and HyQi 02
not affected ~
Oxidation of NO •*
N0£ catalyzed by
heterogeneous ca-
talysis before
absorption
Oxidation with
KMn04 (KHn04
electrolytlcally
reclaimed
Absorption by
molecular sieve,
regeneration of
the sieve by
thermal cycling
Comments
• Consumes natural gas,
uneconomical If high
NOX or 02 content
(also reacts with 02)
• Hay be used In con-
junction with extended
absorption
• Energy recovery
possible
• Works at low or high
pressure
• Uses ammonia, can be
expensive to operate .
• Often used with ex-
tended absorption
• Works at low or high
pressure
• Energy recovery
usually not possible
• Can achieve very low
emission If desired
Limited success
Uneconomical not pres-
ently offered
• High energy and capi-
tal demands
• Hard to fit cycling
of sieve Into, con-
tinuous plant opera-
tion
Licensors
C&I Girdler
D. M. Weatherly
Cherolco
Gulf
Uhde (BASF cata-
lysts)
Mitsubishi
CDL/VITOK
Cams Chemical
Puraslv N
(Union Carbide)
Examples
Olln, Lake Charles, La. (also, Weatherby
plants)
IMC Corp., Strellngton, La. (1976) (with
extended absorption). 817 Mg/d (900 tons/
day ) . 1 977 . Col umbl a N1 trogen ,- Augusta ,
Ga.
Location not available
Nltram plants In Tampa, Fla., Installed
after CDL/VITOK process.
10 plants In U.S.
Plants 1n Europe and Japan
Under development
2 plants 1n Japan, not currently offered
1n U.S.
SO mg/d (55 tons/day) Hercules, Inc.
Bersemer, Ala. 1974
50 mg/d (55 tons/day) U.S. Army, Hols ton,
Kingston, Tenn. (Inoperable, dismantled)

-------
       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 toot 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 NOX 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 N02 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 N0« 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
1s sent to a bleacher where air removes entrained gases and further oxidation occurs.  The bleached
nitric add 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.  NO  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
add plant.
TVA Process
       The Tennessee Valley Authority, at their nitric acid plant in Muscle Shoals, Alabama,  designed
 and Installed refrigeration for NOV 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 ACID-**
                                                        COOLING
                                                         WATER
                                                                •MAKE-UP WATER
                                           PUMP
     Figure 6-5. Schematic diagram of the CDL/V1TOK NOX removal process(Reference 6-8).
                                         &-15

-------
kg/Kg of nftric add (5.5 Ibs/ton).  A flow diagram of their abatement equipment 1s 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 NO  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
tall 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 tall 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. H. Weatherly, Chemico, Uhde and C and I Grtdler (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 NO  concentration 1s above 3000 ppm (Reference 6-6).  There is  nc
                                           A
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
                                                                     , l •
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  Met Chemical Scrubbing
       Met chemical scrubbing uses liquids, such as alkali hydroxides, ammonia, urea and potassium
permanganate to convert NOg 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 

                WATER
(
                                                                                   COOLING
                                                                                WATER HEADER
             r£-
              r
                                                                                             7
COOLER CONDENSER
                                                     NITRIC ACID    GAS FROM AMMONIA
                                                          I
           OXIDIZER
                           Figure 6-6. TVA chilled absorption process (Reference 6-10).
                                                    6-17

-------
                                     TO EXISTING
                                  TAIL GASPREHEATER
I
r___L_
1
i EXISTING [.,, 	
ABSORBER r*1l —
1 1
1
I
|
NITROUS j
1
L --.—
NITRIC ACID }

1
1
i
1
1
|
1
j
"L

LEVEL
CONTROL
i Ml
	 i


	
SECONDARY
ABSORBER
TOWER




RECOVERED
NITRIC
AC
ACID TRANSFER
PUMPS
(ONE SPARE)
^1 ' *v
" W ^
^



S1 ^1
ID








,










^J \







FLOW
CONTROL
'vijL
i
i
i
.. i
PROCESS
WATER
ivn i bit
|j»^"°^™ ^^^
STARTUP^ STARTUP PROCESS WATER
ONLY PUMP RUMP
(EXISTING)
Figure 6-7, Grande Paroisse extended absorption process for NOx treatment.
                               ,6-18

-------
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).

                              HN02  +  CO(NH2)2  £  N2  +  HNCO  +  2H20                    (6-4)

                                 HNCO  +  HN02   *  N2  +  C02 +  HgO                       (6-5)

                              HNCO  +  H20 + H+ t  NH^  +  COg                             (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   *  N2  +  C02  +  NH4N03  +  ti£Q           (6-7)

As shown in reaction (6-7), half the nitrogen in the reaction will form NH^NOj, a valuable byr
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 n'ltric acid.  This chilled gas and condensate passes into
Section A of the MASAR absorber.  Meanwhile, the norma.1 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 COfOOg). N2, and HgO.  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 prdcess temperature in Section B.

                                                6-19

-------

TAIL GAS
- TO ^
NITRIC ACID
PLANT

SPENT MASAR
A (BLOW DOWN)
LIQUID
CHILLER
/ 1
A i ^
CONCENTRATED MASAR f V^V

TAIL GASES
FROM PLANT
*
TAIL GAS
.,„. 	 	 I7HHIFR

fiHIl 1 FR' ««. .__— _.,„.„. .1 V

* PUMP
FEED WATER


SECTION
x~x
l"1 L~^
'•^
SECTION
XX
_n_

/"'^
SECTION
XX
r
^ y
MASAR
ABSORBER

*g .

                TO NITRIC ACID PLANT
                 ABSORBER COLUMN
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 neat
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 acfd 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 1£!0 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 NOX 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 NOX 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 NOX 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 NOV emissions for lower-grade phosphate
                                                                    A                    *
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 NO  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 NO  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 NOo/Mg acid (^210 ppm) standards set for new nitric acid plants.
       The process flow diagram fpr 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 — a 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"
    pH
  RECORD.
 CONTROL.

  '
TREATED
TAIL GAS
  OUT
-^	
        LEVEL
                    ACIDIC
                   SECTION
                    »>  A  /N
   STEAM
 CONOENSATE
                 AMMONIACAL
                  SECTION
                 COLLECTION
                             PRODUCT
                            AMMONIUM
                              NITRATE
                             SOLUTION
                                    LEVEL
                                 " CONTROL
       pH
     RECORD.    ;
                               AMMONIA-
                                   LEVEL
    CONTROL.

   —tS	^
HYDRAULIC
 CONTROL
  VALVE

   pH
 RECORD.
 CON-TROL.
                          4XH"^-i^
   Figure 6-9.  Process flow diagram for trie Goodpastiire process (Reference 6-14),
                                      6-23

-------
 stripped from the solution in the amnoniacal  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
 Dimraitt 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.
                                                                «    •..   I
        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, modem 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  +  3NO£   *  2NaN03  +  NO  +  HgO                  (6-8)
                           2NaOH  +  NO  +  N02 '  *  2NaN02  +  HgO                         (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.
                                                                v
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 NO .  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
the tail gas to form N2, H-0, 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 1n 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 1s about 150K (270F).  For
*
 Relative to hydrogen this is still true; however, the economics of using natural  gas have greatly
 changed in the last three years.

                                                6-25

-------
                             TAIL GAS PREHEATER
                                                                    SET POINT
                                                                  CH4 OFF WHEN
                                                                 NHa'TO CONVERTER
                                                                      IS OFF
                                                            ON-OFF
                                                        SET POINT-OPEN
                                                     A   HIGH 02
                                         CH4/02
                                      CONTROLLER
                                           I	
                                                                           02
                                                                       ANAILYZER
                                                                       CONTROLLER
                            Qfl.QFF
                           HIGH TEMP.
                            SET POIWT
                                                                          CH4
                                                                       ANALYZER
                                                                       CONTROLLER
                                    TEMP.
                                  RECORDER/
                                 CONTROLLER
MOLECULAR SIEVE ~
  DESULFURIZER
POKIER RECOVERY
    TURBINE
                                    COMPRESSOR
                            PROCESS
                             AIR
          Figure 6-10. Nonselective catalytic reduction system (Reference 6-16).
                                        6-26

-------
decolonization, 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 (9QOF) ,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 NO  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 NO  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 $o
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
1n NO  reduction compared to its success in decolorization may reflect that reaction kinetics make
1t much more difficult to reduce NO than NOg.
       Fuel requirements for nonselective abatement with methane are typically 10 to 20 percent
over stolchiometric.  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 NO  to form N~.  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
(41OF and 518F) because above 543K, ammonia may oxidize to form NO ; 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
supplejnent 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.  NO  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 NOg 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 NO--  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 NOg.  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 NOX 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 NO  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
I *-
1
i
I
! i
j i-.
HOT GAS
CONTAINING
DESORBEDN02
1
S
^,

/
^,
NO OXIDATION TO N02
AND N02[,~H20
ADSORPTION

REGENERATION
X.



CLEAN DRY /~~~^
TAIL GAS f ^\
-*-. HEAT
EXCHANGER
j (EXISTING)
1

POWER
RECOVERY
(EXISTING)
FGAS
Figure 6-11.  Molecular sieve system (Reference 6-22).
                       6-30

-------
                         TABLE 6-2.  PERFORMANCE OF HERCULES PURASIV N UNIT  DURING THREE.DAY
                                    RUN (REFERENCE 6-24)
o>

NOV in Effluent
A
Average, ppm
Range, ppm
NOV in Feed
A
Average, ppm
Range, ppm
Average Gas Flow
Tail gas flow, Nm3/sa
Recycle gas flow,
NraVs
Total gas flow, Nm3/s
May 16

2
0-6

2,600
2,000-3,000

2.29
0.45
2.74
May 27

2
0-7

2,400
2,300-2,500

2.19
0.45
2.64
May 28

5
0-25

2,450
2,300-2,500

2.17
0.45
2,62
                          alNmVs = 2118.9 scfm

-------
Ol
I
UI
ro
                         TABLE 6-3.  PERFORMANCE OF U.S. ARHY-HOLSTON  PURASIV N  UNIT DURING
                                     THREE DAY RUN (REFERENCE 6-24)

                                     (PLANT LARGELY DISMANTLED}

NOX 1n Effluent
Average, ppm
Range, ppm
NOV in Feed
A
Average, ppm
Range, ppm
Average Gas Flow .
Tail gas flow, Nm3/sa
Recycle gas flow,
Nms/s
Total gas flow, Nm3/s
August 17

<1
0-2

4,100
3,000-5,700

2.08
0.52
2.60
August 18

6
0-30

3,700
3,500-4,400

2.08
0.52
2.60
August 19

7
0-30

3,900
2,500-4,700

2.08
0.52
2.60
                               Nm3/s  =2118.9 scfm

-------
     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 continuous
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, 1975 costs).  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 (1976 costs).  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 $1.74 Mg ($l.i>8/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 a 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 costs 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

-------
                                         TABLE 6-4.   CAPITAL AND OPERATING COSTS FOR DIFFERENT MOX ABATEMENT
                                          SYSTEMS IS A 270 Mg/d NITRIC ACID PLANT  (References 6-6 and 6-26 P
o>
to

Capital Investment, ($)
Royalty
Operating Labor, (hr/yr)
($/yr)
Maintenance Labor,
(*/yr)
Labor Overhead (Incl. fringe
benefits & supervision, $/yr)
Catalyst or Molecular Sieve
Cooling Water, (l/m1n)
tt/yr)
Steam, (kg/hr)
($/yr credit)
Electricity, (kWh)
($/yr)
. Boiler Feed Mater, (1/mln)
($/yr)
Fuel, m
($/yr)
Nitric Add, (Hg/day)
($/yr)
Urea, (Mg/day)
($/yr)
Ammonium Nitrate, (Hg/day)
($/yr)
Depreciation (11-yr. life)
Return on Investment (@ 202)
Taxes & Insurance (@ 2%)
Total Annual Cost, ($/yr)
Annual Cost, ($/Mg)
Catalyst
Reduction
1,384,000
—
360
2,200
315
2,200
4,400

77,800
—
--
(7,182)
(387,590)
128
20,890
132
12,850
8.3
465,120
—
—
--



125,900
276,800
27,700
628,270
6.79
Molecular
Sieve
1,200,000
—
360
2,200
315
2,200
4,400

45,600
1,892
7,330
113
6,120
322
52,550

—
0.6
32,640
(6.0)
(112,200)
—



109,090
240,000
24,000
413,930
4.48
Grande
Paroisse
1,000,000
Included
360
2,200
315
2,200
4,400

—
1,135
4,420
__
__
90
14,690

—
__
__
(5.4)
(102,000)
—



90,910
200,000
20,000
236,780
2.56
CDL/
Vitok
575,000
None
360
2,200
315
2,200 "
4,400

—
3,861
14,980
324
17,500
265
43,250

—
__
__
(5.4)
(102,000)
—



52,300
115,000
11,500
161,330
1.74
Hasar
663,000
Fee
360
2,200
315
2,200
• 5,975

— .- -
—
—
594
32,070
20
3,260

—
__
__
(4.8)
. (89,760)
1.24C
74,528
(1.13)
(24,500)
60,300
132,600
13,260
195,708
2.12
Goodpasture
425,000
51 ,000
360
2,200
315
2,200
4,400
— *
114
440
--
—
45
7,340
—
—
--
--
—
~
—
--
--
01.8)
(422,000)
38,640
85,000
8,500
(42,290)
(0.46)
                         aThis table is given in Appendix A 1n English units.
                           Investment estimates exclude Interest during construction, owners expenses, aid land costs.
                         clnclude credit for 0.0017 Hg of urea/Hg of nitric acid produced present in the spent
                           solution (D.SITPD).
                          Parenthesis indicate credit taken.

-------
               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 )a

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
T.80-
-
-
13.07
Goodpasture
' -
1.38
-
U38
 This table is given in Appendix A in English units.
                   TABLE 6-6.  BASIS  FOR TABLES 6-4 AND 6-5  (Reference 6-6)a

                                (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
6 $6.1/hr
9 $7.0/hr
@ 100% of labor (including fringe
.  benefits and supervision)
@ $0.008/1000 1
9 $0.20/1000 1
@ $1.90/GJ
@$1.90/GJ
@ 11 yr straight line
@ 2035 of capital  cost
@  2% of capital  cost
@ $ 99/Mg
@ $176/Mg
9 $lJO/Mg

@ $0.02/kWh
@ $173/Mg
aThis 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, 1976  costs).
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.

5.2    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
Adipic 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
m
37
1193
7431
TO3 tons
5324
573
82
40
409
286
122
41
1315
8192
                                 6-37

-------
About 735 kg (1600 Ib) of nitric acid (TOO 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 HNO,.  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
1s 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 NHg 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
KOg. which has very strong oxidizing power.
       Adipic acid (COOH-(CH2)4*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
add were manufactured 1n 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  +  NOX  +  H-0          (6-11)

                        cyclohexanol  +• nitric acid  •+• adipic acid  +•  NO,  +  H-0          (6-12)
        The main nitrogen  compounds formed in the above reactions are NO,  NO-, and  N^O.   The dissolved
 oxides  are stripped from  the adipic acid/nitric acid solution with air and steam.  The NO and N02
 are recovered by absorption in nitric acid.  The off-gas from the NO  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 NOX 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 NOX  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 NOX from organic oxi-
dation processing is the replacement of nitric acid as an oxidant by catalytic processes using air
oxygen.  Tht 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 Kg/d (300 tons/day) adiplc acid plant would be about $2.48 x 10  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 N0«) to produce nitro
compounds or nitrates.  The following equations illustrate the two most common types of reaction:
                           RH + HON02 - RN02 + HgO                                (6-13)
                           ROH + HON02 •* RON02 + HgO                              (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.
      i-
       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 H2SO., 32 to 39 percent HNOj, and 8 percent water are used in batch
operations, which may process 3.785 m* (1000 gallons) to 5.678 m* (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

-------
                                                NITRATING HOUSE
wATrn t ^ nnnwtVFR «_-»i
(ED
;iu 1 	 ,
r " ~ /
^ MIXED-ACID 	 ^ MIXED-ACID . 	 NiTRATOR 	 » SEPAI
^ STORAGE, SCALE TANK * NITRATOR » SEPAf
> i
1
LYCERIN 	 -L GlYCERIN 1 \ GLYCERIN
LYCERIN j GI.Y.fcKlN *SCALETABrK J
GLYCERIN ' y^
HEATER X
HOUSE y^
SPENT ACID WATER
1 L
SPEKT-ACID i L
SO
WA'
UR.
fER
WE! NG
1ATOR -NG-*.
WATER 	
HOT
WATER

SODA
WATER

PREWASH
•*•
S i-
N


t ' 	 >
STORAGE SODA ASH V
SODA
WATER

N
JL_
T
NG
5
1 <-^-

CATCH
TANK
1
WASTE
WATER
J NEUTRALIZING
, HOUSE

EUTRALIZER *J
/WATER
_! '_ „ 	 ^ _^^_^ . " / X
	 , j WASIE
|
NITRIC ACID TOWER
RECOVERY HOUSE |
J_l-1ZJ__^.l_iJ N
' 1 1
BUGGY

NG NG
CATCH
TANK
	 	 	 	 i r_

- — — .
CATCH
TROQLYCERIN TANK
TO POWDER
1
1
RECOVERED   WEAK
    NITRIC   SULFURIC
      ACID   ACID

                                                                   WASTE
                                                                   WATER
Figure 6-12. Batch process for the manufacture of nitroglycerin (NG) (Reference 6-39).
                                    642

-------
     GLYCERIN
       FEED
AGITATOR
                         NITROGLYCERIN
                           M
                                                               WASH
                                                              WATER
                                                            TO BAFFLED
                                                          SETTLING TANKS
MIXED
 Adl
                                                      WATER
                                                    SEPARATOR
                     J
A
                                                       (
                                                   NITROGLYCERIN
                                                     SODA WATER—*»
                                                                          eg.
                                                                          <0
                                                                          3"
                                             AIR
                               V
                                AIR
          TO DROWNING
              TANK
                                WATER
                                   AIR
                       L
                           STORAGE
                             TANK
                                              .NITROGLYCERIN.
                                WATER
(PASSED THROUGH ADDITIONAL    NITROGLYCERIN
WASH COLUMNS IF NECESSARY)
       Figure 6-13. Schmid-Meissner continuous-nitration plant (Reference.6-39).
                                          6-43

-------
     AGITATOR
GLYCERIN
  FEED
   i
          ACID
                                     AGITATOR
                                   WATER
      i
      NITROGLYCERIN
    NITRATOR
        f
   TO-
DROWNING
                LINE OF
              SEPARATION
   SPENT ACID
      TO
LEVELING DEVICE
                           ACID     X
                         SEPARATOR 1
SEPARATOR
 LEVELING
  DEVICE
                           NITRO-  )
                          GLYCERIN'
                            ACID
                                    MECHANICAL
                                      WASHER
                        TO DROWNING
                           TANK
     SPENT ACID
 TO AFTER-SEPARATOR
    AND STORAGE
                                       ~L
                                                                     WASTE
                                                                   -WATER1
                                                         WATER
                                                        SEPARATOR
                                                    AGITATOR
                                              NITROGLYCERIN
                                                      LEVELING
                                                       DEVICE
                                          SODA ASH
                                          SOLUTION
                                              t_
                                                  MECHANICAL
                                                 NEUTRALIZER
                                                i
                                                             NITROGLYCERIN
                                                               EMULSION
                                                 TO ADDITIONAL
                                                 WASHERS AND
                                                 SEPARATORS
                                                               T
                                                     WASH WATER TO SEPARATOR
              Figure 6-14. Biazzi continuous-nitration plant (Reference 6-39).
                                   6-44

-------
                                          GASES TO
                                      ABSORPTION TOWER
                                                        S-BEND
                                                      CONDENSER
SPENT-ACID
FEEDTANK
           CHEMICAL-WARE
             BV.OCKCOCK
                                                             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
1n 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 HNO,, 52 to 56 percent H2S04, and 12 to 20 percent water is  used  in  equipment sized to
handle up to 11.4 m*  (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, 1s 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 HgSO^, and 5 to 3 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 add plants, and nitric acid storage tanks.
       Nitration reactions per se do not generate N0₯ emissions.  NOV is  formed in  side reactions
                                 ~          *          .               A  i
Involving the oxidation of organic materials.  Relatively  little oxidation  and NOV formation occur
                                                                                  X
when easily nitratable compounds, such as toluene, are processed.  Much more severe conditions  are
required 1n processing compounds that are difficult to nitrate, such as dinitrotoluene; more  oxida-
tion takes place and, thus, more NOX is formed.
       Limtied data are available on actual NO  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/Hg (0.18 Ib/ton 1 at a stngle location CReference 6-40).  At the
same location, emissions averaging 7 kg of N02 per Mg of acid were  reported  in manufacturing  specialty
products 1n 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-
uencs, 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 HN03 (14 Ib/ton),
(Reference 6-40) as upper limits for nitrations, the NOX emissions in 1970 would have the range indi-
cated in Table 6-8.  Even using the upper limit, NOX 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 NOg 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 ones 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 NOX 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 NOX incineration are regarded as the most plausible controls for
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.

-------
                                 TABLE 6-8.   ESTIMATED. NOv EMISSIONS FROM ORGANIC NITRATIONS
                                             IN 1970 (REFERENCE 6-42).
O)
Product
Nitrobenzene
Blnitrotoluene
1970 Production
Mg
1 (tons)
233,600
(257,000)
131,542
(145,000)
Estimated HNOa
consumption
Mg
(tons)
126,099
(139,000)
101,151
(111,500)
NOX Emissions
Mg
(tons)
Lower Upper
limit limit
883
(973)
708
(780)

-------
6.2.3.4   Costs
       Fume Incinerator investments  are quoted at $10,000 to $20,000 by one source (Reference 6-43,
1966 costs).  Another suggests that  investments of $75,000 to $150,000 are necessary for flame
abatement facilities 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 .
10* 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-nltramine (RDX), and trinltrophenylmethyl-
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 1in 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:
Ot  +  H,S04* °,N  | Oj  NO^
             CH   +  3HONO   +  HS0* °N    O    NO   +  3H 0  +' H SO            (6-15)
                                                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 ac^'d
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
add.  The product from the third nitration vessel is sent to the wash house at which point asym-
inetrical 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 •* C6H702(ON02}3 + 3 HgO + H2S04           (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

-------
     TOLUENE-
EXHAUST

H
ELECTROSTATIC
PRECIPITATOR


SULFURIC ACID !
CONCENTRATOR
HOT GAS
1
FURNACE
H2S04
TANKS
t
FUEL:
              H2S04
,GO%HN03 BI-AC10 60%HN03 TRI-ACID OLEUM
I f
MONO-HOUSE

1
r t i *
BI-HOUSE
1 | 	 f

TRI-HOUSE
1 41
.SPENT MONO-OIL FUMES BI-OIL TRl'oiL
j FUMES *,
DENITRATOR

r FUMES
OXIDATION
CHAMBER
4
1 ' '
OXIDATION
TOWERS AND!
SEPARATORS

PURIFI
TNT
f
BUBBLE CAP
TOWER
VE
1 '
BUBBLE CAP
TOWER
1
VENT
HN03
TANKS



WASH
HOUSE
ED RED WATER i«


93
H2£
97% HN03 6Q%
| HN03
VENT
1 ,
NITRIC ACID
CONCENTRATOR
i 1 Na2C
LLITE EXHAUST 1 1
LUTION 1 ,BEsinUAL 1
f^ 1 i K2SG4 i
EVAPORATORS
WASTE LIQUOR
MT HOT GAS 1
1 1 t
FURNACE

ROTARY
KILNS
%
.04

03
SELLITE
PLANT
H2fl
STE/
OLEUM EXHAUST
GAS
,
IM
FUEL Na2S04 EXHAUST
                Figure 6-16. Trinitrotoluene (batch process) manufacturing diagram (Reference 6-45).

-------
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 NO  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 NOV emissions.
                                                                  A
       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
1n Klngsport, 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 b
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)
5^(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
cUse 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 te sig-
 nificant because these values are based on only one test result.
6Reference 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).
       NOX emissions are dependent on the quantity of carbonaceous material in the rock, since NO
is formed as nitric acid oxidizes the carbonaceous matter.  The use of calcined rock avoids the
production of NOX.
       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 NOX emissions to only a minor degree.   Although no measure-
ments of NOX emissions are available, brown plumes are said to occur.
       One company has found that the addition of urea to the acidulation mixture reduces NOV emis-
                                                                                            X
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 1n 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
                                                                           I
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 NOX 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 N02 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 NOV 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 NO, 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.
                                                                     . ,, in  , |,    „ i     ,  "•

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, 752(5} -.969, January 3\,
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.

6-25   May! and, 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.
                                                6-5S

-------
6-34   Durocher, D.F.  et al_.,  "Screening Study to Determine Needs  for Standards of Performance for
       New Sources of Dimethyl  Terephthalate and Terephthalic  Acid Manufacturing" Draft Final Report,
       6CA-TR-76-17-6.  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 Cyclohexanft 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, 1%4.

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 Tl-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 al..,  "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 three large cost tables
                                 \

presented in Section 6.  The tables are arranged sequentially in the order in which they appear in


the section and have the same table number as their counterparts except for the prefix "A".
                                                A-l

-------
          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 Investaent,* ($)
Royalty
Operating Labor, (hr/yr)
(5/yr)
Maintenance Labor,
($/yr)
lawsr Overhead (1nc1. fringe
benefits & supervision, S/yr)
Catalyst or Molecular Slave
Cooling Water, (gps»)
(S/yr)
Stew, (1b/yr)
(5/yr credit)
Electricity, (kW)
($/yr)
Boiler Feed Water, (gpm)
(5/yr)
Fuel, (10* 8tu/hr)
(5/yr)
Hltrlc Acid, (tpd)
(S/yr)
Urea, tpd
(S/yr)
AnoonltM Nitrate, (tpd)
($/yr)
Depreciation (11-yr Ufa)
Return on Investment (8 20!1)
Taxes & Insurance, (9 22)
Total Annual Cost, ($/yr)
Annual Cost, ($/ton)
Catalyst
Reduction
V.384,000
—
360
2.200
315
2.200
4,400

77,800
_
MM
(15,8335°
(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
4f|;^§
4.06
Grands
Parolsse
1 ,000,000
Included
360
2,200
315
2.200
4,400

_
. 300
4,420

_
90
14,690
--
_
„
M
(6.0)
(102.000)




90,910
200,000
20.000
2ld,?80
2.32
COL/
V1tok
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
lUSOO
T81T33?
1.53
Masar
663,000
fee
360
2,200
540
3,775
5,975

—
._
i —^
1,310
32,070
20
3,260
.-
_
~
..
(5.28)
(89,760)
1.370
74,528
1.25
(42,500)
60,300
132.600
13.260
193,/od
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
3,500
(42,290)
(Q-.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 1n tne spent
 solution (O.SCTPD).

Parenthesis Indicate credit taken.
                                                A-2

-------
    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
011

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 9 300 tpd and 102.000 tons/yr)
        (March 1975 Dollars, 0JR 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.   Antnbnium Nitrate
   14.   1 kWh - 10,500 Btu
   15.   Electricity
   16.   Amrania
@ $6.1/hr
9 $7.0/hr
@ 100% of labor (including fringe  benefits
  and supervision)
9 $0.03  1,000 gaL
9 $0.75  1,000 gal
9 $2.00/10* Btu
9 $2.00/10* Btu
9 11 yr straight Line
9 Z0% of capital cost
9  2% of capital cost
9 $90/ton
9 $160/ton
9 $100/ton

9 $0.02/kWh
9- $1577ton
*o.s.  GOVEia«m
              OPF1CK: 1983-639-ooi/30o2
                                      A-3

-------

-------
                                             APPENDIX B
                                        PREFIXES FOR SI UNITS
       The names of multiples and submultlples of SI units may be formed by application of these
prefixes:
                Factor by Which
Unit 1s Multiplied
1019
1015
1012
10»
10s
103
102
10
10-1
10-2
10-3
io-s
10-9
1Q-12
10-is
io-18
Prefix
exa
peta
tera
glga
mega
kilo
hecto
deka
deci
centl
mini
micro
nano
p1co
femto
atto
Symbol
E
P
T
G
M
k
h
da
d
c
m
U
n
P
f
a
                                                8-1

-------

-------
                                             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-stoiehiometric 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 largft 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 Roller exhaust
gases are recirculated to the burners to inhibit NO formation.
Flue Gas Treatment -A proce'ss which treats tail gases  chemically to remove NO  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  MM (290GJ/h).
                                                0-1

-------
 Beat Release  Rate — The  rate  of combustion  per unit  volune of firebox, typically  In terms of MH/mJ.
 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  In- the higher heating value even if  it  1s not practically available, 1:e.<- heat of con-*
 dtnslng water vapor.
 la*  Excess Air—  A  combustion modification  1n  which  NOX formation Is Inhibited by reducing the excess
 air  to less than  normal  ratios.
 Lower or  Met  Heating Value  (LHV) — The heat that Is  practically available from a  fuel to generate
 steas or  otherwise  raise the  temperature of the  media  receiving energy.  The net  heating value assumes
 complete  combustion.   It differs from the higher heating value in that heat of vaporizing* water of
 ccB&ustion is considered a  recoverable loss.
 Off-Stoichloflietric  Combustion (OSC) — A combustion modification technique in which burner stoichi-
 ctnetry is altered to inhibit  HOX formation.  Types of OSC Include biased firing, burners out of
 service,  and  two-stage combustion.
 Packaged  toilers  -  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.
 Pplyeycljc Organic  Hatter  (POM) - Organic compounds which exists in condensed phase at ambient tem-
 ptraturt  and  are  emitted as either "carbon  on  particular" or condensed onto emitted particulate.
 Polynuclear Aromatic Hydrocarbons (PKA) - Same as POM.
*        *
 Stoichloroetric 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-stoichiometric 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 MNOX port operation.
 Vatsrtube Boiler  — A steam generator with hea^ 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-450/3-R3-nn?
                             2.
                                                           3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
  Control Techniques for Nitrogen  Oxides Emissions from
  Stationary Sources - Revised  Second  Edition
                                                           5. REPORT DATE
                January  1983
             6. PERFORMING ORGANIZATION CODE
 . AUTHOR(S)
                                                           8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
                                                           10. PROGRAM ELEMENT NO.
                                                           11. CONTRACT/GRANT NO.
                                                                68-02-3513
                                                                Work Assignment 24
12. SPONSORING AGENCY NAME AND ADDRESS
   United  States Environmental Protection Agency
   Office  of  Air Quality Planning and  Standards
   Research Triangle Park, North Carolina 27711
             13. TYPE OF REPORT AND PERIOD COVERED
                  Final
             14. SPONSORING AGENCY CODE

                  EPA 200/04
15. SUPPLEMENTARY NOTES
   This  document  is issued in accordance  with the requirements of  Section 108 of the
   Clean Air Act, as amended, 1977.
16. ABSTRACT
       As  required by Section 108 of  the  Clean Air Act, this revised  second edition
  compiles the best available information on NO  emissions; achievable  control levels
  and  alternative methods of prevention and control of NO  emissions; alternative
                                                           x
  fuels, processes, and operating methods which reduce NO  emissions;  cost of NO
                                                           x
  control methods, installation, and operation;  and the energy requirements  and x
  environmental impacts of the NO  emission control technology.
                                  x
       Each stationary source of NO  emissions is discussed along with  the various
  control techniques and process modifications available to reduce NO   emissions.
  Various combinations of equipment process conditions and fuel types are identified
  and evaluated for NO  emission control.
                       x
       This  revised second edition of Control Techniques for Nitrogen Oxides  Emissions
  from Stationary Sources updates the second edition (EPA-45,0/1-78-001) published in
  January  1978.   The changes are limited  to  revisions of information on emissions and
  emission factors (Chapter 2), combustion modifications (Section 3.1), combustion
  flue gas treatment (Section 3.2), utility  boilers (Section 4.1), industrial boilers
  (Section 4.2),  space heating (Section 5.1), and industrial process heating  (Section 5.
                                             3)
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS  C. COSATI Field/Group
  Nitrogen Oxides Emissions
  Control  Techniques
  Fossil Fuel  Combustion
  Nitric Acid  Manufacturing
  Costs
  Photochemical Oxidants
                                 13b
18. DISTRIBUTION STATEMENT

  Release Unlimited
19. SECURITY CLASS (ThisReport)
     Unclassified
21. NO. OF PAGES
   428
                                              20. SECURITY CLASS (Thispage)
                                                   Unclassified
                                                                         22. PRICE
 EPA Form 2220—1 (Rev. 4-77)    PREVIOUS EDITION is OBSOLETE

-------

-------