United Slates
Environmental Protection
Agency
-------
TECHNICAL REPORT DATA
{Please read Itotpjctions on the reverse before comple
1. REPORT NO,
EPA-60Q/R-92-093c
3!
Ill III III III III
PB93-212868
4. TITLE AND SUBTITLE
Proceedings: 1991 Joint Symposium on Stationary
Combustion NOx Control, Washington, D. C. , March
25~28, 1991, Volume 3. Sessions 6-8
5. REPORT DATE
July 1892
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Carolee DeWitt,
8. PERFORMING ORGANIZATION REPORT NO,
Compiler
9, performing organization name and address
William Nesbit and Associates
1221 Farmers Lane
Santa Rosa, California 95405
10. PROGRAM ELEMENT NO.
n, contract/grant no.
NA (EPR1 Funded)
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
13, TYPE OF REPORT AND PERIOD COVERED
Proceedings; 3/89 - 3/91
14. SPONSORING AGENCY CODE
EPA/600/13
is.supplementary notes _&EERL project officer is Robert E. Hall, Mail Drop 65, 919/541-
2477. Volume 1 includes Sessions 1-3, and Session 2 includes Sessions 4 and 5.
16. abstract proceedings document the 1991 Joint Symposium on Stationary Combus-
tion NOx Control, held in Washington, DC, March 25~28, 1991. Jointly sponsored by
EPRI and EPA, the symposium was the sixth in a biennial series devoted to the
international exchange of information on recent technological and regulatory develop-
ments for stationary combustion nitrogen oxides (NOx) control. Topics covered inclu-
ded the significant increase in active full-scale retrofit demonstrations of low-NOx
combustion systems in the U.S. and abroad over the past 2 years; full-scale oper-
ating experience in Europe with selective catalytic reduction (SCR); pilot- and bench-
scale SCR investigations in the U.S. ; increased attention to selective noncatalytic
reduction (SNCR) in the U.S.; and NOx controls for oil- and gas-fired boilers. The
proceedings are published in three volumes.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Pollution
Nitrogen Oxides
Combustion
Catalysis
Fossil Fuels
Pollution Control
Stationary Sources
Catalytic Reduction
Noncatalytic Reduction
13 B
07B
21B
07D
21D
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport/
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS {Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73J
A-37
4
-------
EPA-600/R-92-093 c
July'19S2 ~
PROCEEDINGS:
1991 JOINT SYMPOSIUM ON STATIONARY COMBUSTION NOx CONTROL
Washington, D.C., March 25-28, 1991
Volume 3. Sessions 6-8
Compiled by
Carolee DeWitt
William Nesbit and Associates
1221 Farmers Lane
Santa Rosa, CA 95405
EPA Project Officer:
Robert E. Hall
U.S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
EPRI Project Manager:
Angelos Kokkinos
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 93404
Prepared for:
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 93404
-------
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
fiTIS is autlwiied lo reproduce awl Mil Wj
leport. Permission for lurthw reprodticllw
must be obtained Irani tee copVTight cwnw.
Copyright (e) 1991, EPR1 GS-7447, Proceedings:
1991 Joint Symposium on Stationary Combustion
NOx Control, Volumes 1, 2, and 3. Since this
work was, in part, funded by the U, S, Government,
the Government is vested with a royalty-free, non-
exclusive, and irrevocable license to publish, trans-
late, reproduce, and deliver that information and
to authorize others to do so.
-------
The 1991 Joint Symposium on Stationary Combustion NOx Control was held in Washington, D.C.,
March 25-28,1991, Jointly sponsored by EPRI and EPA, the symposium was the sixth in a biennial
series devoted to the international exchange of information on recent technological and regulatory
developments for stationary combustion NO^ control. Topics covered included the significant
increase in active full-scale retrofit demonstrations of low-NOx combustion systems in the United
States and abroad over the past two years; full-scale operating experience in Europe with selective
catalytic reduction (SCR); pilot- and bench-scale SCR investigations in the United States; increased
attention on selective noncatalytic reduction in the United States; and N0X controls for oil- and gas-
fired boilers.. The symposium proceedings are published in three volumes.
-------
-------
PREFACE
The 1991 Joint Symposium on Stationary Combustion NOx Control was held March 25-28, 1991, in
Washington, D.C. Jointly sponsored by EPRI and EPA, the symposium was the sixth in a biennial
series devoted to the international exchange of information regarding recent technological and
regulatory developments pertaining to stationary combustion NOx control. Topics discussed
included the significant increase in active full-scale retrofit demonstrations of low-NOx combustion
systems in the United States and abroad over the past two years; full-scale operating experience in
Europe with selective catalytic reduction (SCR); pilot-and bench-scale SCR investigations in the
United States; increased attention on selective noncatalytic reduction in the United States; and NOx
controls for oil- and gas-fired boilers.
The four-day meeting was attended by approximately 500 individuals from 14 nations. Sixty-six
papers were presented by EPRI and EPA staff members, domestic and foreign utility companies,
federal and state government agencies, research and development organizations, equipment
vendors from the United States and abroad, and university representatives.
Angelos Kokkinos, project manager in EPRl's Generation & Storage Division, and Robert Hall,
branch chief, Air & Energy Engineering Research Laboratory, EPA, cochaired the symposium. Each
made brief introductory remarks. Michael R. Deland, Chairman of the President's Council on
Environmental Quality, was the keynote speaker. Written manuscripts were not prepared for the
introductory remarks or keynote address and are therefore not published herein.
The Proceedings of the 1991 Joint Symposium have been compiled in three volumes. Volume 1
contains papers from the following sessions;
¦ Session 1: Background
¦ Session 2: Large Scale Coal Combustion I
¦ Session 3: Large Scale Coal Combustion II
iii
-------
Papers from the following sessions are contained in Volume 2:
Combustion NOx Developments I
Large Scale SCR Applications
Post Combustion Developments 1
Industrial/Combustion Turbines on NOx Control
Papers from the following sessions are contained in Volume 3:
¦ Session 6A: ^ Post Combustion Developments IK
¦ Session"6B: Combustion NQ^Developments II
¦ Session 7A: New Developments I [
¦ Session.7B: New Developments \\
¦ Session-8. Oil/Gas Combustion Applications-—
An appendix listing the symposium attendees is included at the end of Volume 3,
¦ Session 4A:
¦ . Session 4B:
¦ Session 5A:
¦ Session SB:
iv
-------
CONTENTS
Paper VOLUME 1 Page
SESSION 1: BACKGROUND
Chair: 1, Torrens, EPRI
"NOx Emissions Reduction in the former German Democratic Republic," B. Kassebohm
and S. Streng 1 -1
" Top-Down' BACT Analysis and Recent Permit Determinations," J. Cochran and M. Fagan 1-t5
"Analysis of Retrofit Costs and Performance of NOx Controls at 200 U.S. Coal-Fired
Power Plants," T. Emmel and M. Maibodi 1-27
"Nitrogen Oxides Emission Reduction Project," L. Johnson 1-47
"The Global Atmospheric Budget of Nitrous Oxide," J, Levine 1 -65
SESSION 2: LARGE SCALE COAL COMBUSTION I
Chair: B. Martin, EPA and G. Often, EPRI
"Development and Evolution of the ABB Combustion Engineering Low NOx Concentric
Firing System," J. Grusha and M. McCartney 2-1
"Performance of a Large Cell-Burner Utility Boiler Retrofitted with Foster Wheeler
Low-NOx Burners," T. Lu, R. Lungren, and A, Kokkinos 2-19
"Design and Application Results of a New European Low-NO, Burner," J, Pedersen and
M. Berg 2-37
"Application of Gas Reburning-Sorbent Injection Technology for Control of
NOx and S02 Emissions," W. Bartok, B. Folsom, T. Sommer, J. Opatrny, E. Mecchia,
R. Keen, T. May, and M. Krueger 2-55
"Retrofitting of the Italian Electricity Board's Thermal Power Boilers," R. Tarli, A. Benanti,
G. De Michele, A. Piantanida, and A. Zennaro 2-75
"Retrofit Experience Using LNCFS on 35GMW and 165MW Coal Fired Tangential Boilers,"
T. Hunt, R. Hawley, R. Booth, and B. Breen 2-89
"Update 91 on Design and Application of Low NO„ Combustion Technologies for Coal
Fired Utility Boilers," T. Uemura, S. Morita, T. Jimbo, K. Hodozuka, and H. Kuroda 2-109
v
-------
Paper Page
SESSION 3; LARGE SCALE COAL COMBUSTION II
Chair: D. Eskirwi, EPRI and R, Hall, EPA
"Demonstration of Low NOx Combustion Control Technologies on a 500 MWe Coal-Fired
Utility Boiler," S. Wilson, J. Sorge, L. Smith, and L. Larsen 3-1
"Reburn Technology for NOx Control on a Cyclone-Fired Boiler," R. Borio, R. Lewis, and
M, Keough 3-23
"Full Scale Retrofit of a Low NOx Axial Swirl Burner to a 660 MW Utility Boiler, and the
Effect of Coal Quality on Low NOx Burner Performance," J. King and J. Macphail 3-51
"Update on Coal Reburning Technology for Reducing NOx in Cyclone Boilers," A. Yagiela,
G. Maringo, R. Newell, and H. Farzan 3-74
"Demonstration of Low NOx Combustion Techniques at the Coal/Gas-Fired Maas Power
Station Unit 5," J, van der Kooij, H. Kwee, A. Spaans, J. Puts, and J. Witkamp 3-99
'Three-Stage Combustion (Reburning) on a Full Scale Operating Boiler in the U.S.S.R.,"
R, LaFlesh, R. Lewis, D. Anderson, R. Hall, and V, Kotler 3-123
VOLUME 2
SESSION 4A: COMBUSTION NOx DEVELOPMENTS I
Chair: W, Linak and D. Drehmei, EPA
"An Advanced Low-NOx Combustion System for Gas and Oil Firing," R, Lisauskas
and C, Penterson 4A-1
"NO, Reduction and Control Using an Expert System Advisor," G. Trivett 4A-13
"An R&D Evaluation of Low-NOx Oil/Gas Burners for Salem Harbor and Brayton Point
Units," R. Afonso, N. Molino, and J. Marshall 4A-31
"Development of an Ultra-Low NOx Pulverizer Coal Burner," J. Vatsky and T. Sweeney 4A-53
"Reduction of Nitrogen Oxides Emissions by Combustion Process Modification in
Natural Gas and Fuel Oil Flames: Fundamentals of Low NOx Burner Design," M, Toqan,
L. Berg, J, Be6r, A. Marotta, A. Beretta, and A. Testa 4A-79
"Development of Low NOx Gas Burners," S, Yang, J. Pohl, S. Bortz, R. Yang, and W. Chang 4A-105
SESSION 4B: LARGE SCALE SCR APPLICATIONS
Chair: i. Cichanowicz, EPRI
"Understanding the German and Japanese Coal-Fired SCR Experience," P. Lowe,
W. Ellison, and M. Perlsweig
"Operating Experience with Tail-End-and High-Dust DENOX-Technics at the Power Plant
of Heilbronn," H. Maier and P, Dahl
4B-1
4B-17
vi
-------
Paper Page
"S03 Generation-Jeopardizing Catalyst Operation?," R. Jaerschky, A. Merz, and J. Mylonas 4B-39
"SCR Operating Experience on Coal-Fired Boilers and Recent Progress," E. Behrens,
S, Ikeda, T. Yamashita, G. Mittelbach, and M. Yanai 4B-57
'Technical Feasibility and Cost of SCR for U.S. Utility Application," C. Robie, P. Ireland,
and J, Cichanowicz 4B-79
"Application of Composite NOx SCR Catalysts in Commercial Systems," B. Speronello,
J, Chen, M. Durilla, and R. Heck 4B-101
"SCR Catalyst Developments for the U.S. Market," T. Gouker and C. Brundrett 4B-117
"Poisoning Mechanisms in Existing SCR Catalytic Converters and Development of a New
Generation for Improvement of the Catalytic Properties," L. Balling, R. Sigling, H. Schmelz,
E. Hums, G. Spitznagel 4B-133
SESSION 5A: POST COMBUSTION DEVELOPMENTS I
Chair: C. Sedman, EPA
"Status of 1 MW SCR Pilot Plant Tests at Tennessee Valley Authority and New York State
Electric & Gas," H. Flora, J. Barkley, G. Janik, B. Marker, and J. Cichanowicz 5A-1
"Pilot Plant Investigation of the Technology of Selective Catalytic Reduction of Nitrogen
Oxides," S. Tseng and C. Sedman 5A-17
"Poisoning of SCR Catalysts," J. Chen, R. Yang, and J. Cichanowicz 5A-35
"Evaluation of SCR Air Heater for N0X Control on a Full-Scale Gas- and Oil-Fired Boiler,"
J. Reese, M. Mansour, H. Mueller-Odenwald, L. Johnson, L Radak, and D. Rundstrom 5A-51
"N20 Formation in Selective Non-Catalytic NOx Reduction Processes," L. Muzio,
T. Montgomery, G. Quartucy, J. Cole, and J. Kramlich 5A-71
Tailoring Ammonia-Based SNCR for Installation on Power Station Boilers," R. Irons,
H. Price, and R. Squires 5A-97
SESSION 5B: INDUSTRIAL/COMBUSTION TURBINES ON N0X CONTROL
Chair: S. Wilson, Southern Company Services
"Combustion Nox Controls for Combustion Turbines," H. Schreiber 5B-1
"Environmental and Economic Evaluation of Gas Turbine SCR N0X Control," P. May,
L Campbell, and K. Johnson 5B-17
"NOx Reduction at the Argus Plant Using the NOxOUT* Process," J. Comparato, R, Buchs,
D. Arnold, and L. Bailey 5B-37
vii
-------
Paper
Page
"Reburning Applied to Degeneration NOx Control," C. Castaldini, C. Moyer. R. Brown,
J. Nicholson 5B-55
"Selective Non-Catalytic Reduction (SNCR) Performance on Three California Waste-to-
Energy Facilities," B. McDonald, G. Fields, and M. McDannel 5B-71
"Use of Natural Gas for NOx Control in Municipal Waste Combustion," H. Abbasi,
R. Biljetina, F, Zone, R. Usauskas, R. Dunnette, K. Nakazato, P, Duggan, and D. Linz 5B-89
VOLUME 3
SESSION 6A: POST COMBUSTION DEVELOPMENTS II
Chair: D, Drehmel, EPA
"Performance of Urea NOx Reduction Systems on Utility Boilers," A. Abele, Y. Kwan,
M, Mansour, N, Kertamus, L, Radak, and J. Nylander 6A-1
"Widening the Urea Temperature Window," D. Teixeira, L Muzio, T. Montgomery,
G. Quartucy, and T. Martz 6A-21
"Catalytic Fabric Filtration for Simultaneous NOx and Particulate Control," G. Weber,
D, Laudal, P. Aubourg, and M. Kalinowskl 6A-43
SESSION 6B; COMBUSTION N0X DEVELOPMENTS II
Chair: R. Hall, EPA
"Heterogeneous Decomposition of Nitrous Oxide in the Operating Temperature Range of
Circulating Fluidized Bed Combustors," T, Khan, Y.Lee, and L, Young 6B-1
"NOx Control in a Slagging Combustor for a Direct Coal-Fired Utility Gas Turbine,"
P. Loftus, R. Diehl, R, Bannister, and P. Pillsbury 6B-13
"Low NOx Coal Burner Development and Application," J. Allen 6B-31
SESSION 7A: NEW DEVELOPMENTS I
Chair: G, Veerkamp, Pacific Gas & Electric
"Preliminary Test Results: High Energy Urea Injection DeNOx on a 215 MW Utility Boiler,"
D. Jones, S. Negrea, B. Dutton, L. Johnson, J. Sutherland, J. Tormey, and R. Smith 7A-1
"Evaluation of the ADA Continuous Ammonia Slip Monitor," M. Durham, R. Schlager,
M. Burkhardt, F. Sagan, and G, Anderson 7A-15
"Ontario Hydro's SONOX Process for Controlling Acid Gas Emissions," R. Mangal,
M. Mozes, P. Feldman, and K. Kumar 7A-35
"Pilot Plant Test for the NOXSO Flue Gas Treatment System," L. Neal, W. Ma, and R, Bolli 7A-61
vm
-------
Paper Page
"The Practical Application of Tunable Diode Laser Infrared Spectroscopy to the Monitoring
of Nitrous Oxide and Other Combustion Process Stream Gases," F. Briden, D. Natschke,
and R. Snoddy 7A-79
SESSION 7B: NEW DEVELOPMENTS II
Chair; C. Miller, EPA
"In-Furnace Low NOx Solutions for Wall Fired Boilers," R. LaFlesh, D. Hart, P, Jennings, and
M. Darroch 7B-1
"NOx Reduction on Natural Gas-Fired Boilers Using Fuel Injection Recirculation (FIR) -
laboratory Demonstration," K. Hopkins, D. Czerniak, L. Radak, C. Youssef, and J. Nylander 7B-13
"Advanced Reburning for N0X Control in Coal Fired Boilers," S. Chen, W. Seeker, and
R.Payne 7B-33
"Large Scale Trials and Development of Fuel Staging in a 160 MW Coal Fired Boiler,"
H. Spliethoff and R. Doieial 7B-43
"Computer Modeling of N20 Production by Combustion Systems," R. Lyon, J. Cole,
J. Kramlich, and Wm. Lanier 7B-63
SESSION 8: OIUGAS COMBUSTION APPLICATIONS
Chair: A, Kokkinos, EPRI
"Low NOx Levels Achieved by Improved Combustion Modification on Two 480 MW Gas-
Fired Boilers," M, McDannel, S. Haythornthwaite, M. Escarcega, and B. Gilman 8-1
*NOx Reduction and Operational Performance of Two Full-Scale Utility Gas/Oil Burner
Retrofit Installations," N. Bayard de Volo, L. Larsen, L. Radak, R, Aichner, and A. Kokkinos 8-21
"Comparative Assessment of NOx Reduction Techniques for Gas- and Oil-Fired Utility
Boilers," G. Blsonett and M. McElroy 8-43
"Analysis of Minimum Cost Control Approach to Achieve Varying Levels of NOx Emission
Reduction from the Consolidated Edison Co. of NY Power Generation Systems," D. Mormile,
J. Pirkey, N. Bayard de Volo, L. Larsen, B. Piper, and M, Hooper 8-63
"Reduced NOx, Particulate, and Opacity on the Kahe Unit 6 Low-NOx Burner System,"
S. Kerho, D, Giovanni, J, Yee, and D. Eskinazi 8-85
"Demonstration of Advanced Low-NOx Combustion Techniques at the Gas/Oil-Fired Flevo
Power Station Unit 1," J. Witkamp, J. van der Kooij, G. Koster, and J. Sijbring 8-107
APPENDIX A:
LIST OF ATTENDEES
A-1
-------
PERFORMANCE OF UREA NO, REDUCTION SYSTEMS
ON UTILITY BOILERS
Andris R. Abele*, Yul Kwan, and M.N. Mansour
Applied Utility Systems, Inc.
1140 East Chestnut Avenue
Santa Ana, California 92701
N.J. Kertamus and Les J. Radak
Southern California Edison Company
2244 Walnut Grove Avenue
Rosemead, California 91770
James H. Nylander
San Diego Gas and Electric Company
4600 Carlsbad Boulevard
Carlsbad, California 92008
ABSTRACT
Test results from the full-scale application of urea injection for NO, reduction on two utility
boilers demonstrate the sensitivity of urea NO, reduction performance to boiler design,
operating conditions, and urea process variables. The two utility boilers are both gas- and
oil-fired boilers, but of different size and design. The demonstration sites include a
Southern California Edison Company 320 MW tangentially-fired boiler and a San Diego Gas
and Electric Company (SDG&E) 110 MW front wall-fired boiler.
The performance of the urea NO, reduction process at the two sites was dominated by
variables affecting the temperature at the injection location and the mixing of urea with the
combustion products. Varying operating conditions, such as load and firing configuration,
changed the temperature distribution in the boilers as well as initial NO, levels. Such
changes affect the relative location of urea injectors within the urea reaction temperature
window and, thus, the level of NOx reduction achieved. Available injection process
variables, including injector design, solution flow and pressure, injector location and spray
orientation, were used to optimize the distribution of urea within the reaction window at
varying loads to achieve maximum NO, reduction.
Minimum NOt emissions were achieved at both sites by coupling urea injection with
modified combustion conditions. Urea NOt reduction performance at these modified
operating conditions was about 30 percent at NSR = 2.0 over the boilers' load ranges.
Resulting stack NO, emissions at both units were 20 to 45 ppm @ 3% 02 depending on
load, while ammonia slip was less than 20 ppm.
* Currently with the South Coast Air Quality Management District.
6A-1
-------
The urea NO, reduction process is a selective non-catalytic reduction (SNCR) process which
encompasses a sequence of steps. Aqueous urea solution is pumped to injection nozzles
which spray the chemical into a boiler or furnace chamber. The droplets of injected
solution evaporate and the urea thermally decomposes into reactive species. The urea
droplets and released reactive species mix with the NO,-laden combustion products. Urea
species then react with the combustion products at the proper temperatures to reduce nitric
oxide (NO) to elemental nitrogen (N2).
The NO-reducing reactions are temperature sensitive and occur within a narrow temperature
range. If the urea is released at too high a temperature, the chemical species can actually
be oxidized to NO,,. If the urea is released at low temperatures, the NO-reducing reaction
rates are limited and result in poor chemical utilization.
An additional complication in SNCR systems is that these temperature sensitive reactions
must occur not in a well controlled reactor, but in a load-following utility boiler. The
design of these systems must address the issues of temperature variations and mixing
limitations to the extent possible. Since a utility boiler presents a far from perfect reaction
chamber environment, efficient utilization of injected urea is not possible for all boiler
operating conditions. Since the process is imperfect, excess urea must be injected to
maximize the availability of NOx-reducing species within the narrow reaction window
provided within utility boilers. Unutilized ammonia (NH3) will be a result if the injection
temperature is too low. At high injection temperatures, excess NH3 is oxidized to NOlP
defeating the purpose of reducing NO, emissions. Thus, tradeoffs will exist between NOK
reduction and overall process performance.
To understand the effectiveness of the urea injection process, the term Normalized
Stoichiometric Ratio (NSR) was defined as the ratio between the actual amount of urea
injected and the theoretical amount required to react with all the NO present. For example,
a urea flowrate of NSR =1.0 provides the exact amount of urea to react with 100 percent
of the NO present. This stoichiometric ratio of NSR = 1.0 is equivalent to a urea to NO
mole ratio of 0.5, since one mole of urea (NH2CONH2) potentially has two moles of
nitrogen species (e.g., NH;) available to react with NO.
6A-2
-------
COMPARISON OF BOILER DESIGNS
The two boilers used for demonstrating the urea NO, reduction process are different in
design and size. Both boilers are located in Southern California. The primary fuel for each
unit is natural gas, but each unit is also equipped to burn low sulfur fuel oil. Cross-sections
of the two boilers are shown in Figure 1, and their design characteristics are compared in
Table 1.
Encina Unit 2 is a 110 MW Babcock and Wilcox Company boiler. The unit is fired from
the front wall with ten burners arranged in two elevations of five burners. The unit operates
with balanced draft maintained by forced draft and induced draft fans. Flue gas
recirculation (FGR) injected between water tubes on the back wall of the lower furnace is
a primary means of steam temperature control. Final superheat steam temperature is
controlled by spray attemperation. The final reheat steam temperature is controlled by
distribution dampers in the baekpass. A total of sixteen existing observation ports are
available for urea injection in two elevations of the upper furnace. One elevation is located
adjacent to the furnace exit and entrance to the convective pass, while the second elevation
is about 12 feet below, near the arch of the furnace.
Etiwanda Unit 3 is a 320 MW Combustion Engineering boiler. This is a tangentially-fired
boiler with twin furnaces separated by a division wall. Etiwanda Unit 3 operates with a
pressurized furnace. This unit is unique in its downward flow arrangement with the burner
assemblies located at the top of the boiler. The burner assemblies consist of three tiers of
gas and oil burners located in the corners of each furnace. Tilt of the burner assemblies is
a primary means of reheat steam temperature control. FGR is injected into the windbox for
NO, control and for steam temperature control at low loads. Spray attemperators maintain
final steam temperatures. Twelve existing observation ports arranged in two elevations near
the furnace exit were initially used for urea injection. Additional ports were installed based
on initial test results and modeling efforts to improve NO, control performance over a wider
load range.
Etiwanda Unit 3 differs from Encina Unit 2 in a number of ways which can affect urea NO,
reduction performance. These differences include:
• Boiler dimensions and geometry - Etiwanda Unit 3 is physically larger than
Encina Unit 2 with a larger furnace cross-section. In addition, Etiwanda
Unit 3 has a divided furnace which limits access to the furnace cross-section
by urea injectors to two walls rather than three walls as at Encina Unit 2;
• Firing configuration - Etiwanda Unit 3 is tangentially down-fired while
Encina Unit 2 is a conventional front wall-fired boiler. The firing
configuration, and the furnace geometry affect the furnace flow field and
thus can be expected to affect mixing of injected urea with the furnace
gases;
6A-3
-------
• Thermal environment - At full load, gas temperatures in the region of the
furnace exit are significantly higher at Etiwanda Unit 3 (2400°F) than at
Encina Unit 2 (2250°F). Since the urea NO, reduction reactions are
temperature sensitive, differences in injector configurations and resulting
performance can be expected;
• Combustion conditions - The combustion conditions at Etiwanda Unit 3
result in significantly lower initial NO, levels than found at Encina Unit 2,
At full load on gas fuel, for example, NO, emissions at Etiwanda Unit 3 are
as low as 90 ppm (@ 3% 02) compared to 225 ppm (@3% O,) at Encina
Unit 2 with all-burners-in-service (ABIS). This is the result of NO, controls
that have been in place since the 1970's, consisting of FGR and two-stage
combustion achieved with burners-out-of-service (BOOS),
SENSITIVITY OF UREA NO, REDUCTION PERFORMANCE
Results from urea injection trials of both Encina Unit 2 and Etiwanda Unit 3 are indicative
of the key factors affecting NOx reduction potential. While boiler operating conditions
directly affected NO, reduction achieved with urea injection, the injection conditions and
configurations could be adjusted to ultimately minimize stack NO, emissions over a range
of conditions on each unit.
Effect of Operating Conditions
Previous urea injection testing at Encina Unit 2 was conducted with the boiler operating
with ABIS05. Initially, urea injection was evaluated as a cost-effective NO, control
alternative to the combustion modification techniques typically used in the SDG&E system
to meet current NO, regulations. The combustion modification techniques reduce overall
boiler efficiency compared to the higher, NO,-producing ABIS operating mode. With urea
injection, however, NO, emissions could meet existing NO, regulations while operating with
the more efficient ABIS. Subsequent testing has been completed to evaluate urea injection
in conjunction with alternate, or modified, combustion conditions.
The firing configurations evaluated included ABIS, air biasing, BOOS, and fuel biasing.
These alternatives were evaluated to determine the overall NO, reductions possible by
coupling urea injection with modified combustion conditions. ABIS represents conventional
operation with balanced fuel and air for all the burners, resulting in high baseline NO,
emissions. Air biasing was achieved with ABIS by closing the registers to the lower burner
elevation and thus diverting air to the upper level. This results in staged combustion, with
the lower burners operating fuel-rich and the upper burners operating fuel-lean. The effect
of staged combustion achieved with air biasing not only reduced baseline NO, emissions,
but also affected the heat release distribution through the boiler by delaying the mixing of
fuel and air. BOOS operation was achieved by shutting the fuel off to three of the ten
6A-4
-------
burners. This redistributes the fuel to the remaining burners and results in those burners
operating fuel-rich. BOOS operation thus also results in staged combustion and reduced
NO, emissions. Since the fuel distribution is changed with BOOS, the heat release
distribution also changes.
In gas fuel biasing, some of the fuel is diverted from the upper elevation of burners to the
lower elevations. This increases the heat release into the lower furnace and achieves staged
combustion. Compared to air biasing and BOOS operation, which delay fuel and air mixing
by varying air distribution or by discrete changes in fuel distribution, fuel biasing provides
more uniform changes in fuel distribution such that slightly fuel-rich and slightly fuel-lean
zones are created. The result with fuel biasing is a more confined heat release zone due to
more balanced fuel and air mixing and, more importantly, the diversion of fuel to the lower
burner elevation.
The urea NO, reduction performance varied for the different combustion modes at Encina
Unit 2, as the data in Figure 2 illustrate. Corresponding NOx emissions are shown in Figure
3. The data presented in Figures 2 and 3 represent urea NOt reduction performance
resulting from the injection configuration optimized for ABIS operation. No attempt was
made in these trials to optimize performance for each operating condition. Thus, injection
nozzle characteristics and injection configuration were constant. The highest percentage
reductions were achieved with ABIS operation and the lowest with BOOS operation.
Differences in measured performance can be attributed directly to changes in boiler
conditions. The data set presented in the two figures indicates that differences in NO*
reduction performance can be attributed both to the different initial NOx levels produced by
the different combustion configurations and to the effect oh the temperature distribution
through the boiler.
Analogous variations in urea NOs reduction performance with changing operating conditions
were documented at Etiwanda Unit 3®. Figure 4 illustrates the effect of various combustion
conditions on NO, reduction while Figure 5 presents the corresponding NO, emissions
levels. Included in the data presented from Etiwanda Unit 3 are urea injection test results
with normal, as found fuel oil-fired conditions; normal, as found gas-fired conditions; and
modified gas-fired combustion conditions. The modified combustion conditions at Etiwanda
Unit 3 comprised adjustment of burner tilt to horizontal for all loads with increased FGR
flowrate. As in the case for the Encina Unit 2 data set, the urea injection configuration was
not optimized for each operating condition.
The highest NO, reductions achieved at Etiwanda Unit 3 were with fuel oil. Fuel oil-firing
improves NOx reductions due to producing more favorable temperatures in the boiler (due
to differences in heat transfer characteristics between oil and gas fuels). Furnace exit gas
temperatures are about 200°F lower for oil-firing than comparable gas-fired conditions.
NO, reductions over 30 percent were achieved with gas-firing over the load range of 80 to
320 MW. Changes in combustion conditions, however, resulted in variations in NO,
reduction performance. Even with the variations in urea system performance, the lowest
6A-5
-------
N0X emission levels, down to 21 to 45 ppm (@ 3% 02) depending on load, were achieved
by coupling low NO,, modified combustion conditions with urea injection.
Effect of Urea Injection Parameters
Tests to optimize urea injection performance at each unit involved parametric evaluation of
urea injection process variables. The variables considered included; atomizer design,
solution flow and pressure, location and injector orientation at each injection location.
Conclusions from these parametric tests for both units include the following03:
• Atomizer design and the resulting spray characteristics (spray distribution
and angle, droplet size distribution, and injection momentum) affect NOx
reduction performance. The effect of these atomizer specific characteristics
are related to the penetration of urea spray into the furnace flow, the
resulting mixing of urea with the furnace gases, and the rate of evaporation
and the ultimate location of release of urea into the furnace; gases',
• The location of injectors and their orientation can improve NO* reduction
performance by taking advantage of furnace flow dynamics to enhance
mixing of urea with the furnace gases and maximize residence time at
optimum reaction temperatures.
Because of the fundamental differences in the thermal and mixing environments presented
by the two different units, the injector design and performance characteristics (i.e., droplet
size distribution, spray angle, injection momentum, etc.) were significantly different. In
relative terms, the requirements for Encina Unit 2 compared to Etiwanda Unit 3 were
injectors which produced small urea solution droplets; lower injection momentum to cover
the furnace gas flow across the entire cross-section of the boiler; and spray angle, shape,
and location of ports to inject across the cross-flowing stream. These requirements are
consistent with the characteristic differences between the two units, including:
• Favorable furnace gas temperatures in the region of injection at Encina Unit
2 for urea NO, reduction reactions to occur, thus requiring the fast
evaporation and release of urea from small solution droplets;
• Small furnace cross-section dimensions in the region of injection requiring
only relatively low injection momentum for adequate penetration and mixing
of urea droplets with the furnace droplets;
• More uniform furnace gas flow with less cross-mixing due to the front wall
firing configuration compared to the swirling flow field resulting from
tangential firing, requiring use of ports physically spaced across the boiler.
6A-6
-------
The requirements for Etiwanda Unit 3, on the other hand, were satisfied by urea solution
injection characteristics which included large droplets that would delay the evaporation and
release of urea from the high temperatures at the point of injection for reaction in lower
temperature regions, In addition, the injectors and locations were developed to optimize the
distribution and mixing of the urea solution by taking advantage of the furnace flow
dynamics of the tangentially, down-fired configuration. In fact, in a brief series of trials
to establish a direct comparison for urea injection between Encina Unit 2 and Etiwanda Unit
3, the injectors achieving optimum performance at Encina Unit 2 were found to achieve
essentially no NO, reduction at Etiwanda Unit 3 at full load conditions.
OPTIMIZATION FOR VARYING CONDITIONS
The data from these two utility boilers demonstrate that unit design and operating conditions
can affect urea NO, reduction performance. Since urea systems are designed by necessity
for optimum performance at selected, typical operating conditions, NO, reduction
performance will vary. However, the design of urea injection and control systems can
incorporate adjustable parameters to accommodate intermediate or varying conditions. This
potential to control over varying conditions has been demonstrated at both Encina Unit 2 and
Etiwanda Unit 3:
Multiple Level Injection
At Encina Unit 2, for example, simultaneous injection from multiple levels improved NO,
removal at both high and low loads®. In a multiple injection configuration, a reduced
dosage of urea (lower NSR) is injected at each elevation. This improves urea utilization
and, in turn, the overall NO, removal. This improved utilization also reduces byproduct
NHj emissions. Figure 6 compares NO, reduction performance at Encina Unit 2 achieved
with bi-level injection for natural gas and fuel oil-firing. The method of bi-level injection
reduced the sensitivity of NO, removal to load. In addition, similar performance was
achieved for the two different fuels even though the resulting furnace temperature profiles
are distinctly different.
Injection Location and Orientation
Another technique used at both units to adjust for varying operating conditions was adjusting
injection location by varying injector orientation. In practical applications of the urea
injection process, boiler penetrations to accommodate urea injectors will be selected to
provide access into favorable temperature regions for a limited number of conditions or
loads. To maintain urea NO, reduction performance for intermediate loads or changes in
operating conditions, the orientation of the injectors can be used to adjust the relative
location of urea injection. Recent tests were conducted at Encina Unit 2 to evaluate the
optimization of urea injection with the combustion modification technique of fuel biasing
6A-7
-------
to achieve minimum stack NO, emissions. The test results illustrate how varying orientation
from available injection locations can improve performance and how orientation can be used
to maintain NO, reduction performance as operating conditions vary.
Urea NO, reduction performance was evaluated with and without fuel biasing by screening
injection location and orientation. Tests were completed for loads of 80 MW and 50 MW.
At 80 MW with ABIS operation, the highest NO, reduction achieved was 44.3 percent using
the lower level injectors only pointed up and urea injected at a rate of NSR = 2.0. This
reduction resulted in NOx emissions of 50 ppm (@ 3 % 02) from a baseline of 91 ppm. With
fuel biasing at the same load, however, the highest NO, reduction achieved was 29.1 percent
using simultaneous bi-level injection with both the upper and lower elevations of nozzles
pointed up and urea injected at NSR = 2.0. The optimum urea injection configurations thus
shifted for the two different firing modes.
The reasons for this shift appear to be a shift in furnace temperature. Furnace exit
temperatures increased about 40 °F with fuel biasing. As a result, NOx reduction was
improved by injection at a higher, and therefore cooler, elevation for fuel biasing conditions
than for normal ABIS operation. Although relative urea NO, reduction performance was
decreased with fuel biasing compared to ABIS, stack NO, emissions were reduced from 50
ppm (@3% Oj) for ABIS and urea down to 38 ppm (@3% 02) for fuel biasing and urea.
At 50 MW the data indicate that, for ABIS operation, injecting urea through the lower
elevation with nozzles pointed upward achieved the highest NO, reduction. For fuel bias
operation, however, the best configuration was bi-level injection with the upper elevation
injectors pointed down and the lower elevation injectors pointed up. As for the 80 MW
case, the shift in optimum injection configuration for the two operating conditions suggest
contributing affect of a change in furnace gas temperature. The data also indicate that
significant reductions can be achieved for low initial NO, levels, resulting in stack emissions
down to 23 ppm for an NSR = 1.7.
Dilution Water Flow and Injection Momentum
At Etiwanda Unit 3, three elevations of injection ports were determined to provide coverage
over the unit's normal load range, 80 to 320 MW, as shown in Figure 7. However,
Etiwanda Unit 3 is also routinely operated down to 20 MW. Test results demonstrated that
dilution water flow could be used in conjunction with injector elevation and orientation to
adjust the ultimate fate of urea droplets and achieve NO, reductions at loads less than 160
MW. By varying dilution water flow, the solution concentration, injection momentum, and
resulting droplet size distribution is changed. The parameters directly affect the point at
which the urea is released from solution to react with the furnace gases.
The performance of the urea NO, reduction system at Etiwanda Unit 3 is illustrated in
Figure 8. The optimized system is used together with combustion modifications to achieve
NO, levels of 20 to 45 ppm over the entire load range of 20 to 320 MW. This represents
6A-8
-------
significant reductions in NO, compared to normal, as found conditions also shown for
reference. In addition to the NOx reductions achieved, the available data indicate that
byproduct NH3 emissions below 20 ppm could be maintained up to urea flowrates
corresponding to NSR = 2.0. Figure 9 illustrates typical NH3 emissions measured at
Etiwanda Unit 3.
CONCLUSIONS
The effectiveness of the urea NO, reduction process is sensitive to temperature and mixing
phenomena as well as chemical stoichiometry (NSR). Since the urea NO, reduction process
occurs within the boiler furnace, the ultimate performance of the urea process is thus
dependent on boiler design and operating characteristics. Although the design of SNCR
systems must attempt to address these factors, realistic limitations must be imposed on the
range of expected boiler operating conditions (fuel type, load, burner firing pattern, excess
air, FGR flowrate, etc.) over which the system performance can be optimized.
To accommodate differences in boiler design and variations in operating conditions, urea
injection process parameters can be adjusted and optimized. Improvements in urea NO,
reduction performance and, ultimately stack NOs emissions, can be achieved by modifying
combustion conditions, optimizing injection location and orientation, and adjusting injection
nozzle droplet size and injection momentum. NO, reductions of about 30 percent at NSR
= 2.0 could be achieved over the load range of 20 to 320 MW at Etiwanda Unit 3, resulting
in stack NO, emissions in the range of 20 to 45 ppm (@3% 02) when combined with
combustion modifications. At Encina Unit 2, similar reductions and stack NO, levels (23
to 38 ppm @3% 02) could be achieved when urea injection was coupled with the
combustion modification technique of fuel biasing. In general, the data trends suggest that
for these gas- and oil-fired boilers, more confined heat release zones provide a more
favorable furnace environment than deeply staged, delayed mixing conditions.
REFERENCES
1. J.H. Nylander, M.N. Mansour, and D.R. Brown, "Demonstration of an Automated
Urea Injection System at Encina Unit 2," in proceedings of the Joint Symposium on
Stationary Combustion NO, Control, EPRI Report GS-6423, July 1989.
2. A.R. Abele, D.R. Brown, Y. Kwan, M.N. Mansour, and J.H. Nylander,
"Demonstration of Urea Injection for NO, Control on Utility Boilers," in proceedings:
GEN-UPGRADE 90, EPRI Report GS-6986, September 1990.
6A-9
-------
6A-10
-------
70
60
50
40
30
20
10
NOx Removal (%)
NSR = 2.0
ABIS -B-Air Bias -A- BOOS
0 20 40 60 80 100 120
Load (MW)
Figure 2. Effect of Combustion Conditions on Urea
NOx Removal at Encina Unit 2, Gas Fuel.
6A-11
-------
90
80
70 h
60
50
40
30
20
10
0
NOx (ppm @ 3% 00)
NSR ¦ 2.0
-0- ABIS -B- Air Bias BOOS
0 20 40 60 80 100 120
Load (MW)
Figure 3. Effect of Combustion Conditions on
Stack NOx Emission Levels with Urea
at Encina Unit 2, Gas Fuel.
6A-12
-------
60
NOx Removal (%)
NSR = 2.0
50 f-
40
30
20
10
0
-O- Oil-As Found
S- Gas-Comb. Mod.
Gas-As Found
0 50 100 150 200 250 300 350
Load (MW)
Figure 4. Effect of Operating Conditions on Urea NOx
Reduction Performance at Etiwanda Unit 3.
6 A-13
-------
90
80
70
60
50
40
30
20
10
NOx (ppm 3% 00)
NSR m 2.0
Oil - As Found -A- Gas - As Found
~B- Gas - Comb. Mod.
0 50 100 150 200 250 300 350
Load (MW)
Figure 5. Effect of Operating Conditions on Stack NOx
Emission Levels with Urea at Etiwanda Unit 3.
6A-14
-------
80
NOx Removal, Percent
70 -
60
50
NSR ¦ 1.0
40
30 -
20
10
50
Gas Firing "V- oil Firing |
60
70 80
Load, MW
90
100
80
NOx Removal, Percent
40
30
20
10
50
NSR - 2.0
Gas Firing Oil Firing
60
70 80
Load, MW
90
100
Figure 6. Comparison of NOx Reduction with Bi-Level Injection
for Natural Gas and Fuel Oil Firing at Encina Unit 2.
-------
• Urea Injection
Ports
O Unused Ports
Loop 3
Side View
Figure 7. Etiwanda Unit 3
Division Wall
o
O .
o o
o o
o • o
Loop 2
o • o
Loop 2
Front
View
Urea Injection Port Locations
-------
NOx (ppm @ 3% 02)
110
100
90
80
70
60
50
40
30
20
10
0
n NOx
Urea * Combustion Modification NOx
50
100
150
200
250
300
350
Load (MW)
Figure 8. Overall NOx Reduction Performance at Etiwanda Unit 3,
Gas Fuel.
-------
0 1 2 3 4
NSR
Figure 9. Typical NH 3 Emission from Optimized
Urea System at Etiwanda Unit 3, Gas Fuel.
-------
TABLE 1. BOILER DESIGN CHARACTERISTICS
Design
Encina
Etiwanda
Parameter
Unit 2
Unit 3
Capacity (MW)
110
320
Firing
Front
Tangential
Configuration
Wall
Down-Fired;
-
Divided Furnace
Burners
2 Rows x
3 Elev/
5 Burner
Corner
x 8 Corner
Peabody
CE
Dimensions
Height (ft)
77.0
88.1
Depth (ft)
20.0
22,1
Width (ft)
34.0
60.0 (30/30)
Steam Flow (lb/hr)
700,000
2,305,000
SH Temperature (°F)
1000
1050
RH Temperature (°F)
1000
1000
Steam Press, (psig)
1450
2450
6 A-19
-------
Previous Page Blank
WIDENING THE UREA TEMPERATURE WINDOW
D. P. Teixeira
Research & Development Department
Pacific Gas and Electric Company
San Ramon, CA 94583
L. J, Muzio
T. A. Montgomery
G. C. Quartucy
T. D. Martz
Fossi! Energy Research Corporation
Laguna Hills, CA 92653
6A-21
-------
-------
ABSTRACT
The results of laboratory tests to widen the effective temperature range while, at the same time,
minimizing byproduct emissions for the urea injection SNCR process are described. Data are
presented showing the effect of a number of additives {methane, combination of hydrocarbons, carbon
monoxide, ethylene glycol, HMTA, and furfural) and initial NO, level (125 and 250 ppm) on NO,
removal efficiency and byproduct emissions (N H3, CO, N20) as a function of temperature. Several new
phenomenon not previously observed are described. Of particular interest is the strong effect of CO
on N20 emissions during urea injection. In addition, many additives were found to improve NO
reduction but not NO, reduction, In these cases, the presence of additives converted the NO Initially
present to NOz and/or N20.
6A-23
-------
INTRODUCTION
A variety of technologies Is available to control NOx emissions from fossil power plants. One attractive
option is selective non-£atalytlc reduction (SNCR) with urea Q). However, the SNCR process, which
has many attractive features, does have several disadvantages. One drawback is the relatively narrow
temperature "window" over which the process is effective. Another potential disadvantage is the
emission, at least under some operating conditions, of undesirable byproducts such as NH3 or CO.
These issues become even more important for units which are cycled frequently or use multiple
fuels-which is the case for many fossil plants.
Results of a series of laboratory tests to address the issues noted above through the use of additives
to the basic urea Injection process are described in the sections which follow. The effects of additive
type, additive concentration and initial NO, level on NO, removal and byproduct emissions as a function
of temperature are presented.
PROCESS DESCRIPTION
Conceptually, the SNCR process with urea is quite simple. An aqueous solution of urea is injected
into, and mixed with, the flue gas at the correct temperature. After the mixing has been completed,
the urea then reacts selectively to remove the NOr
In practical applications, however, the process (and the equipment required) can be much more
complicated. Non-uniformities in velocity, temperature, and NO, concentration at the point of injection,
along with the variation in the physical location of the effective process temperature range within the
boiler, depend on various operating factors including load, type of fuel fired, and length of time on a
particular fuet. These factors often lead to multiple levels of injection and/or use of additives to
accommodate the shifts in temperature.
6A-24
-------
PILOT SCALE TEST FACILITY
A schematic of the pilot-scale facility used for these tests is shown in Figure 1. The pilot scale
combustor fires natural gas, doped with NH3 to control the initial NO, level. The combustor and test
section are refractory lined with the test section being 15 cm in diameter and 240 cm long. At the firing
rates used for these tests, the residence time in the test section is nominally 0,5 seconds, while the
temperature drop along the test section is nominally 250°C/sec (45QBF/sec). The SNCR solutions were
injected into the combustion products at the combustor throat through a small air assist atomizer,
above the test section. The atomizer was fabricated into a water cooled holder. The atomizer was
located at the center of the throat with the spray directed downward (i.e., co-flowing with the
combustion products). The solutions were pumped with variable speed peristaltic pumps and metered
with rotameters. In order to maintain a constant thermal environment in the test section, the total
amount of liquid injection was held constant at nominally 1 liter/hr. By diluting a concentrated urea (or
other SNCR chemical) solution with distilled water, the amount of chemical reagent was varied while
a total liquid flow rate of 1 liter/hr was maintained.
Gas samples were taken at the exit of the combustor with a water-cooled probe and transported to a
series of gas analyzers (NO/NO,, N20, CO, C02, and 02). The continuous measurement of N,0 was
made using an NDIR based technique (2). NH3 was measured using a selective ion electrode
technique.
The pilot-scale tests investigated the effect of temperature, additives, chemical injection rate, and initial
NO, concentration on NO, removal efficiency and byproduct emissions (specifically NH3, CO, and
Np).
RESULTS
During this study, experiments were carried out at initial NO, levels of 125 ppm and 250 ppm aid
N/NO, molar ratios of 1 and 2. For brevity, most of the results shown in this paper will be from the
tests at an initial NO, level of 125 ppm. Results at 250 ppm will be shown for situations where the
effect of the SNCR chemical, or additive, exhibits different behavior from that observed at the 125 ppm
level.
Baseline Performance of Urea - No Additive
To establish a reference for comparison of results from the various additives, a series of baseline teste
were performed using urea alone. The baseline NO, removal and byproduct emission results over
6A-25
-------
BUSNEH FLOW SYSTEM
Figure 1. Pilot-Scale Combustor Facility
6A-26
-------
the temperature range investigated for initial NO, levels of 125 and 250 ppm and a urea injection rate
corresponding to molar ratios of nitrogen to NO, (N/NO,) of 1 and 2 are shown In Figures 2 and 3.
Figure 2 shows the results for an initial NO* level of 125 ppm. Figure 3 shows the same data but for
an initial NO, level of 250 ppm. The narrow effective process temperature range for NO, removal can
be clearly seen in both figures, as can the increasing levels of NH3 and CO byproducts as temperature
is decreased, Also shown are byproduct levels of N20 produced by the process at the test conditions.
Other investigators have also noted N20 byproducts associated with urea injection (3).
Carton Monoxide Additive
A review of the general combustion chemistry literature showed that CO was a potential compound that
could alter the temperature dependance of the urea injection process. This behavior was also
suggested by the data of reference 4 showing the effect of CO at high concentrations (8000 ppm CO)
on NO, removal. While the use of CO to modify the urea temperature window in power plant boilers
presents several difficult practical application issues, it was felt important to address the effect of CO
since all combustion devices emit some level of CO.
For the data discussed below, the CO additive was introduced by injecting it with the atomizing air.
NO. Removal Temperature Dependance. Figure 4 shows the effect of CO on NO, removal as a
function of temperature at an initial NO, level of 125 ppm and N/NO, ratio of 2. This figure shows
several interesting features:
• CO, even in relatively low amounts, has a significant Impact on the NO, removal efficiency at a
given temperature. As CO levels are increased, the NO, removal versus temperature
dependence shifts to a lower temperature regime. Figure 4 shows that, increasing the CO levels
from O ppm to 1000 ppm shifts the peak NO, removal temperature about 200°F lower.
• As CO levels increase, the effective process temperature range is broadened. For the conditions
of Figure 4 when CO is in the 500-1000 ppm range, the window appears to be broadened by
about 100°F.
• Increasing CO also lowers the peak level of NO, removal possible. Figure 4 shows that peak NO,
removal decreases from about 55% to 45-50% as CO increases from 0 ppm to 500 ppm; it further
decreases to about 45% as CO is increased to 1000 ppm. Similar behavior is noted for the other
conditions .investigated.
CO Byproduct Emissions. The final CO levels resulting from addition of CO to the urea process are
shown in Figure 5. As can be seen, at the lowest temperature evaluated, 1470°F, CO emissions
increase as the initial amount of CO addition is increased. However, for temperatures at or above
1600-1650°F, final CO levels are practically independent of the amount of CO added.
6A-27
-------
I*
9- c
a.g
tn G
C S»
0
•s u
m cz
1 ^
in y
-20
1400
J_
» • I
(a) N/NO, = t
1500 1600 1700 1800 1900 2000 2100 2200 2300
Temperature, °F
g s-
a. c
CL O
W §
c 2
.2
g(T
M§
(b) N/NO, = 2
•20
1400
1500 1600 1700 1800 1900 2000 2100 2200 2300
Temperature, *F
Figure 2. NO, Reduction arid Byproduct Emissions with Urea Injection
(Initial NO, = 125 ppm)
6A-28
-------
E
Q. C
CL O
*S
.2 s
C X
il o
(a) N/NO, = 1
1500 1800 1700 1000 1900 2000 2100 2200 2300
Temperature, °F
£ 3*
Q. c
Q. O
W O
c -3
o 13
•s ®
w rr
p J<
LU O
z
(b) N/NO. = 2
1400 1500 1600 1700 1800 1900 2000 2100 2200 2300
Temperalure, "F
Figure 3. NO, Reduction and Byproduct Emissions with Urea Injection
(Initial NO, = 250 ppm)
6A-29
-------
c
.a
XI
-------
NH, Byproduct Emissions. Since minimum unreacted NH3 from the SNCR process Is desirable both
from an environmental, as well as boiler impact standpoint, measurements of the byproduct NH3 were
made. Figure 6 shows the results of these measurements for an initial NOx level of 125 ppm and
N/NO„ ratio of 2, As expected, NH3 emissions decrease as temperature Increases. However, NH3
levels at any given temperature, were found to decrease significantly as CO levels increased.
N„Q Byproduct Emissions. The most interesting influence of CO on the urea injection process was on
the Nj,0 byproduct characteristics (Figure 7). The effect of CO on N20 is strongly temperature
dependent. At higher temperatures (approximately 1900°F and above), N20 levels tend to merge to
a similar low level for all combinations of CO, initial NO, and N/NO„. At these high temperatures, N20
tends to decrease rapidly to very low levels as temperature is increased.
However, at the lower temperatures investigated (1500-1600sF}, a very different behavior can be seen;
NjO levels increase with increasing CO levels. For example, at an initial NO, of 125 ppm and N/NO,
= 2, NzO increases from about 10 ppm to 35 ppm as CO is Increased from 0 ppm to 1000 ppm.
Although not shown, NzO emissions at these lower temperatures also Increase as the amount of urea
{i.e. N/NOJ and initial level of NO, increase. At the highest initial NO, (250 ppm),N/NOx (2), and CO
(2000 ppm) levels investigated, N20 concentrations approach 100 ppm.
At the intermediate temperatures (between 1500°F and 1900°F), there is a transition from the low
temperature behavior to the high temperature behavior. At the lower CO levels, increasing
temperatures first produce an increase in N.O then a decrease as temperature is increased, with an
obvious maximum in the NsO as a function of temperature. At higher CO levels, N20 initially remains
relatively constant as temperature increases, then drops off abruptly.
Implications. There are several important practical implications regarding the influence of CO on the
urea injection process, in particular the N20 characteristics. First, to minimize' N20 production in the
urea injection process it is important to maintain low CO levels.
Second, when using urea injection, a "coupling" between the combustion process and the urea
injection process can occur, i.e. CO produced in the burner region influences the SNCR performance,
This may be especially true for low NO, burner systems where, as is well known, there are frequently
trade-offs between the NO, reduction and CO levels.
Lastly, the effect of CO on N20 formation may explain some of the differences in N20 levels reported
by various researchers at a recent workshop on N20 (5),
6A-31
-------
<;au 1 r~
CO Addition
200
150
100
50
1100 pprn
A 0 pprn
A 65 pprn
* 125 ppm
0
1400 1500 1600 1700 1800 1900 2000 2100 2200 2300
Temperature, °F
Figure 6. Effocl of CO Additive with Urea on Byproduct NH, Emissions
{Initial NO, = 125 ppm; N/NO, = 2)
CO addition
O 0 ppm
¦ 500 ppm
~ 1000 ppm
1400 1500 1600 1700 1800 1900 2000 2100 2200 2300
Temperature, CF
Figure 7. Effect of CO Additive with Urea on Byproduct N20 Emissions
(initial NO, = 125 ppm; N/NO, = 2)
6A-32
-------
Methane - Additive
Methane (CK<) was also investigated as a potential additive to alter the urea/NO, removal temperature
dependance. The results shown are for tests conducted at 1600°F, N/NO, = 2, initial NQX levels of 125
ppm and 250 ppm, arid CH4/NO, molar ratios of 0, 0.5 and 1, Figures 8 and 9 show the results.
For the initial NO, levels investigated, both NO and NO* (N0+N02) levels decrease with the addition
of urea alone. However, when methane is added, while the NO levels continue to decrease for both
initial NO, levels, the effect on NO, differs. At an initial NO, level of 250 ppm, NO, levels continue to
decrease with the addition of CH4. However, at the lower initial NO, level of 125 ppm, while NO levels
decrease with CH„ addition, NO, levels remain constant. At this lower initial NO, level, the effect of
the CH4 is to oxidize NO to N02, rather than to enhance the SNCR process.
The effect of methane additive with urea on N20 emissions is also included in Figures 8 and 9. At both
initial NO, levels, methane promotes the formation of Ns0 as a byproduct; the N?0 levels increase with
increasing amounts of CH4.
Efforts to explain the significantly different behavior between the two initial NO, cases have to date
been unsuccessful. The possibility of hydrocarbon interference with the N20 measurements, which is
known to occur for the instrument used, was considered but could not explain the results observed.
Multiple Additives
NO. Removal Efficiency. A Japanese patent (6) Identifies multiple hydrocarbon additives used with
urea to broaden the temperature window. A specific example was presented for the following
conditions: urea at N/NO, = 4; initial NO, * 990 ppm; temperature = 1400°F; and additives consisting
of ethylene glycol, propane and carbon. Without the additives (i.e. urea only) the NO, reduction, as
expected, was low, under 10%. With the additives, the NO, reduction was increased to almost 75%.
A series of tests were performed to verify the performance of the multiple additives under the following
conditions: initial NO, = 790 ppm; temperature of 1400°F; N/NO, = 4; ethylene glycol/urea concentration
of 9.5%; propane to urea of 57%; and carbon/urea of 33%. All concentration ratios are on a molar
basis. The tests were conducted sequentially to evaluate the individual, as well as combined, effect.
The results are shown in Figure 10.
As can be seen in Figure 10, the addition of glycol resulted in an increase of NO, removal from about
10% with urea only to about 20%. Addition of propane increased the NO, removal to almost 55%.
6A-33
-------
E tOO
o
CM
z
CJ
o
O
Initial
NO + N02 + N20
NO + N02
Urea Urea + Urea +
0.5 CH4/Uroa 1.0 CH4/Urea
Figure 8. Effect of Methane Additive with Urea Injection on NO, NO?, and JSlsO
(Temperature = 1600°F; Initial NO, = 125 ppm; N/NO, = 2)
300
250
e 200 -
0
CM
z
01
O
O
z
150 -
100
Initial
NQ2 + N20
N02
Urea Urea-t- Urea +
0.5 CH4/Urea 1.0 CH4/Urea
Figure 9, Effect of Methane Additive with Urea Injection on NO, N02 and NgO
(Temperature = 1600°F; Initial NO, = 250 ppm; N/NO, = 2)
6A-34
-------
0
1
3
T)
®
CE
Z
"O
i
o
Ursa
+ Glycol
+ Propane
+ Carbon
Urea
+Glyco!
n . _ +Propane
+ Propane _ ,
r + Carbon
Urea
+Glycol
~ %dNO
IS %ANOx
(Reference 6)
Preseni Tesls
Figure 10. Effect of Multiple Additives on NO, Redueiton with Urea
.2
G
3
E
EC
X
O
2
"O
c
03
O
~ %ANO
¦ %ANOx
Figure 11. Effect of Multiple Additives on NO, Reduction with Urea
(Temperature = 1400°F; Initial NO, = 760 ppm; N/NOx = 1)
6A-35
-------
Further addition of the carbon actually resulted in a small deterioration in NO„ removal. While the 55%
removal did not quite match the 75% value cited in the patent, the results were sufficiently encouraging
that additional tests were conducted. ,
The next series of tests were done under nominally the same conditions as above (initial NO, of 755
ppm; temperature of 1400°F; ethylene glycol/urea of 9.9%; propane/urea of 60%), but at a lower N/NO„
ratio of 1.0. Results of these teste are shown in Figure 11. No improvements in NO, removal were
noted for the case of glycol-only addition. NO, removal increased to about 15% for the glycol plus
propane case. No change was seen in NO, removal when methane was substituted for propane. It
should also be noted that, in the urea plus glycol plus propane, or methane, cases, the NO, removal
was significantly lower than the NO removal.
The final series of tests considered the multiple additive concept at conditions of greatest practical
interest: temperature of 1600°F; initial NO, levels of 125 ppm and 250 ppm; N/NO, of 2; ChyNO,
values of 0, 0.5,1; ethylene glycol/urea of 10%. The results of these tests are shown in Figures 12
and 13. At 125 ppm initial NO, (Figure 12), little benefit of the multiple additives was observed. Some
improvement in the NO removal was noted for glycol addition alone. Very little NOx removal
improvement was noted. As discussed previously, no effect of methane on either NO, or NO removal
was observed other than to increase N20 emissions, A case where methanol was substituted for
glycol at a methanol/urea of 10% was investigated and yielded virtually identical results.
Contrary to the general lack of improvement in NO, performance at 125 ppm, meaningful improvement
in NO, (and NO) removal was observed when glycol was added, and/or when CH4 was added at a
higher initial NO, level of 250 ppm (Figure 13). As will be seen in the next, section, N02 and NaO
formation from the initial NO explains at least a portion of the difference between the NO* and NO
removal levels.
N0,/N,0 Characteristics. Data summarizing the NOz and N80 characteristics of the multiple additive
concept are summarized in Table 1.
6A-36
-------
a
fi
d
Cvj
2
oi
O
O
2
Urea N/NO (molar)
Glycol/Uraa (molar)
Cl-W/Urea (molar)
NO + N02 + N20
N02
Initial
0.0
0.0
0.0
Urea
2.0
0.0
0.0
Urea/Add. Urea^Add. Ursa/Add.
2.0
0.1
0,0
2,0
0.1
0.5
2.0
0.1
1,0
Figure 12. Effect of Multiple Additives (Ethylene Glycol and Methane) on
NO, Reduction with Urea
(Temperature = 160CTF; Initial NO, = 125 ppm; N/NO, = 2)
Q.
CL
O
CJ
CM
o
z
o
2
300
250
200
150
100
Urea N/NO (molar)
E.Glycol/Urea (molar)
CKM/Urea (molar)
NO + N02 + N20
Initial
0.0
0.0
0,0
Uisa Urea/Add. Ursa/Add. Ursa/Add.
2.0
0.0
0.0
2.0
0.1
0.0
2.0
0.1
0.5
2.0
0.1
1.0
Fiaure 13. Effect of Multiple Additives (Ethylene Glycol and Methane) on
NO, Reduction with Urea
[Temperature = 1600°F', Initial NOj = 250 ppm, N/NO, = 2)
6A-37
-------
Table 1
NOj AND N20 - MULTIPLE ADDITIVES
T = 1600°F N/NO = 2 Ethylene Glycol/Urea = 10%
initial NO.
CHJUrea
Final NO„
N,0
N.O + NO,
ppm
molar
gpm
ppm
ppm
125
0
19
13
32
125
0.5
24
21
45
125
1
25
28
53
250
0
33
26
59
250
0.5
37
42
79
250
1
31
63
94
Although not shown, virtually Identical data were collected for methanol under the same test conditions.
As can be seen, a portion of the original NO appears in the products as N02 and NzO. N02 levels
were roughly in proportion to the initial NO„ levels and tended to increase as the CH«/urea increased.
Likewise, N20 increased approximately in proportion to the initial NO, and as CH4/urea was increased.
HMTA/Furfural Additives
A review of the patent literature also indicated that the addition of hexamethylenetetramine, CeHiaN4
(HMTA), and furfural (C6H4Og) to urea results in a broadening of the effective process temperature
range for NO, reduction (7.8,9).
A series of tests were conducted to evaluate the effectiveness of these compounds. The teste
evaluated HMTA addition alone and in combination with furfural. A temperature of 1650°F was used
for these tests. The quantity of additives used in the tests, was estimated based on the data contained
in References 7-9. Test conditions were as follows: initial NO, level of 250 ppm; HMTA/urea of 0.2;
furfural/urea of 3.65 (all on a molar basis). Results of these tests are shown in Figure 14.
Examination of Figure 14 shows that the addition of HMTA alone, or the HMTA/furfural mixture, led to
a meaningful improvement in both NO and NO, removal. However, the improvement in NO, is
considerably lower than the improvement for NO removal. Evaluation of the final NOs levels (Table
6A-38
-------
MNO = 1.0
N/NO *= 1.0
HMTA
Addition
N/NO = 1.0
N/NO = 1.8
HMTA/Furfural
Addition
N/NO = 1.0
N/NO = 1.8
Urea
Alone
~ %ANO
H %ANOX
N/NO = 2,0
N/NO = 2.0
Urea Alone
Urea and Additive
Figure 14. Effect of HMTA and Furfural on NO and NO, Reduction with Urea
(Temperature = 1650°F; Initial NO, = 250 ppm)
6A-39
-------
2) showed that a portion of the initial NO was being oxidized to N02. Unfortunately, data for NsO was
not collected during this test series, so a more complete assessment of the impact of HMTA/furfural
on byproducts could not be done.
Table 2
NOz LEVELS WITH HMTA/FURFURAL ADDITIVE
Temperature = 1650°F Urea/NO, = 1
HMT A/Urea
(Molar)
Furfural/HMTA
(Molar)
Initial NO.
(Ppm)
Final NO,
teml
0,2
0.2
0
3,65
15
15
60
58
Since the nitrogen in the HMTA increases the effective N/NO, ratio from 1 to 1.8, Figure 14 also shows
the NO, removal expected for the urea only case at N/NO, = 2. This allows an alternative comparison
of the behavior of HMTA since one alternative to the use of HMTA additives would be to increase the
N/NO, by increasing the amount of urea injected in place of adding the HMTA. As can be seen,
increasing the amount of urea injected provided a comparable degree of NO, removal when compared
to HMTA, or HMTA/furfural addition.
Future Research
Continuation of efforts to find additives or alternative reducing agents to improve the SNCR process
will be pursued in the future. In' addition, a series of tests to evaluate the effect of CO additive with
NHa as a reducing agent will be conducted and compared to the urea plus CO additive results.
CONCLUSIONS
A number of unexpected results were observed when testing various additives to the urea Injection
process: .
CO shifts and broadens the temperature window even at low CO levels; in addition
significant changes in the byproduct emissions, especially for N20, occur.
6A-40
-------
CH4 exhibits markedly different NO, and NO removal behavior depending on the initial
NO, level. Reasons for this behavior are not understood. CH4 addition also leads to the
conversion of NO to NQ2 (oxidation) and the formation of N20.
As with CH4, the use of multiple hydrocarbon additives leads to different NO, and NO
removal behavior, depending on the initial NO, level. The use of multiple additives also
leads to the conversion of a portion of the initial NO to N02 and N20.
The HMTA and furfural additives lead to the conversion of NO to NQ2, As a result, NO
removal Improves to a greater extent than the NO, removal. Further, it appears that the
improvement in NO, reduction can be attributed to the Increased N/NO, injection ratio
that results from the addition of HMTA.
In addition to the specific conclusions reached above for the individual additives, overall examination
of the results indicates a more general conclusion: The chemistry involved in urea NO, removal is
more complex than previously thought. As a result, when considering employment of the process to
a specific application, careful consideration of the initial NO, level and the levels of trace combustion
product species, including hydrocarbons and CO, is required.
6A-41
-------
REFERENCES
1. Arand, J. K,, Muzio, L. J., Sotter, J, G,, U.S. Patent 4,208,386. June 17, 1980.
2. Montgomery, T. A., et al, "Continuous Infrared Analysis of NaO in Combustion Products",
JAPCA Vol. 39, No. 5, May 1989.
3. Jodal, et al, "Pilot Scale Experiments with Ammonia and Urea as Reductants in Selective
Non-Catalytic Reduction of Nitric Oxide", 23rd International Symposium on Combustion,
Orleans, France, July 1990.
4. Slebers, D. I. and Caton, J. A., "Removal of Nitric Oxide from Exhaust Gas with Cyanurlc
Acid", Paper No. WSS/CI88-66, 1988 Fall Meeting of the Western States Section of the
Combustion Institute, Dana Point, California, October 1988.
5. Second European Workshop on NaO Emissions, Lisbon, Portugal, June 1990.
6. Kuze, T., et al, Japanese Patent 53128023. November 8, 1978.
7. Bowers, E. B.. U.S. Patent 4,751,065, June 14, 1988.
8. Epperly, R. E. and Sullivan, J. C., U.S. Patent 4.770,863, September 1988.
9. Epperly, W, R., Q'leary, J. H., Sullivan, J. C., U.S. Patent 4.780,289. October 25, 1988.
6A-42
-------
CATALYTIC FABRIC FILTRATION FOR
SIMULTANEOUS NOx AND PARTICULATE CONTROL
Greg F. Weber arid Dennis L. Laudal
Energy and Environmental Research Center
University of North Dakota
Box 8213, University Station
Grand Forks, ND 58202
Patrick F. Aubourg and Marie Kalinowski
Owens-Corning Fiberglass
P.O. Box 415
Granville, OH 43023-0415
Prepared for
Electric Power Research Institute
3412 Hill view Avenue
Palo Alto, CA 94303
6A-43
-------
-------
ABSTRACT
The Energy and Environmental Research Center (EERC) at the University of North
Dakota (UND) has been working with Owens-Corning Fiberglas Corporation (OCF) for
several years evaluating Catalytic Fabric Filtration for simultaneous NO, and
particulate control. Early work sponsored by OCF was presented at the 1989
EPRI/EPA NO, Symposium. Since April 1988, the U.S. DOE Pittsburgh Energy
Technology Center (PETC) has funded development activities at the EERC, with OCF
providing catalyst-coated fabric samples for testing.
The work has involved evaluating samples (1 ft2) of catalyst-coated fabric prepared
by OCF using actual flue gas from the combustion of pulverized coal. Dependent
variables included air-to-cloth ratio, ammonia/NO, molar ratio, and coal type
{bituminous, subbituminous, and lignite). Flue gas temperature was maintained at
650"±25CF. Resulting NO, removal efficiency and ammonia slip varied significantly
with air-to-cloth ratio. As the air-to-cloth-ratio increased from 2 to 6 ft/min,
NO, reduction decreased from 85-95% to less than 70% with corresponding ammonia
slip values ranging from 5 ppm to 360 ppm. For the short-term (8-hour) tests
completed, the four coals tested did not appear to have a significant effect on
catalyst-coated fabric performance. Bench-scale tests have demonstrated that 90%
NO, reduction can be achieved with an ammonia slip of <5 ppm.
6A-45
-------
INTRODUCTION & BACKGROUND
In 1990, the first major reauthorization of the Clean Air Act since 1970 was
enacted by Congress and signed into law by the President of the United States,
Although S0Z emissions are still the primary focus of acid rain control, studies in
Europe and the United States investigating the role of NO, in acid rain formation
and ozone chemistry have resulted in NO, control being an important component of
the new Clean Air Act (1,2). Specifically, the Clean Air Act Amendments of 1990
require a two million-ton reduction in NO, emissions by January 1, 1995.
Expectations are that NO, emissions will be regulated, more strictly at the local
level (state and local regulatory agencies) than as currently addressed under the
reauthorized Clean Air Act. Therefore, technology capable of achieving higher
levels of NO, control than those demonstrated by low NO, burners must be developed..
For the past six years, the Energy and Environmental Research Center (EERC), using
fabrics developed by Owens-Corning Fiberglas (OCF), has pursued the development of
the catalytic fabric filtration concept as an advanced NO, control technology. The
overall objective of the project is to evaluate the potential of a catalytic
fabric filter for simultaneous NO, and particulate control. Specific goals include
the following:
• 90% NO, removal efficiency with <25 ppm ammonia slip.
• A particulate removal efficiency of >99.5%.
• A bag/catalyst life of >1 year.
• A 20% cost savings over conventional baghouse and SCR control
technology.
• Compatibility with S02 removal systems.
• A nonhazardous waste material.
Even though promising results were obtained in the early bench-scale work funded
by OCF, a continued effort was needed to further develop the product that would
give the best combination of high NO, removal capability, low ammonia slip, high
particulate removal efficiency, and long catalyst/bag life.
Specific activities have progressed from bench-scale experiments using simulated
flue gas (Task A) and flue gas from a pc-fired source (Task B) to pilot-scale
experiments with catalyst-coated bags. Specific parametric and fabric-screening
tests using simulated flue gas (Task A) were conducted in which the fabric weave,,
coating composition, and coating process were adjusted to develop acceptable
fabrics for further testing. Task B, which is the focus of this paper, involved
the testing of ten catalyst-coated fabric samples developed by OCF using a
6A-46
-------
slipstream of flue gas from EERC's Particulate Test Combustor (PTC). Based on the
results of these bench-scale experiments, tests with catalyst-coated filter bags
are scheduled to begin in the summer of 1991.
RESULTS & DISCUSSION
The purpose of Task B was to further evaluate catalyst-coated fabric samples in
the presence of flue gas generated during pulverized coal combustion. This was
considered necessary to begin evaluating the potential effects of fly ash on
catalytic performance: specifically, the effects of submicron particles, volatile
species, and trace elements that could not be addressed using synthetic flue gas.
Ten catalyst-coated fabric samples (Fabrics #2, #3, #4, #5, #7, #13, #14, #15,
#17, and #18) developed by OCF were selected for testing. The criteria for
selecting these fabric samples for further evaluation were high NO, removal
efficiency and/or low ammonia slip, based on Task A results. Detailed
descriptions of eight catalyst-coated fabric samples were presented in a previous
report (3). Fabrics #17 and #18 were catalyst-coated fabric samples recently
developed by OCF. Fabric #17 was similar to previously tested Fabric #2, except
that a different vanadium source was used to prepare the coating, and
modifications were made to increase the surface area.
The catalyst coated on Fabric #18 was a new iron-based catalyst. Iron compounds
have been shown to be effective catalysts for reducing NO, (4). In addition, it
may broaden the temperature window for the NO, reduction reactions.
Four coals were selected for Task B testing, a medium-sulfur washed Illinois #6
bituminous (the baseline coal), a high-sulfur Pyro Kentucky bituminous, a Jacobs
Ranch subbituminous, and a South Hallsville, Texas, lignite. Each of the ten
fabrics was tested with the washed Illinois #6 bituminous coal at air-to-cloth
ratios of 2, 3, 4, and 6 ft/min. Ammonia slip and S03 measurements were made at
each air-to-cloth ratio. The ammonia/NO, molar ratio was to be held constant at
0.9; however, due to an error in calculating an orifice coefficient, several tests
were conducted at an ammonia/NO, molar ratio of 1.1. Cloth weight in all instances
was 14 ounces per square yard.
Based on the results of the first eight fabric-screening tests, two fabric
samples, #2 and #13, were selected to be tested using the remaining three coals.
For the first 6 hours of the test, the air-to-cloth ratio was held constant at 3
ft/min. However, near the end of each test, the air-to-cloth ratio was adjusted
to 2 ft/min for 1 hour and then 4 ft/min for 1 hour. The ammonia/NO, molar ratio
was held constant at 0.9. The slipstream sample system used to perform the tests
is shown in Figure 1.
The results of the Task B fabric-screening tests are presented in Table 1. These
results are consistent with the values reported for Task A. As expected, there
was a marked decrease in NO, removal efficiency with increased air-to-cloth ratio.
An example of this is shown in Figure 2. Although there was some variability in
the operation of the combustion system, NO, removal efficiency was relatively
constant with time. Fabric f2 appeared to have demonstrated the best overall
performance of the first eight fabric samples tested, with respect to high NO,
removal and low ammonia slip.
6A-47
-------
The results for Fabric #17, with the new vanadium source, compared favorably to
Fabric #2, which is similar in all other respects. The two fabrics are compared
directly in Figure 3, As can be seen, with the exception of the ammonia slip at
an air-to-cloth ratio of 2 ft/min, the results are very similar. Figure 4 shows
the actual ammonia/NO, molar ratio as a function of time for Fabric #17. As is
shown in the figure, the ammonia/NO, molar ratio averaged about 0.95 for the test
at an air-to-cloth ratio of 2,2 ft/min. This may have been the reason for the
higher ammonia slip at the lowest air-to-cloth ratio. Figure 4 data are typical
of the variability in ammonia/NO, molar ratio for all the tests.
For Fabric #18, the results did not seem to be very impressive (an NO, removal
efficiency of 64% at an air-to-cloth ratio of 2 ft/min); however, this is
promising, as the coating process for iron has not been optimized. As stated
earlier, iron presents several potential advantages over vanadium; however,
further development by OCF will be necessary to improve its performance.
From the fabric-screening data, the maximum air-to-cloth ratio that can be used
and still obtain >85% NO, removal efficiency is 3 ft/min, which is consistent with
the bench-scale results using simulated flue gas (Task A). For all the catalyst-
coated fabric samples, there was a marked decrease in catalytic performance at
air-to-cloth ratios of 4 and 6 ft/min.
Following completion of the first eight fabric-screening tests, fabric samples #2
and #13 were chosen to test the effects of coal type on fabric performance. Both
fabrics were tested using, the three remaining coals: South Hallsville, Texas,
lignite; Jacobs Ranch subbituminous; and a Pyro Kentucky bituminous at an air-to-
cloth ratio of 3 ft/min, ammonia/NO, molar ratio of 0.9, and temperature of 650°F.
Table 2 summarizes the results from these tests as well as data from the previous
screening tests using the washed Illinois #6 bituminous coal. The data are also
represented graphically in Figures 5 and 6.
From the data, it appears that NO, removal efficiency with Fabric #2 was similar
(85% to 90%) for three of the four coals fired in the pilot-scale combustor. The
exception was observed when firing the South Hallsville, Texas, lignite. Although
an obvious explanation of this result (80% NO, removal efficiency and 121 ppm
ammonia- siip) is not apparent, EERC believes that the filtration characteristics
of the South Hallsville fly ash may have contributed to the observed result.
Specifically South Hallsville, Texas, lignite is known to produce an ash that is
difficult to collect in a fabric filter (5), A large number of pinholes were
present in the dust cake at the conclusion of the test. Pinholes may result in
localized areas of very high air-to-cloth ratios which, depending on the number
and size of the pinholes, can limit contact between the flue gas and the catalyst,
resulting in decreased N0„ removal efficiency and increased ammonia slip.
For Fabric #13, the results using South Hallsville, Texas, lignite were more
successful, as excessive pinholing did not occur. Although the NO, removal
efficiency was somewhat lower, about 83% compared to 86% and 90% for the Jacobs
Ranch and Illinois #6 coals, respectively, the data is not conclusive. Therefore,
the effect of coal type, if any, on catalyst-coated fabric performance has not yet
been determined. The results using the Pyro Kentucky bituminous coal with
Fabric #13 are suspect due to an upset in the pilot-scale combustion system.
6A-48
-------
Excessive slagging resulted in an unstable flame in the burner, causing an early
shutdown of the test.
Table 3 presents surface area and catalyst data for each of the catalyst-coated
fabric samples tested. • Both were measured prior to exposure to the flue gas and
after completion of the reactivity tests. In all cases, there was a substantial
decrease in surface area after exposure to flue gas. But, for most of the fabric
samples tested, the catalyst concentration decreased only slightly or remained
constant with exposure to flue gas. However, this indicates that the decrease in
surface area is not due to sluffing of the catalyst from the fabric surface. The
decrease in surface area may be due to a slight sintering effect, possible
plugging of the surface pores by submicron aerosols or fly ash particles, or due
to residual carbon burnout in the coatings.
The initial BET surface area for both Fabrics #17 and #18 was higher than previous
fabrics. However, the surface area for Fabric #17 after exposure to flue gas
(which gave results very similar to Fabric #2) decreased to a level that was
essentially the same as that observed for Fabric #2, For Fabric #18 (iron
catalyst), there seems to have been almost a complete collapse of surface area.
The reason for this is not known at this time; however, it was speculated by OGF
that there may be some temperature effects. Figure 7 shows the NO, removal
efficiency as a function of the surface area after exposure to flue gas. One
surface area point does not fit the curve. This data point represents Fabric #7,
and a final determination concerning its validity has not been made. Fabric #7
may be tested again during upcoming pilot-scale activities. Although other
factors such as weave texturization may also be important, the figure shows that
NO, removal efficiency is directly proportional to the surface area. Based on this
data, the minimum BET surface area needed to achieve 85% NO, removal efficiency at
an air-to-cloth ratio of 3 ft/min is about 4-5 m2/g.
For Fabrics #17 and #18, N20 was measured at the inlet and outlet of the catalyst-
coated fabric. The measurements are shown in Table 4. Within the limits of the
instrument, the table shows that there is no apparent conversion of NO, to Nz0
across the catalyst-coated fabric. Downstream N20 values ranged from 4 to 6 ppm.
This is consistent with results presented by other researchers (6,7) for a
pulverized coal-fired boiler. Additional measurements will be made when pilot-
scale bag tests begin.
6A-49
-------
CONCLUSIONS
Based on the results of Task B testing, several conclusions can be made,
1. There was a substantial decrease in NO,, removal efficiency with increased
air-to-cloth ratio for all the catalyst-coated fabric samples tested. It
appears that for the 14 ounce per square yard fabric samples tested, in
the bench-scale system, the maximum air-to-cloth ratio at which 85%-90%
NO, removal can be achieved is 3 ft/min.
2. Although there was some variability in the data, the NO, removal
efficiency appeared to be constant with time over the short (eight hours)
duration of these tests.
3. Of the fabric samples tested, Fabrics #2 and #17 appear to provide the
best performance with respect to NO, removal efficiency and ammonia slip.
4. Although three of the coals, the two bituminous coals and" the
subbituminous coal, resulted in similar catalyst-coated fabric
performance, there appeared to be a reduction in NO, removal efficiency
for the South Hallsville, Texas, lignite. This may have been a result of
pinhole formation.
5. When the catalyst-coated fabric is exposed to flue gas, there is a
decrease in the total surface area. A minimum BET surface area after
exposure to flue gas of 4 to 5 nf/g is necessary to provide good catalyst-
coated fabric performance. Therefore, in order to improve performance,
it would be beneficial to increase the surface area of the catalyst or
the catalyst-coated fabric.
6. There does not seem to be any decrease in catalyst-coated fabric
performance using the new vanadium source. Although the NO, removal
efficiency using the iron catalyst is relatively low, it does show
promise, as the coating process for the iron catalyst has not been
optimized.
7. For these initial tests, there is no apparent conversion of NO, to N20
across the catalyst-coated fabric.
REFERENCES
1. Hjalmarsson, A.K.; Vernon, J. "Policies for NO, Control in Europe,"
Presented at: 1989 EPRI/EPA Joint Symposium on Stationary Combustion NO,
Control, San Francisco, CA, March 1989.
2. Bruck, R.I, "Boreal Montane Ecosystem Decline in Central Europe and the
Eastern United States: Potential Role of Anthropogenic Pollution with
Emphasis on Nitrogen Compounds," Presented at 1985 EPRI/EPA Joint
Symposium on Stationary Combustion NO, Control, Boston, MA, May 1985.
6A-50
-------
3. Weber, G.F.; Laudal, D.L. "Final Technical Project Report for April 1988
through June 1989 for Flue Gas Cleanup," Work performed under DOE
Contract No. DE-FC21-86MC10637, Grand Forks, NO, November 1989.
4. Kato, A.; Matsuda, S.; Nakajima, H.I.; Watanabe, Y. "Reduction of Nitric
Oxide on Iron Oxide-Titanium Oxide Catalyst," Journal of Physical
Chemistry 1981, 85, (12), 1710-1713.
5. Miller, S.J.; Laudal, D.L. "Flue Gas Conditioning for Improved Fine
Particle Capture in Fabric Filters: Comparative Technical and Economic
Assessment," Vol II. Advanced Research and Technology Development, Low-
Rank Coal Research Final Report, Work performed under DOE Contract No.
DE-FC21-86MC10637, Grand Forks, ND, 1987, Vol.III.
6. Aho, M.J.; Rantanen, J.T.; Linna, V.L. "Formation and Destruction of
Nitrous Oxide in Pulverized Fuel Combustion Environments between 750° and
970°C," Fuel 1990, 22, 957-1005.
7. Kokkinos, A. "Measurement of Nitrous Oxide Emissions," EPRI Journal
1990, April/Hay, 36-39.
6A-51
-------
Thermocouples
To Baghouse
To Gas Pump and
Dry Gas Meter
To Sample Conditioner
for Flue Gas Analysis
Figure 1. Slipstream Sample System
Fabric 12 (NHj/NO, Molar Ratio
o - A/C
a - A/C
• - A/C
0 - A/C
50
2 ft/min.
3 't/min
4 ft/min
6 ft/min
100
Time (min)
150
200
Figure 2. NO, Removal Efficiency as a
Function of Time and Air-to-Cloth
Ratio for Fabric #2
6A-52
-------
A/C = 2 ft/min A/C = 3 ft/min A/C = 4 tt/min A/C = 6 ft/min
NH3/N0* Molar Ratio «= D.9
Figure 3. Comparison of the NO, Removal Efficiency as
a Function of Air-to-CIoth Ratio for Fabrics #2 and #17
Time (min)
Figure 4. Ammonia/NO, Molar Ratio as a Function
of Time for Fabric #17
6A-53
-------
Fabric #2 Air-to-Cloth Ratio (ft/min) ¦
Illinois #6 Jacobs Ranch Pyro Kentucky South Hallsvilla
NH3/NO* Molar Ratio = 0,9
Figure 5. Comparison of the Catalytic Performance
Using Four Different Coals for Fabric #2
Fabric #13 Air-to-Cloth Ratio (ft/min) ¦¦ 2 |§s§ 3 mo 4
Illinois #6 Jacobs Ranch Pyro Kentucky South Hallsville
NH3/NO* NH3/NO* Molar Ratio ¦ 0.9
Molar Ratio
¦ 1.1
Figure 6. Comparison of the Catalytic Performance
Using Four Different Coals for Fabric #13
6A-54
-------
10
Air-to-CisUi Ratio = 3 f'./min
NH3/N0s Molar Ratio = b.9
0 -\ , 1 1 j 1
50 eo 70 80 90 IOC
NO^ Removal. Efficiency (%)
Figure 7. NO, Removal Efficiency as a Function
of Catalyst-Coated Fabric Surface Area after Exposure
to Flue Gas
6A-55
-------
Table 1
• RESULTS
FROM TASK B
— BENCH
-SCALE
FABRIC-SCREENING TESTS s'b
NO, ,
Particulate
A/C
NH,/NQ,
NO.
NO,
Removal
Amnion i a
Removal
Fabric
Ratio
Molar
Inlet
Outlet
Efficiency
SI l p
Efficiency
No.
(ft/mi it)
Ratio
(PPfflS
fppra]
m
(pom)
(%)
2
2
1.1
765
20
97.4
187
2
3
1.1
716
30
94.7
. 63
2
4
1.1 .
740
63
86,8
129
99.8
2
4.5
1.1
735
64
91.3
121
2
2
0.9
540 '
58
89.3
5 .
2 .
3
0.9
550
83
84.9
7
2
4
0.9
590
112
81.0
22
99.S
2
6
0.3
630
175
72.2
76
3
2
' 0.9
760
22&
70.3
NO
3
4
0.9
710
390
45.1
' ND
90.4
3
S
0.9
720
490
31.9
357
4
2
0.9
715
171
76.1
87
4
3
0.9
695
235
66.2
127 ,
4
4
0.9 .
675
310
54.1
' 179
99.5
4
6
0.9
645
436
32.4
288
5
2
0.9
730
90
8?.7
28
S
3
0\9
700
125
82.1
54
5
4
0.9
760
190
75.0
76
99.9
5
6
0.9
730
305
58.2
163
7
2
0.9
700
75
89,3
. 4
?
3
0.9
675
95
85,9
13
7
4
0.9
650'
175
73.1
33
99.8
7
6
0.9
660
200
69,7
50
13
2
1.1
673
34
94.9
64
13
3
1.1
666
64
90.7
58
13
4
1.1
688
126
81.7
88
99,4
13
6
1.1
671
209
68.9
108
14
Z
1.1
. 703
89
67.3
107
' 14
3
1.1
729
151
79.3
153
14'
4
1.1
772
229
70.5
256
99.8
14
6
1.1
83S
433
48,3
179
15
2
1.1
647
40
95.3
57
15
3
1.1
789
68
91.4
58
15
4
1.1 '
761
98
87.1
104
99.9
15
6
1.1
656
193.
70.6
122
17
2.2
0.9
306.
27
91.2
45
99.8
17
3
0.9
292
31
89.4
17
17
4
0.9
258
65
75.4
28
17
5.5
0.9
287
101
64.8
73
18
2
0.9
372
134
64.0
102
99.9
18
3
0.9
401
166
58.6
68
18
4
0.9
413
212
48.7
122
18
6
0.9
381 '
224
41.2
172
Each catalyst-coated fabric sample was evaluated using a slipstream of flu# gas from a pc-firad pi lot-scale
combustor firing a washed Illinois #6 bituminous coal.
"ND" denotes data that are not available due to problems encountered with the sampling system.
6A-56
-------
2
2
2
13
13
13
2
2
2
13
13
13
2
13
13
13
2
2
2
13
Table 2
RESULTS FROM TASK B — EFFECTS OF COAL TYPE
A/C
Ratio
fft/mi
NHj/NO,,
Molar
ratio
NO,
Inlet
(PPm) .
NO,
Outlet
IfiEffll
NO,
Removal
Efficiency
W
Ammon i a
Slip
(PP"0
Washed Illinois #6 Bituminous
Jacobs Ranch. Wyoming. Subbituminous
Pyro Kentucky Bituminous
Particulate
Removal
Efficiency
m
2
0.9
540
58
89.3
3
0.9
535
81
84.9
7
99.8
4
0.9
590
112
81.0
2
1.1
673
34
94.9
3
1.1
686
64
90.7
58
99.4
4
1.1
688
126
81.7
0.9
785
59
92.5
0.9
760
75
90.1
86
99.9
0.9
800
90
88.8
0.9
645
80
87.6
0.9
680
105
84.6
99
99.9
0.9
675
195
71.1
South H.al 1 sville. Texas.
Liqnite
0.9
900
175
80.6
121
0.9
820
110
86.6
0.9
810
140
82.7
75
99.8
0.9
825
195
76.4
2
0.9
970
93
90.4
3
0.9
930
130
86.0
10
99.7
4
0.9
925
178
80.8
3
0.9
810
170
79.0
30
99.6
6A-57
-------
Table 3
CATALYST CONCENTRATION AND BET SURFACE AREA
FOR EACH OF THE CATALYST-COATED FABRICS TESTED*
Catalyst Concentration1' Surface Area'
Unexposed
exposed
Change
Unexposed
(nf/ql
Exposed
(m /q)
Change
Fabric No.
(ma/a)
fma/al
(%)
(X)
Blank
0.03
0.56
2
9.1
9.0
1.1
9.50
6.19
34.8
2
8.4
8.3
1.2
10.68
5.11
52.2
3
4.7
3.7
21.3
3.31
1.54
53.5
4
4.7
4.2
10.6
4.28
2.02
52.8
5
5.5
5.4
1.8
5.79
3.74
35.4
7
7.6
6.3
17.1
6.62
2.74
58.6 .
13
6.8
6.1
10.3
5.76
4.04
29.9
13
8.4
8.0
4.8
6.52
4.00
38.7
14
3.4
3.6
(5.9)"
3.09
1.90
38.5
11
7.7
5.7
26.0
6.24
3.79
39.3
17
13.2
13,4
(1.5}"
14.61
5.05
65.4
18
7.1
7.4
(4.2 r
16.60
2.19
86.8
* Unexposed and exposed refer to exposure to flue gas.
° Catalyst concentration is mg catalyst per g of coated fabric.
c Fabric surface area is m2 per g of coated fabric {BET surface area).
3 { ) Indicates .there was a measured increase in catalyst concentration.
Table 4
N20 CONCENTRATION IN THE FLUE GAS
Air-to-cloth Inlet N20 Outlet N2Q
Ratio Concentration Concentration
fft/min) (mm) BEm3
• ¦ Fabric #17
2.2 4.0 5.0
3 3.5 4.5
4 4.0 4.5
5.5 4.0 4.5
Fabric #18
2 5.5 6.0
3 4.5 5.0
4 4.0 4.5
6 . 3.5 4.0
6A-58
-------
HETEROGENEOUS DECOMPOSITION OF NITROUS OXIDE IN THE OPERATING
TEMPERATURE RANGE OF CIRCULATING FLUIDIZED BED COMBUSTORS
T. Khan
Y.Y. Lee
L. Young
AhJstrom Pyropower Inc.
8970 Crestmar Point
San Diego, California 92121
ABSTRACT
There is growing concern over the increasing atmospheric nitrous oxide concentration. This concern
stems from the realization that nitrous oxide contributes to the depletion of the ozone layer and to
the greenhouse effect. A research program has been developed at Ahlstrom Pyropower Inc. to study
the emission of nitrous oxide from circulating fluidized bed combustors (CFBCs). The program
involves, in part, an investigation into the mechanism of nitrous oxide formation and destruction in
the operating temperature range of CFBCs. This paper describes a study directed at understanding
the decomposition of nitrous oxide on solid materials known to be present in the combustor.
An electrically heated tubular quartz reactor (2.3 cm I.D.) was used to study the decomposition of
nitrous oxide on six different solid materials; alumina, silica, ceramic beads, sulfated limestone,
calcined amorphous limestone and calcined crystalline limestone. Approximately 10 cm1 of each
solid material was placed in turn in the reactor and a mixture of nitrous oxide (200 pprn) in helium
was passed through the reactor. The concentration of nitrous oxide at the reactor outlet was
measured to determine the extent of N20 decomposition. As a basis for comparison, the
homogeneous phase decomposition of nitrous oxide in the reactor was also studied.
Results showed that a significant amount of NaO decomposed even in the absence of any solid
material in the reactor. It was observed that the presence of solid materials in the reactor enhanced
the decomposition of nitrous oxide and that the degree of enhancement was dependent on the solid
material being tested; calcined limestone, for example, was seen to be highly effective in
decomposing nitrous oxide while ceramic beads showed little or no effect.
6B-1
-------
INTRODUCTION
There is growing concern over the increasing concentration of atmospheric nitrous oxide. This
concern stems from the realization that nitrous oxide contributes to the greenhouse effect and to
the depletion of the ozone layer. The mean concentration of N20 in the atmosphere is 330 ppbv
and it is estimated that it is increasing at a rate of Q.2% per year fll.
It has been suggested that fossil fuel combustion is a major contributor to the atmospheric nitrous
oxide inventory. Measurements £21 show that nitrous oxide emissions from circulating fluidized
bed combustors (CFBCs) range from 20 to 120 ppm. Based on these emission values, it is doubtful
that nitrous oxide emissions from fluidized beds contribute more than a minor fraction to the global
inventory. Nonetheless, in accordance with its dedication to developing an environmentally safe
product, Ahlstrom Pyropower Inc. has instituted a project directed at the reduction of nitrous oxide
emissions from AHLSTROM PYROFLOW® boilers. The project involves, in part, an investigation into
the formation and destruction of nitrous oxide in circulating fluidized bed combustors.
Knowledge of the principal reactions involved in the formation and destruction of nitrous oxide in
fluidized bed environments is limited at best. In order to minimize nitrous oxide emissions it is
necessary that:
1, reactions that play a dominant role in the formation and destruction of nitrous oxide be
identified and that
2. the effect of process parameters on the kinetics of these reactions be studied in detail.
Studies f3.41 indicate that hydrogen cyanide (HCN), released during the devolatilization of coal, is
a major precursor of nitrous oxide. It is believed that HCN undergoes oxidation to NCO which in
turn reacts with nitric oxide (NO) to form nitrous oxide (NjO), There is relatively little debate
about the importance of this reaction path as a means of formation of nitrous oxide. Doubts about
it being the only major nitrous oxide formation path have however been expressed. De Soete £5J
and Axnand and Andersen £61 have reported the formation of nitrous oxide by the reduction of NO
on char surfaces. De Soete £51 has also reported that nitrous oxide may be formed by the oxidation
of char nitrogen (1-5%) during combustion.
Nitrous oxide destruction in the fluidi2ed bed environment may occur through both homogeneous
and heterogeneous phase reactions. Kramlieh et al. £41 and Ernola et al. £31 have suggested that
the principal nitrous oxide destruction reaction is its homogeneous phase reduction to nitrogen by
hydrogen radicals. Relatively very little is known about the heterogeneous phase destruction of
6B-2
-------
nitrous oxide. It is believed £5} that nitrous oxide reduction on char is one of the major
heterogeneous N,0 destruction pathways. Little or no information currently exists on the interaction
of nitrous oxide with solids, other than char, present in a fluidized bed environment.
This paper describes a study directed at investigating the heterogeneous decomposition of nitrous
oxide in the operating temperature range of a CFBC. An electrically heated tubular quartz reactor
(2.3 cm l.D.) was used to study the decomposition of nitrous oxide on six different solid materials;
alumina, silica, ceramic beads, sulfated limestone, calcined amorphous limestone and calcined
crystalline limestone. Approximately 10 cc of each solid material was placed in turn in the reactor
and a mixture of nitrous oxide (200 ppm) in helium was passed through the reactor. The
concentration of nitrous oxide at the reactor outlet was measured to determine the extent of N20
decomposition. As a basis for comparison, the homogeneous phase decomposition of nitrous oxide
in the reactor was also studied.
EXPERIMENTAL SET-UP
The experimental set-up used in the course of this study (Fig, 1) consists of three major components:
1. An electrically heated quartz tube that serves as a reactor.
2. Mass flow controllers used to deliver a measured amount of a nitrous-oxide/helium
mixture to the reactor.
3. A HORIBA Non-Dispersive Infrared nitrous oxide analyzer.
Reactor
The reactor, for the major part, is a 91.5 cm long, 2.3 cm l.D. quartz glass tube. Caps at the end
of the tube house ports for the inlet and the outlet of reactant and product gas mixtures. The end
caps also house inlet ports for thermocouples used in measuring and controlling the reactor
temperature. A sintered quartz glass filter is provided 50.8 cm from one end of the tube and serves
to support a bed of the solid material being tested. The reactor is heated by a three zone, 61 cm
long electric furnace. The two outermost zones of the furnace are 15.25 cm long and the central
zone is 30.5 cm in length. Each furnace zone is independent of the others in its temperature control.
Thermocouples inside the reactor serve as sensors for controllers that control the temperature of each
furnace zone.
Mas 5 Plow Controllers
Two mass flow controller modules, one for a nitrous-oxide/helium mixture (0.4% N20) and the
6B-3
-------
other for pure helium were used in the course of this study. Using these controllers it was possible
to feed mixtures of nitrous oxide in helium at predetermined concentrations and flow rates to the
reactor. It may be mentioned here that helium was chosen as a "balance gas' due to its chemical
inertness and its high thermal conductivity. The high thermal conductivity was necessary to minimize
radial temperature gradients and the heat up zone within the reactor.
Nitrous Oxide Analyzer
A HORIBA Non Dispersive Infrared N20 analyzer was used to measure the concentration of nitrous
oxide in the inlet and outlet gas streams of the reactor. The analyzer was equipped with a 7.8 jan
wavelength filter.
EXPERIMENTAL PROCEDURE
Homogeneous Phase Decomposition Study
The reactor was heated to the desired temperature and a 200 ppm mixture of NzO in helium was
fed to the reactor at three different flow rates (500, 1000 and 1500 cmVmin). At each condition,
the concentration of nitrous oxide at the outlet of the reactor was measured to determine the extent
of nitrous oxide decomposition. This procedure was repeated for six reactor temperatures; 650, 700,
750, 800, 850 and 900°C. The results obtained are presented in the following section.
Heterogeneous Phase Decomposition
Approximately 10 cm3 of the material being tested (250um>mean particle diameter>125jtm) was
placed in the reactor and the reactor was heated to 850*C. A 200 ppm mixture of N20 in helium
was fed to the reactor at a flow rate of 500, 1000 and 1500 cmVmin. At each condition, the
concentration of nitrous oxide at the outlet of the reactor was measured to determine the extent
of nitrous oxide decomposition. A comparison between the results obtained for each solid material
tested is presented in the following section,
RESULTS
Results of the homogeneous phase nitrous oxide decomposition study are shown in Table 1. As
may be seen from the data, nitrous oxide does not decompose to any significant extent below 700°C.
It is also evident that the rate of nitrous oxide decomposition increases with reactor temperature and
residence time. It is most likely that the products of the nitrous oxide decomposition were nitrogen
and oxygen; no nitric oxide (NO) was detected in the outlet stream from the reactor.
Reaction rate constants for the homogeneous phase decomposition of nitrous oxide were calculated
6B-4
-------
from the obtained data. It was assumed, in the calculation, that the decomposition of NaO occurs
via a first order reaction. The reaction rate constant, k, is presented as a function of temperature
in Table 2. Fig. 2 is a plot of -ln(k) versus 1/T. As may be seen, the plot is a
straight line. This indicates that the assumption that nitrous oxide undergoes a first order
decomposition reaction was correct. The rate of homogeneous phase nitrous oxide decomposition
may thus be written as;
d[NsOj/dt = -k [NjO]
where, [N20] is the nitrous oxide concentration at time t. The reaction rate constant, k, a function
of temperature, may be expressed as:
k = kcexpt-E/RT]
The value of the activation energy, E, derived from the slope of the plot (E/R) is 246.6 kJ/mol.
The frequency factor, ko, may be derived from the y-intercept of the plot, -lnCkJ, and is equal to
2.813 x 10'° sec"'.
The results of the heterogeneous phase N20 decomposition studies are shown in Table 3. Also
included in the table, for the purpose of comparison, are the results from the corresponding empty
tube (homogeneous phase) experiments. The results show the fraction of nitrous oxide that
decomposes on passage through the reactor. The residence times entered at the top of the table
are the residence times of the gas mixture in the entire heated length of the reactor. The numbers
(1) and (2) are used to distinguish between the two types of limestones used; respectively, the
calcined crystalline limestone and the calcined amorphous limestone. The variation of nitrous oxide
decomposition with total reactor residence time, is shown, for each solid material and the empty tube
experiment, in Fig. 3.
As may be seen from the results, the presence of ceramic beads or sulfated limestone in the reactor
does not significantly affect the decomposition of nitrous oxide. The presence of silica sand or
alumina enhances the decomposition of nitrous oxide to a small extent. The most dramatic results
are those obtained in the presence of calcined limestone. It may be seen that nitrous oxide
decomposes completely in the presence of the calcined crystalline limestone at 850"C. As may be
seen from the graphical results, the conversion in the presence of calcined limestone is dependent
on the kind of limestone used. There is an almost 50% difference in the conversions for the two
types of limestones used at a reactor residence times of 3.2 sec. As in the case of the homogeneous
6B-5
-------
phase decomposition studies, no NO was detected in the outlet stream from the reactor.
DISCUSSION AND CONCLUSIONS
Based on the observations and results described in this paper, the following conclusions may be
drawn.
1. The homogeneous phase thermal decomposition of nitrous oxide is a very important pathway for
nitrous oxide destruction in a fluidized bed combustor. A simple calculation shows that under
typical operating conditions in a circulating fluidized bed, that is, a gas residence time of 6 .
seconds at an average furnace temperature of 870° C, over 60 percent of the nitrous oxide present
at the bottom of the combustor would be destroyed before it reached the combustor exit.
Furthermore, .if the average operating temperature of the unit were to be increased by 10°C, the
extent of N20 decomposition would be increased to 70%. It has been seen in measurements on
commercial scale CFBCs |2J that the nitrous oxide emission level does in fact decrease
significantly with increasing bed temperature. It must be realized, of course, that the rate of
nitrous oxide formation is also temperature dependent.
2. Of the solid materials tested, calcined limestone was seen, to decompose nitrous oxide most
efficiently. Alumina and silica sand were seen to slightly enhance the decomposition of nitrous
oxide and ceramic beads and sulfated limestone were seen to have virtually no effect on the
extent of nitrous oxide decomposition. One would expect, in the light of these observations, to
see a dramatic decrease in nitrous oxide emissions with increasing feed Ca/S ratio in a CFBC.
This, however, has not been the case. Studies on a 0.8 MWm Ahlstront Pyroflow pilot plant T21
show only a slight reduction in nitrous oxide emissions with increasing feed Ca/S ratio; no
definite relationship between nitrous oxide emissions and feed Ca/S ratio could be detected for
a similar study £21 carried out on commercial scale units.
3. The efficacy of calcined limestone in decomposing nitrous oxide was dependent on the type of
limestone used. Calcined crystalline limestone was seen to decompose nitrous oxide more
effectively than was calcined amorphous limestone. At a reactor residence time of 3.2 seconds,
the calcined crystalline limestone was seen to completely decompose the nitrous oxide, where, the
calcined amorphous limestone was seen to decompose only 50% of the inlet nitrous oxide.
ACKNOWLEDGEMENT
The authors gratefully acknowledge partial funding of the described study by the Finnish Ministry
of Trade and Industry through the L1EKKI program.
6B-6
-------
REFERENCES
1. R.F. Weiss, Journal of Geophysical Research, vol. 86, 1981, p. 7185.
2. M. Hiltunen, P. Kilpinen, M. Hupa and Y.Y. Lee, "N20 Emissions from CFB Boilers: Experimental
Results and Chemical Interpretation." To be presented at the IP Int. Conf. on Fluidized Bed
Combustion. Montreal, 21-24 April, 1991.
3. P. Emola & M. Hupa, "Kinetic Modelling of Homogeneous NaO Formation and Destruction in
Fluidized Bed Conditions." Proceedings of the Joint Meeting of the British and French Sections
of the Combustion Institute. Rouen, 1989, p. 21.
4. J.C. Kramlich, J.A. Cole, J.M. McCarthy, W.S. Lanier & J.A. McSorley, "Mechanisms of Nitrous
Oxide Formation in Coal Flames." Combustion and Flame. 1989, vol. 77, p. 375.
5. G.G. De Soete, "Heterogeneous NO and N20 Formation from Bound Nitrogen during Char
Combustion." Proceedings of the Joint Meeting of the British and French Sections of the
Combustion Institute. The Combustion Institute, 18-21 April, Rouen, 1989, p. 9.
6. L.E. Amand & S. Andersen, "Emissions of Nitrous Oxide (N20) from Fluidized Bed Boilers."
Proceedings of the 1989 International Conference on Fluidized Bed Combustion, vol. 1, pp. 49-
56.
6B-7
-------
Figure 1. Experimental Setup for Quartz Tube Reactor Studies
Figure 2. —In(k) Vs. 1/T
6B-8
-------
1.0
0,9
0.8
0.7
0.6
0.5
0,4
0.3'
0.2,
¦v Empty Tube
o.t
o — o Alumina
o — o Ceramic Beads
¦ — ¦ Silica Send
6 — a Sulfated Limestone
~ — ~ Calcined Limestone (1)
• — • Calcined Limestone (2)
0.0
Reactor Residence Time (sec)
10
Figure 3. Fractional N^O Decomposition vs. Reactor Residence Time
6B-9
-------
Table 1
HOMOGENEOUS PHASE DECOMPOSITION OP NITROUS OXIDE
Reactor Pressure 3 psig Inlet N20 Concentration : 200 ppm
Temperature
Residence' Time
»zO Outlet concentration
( C)
(sec)
11.7
200
650
5.8
200
3.9
200
11.1
197
700
5.5
200
3.7
200
10.5
18 5
750
5.3
200
3.5
200
10.0
150
800
5.0
174
3.4
182
9.6
78
850
4.8
125
3.2
148
9,2
12
900
4.6
52
3.1
81
Table 2
HOMOGENEOUS PHASE NjO DECOMPOSITION REACTION RATE
CONSTANT VS. TEMPERATURE
Temperature
<*e)
* 1
(sec )
700
0.001350
750
0.007399
800
0.028640
850
0.097444
900
0.301750
6B-10
-------
Table 3
FRACTIONAL N20 DECOMPOSITION VS. TOTAL REACTOR RESIDENCE TIME
Reactor Temperature : 850°C Reactor Pressure : 3 psig
Material
t—9.€b
t=4.8s
t=3.2s
Alumina
0.65
0.41
0.29
Ceramic Beads
0.61
0.38
0.26
Silica Sand
0.63
0.40
0. 28
Sulfated Limestone
0.61
0.38
0. 26
Calcined Limestone (1)
1.00
1.00
0.98
Calcined Limestone ('2)
0.92
0.65
0.50
Empty Tube
0.61
0.38
0. 26
6B-11
-------
-------
NOx CONTROL IN A SLAGGING COMBUSTOR FOR A
DIRECT COAL-FIRED UTILITY GAS TURBINE
P. J. Loftus and R. C. Diehl
Energy Technology Office/Textron Defense Systems
(Formerly AVCO Research Laboratory)
2 385 Revere Beach Parkway
Everett, MA 02149
and
R. L. Bannister and P. W. Pillsbury
Westinghouse Electric Corp.
The Quadrangle, 4400 Alafaya Trail
Orlando, FL 32826-2399
6B-13
V...
-------
-------
ABSTRACT
A modular combustion concept, which emphasizes controlled coal
thermochemistry, has been developed for application in direct coal
firing of utility gas turbines. The approach under investigation ,is
based on a multi-stage, slagging combustor, which incorporates N0X,
S0K and particulate emissions control. This approach allows raw
utility grade coal to be burned, thereby maintaining a low fuel cost.
The- cost of electricity from combined cycle plants incorporating a
direct coal-fired gas turbine is expected to be significantly lower
than that from conventional pulverized coal steam plants.
The first stage, the primary combustion zone, is operated fuel-
rich to inhibit NOx formation from fuel-bound nitrogen and has a jet-
driven, toroidal vortex flow field, which provides for efficient,
stable and rapid combustion at high heat release rates.. Impact
separation of molten mineral matter is accomplished in the second
stage, which is closely integrated with the primary zone. The second
stage may also include a slagging cyclone separator for additional
slag removal. This is a novel application for a cyclone separator.
Final oxidation of the fuel-rich gases and dilution to achieve the
desired turbine inlet conditions are accomplished in the third stage.
6B-15
-------
Rapid quenching and good mixing with the secondary air are employed to
avoid thermal" NO* formation by minimizing peak flame temperatures and
residence times in the third stage.
The combustor concept has been extensively tested at a thermal
input of 3.5 MWt (12 MM Btu/hr) and, a pressure of 6 atmospheres. Both
pulverized coal and coal-water mixtures have been successfully fired.
The combustor has demonstrated stable and intense combustion, with
excellent carbon conversion, efficient slag capture, retention of most
of the coal alkali in the slag and low pressure and heat losses. The,
staged combustion N0X control strategy has proved very effective: N0„
emissions are approximately one fifth,of the New Source Performance
Standards requirements.
BACKGROUND 5
The authors' companies are working under Department of Energy
sponsorship to develop the technology base for direct coal-firing of
utility gas turbines. The approach under investigation is based on a
multi-stage, slagging combustor, which incorporates N0X, SOK and
particulate emissions control. This approach allows raw utility grade
coal to be burned, thereby maintaining a low fuel cost. The cost of
electricity from combined cycle plants incorporating a direct coal-
fired gas turbine is expected to be significantly lower than that from
conventional pulverized coal steam plants with flue gas
desulfurization
-------
re-circulation which leads to extremely intense and very stable
combustion of a wide variety of fuels. The toroidal vortex design
gives very high volumetric heat release rates for coal combustion (up
to 40 . These heat release rates are some three to four times
that for cyclone-type combustors, leading to smaller combustor sizes
and lower wall heat losses. Fuel-rich conditions in the primary zone
inhibit NOx formation from fuel-bound nitrogen. Extensive use was
made of 3-D combustion modelling'techniques in the. preliminary design
of the combustor (Chatwani and Turan, 1988, Loftus et al., 1988).
The toroidal vortex provides the mechanism for flame
stabilization and also for inertial separation of larger ash/slag
particles. Partial separation of mineral matter and char at the top-
ox the toroidal vortex zone results in initiation, of wall slagging
there, with continued deposition and flow over all exposed wall
surfaces. In order for successful slag deposition in the dome region,
enough coal particle residence time and combustion product re-
entrainment must be provided to ensure rapid coal particle burnout,
resulting in molten, free mineral matter. Larger, partially
devolat11ized coal particles will continue to burn, either in
suspension or in the wall slag layer. The primary zone was designed
for approximately 9 0 percent suspension burning and 10 percent wall
burning. The primary zone particle residence time is of order 100
msec for a 75 micron diameter particle. The.primary zone slag layer,
provides thermal and erosion protection- for the combustor walls, in
addition to a mechanism for oxidation of deposited char. The slag
layer formed from this, portion of the mineral matter eventually
reaches the impact separator, where it is collected in the slag
bucket.
The major fraction of mineral removal from the gas is obtained in
the second stage impact separator which is at the exit from the
primary zone. The separation of combustion and slag removal duties
between the two stages has two- substantial benefits. First, it
results in maximum removal of carbon free slag: at the primary zone
exit there is a very high carbon conversion fraction—essentially all
the-coal char has been oxidized, leaving free mineral matter behind.
Second, due to the low density of the combustion products, a simple
impact or allows high efficiency separation of fine mineral matter
particles at low cost in pressure drop-. Overall, the air, pressure
drop is optimally distributed, first for combustion stabilization and
second for mineral matter removal.
Pulverized limestone sorbent is used for control of sulfur
emissions. The sulfur control technique used is based on related ETO
work on super-equilibrium sulfur capture (Abichandani et al., 1989).
The sorbent is injected into the primary zone combustion products,
which generally contain .a mixture of S02, H2S and COS, at the exit of
the primary zone, just upstream of the exit nozzle. Reacted sorbent
is removed along with the coal ash in the second stage impact
separator.
Final oxidation of the fuel-rich gases and dilution to achieve
the desired turbine inlet conditions are accomplished in the third
stage. Rapid quenching and good mixing with the secondary air are
employed, to avoid thermal NO, formation by minimizing peak flame
temperatures and residence times in the third stage.
6B-17
-------
N0X CONTROL APPROACH
Emissions of nitrogen oxides in the products of combustion are
controlled by adopting the following approach:
* Sub-stoichiometric .(fuel-rich) combustion conditions in the
first combustor stage.
.• Effective control of the gas temperature and stoichiometry
histories during final oxidation/dilution in the third
combustor stage.
The main source for formation of nitric compounds- in the
combustion of coal.is fuel-bound nitrogen. Part of the fuel nitrogen
is released with the volatiles in the early stages of combustion and
the remainder is retained by the char residue and released during
subsequent char oxidation. Nitric oxide can be produced by the
oxidation of the nitrogen in the volatiles or in the char. N0X
formation from fuel bound nitrogen is very sensitive to the combustion
stoichiometry. It is known that HCN and NK3 are formed in the gas
upon evolution of coal nitrogen. These can subsequently be oxidized
to NO or can react with NO to form harmless molecular nitrogen,
depending upon the availability of oxidants in the gas. Fuel-rich'
operation promotes the formation of molecular N2 as the end product of
the fuel nitrogen, whereas fuel lean operation, with the availability
of oxidants, results in NO formation.
Volatile nitrogen is defined as that which is produced from the
volatile coal fractions and reacts in the gas to form N2/ NO, HCN or
NH3. Char nitrogen is-that which is associated with a solid, either
as a pyroly-sis product of tars or as the original coal char. The
distribution of nitrogenous species, between volatile nitrogen and char
nitrogen is critically dependent on the coal particle heating rate,
the peak temperature, the residence time at high temperature and the
nitrogen distribution within the coal (Smart and Weber, 1989) . Fuel-
bound nitrogen is the major source of N0X in conventional PC
combustion, typically accounting for more than 80 percent of total N0X
emissions (Pershing and Wer.dt, 1979) .
For staged combustion to be effective, it is important to avoid
the carry over of either volatile or char nitrogen to the final
oxidation zone, where these"can be converted to NO. The intense and
rapid mixing produced by the toroidal vortex design leads to rapid de-
volatilization of the coal, homogeneous combustion conditions and
efficient oxidation of the char to a fuel-rich gas in the first stage.
These conditions favor conversion of fuel bound nitrogen to molecular
nitrogen ana minimize the possible carry-over of volatile or char ¦
nitrogen to the third combustor stage.
For the case of PC combustion, the calculated equilibrium
concentrations of nitrogen oxides in the combustion gas are shown in.
Figure 2 for various primary zone fuel air equivalence ratios and
temperatures. This plot includes NH3 and HCN, which have been
converted to total NO„ and included in the concentrations shown. (The
contributions from these species are typically small.) NO„
concent rations at the adiabatic flame temperature and at 100 K and 200
K below the adiabatic flame temperature are shown. The NOx
concentrations in the gas corresponding to the NSPS limits for
6B-18
-------
bituminous coal (0.6 lbs per MM Btu) and: sub-bituminous coal (0.5 lbs
per MM Btu) are also shown as a function of fuel-air equivalence
ratio. The strong temperature dependence of NOx is clearly seen: a
temperature drop of 200 K typically reduces the equilibrium NOx by a
factor of three or' more. The equilibrium NOx concentration in the gas
becomes less than the NSPS standard at fuel-air equivalence ratios-
greater than about 1.2 and at the primary zone nominal operating, point
(equivalence ratio in the range 1.3 to 1.4) the equilibrium- NOx in the
primary zone' is more than 'a factor of ten less than the NSPS
requirement.
The control of stoichiometry and temperature in the'third
combustion stage is key to minimizing- the formation of thermal NOx.
The formation of thermal NO„ is governed by the highly temperature
dependent reactions between nitrogen and oxygen, the Zeldovich chain
reactions. The rate of formation is significant only at temperatures
above approximately 1900 K (3Q0Q°F) and increases with increasing
oxygen concentration. Thus temperatures in the final oxidation zone
should be maintained below this value to avoid thermal NOx formation.
The secondary air for final,combustion in the' last combustor stage is
added in such a manner that the gas is rapidly quenched and maintained
at a temperature below .which thermal NOx can form, while final
oxidation of the unburned species in the gas is completed. As soon as
the final oxidation is complete, the dilution air is introduced, again
with rapid and complete mixing, in order to quench all further NQX
generation.
Kinetic calculations were performed to determine the desired
temperature and operating conditions during final oxidation and the
appropriate split between quench/final oxidation air and dilution air.
These calculations showed that - an adiabatic flame temperature of about
1800 K is reached for a fuel air equivalence ratio of 0.6 in the
intermediate zone and that the final oxidation of the gas is completed
within a few milliseconds, see Figure 3. At these conditions thermal
NO„ formation is insignificant and the predicted final NO,
concentration in the gas will be only a small fraction of the"NSPS
limit. It is important to obtain effective mixing of the secondary
combustion air with the hot fuel-rich primary gases.
Three-dimensional aero-thermal calculations and analysis of the
mixing process in both the intermediate and dilution zones of the
third combustor stage were conducted. The number, size and
orientation of the intermediate -and dilution zone jets were varied to
arrive at the optimum mixing performance, expressed as a minimum exit
temperature pattern factor. The final design analysis involved
extending the three-dimensional aero-thermal flow modelling of the
third stage to the full reacting flow field. However, no attempt was
made to optimize the lean-burn combustor from the point of view of NO„
control. The principal purpose of the experimental work was to tackle
the major technical issues in this development effort, which are
related to Obtaining efficient primary zone and slag separator
performance.
TEST ARRANGEMENT AND COMBUSTOR OPERATION
The combustor concept has been extensively tested at a thermal
input of 12 MM Btu/hr (3.5 MWt) and a pressure of 6 atmospheres.
Tests have been conducted with both pulverized coal (PC) and coal-
6B-19
-------
water mixture (CWM) fuels. A photograph of the subscale slagging
combustor test arrangement as currently installed at ETO's Haverhill•
test site, is shown in Figure 4. The nominal test conditions for the
subscale combustor are as listed in Table 1. An oil fired air pre-
heat er is used to heat the combustion air in order to correctly
simulate the gas'turbine compressor discharge conditions. A
downstream sonic orifice is used to control the combustor chamber
pressure. After pressure let-down, the combustor exit gases are water
quenched before being led to an exhaust stack. The subscale combustor
is water cooled, the cooling water being re-circulated via a cooling
tower. All fuel-rich zone components are lined with a high alumina
refractory. This is both to reduce heat losses in this small scale
experimental combustor and to promote slagging during the relatively
short, tests.
Start-up and operation of the system proved to be simple and
reliable. After establishing the correct air flow rates_through the
system and allowing the air pre-heater to come up to design
temperature, a methane/oxygen torch in the primary zone is ignited.
The torch is used to ignite a fuel oil flame and is then extinguished.
Fuel oil is then burned for approximately 15 minutes, in order to pre-
heat the refractory liner. The oil is injected via two spray nozzles
in the primary zone. After the refractory liner has reached operating
temperature, the coal (PC or CWM) flow is started, and the fuel oil
flow is stopped. In PC testing, a pneumatic conveying system is .used
to feed coal to the primary zone. For CWM' testing, a Moyno
progressing cavity pump is used to supply CWM to the combustor. The
CWM atomizers are Parker-Hannifin air-assist atomizing nozzles.
Atomizing air for CWM tests is- supplied from a high pressure tube
trailer via a heat exchanger. The heat exchanger warms the expanded
high pressure air back up to approximately ambient temperature.
A detailed fuel specification for the proposed application was
prepared by AMAX Extractive Research and Development. Choice of coal
(and consequently of mineral matter composition), coal particle size
and- CWM composition affects certain primary design constraints for the
slagging combustor. These include liquid slag formation, combustion
efficiency, downstream corrosion, erosion and deposition- and pollutant
generation. The primary.zone of the combustor was designed to operate
at highly fuel-rich (i.e.-low flame temperature) conditions. The
flame temperature is obviously even lower for CWM fuels. Consequently
a low ash fusion temperature coal was desirable. The ratio of ash to
sulfur content is of interest: the higher the coal sulfur content, the
higher the ratio of limestone sorbent to ash in the slag to be
separated and the greater the effect of sorbent injection on slag
properties. The preferred -coal fuels were determined to be high
volatile eastern bituminous coals.. These coals have the advantages' of
a high heating value, leading to favorable,combustion characteristics
with high flame temperatures and rapid combustion. They also tend' to
have low to medium sulfur contents and soluble alkali contents below
0.05 percent,. From this general specification, several specific seams
were identified for use in the subscale combustor testing. These
included a low and a high sulfur eastern bituminous coal and a low
sulfur western sub-bituminous coal. Detailed coal specifications are
given in Table 2. . The CWM fuels tested were prepared" from close to
standard grind (95 percent through 200 mesh)- coals.
A full test program was conducted with PC feed before switching
6B-20
-------
TABLE 1
SUBSCALE SLAGGING COMBUSTOR NOMINAL TEST CONDITIONS
Coal Thermal Input
3.5 MVV, (12 MM Btu/hr)
Coal Feed
95% < 200 mesh PC
, 95% < 200 mesh, 60% solids CWM
Atomizing Air/CWM Mass Flux Ratio
1.0
Oxidizer
620 K (650°F) pre-heated air
Primary Zone Equivalence Ratio
1.3 to 1.4
Total Mass Flow Rate-
¦ 3.2 kg/s (7 lb/s)
Exit temperature
1300 K (1850°F)
Pressure
6 atm
Sorbent
-325 mesh limestone
Sorbent Molar Ratio
Ca/S = 2
over to CWM feed. From the outset of combustor testing, a stable,
flowing slag layer was formed on the primary zone dome and walls.
Some dissolution of the refractory layer was observed in the early
runs, but after a few hours of operation an equilibrium insulation
layer of slag and refractory was formed. Equilibrium,slag layer
thicknesses in the primary zone, where heat fluxes are high, are on
the order of 1 mm. The corresponding thickness in the slag separator
is on the order of 3 mm. No obstruction or fouling of any of the
primary zone coal/air injectors or of the relatively small diameter
primary zone exit nozzle with slag was observed. The impact separator
worked as planned, and a flowing slag layer was observed on the top
and sides of the impactor centerbody and on the slag bucket walls.
TEST RESULTS
A full series of tests with PC fuels demonstrated that the
combustor primary zone produces excellent carbon conversion
performance, see Figure 5. At the nominal primary zone operating
point (fuel/air equivalence ratio of 1.3 to 1.4) the carbon conversion
for PC firing is better than 99 percent. For PC firing, better than
98 percent carbon burnout in the primary zone was obtained for
fuel/air equivalence ratios as high as 1.6. In order to obtain good
carbon conversion performance on CWM fuels, it was necessary to
increase the primary zone aspect ratio (length/diameter). .For PC
firing the aspect ratio of the primary 2one was- 1.25
-------
TABLE 2
ANALYSES FOR COALS TESTED TO DATE
Coal Analysis As
Received
Dorchester,
VA
Pittsburgh #8
Hanna Seam,
WY
% Moisture
1.00
1.49
9.09
% Ash
6.24
7.59
5.37
% Volatile Matter
33.30
38.28
38.33
% Fixed Carbon
59.46
.-52.64
47.21
% Sulfur
0.96
2.35
0.57
% Chlorine
0.04
0.14
0.05
% Carbon
80.43
76.73
67.W
% Hydrogen
4.79
5.21
5.06
% Nitrogen
1.72
1.34
1.44
% Oxygen
' 4.82
5.15
11,33
MJ/kg (Btu/lb)
32.95 (14,234)
32.00 (13,822)
27.28 (11,784)
for PC, but this is to be expected, given the lower heating value and
f x sine temperatur©s of CWM fuels. Measured f 1 ante tenio©r 31 ur e s in the
dome region of the primary zone for PC firing are shown in Figure 6.'
The primary zone temperatures at the nominal primary zone operating'
point are 210-0 to 2000 K (3320 to 3l40°F) for PC firing and some 150
to 200 K (270 to 360°F) lower than this for CWM firing.
Figure 7 shows measured primary zone N0K concentrations .for
pulverized coal firing. These measurements were made at the exit from
the primary zone, just upstream of the main exit nozzle. The
measurements are compared both with' calculated equilibrium NO, levels
for PC firing and also the NSPS limits, as described above. The limit
of resolution of the cnemiluminescent analyzer used in making these
measurements is of order 10 ppm. The measured NO* concentrations, are
well below the NSPS limits and are in general agreement with the
calculated equilibrium concentrations at 100 to 200 K below the
adiabatic flame temperature. The measured flame temperatures, shown
in Figure 6,-are typically 150 to- 200 K below the adiabatic flame
temperature.
Corresponding primary zone results and equilibrium calculations'
for the case of 60 percent solids CWM firing are shown in Figure 8.
The results for CWM firing are substantially different from those for
PC firing. -While the calculated equilibrium NOK concentrations for
CWM firing are lower than those for PC firing, because of the lower
flame temperatures, the measured NO, concentrations at the primary
zone exit for CWM firing are considerably higher than'those for PC
firing. The CWM measurements are also considerably higher than the
6B-22
-------
calculated equilibrium- concentrations for CWM1 firing. . •
This increase - in NO* concentrations for CWM firing is also
reflected in the lean zone exit, or exhaust emissions, measurements'.
Measured N0X exhaust emissions, corrected1 to 15 percent oxygen, for
three PC fuels and for 60 percent solids Dorchester CWM are plotted as
a function of primary zone fuel-air equivalence ratio in Figure 9.
These measurements- were made at the combustor exit, downstream of the
lean-burn zone. The overall combustor fuel-air equivalence ratio at
the lean zone exit was fixed as the primary zone equivalence ratio was
varied. The dramatic reduction of N0X levels with increased staging
of the combustion is clearly illustrated. The NSPS limit (0.6 lb/MM
Btu for bituminous coals), scaled for the combustor exit conditions,
is shown for reference. At the nominal design operating point, the
combustor NOx emissions for both PC and CWM firing are well below the
NSPS limit. Not enough information is available to partition the
exhaust NOx emissions between contributions from (1) primary zone NO
generation; (2) lean-burn zone oxidation of volatile or char nitrogen
carried over from the fuel-rich zone; and (3) generation of thermal
NOx in the lean-burn zone. It is obvious, however, that NOx is
generated in the lean-burn zone. For example, NOx levels at the rich-
zone combustor exit at equivalence ratios an the range 1.3 to 1.4 (the
nominal primary zone operating point) for PC firing have been measured
at 20 to 40 ppm. The primary zone typically has one third of the
total gas mass flow rate. If no NO„ was generated in the lean-burn
zone, the primary zone NO, would therefore be diluted by a factor of
approximately three, giving emissions on the- order of 10 to 15 ppm.
The actual emissions at this condition are of order 30 to 50 ppm-.
Thus some 20 to 4 0 ppm N0„ are generated in the lean-bu-rn zone. These-
20 to 40 ppm are either from thermal NOx in the lean-burn zone or from
lean zone oxidation of char of volatile nitrogen carried over from the
primary zone.
The exhaust NO, emissions for CWM firing are slightly higher than
those for PC firing. At the nominal primary zone operating point, the
CWM- emissions are in the range 60 to 80 ppm, compared with 30 to 50
ppm for PC f-iring. While the precise mechanisms leading to the higher
levels of NO* with CWM firing are not clear at present, several
contributing factors may be identified. As discussed above, the
measured NOx levels at the primary zone exit for CWM firing are much
higher than those measured at the same location for PC firing. In
fact, for CWM firing the measured NOx is in excess of the
thermodynamic equilibrium level. Thus NO- destruction would be
expected downstream of the primary zone exit. This indeed appears to
be the case: if the assumption of no NO„ generation in the lean-burn
zone is again made, and the primary zone NO, concentration is diluted
by a factor of three, the NO„ concentration "so obtained is of order
160 ppn, considerably in excess of the measured NOx emission for CWM
firing of 60 to 80- ppm. This suggests that NO is destroyed between
the primary zone exit and the lean zone inlet.
The large differences in primary zone NOx between PC and CWM
firing are' indicative of significant differences in temperature,
heating rate and stoichiometry histories in the fuel-rich primary zone
for the two fuels. As discussed above, the partition of the fuel-
bound nitrogen between volatile and char nitrogen and the subsequent
conversion of NO, precursors to molecular nitrogen are strongly
dependent, on such parameters. Because' of its high moisture content
6B-23
-------
and large size, a CWM droplet will experience both a lower heating
rate and'a lower final temperature than a pulverized coal particle.
This may lead to both less complete evolution of fuel-bound nitrogen
and also less efficient conversion of released fuel-bound nitrogen to
molecular nitrogen and consequently to higher N0X emissions.
The post-run appearance of the slag layer in the primary zone
would also indicate that more wall burning occurs for CWM firing than
for PC firing, possibly because of the production of relatively large
coal particle agglomerates on evaporation of the moisture in the CWM
droplet. These larger coal particles will be inertially separated
from the toroidal vortex onto the slagged wall before burning out
completely. Thus the gas phase stoichiometry for CWM burning is
leaner than the global stoichiometry based on air and fuel inputs.
NOx levels at the exit of the primary zone may therefore reflect the
equilibrium levels at .leaner conditions, and given enough residence
time, would eventually be reduced to levels indicative of the global
stoichiometry.
Figure 10 shows the exhaust NOx emissions plotted as a function
of the combustor outlet temperature. The nominal design outlet
temperature is 1850°EV at which temperature the NO* emission is some
4 0' ppm. There is only a moderate increase in NO„ emissions as the
outlet temperature is increased to 2000-°F.
NOx generation and destruction in staged combustion are
controlled by an extremely complex series of 'phenomena. Given the '
limited amount of experimental information available from a practical
staged slagging combustion system such as the one currently being
tested, it is dxffi.cult 'to completely identify the precxse mechanisms
responsible for the results obtained.' However, the general' concept of
stagea combustion for NOx reduction has worked extremely well in this
application, leading to NO* emissions on the order of one fifth of the
NSPS requirements.
CONCLUSION
A three-stage combustion concept has been developed for
application to direct coal-firing of utility gas turbines. A key
aspect of combustor performance is the effective control of NO„
emissions. A subscale combustor (3.5 MWC, 6 atm) is currently being
tested. Results for various coal fuels fired as either PC or CWM have
shown extremely good coal particle burnout leading to effective
slagging in the primary zone. The combustor employs staged combustion
(fuel-rich conditions in the primary zone to inhibit NOx production
from fuel-bound nitrogen; rapid quench/good mixing in lean-burn zone
to reduce peak flame temperature and minimize thermal KOx production)
for NO„ emissions control. For primary zone fuel-air equivalence
ratios greater than approximately 1.1 for PC firing and 1.15 for CWM
firing, the subscale slagging combustor NCX emissions are well below
the NSPS limit.- Given the high levels of fuel-bound nitrogen in the
coals burned (typically 1.3%), the staged combustion has worked
extremely well to control NO„ emissions.
ACKNOWLEDGEMENTS ;
The work described in-this paper is sponsored by the U. S.
Department of Energy through the Morgantown Energy Technology Center
6B-24
-------
under Contract .No. DE-AC21-86MC23167. Mr. Donald W. Ceiling is the
METC Program Manager.
REFERENCES
Abichandani, J. S., Loftus, P. J., Diehl, R. C., Woodroffe, J. A., and
Hclcombe, N. T. (1989) "Nonequilibrium Sulfur Removal from High
Temperature Gases," Proceedings: Sixth Annual Pittsburgh Coal
Conference, Pittsburgh, PA, September, 1989.
Chatwani, A. U., Turan, A., and Stickler, D. B. (1988) "Design and
Sizing of the Primary Stage of a Toroidal Vortex Gas Turbine Combustor
Using a 3-D Flow Field Modelling Cede," Western States Section Meeting
of the Combustion Institute, Salt Lake City, UT, March, 1988.
Loftus, P. J., Chatwani, A. u., Turan, A., and Stickler, D. B, (1988)
"The Use of 3-D Numerical Modelling in the Design of a Gas Turbine
Coal Combustor, " Heat Trans fer in Gss Turbine Sngx ne s and Three—
Dimensional Flows, ASME HTD-Vol. 103, pp. 95-105, edited by E. Elovic,
J. E. O'Brien, and D. W. Pepper, New York. Also presented at ASME
Winter Annual Meeting, Chicago, IL, December, 1988.
Mattsson, A. C. J., and Stankevics, J.,0. A. (1985) "Development of a
Retrofit External Slagging Coal Combustor Concept," Proceedings:
Second Annual Pittsburgh Coal Conference,' Pittsburgh, Pennsylvania.
Pershing, D. W. and Wendt, J. 0. L. (1979) "Relative Contributions of
Volatile Nitrogen and Char Nitrogen to N0„ Emissions from Pulverized
Coal Flames," Industrial and Engineering Chemistry: Process Design and.
Development, 18 (1); 60—66, 1979.
Pillsbury, P. W., Bannister, R. L., Diehl, R. C., and Loftus, P. J.'
(1989) "Direct Coal Firing for Large Combustion Turbines: What Do
Economic Projections and Subscale Combustor Tests Show?" ASME Paper
89-JPGC/GT-4, Joint ASME/IEEE Power Generation Conference, Dallas,
Texas, October, 1989.
Smart,. J. P. and Weber, R. (1989) "Reduction of NOK and Optimization
of Burnout with an Aerodynamically Air-Staged Burner and an Air-Staged
Precombustor Burner," Journal of the• Institute of Energy, December
1989, pp 237-245.
Stankevics, J. 0. A., Mattsson, A. C. J., and Stickler, D. B. (1983)
"Toroidal Flow Pulverized Coal-Fired MHD Combustor," Third Coal
Technology Europe Conference, Amsterdam, The Netherlands.
6B-25
-------
Figure 1 Schematic diagram of three stage slagging combustor
concept including slagging cyclone separator
NOx (ppm)
1 1.1 1.2 1.3 1.4 1.5 . 1.6
Fuel-Air Equivalence Ratio
Figure 2 Calculated thermochemical equilibrium N0X
concentrations in the fuel-rich zone as a function of
fuel-air equivalence ratio for three gas temperatures:
adiabatic flame temperature (AFT), 100 K below AFT, and
200 K below AFT
6B-26
-------
Time (msec)
Figure 3 Variation of species concentrations showing final CO
burnout and NO generation in lean burn zone at a fuel-
air equivalence ratio of 0.6
Figure 4 Photograph of subscale slagging combustor test
arrangement
6B-27
-------
Carbon Conversion (%)
Primary Zone Fuel-Air Equivalence Ratio
Figure 5 Measured, primary zone carbon conversion for PC and CWM
firing as a function1 of fuel-air equivalence ratio
Measured Flame Temperature (K)
Primary Zone Fuel-Air Equivalence Ratio
Figure 6 Measured primary zone flame temperatures for PC firing
as a function of fue1~air equivalence ratio
6B-28
-------
NOx (ppm)
1 1.1 1.2 1.3 1.4 1.5 1.6
Fuel-Air Equivalence Ratio
Figure 7 Measurements of N0X concentrations at exit of primary
zone for PC firing and calculated equilibrium N0X
concentrations for PC combustion as a function of fuel-
air equivalence ratio
NOx (ppm)
Fuel-Air Equivalence Ratio
Figure 8' Measurements of NO* concentrations at exit of primary
zone for CWM firing and calculated equilibrium NOK
concentrations for CWM combustion as a function of
fuel-air equivalence ratio
6B-29
-------
600
500
400 r
300
200
100 r
NOx Eppmv, dry, corrected to 15% 02)
0 PC L/D ¦ 1.26
* CWM L/D • 1.50
0
\ tl ^
'
-
00 ^
X * \
NSPS umi-t
_
.
V s*
^ —%lt
0
1 t
0 w v " 0 .
i i i r
0.8 0.9 1 1.1 1,2 1.3 1.4 1.5 1.6
Primary Zone Fuel-Air Equivalence Ratio
1.7
1.8
Figure 9 Measured lean-zone exit NO, concentrations (dry,
corrected to 15 percent oxygen) for PC and CWM burning
as a function of primary zone fuel-air equivalence
ratio
100
80
60
40
20
NOx (ppmv, dry, corrected to 15% 02}
0 1—
1500
A
~£r
1600 1700 1800 1900 2000
Combustor Outlet Temperature (deg F)
2100
Figure 10 Measured lean-zone exit N0X concentrations (dry,
corrected to 15 percent oxygen) as a function of lean-
zone outlet temperature
6B-30
-------
LOW NQX COAL BURNER
DEVELOPMENT AND APPLICATION
J. W. Allen
I-International Combustion Ltd.
Sinfin Lane,
Derby, Engl and DE2 9GJ
6B-31
-------
-------
ABSTRACT
The paper describes the development arid application of a front wall low N0x coal
burner in the U.K. power industry.
Target NCx emission levels set by European Community Directives, for the U.K.
industry, were met both in full scale single burner thermal trials and in the
multi burner boiler operation.
The paper highlights the basic differences between test rig and boiler
installations, not only in Combustion performance but also in the boiler
operational effects which influence the selection of materials of construction
for the critical burner parts.
In order to optimise the boiler performance, the characteristics of the low N0x
burner must be recognised in the boiler operating procedures.
6B-33
-------
INTRODUCTION
Current UK NO emission targets for large combustion plant (i.e. plant with heat
A
input greater than 50 MW thermal^ are based on a European Economic Community
(EEC) Directive (88/609/EEC) issued in December, 1988 (1). The Directive
stipulates limits for new large plant and also N0x reduction targets to be
achieved by the various EEC countries over the decade to 1998. NO limits for
the various fossil fuels are given in Table 1.
Although these N0x levels refer to new plant they have become target norms for
the retrofitting of power generation boilers in the UK for low N0x operation.
Furthermore the UK is required to reduce NO emission levels by 1554 prior to
1993 and 30% prior to 1998, based on N0X emission levels in 1980.
European units for N0x concentrations are frequently quoted in mg/Nm3, although
most concentration measurements are made in terms of parts per million (ppm).
For comparison purposes, the ppm concentration is referred to either a 3% or 6%
dry waste gas oxygen concentration. Table 2 gives the interconversion factors
for terms commonly used for the expression of N0x concentrations.
Table 1
EMISSION LIMIT VALUES FOR N0X FOR NEW PLANTS
Type of Fuel
Limit Values (mg/Nm3}
Sol id in General
Solid with less than 10& volatiles
Liquid
Gaseous
650
1300
450
350
6B-34
-------
Table 2
INTERCONVERSION OF N0X CONCENTRATION TERMS
To convert
From
To
>
Multiply by
nig/Nm3 ppm
lbs/106 Btu
mg/Nm3
ppm
lbs/106 Btu
2.05
1230
0.487
8.14 x 10"4
1.67 x 10"3
598
Table 2 is based on coal combustion with a dry flue gas 02 content of 6%. To
correct NO concentrations, at differing Oz levels, the following formula can be
A
used
Prior to the privatisation of the electricity industry in the UK the CEGB
announced a £ 170M programme in order to achieve the reductions in N0X emission
levels as required by the EEC Directive. The two major privatised power
generators, National Power and PowerGen, are continuing with this programme.
Progress in the conversion of corner fired units, in the UK has proceeded quickly
following the successful demonstration of the .'Low N0x Corner Firing System
(LNCFS)' installed in a single 500 MW boiler in the CEGB, North Western Region,
in 1985(2),(3). The 500 MW+ corner firing capacity of both National Power and
PowerGen is committed to this low N0X system.
Conversion of the wall fired units has proceeded more slowly, at the time of
writing around 25-301 of the UK wall fired coal capacity has been converted or
committed to low N0X burner retrofit. This slower progress has enabled the
power generating and manufacturing organisations to proceed via a well defined
programme based on isothermal and mathematical modelling, single burner full
scale rig testing and the testing of individual burners within an actual boiler
environment, before commencing a full boiler commercial retrofit. All the low
N0x burner developments, including corner firing, have been based on combustion
staging techniques, which have been demonstrated as capable of achieving the N0X
N0x (ppm at O2j) m 21 - n-
N0x (ppm at O2 )
6B-35
-------
reduction requirements of the EEC- Directive, The burner development and
operations described' in this paper* relate to a front wall low NQx coal burner
incorporating both.fuel and air staging into the basic design. Although these
burners are capable of meeting the NCx emission requirements up to 1998 it is
anticipated that a tightening of the regulations within the EEC will occur
before that date. Improved1 internal staging, furnace staging and, perhaps, post
combustion No reduction techniques will have to be introduced to meet these
A
more strict emission limits.
If post combustion reduction techniques are eventually required, an accepted
basic low N0X burner system will enable any future emission regulations to be
met effectively both in terms of speed of implementation and minimum capital
cost.
PRINCIPLES OF BURNER DESIGN AND DEVELOPMENT
The current NEI-ICL low NO^ wall burner design is shown in Figure 1. Air
staging is achieved by splitting the combustion air into independently swirled-
secondary and tertiary streams. Fuel staging is achieved by means of fuel flow
redistributors (FFR) located in the, pu1verised coal/primary air stream close to
the burner exit. Situated in this location the FFR produce a fuel lean/fuel
rich profile at the burner mouth. Ignition, of the main pulverised coal fuel
(PF) is achieved via a centrally located oil burner with its integral combustion
air supply fan. PF is supplied, from the PF supply piping, via a tangential
inlet and scroll distribution system to the annular burner fuel duct. The
design concepts were developed using isothermal modelling techniques, to examine
both the flow of fuel and the air distribution within the burner - system. Fuel
flow work addressed the problem of roping within the burner fuel annul us and
produced an evenly distributed flow into the FFR system which then produced the
required fuel staging effect at the burner exit. Various forms of FFR devices
were tested using flow visualisation techniques. Air distribution and air swirl
were studied in relation to the recirculation and general mixing patterns
produced both in the near burner region and further downstream. Figure 2
illustrates a typical recirculation pattern from an early burner design.
Following the isothermal model work a series of potential low NO burner design
configurations were selected for thermal testing, at full scale, in the 88 MW
NEI-ICL burner test facility. The initial full scale tests related to a 37
burner design which would be required for several 48 burner 500- MW front wall
6B-36
-------
coal fired units, in. the UK. During this work, operating parameters, such as
those relating- combustion air preheat and heat input to NO levels, were
established (Figures 3a and 3b). In this work the principle of good flame
retention at the burner mouth, as a pre-requisite of low NO operation, was also
established. This,and the effectiveness of the FFR, in controlling overall N0x
emission levels, is illustrated in Figures 4a and 4b.
In order to relate the test rig burner performance to site boiler performance,
particularly with respect to NO emissions, the test rig was refractory lined in
a pattern determined by computer calculations, such that the rig centre line
temperature was similar to the boiler centre line flame temperature, as shown- in
Figure 5 (4). To demonstrate the effectiveness of .this approach a standard
burner from a 500 MW boiler was rig tested under these conditions and did.
reproduce site N0x levels of around 700 ppm at 3% Oz . Thus a 1:1 rig factor in
respect of NO emission levels was established.
Further work was carried out on flame retention, which resulted in successful
patent applications, for the burner design(5) and also up-rating of the design
from 37 MW, , to 58 MW., without an increase in NO emissions. The 58 MW burner
th th x
was also' required to operate with a primary air to pulverised coal ratio of
1.2:1 compared to the more usual 1.5/2:1 range. Furthermore the primary air was
vitiated by the use of recycled flue gas into the ball mills for coal drying
purposes. This primary air vitiation and low pa:pf could aid low NO
A
performance of the burner but also adversely affect flame stability and burn
out.
Figure 6 demonstrates the NO performance of this larger burner showing not only
the usual trend, of increasing NO with waste gas Oa content (with a N0„ level of
3 x x
375 ppm at 3% O2), but also that the burner can operate at lower excess air
rates than normally used for coal firing without the generation of high CO
levels. Corresponding with CO levels below 100 ppm the carbon in dust levels
measured on the rig tests were a maximum of 21. During the thermal trials the
opportunity was taken to collect in.-f 1 ame gas samples and temperature
measurements. Contour plots of gas and temperature variations are shown in
Figures 7a-7d. These emphasise the importance of the near burner region
aerodynamics in establishing a centrally located reducing atmosphere with the
flame envelope which encourages the formation of Ni rather than N0X from the
nitrogen contained in the fuel. High N0X levels were produced in the outer
regions of the flame, close to the burner, corresponding with the mixing of
6B-37
-------
secondary air and the outer layers of the fuel stream. This N0X mixed later in
the flame with the reductants produced in the flame core, thus producing a low
overall N0x emission from the flame.
Depending on the particular conditions rig NO levels were in the 300-400 ppm
range (related to 31 02, dry)'which represents an approximate 501 reduction in
N0x.
PERFORMANCE OF BURNERS IN SITE INSTALLATIONS
Prior to the possible retrofitting of a full boiler set of low NO burners it
A
was considered prudent to replace just one or two standard burners, with the Tow
NQX designs, in an operating boiler. This preliminary installation would enable
the compatibility of the low NGX burners, within a hot multi-burner furnace
environment, to be assessed from an operational and durability standpoint. Two
37 MW^ low NO^ burners were installed, on a 48 burner 500 MW boiler, in the
wing and centre top row locations and a single 58 MW^ burner installed in the
centre top row position of a 32 burner 500 MVJ boiler. The centre top row
location was considered to give the most hostile conditions regarding burner
component temperatures, particularly in the non-firing mode. The wing position
enabled a qualitative assessment of the burner, in- operation, to be made. The
centre top row burners were inspected, in-situ, using a water cooled periscope
inserted into the burner via de-ashing ports, critical components of the burner
were instrumented with thermocouples to provide burner . metal temperature
variations in both the firing and non-firing operational modes.
Temperatures recorded from the single low NO burner, installed in the standard
burnered furnaces, gave some cause for concern, as in the non-firing mode,
temperatures approaching recommended limits for the material used in the
critical burner areas were recorded, with the normal 10-151 MCR cooling air
equivalent passing through the burner.
Computer calculations of heat flux based on test rig data, of low NO burner
A
operation, showed that with a full boiler set of low N0X burners the temperature
of the critical burner components would be satisfactory. The main reason for
this was the lower peak flame temperature of the combustion staged low NO
burner system which also occurred further down stream from the burner exit.
There was also a change in the gas recirculation pattern at the furnace front
wall as a result of the low N0x burner design.
6B-38
-------
Peri scope observations indicated, the possibility of some ash deposition in the
low NO burner installed in a conventional boiler burner system. From both test
X
rig experience and computer predictions it was postulated that the change in
front wall flow patterns from
eliminate this possibility,
a full boiler set of low N0„ burners would
Although both rig operating experience and computer predictions indicated that
neither high material temperature or ash deposition would be a problem with a
full set of low NO burners, material specifications for the critical burner
X
components were selected and a minor modification made to the secondary air
stream aerodynamics to provide further assurance. In practice, with the full
boiler set of low NQX burners, the computer and test rig predictions, regarding
critical -burner metal temperatures and ash deposition, were verified. By
carrying out these investigations a considerable data bank was compiled on
potential materials for burner construction covering fabricated, cast materials,
coated materials and ceramics.. Data on erosion resistance of these materials
exposed to flowing pulverised coal streams were also obtained. Table 3 compares
the temperatures measured in the single low N0X burner and the multiple low NC'x
burners after the boiler modification.
Table 3
BURNER METAL TEMPERATURES COMPARISONS
BEFORE AND AFTER LOW N0X BOILER MODIFICATIONS
Burner Component
Temperature °C
Before Modification
mean peak
After Modification
mean peak
Terti.ary Air Duct
Secondary Air Duct
880
870
980
950
868
838
1011
915
Primary Air Duct)
Flame retainer )
Oil burner core tube
1070 1170
730 810
917 940
707 792
Temperatures in Table 3 relate to the non-firing mode with 10-15% of normal
firing air supply passing through the burner.
6B-39
-------
Also, prior to the installation of a full boiler set of low N0x burners, the N0x
and CO levels were measured on an unmodified boiler (6). The results are shown
in Figure 8. In. general the unmodified boiler NO emissions were in the range
550-730 ppm (related to 3% O2, dry), depending upon the excess air level, with a
mean level of 633 ppm at 3% O2. Thus a 30% reduction in N0X would require the
boiler to operate at a mean figure of 443 ppm well within the capacity of the
burner, from the rig test data (see Figure 6). Carbon in dust from the
unmodified boiler was in the range 0-6? - 3.3% (mean 1.93%) depending upon mill
groups in operation and excess air levels, under similar conditions CO levels
were recorded in the 60-200 ppm range.
Figure 9,, shows the results from the initial commissioning, trials of the full
boiler set of low N0X burners, covering the whole range of mill groups and
excess air levels, equivalent to the 2-5% waste gas O2 range and compares them
wit te test rig burer performace. Summarising these early results from te
boiler, the low N0V burners, in combination, can operate under the conditions
outlined in Table 4.
Table 4
INITIAL COMPLIANT OPERATING RANGE
OF LOW N0X BURNERS
Oz level
3% n
N0X ppm 330 430
CO ppm 25 10
C in Dust % 5 2
N0X levels in Table 4 refer to ppm at' 3% Oz dry.
The results confirm the 1:1 rig factor to boiler factor relating to N0X
emissions, in the 3-41 waste gas Oe range. The CO emission results in Figure 9
indicate that the 100 ppm CO level would not be exceeded until excess air levels
equivalent to 1.8% O2 were obtained, this compares to 2.6% Oa in the unmodified
boiler. Over the 3-41 waste gas O2 range the CO levels in the boiler were
similar to those in the rig tests, however there is a tendency for a more rapid
increase in CO generation, below 3% O2, in the boiler compared to the test rig.
6B-40
-------
The average of all the boiler N0x level results gave 399 ppnr N0X which
corresponds to a 37% reduction' in N0x compared to the mean level of N0x from the
unmodified boiler. This reduction should be even greater when burner
optimisation is complete to. enable the burners to operate at lower 0« levels
without excessive CO generation in the boiler. Carbon in dust levels increased
in the low N0x burnered boiler to an average of around 51 (at 3% Qz) compared to
2% in the unmodified boiler (see Figures 8&9). The general practice with this
boiler is to over-fire on the bottom rows of burners in the unmodified boilers,
as a means of controlling superheater temperatures and this practice has been
continued on the modified boiler. Some burners are therefore operating at lower
overall air.to fuel ratios, however, the increased swirl and hence shorter flame
length of the unmodified burners produces sufficient in furnace time and
turbulence to produce a low C in dust loss overall.
As a result of staged combustion effects low NQX burners have a low overall
swirl producing increased flame lengths and low furnace turbulence levels. We
now know that higher carbon in dust levels are generated from the burners which
are operating at lower overall air levels. The time, temperature and mixing
history {Oz availability), which controls the combustion reactions within the
boiler, including NQX emissions is influenced by furnace geometry and air
quality. The 10 m depth (with an approximate 3:1 width:depth ratio) of the
boiler coupled with the use of vitiated air for coal conveying have an adverse
effect on the final burn-out characteristies. Optimisation of the boiler and
burner performance, fully recognising the low swirl characteristics of the low
NOx burners, should improve this situation.
CONCLUSION
Single full scale burner test facilities can be used to indicate multi-burnered
boiler N0x emission levels. Combustion staged low N0x burner designs are
capable of meeting current legislation relating to NQX emission levels.
Front wall environments are less hostile to burner components in a low N0x
system compared to a conventional front wall coal burner system.
Low NQx burner characteristics must be fully recognised in the optimisation of
low N0X front wall burner boiler operations.
6B-41
-------
ACKNOWLEDGEMENTS
Thanks are due to the Directors of NEI-ICL for permission to publish this paper
and' to many colleagues within NEI-ICL responsible for providing both test rig
and boiler commissioning data.
Thanks are also due to PowerGen Technical and Station personnel for the
provision of boiler operating data and continued enthusiastic interest in the
project,
REFERENCES
1, Official Journal of the European Communities L336 "Council Directive
88/609/EEC of 24th November, 1988 on the Limitation of Emissions of
Certain Pollutants into the Air from Large Combustion Plants"
7th December, 1988.
Z. J. W. Allen, W. J. D, Brooks, N. A. Burdett, F. Clarke and G. Foley.
"Reductions in NO Emissions from a 500 MW Corner Fired Boiler." Joint
x .
Symposium on Stationary N0„ Combustion Control. New Orleans, 1987..
— —X —
3. J. W. Allen "N0X Reductions in Coal fired Boilers."- Modern Power -
Systems. June, 1987.
4. Private Communications. M. J. Sargeant, S. Cooper - CEGB, Marchwood
Engineering Laboratories, 1984.
5. UK Patent 8805208
USA Patent 317743
European Patent 89302101.4
6. Private Communication. CEGB
6B-42
-------
Secondary air
control damper
Outer back plate
Secondary air
. swirl varies
Tertiary
air vanes
Conical liner
Core air tube
PA/PF inlet
Secondary
air tube
Tertio ry
air tube
Entry chamber
Rod'ding tube
Secondary/tertiary
;;ir shut off damper
Fuel flow
redistributors
Figure 1. Low NlOx Front Wo I! Coal Burner.
Axial distance (m)
1.5'
Burner centre line
1,0 s
0.5'
v W w
Flame
boundary
Central •
recirculation
zone
V
.6 ,4 .2
~
—! 1 1 r
.2 ,4 .6 .8
Figure 2. Low NOx Coal Burner Model.
Typical Recirculation Pattern.
6B-43
-------
NOx (ppm)
500
400 -
300
200
100
0 100 200
Air preheat temp. (°C)
3a Effect of Air Preheat on
3% 02)
300
NOx (ppm)
500
400
300 ¦
200
100 •
1
50%
Burner load
100%,
3b Effect of Burner Load on
NOx (Excess Air = 3% C2)
100% Load = 58MW.
NOx (Excess Air
Figure 3. Effect of Air Preheat and Burner Load an NOx
NOx (ppm')
700
600
500
400
300
200
100
0
Fully lifted
flame
Well anchored
flame
0 1 2 3 4 5
% 02 in waste gas
4a Effect of Flame Retention on NOx
NOx (ppm)
7Q0
600 ¦
500
400
300
200 ¦
100 •
Burner
without
FFR
Burner
with FFR
0 1 2 3 4 5
% Qa in waste gas
4b Effect of Fuel Staging an NOx
Figure 4. Effect of Burner Parameters on NOx
6B-44
-------
Centre cell '
gas temp. (°K)
2000 ¦
1750 •
1500 ¦
1 250 •
1000 ¦
750
500 •
X.
Test rig
Boiler
1
6
I I I I I I !
8 10 12 14 16 18 20
Axial distance (m)
Figure 5. Comparison of Refractory Lined Rig and
5Q0MW Boiler Centre Line Temperatures.
% 02 in waste gas
Figure 6. Test Rig Performance of 58MW (Thermal)
Front Wall Coal Burner,
6B-45
-------
CD
CD
¦
-fc.
a>
0-1- 250
1 -
2-
3-
Burner centre line
4-00
0
1 -
2-
0
I
2
~r-
4
I
6 8 10 12
Distance along axis (m)
7a NOx Contours (ppm)
n T
Burner centre line
1
1 4
r~
16
18
~ ~ I l
6 8 10 12 14 16 18 0 2 4 6 8 10 12 14 16
Distance along axis (m) Distance along axis (m)
7c CO Contours (%) 7d Temperature Contours (°C)
Figure 7. In —Flame Gas and Temperature Contours.
18
0 —i
Burner centre line
3-
s *
2.5
~i I r~
10 12 14
4 6 8
Distance along axis (m)
7b 02 Contours (%)
Burner centre line
! 6 18
1 -
900
800
4
I
8
"i r
18
-------
NOx (ppm)
700 '
NOx
~I
'5
4
% 0? at economiser
CO (ppm)
100
CO
! r r
2 3 4
% O, at economiser
I
5
% C
8
Unburnt Carbon
Figure 8.
5
% 02 at economiser
Unmodified Boiler Performance
6B-47
-------
NOx (ppm)
600
KEY
NOx
o Boiler
x Test rig
i
¦4
CO (pom) CO
% 02 ct economise'
% c Unburnt Carbon
% Oj at economiser
Figure 9, Modified Burner Performance on Boiler
During Commissioning, Compared to Single Burner
Test Rig Performance
6B-48
-------
Preliminary Test Results
Hlgfa Energy Urea: Injection DeNQx on a 215 MW Utflily Bailer
by
Dale G. Jones, Ph.D., P.E., Noell, Inc.
Stefan Negrea, P.E., Noell, Inc.
Ben Button, Noell, Inc.
Larry W. Johnson, P.E., Southern Calif. Edison Co.
J. Paul Sutherland, P.E., Southern Calif. Edison Co.
Jeff Tormey, Southern Calif. Edison Co.
Randall A. Smith, Fossil Energy Research Corporation
ABSTRACT
Initial tests of a high energy urea injection SNCR DeNOx system have been
completed at Southern California Edison's Huntington Beach Unit 2 gas- and
oil-fired boiler. The SNCR DeNOx temperature window in this 215 MW utility
boiler occurs In narrow cavities and between boiler convection sections. The
Huntington Beach SNCR DeNOx project is a demonstration of high energy
urea Injection in narrow cavities to evaluate various DeNOx alternatives and
to bring such installations in compliance with South Coast Air Quality
Management District regulations for the metropolitan area.
Following contract award In June, 1990, Noell proceeded with injection system
design, installation and start up. Initial tests of high energy injection into the
2nd cavity and other boiler zones were conducted between Jan. 15 and March
5, 1991. Pressurized urea-water mixtures were injected into cross-flowing flue
gas using high velocity air-driven nozzles. Initial 2nd cavity Injection tests
showed that 25% to 40% DeNOx Is achieved at full load despite adverse
conditions of short cavity residence times (i.e. 40 milliseconds) and floor-to-
roof adverse temperature gradients (i.e. about 200 F). Such adverse conditions
in the 2nd cavity also caused unacceptably high levels of NHs slip.
Additional tests were therefore performed to Investigate urea injection Into the
1st cavity where the full load temperature is about 2050 F. Using only four (4)
sidewall Injection nozzles, 20% to 25% full load DeNOx was obtained at urea
feedrates from NSR = 2 to NSR = 4 (NSR is moles of NHl injected vs. moles of
initial NOx). Under these conditions, NH3 slip measured upstream from the air
preheater averaged less than 3 ppm, or less than about 1.5% of NHl feedrate,
Noell Is proceeding with further development of advanced injection systems to
be considered for installation and additional testing at Huntington Beach.
7A-1
I
-------
1.0 Introduction and Background
Initial tests of a high energy urea injection SNCR DeNOx system have been
completed at Southern California Edison's Huntington Beach Wnlt 2 gas- and
oil-fired boiler. The SNCR DeNOx temperature window in this 215 Mw utility
boiler occurs in narrow cavities and between boiler convection sections. The
Huntington Beach SNCR DeNOx project is a demonstration of high energy
urea injection in narrow cavities to evaluate various DeNOx alternatives to
comply with South Coast Air Quality Management District regulations.
Urea (NH2.CO.NH2) reacts at high temperatures with NOx in combustion flue
gases, approximately as follows;
2 NO ~ NH2.CO.NH2 + 0.5 02 « 2 N2 ~ 2 H20 + CO2
Amine radical (NH2) resulting from thermal decomposition of the urea reacts
with NO. The chemical feedrate vs. quantity of NOx is called the normalized
stoichiometric ratio (NSR), defined as the molar ratio of NHi being Injected
divided by initial NOx. At Isothermal conditions, the SNCR DeNOx process
operates best over a narrow 'temperature window* between 1600 F and 1900 F.
If the flue gas temperature is too hot, some of the NH2 radicals form additional
NOx and DeNOx performance decreases. If the flue gas temperature is too
cold, some of the NH2 radicals form byproduct NH3, called 'ammonia slip* and
DeNOx performance goes down. Thus, a 'temperature window" exists.
This narrow temperature window is the primary drawback of boiler injection
SNCR DeNOx technology. When boiler operations change, temperatures at an
injection location also change. Therefore, multiple levels of injection are
usually required to provide good DeNOx performance over a range of boiler
conditions. At low load, the temperature may be too cold, and Injection
should occur at a location closer to the furnace. At high toad, the temperature
may be too hot, and injection should be at a location further from the furnace.
Noell's boiler injection DeNOx system uses high velocity injection jets to
provide Intense flue gas mixing. These jets can overcome distribution problems
typically observed, such as non-uniformltles in temperature, flowrate, and/or
composition of the flue gas. As in any chemical process, intimate and complete
mixing is important. By proper design and operation of the injection system,
close approximation to a well-mixed reactor can be achieved. Noell's boiler
injection Jets are used for flue gas mixing and operate independently from
chemical feeding, accomplished using feed pumps for higher or lower levels of
DeNOx. Chemical distribution occurs first In the Injection Jet, and then as the
Injection jet(s) mix with cross-flowing flue gas. Noell's boiler Injection concept
is illustrated in Figure 1, which provides results of Jan, 1988 injection system
flow model testing for the KVA/ Basel MSW Incineration plant The left picture
shows 'channelling*, where a smoke stream passes through the flow path
without much mixing. The right picture is similar except that scaled-down
injection jets were Installed Into the sidewall(s) of the flow model to determine
effects on mixing. As can be seen, such high energy injection jets have a major
Impact on flue gas mixing. Similar full-size injection jets were subsequently
Installed in the 330 TPD Basel MSW incinerator. At maximum boiler output at
330 TPD Incinerator feed rate, NOx removal of 70% was obtained at urea NSR
a 1.3, along with relatively low levels of NH3 slip. (Reference 1).
7A-2
-------
Figure 1; Photographs of Flow Model Test Results
KVA Basel 330 TPD MSW Incineration Furnace
January, 1988
•Channelling-Effect Injection Jet Effect
fleft.hand picture! (right-hand picture)
7A-3
-------
Noell has also installed Its high energy boiler Injection SNCR DeNOx process
at the 325 MW coal-fired power plant of BKB/Offleben in Germany, which was
started up for commercial operation in Sept, 1989. In this coal-fired boiler,
Noell's steam-driven nozzles are used for urea injection to achieve 95 ppm NOx
at full load. At full load, the urea NSR is about 0,64, corresponding to about
32% DeNOx with NH3 slip of less than 1.0 ppm. Due to the S02 content of the
flue gas, the Offleben requirement is less than 5.0 ppm NH3 slip to avoid
forming ammonium bisulfate deposits in the air preheater. (Reference 2)
In more recent developments, Noell has been awarded a contract by the Public
Service Company of Colorado (PSCC) to design and procure boiler injection
SNCR DeNOx equipment for a Clean Coal III project at PSCC's Arapahoe coal-
fired station. This boiler injection SNCR DeNOx project is being co-sponsored
by the U.S. DOE and by EPRI. Noell has also been awarded a contract by the
Tennessee Valley Authority (TVA) to conduct perform field testing of flue gas
temperatures, and conduct boiler flow model testing of injection system
options for a project being considered by TVA to demonstrate boiler injection
SNCR DeNOx at a large coal-fixed power plant
2.0 Description of Huntington Beach Unit 2 Boiler
This gas- and oil-fired 215 MW boiler incorporates a pressurized furnace with
front wall-fired burners arranged 6 wide by 4 high. The drum-type natural
circulation steam generator includes pendant secondary superheater and
reheat superheater convection sections. It is in the area of these pendant
sections that flue gas temperatures at full load on gas fuel reach levels of
interest for SNCR DeNOx. Full load superheater outlet conditions are
1,560,000 lb/hr at 2450 psig and 1050 F. Flue gas from the furnace passes
horizontally through the secondary superheater, a water screen formed by the
rear wall tubes of the furnace, the reheater, and the pendant loop portion of
the primary superheater. Following the rear cavity, the flue gas then passes
vertically downward through the balance of the convection sections, air
preheater and stack. Flue gas recirculation fans are provided for accurate
control of superheated steam temperatures. At full load on gas fuel, about 8%
of the flue gas is recirculated to the furnace bottom hopper. A side sectional
elevation of the boiler is shown In Figure 2. The furnace cross section in the
vertical upflow direction Is 24 ft wide by 50 ft. deep.
Detailed description of the boiler convection sections goes beyond the scope of
this report. It is sufficient to say that the flue gas velocities at full load on
gas fuel are such that the residence times in the 1st and 2nd cavities between
convection sections are on the order of 40 milliseconds (msec), and that flue
gas temperatures initially decrease at a rate of about 4 F/msec In the first
pendant section, and then at a rate of about 2 F/msec in the second and third
sections. These narrow cavities and very short residence times are typical tor
many gas- and oil-fired boilers, and offer perhaps the most difficult type of
challenge for application of boiler injection SNCR DeNOx. An earlier
publication by Mittelbaeh, et.al. indicates that at 1800 F or above, flue gas
residence times of about 100 msec would be sufficient to complete most of the
SNCR DeNOx reactions (Reference 3). In the case of the Huntington Beach
Unit 2 boiler, this expectation was overly optimistic.
7A-4
-------
Figure 2: Side Sectional Elevation, Huntington Beach Unit 2 Boiler
Southern California Edison Company
7A-5
-------
3.0 Flue Gas Temperatures
Prior to design of the injection system, flue gas temperature data was obtained
using HVT probes at the upper furnace front and side-wall observation doors,
and by acoustic pyrometer to obtain average flue gas temperature at the inlet
of the first pendant tube section. The various field measurements of flue gas
temperatures were compared with boiler manufacturer design data as follows:
Table 3.1 COMPARISON OF FLUE GAS TEMPERATURES
Huntington Beach Unit 2 at Full Load (Gas Fuel)
' Source of Data SSH Inlet Ipt Qvlty 2nd Cavltv
HVT Probe @ Observation Doors 2230 F n/a n/a
Acoustic Pyrometer @ Obs. Doors 2280 F n/a n/a
HVT Probe @ Manway Doors n/a 1910 F I?) 1760 F
Manufacturer Design Sheets 2340 F n/a 1775 F
The field data seemed to be in reasonable agreement with boiler manufacturer
data. Computer-generated prediction of 2nd cavity temperature contours (full
load on gas fuel) were also provided by the boiler manufacturer, which
indicated cooler zones averaging 1700-1800 F near the 2nd cavity floor, hotter
zones of about 1850-1950 F in the middle, and then 1800 F or above nearly all
the way to the 2nd cavity roof. Based on the foregoing, there was no reason
to doubt that the 2nd cavity was the preferred injection zone. The 2nd cavity
measures approximately 16 ft. high by 50 ft wide in cross-section.
Following installation of the 2nd cavity injection nozzles, further data was
obtained. Temperature profiles from HVT measurements In the 2nd cavity are
provided in Figures 3 and 4, where the strong Influence of burner patterns
under otherwise identical operating conditions Is easily seen. Burner pattern
adjustment caused average flue gas temperatures to increase (or decrease) up
to 100-150 F. The entire SNCR DeNOx temperature window is only 300 F, and
changes of 100-150 F are quite significant As seen In Figures 3 and 4, flue
gas temperatures also decreased up to 200 F from the floor to the roof. This
adverse temperature gradient substantially shortened the 2nd cavity injection
residence times within the 1600-1900 F SNCR DeNOx temperature window.
4.0 Description
-------
Figure 3; 2nd Cavity Flue Gas Temperatures Near Boiler Centerllne
Huntington Beach Unit 2 Boiler, Full Load, Gas Fuel
Southern California Edison Company
KVT Tefflpermtu re ifl
1000
4 * 0 io ia h i*
Height Above floor fll 2nd Cavity, ft
KVT Temperature (R
ft
JTt
r
Pre
mm
Xmk
—
es «
* f
. its
¦ SI
lewi
L
-3r<
/0(
L>
-Sf
S-E
J/Oi
L*t"J
>s-^
'est
X
s
\
/
K
V
B
DOl
-We
"/
r
*
iv
1
¦«
Sj
/
\
so<
3^
ut
V
v
n
v
n
'
<9 13 M l«
Helgbt Above Fleer of 2nd Cavity, ft
Figure 4; 2nd Cavity Flue Gas Temperatures Near Boiler Walls
Huntington Beach Unit 2 Boiler, Full Load, Gas Fuel
Southern California Edison Company
7A-7
-------
5.0 Results of 2nd Cavltv Injection Tests
System tests involved selection of pump settings for controlling the urea-water
mixture ratio. The liquid mixture was tlien pusiped to tlxe boiler level and
Injected Into the cross-flowing flue gas using air-driven nozzles operating at
sonic jet velocities. A number of hign velocity injection nozzles were installed
In the floor zone of the 2nd cavity. By means of aspirated ports, these nozzles
could be extended or retracted up to 8 ft, into the pressurized flue gas zone,
without influencing boiler operations. Two (2) air orifice sizes were tested, the
larger orifice(s) requiring an Injection air flowrate of about 2.1% of the full
load flue gas flowrate, and the smaller orlflce(s) requiring about 1.2%.
Figure 5 shows the effect of boiler load and burner pattern on percentage
DeNOx for 2nd cavity Injection at NSR = 2 for the two (2) sizes of injection
nozzles. As can be seen, the effect of increasing boiler load with ABIS fall
burners In service) was to increase the DeNOx performance. With normal
BOOS (burners out of service), increasing boiler load at a constant urea
feedrate for NSR = 2 at full load caused a decrease in DeNOx performance.
With the smaller nozzles, reduced DeNOx performance especially at full load
was partially caused by reduced flue gas mixing at higher flue gas velocities.
Figure 5 illustrates the effect of adjusting the burner pattern from normal
BOOS to ABIS, which causes increased flue gas temperatures (Figure 3 & 4).
The increased flue gas temperatures, in turn, caused a full load DeNOx
performance increase from 27% to 40%. Since the change in burner pattern
caused 2nd cavity flue gas temperature changes of 100-150 F, and since the
resulting DeNOx increase (at otherwise identical conditions) was relatively
large, it was concluded that SNCR DeNOx in the 2nd cavity at full load was
operating at the colder edge of the 1600-1900 F temperature window. The
injected urea behaved as if the isothermal temperature was about 1600 F,
regardless that full load HVT temperatures in the 2nd cavity itself averaged
1720-1780 F. These Initial full load results up to 40% DeNOx were achieved
despite adverse conditions of short cavity residence time (i.e. 40 milliseconds)
and 2nd cavity floor-to-roof adverse temperature gradient fi.e. about 200 F).
Despite moderate DeNOx levels which were achieved, such adverse conditions
In the 2nd cavity caused unacceptably high levels of NH3 slip.
Further analysis of these Initial test program results showed that the hotter
1st cavity or upper furnace zones offered better locations at full load for high
energy SNCR DeNOx injection than the 2nd cavity.
6.0 Tests of 2nd Cavltv Injection Nozzle Supply Pressure
Additional tests were conducted using the larger 2nd cavity nozzles. In these
tests, the boiler was held at full load, and urea NSR feedrate was increased to
determine DeNOx vs. NSR. The results are presented in Figure 6, where it is
seen that with a tower nozzle pressure, the DeNOx cannot be increased beyond
about 20% regardless how much the chemical feedrate is increased. This type
of response curve is indicative of relatively poor flue gas mixing, where the
SNCR DeNOx process become mixing limited. At the higher nozzle pressure,
there is a continuing Increase In DeNux performance up to about 37% as NSR
is increased up to about 5. This second type of response curve Is indicative of
relatively good flue gas mixing.
7A-8
-------
Figure 5; Effect of Boiler Operations on 2nd Cavity Injection DeNOx
Huntington Beach Unit 2 Boiler, Gas Fuel
Southern California Edison Company
Percent NOx Removal
HBGS Unit 2 Boiler Load, Gas Fuel. MW (Net Output)
7A-9
-------
7.0 Results of 1st Cavity Infection Tests
Additional tests were performed to Investigate 1st cavity Injection at higher
full load temperatures, which averaged about 2050 F in the 1st cavity. This
was several nundred degrees Fahrenheit hotter than the average full load
temperature in the 2nd cavity. The existing 1st cavity sootblowers were
removed and air-driven nozzles were installed Into these existing membrane
wall aspirated ports. Using four (4) sldewall nozzles with known limitations
in flue gas cross-sectional coverage, 20% to 25% full load DeNOx was
obtained with urea feedrates from NSR = 2 to NSR = 4 (Figure 7). For these
same urea NSR feedrates and operating conditions, NH3 slip as measured
upstream from the air preheater was well below 1.5% of the NHl Injection rate,
and averaged less than 3 ppm. Despite the very high 2050 F temperature, the
SNCR DeNOx process operated beyond expectations, especially considering the
relatively poor flue gas cross-sectional coverage and mixing afforded when
using only four (4) siaewall nozzles.
8.0 Results of Upper Furnace Inleetlon Tests
Further tests were also performed to determine upper furnace injection DeNOx
as a function of boiler load. Again, only four (4) sldewall nozzles were used
where existing observation doors (aspirated) were available. The chemical
feedrate during these tests was maintained at a constant value which
provided NSR a 2 at full load conditions. As shown in Figure 8, the
percentage DeNOx decreased from a maximum of about 40% at a reduced load
of 120 Mw. At full load on gas fuel, the flue gas temperatures are about 2300
F at the inlet of the first boiler tube bank. This is too hot for SNCR DeNOx.
and as shown in Figure 8, the DeNOx decreased down to about 5% at full
load. NH3 slip characteristics are also shown in Figure 8, where It Is seen that
at about 145 MW or 150 MW boiler load, upper furnace flue gas temperatures
are most favorable for optimum SNCR DeNOx performance.
9.0 Further Work In Progress
Noell is proceeding with further development of advanced injection systems to
be considered for Installation and additional testing at Huntington Beach.
7 A-10
-------
Figure 6: Effect of Nozzle Pressure on 2nd Cavity Injection DeNOx
Huntington Beach Unit 2 Boiler, Full Load, Gas Fuel
Southern California Edison Company
Per cat DeNOx Uom lolUai NO* * 110 ppm
225 MW, Cu fuel
40 "
j
I
13
J»!g
-~
5.3 pi
lg
/i-
rH
/
/
/
/
/
f
35
30
5
"ftrwat
De*j0K
33
NO%{*
3«
33
IS
¦
•
/
/
/
_ _
•
-
13 3.
Normalized Stmchiomttrtc Ratio (NSRJ
fj
tr
Figure 7: 1st Cavity Sidewall Injection DeNOx vs. NSR
Huntington Beach Unit 2 Boiler, Full Load, Gas Fuel
Southern California Edison Company
7 A-11
-------
Figure 8: Effect of Boiler Load on Upper Furnace Injection DeNOx
Huntington Beach Unit 2 Boiler, Gas Fuel
Southern California Edison Company
5
r
%
I
7A-12
-------
Results and Conclusions
1. Narrow cavities and very short residence times In many gas- and oil-
fired boilers offer perhaps the most difficult challenges for application of
boiler injection SNCR DeNOx.
2. Flue gas temperature variations caused by normal boiler operations can
and will have significant effects on boiler injection SNCR DeNOx,
even when there are no changes In boiler steam production or load.
Successful load-following SNCR DeNOx systems must have multiple
injection zones and relatively sophisticated controls.
3. Detailed field temperature measurements and flow model optimization
tests of injection jets are considered prerequisites for the design of high
performance (boiler-specific) SNCR DeNOx injection systems
4. Despite adverse time/temperature conditions In narrow cavities between
adjacent convection sections to the Huntington Beach gas-fired boiler,
full load DeNOx performance was obtained as follows*.
Injection Zone Nozzle Posltlonfsl. DeNOx NH3 Slip
2nd Cavity Multiple Floor Nozzles 25%-40% high
1st Cavity Sidewall Nozzles (4) 20%-25% low < 3 ppm
Upper Furnace Sidewall Nozzles (4) 0%-5% zero
5. This Initial Huntington Beach test program has shown that SNCR
DeNOx is a function of available DeNOx reaction time plus Injection
system cross-sectional coverage and mixing. In this application at full
load with short residence times, injection into the 1st cavity at a flue
gas temperature of about 2050 F appears to provide the best SNCR
DeNOx results.
6. Noeli is proceeding with further development of advanced injection
systems to be considered for installation and additional testing at
Huntington Beach.
References
1. Jones, D.G., et. al,. 'Two-Stage DeNOx Process Test Data from
Switzerland's Largest Incineration Plant*, EPA/EPR1 Symposium
on Stationary Combustion NOx Control, San Francisco, California,
March 6-9, 1989.
2. Negrea, S., et. al., 'Urea Injection NOx Removal on a 325 MW Brown
Coal-Fired Electric Utility Boiler in West Germany", 52nd Annual
Meeting, American Power Conference, Hyatt Regency Chicago,
April 23-15, 1990.
3. Mittelbach, G., et. al., 'Application of the SNCR Process to Cyclone
Firing', Special Meeting on NOx Emissions Reduction of the VGB,
German Power Industry Association, June 11-12, 1986.
7 A-13
-------
-------
EVALUATION OF THE ADA CONTINUOUS AMMONIA SLIP MONITOR
Michael D, Durham, Richard J. Schlager, Mark R. Burkhardt,
Francis J. Sagan and Gary L. Anderson
ADA Technologies, Inc.
304 Inverness Way South, Suite 110
Englewood, CO 80112
,7 A -15
-------
-------
ADA Technologies, Inc. has developed a continuous emissions monitor for use with
advanced NO control technologies that is capable of simultaneously monitoring ppm levels
of NH3 and NO In flue gas. The instrument can also measure S02 when it is present in the
flue gas. The instrument is based on ultraviolet light absorption using a photodiode array
spectrometer. It has unique advantages over other ammonia instruments as it directly
measures ammonia as opposed to the indirect chemiluminescent techniques which must
infer the NH3 concentration from the difference between two large numbers, The monitor
has undergone extensive laboratory and field evaluation and data are presented which
demonstrate sensitivity, accuracy and drift of the instrument. The analyzer has been field
tested at a gas turbine with SCR, a coal-fired circulating fluidized bed with ammonia Injection,
a refinery boiler with SNR, and a utility boiler with urea injection. The accuracy of the
instrument was determined by comparison with extractive wet chemical measurements.
7A-17
-------
I, INTRODUCTION
ADA Technologies, Inc. has developed a continuous, reai-time analyzer for measuring part-
per-million levels of ammonia (NH3) end nitric oxide (NO) in flue gas associated with
advanced NOx reduction systems. A two-year long development program sponsored by the
U.S. Department of Energy resulted in an analyzer that is specific to ammonia, reliable, and
accurate. Other common flue gas components do not interfere with the measurement of
nh3.
This instrument filis the need created by advanced NOx control technologies for an ammonia
slip monitor which can be used as part of the process control system. Ammonia is a primary
ingredient in virtually all of the advanced NOx control processes such as selective catalytic
reduction (SCR) and selective non-catalytic reduction (SNR) technologies. However,
because of severe problems related to the penetration of unreacted NH? through the flue
gas treatment system, it is extremely important to measure and control the downstream
concentrations of NH3.
The instrument is an effective diagnostic tool for optimizing De-NOx systems, and will be a
valuable component of NOx control equipment in many applications including: coal-, oil- and
gas-fired utility boilers, co-generation plants, refineries, municipal solid waste incinerators,
and research programs.
The monitor has been operated as both an in-situ and extractive instrument. The extractive
mode of operation allows a testing team to evaluate the stratification of NH3 gas across the
diameter of a duct. This capability is particularly important in evaluating whether ammonia is
dispersed uniformly within the flue gas of a SCR or SNR De-NOx system.
II. MEASUREMENT PRINCIPLE
A. MEASUREMENT PRINCIPLE
Ammonia and NO absorb light in the ultra violet (UV) range at specific wavelengths, and the
shape of the absorption spectra are characteristic of the identity of the particular gas. Figure ¦
1 shows absorption spectra for NH3 and NO in a selected UV wavelength region. In this
spectral range, NO absorbs at two characteristic wavelengths, and NH3 absorbs at four
characteristic wavelengths. The two large doublet peaks identify the absorption due to NO,
and the four smaller peaks, which include two characteristic doublets, are due to NH3. The
quantity of light absorbed by a gas is proportional to its concentration, as defined by
principles of Beer's Law. Since the NO doublet located near diode 400 overlaps with one of
the ammonia peaks, this region cannot be used for analysis. However, the NO peak at
diode 850 and the ammonia peaks at either diode 200 or diode 600 do not interfere and
therefore can be selected and analyzed to determine the concentrations of these two gases.
The data available from the multichannel spectrometer allow measurement of these two
gases directly and accurately.
7A-18
-------
Figure 1. Absorbance spectra for ammonia and ammonia/nitric oxide mixture.
B. DETECTION SYSTEM
Photodiode array detectors provide a technology to improve upon the design of
conventional scanning monochromator-based spectrometers. The improvement involves
the placement of a series of detectors across the focal plane of a polychromator, each with
its associated readout electronics. The most advanced of these systems use a linear
photodiode array (LPDA) detector. The LPDA is a large-scale integrated circuit fabricated on
a single monolithic silicon crystal. It consists of an array of diodes, or pixels, each acting as
a light-to-charge transducer and a storage device. These detectors are ideally suited for use
in UV spectrometers because they have a large quantum efficiency (40-80%) throughout the
range as well as geometric,, radiometric, and electronic stability. The array itself can be
mounted and operated so as to be tolerant of high temperature, humidity, vibration, and
electrical and magnetic fields.
An LPDA spectrometer system, shown schematically in Figure 2, operates by passing a
continuous light source through the sample and into the polychromator. The polychromator
disperses the light across the LPDA, which has replaced the exit slit of a conventional
spectrometer. The array is located in the focal plane of the polychromator so that each
diode corresponds to a particular wavelength resolution of the UV-VIS spectrum. The diode
array provides an almost ideal sensor for the digital acquisition of spectra, as the array itself,
by its presence in the focal plane of the spectrometer, digitizes the spectrum into discrete
intervals. Unlike the scanning spectrometers, whose wavelength accuracy is mechanically
limited, the LPDA spectrometer is limited only by geometric constraints of the detector itself,
7A-19
-------
Deuterium
tanp
Gas Inlet
Gas Cell
Gas Outlet
Figure 2. Schematic diagram of Linear Photodiode Array spectrometer system.
by vibration and thermal expansion of the optical components, and by the stability of the
source. Wavelength accuracy is equivalent to the diode spacing multiplied by the linear
dispersion of the spectrograph. Its geometric registration and, therefore, its wavelength
accuracy and precision, are greater than any mechanically scanned spectrometer
With the PDA detector it is possible to develop algorithms which use the unique structure of
the absorbance spectrum to quantify the concentration of the gas. This approach eliminates
the need to maintain the initial intensity (y reference and simplifies and speeds the
calculation. Since the analysis procedure searches for characteristic features of the
absorption spectrum rather than a fixed wavelength, it is less sensitive to drift or lamp
intensity fluctuations.
The photodiode array detector has unique advantages over all the other ammonia
instruments. It provides a direct measurement of ammonia and is, therefore, inherently more
sensitive than the indirect chemiluminescent measurement techniques which must infer the
NH3 concentration from the difference between two large numbers. In addition, the
photodiode array spectrometer has the following unique features.
• The instrument can be built with no moving parts which will reduce
maintenance and increase reliability in an industrial environment.
• The software is written to provide built-in checks for alignment of the optics.
• Changes in light intensity to do create a drift problem.
• Finally, the interferences are well known and can be accurately handled by
the PDA detector.
7A-20
-------
III. LABORATORY EVALUATIONS
A. TEST SET-UP
Performance parameters of the analyzer were determined in a series of laboratory tests.
Gases used In the evaluation were supplied in cylinders containing the individual gases in a
background of nitrogen gas. The concentrations of the gases were certified by the
manufacturer through analysis. Gases were mixed in various combinations and
concentrations using mass flow controllers and manifold system. The gas flow was then
metered into the analyzer for evaluating performance.' Tests were conducted using a gas
cell with a path length of 90-cm. The cell was heated to maintain an internal gas temperature
of 300 °F. Results of the evaluation follow.
B. LINEARITY OF NH3, NO, AND S02
The linearity of'the response of the analyzer was evaluated by initially calibrating the analyzer
using nitrogen and a span gas for each component of interest. Gas concentrations were
then decreased in steps and resulting analyzer measurements noted. Results of the linearity
evaluation for NH3, NO, and S02 are shown in Figures 3 through 6.
Ammonia results are shown for two ranges of measurement, 0 to 70 ppm and 0 to 10 ppm.
Figure 3 shows that when calibrated at 70 ppm, measured concentrations are within 1 ppm
of the input concentration. For the low range, Figure 4 shows that measured concentrations
are within 0.5 ppm of the input concentration.
Prior to measuring the linearity of the NO, the instrument was calibrated using two
concentrations of NO because the absorbance of NO requires a second order equation to fit
the calibration curve. Using this technique, the linearity of the instrument is within 2% of the
actual concentration over a concentration range of 0 to 200 ppm as shown in Figure 5. If
only a single gas is used for calibration, there is a maximum 10% deviation from linearity in
the middle of the range.
Figure 6 shows the linearity of the analyzer for S02 calibrated at 80 ppm, For all gas
concentrations, the measured values are within 1 ppm of the input concentrations. The
capability to accurately measure sulfur dioxide provides the basis for eliminating its
absorbance as an interference to the measurement of NO and NH3.
C. LONG-TERM NOISE AND DRIFT
Analyzer noise and drift were estimated by observing instrument readings over a 36 hour
period of time as a mixed gas stream of fixed composition was passed through the
measurement cell. Analyzer measurements for, NH3, NO, and S02 are shown in Figures 7.
The composition of the gas stream was 10 ppm NH3,55 ppm NO, and 80 ppm S02.
7A-21
-------
0 to 20 30 40 50 60 70 80 90 100
Input NH3 Concentration (ppm)
Figure 3. Linearity of NH3 measurements when analyzer is calibrated using 70 ppm
standard gas.
Figure 4, Linearity of NH3 measurement when analyzer is spanned using 10 ppm
calibration gas.
7A-22
-------
Figure 5. Linearity of analyzer to NO input concentrations when calibrated using two span
gas concentrations.
Figure 6. Linearity of S02 measurements when analyzer calibrated using 80 ppm span gas.
7A-23
-------
90
ft?0-
¦A
a 60'
o
•3
g 50'
M
a
O 40 ¦
d
o
U 30'
sn
eS
O 20-
^— 4—^- 0
WW BO ppm SO, Input
-o-
OOOOQ 55 ppm NO Input
***** 10 ppm NHj Input
0 j i i—i—| I 'i—i—|—i—i—i—j—i—p—i—!—r—i—i—|—i—i—i—|—i—i—i—|—i
0 4 8 12 16 20 24 ' 28
Measurement Period (hr)
i j i r—r-
32 36
Figure 7. Noise and drift characteristics of NH3, NO, and S02 measurements over 36-
hours.
As can be seen from Figure 7, with unattended operation, the output is extremely stable for
all three gases. Analyzer noise is defined as the short-term peak to peak signal variation,
and is equal to_± 0.3 ppm for NH3, _+ 0.15 ppm for NO, and _+ 0.1 ppm for S02. Analyzer
drift is defined as the long-term variation in analyzer signal around an average value.
Analysis of the measurements shows that the drift is _+ 0.3 ppm for NH3, and j: 0.3 ppm for
NO drift, and is _+ 0.4 ppm for S02- These noise and drift measurements are well within the
accuracy capabilities of the gas flow delivery system using the mass flow controllers.
D. RESPONSE TIME
The response time of the analyzer is a function of how quickly a sample of gas is delivered to
the light path and the time it takes to process the spectral information into gas concentration
units. Since the data processing time is very short, on the order of a few seconds, the rate of
response becomes directly related to the volume of the gas cell and the flow rate of the gas
through that cell. For example, 90% of full scale response is achieved to a known NO
calibration gas input within five equivalent volume changes of the cell. This rate of gas flow
through the sample system is typically done within 1-minute. The response to ammonia gas
is slightly slower than observed for NO, due to the nature of ammonia gas which requires
conditioning of tubing surfaces during its travel to the measuring cell.
7A-24
-------
E. MINIMUM DETECTION LEVELS
The minimum detection level for a particular gas is defined as twice the noise value. Based
on data shown in Figure 7, the minimum detectable level using a 0,9 meter log gas cell is 0.6
ppm for NH3 and 0.3 ppm for NO.
The minimum detectable level and maximum concentration measurable using absorption
spectroscopy are a function of the path length that the light travels through a gas sample.
Higher gas concentrations can be measured using a shorter path length, but minimum
detection levels increase in proportion. In actual practice, gas measuring cells lengths are
specified based on the particular application and accuracy requirements.
F |NTERFERENCES
Several gases that are typically found in flue gas absorb light in the lower UV region and
present a potential for interfering with the measurement of NH3 and NO. However,
experiments were conducted which demonstrated that at typical flue gas concentrations,
N02, CO, C02, 02, and H20 did not interfere with the measurement of NO and NH3< The
most predominant interference is SO^ which, depending upon the concentration, can be
accounted for using spectral subtraction which has been described previously (Durham et
al., 1989). The maximum S02 concentration that can be accurately subtracted from the
absorbance spectrum depends upon the length of the gas cell. For example in a 0.9 meter
cell, the maximum concentration of S02 is 80 ppm. If the cell is reduced to 4 cm, then the
maximum concentration increases to 1800 ppm S02. However, with the smaller cell the
minimal detection limit for NH3 is increased to 13 ppm. Therefore, a gas cell needs to be
selected for the specific application.
IV. FIELD EVALUATIONS
A. GAS TURBINE WITH SCR
I
The ADA Analyzer was used to evaluate the De-NOx system of a gas-fired co-generation
facility. At this site, the Analyzer was evaluated as both an in-situ and an extractive
instrument. The in-situ instrument avoids sample biasing and minimizes the operating and
maintenance requirements. The extractive version is designed for traversing ducts
downstream of the NOx control system to optimize the ammonia injection configuration.
At this site, ammonia is injected upstream of a selective catalytic reduction (SCR) bed to
control the NOx emissions. The plant did not have an ammonia detector but did monitor the
concentration of NOx at the inlet and outlet of the SCR and measured the quantity of NH3
that was injected. The target NOx emission from the facility was 18 ppm.
7A-25
-------
Verification of the Accuracy of the Instrument
The measurement accuracy of the analyzer was determined by comparing instrument
emission measurements against a standard wet chemical technique. This manual technique
involves extracting a sample of the flue gas from the stack and bubbling it through an acidic
solution which collects the ammonia. The solution is then analyzed in a laboratory using a
selective ion electrode to determine the quantity of NH3 collected. Although this technique is
very manpower intensive, accurate measurements can be obtained if the procedures are
followed carefully. An experienced third party testing firm was contracted to perform the wet
chemical measurements.
Several wet chemical tests were conducted while the analyzer continuously measured NH3
concentrations. The analyzer was used in-situ, while wet chemical tests were conducted
from a different, neighboring port on the duct In spite of the fact that the measurements
were made at different points in the stack, there is excellent agreement between the two
methods. Figure 8 shows a comparison of the ammonia concentrations measured by the
continuous analyzer and the manual method. The straight line represents a 1:1 correlation.
The numbers inside the data points are the ports where the extractive measurements were
made. The ADA instrument was operated at a port midway between the two orthogonal
ports 1 and 4. The different ammonia levels in the stack were achieved when the facility
operator manually adjusted the ammonia injection rate. This data demonstrates that the
instrument is capable of accurately measuring the concentration of ammonia in a flue gas
stream.
Figure 8. Comparison of NH3 measurements using the ADA In-Situ monitor and extractive
wet chemical analysis at a co-generation facility.
h-
0
1 1 1 : i I ; I 1 1 I t r - i I 1 « ' 1 I { II | 1 I i "¦ I i r *
5 10 1S 20- 25 30
NHj CONCENTRATION (pprr) SY WET CHEMICAL ANALYSIS
7A-26
-------
Continuous Operation
The instrument was operated on a 24-hour per day basis during the test week, Algorithms
were written to eliminate any detrimental effects due to fouling of the lenses or mirror. During
the operation of the instrument some fouling of the mirror did occur due to the deterioration
of the purge blower. This resulted in a reduced magnitude of light detected by the
photodiode array. However, the algorithms operated as designed to account for loss of light
level, and the fouling had no effects on the measurements of NHg and NO concentrations.
Figure 9a shows a plot of the data obtained during a 24-hour period. The trends in the NH3
and NO measurements show a gradual decline in the NO concentration while the ammonia
slip is increasing. Whenever a sharp change in NO level occurs, there is a corresponding
change in the opposite direction for NH3. The ammonia injection rate is plotted in Figure 9b.
As can be seen there is a strong correlation between the ammonia injection rate and the
ammonia slip. This data indicates the variability that occurs in even a stable combustion
system such as the gas turbine combustor.
Evaluation of the SCR System
The data obtained during the continuous in-situ measurements were reduced to determine
the relationship between the NO level and the NH, slip. These data, which are plotted in
•Figure 10, provide very valuable information relative to the performance of an SCR system. It
can be seen that for higher concentrations of NO there is very little slip and the amount of slip
increases as the NO is reduced. However, at some point any further decrease in NO can
only be achieved with a significant increase in ammonia slip.
This data is extremely important relative to the cost-effective operation of an SCR and the
resulting emissions. If the facility is operating under a permit that specifies only a maximum
NO concentration, without considering the ammonia slip, the minimum level of emissions will
not be obtained. In this example, in order to obtain a 2 ppm reduction in NO from 19 to 17
ppm, the NH3 slip will increase by 20 ppm. It would be more desirable to operate at the knee
of this curve to minimize the total release of pollutants.
Operating at this point would also make economic sense. At an ammonia slip level of 25
ppm, half the injected ammonia is going up the stack unreacted. This means that the cost of
the ammonia is double what it would be if the system: were controlled with the slip as a
parameter. This data also demonstrates the importance of a continuous ammonia slip
monitor. Since the performance of the catalyst in the SCR is going to change over time, the
continuous monitoring of the flue gas can be used to identify the optimum operating
conditions at all times.
Evaluation of the Extractive Analyzer
The analyzer was also used in an extractive mode in order to measure gas concentration
gradients in the system. A probe was used to draw samples of flue gas from discreet points
across the diameter of the stack and into the analyzer. Since there was no access
immediately downstream of the catalyst, a traverse was made at the stack. The traverse was
7A-27
-------
Figure 9a. Continuous NH3 and NO measurements from a co-generation facility.
Figure 9b. Ammonia injection rates during emissions measurements.
7A-28
-------
30
25'
,20'
Q.
Q.
V)
n
5 10
0 I I I I I 1 I i ¦¦ r ) i ' i t I 1 I I I I t i •»¦¦ I I I I I ' I r i ' 1"< I ' I1 i ' ' I
0 5 10 15 2G 25 30 35 40 45
50
OUTLET NO CONCENTRATION (ppm)
Figure 10. Nitric oxide emissions as a function of ammonia slip.
made parallel to the ammonia injection grid. The results presented in Figure 11 show the
presence of strong gradients in both NO and NH3 concentrations across the stack. The
higher levels of NO correspond with lower levels of NH3. Both the gradients and the inverse
relationship between NO and NH3 are due to an improper balancing of the ammonia
injection valves. This shows the usefulness of the extractive instrument in providing a means
to optimize the ammonia injection system.
B. COAL-FIRED FLUIDIZED BED WITH SNR
The ADA Continuous Ammonia Analyzer was field tested at a 49.5-MW coal-fired circulating
fluidized bed co-generation facility. The plant injects ammonia into the primary cyclone for
control of NO . The on-site CEM system incorporates a chemiluminescent instrument to
beasure both NH3 and NOx levels using a thermal converter for ammonia. Flue gas samples
are withdrawn from the center of the stack (approximately 100 feet above ground level) via a
heated in-situ probe. The flue gas is pulled down approximately 100 feet of heated sample
line to an instrument enclosure. Moisture is removed from the flue gas sample before it
entered the NO^ analyzer. In the NH, measurement mode, a solenoid valve is activated
periodicafly, forcing the flue gas through a thermal converter which converts the NH~ to NO.
The signal generated from the flue gas that by-passes the thermal converter is subtracted
from the signal generated when the flue gas passes through the thermal converter to obtain
the NH3 concentration present in the sample.
7A-29
-------
Injection
Valves
Figure 11, Measured concentration gradients for NH3 and NO.
The field test program was performed to determine the accuracy of the ADA Continuous
Ammonia Analyzer for measuring NH3, S02 and NO in a flue gas environment containing low
levels (5-40 ppm) of S02. As was done in the previous field study, the NH3 concentrations
measured by the ADA monitor were compared with those obtained using the standard
ammonia wet chemical technique performed by a third party. In addition, a comparison
between the ADA ammonia monitor and the chemiluminescent ammonia monitor determined
how well the two techniques agreed with each other and with the standard wet chemical
method.
Simultaneous NH3 measurements were taken using the wet chemical method, the ADA
ammonia monitor, and the chemiluminescent ammonia monitor. The chemiluminescent
ammonia monitor took samples from the center of the stack through a heated sample probe.
The ADA ammonia monitor measured NH3 directly in the stack through a port positioned at a
90° angle from the chemiluminescent monitor sample probe. The wet ammonia
measurements were performed by positioning the wet ammonia sample probe adjacent to
the ADA in-situ probe. This was done by placing the sample probe through the sample port
90° from the ADA monitor (180° from the chemiluminescent ammonia monitor) and then
bending the sample line to physically contact the ADA in-situ probe.
Figure 12 shows the comparison of the NH3 concentrations measured by the ADA ammonia
monitor, the chemiluminescent ammonia monitor, and the wet chemical ammonia method.
All data were corrected for 7.8% moisture and 5% oxygen. These conditions were measured
in the stack at the time of sampling. The sample points are averages taken over the wet
ammonia method sampling time. Measurements of different NH3 levels were attempted
7A-30
-------
when the facility operators manually adjusted the ammonia injection rate, However, the
vaporizers were not functioning properly at the time of the test, and the ammonia control
valves were opened fully.
As shown in Figure 12, the wet chemical and the ADA methods agree well. This test also
shows the effectiveness of the ADA processing package in eliminating the interfering effects
of S02 on the NH3 measurements. The chemiluminescent ammonia monitor response was
approximately 3-5 times higher than the standard wet chemical method. This inaccurate
measurement of the ammonia slip could result in the injection of an insufficient quantity of
ammonia to react with NOx.
Time (Hours)
Figure 12. Ammonia slip measurements on a coal-fired fluidized bed boiler using three
methods.
C. REFINERY BOILER WITH SNR
The ADA analyzer was used to measure NH3 and NO emissions from a thermal De-NOx
system used on a refinery boiler gas stream. Ammonia gas was injected into the hot exhaust
gas from a furnace in order to effect the NOx reduction reaction. The gas stream contained
several hundred parts per million SO„. Therefore, a gas measuring path length was chosen
to most effectively accommodate the flue gas S02 content, while still providing the necessary
degree of accuracy for NH3 and NO measurements.
7A-31
-------
Accuracy Determination
The analyzer was again used in both an in-situ and extractive mode to gather data. The
facility performed several wet chemical NH3 evaluations while the analyzer operated in-situ.
These results compared as follows:
These results indicate good agreement between the methods, especially given the rapid
short-term changes in NHg emission levels observed in the flue gas stream using the real-
time analyzer,
De-NOx System Evaluation
Ammonia slip and NO emissions data were collected as De-NOx system variables were
adjusted. Figure 13 shows the relationship between NO emissions and NH3 slip measured
over a range of operating conditions. Because of the proprietary nature of the information,
the data are plotted in relative concentration terms. This figure has a very similar shape as
the plot obtained from the SCR tests in that there is a point of diminishing returns relative to
the amount of ammonia injected. This is represented by the point where only minimal
reduction in the concentration of NO is obtained at the expense of significant Increases in
ammonia slip . Figure 14 shows the relationship between NH- slip and NH-j injection rates.
Data such as these, when collected in combination with other process information, can
produce a significant data base for use in characterizing a De-NOx system, and for
troubleshooting purposes.
The data presented on the De-NOx system evaluation were collected in only a few days of
testing. These results demonstrate the ability of a real-time analyzer for effectively
characterizing emissions from a full-size control system.
D. UTILITY BOILER WITH UREA SNR
The final field test program was conducted during a demonstration of urea injection into a
utility boiler. This program was conducted during October to December, 1990 and is
described In the paper by Abele (1991) which is presented at the 1991 NOx Control
Symposium. During this program, the instrument was successfully operated during the test
program. The calibration of the instrument was checked at the beginning and end of the
program. After nearly two months of operation, the calibration constants had drifted less
than 2%.
Wet Chemistry
ADA Analyzer
51 ppm
171 ppm
60 ppm
225 ppm
7A-32
-------
Figure 13. Nitric oxide emissions as a function of ammonia slip at a refinery boiler.
12-
10-
¦i ¦ I i i i i
0 2 4 6 8 10 12
Ammonia Injection Rate, Relative
Figure 14, Relationship between ammonia slip and ammonia injection rate for refinery SNR
system.
7A-33
-------
V. STATUS
ADA continues to provide testing services and analyzers for evaluations of De-NO^ systems.
ADA has been working toward commercialization of the analyzer technology with instrument
manufacturers. ADA will be participating in a round-robin performance evaluation of
commercially available analyzers with regulatory agency involvement beginning in March.
ADA highly endorses such programs and will report results at upcoming meetings.
VI. REFERENCES
Durham, M.D., T.G. Ebner, M.R. Burkhardt, and F.J. Sagan (1989). "Development of an
Ammonia Slip Monitor for Process Control of NH~ Based NOx Control Technologies",
presented at the AWMA International Specialty conference on Continuous Emission
Monitoring-Present and Future Applications, Chicago, !L November 12-15.
Abele, A. (1991). "Performance of Urea NOx Reduction System on Utility Boilers", EPRI-EPA
1991 Joint Symposium on Stationary Combustion NOx Control, Washington D.C.,
March 25-28
7A-34
-------
ONTARIO HYDRO'S SONOX PROCESS FOR
CONTROLLING ACID GAS EMISSIONS
R, Manga! and M.S. Mozes
Ontario Hydro Research Division
800 Kipling Avenue
Toronto, Ontario
MiZ 5S4 Canada
and
P.L, Feidman and K.S. Kumar
R-C Environmental Services and Technologies
US Highway 22 West
Branchburg, New Jersey
USA 08876
1A-35
-------
-------
ABSTRACT
An in-fumace slurry injection process for the simultaneous control of SOj and NO, from power plant flue gases has been
developed at Ontario Hydro's 640 MJ/h (0.6 x 10" BTU/h) Combustion Research Facility, The process known as SONOX
involves the injection of an aqueous slurry of a calcium-based sorbent such as limestone, dolomite, hydrated lime, etc and
a nitrogen-based additive into the furnace at temperatures ranging between 900 to 1350°C, Coals varying in sulphur
content from 0,54 to 2.8% with NO, emission levels of 450-620 ppm were studied. Operating parameters have been
optimized far maximum SO, and NO, capture. Under optimized operating conditions the technique removes up to 85%
of the S02 and effective NO, removal is 63-80%. The specific removal levels obtained depend upon the type of sorbent
and nitrogen-based additive, temperature, stoic biometry and coal. The effluent gas stream has been characterized for NH,t
HCN and N,0. The solid waste produced is composed of fly ash, CaSO, and CaO which can be collected by the ESP.
Due to the high dust loading that results from the process, the ESP performance deteriorates somewhat. A levelized cost
estimate indicates that a SONOX system is about half the cost of a wet FGD system to own and operate. Negotiations
are in progress to demonstrate this process on full scale boilers.
INTRODUCTION
In December 1985, the Ontario government announced a more stringent acid gas emission policy limiting Ontario
industries in atmospheric emission of SOj and NO,. Ontario Hydro's limits were reduced to 430,000 tonnes/year starting
in 1986 and to 215,000 tonnes/year starting in 1994. This regulation is challenging in thai Ontario Hydro must stay below
the regulated tonnage limit regardless of changes in the demand for energy or in other forms of generation. Although the
regulation limits the amount of SO, emissions, the level of NO, emissions Is not specifically regulated and Ontario Hydro
is free to trade between SOa and NO, as long as the aggregate emissions of the two (SO, and NOJ does not exceed
215,000 tonnes/year and no more than 175,000 tonnes/year may be SOj(l,2). Specific NO, legislation is now being
negotiated between the Federal and Provincial Ministers of the Environment.
Consequently, Ontario Hydro embarked on a program to curtail acid gas emissions from its coal burning plants. This
program was initiated to meet the above mentioned regulations.
Several options are being considered to reduce both SO, and NO,. In the oik of SO,, some options include: sorbent
injection processes, burning low sulphur coals with flue gas conditioning, wet flue gas desulphurization and the limestone
dual alkali process. Ontario Hydro is committed to two scrubbers being in operation at the beginning of 1994. For NO,
control, the options can be classified as non-retrofit and retrofit technologies. Non-retrofit options would be to reduce
NO, emissions by installing fossil replacement generation that has lower NO, emission rates than are currently generated
by existing stations and to reduce coal generation. Burning natural gas is an example. Retrofit options include: low NO,
burners, selective catalytic reduction and selective non-catalytic NO, reduction fffocesses-
-------
Of the options considered to meet the above regulations in-furnace sorbent injection and selective non-catalytic NO.
reduction processes were investigated extensively at Ontario Hydro's 640 MJ/h Combusuon Research Facility. As a result
Ontario Hydro's SONOX process which injects a calcium-based sorbent slurry and an additive to simultaneously abate
SO, and NO, was developed and is the subject of this paper.
The SONOX process is an in-fumace injection technique of an aqueous slimy of a calcium-based sorbent and a soluble
additive injected ai temperatures ranging between 900 and 1350°C. The calcium-based sorbent reacts with SO) and the
additive reacts with NO,. The furnace which serves as the chemical reactor provides sufficient residence time and
favourable temperature few the reactions. The following reactions represent globally, the SOj/NO, (SONOX) removal
paths:
CaC0} CaO + COj
CaO + SOs -i- 1/2 O, > CaSO.
NO + Reagent (Additive) > N, ~ HjO
The technique provides excellent distribution and mixing with the flue gas for the above reactions to be efficiently
completed(3). A schematic of the process is shown in Figure la. The process steps can be visualized as follows:
• Atomization of Ca sorbent and additive;
• Water droplet evaporation;
Panicle disintegration for the Ca sorbent and thermal cracking of the additive;
Calcination of the Ca sorbent;
Development of reactive sorbent and additive (CaO and NHj);
• SOj and NO, capture.
The above steps are illustrated in Figure lb for limestone.
EXPERIMENTAL
Combustion Research Facility
The study was conducted at Ontario Hydro's Combustion Research Facility (CRF) designed for a maximum coal feed rate
of about 20 kg/h (44 lb/h) U.S. bituminous coal at a firing rate of 640 Ml/h (0.6 * 10s BTUAi) (Figure 2). The furnace
is a refractory-lined cylindrical chamber, fully equipped for monitoring gas and wall temperatures. There are multiple
ports for flame observation and for insertion of solid sampling probes. There are also probes to determine slagging and
fouling rates. The pulverized coal is delivered down-draft to the burner with the combustion air which can be electrically
preheated to temperatures up to 350°C (662®F). Gas bumen on each side of the coal burner are used to heat the furnace
to operating temperatures before beginning to feed the coal.
The coal burner, designed and constructed by Research Division staff, is equipped with a vortex generator and four air
vanes to assure good mixing and adequate residence time of the fuel-air mixture in the combusuon zone. The combustion
gases in the furnace are cooled by water and/or air circulating in the cylindrical Inconel jacket around the furnace. This
cooling system is equipped with temperature sensors and flow meters to control furnace quenching rates.
The combustion gases leaving the furnace are further cooled by a series of air-cooled heat exchangers prior to entering
the resistivity probe housing and ESP. The ESP consists of a cubic stainless steel chamber, and is equipped with two sets
of interchangeable cells. One set has an 11-plate electrode with 2.5 cm (1 In) spacing, the other a 5-plate electrode with
5 cm (2 in) spacing. The design specific collection areas (SCA, m'/mVs) for the two sets of cells are 39 (0.2 ftVcfm) aid
17 (0.09 ftVcfm) respectively for baseline firing conditions using s high volatile U.S. bituminous coal.
7A-38
-------
The CRF instrumentation permits systems temperatures, and flue gas composition (O,, CO,, CO, SO, and NO.) to be
monitored continuously. Gas temperatures in the furnace are measured with a suction pyrometer and flame temperatures
with ail optica! pyrometer. Row rates and pressures are measured by (low meters and manometers. All measuring and
monitoring systems are linked to a computerized data acquisition system. Particulate mass loading in the flue gas before
and after the ESP is measured with an isokinetic sampling system. In-situ resistivity is measured with a point-plane
resistivity probe situated in the resistivity probe housing and particle size distribution of the ash is measured with a Pilot
Mark 3 Cascade Impactor.
A more complete description of the facility is given elsewhere/4/.
SONOX Hardware
A general overview of the hardware used is shown in Figure 3. A positive displacement pump pumps the slurry/additive
mixture from a continuously slimed mixing tank under a pressure of 650 to 720 kPa. Recirculation and a static mixer
upstream of the furnace kept the particles in suspension and prevented settling. A small metering pump delivered the
sluny/additive mixture to the atomizer through which Tine droplets were injected into the flue gas stream.
Injection was in the middle of the furnace through a rwin-fluid high pressure nozzle (5 or 3 mm) with an internal mixing
chamber, shown in Figure 4, Operating pressures range between 40 to 60 psig. The stainless steel nozzle was purchased
from Turbotak Inc. The MMD of the droplets was about 12 pm for the 5 mm nozzle and approximately 6 Jim for the
3 mm nozzle. The nozzle was equipped with a cooling jacket which was necessary to avoid evaporation of the water and
hence drying of the slurry causing deposition of particles.
Fuels and Sor bents
Several coals ranging in sulphur content from 0.54% to 2.8% were used with the SONOX technology. These coals
include a 0.54% beneftciated western Canadian coal, supplied by Unocal Canada, a 1.1% S coal resulting from a blend
of western Canadian and eastern U.S., a 1.7% S eastern U.S. bituminous and a 2.8% S coal from Nova Scotia, Canada.
The proximate and ultimate analyses of the coals are shown in Table 1.
Sorbenis used include two local calcitic limestones from Ontario, namely BeachviUe and PL Anne. A Beachville hydrated
lime was also studied. Also from Ontario, a dolomitic sione was used supplied by E.C. King. A Mosher limestone from
Nova Scotia was used with the Nova Scoria coal. The chemical and physical properties of the raw sorbenis are shown
in Table 2, These analyses were performed by ORTECH International • a research foundation in the province of Ontario.
Of the additives used to remove NO,, the three best are described in this paper and are labelled A, B and C.
Procedures
After steady state was achieved with the baseline coal, injection of the sorbent slurry/addinve into the middle of the
furnace was initialled. Temperature-time and radial profiles simulating Lakeview and Umbton TGS were studied.
Changing the quenching rale allowed ihe effect of residence time to be studied. Data collected during each test include
system temperatures, and pressures, slurry/additive-feed rales and stoichiomeuy, flue gas constituents concentrauons (CO,,
03. CO, SO, and NOJ, in-situ ash resistivities and particle size distribution. Coal, sorbents feed and fly ash samples were
collected during the tests. Analysis of samples include chemical composition and panicle size distribution. In selected
runs, NH,. NjO and HCN were monitored. Calcines and sulpha ted calcines were analyzed for CaO, CaCOj and CaSO,
content. Porosity, mass median diameter and BET surface area of some samples were also determined. The analytical
methods used are described in referenced).
7A-39
-------
RESULTS AND DISCUSSION
The most important parameters thai were found to affect process performance (SO, and NO, capture) are classified under
the following categories:
Sorbeni/Additive
Chemical and physical characteristics;
• Concentration; and
• Addition rate (stoichiometric ratio).
Injection Parameters
« Mode of injection;
• Droplet size, distribution and mixing with the flue gas;
• Temperature; and
• Residence time.
Coal
• SO, and NO, concentration.
These parameters were optimized for maximum SOjMO, capture on the pilot furnace. It is important however, to address
some of the advantages of the SONOX process and to mention thai negotiations are in progress to demonstrate SONOX
on the full scale. Some of the advantages are:
SONOX provides a low cost solution to the removal of acid gas from flue gases;,
¦ SONOX is suitable for retrofit application;
• SONOX is applicable to coals with various SO, and NO, levels; and
» SONOX requires short lead time for installation.
Sorbents Comparison
For SO, control using alkaline-based sorbents, sorbent composition and physical properties ire important factors in
determining overall capture performance^,6,7,8,9). Significant variability in the reactivity of the various sorbents has been
observed and it was recognized that surface area and porosity play a vital role in sorbent reactivity. Figure S illustrates
the effect of porosity on sulphur capture for various sorbents, Pl Anne limestone with an initial porosity of 55% gave
significantly higher removal than Beachville limestone with an initial porosity of 17% (70% removal for Pi Anne
compared to 55% for Beachville) at a Ca/S ratio of 3.0, The Nova Scotia limestone slurry was used with the Nova Scotia
coal. Thus a direct comparison of process performance between this sorbent and the local calcitic stones was not possible.
Data indicate, however, that similar sulphur capture can be obtained with Nova Scolia limestone (porosity 57%) and Lhe
Pi. Anne limestone (porosity 53%) even if they are used for two different coals (2.8% and 1,7% sulphur content).
Since the additives for NO.iemoval are water soluble, only the effect of concentration and chemical composition were
evaluated.
Effect of Injection Parameters
Injection parameters thai influence SOj/NO, capture include: atomizer type, injector location, atomizing air pressure, and
particle size distribution or mass median diameter (MMD) of (he atomized droplets. High atomization air pressure
improves the quality of atomization and promotes an early release of the sorbeni/addirive to engage in the suiphation/NO,
reduction reactions. It also increases the discharge momentum of the droplets leading to enhanced penetration and mixing
with the flue gas stream. These experiments were conducted with the Turbotak nozzle.
The effect of atomizing air pressure on droplet size is illustrated for limestone slurry in Figure 6. SO, capture was found
to be a function of droplet size distribution, and quality of atomization and mixing with the flue gas. The best mixing
was observed while spraying a 40% aqueous Pl Anne limestone slurry into the furnace cocunently at an injection location
which was close to the flame zone where increased turbulence exists. Increasing ilie atomizing pressure from 40 psig to
70 psig reduced droplet MMD from 12 pm to 6 ym and improved SOj capture from about 62% to 70% at Ca/S ratio of
3.0.
7A-40
-------
Effect of Temperature and Injection Mode
(a) Slurry Injection for SO, Control
The effect of temperature on SO, capture was evaluated for the different sorbents (Pi. Anne limestone, BeachviUe
limeitone, BeachviUe hydrated lime, Nova Scotia limestone and E.C. King dolomite; while burning the 1.7% S
eastern U.S. coal, the 1.1% S eastern U S/westem Canadian coal blend and the 2.8% S Nova Scotia coal. The
results are shown in Figure 7a. Cocunent injection gave higher SO. capture than the countercurrent mode and the
optimum injection temperature for the coeurrent mode was found to be 1200°C. The comparative performance for
the different coal/sorbent pairs was done with the Turbotak 3 mm nozzle as is illustrated in Figure 7a.
The highest capture, 85% was observed with hydrated lime to be followed by 83% with the E.C. King dolomite, 65-
70% with the porous Pi. Anne limestone and 55% with the BeachviUe limestone at a Ca/S ratio of 3.0 while burning
the 1.7% S U.S. coal. Under ihe same operating conditions, using the same limestone, SOj capture from the western
Canadian/U.S. coal blend was slightly less than from the U.S. coal as is shown in Figure 7a. Injecting the Pt Anne
limestone with the high sulphur Nova Scotia coal (2.8%) resulted in 76% S02 removal at a Ca/S ratio of 3.0.
Sulphur removal efficiency was 58 to 63% using a 2.8% S Nova Scotia coal with the porous Nova Scotia limestone,
at a Ca/S ratio of 2.2. (Because of the presence of grits with this limestone and limited pump capacity, this was the
highest rate at which this sorbent could be fed to the furnace,} However, this compares favourably well with the
60% capture obtained at a Ca to S ratio of 2.5. using the porous Pl Anne limestone with the 1.7% S U.S. coal.
Replacing 5% of the calcium from the Pt Anne limestone by an equivalent amount of dolomite (dolomite doping)
resulted in 80% SOj capture, up by 10% from what was achieved with pure Pt. Anne limestone.
(b) Additive Injection for NO. Control
The effect of temperature on NO, removal is shown for the three additives. A, B and C, in Figure 7b while they were
being injected cocurremJy only. The data indicate that additives A and B show a common optimum at around
1100°C, while additive C shows a "flat" profile between 975 to 1100°C. At 1 lOO'C, additives A and B removed
90 and 84% NO, respectively, while between 975 to 110G°C additive C removed 77 to 80% NO,. This can be quite
a desirable feature for full scale boilers where load is constantly changing resulting in changing temperatures. The
reason for additive C behaving differently from the others is not fully understood and further studies may be able
to provide an explanation.
Sl'P Gases
The concentration of nitrogen containing species such as ammonia (NH,), hydrogen cyanide (HCN) and nitrous oxide
(N,0) in the slip gases during additive injection has been investigated.
Results indicate that NH, slippage for additive A ranged between 7 - 26 ppm and for additive C up to 49 ppm, HCN
was found to be between 3-9 ppm. With no NO, removal additive present the N,0 produced ranged from 10-25
ppm at an initial NO, concentration of - 550 ppm. Decomposition of additive A has a side reaction which could
lead to the formation of NjO. The amount of NsO produced when additive A was injected ranged from 59 • 150
ppm at 1100°C and an additive/NO stoichiometric ratio of 2.0. These data demonstrate thai 11 to 27% of the NO,
is convened to N,0 thus the effective NO, removal for additive A is 63 to 80% instead of 90%. It was found that
NjO formation is affected by injection temperature, additive stoichiometry and NO, level in the flue gas. More
studies are required to optimize operating conditions for minimum conversion of NO, to N:0. Additives B and C
showed an increase in N,0 levels of 5 - 15 ppm from the baseline.
(c) SONOX Process for SOJNQ. Control
Simultaneous capture of SO, and NO, was undertaken by adding additive A to an aqueous slurry of Pl Anne
limestone and dolomite doped Pl Anne limestone while bunting the 1.7% S eastern U.S. bituminous coal with an
initial SO, concentration of 1350 • 1400 ppm and NO, concentration of 550 ppm. The results are illustrated in
Figure 7c for the following optimized conditions:
7A-41
-------
40% aqueous calcium-based slurry (Pi. Anne and dolomite doped)
Ca/S ratio = 3.0
Additive A concentration of 13 5% (w/w) in slurry
Additive/NO mole ratio ¦ 2.0
Injection mode: cocurrent
Nozzle: Turbotak 3 mm, droplet size = 6 jim MMD
The graph of Figure 7c shows the effect of temperature on SOj/NO, capture for additive A combined with Pl Anne
and dolomite doped Pl Anne slurries. S02 capture for the Pl Anne slurry and additive A at the optimun
temperature of 1200°C is 70% and nominal NO, capture is 90%. With the 5% dolomite doped Pt Anne slurry and
additive A, S03 capture is 80% and nominal NO, capture is still 90%,
Effect of Siolcbiometry
(a) Ca/S Ratio for SO, Control
The effect of Ca/S ratios on sulphur capture and sorbent utilization was studied while using the Pi. Anne (porous)
limestone with ihe U.S. coal, the U.S.-western Canadian blend and the Nova Scotia coal. Hie Beachville (non-
porous) limestone, dolomite and hydra ted lime were studied only with the U.S. coal. Injecting the Pi. Anne
limestone wiih the U.S. coal was done at 1200°C and BOOT while all other coal-sorbent combinations were done
at 1200CC. In all cases injection took place cocurrenily using a 40% aqueous slurry. Ca/S ratios varied from 1.5
to 3.0 and the furnace quenching rate was held constant at 500°C/s. The results are shown in Figure 8a. Sulphur
capture and sorbent utilization are plotted vs Ca/S ratios for the various sorbent-coal pairs.
Sulphur capture decreases, but sorbent utilization increases with decreasing Ca/S ratios for all coal-sorbent pairs
tested. At the optimum temperature of 1200°C, dolomite and hydrated lime showed the highest capture. Dolomite
removed 78% of the SO, at a Ca/S ratio of 1.5 for a calcium utilization of 52%, while hydrated lime removed 75%
and 83% SOi at Ca/S ratios of 1.5 and 2.5 respectively. Sorbent utilization was 50 and 33%.
At all Ca/S ratios, the more porous Pl Anne limestone outperformed the less porous Beachville limestone both in
terms of sulphur capture and sorbent utilization. At 1200°C using the U.S. coal with the Pi Anne limestone at a
Ca/S ratio of 3.0, sulphur capture and sorbent utilization were 65 to 70% and 22 to 23% respectively as compared
to 55% and 18% with the Beachville limestone. Using the Pl Anne limestone with the high sulphur Nova Scoua
coal, sulphur capture at a ratio of 2.0 is 72% and at a ratio of 3,0 is 76%. Under the same operating conditions at
a Ca/S ratio of 1.5 sulphur capture for the Pt. Anne and Beachville limestones dropped to 50 and 31 respectively,
but utilization increased to 33 and 21%. With the Nova Scotia coal and Pt. Anne limestone at a Ca/S = 1.5, sulphur
capture is 64% with a sorbent utilization of 43%.
(b) Additive/NO Ratio for NO. Control
The effect of additive normalized stoichiometric ratio, NSR (NSR ¦ moles of additive injected to the theoretical
moles required to remove 100% NO,) for the three additives. A, B and C, was studied while burning the eastern U.S.
bituminous coal. In all cases injection of each additive took place cocurrenily at 1100°C while NSR was varied from
1.2 to 3.0. The concentrations of the additive solutions were as follows: A -13.5% by weight, B - 5.6% by weight,
and C - 16.1% by weight. The baseline NO, from the U.S. coal was 500-550 ppm.
NO, capture is illustrated in Figure 8b. NO, capture by A and C increases with increasing NSR up to 1.7 to a
maximum of 90% (nominal) and 80% respectively, and by B up to NSR = 2.0 to a maximum of 84%. Reagent
utilization drops with increased stoichiometry for all three additives. The best utilization with A was 55-56% at an
NSR of 1.2 to 1.5, with B, 56% at a NSR of 1.0 and with C, 41 to 42% u a NSR of 1.5 to 1.7.
7A-42
-------
(c) Ca/S ¦ Add/NO for SONOX
The effect of Ca/S mole ratio and additive/NO normalized stoichiometric ratio was studied by injecting the 5%
dolomite doped Pt. Anne limestone combined with additive A. The coal burned was the 1.7% S eastern U.S.
bituminous and injection was carried out cocurrently at the optimum temperature of 1200°C. The results in Figure 8e
sho«. that at a Ca/S ratio of 3.0. 80% SOs capture is achieved and at an additive to NO stoichiometric ratio of 11
to 2.0, a nominal NO, capture of 90% is achieved.
Low Sulphur Coal Application
The development of the SONOX technology has been earned out mainly on a medium S (1.7%) eastern U.S. bituminous
coal and a high S (2.8%) coal from Nova Scotia with SOj emissions of 1350-1400 and 1700-1725 ppm and NO, emissions
of 550 and 450-520 ppm respectively.
The effectiveness of the SONOX process was also demonstrated on a western Canadian Obed coal sample, prepared by
Unocal Canada. The sulphur content of the coaJ is 0.54% with initial SO, concentration of 349 ppm. NO. level initially
measured 620 ppm. A 40% aqueous dolomite doped Pi. Anne limestone slurry (10% dolomite) with additive A was
injected cocurrently in the pilot furnace and the effects of injection temperature and stoichiometry observed. The results
are illustrated in Figure 9.
In Figure 9a, SO^NO, capture as a function of injection temperature is plotted for constant stoichiometrics. Ca/S = 3.0
and additive/NO normalized stoichiometric ratio of 3,0. The results indicate that the optimum temperature was around
1100T for both pollutants with SOj removal being 81% and nominal NO, removal being 89%.
The effects of Ca/S ratio and additive/NO stoichiometric ratio is shown in Figure 9b. Removal of both acid gas
components increases with increasing Ca/S and add/NO ratios. Optimum Ca/S ratio for S02 is 2.0 to 2.5 and for NO,,
optimum add/NO stoichiometry is 2.0. Utilization of both sorbents improves with decreasing addition ratios as is shown
in Figure 9b. Under optimized operating conditions (injection temperature * 1100°C, Ca/S = 2.0-2.5 and add/NO = 2.0)
80% SO, and 85% NO, was removed from the flue gas stream. Sorbents utilization and 32-40% and 43% respectively.
These results indicate that the SONOX technology is applicable to coals with various levels of sulphur content and NO,
levels.
Impact on Ash Characteristics, Collectibility and Deposition
The SONOX process produces increased amounts of waste composed mainly of CaSO„ unreacted CaO and fly ash. Any
impact on ESP performance and deposition on the radiant section and convective passes will depend on the type and
chemical composition, the particle size distribution and amount of Ca-based sorbent injected and waste produced.
Waste Characteristics
Particle size distribution of isolrinetically collected waste samples from the injection of various limestone sorbent slurries
while buming a 1.7% S U.S. bituminous coal are compared to that of an ash sample from the same coal in Figure 10.
The mass median diameter of the baseline ash is about 8 |xm compared to 6 tint for the Pt. Anne and 9 )im for the
Beachville limestone slurry. The slightly finer waste resulting from the injection of the very fine Pi. Anne limestone is
not expected to affect panicle migration velocity and ESP collection efficiency.
In Table 3 a typical waste from slurry injection is compared to the baseline ash and to a waste from dry sorbent injection.
High levels of calcium compounds and the quantity produced must be considered for handling and disposal. CaO content
of a typical slurry waste is 302 g/kg and CaSO, content is about 220 g/kg.
Dust Electrical Resistivity and ESP Performance
The resistivity of the baseline fly ash measured in-situ with about 10 ppm SOj naturally occurring in the flue gas from
the medium sulphur eastern U.S. bituminous coal is about ltf ohm .cm. During injection of all slurries, resistivities
consistently increased by one to two orders of magnitude to 10® to lO16 ohm.cm yet the electrical operating conditions of
7A-43
-------
the ESF were not seriously affected and collection efficiencies were not seriously degraded (see Table 4) from a baseline
level of 89% during slurry injection. Dry injection on the other hand results in a resistivity of 10" ohm.cm and an 8%
drop in collection efficiency. It is possible that due to the increased moisture level in the flue gas (up to 23% relative
humidity) a thin acidic film forms around the particles and acts as a conditioning agent aiding the ESP in its performance.
Inlet mass loading to the ESP has increased 2 fold from a baseline level of 1.4 g/m1 with a resulting increase in paniculate
emissions by a factor of about 2 times during slurry injection. Thus the main problem wuh ihe SONOX process is the
high dust loading to the ESP which depends on ihe Ca/S ratio.
Slagging and Fouling Properties of the Waste
Soft deposits, which form at low temperatures and are generally characteristic of deposits found on air heaters and
economizers were observed on the furnace walls and heat exchanger surfaces. These deposits could be easily blown away
by compressed air suggesting that, conventional soot blowing equipment may suffice few full scale application of the
SONOX process.
SONOX COMMERCIALIZATION ISSUES
Electrostatic Precipitator Performance Following SONOX Application
The application of the SONOX technology in the upper furnace region affects the nature of particulate matter entering
the existing electrostatic precipitator. While the additives for NO, control do not add to the paniculate content entering
the ESP. the calcium sorbents for SO, control in the furnace result in higher paniculate loading depending on coal sulphur
content and Ca/S ratios. The precipitator inlet loading can double for most applications. In addition to the increase in
inlet paniculate loading, an increase in particulate resistivity is to be expected because of ihe uptake of SO, from the flue
gas by free lime in entrained solids. While dry sorbent injection technologies increase paniculate resistivity from about
10* ohm.cm to the 1011 levels, particulates from the slurry injection process show resistiviiy levels of about 10* - 10'a
ohm.cm due to the higher moisture content in the flue gas. Hcrice, the electrical operation of the ESP is expected to
remain unaffected and only the solids loading will have to be dealt with.
Precipitator upgrades will be needed in most cases following sorbent injection in order to handle both high loadings and
increased resisuvity. Research-Cottrell has conducted a detailed study on behalf of the Electric Power Research Institute
and proposed solutions for the precipitator degradation problems following furnace sorbent uijection(lO). The most
economical solution is humidification and subsequent evaporative cooling of flue gas to restore resistivity to pre-injection
levels. At the lower temperature, due to increased gas density, the precipitator can be operated at increased power
compared to the pre-injection level operation at 1S0°C. The precipitator can thus be operated at higher collection
efficiency to overcome the increased loading effect The humidification concept for restoring precipitator operation has
been successfully carried out at two full-scale plants by EPRI and DOE(ll). The humidificanon concept has also been
demonstrated earlier by Research-Cottrell at the pilot scale in a CONOCO supported program.
The requirements for cooling to restore ESP performance are significantly reduced for the SONOX process because of
reduced paniculate resistiviiy. We expect the stack paniculate emissions to be restored to pre-injection levels by operating
at ESP inlet gas temperature between 110 to 120°C.
Economics
Economic analysis of the SONOX technology indicates that capital costs can vary between 30 to 60 S/KWe, including
moderate precipitator upgrade costs, for combined SO, and NO, removal rates at SO to 70% each. This can be compared
to the wet FGD capital costs of 150 to 400 S/KWe, the higher cost numbers being applicable to smaller plants in the 150
MW size range. The operating costs of SONOX will be higher because of higher sorbent consumption when compared
to wet FGD. A levelizcd cost estimate, however, indicates that a SONOX system is about half the cost of a wet FGD
system to own and operate.
SONOX technology has been demonstrated at the pilot plant level. Application of the SONOX concept on a full-scale
coal-fired boiler does impact the overall system and the following questions need to be addressed to assure a successful
commercialization path:
7A-44
-------
* what is the optimum nozzle array configuration and slurry size distribution to assure proper gas-slurry
contact?
what is the optimum sulphation and NO, removal temperature window in the upper furnace region?
what is the effect of increased solids loading on boiler tube erosion?
what is the effect of increased loading and resistivity on ESP performance, and what is the best precipitator
upgrade approach?
* what is the best approach to increased solids handling of the calcium-rich ash?
Many of the answers to the above questions can be obtained from the experience with full-scale dry furnace sorbent-
injection systems already operating in Germany and other parts of Europe. Ontario Hydro/Research Cornell are currently
seeking to demonstrate the SONOX technology on a full-scale coal-fired utility boiler.
SUMMARY and conclusions
The SONOX process, an in-fumace injection of a calcium-based sorbent and a ninogen-based additive is a very efficient
way of removing SOj and NO, from flue gases. This technique facilitates improved distribution and mixing of the
sorbent/additive with the gas flow, reduces deactivation of the sorbent/additive and allows sufficient residence time at
favourable temperatures for the reaction between CaO and SOj, and NTH, and NO to be efficiently completed. The process
was developed at Ontario Hydro's 640 MJ/h (0.6 x 10* BTU/h) Combustion Research Facility. Coals studied ranged in
sulphur content from 0.54 to 2.8% and calcium sorbems used include two local calcibc limestones and a hydrated lime
from Ontario, a local dolomitic stone and a limestone from Nova Scotia. NO, levels in the flue gas ranged between 450-
620 ppm and several nitrogen-based additives were investigated. The following is a summary of the findings:
Sorbents chemical and physical properties are very important in determining the degree of SO/NO. removals.
Dolomite with a high magnesium content was very effective in removing SOs as was the case for hydrated
lime. Pl Anne limestone with an initial porosity of 55% was superior to Beachville limestone with an initial
porosity of 17%, Five percent dolomite doped Pt. Anne limestone increased SO- capture from 70% to 80%.
The nitrogen-based additives did not vary substantially in their ability to remove NO,.
• Injection parameters were found to be also very important in removing SO, ami NO,. High atomizing air
pressure which improves the quality of atomization, promotes and early release of the sorbenl/addiiive mixture
and increases the discharge momentum of the droplets, increased SO^NO, capture sipificanily. In the case
of SOj removal, increasing the atomizing air pressure from 40 to 70 psig increased SOj capture from 62 to
70% for the PL Aiute limestone.
The optimum injection temperature for SO, control was 1200°C while NO, was 1100°C. However, with the
SONOX technology (simultaneous control of both SO, and NOJ the optimum temperature was found to be
1200°C. Injecting 5% dolomite doped Pt Anne limestone slurry aid additive A at the optimum temperature
of 1200°C resulted in 80% SO, capture and nominal NO, capture is 90%. However, the effective NO,
removal is 63 to 80% because U to 27% of the NO, is converted to N,0. Hydrated lime removed up to 85%
SOa from the flue gas.
• Both SO, and NO, improves with increasing Ca/S and Add/NO stoichiometric ratios. Optimum Ca/S and
Add/NO stoichiometric ratios were found to be 2.5 to 3.0 and 1.5 to 1.7 respectively. Burning the 1,7% S
eastern U.S. bituminous coal and injecting 5% dolomite doped Pl Anne limestone at a Ca/S ratio of 3.0 and
additive A at a normalized stoichiometric ratio of 1.7 removed 80% SO, and nominally 90% NO, at the
optimum temperature of 1200°C.
7A-45
-------
SONOX was also found to be very effective for low sulphur coal application. Firing a low sulphur western
Canadian Obed coal supplied by Unocal Canada with a sulphur content of 0.54% and injecting 10% dolomite
doped Pt. Anne limestone slurry and additive A (Ca/S = 2.0-2.5 and add AvNO = t .7-2.0), removed 80% SO,
and nominally 85% NO, from the flue gas.
Panicle size distribution of the waste from the Pt. Anne slurry was slightly finer than the baseline ash. The
waste contains fly ash and calcium compounds (CaO, CaSQ,. etc) and the quantity produced must be
considered fa handling and disposal systems.
Ash resistivities increased by one to two orders of magnitude from 10* ohm cm to 10s to 1010 ohm.cm but
ESP collection efficiencies were not seriously affected. The increased level of the flue gas moisture is
believed to act as a conditioning agent.
Slagging does not appear to be a problem and the soft deposit formed on the furnace walls and heat
exchanger surfaces was easily removable.
A levelized cost estimate indicates thai a SONOX system is about half the cost of a wet FGD system to own
and operate and negotiations are in progress to demonstrate this process on the full scale.
FUTURE WORK
Studies are planned whereby other nozzles will be investigated. Other additives that have the potential for high NO,
removal while at the same time ensuring cost effectiveness of the SONOX technology will be studied. Fundamental
studies to beuer understand the SO^NO, removal paths will be undertaken. Activating and recycling waste from the
process is being investigated and utilization studies are being conducted at the University of Calgary.
More importantly, negotiations are in progress to demonstrate this process on full scale boilers.
The work described in this paper was not funded by the U.S. Environmental Protection Agency and therefore the contents
do not necessarily reflect the views of the agency and no official endorsement should be inferred.
ACKNOWLEDGEMENTS
The authors wish to express a special thanks to Ontario Hydro's New Business Ventures Division for their dedicated
efforts in conducting negotiations to commercialize the SONOX technology. In particular, we recognize the efforts of
Mr. F. Schneider and Mr. R. Kozopas.
REFERENCES
1. Taborek, R„ Dawson, C.W., and Stuan-Sheppard, IJ., "Acid Gas Emission Control - The Requirements, Technology
and Hardware". Ontario Hydro, Design and Development Division, Special Report, March 1986, 3799H.
2. Burn ham, C., "Ontario Hydro's Acid Gas Control Programs". Paper presented to the Standing Committee on General
Government. June 15, 1989.
3. Man gal, R., Mozes. M.S.. Thampi, R„ and Mac Donald, D., "In-Fumace Sortoent SlurTy Injection for SGj Control''
Presented at the Sixth Annual International Pittsburgh Coal Conference, September 25-29, 1989. Pittsburgh, Penn.
4. Mozes, M.S., Mangal, R., Thampi. R., and Michasiw, DJ.., "Pilot Studies of Limestone Injection Process Phase 1:
Simulating Lakeview TGS Quenching Rate*. Ontario Hydro Research Division Report No 86-62-K, May 30,1986.
7A-46
-------
5. Kirchgessner, D.A., Gullen, B.K., and Lorain, J.M., "Physical Parameters Governing the Reactivity of Ca(OH), with
SO,". Presented at the 1986 Joint Symposium on Dry SOj and Simultaneous SOj/NO, Control Technologies, June
2-6, 1986, Raleigh, North Carolina.
6. Dismukvs, E.G., Bentel. R., Gooch, JP., and Rakes, S.L.. "Sorbent Development and Production Studies". Presented
at the 1986 Joint Symposium on Dry SO, and Simultaneous SQyNO. Control Technologies. June 2-6, J 986, RaJeigh.
North (Carolina.
7. Siekely, J,, Evans, J.W., and John, H.Y.. "Gas Solid Reactions". New York, Academic Press, 1976.
8. Simmons, G.A., "Rate Controlling Mechanism of Sulphation", Proceedings 1986 Joint Symposium on Dry SO, and
Simultaneous SOj/NO, Control Technologies, Vol 2, EPRI CS-4966, December 1986.
9. Mozes, M.S., Mangai, R., and Thampi, R., "Sorbent Injection for SO, Control: (A) Sulphur Capture by Various
Sorbents and (B) Humidifkation. Ontario Hydro Research Report No 88-63-K. July 1988.
10. Helfritch, D.J., ei al,, "Electrosiauc Precipitator Upgrades for Furnace Sorbent Injection", EPRI Final Report
GS 6282, April 1989.
11. Altman, R.F., "Precipitation of Particles Produced by Furnace Sorbent Injection at Edgewater", 8ih Symposium on
the Transfer and Utilization of Particulate Control Technology, March 1990. San Diego, California,
7A-47
-------
Sofbent.
Slurry
-f
Additives
00 Water Drop
Evaporation
0 o
. Heat
3 *¦
D Calcination
Dry Limestone
Particles
Heat
>¦
Sulphation
Limestone Slurry
Atomization
Particle
Disintegration
•Calcination
-High Pore Structure
Development
-Sintering Process Avoided
Sulphation
b) Chemical and Physical Steps
FIGURE 1
SONOX PROCESS
7A-48
-------
© Furnace
(£) Burner AssemtVy
(J) ^ Supply
(5) Heat Exchangers
(D Filter UnH a Coal Bin
© Door To Control Room
(J) Resistivity Housing
(a) Electrostatic Precipitator
(9) To Eihaust
@ Propane Gas Control
(n) Sortwnt Injection System
(l|) Isokinetic Sampling System
(T3) Water Infection System
(m) Furnace Ouencttng pipes
(T3) Humktiltcatton Chamber
FIGURE 2
COMBUSTION RESEARCH FACILITY
-------
Air In
FIGURE 3
SONOX HARDWARE
Air In
Water In
Slurry In i
f-
Watm Out
gz
Innrml
Mixing Chamber
FIGURE 4
TURBOTAK "EXTENDED" NOZZLE
7A-50
-------
70
Ca«
Coal
Sorbent
3.0
2.2-2,5
•
e •
US
PA
~
o
US«W,Can.
P.A.
A
N.S.
N.S.
¦
a
US
B.
60
f
m
O
&
tn
S?
50
40
10 20 30 40
Porosity
50
60
FIGURE 5
SOj CAPTURE VS LIMESTONE POROSITY
7A-51
-------
Turfaotak 3 mm Nozzle,
40 % Aqueoul Slurry ol Pt. Anne Limestone
Ca/S - 3.0
Slurry Flowrale 70 ml/min
40 50 60
Atomizing Air Pressure, psig
4 6 8
Slurry Droplet Size, jim
a) Droplet Size vs Atomizing Air Pressure
b) SC>2 Capture vs Droplet Size
FKiURR 6
SO 2 CAPTURE VS SLURRY DROPLET SIZE (ATOMIZINd MEDIA -
AIR )
-------
Coal
%S
Sorbenl
Ca/S Ratio
U.S.
t.7
PI. Anna
3.0
U.S.
1.7
Beachville
3.0
U.S.
1.7
Hydralad Lima
30
—A—
U.SWC
1.1
Pi. Anne
30
—O—
Nova Scotia
2.8
Nova Scotia
2.2
—m—
Nova S«otia
2.8
Pi. Anna
3.0
^o-
U.S
1.7
Dolomite
1.5- 3.0
40 % Umestons Slurry jPi.Anno and Beachville) & Hydralad Lima
20 % Nova Scolia Umastone Slurry
Droplet Size - E jim MMD ( Turbotak 3 mm Nozzle)
80
70
-vl
>
cn
03
f
60
3
I
CM
50
O
cn
s?
40
30
Counter Current Injection
90
80
70
®
3
Q.
ra
O
rg
o
in
20
900
1000
1100 1200
Temperature. °C
1300
60
50
40
30
Cocurrent Injection
20'—
900
1000 1100 1200
Temperature. °C
1300
FIGURE 7a
S<)2 CAPTURE - EFFECT OF TEMPERATURE AND INJECTION MODE
-------
US Coal (1.7% S)
Initial NO„ Cone. - 500 - 550 ppm
1Qi
900 1000 1100 1200 1300 1400
Temperature ,°C
" Effective NOx Removal 63-80 % due to N20 formation
FIGURE 7b
NO. CAPTURE - EFFECT OF INJECTION
TEMPERATURE
US Coal (1.7% S)
Initial NO, Cone. - 500 ¦ 550 ppm
Ca/S Ratio - 3.0
900 1000 1100 1200 1300 1400
Injection Temperature, °C
" Effective NOx Removal 63-80 % due to N20 formation
FIGURE 7c
SONOX PROCESS
SO»/NO, CAPTURE • EFFECT OF INJECTION
TEMPERATURE
-------
40 % Aqueous Slurry
Co Currant Injection
Droplet MMD - 6fim
—o- 1200-C
—1300°0
BHL -Beachville Hydrated Lime
B - Beachville Limestone
PA - Pt. Anne Limestone
US -U.S.Coal
O
CTJ
N
=>
rd
O
U.S. - WC - U.S. Western Canadian Coal Blend
D - Dolomite
NSC - Nova Scotia Coal
D-US
PA-NSC
PA-US
PA - US - WC
BUS
Ca/S Ratio
Ca/S Ratio
FKJURE 8a
SO2 CAPTURE - EFFECT OF Ca/S RATIO
-------
U S Coal
Initial NO, Cone. - 500 550 ppm
Injection Temperature - 1100 °C
** Effective NO„ Removal 63-80 % due to f^ O formation
FIGURE 8b
NO. CAPTURE - EFFECT OF ADDITIVE
STOICHIOMETRY
100
5% Dolomite Doped Pt Anne Limestone/Add A
U.S. Coal (1.7% S)
Initial NO, Cone. ¦= 500 - 550 ppm
Injection Temperature - 1200 °C
100
1.0 2.0 3.0
Sorbents Stoichiometry (Ca/S and Add A/NO)
" Effective NOx Removal 63-80 % due to f^O formation
FIGURE 8c
SONOX PROCESS
S02 / NO CAPTURE - EFFECT OF SORBENTS
STOICHIOMETRY
-------
Unocal Coal. S = 0 54%
10 % Dolomite doped PI. Anne Slurry + Add A
Initial NO, conc. = 620 ppm
100
¦vj
>
I
cn
-si
<0
>
O
O" 60
z
O
v>
20
° SOj Removal
* NO, Removal
1000 1100 1200
Injection Temperature, °C
100
60
n
>
®
oc
O" 60
z _
CM
O
CO
i- 40
20
° SOj Removal
• NO, Removal
1300 1 2 3
Ca/S and Add/NO Stoichiometric Ratio
c
Q
<3
40 =>
a>
€
<8
- 30
20
a) Effect of Injection Temperature
Ca/S = 3; AdtVNO = 3
b) EHecl ol Sorbent Stoichiomelrv
Injection Temperature: 1100 °C
Effective NO* Removal 62-78 % due to N?0 formation ** Effective NO, Removal 62-78 % due to N2O formation
FIGURE 9
SONOX PROCESS
-------
FIGURE 10
PARTICLE SIZE DISTRIBUTION OF BASELINE
FLY ASH AND SLURRY WASTES
7A-58
-------
TABLE I
CHARACTERISTICS OF COALS
UNOCAL
Coal
673
50
15
5
135
122
TABLE2
CHEMICAL AND PHYSICAL PROPERTIES OF SORBENTS
Beachville
Limestone
Pt. Anne
Limestone
Mosher
Limestone
(Nova Scotia)
E.C.King
Dolomite
Beachville
Hydra ted
Lime
LijO. gAg
<003
.
0002
NajO
003
0.1
01
04
03
KzO
06
05
1.1
<10
<005
MgO
80
48
212.1
76
CaO
524.0
53S.4
5380
3009
141 0
F®2°3
01
21
6.4
2.3
15
22 0
43
6.5
10
22
Si02
<12.0
21.0
25.0
79
5.1
Ca(OHl2
7880
LOI
4340
4610
BET area, m^/g
13
29
186
06
12 6
MMD. |im
86
39
110
330
82
p. gtan ®
?-6
23
25
25
2.1
POROSITY. X
170
550
57 0
42 0
264
Proximate Analysis, g/kg
US
Nova Scdia
WOtftUS
UNOCAL
Coal
Coal
Co* Bend
CoH
Moisture
14
12
32
33
Ash
80
96
. 99
135
Volatile Mailer
357
314
321
367
Fined Carbon
548
577
548
465
Healing Value
32
31
29
28
MJAg
Ultimate Analysis, tykq
Caiton
Hydrogen
Nitrogen
Sulphur
Ash
Oxygen
US
Coal
756
57
16
17
80
74
Nova Scotia
Coal
756
50
12
2B
96
58
WCan US
Coal Blend
751
47
14
11
71
101
TAKI.E3
WASTE COMPOSITION
1.7% S US Coal with Limestone Sluiry
Ca/S = 2.5
Temperature = 1200 °C
Baseline
g/kg
Sorbent Slurry Injection Waste
g/kg
Dry Injection Waste
g"
-------
-------
PILOT PLANT TEST FOR THE NOXSO FLUE GAS TREATMENT SYSTEM
L.G Neal
Warren T. Ma
NOXSO Corporation
P.O. Box 469
Library, Pennsylvania 15129
Rita E. Bolli
Ohio Edison
76 South Main Street
Akron, Ohio 44308
7 A-61
-------
-------
ABSTRACT
The NOXSO process is a FGT system that employs a reusable sorbent. A fluidized bed of sorbent
simultaneously removes S02 and NO, from flue gas. The spent sorbent is regenerated by treatment
at high temperature with a reducing gas. Adsorbed NO, is evolved on heating the sorbent to
regeneration temperature. The concentrated stream of NO, produced is returned to the boiler with the
combustion air.
NOXSO Corporation, MK-Ferguson, W.R, Grace & Co., and Ohio Edison will conduct a pilot test
of the NOXSO system at Ohio Edison's Toronto station. The plant treats 12,000 SCFM of flue gas
containing 2300 ppm S02 and 350 ppm NO,, which is roughly 1/20 the size of a commercial module.
The paper summarizes the system design..
An additional test of the NOx recycle concept will be conducted at the Babcock & Wilcox Research
Center in Alliance, Ohio. The test apparatus is a 6 million Btu/hr small boiler simulator. It is a
scaled-down version of B&W's single cyclone front wall fired boiler design. The proposed test plan
and the data from previously reported NOx reduction tests using a pc-fired system at the Pittsburgh
Energy Technology Center are included.
7A-63
-------
INTRODUCTION
The NOXSO Process simultaneously removes S02 and NOx from the flue gas of coal-fired boilers
using a dry, regenerable sorbent. Three previous tests of the NOXSO Process have been conducted.
The first was a bench-scale program conducted at TVA's Shawnee Steam Plant for the purposes of
establishing process chemistry and kinetics, quantifying sorbent attrition rates, and establishing the
corrosion properties of different metals for use in specific applications within the NOXSO Process.
The kinetic tests were all performed in a fixed bed reactor (1*2). Funding was provided by NOXSO
and by the U.S. Department of Energy's (DOE) Pittsburgh Energy Technology Center (PETC). The
second and third test programs were funded and conducted by DOE at PETC with technical guidance
provided by NOXSO Corporation. The second test program was designed to test laboratory data in
a scaled-up process, 3/4 MW in size (3). The third test program was a life-cycle test to determine
sorbent physical and chemical performance over repeated cycles of adsorption and regeneration (4).
The current test program is a 5 MW pilot plant that will provide the data necessary to scale up to a
full size (100 MW) module (5). The pilot plant is currently under construction at Ohio Edison's
Toronto Station and is scheduled to begin operation in May 1991. NOXSO Corporation is responsible
for operation of the pilot plant while funding comes from DOE, the Ohio Coal Development Office,
NOXSO Corporation, W.R. Grace & Co., and MK-Fcrguson Co. A brief comparison of these four
test programs is given in Table 1. Detailed information on test facility design, test results, and data
analysis can be obtained from the previously referenced reports.
PROCESS DESCRIPTION
The NOXSO Process is a post-combustion flue gas treatment technology that simultaneously removes
both S02 and NO, from the flue gas generated by coal-fired utility boilers. The process utilizes a high
surface area ^-alumina substrate impregnated with sodium to achieve removal efficiencies of 90% for
S02 and 70%-90% for NO,. A process flow diagram is shown in Figure 1, and a description of the
process is given below.
Flue gas leaving the boiler passes through the combustion air preheater, the electrostatic precipitator,
and into the NOXSO flue gas treatment system. The flue gas is first cooled to 120°C by vaporizing
a water stream sprayed directly in the ductwork. The cooled flue gas is then passed through a
fluidized bed of sorbent where the S02 and NO, are simultaneously adsorbed. The clean flue gas
7A-64
-------
flows through a cyclone where attrited sorbent is separated and returned to the adsorber bed. Finally,
the flue gas is returned to the power plant duct and exhausted through the stack.
After the sorbent is loaded with S02 and NO„, it is removed from the adsorbers and pneumatically
conveyed to a sorbent heater. The sorbent heater is a three-stage fluidized bed where a hot air stream
is used to heat the sorbent to 660°C. During the heating process, NOx and loosely bound S02 are
desorbed and transported away in the heating gas stream. The hot air stream exiting the sorbent heater
is recycled back to the boiler replacing a portion of the combustion air while providing an energy
credit to the NOXSO Process. At normal boiler operating conditions, the recycled NOx will either be
reduced by hydrocarbon fuel or suppressed by the formation of additional NO, so that a steady-state
equilibrium concentration of NO, is attained.
Once the sorbent reaches a regeneration temperature of 660°C, it is transported from the sorbent
heater to a moving bed regenerator. In the regenerator, sorbent is contacted with natural gas in a
countercurrent fashion. The natural gas reduces sulfur compounds on the sorbent (mainly sodium
sulfate) to primarily S02 and H2S with some COS also formed (less than 2% of total inlet sulfur).
Approximately 48% of the sodium sulfate is reduced to sodium sulfide which must subsequently be
hydrolyzed in the steam treatment vessel. The moving bed steam treatment is obtained from the
reaction of steam with Na2S. The regenerator off-gasses are sent to a Claus plant where S02 and H2S
are reacted to form elemental sulfur. The sulfur is sold as a by-product of the NOXSO Process.
From the steam treatment vessel, the sorbent is fed to a sorbent cooler. The cooler is a three-stage
fluidized bed where the sorbent is cooled to 120°C using an ambient air stream. The warm air exiting
the cooler is further heated in a natural gas Fired heater before being used to heat the sorbent in the
fluidized bed heater. The cooled sorbent is returned to the adsorber completing one full cycle.
PROCESS CHEMISTRY
The NOXSO sorbent is prepared by spraying Na2C03 solution on the surface of 7-alumina sphere (1.6
nominal diameter). Both sodium and alumina contribute to the NOXSO sorbent's capacity to adsorb
S02 and NOx from flue gas. Our laboratory tests show that the presence of steam in the flue gas helps
the S02 sorption. The reaction of the sodium can be described as follows:
7A-65
-------
Na2C03 + Alz03 2NaAl02 + C02 W
2NaAl02 + H20 <—» 2NaOH + Al203 ' (2)
2NaOH * S02 * -|o2 — Na2SOi + H20 (3)
2NaOH + 2JV0 + -|°2 ' 2NaN03 + Jf20 (4)
2NaN02 + tf20 (5)
Adsorbed nitrogen oxides are decomposed and evolved on heating the spent sorbent to regeneration
temperature. The concentrated stream of NO, produced on heat-up is returned to the boiler with the
combustion air. This results in no significant increase of NO* concentration in the boiler flue gas
because of the reversibility of NOx formation in the boiler (1.2).
The spent sorbent can be regenerated at high temperature with many kinds of reducing gases, such as
H2S» CO, H2, natural gas, etc. The regeneration reaction, for example, using natural gas at 610°C
is described below:
ANa2SOi + CHa — > 4NazSO. + C02 + 2H20 (6)
4Na2S02 + 3CHt —> 4Na2S +3C02 + 6H.0 (7)
Al203 + l\ra2S03 <— 2NaAl02, + S02 (8)
Al203 + Na2S + H20 <—> 2NaAl02 + H2S (9)
Although sulfite has not been identified in our studies, it is a likely intermediate in sulfate reduction.
A detailed discussion on the existence of sulfide during regeneration had been given by Gavalas it.al,
(6) who used CO to study the regeneration of alkali-alumina. The S02 and H2S produced from
regeneration are then converted to elemental sulfur in a Claus-type, reactor.
7A-66
-------
S02 + 2H2S <—> xS3/x + 2H20
(10)
The sulfur produced is a marketable by-product of the process.
PROOF-OF-CONCEPT PILOT TEST
Background
On May 10, 1989, a consortium assembled by NOXSO Corporation signed a cost-shared contract with
the DOE/PETC to conduct a POC test of the NOXSO process. The consortium consists of NOXSO,
MK-Ferguson, W.R. Grace & Co., Ohio Edison and the Ohio Coal Development Office. The POC
project will take approximately three years to complete, and the test will be conducted at a coal-fired
Ohio Edison plant in Toronto, Ohio.
POC Test Site
The POC unit will treat flue gas from either Boiler #10 or Boiler #11 at Ohio Edison's Toronto
Station. Two sources of flue gas will be tapped so that the POC test can continue as long as one of
the boilers is operating. A slipstream of flue gas will be taken downstream of the Toronto plant's
electrostatic precipitators. The Toronto boilers are pc-fired and bum Ohio coal containing 3.7%
sulfur. The flue gas typically contains 2300 ppm S02 and 350 ppm NO*.
POC Test Schedule
Detailed design engineering has been completed and the major pieces of equipment have been ordered.
Construction began in April 1990 and will be completed in May 1991. The test will run through
February 1992.
POC Process
The process flow diagram for the POC has shown in Figure 1. The system differs from a commercial
application of the NOXSO technology in two important areas. First, the POC facility does not include
a Claus plant, which in the commercial design would be used to produce a sulfur by-product from the
concentrated stream of S02 and H2S produced in the regenerator. This is because Claus technology
is commercially available and therefore does not require testing at pilot scale. Second, the POC does
not include NO, recycle to the coal combustor. In the commercial design, NO, in the air leaving the
sorbent heater is recycled to the combustor as part of the combustion air. Since NOx formation in the
7A-67
-------
coal combustor is a reversible reaction, addition of NOx to the combustion air suppresses the formation
of NO, in the combustor. However, NO* recycle is impractical in the POC test since the POC treats
less than 10% of the flue gas produced by Toronto Unit 10 or 11.
POC Test Unit Design
Data from three previous tests of the NOXSO process were used to design the POC test facility. A
comparison of the three previous test programs was given in Table 1. The design specifications for
the major equipment in the POC test facility are listed' in Table 2.
Materials of Construction
During development of the NOXSO process, some corrosion problems were encountered, particularly
in the regenerator. Different materials of construction were tested to withstand the high temperature
environment of S02, H2S, elemental sulfur, and sulfated sorbent. Corrosion results were documented
in an earlier report (2), the practical results of the test program are discussed here.
In tests performed at the Shawnee Steam Plant, sorbent was heated with electrical resistance heaters
made of Inconel 600, Monel 400, type 316 and type 316L stainless steel (SS). A11 these materials
exhibited severe corrosion in areas of sorbent contact attributed to hot sulfation of nickel. It should
be noted that the temperature of the heating elements themselves were substantially higher than the bed
temperature of 600°C. The reactor, made out of either type 316 or type 316L SS, showed scale on
the inside surfaces after use. When the reactor was made of type 446 SS or alonized type 316L SS,
there was no scale and only a slight discoloration of the metal surfaces observed.
In the LCTU, the regenerator was made of alonized type 304 SS and showed no visible evidence of
corrosion at the end of 330 regeneration cycles. Based on these results, it was felt that either 446 SS
or alonized 304 or 316L SS would be satisfactory for the POC regenerator.
The sorbent heater also encounters hot sulfated sorbent and will therefore be made of type 304 SS.
The bottom bed of the sorbent heater where the temperature is 660°C will be alonized. All other
vessels will be made of standard A-285 or A-283 grade C carbon steel, as no corrosion problems are
anticipated.
7A-68
-------
The other area in the NOXSO process that requires special consideration for materials is between the
flue gas cooler and the adsorber. In this area, sub-acid dewpoint corrosion can occur. All previous
NOXSO tests have cooled the flue gas indirectly while at the POC the flue gas will be cooled by a
direct water spray in the ductwork. The flue gas temperature in this portion of the system will be
112°F so that an acid-resistant epoxy coating will be used to line the ductwork from the cooler to and
including the bottom of the adsorber. This epoxy has not been tested previously by NOXSO, but there
exists ample literature that supports its use as an acid resistant material in other similar applications.
NO, RECYCLE
NO, recycle will be implemented at the full-scale commercial demonstration plant. The concept of
NO, recycle has been tested previously using the 500 lb/hr coal combustor used for the 3/4 MW tests
and also using a tunnel furnace capable of being fired with a variety of fuels including gas, fuel oil,
coal, and coal-water mixtures.
Previous NO. Recycle Results
NO, recycle was tested by spiking the combustion air with varying concentrations of bottled NO, and
measuring the outlet NO, concentration from the combustor. The net NO, production rate was
determined by a material balance on the combustor as shown schematically in Figure 2. The NO, flow
rate at the exit of the combustor minus the NO, feedrate into the combustor equals the rate that NO,
is produced in the combustor, which is defined as the net NO, production rate (R). For data reduction
purposes, the NO, production rate (R) and the NO, feedrate (F) were normalized with respect to
conditions at zero NO, feed according to R*=R/R„ and F*=F/R„ where Rc is the NO, production rate
at F = O. Results from the 500 lb/hr combustor are compiled in Table 3. The measured data are
NO, concentration at the exit of the combustion system and the flow rate of NO, fed into the
combustor with the combustion air. Data provided in the other columns were calculated.
A plot of R* versus F* is shown in Figures 3 and 4 for both the 500 lb/hr combustor and the tunnel
furnace, respectively. In each case, the data fall in a straight line, but with different slopes. The two
lines are described by the equation R* = 1 - aF*. The parameter "a" is the slope of the line and also
represents the fraction of NO, fed to the combustor that is destroyed, The value of "a" is 0.65 for
the 500 lb/hr combustor and 0.75 for the tunnel furnace. The data for the tunnel furnace include both
natural gas combustion and coal-water slurry combustion.
7A-69
-------
These results demonstrate that the nature of the fuel has no affect on the effectiveness of the
combustion system to reduce NO* fed through the combustion air. Also, the NOx reduction capability
of a combustion system is independent of the amount of NO, fed with the combustion air. Finally,
the most important variables are those associated with the combustor design. NO, recycle will be
extensively studied at the Babcock & Wilcox Research Center in Alliance, Ohio.
Pilot-Scale NO. Recycle Test
The power plant selected for the NOXSO full-scale demonstration (Ohio Edison's Niles Station, Niles,
Ohio) uses cyclone burners. Since the destruction efficiency of NO, recycle has not previously been
tested with cyclone type burners, a demonstration of NO, recycle with this type of coal combustor is
necessary for the proper design of the NOXSO full-scale plant.
Pilot-scale NO, recycle tests will be done using Babcock & Wilcox's 6 million Btu/hr Small Boiler
Simulator (SBS) shown in Figure 5. The water-cooled furnace is a scaled-down version of B&W's
single-cyclone, front-wall fired boiler design. The cyclone has been in operation since February 1985.
The SBS cyclone furnace simulates a large cyclone unit very well. A comparison between the SBS
cyclone furnace and commercial units is given in Table 4.
The NO, recycle tests will begin with three loads and three excess air levels to establish the baseline
of the NO, emission from the SBS furnace, NO will then be injected in multiples of the baseline NO,
production levels. The NO concentration at the air inlet duct to the cyclone will be measured to
document the inlet level. Stack NO, will be measured to determine NO, destruction occurring in the
flame. The series of tests with different NO injection rates will also be performed at three furnace
loads and three excess air levels. This test result will assist the determination of a second injection
location for the next series of tests.
In the second series of tests, NO and N02 will be injected separately for two furnace loads and two
excess air levels. Volumetric flowrate of the injected NO and N02 will be based on the proportion
of these gasses that are present in the NOXSO sorbent heater off-gas. The addition of methane to the
air stream to assist the NO, destruction (7) will also be tested. The NO, recycle test will be finalized
by burning the coal from the Niles plant in the SBS furnace. Since the coal-ash slagging
characteristics are important to the power plant operation, the use of Niles plant coal will assess the
change of the coal ash's "flowability" in the Niles plant when the NO, recycle stream is installed.
7A-70
-------
FUTURE WORK
On December 21, 1989, NOXSO Corporation, in association with MK-Ferguson Company, W.R.
Grace & Co., and Ohio Edison, received an award from DOE's Clean Coal Technology Program to
conduct a $66 million, full-scale commercial demonstration of the NOXSO technology. The U.S.
DOE will provide 533 million and the remaining funds will be provided by the Ohio Coal
Development Office, the Electric Power Research Institute, the Gas Research Institute, the East Ohio
Gas Company, and the aforementioned NOXSO development team. The 115 MW demonstration plant
will be installed at Ohio Edison's Niles Power Plant in northeastern Ohio. Construction is scheduled
to begin in early 1993 with plant startup scheduled in May 1994. This project is the final step in the
program to commercialize the NOXSO technology.
REFERENCES
1. J.L. Haslbeck, C.J. Wang, L.G. Neal, H.P. Tseng, and J.D. Tucker. Evaluation of the NOXSO
Combined N0x/S02 Flue Gas Treatment Process. NOXSO Corporation Contract Report
submitted to U.S. DOE Report No. DOE/FE/60148-T5. November 1984.
2. J.L. Haslbeck, L.G. Neal, C.J. Wang, and C.P. Perng. Evaluation of the NOXSO Combined
NO^/SOj Flue Gas Treatment Process. NOXSO Corporation Contract Report submitted to U.S.
DOE Report No. DOE/PC/73225-T2. April 1985.
3. J.L. Haslbeck, W.T. Ma, and L.G. Neal. A Pilot-Scale Test of the NOXSO Flue Gas Treatment
Process. NOXSO Corporation Contract Report submitted to U.S. DOE Contract No. DE-FC22-
85PC81503. June 1988.
4. J.L. Haslbeck, J.T. Yeh, W.T. Ma, J.P. Solar, and H.W. Pennline. Life-Cycle Test of the
NOXSO Process: Simultaneous Removal of NO, and S02 from Flue Gas. Presented at the 1989
AWMA Annual Meeting, Anaheim, California. June 1989.
5. J.L. Haslbeck, M.C. Woods, R.E. Bolli, R.L. Gilbert, and C.P. Brundrett. Proof-of-Concept
Test of the NOXSO Flue Gas Treatment System, Presented at the EPA/EPRI 1990 S02 Control
Symposium. New Orleans, Louisiana. May 8-11, 1990.
6. G.R. Gavalas, S. Edelstan, M. Flytzani-Stephanopoulous, and T.A. Weston. Alkali-Alumina
Sorbents for High-Temperature Removal of S03. AIChE Journal Vol. 33, No. 2, p. 258. 1987.
7. J.T. Yeh, J.M. Ekmann, H.W. Pennline, and C.J. Drummond. New Strategy to Decompose
Nitrogen Oxides from Regenerable Flue Gas Cleanup Processes. Presented at the 194th ACS
National Meeting. New Orleans, Louisiana. Aug. 30 - Sept. 4, 1987.
7A-71
-------
NOXSO PROCESS FLOW DIAGRAM
FIGURE 1
7A-72
-------
(1-E,KR+F)
FIGURE 2. SCHEMATIC DIAGRAM OF NITROGEN
OXIDE RECYCLE.
7A-73
-------
NORMALIZED NOx feed rate, f*
FIGURE 3. NORMALIZED NOx REDUCTION
DATA-PC COMBUSTOR.
I
0 5 10 15
NORMALIZED N0X FEED RATE, F*
FIGURE 4. NORMALIZED N0X
REDUCTION DATA-
TUNNEL FURNACE.
7A-74
-------
STACK
STEAM
REHIATIR
DEPOSITION
PROBE
SUPERHEATER
FOULING TUBE
DEPOSITION PROBE
FLUE GAS
RECIRCULATION
FURNACE ARCH
PRIMARY AIR
AND COAL
TERTIARY AIR
SECONDARY
AIR
SLAG TAP
MOLTEN SLAG
SLAG COLLECTOR
AND FURNACE
WATER SEAL
FIGURE 5. SMALL BOILER SIMULATOR (SBS) SCHEMATIC
7A-75
-------
Table 1. Comparison of NOXSO Test Programs
Test Program
Operating Parameter
TV A
3/4 MW
Coal Burned, lbs/hr
NA
500
Flue Gas Volume, SCFM
0.35
1200
Adsorber Type
Fixed Bed
Fluid Bed
S02 Inlet Concentration, ppm
2300
1465-5000
NO* Inlet Concentration, ppm
600
470-720
S02 Removal Efficiency, %
90
90-99*
NO, Removal Efficiency, %
90
80-92*
Reducing Gas for Regeneration
H2S, h2, CO
h2, h2+co, CH<
Operating Mode
Batch
Batch
Test Program
Operating Parameter
LCTIJ
POC
Coal Burned, lbs/hr
40
NA
Flue Gas Volume, SCFM
120
12000
Adsorber Type
Fluid Bed
Fluid Bed
S02 Inlet Concentration, ppm
1450-3000
2300
NO, Inlet Concentration, ppm
240-800
350
S02 Removal Efficiency, %
60-90*
**
NO, Removal Efficiency, %
60-90*
**
Reducing Gas for Regeneration
TT pTJ
Natural Gas
Operating Mode
Continuous
Continuous
NA = Not applicable, i.e., small- slipstream was drawn from
the power plant ductwork.
* = Iu the 3/4 MW and LCTU tests, removal efficiencies cover a wide range since
operating conditions were intentionally varied to study their effect on process
performance.
** = Pilot plant is under construction.
7A-76
-------
f
Table 2. POC Major Equipment Specifications*
Fluidized Bed Adsorber
Diameter 10.5 ft
Temperature 120°C
Settled Bed Height 2 ft
Sorbent Residence Time 45 min
Superficial Gas Velocity 3 ft/s
Transport Disengaging Height 7,7 ft
Material of Construction Carbon Steel
Fluidized Bed Sorbent Heater
Number of Stages 3
Diameter 7.7 ft
Settled Bed Height 0.9 ft
Sorbent Residence Time 30 min
Superficial Gas Velocity 3 ft/s
Transport Disengaging Height 2.8 ft
Material of Construction Type 304 SS
Fluidized Bed Sorbent Cooler
Number of Stages 3
Diameter 5.7 ft
Settled Bed Height 1.2 ft
Sorbent Residence Time 20 min
Superficial Gas Velocity 3 ft/s
Transport Disengaging Height 4.3 ft
Material of Construction ^ Carbon Steel
Moving Bed Regenerator/Steam Treater
Diameter 4 ft
Bed Height 10.3 ft/6.8 ft
Sorbent Residence Time 30 min/20 min
Material of Construction Alonized Type 304H SS
Air Heater
Design Flow (Air)
Temperature Rise
Type
Pneumatic Conveyor
Sorbent Circulation Rate
Lift Distance
14,300 Ibs/hr
330°C
Natural gas fired in duct burners
9,994 lbs/hr
80 ft
Adsorber Cyclone
D-50
D-100
Gas Flowrate
* At base case operating conditions.
20 Mm
100 £im
16,257 ACFM @ 120°C
7A-77
-------
Table 3, NO, Reduction Data; 500 Ib/hr Combustor (3)
F
Test
NO, Exit
NO, Exit
NO, Fed
R
No.#
ppm
lb/hr
lb/hr
lb/hr
R*
F*
1
550
3.59
0
+3.59
1.0
0
2
1370
8,94
14.09
-5.15
-1,43
3.92
3
875
5.71
8.29
-2.58
-0.72
2.31
4
650
4.24
0
+4.24
1.0
0
5
850
5.55
4.66
+0.89
0.21
1.10
6
930
6.07
5.49
+0.58
0.14
1.29
7
700
4.56
0
+4.56
1.0
0
8
1100
7.17
6.64
+0.53
0.12
1.46
9
1200
7.82
7.98
-0.16
-0.04
1.75
10
820
5.34
1.60
+3.74
0.82
0.35
Tests 1-3. Coal feedrate = 223 lbs/hr, Flue gas flowrate = 122.1
moles/hr (dry), and Temperature = 2500°F.
Tests 4 - 6. Coal feedrate = 352 lbs/hr, Flue gas flowrate = 160.0
moles/hr (dry), and Temperature = 25Q0°F.
Tests 7 - 10. Coal feedrate = 431 lbs/hr, Flue gas flowrate = 180.4
moles/hr (dry), and Temperature = 2500°F.
Table 4. Comparison of Baseline Conditions Between
the SBS Facility and Commercial Units
Cyclone Temperature
Residence Time at full load
Furnace Exit Gas Temperature
NOx Level
Ash Retention
Unbumed Carbon
Ash Particle Size (MMD; Bahco)
SBS
>3000°F
1.4 sec
2265°F
900-1200 ppm
80%-85%
< 1 % in Ash
6-8 microns
Typical
Cvclone-Fired Boilers
>30oo oF
0.7-2.0 sec
2200C-2350°F
600-1400 ppm
60%-80%
1 % -20 %
6-11 microns
7A-78
-------
THE PRACTICAL APPLICATION OF TUNABLE DIODE LASER INFRARED
SPECTROSCOPY TO THE MONITORING OF NITROUS OXIDE AND OTHER
COMBUSTION PROCESS STREAM GASES
Frank E. Briden
Air and Energy Engineering Research Laboratory
U.S. Enviornmental Protection Agency
Research Triangle Park, North Carolina 27711
David F. Natschke
Richard B. Snoddy
Acurex Corporation
4915 Prospectus Drive
Durham, North Carolina 27713
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved lor
presentation and publication.
7 A- 79
-------
-------
ABSTRACT
There are a number of gases associated with combustion process streams which are difficult
to monitor because of their physical properties and interferences from other gases. Tunable
diode laser infrared (TDIR) spectroscopy offers a reliable, specific means for the continuous
monitoring of many of these gases. Some of the gases that can be efficiently monitored by
this technique are N20, NO, N02, H20, H202, 03, NH3, HON, S02, S03, OCS, C02, CO,
HCI, HBr, HF, CH3CI, CH4, CH3OH, and C2H5OH, to name a few.
This technique requires the use of sophisticated electronic components, but provides an
extremely rugged, simple to operate, stable, sensitive, and reliable instrument. This paper
describes how the Air and Energy Engineering Research Laboratory of the Environmental
Protection Agency at Research Triangle Park, NC, designed, built, and tested, with a coal
burning furnace, a TDIR monitor for NzO. The present diode mount is limited to the
simultaneous use of only two 2 diodes and therefore only two analyte gases per optical cell.
Newer mounts allow the simultaneous use of four diodes. The conversion of the system for
other gases will be described. TDIR in-stack monitoring and long-range atmospheric
monitoring will also be discussed.
7A-81
-------
INTRODUCTION
The measurement of atmospheric N20 and its sources is of great interest since it is a
potential contributor to global warming and its atmospheric concentration is increasing. The
principal sampling method uses an evacuated container to collect a grab sample of the gas
stream of interest, so the containers could be taken back to a laboratory and analyzed later.
The original data indicated a linear relationship between the concentrations of N20 and NOx
in the stack gases. The validity of this data began to be questioned in the mid-1980s when
studies showed the detection of N20 when none was expected. Muzio et al. reported on the
formation of N20 as a sampling artifact while studying natural gas flames injected with S02
and NH3. ^ Another report showed that the artifact could be reduced by drying the gas
before sampling, and the artifact could be effectively eliminated by removing the S02. It
was evident that a gas-phase aqueous reaction between S02, NOx, and H20 was generating
N20 in the sample container. These reactions have been known since the 18th century and
reported as early as 1924.(1) Discovery of this sampling artifact led to research on the
development of sampling and analysis techniques which would provide accurate results.
One project in this area, by the Air and Energy Engineering Research Laboratory, used a
heated sample line and then filtered and desiccated the gas before it was analyzed by an on-
line GC/ECD (for N20) and continuous emission monitors (for 02, C02, CO, and NO). This
research indicated that the N20 concentrations were less than 5 ppm and were not a
function of the N0X concentration. (3>
This project was undertaken to demonstrate the ability of a laser diode system to accurately
and correctly measure the concentration of N20 in stack gas in real time, and to verify the
lower N20 concentrations reported with modified sampling methods.
7A-82
-------
EXPERIMENT
The detection of trace gases using second-derivative spectroscopy was first used in 1978 by
Reid et al. at McMaster University.(4* Second-derivative or modulation spectroscopy
consists of using a modulated source to scan the absorption line of interest. The detector
output is amplified using a phase-sensitive amplifier referenced to twice the modulation rate.
In addition to significantly reducing the background noise by rejecting all signals which are
not in phase with the reference signal, operating the amplifier at twice the modulation, rate
produces a pseudo-second-derivative signal as the output. This signal is proportional to the
absorption of the line being scanned but the signal must be calibrated for each line of
interest. A beamsplitter, lock cell, and a second detector are used to provide a feedback
signal to correct for any drift in the source. For the feedback circuit, a phase-sensitive
amplifier referenced to the modulation frequency reduces the noise level and provides the
stabilization signal.In this system, an infrared diode laser modulated at 2000 Hz was
used as the source. The output frequency of a diode laser can be broadly tuned by adjusting
its operating temperature and finely tuned by varying the applied current. This particular
diode is tunable over the range 2200 to 2215 cm'1. Figure 1 diagrams the optical system.
The cold head contains part of the cooling system for the diode and also provides an
insulating vacuum for the diode since it is operated at 28 K. A monochrometer is used to
isolate the laser line of interest. The beamsplitter deflects a portion of the laser light through
a lock-cell containing a high concentration of N20, arid into a detector to generate the
stabilization signal. The rest of the laser energy passes through the beamsplitter, into the
analytical cell, and then into the analytical detector. The analytical cell is a two-pass, 0.5 m
cell with an external retroreflector. Both detectors are single element mercury cadmium
telluride photoconductive detectors with low noise preamplifiers. The first and second
derivative signals are generated by setting the reference channel of the phase-sensitive
amplifier to either the T (first derivative) or "2f (second derivative) mode. In the "2f mode,
the reference channel of the phase-sensitive amplifier is driven internally at twice the input
frequency, eliminating the need for an external, stabilized 4000 Hz reference signal. The
output signal from the amplifier (for either mode) is a pure DC signal reflecting the magnitude
7A-83
-------
of the input signal which is in phase with the reference signal. Any AC component is the
result of noise in the system and is reduced by the AC filter at the output. This AC filter has
a variable time constant which can be adjusted from 1ms to 100 sec. A higher setting of the
time constant will reduce the noise level, but will also eliminate the corresponding time
variations in the analytical detector signal. The output signal is then displayed on the chart
recorder. The change in the magnitude of the signal, as measured from the baseline
(determined using dry nitrogen gas), is directly proportional to the concentration of N20 in the
analytical cell.
Before beginning the tests, the N20 line with the least interference from the other gases
commonly found in stack gas (H20, C02, CO) at various pressures, temperatures, and
concentrations was determined. Theoretical spectra were calculated using the FASCODE
algorithm which was developed by the Air Force Geophysics Laboratory.(5) Examples of
these spectra are shown in Figures 2, 3, and 4. During this work, the gas pressure in the
analytical cell was maintained at 5000 Pa by continuously pumping on the outlet side of the
cell with a vacuum pump and limiting the flow at the cell inlet port. This kept the pressure-
broadening of the lines to a minimum and, during sampling of furnace gases, cooled the
furnace gases to reduce thermal-line-broadening. The line at 2208.75 cm"1 was chosen for
this work. Initial tests using mixed gases from cylinders verified the detection of N20 and no
response to the C02, CO, S02, and H20 vapor.
The equipment was moved from the laboratory and connected to the Innovative Furnace
Reactor (IFR), a furnace designed to evaluate various methods of scrubbing stack gases. It
is a down-fired, tunnel-fired furnace burning powdered coal. Figure 5 diagrams the system .
During these tests, the IFR was being used to evaluate the efficiency of powdered lime to
reduce S02 emissions. The stack gases were sampled at two different positions (see Tables
1 and 2), one at the end of the furnace before the gas is filtered in the bag house, and the
other near the roof just before the gases were vented to the atmosphere. These are
indicated in Figure 5 as #1 and #2, respectively. The gases at the two sampling positions
are significantly different. At position # 1, the gases reflect the actual combustion products of
the furnace. After leaving the furnace, the gases are diluted and cooled to protect the bag
house filter elements and the roof-mounted blower from damage due to excessive heat.
Although the gases sampled at position # 2 reflect what is discharged to the atmosphere, the
gases have been diluted, cooled, and filtered.
7A-84
-------
The operating parameters of the TDIR system are listed in Table 1.
These operating parameters are typical for each sampling position, but the actual values
were adjusted slightly to optimize the system each day. The system was calibrated each day
using N20 in dry nitrogen at concentrations of 0.108, 0.514, 0.970, 1.99, and 4.82 ppm. A
sample of the data collected from sample position #2 is in Figure 6. This section of the chart
paper shows the time variations in the N20 concentration which is attributed to fluctuations in
the coal feed rate. Also visible are the areas where dry nitrogen is used to verify the
baseline. The addition of powdered lime to the stack had no effect on the measured N20
concentrations. The average concentration is 0.9 ppm with a maximum excursion of 1.0 ppm
and a minimum of 0.8 ppm. Figure 7 shows data collected at sample position #1. There are
several differences evident in this chart. First, the level of N20 is much lower, about 0.3
ppm. Second, as the system is switched from sampling dry nitrogen to stack gas, there is a
spike in the NaO concentration which is not seen in the data from position # 2. Third, the
two spikes at the end of the trace are observed each time the coal feed is stopped and only
air is blown into the burner section of the furnace.
INTERPRETATION
The calibration data were fitted using a linear function to correlate a given deviation from the
baseline to concentration. The results are summarized in Table 2. These concentrations are
much lower than those in the original NzO database and are also lower than the more recent
data indicated. The higher concentrations in the stack at position #2 are caused by the
formation of N20 in the baghouse. The concentration is reduced, by dilution of the gas
stream in the baghouse and after the baghouse, to cool the gas before it is vented. The
data from position #1 is a more accurate measure of N20 produced by the furnace since it is
sampled before there is any chance of dilution and the gas temperature (300 °C) is high
enough to keep the water in vapor form. It is assumed that both the higher temperature and
the reduced time between sampling and analysis work to reduce the amount of N20
generated as an artifact.
The spikes in the data from position #1 are the result of N2Q generation in the filter unit and
the short section pipe connecting the heated sample line to the furnace. The filter and
connector pipe were not heated and would cool off when the furnace gases were not flowing
7A-85
-------
through them. This permits water condensation and the formation of N20 in these unheated
parts. When gas was subsequently drawn from the furnace, the small volume of gas in the
pipe and filter would precede the hot furnace gases into the analytical cell and cause a spike
in the output. The fact that this effect was not observed in the data from position #2
indicates that the gas components had already interacted, producing N20, and could not
generate more N20 in the filter set. It is assumed that this reaction most likely took place in
the baghouse where the ash and lime reaction products were collected and the temperature
fell below 100 °C causing the water vapor to condense and initiate the reaction.
The fluctuations in the N20 concentration both during furnace operation and at the end,
when the coal feed unit was turned off, were well correlated to similar fluctuations In the
concentration of CO which was continuously measured as part of the S02 scrubbing tests.
This may indicate that the N20 is a result of a lower concentration of oxygen in the furnace
which also generates more CO.
CONCLUSIONS
In this study, it was found that the NzO concentration, immediately after the combustor
(position #1, Figure 5) varied above and below ambient which was measured at 280 ppb.
However, conditions in the baghouse caused an increase of N20 up to about 3 times
ambient (position #2, Figure 5). The major source of N20 in the stack gas appears to be its
formation when the water vapor condenses and reacts with other components of the stack
gases.
This study also shows great promise for the use of laser diode modulation spectroscopy for
other applications where continuous monitoring of one or more trace gases is required. The
system is easily modified to monitor other gases by replacing the diode with one that will
operate in the spectral region of interest. By operating both diodes simultaneously and
adding more optical components, the current system can be configured to simultaneously
monitor two gases in the sample stream. There are also cold head systems available which
will allow the use of four diodes simultaneously and therefore the monitoring of four distinct
trace gases.
This method may also be used to directly measure species concentrations in the stack by
using optical windows mounted in the stack access ports (e.g., the sulfuric acid
measurements of Pearson and Mantz ^). Measurements of atmospheric contaminants over
7A-86
-------
long path lengths are feasible and could provide significant information on the generation,
distribution, and dissipation of pollutants which are not generated from single sources. It is
proposed to use this technique to monitor methane emissions from landfills or pasture land.
7A-87
-------
REFERENCES .
(1) L.J. Muzio, et al. "Errors in Grab Sample Measurements of N20 from Combustion
Sources." JAPCA, Vol. 39 No. 3, 1989, pages 287-293.
(2) L.J. Muzio and J.C. Kramlich. "An Artifact in the Measurement of N20 from
Combustion Sources." Geophysical Research Letters, Vol. 15 No. 12, 1988, pages
1369-1372.
(3) W.P. Linak, et al. "N20 Emissions from Fossil Fuel Combustion," In Proceedings;
1989 Symposium on Stationary Combustion NOx Control, San Francisco. CA, March
6-9, 1989, Volume 1, EPA-600/9-89-062a (NTIS) PB89-220529), June 1989.
{4) R.S. Eng, et al. "Tunable Diode Laser Spectroscopy: An Invited Review." Optical
Engineering, Vol. 19 No. 6, pages 952-953.
(5) FASCODE - Fast Atmospheric Signature Code (Spectral Transmittance and
Radiance), H.J.P. Smith et al., AFGL-TR-78-0081 Air Force Geophysics Laboratory,
Air Force Systems Command, United States Air Force, Hanscom AFB, MA 01731.
(6) E.F. Pearson, A.W. Mantz. "A Tunable Diode Laser Stack Monitor for Sulfuric Acid
Vapor" EPA-600/2-80 174 (NTIS PB80-202 690), 1979.
7A-88
-------
Figure 1. Laser diode setup
7A-89
-------
-J
>
¦
CO
o
m
b
8.0-1
7.0-
6.0-
jr 5.0-
&
Q
_J 4.0-^
a
E
O 3.0-1
2.0-
1.0-
0.0
1
2200
2201
N20 1 PPM, 50 MB, 25 C
-r-*
2202 2203 2204
WAVENUMBER (CM1)
Figure 2. N20 Spectrum at 25 °C
-------
2200
2201
N20 1PPM, 50MB, 100 C
li
JL
2202 2203
WAVENUMBER (CM1)
Figure 3. N20 Spectrum at 100 °C
-------
N20. 002. H20, CO AT 25 C. 50 MB
10.0
-si
>
CO
i\i
o
i ^ f
2208 2208.2
22084 22088
WAVENUMBER (CM1)
22088
1v
2209
Legend
tog pro wan
mew
-------
Coal Feeder
-¦J
>
I
CO
u
Sorbenl Feeder
JZ
Sorbenl/Slurry
Injection Probe
Sampling Ports
Root
N20 Sampling Port #2 —
SO Sampling Port
1
t:
o
CL
O)
c
Q.
E
CO -j
CO I"
O
Fan
o
CEM Sampling Port 1
h
-DHL
Slack
±
±
Figure 5. Innovative Furnace Reactor
-------
7A-94
-------
7A-95
-------
TABLE 1. OPERATING PARAMETERS
Parameter
Sample.Position #1
Sample Position #2
Current, mA
217
217
Temperature, K
28
28
Frequency, Hz
2000
2000
Scan Width*, mA
• 5
5
Sensitivity, mV
0,01
0.025
Time Constant, sec
3
3
* A current scan width of 5 mA equates to a frequency shift of 0.75 cm'1
TABLE 2. OBSERVED N20 CONCENTRATION
Data
Sample Position #1
Sample Position #2
ppm
ppm
Average
0.30
0.74
Maximum
0.46
1.27
Minimum
0.14
0.75
(± 0.053}
(± 0.025)
7A-96
-------
In-Furnace Low NOx Solutions for Wall Fired Boilers
by
R. C, LaFlesh, D. Hart, and P. Jennings (ABB Combustion Engineering)
Michael Darroch (City of Jamestown, New York)
Abstract
Since the early 1940's, several thousand Type R pulverized coal burners have
been installed and are operating in wall fired boilers ranging up to 160 MWe
in size. In response to the low N0X Emission requirements, ABB Combustion
Engineering Services, Inc. has undertaken development of the R0-11 coal
burner based on proven Type R wall firing technology with additional N0X
control capability.
Extensive laboratory tests were conducted at a large pilot scale {50 x 10®
Btu/hr) in order to optimize the RO-II coal burner configuration.
Specifically, a number of coal nozzle/air register configurations were
evaluated in terms of their combined ability to meet specific emissions and
operational performance criteria. The RO-II burner reduced NGX from a
baseline uncontrolled level of 0.9 #/106 Btu to 0.5 #/106 Btu during the
laboratory trials.
This paper will review laboratory development activities and report on RO-II
field demonstrations currently in progress.
Background
As a result of the recent Clean Air Act and specific local regulations,
boiler operators are addressing the need to reduce stack gas emissions.
Current attention is focused upon controlling acid rain precursors, oxides of
nitrogen (N0X) and sulfur dioxide (SO^). Under Phase I of the Clean Air Act,
a number of pre NSPS - coal burning wall fired boilers will be required to
reduce their NO* emissions by the mid 1990's. The proposed Federal upper
limit for N0X emissions from wall fired units is 0.50 Ib/HBtu fired.
ABB Combustion Engineering Services, Inc. (ABB-CE) has been actively
developing and commercially demonstrating low N0X technologies for coal fired
tangential and cyclone boiler arrangements. In order to meet the N0X
reduction needs of coal wall fired boilers, ABB-CE has embarked on an
extensive low N0X coal burner development and commercial demonstration
program building on its substantial wall fired experience base with the
ABB-CE Type R burner.
The Type R horizontal burner was developed by Combustion Engineering Inc. in
the early 1940's. This burner was designed to burn pulverized coal, oil, or
7B-1
-------
gas, is applicable to single wall or opposed wall firing in either single or
multiple burner arrangements. In terms of experience, several thousand Type
R burners have been installed and operated in a wide variety of boiler
configurations ranging up to 160 MWe in capacity. Individual burner
capacities have ranged from 20 MBtu/hr to 120 MBtu/hr, As a result of this
extensive field experiences ABB-CE has established Type R design standards
which delineate proven materials of construction and fabrication techniques,
Type R operating procedures are also firmly established.
The Type R coal burner, illustrated in Figure 1, has several key hardware
features. The centrally located coal nozzle has spiral rifling along the
inner walls to promote swirl of the pulverized coal/primary, air stream which
is initially established by a tangential inlet nozzle. A convergent nozzle
tip is located at the end of the coal nozzle. Five (5) deflector vanes,
located near the tangential inlet nozzle, can be adjusted in terms of
incident angle to vary coal/primary air stream swirl which in turn,
influences final luminous flame shape. On the combustion air side, the total
combustion air flow passes through an adjustable angle flat blade swirler
assembly. Combustion air angular momentum can be varied to optimize the
burner's flame stabilizing aerodynamic recirculation zone, directly
Influencing both flame stability and flame shape.
Laboratory Development Program
In order to respond to low NQX requirements for wall fired-coal boiler
retrofit market, ABB-CE embarked on a laboratory development program with the
objective of developing a new low N0X wall fired burner product.
The new burner., named the RO-II burner, would be capable of meeting the
following performance targets;
• N0X less than O.50/1O6 Btu Fired
• Zero or nominal increase in carbon loss and/or CO emissions under
low N0X conditions.
• Acceptable flame envelope (length).
• Zero or nominal increase in fuel system or combustion air windbox
static pressure(s).
At the onset of the development program, ABB-CE assessed the N0„ reduction
potential of the Type R burner design; upon review it was decided to
7B-2
-------
incorporate certain key design features of the Type R design into the new
RO-II burner design. These features specifically included the tangential
inlet, spirally rifled coal nozzle and an adjustable coal stream deflector
vane assembly. The Type R combustion air register assembly was determined to
not offer any advantages in terms of reducing total N0X so alternative air
register assemblies were reviewed for incorporation into the new low N0X RO-
II burner design.
ABB-CE selected a patented, commercially available, air register for
incorporation into the RO-II burner. Key features of the register are
highlighted in Figure 2. These features include:
1. Two separate plenums which permit staged introduction of combustion
air.
- pilot air which is introduced concentrically adjacent to the
centrally located coal nozzle
- main air which surrounds the pilot air stream
2. Involute (spirally shaped) air inlets for each plenum which swirl
total combustion air flow.
3. Separate flow control dampers for both the pilot and main air
streams.
4. Integral instrumentation which permits burner operators to balance
combustion air flow to multiple burner arrays located within a
common windbox.
5. Unique helical flow vane assembly which enhances combustion air
swirl and improves air distribution within the register.
6. A shadow vane assembly which enhances combustion air swirl but more
importantly protects the flow vane assembly and fuel nozzle from
damage due to flame radiation in multiple burner installations.
Photo 1, an end-on view of the RO-II register assembly, highlights the
involute (spirally shaped) air plenum, for both the pilot and main combustion
air streams, and the shadow vane assembly. Photo 2 highlights the flow vane
assembly utilized in the RO-II register. The helical vane arrangement is
shown separate from the air register. Note that the pilot combustion air
stream passes through six (6) vanes at the rear of the burner (i.e. the
widest part of the vane assembly), the main combustion air stream passes
through eight (8) vanes near the burner front (i.e. the narrowest part of the
vane assembly). It should also be noted that the register design requires
minimal maintenance since the only moving parts are the pilot and main air
7B-3
-------
dampers. These same dampers also provide the register with the ability to
compensate for burner to burner combustion air flow imbalances in multiple
burner/common windbox arrangements.
The RO-II development program was largely comprised of extensive combustion
trials of potential RO-II firing system hardware. These trials were
conducted in one of ABB-CE's front wall fired large scale laboratory test
furnaces. ABB-CE's development philosophy was to conduct tests with hardware
designed to operate at a heat input rate of 50 x 1G6 BTU/HR. This rate is
identical to the design heat input rate of the burners to be installed in two
units in Jamestown, NY. By adopting this development philosophy, ABB-CE
could confidently accelerate the process of transitioning laboratory hardware
developments into commercial application.
Prior to conducting the laboratory combustion trials, ABB-CE evaluated the
air register's near-field aerodynamics. The objective of these tests was to
define key aerodynamic characteristics of the register in order to support
the design of compatible coal nozzle configurations. Recirculation zone size
and strength as well as the air register's potential to control stoichiometry
in the burner near field (through internal air staging) were assessed. These
aerodynamic properties were consistent with the low N0X objectives of the RO-
II development program.
Laboratory combustion trials began following the register aerodynamic study.
The focus of these trials was to evaluate the combustion performance of a
variety of air register/coal nozzle configurations. The performance of each
configuration was evaluated in comparison with the overall performance
targets for the RO-II burner. It should be noted that the air register
configuration remained fixed throughout the trials. Development activities
concentrated on combining advanced low N0X Type R coal nozzle arrangements
with the existing air register design.
The combustion trials generated the data necessary to assess RO-II burner
performance, flue gas 02, N0X, and CO concentrations were measured at each
test condition, along with coal/primary air static pressure at the coal
nozzle inlet, windbox and furnace static pressures, and total combustion air
and primary air mass flows. Qualitative assessments of flame shape, length,
and stability were also made throughout the development program. In
addition, flyash samples were taken for subsequent carbon in ash analysis.
7B-4
-------
Furnace horizontal exit gas temperatures were also quantified using suction
pyrometry.
The combustion test program pararr.etrlcally evaluated a number of key RO-II
design and operating variables. Some of the variables investigated included,
coal nozzle/tip configurations, firing rate (MCR and reduced load), excess
air level, coal/primary air velocity at the coal nozzle tip exit, pilot and
main air damper position (pilot/main air flow split) and coal stream swirl.
All laboratory trials were conducted with a Pennsylvania bituminous coal
having 10% ash, with a fixed carbon to volatile ratio of 1.65 and a fuel
nitrogen content of 1.5% by weight. Coal preparation for the laboratory
tests was consistent with typical utility practice; the pulverized coal grind
averaged 70.3% through 200 mesh {75 microns), with 0.6% remaining on a 50
mesh (300 microns) screen.
The laboratory test furnace utilizes a dilute phase (1.5 - 2.0 # primary
air/# coal) indirect coal feed system. A schematic of the feed system is
shown in Figure 3. Figure 3 highlights the fact that a gravimetric feeder is
employed to accurately quantify coal feed rate. The figure also illustrates
the location of static pressure taps in the coal feed system. These
pressures were documented throughout the test program for comparison to
performance targets.
Photo 3 shows the installed RO-II Burner register as viewed from outside the
furnace. Note the use of the tangential entry fuel nozzle inlet,
characteristic of both the Type R and RO-II burner designs. Photo 4 shows
the installed RO-II from the furnace side and highlights the shadow vanes and
divergent refractory quarl similar to typical field installations.
Note also in Photo 4 that there is refractory material on the furnace walls.
The laboratory test furnace has atmospheric pressure water cooled walls. The
furnace gas temperatures and heat release profile are adjusted by altering
the refractory configuration depending upon test objectives. The refractory
configuration selected for these trials was chosen to create a furnace
thermal environment where relatively high levels of thermal N0X would be
generated. In addition to refractory modifications, the test furnace was
intentionally operated at a volumetric cubic heat release rate of 39,800
Btu/hr/ft3. This volumetric heat release rate in effect far exceeded a
7B-5
-------
typical coal-designed boiler's volumetric heit release range of 9,000-16,000
Btu/hr/ft3. As a result of this (and combined with the refractory insulation
thickness and pattern in the furnace), measured furnace gas outlet
temperatures (horizontal furnace outlet plane) were in the 2500 - 2700°F
range, far exceeding typical boiler horizontal furnace gas outlet
temperatures of 1900 - 2000°F. The implication of high temperature furnace
operation during the RO-II laboratory trials is that N0X generated thermally
via the Zeldovich mechanism (1) was projected to be conservatively higher
than would be expected in subsequent field RO-II installations.
Eleven different coal nozzle configurations were evaluated during the
combustion trials. Baseline tests were conducted with a conventional Type R
nozzle; ten advanced Type R nozzle configurations were also evaluated. The
baseline nozzle (Figure 4) was comprised of the tangential fuel inlet, coal
stream deflector vanes, and a spirally rifled nozzle with a convergent tip.
A furnace side view of the baseline Type R coal nozzle is shown in Photo 5.
Combustion test data from the "Baseline" RO-II configuration is shown in
Figure 5 which depicts N0X (ppm corrected to 3% 02) as a function of flue gas
02 concentration. As is characteristic of a diffusion flame burner, NQX
increases with increasing excess air level. The primary point of the figure
is that at a nominal excess air level of 20% (approx. 3.5% 02), measured N0X
was approximately 650ppm (approx. 0.9 #/MBtu). Under all excess air
conditions, N0X exceeded the target value of 0.5 f/MBtu.
The most optimum coal nozzle arrangement of the ten tested is shown in
Figure 6. As shown in the schematic, the optimum RO-II coal nozzle retains
the tangential fuel/primary air inlet, deflector vane assembly, and spirally
rifled nozzle of the Type R design. The optimum RO-II arrangement includes
the addition of a venturi diffuser assembly, which is a channeled flow
control device, and a convergent nozzle tip with axial rifling vs. spiral
rifling as in the baseline case.
Photo 6 is a "furnace side" view of the optimum nozzle arrangement.
Figure 7 graphically depicts the NQX emission performance of a number of the
tested RO-II coal nozzle concepts. Data in this figure highlights the fact
that the coal nozzle design employed had a dominant influence on N0X levels
observed. One can summarize the data contained in Figure 7 by directing
7B-6
-------
attention to the solid line plotted in the center of the graph. All data
below the solid line represents the N0X performance of the venturi diffuser
concept, all data above the line represents alternative tested concepts.
Clearly, the venturi diffuser concept generated lower total N0X at any given
operating excess air level, as compared with all other tested coal nozzle
concepts. Host importantly, at a nominal flue gas 02 concentration of 3.5%
(20% excess air), total measured NO* was 360 ppm (corrected to 3% 02),
meeting the overall project goal of 0.5 #/MBtu N0X. The venturi diffuser
coal nozzle assembly, as a result of Its success in meeting the N0„ reduction
target established for the project, has been chosen as the coal nozzle design
to be utilized in the RO-II burner.
Beyond its N0X reduction capability, the RO-II burner met all other
established performance targets. These targets were set to ensure that the
firing system hardware developed in the laboratory would be retrofitable to
most existing wall fired boiler arrangements. Host units, for example, have
fan limitations in terms of achievable windbox to furnace delta static
pressure. The RO-II coal burner is capable of operation at less than 3.0"
W.C. static windbox to furnace delta pressure at MGR. Most existing boiler
F.D. fan systems are capable of achieving at least that pressure differential
at MGR. In a similar vein, primary (coal transport) air static pressure at
the coal nozzle inlet is a critical factor from a retrofit standpoint. Any
low N0„ burner installation should operate within existing coal feed system
pressure limitations. The RO-II burner operated at MCR with a primary air
static pressure at the nozzle inlet of less than 4.5 inches W.C., an
acceptable operating primary air static pressure for most existing wall fired
installations.
Many low N0X coal firing system laboratory tests and field demonstrations to
date have reported that, under low N0X conditions, carbon in fly ash levels
tend to increase (2, 3, 4). In some cases, CO emissions also increase under
low N0X conditions. These results are, of course, very dependent on coal
type, coal particle size distribution, and furnace configuration. In
practical terms, most low N0X coal firing systems must strike an acceptable
balance between N0X reductions and carbon in fly ash/CO increases. In the
case of the RO-II coal burner, operated at 0.5 I/106 Btu, both carbon in fly
ash and CO emissions did increase, however, the final emission levels
documented were within acceptable operating ranges. For example, under
baseline, high NOx conditions, carbon in fly ash and CO were 1-2% and 30-
7B-7
-------
50ppm, respectively. Under low (0.5 # MBtu) N0X conditions, carbon in fly
ash and CO increased to 3% and 40-70 ppm, respectively. These laboratory
results indicate that nominal increases in carbon in flyash may be expected
^ in RQ-II field applications, again dependent on coal type and furnace
configuration.
Several low N0K coal firing systems evaluated to date for wall fired boiler
applications have experienced increased flame lengths as compared to pre-
retrofit cases (5,6). As in the case of the relationship between N0„, carbon
loss, and CO, one must in most situations strike a balance between N0X
reductions and increasing flame length. Operating experience with the RO-II
coal burner to date is good in this regard. Baseline (high N0X) conditions
produced a luminous, stable flame about 12' long. Under low NGX (0.5 #/MBtu)
conditions, flame length increased to approximately 16'-18' long. The
increase in flame length was deemed acceptable because since the field units
targeted for the first RQ-II coal demonstrations can accommodate a similar
increase in flame length without direct flame impingement on rear wall tube
surfaces. Future boiler retrofits will be assessed on an individual basis
not only to ensure compatibility between furnace depth and the luminous flame
volume of the RO-II low N0X coal burner, but also to ascertain potential for
changes in post-retrofit boiler thermal performance.
Field Experience
Following successful laboratory development trials, the RO-II coal burner has
presently been retrofitted to three (3) field installations. Figure 8 is a
schematic of the as-installed RO-II coal burner. The tangential inlet,
spirally rifled coal nozzle with venturi diffuser assembly and convergent
nozzle tip can be seen in the figure. The pilot and main air plenums,
helical flow vanes, and shadow vanes are also depicted.
The current RO-II field installations are listed in Figure 9 with other
pertinent information. City of Jamestown Unit 10 and BPU Kansas City are
currently undergoing start-up and demonstration testing.
Conclusions
ABB-CE's RO-II coal burner, specifically designed for retrofit wall fired
7B-8
-------
boiler applications, has undergone extensive laboratory testing and is now
commercially available. In laboratory trials, the burner was shown to meet
the N0„ target of 0.5 #/MBtu firing Eastern U.S. bituminous coal while
limiting increases in carbon loss, CO, and flame length to commercially
acceptable levels. The burner also demonstrated the ability to operate
within the capacity of most existing boiler combustion air fan and coal
delivery systems in terms of static pressure requirements. The RO-II burner
offers advantages in terms of its simplified construction and operation. In
addition, the RO-II burner has the ability (via adjusting the main/pilot air
damper system) to equalize burner to burner combustion air flow imbalances in
multiple burner/common windbox plenum arrangements.
References
1. Zeldovich, Y. et al. (1947), Oxidization of Nitrogen in Combustion,
Academy of Sciences of the USSR, Institute of Chemical Physics,
Moscow-Leningrad, Translated by M. Shelf, Scientific Research
Staff, Ford Motor Co.
2. Beard, P. et al "Reduction of N0X Emissions form a 500 MM Front
Wall Fired Boiler" 1989 Joint EPA/EPRI Symposium on Stationary
Combustion NOx Control.
3. Grusha, 0. and McCartney M., "Development and Evolution of the ABB
Combustion Engineering Low N0X Concentric Firing System - 1991
Joint EPA/EPRI Symposium on Stationary Combustion NO, Control.
4. Kinoshita, et al "New Approach to N0X Control Optimization and
Unburnt Carbon Losses" - 1989 Joint EPA/EPRI Symposium on
Stationary Combustion N0X Control.
5. Clark, M.J. et al "Large Scale Testing and Development of the B&W
Low N0„ Cell Burner" - 1987 EPA/EPRI Symposium on Stationary
Combustion Nitrogen Oxide Control.
6. LaRue, A. et al "Development Status of B&W's Second Generation Low
N0X Burners - The XCL Burner" - 1987 EPA/EPRI Symposium on
Stationary Combustion Nitrogen Oxide Control.
7B-9
-------
Figure 3: Coal Feed System Schematic
7B-10
-------
Photo 3: Installed RO-ll Burner Register as Viewed
from Outside the Fumace
Photo 4: installed RO-ll Burner Irom the Fumace Side
Photo 5: Fumace Side View of the "Baseline" Type R
Coal Nozzle
Figure 5: "Baseline" Nozzle Assembly, NOx vs. 0,
TANGENTIAL
RjegPONtflYMR
MJET
'4
VEJNPTUR) DVFUBEH
"7
I
SnJFfTMQ BOX FOR—*]
:AWW. N&2Lg ADJLSTMSMT
» NOZZL£"nPWpTH
AJOALftFUN®
V\
VOfTUR GWU8ER
AEXJU5TMEHF RGQS
0&LECT0^VA«E3
Figure 6: Venturi Diffuser Nozzle Assembly, Test
Equipment Schematic
7B-11
-------
Photo 6: "Furnace Side" View ol the Optimum Nozzle
Arrangement for the RO-II Burner
Cu*tom«r
Unit
Unit
7yp*
Steam Plow
Ib/hr
No. of
Surnera
Fueia
04 ol Public UtIL
City of
s
CE-VU40
165,000
4
E.BU
Bd *1 Pufclte UliL
City of J&m««icwn
10
CE-VU40
166,000
4
£. en
Bd o» Public Utkl
Kanaaa City
Qutndare
Jnlt 2
Riley
1,126,000
S
Sub-Sir.
N&lurti Gaa
Propana
Figure 9: RO-II Experience List
Figure 8: RO-II Burner
7B-12
-------
NQX REDUCTION ON NATURAL GAS-FIRED BOILERS
USING FUEL INJECTION RECIRCULATION (FIR) -
LABORATORY DEMONSTRATION
Kevin C. Hopkins, David 0. Czerniak
Carnot
15991 Red Hill Ave., Suite 110
Tustin, CA 92680-7388
Les Radak
Southern California Edison Company
2244 Walnut Grove Avenue
P.O. Box 800
Rosemead, CA 91770
Cherif Youssef
Southern California Gas Company
3216 North Rosemead Blvd.
El Monte, CA 91731
James Nylander
San Diego Gas & Electric Company
P.O. Box 1831
San Diego, CA 92112
7B-13
-------
-------
ABSTRACT
Increasingly stringent NOx regulations on industrial and utility boilers may require
implementation of expensive post-combustion NOx control techniques. Fuel Injection
Recirculation (FIR) is a relatively low cost NOx reduction strategy for natural-gas
fired boilers in which the fuel is diluted prior to combustion with air, steam, or
flue gas. This technique is different from conventional flue gas recirculation (FGR)
because it is conceptually believed to impact prompt as well as thermal NO formation
mechanisms and is therefore capable of greater NOx reductions. Furthermore, the two
technologies when applied in conjunction are additive is terms of NOx reduction.
As a preliminary step towards full scale implementation of FIR, a laboratory
demonstration was performed to determine the feasibility of the technology. FIR was
demonstrated on a 2.0 MMBtu/hr test facility designed to simulate burners used on full
scale utility boilers. The test facility employed combustion air preheat, FGR,
staged-air firing, and was modified to inject flue gas, air, or saturated steam into
the fuel stream prior to combustion. The effectiveness of FIR was determined at
varying injection rates, firing rates, air preheat levels, FGR rates, and excess 02
conditions.
Results show that FIR is more effective that FGR in reducing NOx, and that a
additional 50% NOx reduction was achieved when FIR is used in conjunction with FGR.
The test program demonstrated that in a full-scale application, FIR may be capable of
reducing NOx to low levels, at an attractive cost relative to post-combustion control
retrofits.
7B-15
-------
INTRODUCTION
Carnot was contracted by the Southern California Gas Company, the Southern California
Edison Company (SCE), and the San Diego Gas and Electric Company (SDG&E) to perform
a laboratory demonstration of a potential new NQX reduction technology for gas-fired
boilers which has been designated Fuel Injection Recirculation (FIR). As a
preliminary step towards full-scale implementation, this demonstration program was
performed to determine the feasibility of the technology.
Fuel Injection Recirculation involves recirculation of a portion of the boiler flue
gas and mixing it with the gas fuel at some point upstream of the burner.
Additionally, the FIR concept can be expanded to include the fuel injection of any
inert diluent such as steam or air. This method conceptually is believed to be
capable of greater N0X reductions than can be achieved by conventional Flue Gas
Recirculation (FGR), which is mixed with the combustion air. Furthermore, it is
anticipated that when implemented on a utility boiler, the two technologies would be
to some extent, additive in terms of N0X reductions, ultimately resulting in very low
N0X emissions. The principal motivation for pursuing this concept is the potential
cost benefit in comparison post-combustion NOx control technologies such as SCR and
urea injection, which are presently being considered to meet the stringent new NQX
limits specified in the South Coast Air Quality Management District Rules 1135 and
1146. The FIR concept is also attractive because full-scale application of FIR would
require relatively few modifications to existing equipment.
The approach taken for this laboratory demonstration program was to apply the FIR
technology on a test facility which incorporates many key design and operational
attributes of burners in use on utility boilers, and which employs N0X control
techniques commonly used in these large scale boilers. The primary emphasis of the
this feasibility study was a practical evaluation of FIR over ranges of important
operating conditions such as firing rate, air preheat, overfire air, and FGR.
7B-16
-------
TECHNICAL OBJECTIVES
Throughout this study, FIR was evaluated primarily in terms of flue gas concentrations
of NQX, 02» COz, and CO, and in terms of burner stability and flame characteristics.
The specific technical objectives of the investigation were as follows:
1. Evaluate the NO^, reduction effectiveness of FIR using a
laboratory-scale burner similar in design and thermal
environment to burners used on electric utility boilers.
2. Evaluate the NO reduction efficiency of FIR alone, and in
combination with FGR.
3. Evaluate the effect of FIR on minimum operable 02 level, and
on burner stabi1ity.
4. Evaluate the effect of reduced firing rate on the
effectiveness of FIR.
5. Evaluate the effect of air staging on the effectiveness of
FIR.
6. Compare the effect of air relative to flue gas as the FIR
diluent.
7. Compare the effect of steam relative to flue gas as the FIR
diluent.
BACKGROUND
Fuel Injection Recirculation (FIR) is a potential new N0X control strategy for natural
gas-fired boilers which is defined as the injection of any inert diluent into the fuel
gas at some point upstream of the burner. The concept originally involved the
extraction of flue gas from the exit of the boiler, cooling it If necessary, and
finally compressing it for injection at gas header pressures into the fuel line.
Operating expenses and equipment costs may be reduced by injecting other diluents such
as air or steam, or by lowering gas header pressures through burner modifications.
FIR and Prompt NO Formation: N0X formation in natural gas-fired boilers is associated
with two mechanisms known as thermal NO and prompt NO. Thermal NO refers to the high
temperature reaction of nitrogen and oxygen from the combustion air. This mechanism,
which is commonly termed the "Zeldovich" mechanism after its discoverer, is thought
to occur in the post-flame or burned gas zone. Low excess air firing, flue gas
recirculation, burners-out-of-service (BOOS), and air staging are commonly used on
utility boilers to control thermal NO formation.
The existence of another NO formation mechanism was first suggested by Fenimore whose
measurements showed that reactions other than the Zeldovich mechanism were taking
place, and that some NO was being formed in the flame region. Because of the early
7B-17
-------
formation of NO, Fenimore coined the name "prompt" NO. Fenimore proposed that C2 and
CH radicals present in hydrocarbon flames contribute to the formation of prompt NO.
The formation of prompt NO is greater in fuel-rich flames, and decreases with the
increase in local 02 concentrations. Similar experiments have shown that prompt NO
formation is a function of flame temperature as well as stoichiometry. Other
measurements made in flat flame burners demonstrate that prompt NO can account for 10-
40 ppm of the total NO formed. In utility boiler systems, prompt NO is assumed to be
less than 50 ppm while the thermal NO contribution can be as high as 125-200 ppm.
Thermal NO control techniques such as FGR and BOOS can decrease NO to concentrations
approaching prompt NO concentrations. The South Coast Air Quality Management District
Rule 1135 for utility boilers will require N0X emission limits translating to about
25 ppm, and therefore the control of prompt NO formation is important if new emissions
limits are to be met without installation of expensive post-combustion control
techniques.
FIR appears to be a effective and relatively inexpensive technique for the control of
prompt NO formation. It is believed that FIR reduces prompt NO formation by diluting
the fuel prior to combustion thereby reducing the concentration of hydrocarbon
radicals which, produce prompt NO. In addition, FIR also acts like FGR in reducing
thermal NO production. It is anticipated that FIR in combination with FGR, could
reduce NOx emissions to levels below 25 ppm by controlling both NO formation
mechanisms.
TEST DESCRIPTION
Test Facility: The laboratory facility selected for this evaluation of FIR was an 80
hp Scotch fire-tube boiler. This boiler was modified to incorporate many significant
components of a full-scale utility boiler furnace. The test facility comprised the
fire-tube boiler, which is capable of firing up to 3.0 x 106 Btu/hr on natural gas,
a forced draft fan, a separately fired air preheater (APH), a 5 1/2" diameter gas fuel
ring, a ceramic quarl, and a windbox with a sixteen blade variable air register. Off-
stoichiometric firing was achieved by diverting a portion of the pre-heated combustion
air to the overfire air (0FA) ring placed downstream of the burner face. A separate
fan was used to recirculate a portion of the flue gas to the combustion air (FGR).
The FGR flowrate was determined by measuring the windbox 02 concentration along with
the flue gas 02 concentration. The mass flowrate of the flue gas recirculated was
subsequently determined from stoichiometric calculations.
Natural gas was supplied to the boiler via a 10 psig supply, and metered using a
rotameter. The maximum firing used in this study was 2.0 x 106 Btu/hr. The burner
consisted of 3/8 inch ring with 11 equally spaced holes drilled radially, each of
7B-18
-------
0.189 inches diameter. The ceramic burner quarl, six inches long with a nine-inch
exit diameter, was geometrically similar to those used on small Peabody ring burners
in utility boilers. The air register vanes were set initially to target a baseline N0X
level characteristic of full-scale units. The air register vanes were set at 20° off
radial and were not varied throughout the remainder of the tests.
The FIR concept was tested using three fuel diluents; flue gas, air, and saturated
steam. Most of the testing was performed using flue gas as the diluent. The flue gas
injection system consisted of a 5 hp rotary lobe type compressor capable of a delivery
pressure of up to 8 psig at a flow rate of 30 scfm of flue gas.' Flue gas, extracted
at the stack plenum, was compressed and injected into a 2 inch fuel line through a
sparger. FIR tests with air injection were performed using the same configuration as
above with the inlet to the blower disconnected from the stack plenum.
Steam injection was accomplished using a separately fired 2-1/2 hp Parker Boiler
providing saturated steam at approximately 180 psig. The flow rate was controlled
using a gate-valve and was metered using an Annubar flow sensor. Steam was injected
through the sparger into a heat-traced fuel line.
Test Conditions: The principal objective of this laboratory demonstration program was
to determine the effectiveness of FIR in reducing NQX at conditions characteristic of
large industrial or utility boilers. Conditions and parameters which significantly
impact NO^ on full-scale units include combustion air temperatures, off-stoichiometric
firing, excess air levels, load variations, flue gas recirculation to the combustion
air, burner configuration, and air register orientation. It was not practical to
systematically investigate the influence of each of these characteristics in the
laboratory facility. Once baseline configurations were established, the burner
hardware and the air register orientation were not changed throughout the testing.
Excess Air Levels: Tests were performed at a "minimum" or "nominal" excess 02
condition. The minimum 02 condition was defined by the following criteria:
1. the excess air level producing 200 - 400 ppm CO, or
2. an excess 02 concentration of « 0.3 %
The second criteria was necessary because at some test conditions, CO did not exceed
100 ppm even at extremely low 02 concentrations. The 0.3 % 02 concentration was
necessary as a lower safety limit for those tests where CO remained below 100 ppm.
The nominal 02 condition was defined as the amount of excess air necessary to increase
the minimum 02 concentration by 0.5 %.
Flame Characteristics: Since- an important objective of this test program is to
determine the limits of applicability of FIR with respect to flame characteristics,
7B-19
-------
the test series involving fuel dilution with flue gas, steam, or air, the diluent was
added to the point of flame instability. Flame stability and general flame
characteristics were determined primarily form observations. The flame was considered
to be unstable if any of the following was observed:
1. Any tendency for the flame to lift-off from the burner face
and re-attach downstream on the OFA ring.
RESULTS AND DISCUSSION
The results of the Fuel injection Recirculation (FIR) test program are presented in
this section. The NO, results presented below are expressed in ppm corrected to 3%
0-, on a dry basis. The N0X reductions achievable, and the limitations in terms of
flame stability are considered for FIR used in conjunction with varying firings rates,
flue gas recirculation rates, air preheat levels, and both with, and without overfire
air. For each test series, the Injection rate of flue gas was increased until the
limit off flame stability was reached. The flame stability limit is defined as the
maximum injection rate at which the flame remains attached to the burner face.
(Higher injection rates would cause the flame to detach from the burner face and re-
attach to the overfire air ring).
For the purposes of later comparison, the "baseline" condition is defined by the
following parameters:
The baseline N0X concentration for this test facility was 87.6 ppm § 3% 02. Without
OFA, the N0X concentration was 167.6 ppm 9 3% 02. The use of OFA reduced N0X by 48%.
This is consistent with full-scale N0X reductions attainable using N0X ports and/or
burners-out-of-service (BOOS). The effects of other parametric variations are
presented below.
Summary of Baseline Characteristics
2. Excessive fluctuations in furnace draft
3. Excessive fluctuations of N0X, CO, or 02 concentrations.
• firing rate:
• 0? condition:
2,000,000 Btu/hr ± 2 %
minimum (defined by CO « 200-400 ppm)
nominal (defined by « 10% of total air)
480 - 491 °F
0 %
• OFA condition:
• APH temperature:
• Windbox FGR:
approximately 10% overfire air with a combustion air
temperature of approximately 490 *F.
• N0X is very sensitive both to excess air level and to
combustion air temperatures, especially at lower FGR rates.
7B-20
-------
• The measured N0X vs FGR relationship is typical of full-
scale units.
• The NO vs firing rate relationship is characteristic only
of smaller industrial boilers.
Flue Gas as FIR Diluent
The effect of Fuel Injection Recirculation using flue gas as the diluent is presented
in this section. The amount of FIR injection is expressed in two ways. First as a
percent fuel dilution defined as the percentage of the volume of flue gas injected to
the total volume flow through the burners. Alternatively, for the purposes of
comparison to conventional flue gas recirculation, it is expressed as the percent of
the weight of the flue gas injected to the total weight of the air and fuel.
FIR vs Windbox FGR: The effect of FIR without windbox flue gas recirculation (FGR),
and at an optimum and maximum FGR rates are presented in this section. The firing rate
is 2.0 x 106 Btu/hr with nominal OFA at the minimum 02 condition. The results are
shown in Figure I-A and 1-B.
Figure 1-A shows N0„ concentration vs FIR injection rate expressed as percent fuel
dilution. N0X decreases uniformly with increasing FIR injection. With no windbox
FGR, the rate of decrease is approximately 1.7 ppm per % fuel dilution. At higher
windbox FGR rates, the rate of decrease is approximately 0.6 ppm per % fuel dilution.
The decreasing effectiveness at higher windbox FGR rates indicates that FIR reductions
are partially thermal NO, reductions and that the two techniques are to some extent
redundant. However, since further decreases are measured even at the maximum windbox
FGR rate, the two techniques also appear to be additive.
This additive effect can be more clearly seen in Figure 1-B where the effect of FIR
on N0X is plotted as a function of the total flue gas recirculated (to windbox and to
fuel). For each of the three data sets shown on the graph, the windbox FGR is held
constant while the FIR flowrate is increased. The dotted line on the graph defines
the relationship between NQX and the windbox FGR alone. At both the 15% and 23%
windbox FGR rates, FIR injection is capable of additional reductions of approximately
50%. Table 1 summarizes the maximum reductions achievable with FIR when used in
conjunction with FGR. Furthermore, it is evident that FIR alone is more effective
than FGR: 5% of the flue gas injected into the fuel results in lower NOx than 23% flue
gas injected into the combustion air. This is shown graphically in Figure 2 where N0X
reduction is plotted vs the total flue gas recirculated. The N0X reduction curve
rises more steeply with FIR than without. It should be re-stated here that the flue
gas recirculation to the fuel requires significantly higher compression that
recirculation to the combustion air.
7B-21
-------
As postulated earlier, FIR is believed control prompt, as well as thermal NO, both by
reducing peak flame temperatures and by lowering the concentration of certain
hydrocarbon radicals which are thought to produce prompt NO. The concentration of
prompt NO formed in utility combustion systems is thought to be 25 ppm or less. The
tests performed in the present study are not intended to distinguish between prompt
NO reductions and thermal NO reductions, or even to confirm the existence of prompt
NO. It is not possible to conclude whether the additive N0X reductions are due to
more efficient mixing of flue gas with the fuel and air, or whether FIR actually
suppresses prompt NO formation. What can be concluded however is that FIR is more
effective than windbox FGR, and that together there is a measurable additive benefit.
The use of FIR does not significantly affect flame stability up to a fuel dilution
ratio of approximately 35% Higher injection rates create a tendency to lift off the
burner face creating a "boiler rumble" and large fluctuations in N0X and 02 and furnace
draft. At lower injection rates, the appearance of the flame is not significantly
different from the flame appearance with no FIR injection, other than decreased
brightness which is indicative of lower peak flame temperatures.
The Effect of Overfire Air on FIR: The effect FIR when used without overfire air is
shown in Figure 3-A and Figure 3-B. FIR is equally effective with, or without
overfire air. Without OFA, FIR reduces N0X concentrations by 60% at 0% FGR and 15%
FGR.
It was also expected that overfire air would affect flame stability by decreasing the
burner throat velocities. The tests demonstrated that overfire air does not affect
flame stability. Figure 3-A shows that the limit of flame stability is approximately
at 35% fuel dilution regardless of the OFA rate. Figure 3-8 shows that the effect of
overfire air has a decreasing effect at higher FGR rates. For example, at 15% windbox
FGR with the maximum FIR injection rate, 10% air staging results in less than a 5 ppm
N0X reduction.
The Effect of Firing Rate On FIR: The effect of FIR at three firing rates is
presented in Figure 4, FIR injection using flue gas results in approximately the same
N0X reductions at 1.0, 1.5, and 2.0 x 106 Btu/hr. The slopes of the curves on Figure
4 are not a function of the firing rate. With no windbox FGR, FIR reduces N0X at
approximately 5 ppm/%fuel dilution up to 35% fuel dilution. At an optimum windbox FGR
rate, the slope decreases to .6 ppm/%fuel dilution up to 35% fuel dilution.
It is important to note that reduced firing does not significantly affect flame
stability. The limit of flame stability occurs at approximately 35% fuel dilution at
each firing rate tested. It is difficult to extrapolate this characteristic to the
full-scale application primarily due to the non-characteristic N0X vs firing rate
7B-22
-------
relationship, i.e. the relative increase in N0X at the mid-firing rate. It is also,
important to remember that the minimum 02 condition at the lower firing rates results
in significantly higher 02 concentrations. The air register vane setting is likely
to affect flame stability and the minimum 02 condition, however the effect of air
register adjustments was not examined during this test program.
Air As FIR Diluent
The original concept of Fuel Injection Recirculation involved injecting flue gas into
the fuel. In principle any diluent could have the same affect on prompt NO formation.
The advantage of using air as a fuel diluent is that compressing dry air up to fuel
pressures is less expensive than compressing hot flue gas. In addition, problems with
moisture condensation in the fuel delivery system are eliminated if air is used
instead of flue gas. The effectiveness of air injection was explored in a limited
test matrix intended to compare air to flue gas as FIR diluents.
Air was injected as an FIR diluent at the following conditions: high combustion air
temperatures, at a nominal overfire air rate, and at two FGR rates. The results are
shown in Figure 5, where the results for flue gas injection are re-plotted for
comparison. These results demonstrate that air injection is not as effective as flue
gas injection in overall N0X reductions. For the 0% FGR case, NOx actually increases
at low air injection rate. The characteristic is not measured at the 15% FGR
condition. Table 2 shows that the overall N0„ reductions achieved using air injection
are less than half of the reduction measured using flue gas injection.
Steam as FIR Diluent
Steam is another fuel diluent which in principle should reduce NOx much the same way
as flue gas. The use of steam as an FIR diluent for full-scale application may be
attractive on a cost basis since it would require no additional compressors. Provided
that steam could be extracted at relatively low pressures, the impact on boiler heat
rate should not be prohibitive. The use of steam injection as a means of NOx control
on large boilers is not a new technique. However, it is usually injected into the
combustion air upstream of the burner rather than into the fuel.
Particular experimental difficulties precluded a more expanded test matrix with steam
injection. The primary difficulty was the high fluctuation in steam flow: the
flowrate fluctuated by approximately 25 %. This made measurement of steam flow rate
difficult and caused high fluctuations of N0X and especially CO. Figure 6 shows a
example time trace taken from data logger records. Note that the N0X has been
corrected to 3% 02. NOx, CO, and 02 fluctuated regularly at the same frequency of the
steam generator fluctuation. The period of the fluctuation was approximately 4
minutes. As the steam flow cycled to a maximum, about 62 Ib/hr, the N0X reached a
7B-23
-------
minimum, and CO was in excess of 1000 ppm. At the minimum steam flow, about 48 Ib/hr,
the N0X reached a relative maximum, and CO reached a minimum. Since the fuel flow
could not be adjusted for changes in back pressure, the fuel flow also cycled causing
small fluctuations in 02. Despite the fact that the steam generator flow rate could
not be held constant, the results generated are still valuable. The steam flow cycled
in a very regular, repeatable manner, and accurate data were obtained by averaging the
continuous emissions data over many cycles,
The results of the steam injection test are presented in Table 4-11 and in Figure 7.
The steam injection tests were performed without overfire air. When steam injection
was used in conjunction with overfire air, excessively high CO emissions resulted as
well as poor flame stability. Overall NO^ reductions are 54% without FGR, and 36%
with 15% FGR. Figure 7 presents a comparison of steam injection and flue gas
injection. Also shown on this figure are the minimum and maximum NQX concentrations
corresponding to the maximum, and minimum steam flow rate. The results show that with
no overfire air, steam injection is nearly as effective as flue gas injection.
CONCLUSIONS
Fuel Injection Recirculation (FIR) was demonstrated on a laboratory scale test
facil ity designed to simulate the significant combustion characteristics of full-scale
utility natural gas burners. FIR was evaluated in terms of N0X reductions and burner
stability. While, the absolute values of NO, emissions results presented in this
report should not extrapolated directly to full-scale boilers, relative N0X reductions
and general trends measured on the sub-scale facility, should be representative of
results expected on full-scale units. The major conclusions drawn from the laboratory
evaluation are presented below:
Baseline Characteristics
• At test conditions typical of utility boilers, the baseline
N0X concentrations on the . sub-scale facility are
representative of full scale units.
• The measured NQX dependencies on FGR, air staging, air
•preheat temperatures, and excess , air Tevels are
representative of trends seen in full scale units.
• The measured relationship between N0X and firing rate is
typical of smaller package boilers. *
Flue Gas as FIR Diluent
• FIR is an effective NQX reduction technique to be applied to
natural gas-fired boilers, and N0X reductions achieved using
this technique are additive to those achieved by windbox FGR
and air staging.
7B-24
-------
• FIR is more effective than windbox FGR, per pound of flue
gas recirculated, in reducing N0X emissions.
• FIR in combination with air staging and windbox FGR results
in additional NO reduction of approximately 50%. NO
concentrations below 25 ppm were achieved at full load with
nominal air staging, 15% FGR and 35% fuel dilution.
• FIR has no adverse effects on maintaining minimum 02 levels.
• FIR is equally effective at reduced firing rates and when
used without overfire air.
• FIR operates with good flame stability at high combustion
air temperatures and nominal air staging at FIR levels up to
35% fuel dilution. However, the maximum level of FIR
consistent with acceptable burner stability decreases with
decreasing combustion air temperature.
• With no air staging, FIR operates with good flame stability
at low combustion air temperature up to a 35% fuel dilution.
Air as FIR Diluent
• Air as an FIR diluent is less effective than flue gas and
leads to flame instabilities at lower injection rates.
Steam as FIR Diluent
• Steam as an FIR diluent when applied in combination with air
staging results in poor flame stability and high CO
concentrations.
• Steam when applied with no air staging is nearly as
effective as flue gas as an FIR diluent.
• CO concentrations are generally higher with steam than with
air, or flue gas as the FIR diluent.
7B-25
-------
STEAM INJECTION, O OFA, 0% FGR
1.1
1.0
0.9
a
0.8
<
o
0?
06
-J
3
05
u.
u.
0.4
O
ss
0.3
0.2
01
0.0
TEST 186,107
19 November 1990
1026
10:40
10:55
11:09
TIME
Figure 6. Example Emissions Time Traces with Steam Injection
1B0
160
140
_
N. \ \
-
o
\S-\
\ •average
0 \
-
15% FGR
MINIMUM \
steam row V
_ • 'I MAXIMUM
¦v —STEAM FLOW -
-
1 > 1 • i 1
.... i , i ... .
1. . , , 1
\
* MAXIMUM
STEAM FLOW
i , i . , .1 , . • , . '
0 5 10 15 20 25 30 35 40 45 SO
PERCENT FUEL DILUTION
Figure 7. Effect of Steam Injection
7B-26
-------
110
100
90
80
70
60
50
40
30
20
10
0
_l r
_AIR INJECTION
FIRING RATE: 2.0 x 10® Btu/hr
APH: 490 °F
NOMINAL OFA
MINIMUM 02
—At AIR INJECTION
—A— FLUE GAS INJECTION
10 15 20 25
30
35
40 45
50
PERCENT FUEL DILUTION
Figure 5. Effect of Air as FIR Diluent
7B-27
-------
PERCENT FUEL DILUTION
Figure 4. Effect of FIR at Three Firing Rates; NOx vs. Dilution
7B-28
-------
180
160
140
120
100
80
60
40
20
0
"I i. i' i"'i
i i-i I, |11 i i r i1
O 0% WB FGR
X
-s.
x
FIRING RATE: 2D MMBtu/hr
APH: 4S8=F
MINIMUM Qj
NOMINAL OFA
NO OFA
15*
15% WB
10 15 20 25 30 35 40 45 50
PERCENT FUEL DILUTION
Figure 3A. FIR With and Without Overfire Air
160
160
140
O 120
S?
F>
@ 100
a. so
a
O 60
z
40
20
0
0* WB FGR
I ,
FIRING RATE: 2.0 MMBtu/hr
APH: 4ae°F
MINIMUM 02
-A- NOMINAL OVERFIRE AIR
—O— NO OVERFIRE AIR
\0 15% WB FOR
\D
10
15
20
.25
30
PERCENT FLUE GAS RECIRCULATION
(FGR + FIR)
Figure 3B. Effect of FIB with and Without Overfire Air
7B-29
-------
PERCENT FLUE GAS RECIRCULATION
Figure 2. Maximum NOx Reduction with FIR
7B-30
-------
110
100
90
BO
70
60
50
40
30
20
10
0
_ T r j—r 1 , j r 1 , !
FIRING RATE: 2.D MMBluflw
—
APH: 668±e0F
NOMINAL OFA
1
MINIMUM O2
0* WB FGR
-i
:
_Z
A :
- U\15% WB FGR \
v.
A
;
~ j
23% WB FGR _
1 i i , 1 , 1 , 1 , 1 , 1 ,
1 , 1 , 1 1 1 , :
0 S 10 15 20 25 30 35 40 45 SO
PERCENT FUEL DILUTION
Figure 1A, Effect of FIR at Threa FGR Rates; NOx vs. Percent Fuel Dilution
110
100
SO
80
70
60
50
40
30
20
10
0
0% WB FGR
FIRING RATE: 2.0 MMBlu/hr
APH; 48S ± B°F
NOMINAL OFA
MINIMUM 0 2
15% W0 FGR flSc
23% WB FGR
10
15
20
25
30
PERCENT FLUE GAS RECIRCULATION
(FGR + FIR)
Figure 1B. Effect of FIR at Three FGR Rates; NOx vs. Total FGR
78-31
-------
TABLE 1
MAXIMUM NO, REDUCTIONS WITH FIR
Windbox NO @ 3% O,
FGR,% 0% FIR Max FIR % Reduction
0 ¦ 89.2 37,8 57.6
15 40.6 21.9 46.1
23 35.3 17.0 51.8
TABLE 2
COMPARATIVE NO, REDUCTIONS;
AIR INJECTION VS FLUE GAS INJECTION
Air Injection Fiue Gas Injection
0 MAX , 0 MAX
FIR FIR ^Reduction FIR FIR ' %Reduetion
0% FGR 94.1 73,8 21.6 89.2 33,1 62.9
15% FGR 41.0 31.2 23.9 40.6 21.9 46.1
NOTES 1. Firing Rate = 2.0 x 10' Btu/hr
2. Nominal OFA '
3/ Minimum 02
4. APH = 490 °F
7B-32
-------
ADVANCED REBURNING FOR NO- CONTROL
IN COAL FIRED BOILERS
S, L. Chen
W. R. Seeker
R. Payne
Energy and Environmental Research Corporation
18 Mason
Irvine, California 92718
(714)859-8851
7B-33
\
-------
s
-------
ABSTRACT
This paper summarizes an experimental study which was conducted to investigate the
chemical constraints of the reburning process and identify advanced reburning
configurations for optimal N0X reduction in coal-fired boilers. Tests were performed
initially on a bench scale tunnel furnace to characterize and optimize the fuel-rich
reburning zone and fuel-lean burnout zone independently. Based on the results, an
advanced reburning process was designed which integrated reburning with selective
reducing agent injection to enhance the burnout zone efficiency. The concept was
subsequently tested in a pilot scale facility and yielded over 80 percent reduction in
N0X emissions.
7B-35
-------
INTRODUCTION
Reburning is an NOx control technology which uses fuel to reduce NO.1"* The main heat
release zone can be operated normally to achieve optimum combustion conditions without
regard for N0K control. With reburning, a fraction of the fuel is injected above the
main heat release zone. Hydrocarbon radicals from combustion of reburning fuel react
with nitric oxide to form molecular nitrogen. This process occurs best in the absence
of oxygen. Thus sufficient reburning fuel, between 15 and 20 percent of the total heat
input, must be added to produce an oxygen deficient reburning zone. Subsequently, air
is provided to combust fuel fragments which remain at the exit of this zone. Since
reduced nitrogen species NH3 and HCN are also present, air addition may allow a further
N0X reduction.
Previous studies showed that 60 percent reduction in N0X emissions could be achieved
with natural gas reburning.5 Recently research has been conducted to examine and
enhance the N0X reduction chemistry in the burnout zone.6 The burnout zone can be
considered as an excess-air "flame" burning the remaining fuel fragments from the
reburning zone. Oxidation of the fuel fragments, particularly CO, could generate a
significant amount of radicals via chain branching:
CO + OH = C02 + H
H + 02 = OH + 0
0 .+ HaQ = OH + OH
These radicals play an important role In the conversion of XN species to N2 or NO
during burnout.
Figure 1 is an experimental examination of the burnout zone chemistry, in particular,
the conversion efficiency of NH3 to N2. The rich zone was assumed to supply 600 ppm
each of NO and NH3, or an N to NO ratio of 1.0. Under excess air conditions, ammonia
gas was mixed with various amounts of CO and injected at temperatures between 1300 and
2200°F. The solid symbols represent the injection of NHS alone, which is basically a
7B-36
-------
simulation of Thermal De-NOx. For the open symbols, 0.2 percent CO was included with
NHj, thereby yielding a burnout like environment. The presence of CO lowered the
optimum temperature for N0X reduction from 1800°F to 1500°F. It is readily apparent
that a reduction in the burnout temperature from the 2200-2400°F normally employed in
the reburning process would increase the conversion efficiency of NH3 + NO to N2
because of the presence of CO.
This paper summarizes the results of a pilot scale study which was undertaken to
investigate the possibility of positive synergism between the injection of selective
reducing agents, such as ammonium sulfate, to provide the reducing specie NH3,and
combustion modifications, such as reburning,to serve as the source of CO.
EXPERIMENTAL
The 3.0 MWt, down-fired tower furnace5 used in the pilot-scale investigations was
refractory-lined and water jacketed with inside dimensions of 1.2 x 1.2 x 8.0 m. The
four main diffusion burners each consisted of an inner pipe for axial primary fuel
injection and an outer pipe, equipped with swirl vanes, for the main combustion air.
This four burner array produced relatively uniform velocity and composition profiles
at the primary zone exit. The furnace contained seven rows of ports for reburning fuel
and burnout air injection. The temperature profile was manipulated by insertion of
cooling panels, positioned against the furnace walls. The reburning fuel and burnout
air injectors were designed to maintain jet mixing similarity between the pilot-scale
furnace and a full scale boiler based on empirical correlations for entrapment rate
and jet penetration.
Exhaust gas samples were withdrawn through a stainless steel, water-jacketed probe and
analyzed for N0X (chemiluminescence), 02 (paramagnetic), C0/C02 (NDIR), and S02 (NDUV).
A water jacketed probe with an internal water quench spray near the front end was used
for extracting in-flame samples. Gas phase HCN and NH3 species were collected in a gas
washing unit and subsequently analyzed for CN" and dissolved ammonia using specific ion
electrodes. Gas temperatures were characterized with a suction pyrometer.
RESULTS
Recent studies6 have suggested that the key parameters for the enhancements of burnout
zone chemistry in staged combustion or reburning are:
• Reaction temperature (850°C)
• CO levels (0.5% or less), and
• NHj species.
7B-37
-------
Advanced Rebornino
Apparently the conventional reburning process does not provide the required
environment. An advanced reburning process, which combines reburning with selective
N0X reduction (SNR) via ammonium sulfate injection, was designed. Figure 2 shows two
hybrid schemes with 20 percent and 10 percent gas reburning, respectively. With 20
percent reburning (SR2 ¦ 0.9), the burnout air was divided into two streams to yield
an SRj of 1.03 and an SRt of 1.15. With 10 percent reburning, the reburning zone
stoichiometry (SR2) was 1.03 and the burnout air stoichiometry (SRt) was 1.15. In both
cases, an aqueous solution of ammonium sulfate was atomized with the final burnout air
an injected at 850°C at an N to NO molar ratio of 1.5. . '
Verification Tests
Figure 3 shows the advanced reburning results obtained with natural gas as the primary
fuel. The natural gas fired at 4.5 x 10s Btu/hr was doped with NH3 to yield primary
N0X levels of 600 and 400 ppm (dry, 0 percent 02). Twenty and ten percent advanced gas
reburning were applied, respectively. Similar final emissions, approximately 125 ppm
NOx, were achieved with both concepts. Experiments were subsequently carried out with
an Indiana coal as the primary fuel. The Indiana coal produced an uncontrolled N0X
emission of 800 ppm (dry, 0 percent 02) at 15 percent excess air. The primary N0X at
SR, = 1.13 was 680 ppm. Figure 4 presents the results and indicate that as seen in
the bench scale studies6, both advanced concepts were equally effective in NQX
reductions. It is apparent that there exists a tradeoff between natural gas premiums
and the cost of ammonium sulfate.
Ammonia Slip and SO.. Emissions .
The injection of ammonium sulfate into the furnace has a potential of producing
unwanted emissions such as NH3 and S02/S03. A series of exhaust measurements were made
to evaluate the slip of ammonia using selective ion electrode and the emissions of S02
and SOj via controlled condensation during the Indiana coal tests. Exhaust NH3
concentrations were negligible in all cases including those obtained with Utah coal and
natural gas as the primary fuel. Higher S02 emissions were obtained with 10 percent
gas reburning. However, the uncontrolled S02 level was maintained with 20 percent gas
reburning due to dilution. No increase in S03 emissions was observed for both cases,
suggesting favorable conversion of the sulfate to S02.
Thus, there exists a control strategy to prevent the increase in S02 emissions due to
injections of ammonium sulfate. For the application of advanced reburning to high
7B-38
-------
sulfur coals, 10 percent gas reburning is recommended, whereas for low sulfur coal
applications, the 20 percent gas reburning concept is preferred.
CONCLUSIONS
In summary, these results suggest that selective reducing agents can be combined with
combustion modification techniques to provide NO^ reductions that are larger than those
that are possible by applying the technologies simultaneously but separately. By using
the stoichiometry control associated with reburning to produce a slightly fuel rich
region for selective reducing agent injection, reductions can be achieved at relatively
Tow temperatures without the use of stainless steel or other catalysts.
ACKNOWLEDGEMENTS
This work was primarily supported by the U.S. Department of Energy, Pittsburgh Energy
Technology Center (Contract No. DE-AC22-86PC91025) with Dr. Richard Tischer as the
Project Manager. We also would like to acknowledge the contributions of our colleague
Mr. Loc Ho in conducting the experiments.
DISCLAIMER
Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise, does not necessarily constitute or imply its
endorsement, recommendation, or favoring by the United States Government or any agency
thereof. The views and options of authors expressed herein do not necessarily state
or reflect those of the United States Government or any agency thereof.
REFERENCES
1. Myerson, A. L., et al., Sixth Symposium (International) on Combustion, The
Combustion Institute, 1957, p. 154.
E. Reed, R. D., "Process for the Disposal of Nitrogen Oxide." John Zink Company,
U.S. Patent 1274637, 1969.
3. Wendt, J. 0. L., et al., Fourteenth Symposium (International) on Combustion, the
Combustion Institute, 1973, p. 897.
4. Takahashi, Y., et al., "Development of Mitsubishi 'MACT' In-Furnace N0„ Removal
Process." Presented at the U.S.-Japan NQX Information Exchange, Tokyo, Japan, May
25-30, 1981. Published in Mitsubishi Heavy Industries, Ltd. Technical Review,
Vol. 18, No. 2.
5. Chen, S. L., et al., 21st Intl. Symp., Combustion Institute, 1986, p. 1159,
6. Chen, S. L., et al, JAPCA. Vol. 39, No. 10 (1989).
7B-39
-------
100
SR - 1.1
(NO]^p - 600 PPM (DRY, 0% 02>
- N/(NOx)p - 1.0
O NH3 + 0.2% CO
• NH3 ONLY
1400 1600 1800 2000 2200
PEAK INJECTION TEMPERATURE <°F)
Figure 1. conversion in the "burnout zone.
FUEL + AIR
X
20%
NAT. GAS
AIR —•
AIR AND
(NH^2 SO4
110
0.90
1.03
115
T
10%
NAT. GAS
FUELj* AIR
AIR AND^
(NH4)2SO^
Figure 2. Advanced reburning concepts.
7B-40
-------
I 1 Rebuming
V77X Advanced Reburning
600
500
400
300
200
100
Primary NOx
o
3
¦o
<1>
QC
o
CO
i
Primary N0X
e
o
a
¦o
«>
QC
20% Gas
10% Gas
Figure 3. Results obtained with natural gas as primary fuel,
7B-41
-------
/
800
(S1
*
o
600
>
DC
9 400
3E
£L
Q.
X*
o
Z 200
UNCONTROLLED NO
JJL
INDIANA COAL
eft
<
0
#
o
CM
m
<
0
#
o
z
o
I""
o
III
cc
*
CO
CO
1
Figure 4. Pilot scale results with Indiana coal
7B-42
-------
LARGE SCALE TRIALS AND DEVELOPMENT OF FUEL STAGING IN A 160 MW COAL
FIRED BOILER
H. Spliethoff
Universitat Stuttgart
Institut fUr Verfahrenstechnik und Darapfkesselwesen
Prof. Dr. techn. R. Dolezal
Pfaffenwaldring 23
7000 Stuttgart 80, Germany
733-43
-------
-------
ABSTRACT
In a study under the contract of the Saarbergwerke AG it is planned to achieve N0X
emissions near 200 rag NO^/m3, i.e. 98 ppm NO without expensive DENOX technology.
By application of retrofit primary methods (air staging, flue gas recirculation)
the N0X emissions from the coal fired boiler Fenne 3 (slag tap furnace, 160 MW
electric power) could be reduced from 900 to 520 ppm NO at 5% 02- In the year 1988
the boiler was equipped with an arrangement for fuel staging. Reburning fuel is
coal gas with 50 % Hg and 25 % CH4.
Experiments from September 1988 to July 1990 showed that reburning can reduce N0X
emissions from 520 ppm to 180 ppm NO (5% Oj). The influence of different parameters
(primary zone stoichiometry, reducing zone stoichiometry etc.) was investigated.
The reduction zone stoichiometry and the reburn fuel mixing were pointed out to be
the most important parameters for low N0X emissions by reburning /l/.
In order to optimize reburning the following work has been done:
• distribution of flue gas concentrations was measured (primary zone,
reducing zone, burnout zone),
• reburning fuel mixing was optimized by three-dimensional fluid flow
computations,
• fuel staging with synthetic gases was examined in a 0.5 MW test facility
and
• the influence of ammonia addition into the reduction zone was investigated.
By optimizing the reburning gas injection and by addition of ammonia to the
reduction zone the N0X emissions could be reduced to a minimum of 130 ppm NO (5%
O2) up to now. Reburning has only a slight impact on the burnout of the coal. The
carbon content in the fly ash is less than five percent.
78-45
-------
INTRODUCTION
In the last years there had been large efforts to lower N0X emissions from
stationary combustion sources. For combustion systems with a thermal load of more
than 300 MW N0X emissions of 200 mg/m^ NO calculated as NO2 (98 ppm NO) at 5% O2
(molten ash furnace) or 6% Og (try ash furnace) are demanded in Germany,
Applied and commonly used techniques for N0X abatement can be devided in
t combustion modifications,
• selective non catalytic reduction (SNCR) by ammonia or urea and
• selective catalytic reduction (SCR) by ammonia.
Due to the short period for retrofitting existing power plants and equipping new
power plants with NQX abatement techniques, most German hard coal fired power
stations are or will be soon equipped with the SCR DENOX technology.
Measures to influence the N0X emissions of coal furnaces by combustion
modifications are:
t optimized boiler operation (low oxygen operation),
• flue gas recirculation,
• air staging (single burner or in the furnace) and
• fuel staging, reburning (single burner or in the furnace).
In the past years air staging has proved to be an effective method for N0X
reduction. For German lignite it seems possible to achieve the required N0X
emissions without expensive DENOX-technology by improved air staging in the furnace
/2/. A further technique of minimizing NQX emissions is a method called fuel
staging, reburning or In-Furnace N0X Reduction. Results of fuel staging in test
facilities are very promising.
A published application of reburning to coal combustion furnaces is the MACT
process. By fuel staging at a coal dust furnace N0X emissions of less than 150 ppm
could be achieved /3/.
Figure 1 shows the principle of fuel staging. In the first zone, which is the main
heat release zone, the fuel can be burnt under fuel lean conditions to ensure
complete burnout. The addition of reburning fuel creates a fuel rich, N0X reduction
zone. The reduction of nitrogen oxides is initiated by hydrocarbon radicals. In the
final zone the combustion is completed by addition of air.
DESCRIPTION OF THE PROJECT "BRENNSTOFFTRENNSTUFUNG (BTS)"
To lower the NOx emissions in coal dust furnaces the project "Combined minimizing
of N0X production and reduction of formed N0X - Brennstofftrennstufung (translated;
Fuel Splitting and Staging)" has been initialized.
Coal is divided by a devolatilization process in a reduction gas with volatile
nitrogen and the remaining coal (char). Both fractions are burned in a fuel staged
combustion with char as primary fuel and pyrolysis gas as reburning fuel.
7B-46
-------
The project consists of several steps:
• Investigation of reburning at a 0,5 MW gas fired combustion facility with
synthetic fuel,
• Large scale tests of reburning with coal gas as reburning fuel in a slag
tap furnace,
• Investigation of the process "fuel splitting and staging" in a small scale
test facility.
The investigations of reburning in the 0.5 MW combustion facility with synthetic
fuels and the trials at the 160 MWe] slag tap furnace are subjects of this report.
Results of performance and emissions of the process "Fuel Splitting and Staging"
in a small scale test facility are soon expected.
MECHANISMS GOVERNING N0X PRODUCTION AND REDUCTION AT FUEL STAGING
Figure 2 shows the N0X production and NOx reduction mechanism for the three zones
of a fuel staged combustion with coal dust as primary fuel and gas as reburning
fuel.
In the main heat release zone the formation of N0X is mainly due to the fuel
nitrogen. During devolatilization of coal a part of fuel nitrogen is released with
the pyrolysis gases, the other part remains in the coal char. The amount of
nitrogen released with the pyrolysis products depends on coal properties (volatile
matter content) and temperature. The volatile nitrogen and char nitrogen are
converted to N0X in a different way and in different amounts.
The volatile nitrogen quickly forms the intermediate species HCN, which is then
converted in a slow reaction to NH3. Depending on the fuel/air ratio and on
temperature, NH3 is either reduced to molecular nitrogen or it forms NO. The degree
of nitrogen oxide formation from the volatile fuel nitrogen can be affected by
primary combustion modifications, such as air staging or flue gas recirculation.
The production of NQX from Char-N is generally low with conversion rates between
10 and 20 percent. The heterogeneous production of nitrogen oxide is less sensitive
to process parameters as the formation from volatile sources. Therefore it is
assumed, that Char-N is responsible for minimum N0X emissions, which can not be
lowered.
In the reduction zone the nitrogen oxides formed in the main heat release zone are
reduced by homogeneous reactions. If the reburning fuel contains hydrocarbons, the
gas phase reduction of NO is initiated by CH^ in a fast reaction
NO + CH-j —> HCN + products, (1)
This fast step is followed by the relatively slow conversion of HCN to NH-j. This
reaction is significant for the overall reduction.
NH-j then either forms NO by reaction with 0 or OH radicals
NHj + 0 / OH —> NO + products (2)
or is reduced by NO to N2
7B-47
-------
NH-j + NO —> Ng + products (3).
Because of the fuel rich atmosphere in the reduction zone reaction (3) is
predominant. .
Investigations of Bose /4/ confirm, that the gas phase reactions are dominant in
fuel rich combustion zones' and that the heterogeneous reduction is of minor
importance for coal dust combustion.
The gas phase nitrogen reactions in the first and second stage are quite the same,
as to be seen in figure 2.
By addition of air the N-containing species NO, HCN and NH-j are converted to N0X
in the burnout zone. NO and HCN are almost completely transformed to N0X, NH-j only
in a very small amount /5/. If the burnout air is added to the flue gas at
temperatures of about 900 °C» a further NOx reduction is possible.
REBURNING WITH SYNTHETIC COAL GASES IN A TEST FACILITY
In order to study the reduction efficiency with a pyrolysis gas as reburning fuel
experimental investigations were carried out under the contract of the
Saarbergwerke in a gas fired combustion facility at the University of Karlsruhe,
The synthetic pyrolysis gas consists of 60% Hj> and 30% CH4. The watercooled
combustion chamber is described elsewhere /6/. The residence time in the reducing
atmosphere is about one second, the flue gas temperature at the location of gas
injection is about 1300 eC, at the location of air injection about 900 "C. The
stoichiometric ratio of the first fuel lean zone is X\ = 1.1 with a measured NOx
level after the first stage of 600 ppm. The overall stoichiometric ratio was kept
constant at A3 = 1.2.
The keypoint of the tests was to evaluate the influence of ammonia addition to the
reburning fuel, as pyrolysis gases contain nitrogen species such as NH3.
Furthermore the pilot scale results are compared to the results of reburning in the
slag tap furnace in order to demonstrate optimization potential for the large scale
application. The experiments at a pilot scale test facility allow the variation of
parameters which cannot be changed at a utility power plant.
Earlier investigations showed, that the addition of a nitrogen species such as NH3
to a reburn fuel makes no difference at the optimum stoichiometry X£, but outside
this optimum the N containing reburn fuel resulted in higher NQX emissions /I/.
Figure 3 shows the final NQX emissions and the corresponding measured nitrogen
species after the reduction zone for using a reburn fuel containing no NH3, 1.5%
and 3 % NH3. For pure pyrolysis gas (0% NH3), NOx is reduced from 600 ppm (5% O2)
after the primary zone to 115 ppm after the burnout zone at \% = 0.85. The addition
of 3 Vol* ammonia results in a shift of the optimum stoichiometry to >>2,opt = 0-89
and a further reduction of the total NGX emissions to 60 ppm N0X (5% 0?). The
corresponding N-species of the reduction zone show an increased reduction of N0X,
the concentration of NH3 rises drastically for X2 < *2,opt> while the HCN emission
7B-48
-------
is not affected by the increased NH3 input. At the optimum stoichiometry without
ammonia, N-species of 130 ppm NO and 20 ppm NH3 are converted to 115 ppm N0X in the
burnout zone. For the maximum NH3 addition (3%) 60 ppm NO and 100 ppm NH3 form 60
ppm final N0X emissions.
Further experiments at the University of Karlsruhe /8/ outside this project with
natural gas as reburn fuel showed similar trends as in the case of ammonia
addition.
In contrast to other investigations the addition of ammonia to the reburning gas
enhances the reduction efficiency of reburning significantly. The discrepancy of
the presented results to those of other authors are believed to be caused by the
high temperature of about 1300 °C in the reduction zone, optimized mixing injection
and a residence time of one second. These conditions favour the formation of NH3
rather than HCN in the reduction zone for all three cases studied. While the NO of
the reduction zone is completely converted to NDX in the burnout zone, the
conversion of NB3 to N0X is small. The high conversion of HCN to N0X can be
avoided. This is in agreement to Tagaki, who reports a low conversion rate of NH3
to N0X and a high rate of HCN to N0X /5/.
INVESTIGATION OF REBURNING IN A 160 MW SLAG TAP FURNACE
In order to show the effectiveness of N0X reduction with pyrolysi s gas as reburning
fuel and to find out the main parameters, the fuel staged combustion was applied
to a 160 MWe] power plant.
Furnace design and performance of the trials
Figure 4 shows the furnace of the steam generator and the zones of the fuel staged
combustion. The furnace consists of two molten ash chambers. The two burner rows,
consisting of four air staged burners, are arranged in two stages at each chamber.
To lower the N0X emissions of the molten ash chambers, the old unstaged burners had
been retrofitted by air staged burners. As a second method to reduce N0X by primary
measures, flue gas. recirculation to the pulverizer mills had been installed. The
achievable NOx emissions by primary N0X reduction had to be evaluated as the basic
emission level before starting reburning.
After the fuel lean combustion of coal dust in the molten ash chambers reduction
gas can be injected to the flue gas by twelve nozzles for each chamber. The
arrangement of reburning fuel injection is shown in figure 5, The flue gas at the
end of the first zone has a temperature of about 1400 - 1500 °C. The injected fuel
is coke oven gas, which mainly consists of Hg (50%) and CH4 (25%). The addition of
reburning fuel causes the formation of fuel radicals, which start the N0X reduction
process. The residence time of the flue gas under fuel rich conditions in the
reduction zone is about one second at maximum thermal load.
By addition of burnout air at the end of the separated flue gas channels behind the
chambers the combustion is completed.
7B-49
-------
The entire experimental program from September 1988 till September 1990 included
trials without reburning to determine the initial emissions, trials with coal gas
as reburning fuel and experiments with ammonia addition into the reduction zone and
to the burnout zone. During the experiments about 100 process variables were
measured for On-line monitoring and stored for later data analysis. Besides the
operational flue gas analysis in the furnace and at the stack, flue gas
concentrations and temperatures were measured in cross sections behind the
chambers, in the reduction zone and in the burnout zone for a better understanding
of N0X formation and destruction and to point out possibilities for optimization.
As the results of N0X emissions by reburning are a function of the stoichiometry
of the main heat release zone, the reburning zone and the burnout zone, the
stoichiometrics of the zones had to be calculated accurately. While the air flows
and the reburning gas flows were measured, a measurement of the pulverized coal
flow was not available.
The air stream, necessary for the stoichiometric combustion of coal, is
proportional to the ratio of thermal power and the efficiency of steam generation.
Vair,stoich. " A * Pth ^
The thermal power can be calculated by the superheater and reheater Jetstream
and the temperatures and pressures necessary for determining the corresponding
enthalpies. The efficiency of steam generation is dominated by the heat loss of
the flue gas. The variable A gives the necessary air for combustion of coal with
a thermal input of 1 MW. A is constant for a large range of coals and not varying
with changing water or ash contents of the coal.
The stoichiometrics computed by this method were verified by comparison with the
stoichiometries calculated from flue gas composition.
Results
Primary methods. The results of the primary N0X reduction (air staging at the
burner, flue gas recirculation) are summarized in figure 6. The NO emissions are
plotted as a function of the recirculated flue gas stream. Each point in figure 6
relates to a value, measured every ten seconds.
The application of air staging is for this slag tap furnace the more effective
method for reducing N0X emissions than the application of flue gas recirculation.
By air staging at the burner without flue gas recirculation the N0X emissions could
be lowered from 644 ppm to 500 ppm NO (5% 0?). When 10% of the whole flue gas was
recirculated to the mills, air staging caused a reduction from 560 to 490 ppm NO.
By application of different methods for N0X reduction at the same time the
effectiveness of the single measure decreases.
7B-50
-------
The initial emissions for the reburning trials were 500 - 550 ppm, which could be
obtained by air staging at the burner and by flue gas recirculation. The initial
emissions refer to an unstaged operation in the furnace, what means that the
stoichiometry of the chambers and the overall stoichiometry were kept constant at
1.2.
Reburning results. Figure 7 shows the. result of reburning with varying gas streams.
Each value corresponds to a trial of at least two hours. At a steam generation
power near full load the N0X emissions without reduction gas are 520 ppm for a
stoichiometry of 1.2. By air staging in the furnace and at a constant thermal load
the N0X emissions could be lowered to 460 ppm (Aj = 1.1, A3 = 1.2). The reduction
of the thermal power caused in this test no significant change of NO emissions.
Other tests showed a maximum influence of reduced thermal load of 20 ppm NO for the
staged case. The reduction of the thermal power corresponds to the heat input of
the maximum gas stream.
By increasing the gas stream at a constant first zone stoichiometry, the NO
emissions decrease sharply. By supplying twenty percent of the total heat input by
the reburning fuel, NQX emissions of 180 ppm (5% Oj) could be achieved. The unburnt
carbon in the fly ash was 4%.
The dominating parameter for reburning is the stoichiometry of the reduction zone.
Figure 8 shows NO emissions for trials in 1989 and 1990 without measures for an
improved reburning as described later. The trials were performed at different
primary zone stoichiometrics, burnout zone stoichiometrics and different thermal
loads. If sufficient air is provided for the coal combustion in the molten ash
chambers, reburning caused no increase of unburnt carbon in the fly ash. The
operation of the first zone with a stoichiometry greater than 1.09 for the
existing, non optimized coal dust distribution to the burners secured a
satisfactory burnout of the coal below the 5% threshold value.
Figure 9 compares the measured NO concentrations in the reduction zone without
reburning gas and with a reburning fuel of 20% of the total thermal input. Without
reburning gas an uniform distribution of NO concentrations of 550 ppm (at 0% O2}
was measured in the cross section before burnout injection. By addition of
reburning fuel of 20 % the cross section measurements showed NO concentrations
between 100 and 300 ppm NO. The concentrations of NO are corresponding to the
measured concentrations of CO, Hg and CmHn (Figure 10). Near the furnace wall on
the side of the gas injection (left side in figure 9 and 10) and in the middle of
the furnace the concentrations of the combustible species are maximum. The non-
uniform distribution is mainly caused by an incomplete mixing of the reburning gas
with the flue gas from the molten ash chambers. Further cross section measurements
of flue gas concentrations behind the chambers show that the coal dust distribution
to the burners also contributes to an unbalanced distribution in the reduction
zone. In the scope of the investigations the coal dust/air distribution was not
optimized, but it is assumed that a control of coal dust supply to the individual
burners can contribute to obtain lower N0X emissions.
7B-51
-------
Mixing calculations. Experimental investigations of Kolb /6/ with natural- gas as
reburning fuel pointed out the influence of mixing on the N0X emissions for a fuel
staged combustion. By an optimized mixing of reburning gas he could achieve a 50%
reduction compared to the case without optimization. The effect of mixing phenomena
on the results at the test facility of the University of Karlsruhe was minimized
by an optimized mixing. The reduction zone in the slag tap furnace "Fenne 3" at an
optimum mean stoichiometry consists of areas with stoichiometrics, which differ
from the optimum stoichiometry, so resulting in higher NOx emissions.
In order to improve the mixing of the reburning gas and to optimize NO reduction,
mixing of the reburning fuel was calculated by three-dimensional fluid flow
computations.
The grid used for the computations is shown in figure 11. Because of the symmetry
of the furnace the fluid flow was calculated for a half of one chamber. As the
combustion of coal dust is mainly completed in the chambers, the computation
disregards heat transfer processes by reaction and radiation.
The choice of the computation domain considers the asymmetric distribution of the
velocities (Figure 12) at the location of reduction gas injection. This is caused
by the return of flue gas from the chambers to the upstreaming gas in the first
furnace duct. In the cross section above gas addition an non-uniform distribution
of velocities can be seen with maximum velocities near the side wall and the wall
opposite to the gas nozzles. At the wall near the gas nozzles recirculation takes
place. In the following cross sections the velocities are more balanced, but still
showing basically the same tendencies.
The calculated stoichiometrics in figure 13a confirm the measured distribution at
a cross section at the end of the reduction zone. As it was evaluated in the test
facility with a reburning fuel containing ammonia, N0X reduction is optimum at X?
- 0.9 and satisfactory for a reduction 0.82 < X2 < 0.92.
The computations indicate that the area with a stoichiometry for a satisfactory
reduction covers only 15% of the cross section. In 35 % of the cross section the
flue gas atmosphere is fuel lean.
In order to improve gas injection the cooling air duct of the gas nozzles should
be connected to the existing flue gas recirculation. Before installation the
influence of an increased mixing momentum on the stoichiometry distribution was
computated, as shown in figure 13b.
With flue gas as additional mixing momentum the area with a satisfactory reduction
covers 60 % of the cross section at the end of the reduction zone. These results
of calculation were the reason to install a provisional connection of the flue gas
recirculation to the gas nozzles. A comparison of measured and calculated
stoichiometrics showed a good agreement /9/.
Trials of improved reburning. The impact of an increased mixing momentum on the
final NO emissions is shown in figure 14. The decrease in NO emissions in this test
was about 25 ppm. The effect of the more uniform distribution of reduction gas on
7B-52
-------
the NO concentrations measured at the end of the reduction zone is depicted in
figure 15. With flue gas as additional mixing momentum the average NO
concentrations are reduced by 40 ppm. An increased reduction of local HO
concentrations seems to be equalized by an increased conversion of the N-species
of the reduction to NO in the burnout zone.
The recirculation of flue gas provided the possibility of ammonia addition into the
reduction zone. In order to quench the flue gas, water or ammonia water can be
injected into the flue gas. For these tests a 15% NH3 concentration was used.
The results confirmed the positive effect of ammonia on N0X reduction. The
experiment shown in figure 16 was carried out at a reduced thermal load in order
to examine a wider range of reducing zone stoichiometrics. In the case without
ammonia addition (with flue gas) no N0X minimum could be determined, with ammonia
injection the N0X emissions were minimum at X2 = 0.89. Only for very fuel rich
conditions in the reduction zone A3 < 0.85 (reduction gas fraction > 25%) ammonia
addition leads to higher NO emissions. Figure 16 also demonstrates the effect of
burnout stoichiometry A3, A decrease of A3 from 1.2 to 1.1 causes a decrease in the
NO emissions for the case with and without ammonia addition. The unburnt carbon in
the fly ash was less than 4%.
Laser measurements /1Q/ of NH3 concentrations in the flue gas at the end of the
furnace detected in no case a measurable ammonia slip.
The addition of ammonia to the burnout air had only a positive effect for higher
NO emissions or stoichiometries A2 > 0.92 (Figure 17). The temperature of the flue
gas after burnout air injection is between 1000 and 1150 *C, measured at full
thermal load over the complete cross section of the furnace.
The reported results refer to a two chamber operation. In one chamber operation
lower emissions could be determined, as shown in figure 18. Each value in figure
18 corresponds to one test over several hours.
The difference between one chamber and two chamber operation is the possible use
of an air stream to the chamber out of operation as a further burnout air, so that
in one chamber operation the burnout air can be added in two stages. In one chamber
operation minimum emissions of 130 ppm at 5 % Og could be obtained at
stoichiometries of the burnout zone beetween A3 = 1.05 - 1.1 (without regarding the
air from the chamber out of operation).
In figure 19 the unburnt carbon in the fly ash is plotted as a function of the
reduction zone stoichiometry for the one chamber tests.
CONCLUSIONS
By application of reburning to a slag tap furnace a NO reduction from 520 ppm to
minimum emissions of 130 ppm were obtained. The investigations pointed out the
strong influence of reduction zone stoichiometry on the NO emissions. Mixing of
reburn fuel has to be optimized and burnout zone stoichiometry should be as low as
possible to achieve low N0X emissions.
7B-53
-------
For the slag tap furnace "Ferine 3" there exists a further N0X reduction potential
by
• optimizing the reburn fuel mixing into the reduction zone,
• optimizing of the coal dust distribution to the burners,
• arranging the burnout air injection in at least two stages and by
• addition of ammonia above the reburning gas injection.
Measures to increase the fineness of the coal dust would allow to minimize the
reburning fuel fraction.
ACKNOWLEDGEMENTS
This work was conducted under the contract of the Saarbergwerke AG with financial
support of the federal Ministry of Research and Technology (BMFT), Germany.
REFERENCES
1. H. Spliethoff. "NOx-Minderung durch Brennstoffstufung mit kohl estammigen
Reduktionsgasen." VDI-Bericht 765, 1989, pp. 217-230
2. K.R.G. Hein, D. Kallmeyer. "Stand der NOx-Minderung bei braunkohlebefeuerten
GroBkesselanlagen." VGB Kraftwerkstechnik, June 1989, pp 591-596
3. M. Araoka, A. Iwanaga, M. Sakai. "Application of Mitsubishi "Advanced MACT "
In-Furnace Removal Process." 1987 Joint Symposium on Stationary Combustion NOx-
Control, New Orleans 1987
4. A.C. Bose, 0.0.L. Wendt. "Pulverized Coal Combustion; Fuel Nitrogen Mechanics
in the rich Post-Flame." 22ndt Symp. (Int.) on Combustion, The Combustion
Institute, 1988, pp 1127-1134
5. T. Tagaki, T. Tatsumi, M. Ogasawara. "Nitric Oxide Formation from Fuel Nitrogen
in Staged Combustion: Roles of HCN and NHi." Combustion and Flame 35, 1979, pp
17-25
6. T. Kolb, W. Leuckel. "Reduction of NOx Emission in Turbulent Combustion by Fuel
Staging / Effects of Mixing and Stoichiometry in the Reduction Zone."
22nd Symp. (Int.) on Combustion, The Combustion Institute, 1988, pp 1193-1203
7. S.L. Chen, J.M. McCarthy, W.D. Clark, M.P. Heap, W.R. Seeker, D.W. Pershing,
"Bench and Pilot Scale Process Evaluation of Reburning for In-Furnace NOx-
Reduction"
21st Symp.(Int) on Combustion, The Combustion Institute, 1986, pp. 1159-1169
8. J. Ritz, T. Kolb, P. Jahnson, W. Leuckel. "Reduction of NOx Emission by Fuel
Staging Effect of Ammonia Addition to the Reburn Fuel," Joint Meeting of the
British and French Section of the Combustion Institute (1989), Rouen, France
9. H. Spliethoff, B. Epple, D. Renner, "Einmischung von Reduktionsbrennstoff oder
Reduktionsmitteln in technische Feuerungen" 6. TECFLAM Seminar, Stuttgart 1990
10. H. Hemberger, H. Neckel, J. Wolfrum. "LasermeStechnik und mathematische
Simulation von SekundarmaBnahmen zur NOx-Minderung in Kraftwerken." 3. TECFLAM
Seminar, Karlsruhe 1987
7B-54
-------
Main Fuel / Air
Primary Zone
X > 1
Reduction Fuel
Reduction Zone
X < 1
Burnout Air
Burnout Zone
X > 1
Figure 1. Principle of fuel staging {reburning)
MAIN HEAT RELEASE ZONE
REDUCTION ZONE
BURNOUT ZONE
COAL DUST AIR
REDUCTION GAS
BURNOUT AIR
i
i
1 r-r
J Char N i
1 fM"
1 —\ N
Fuel N J (NO) —
> -A /
(Volatile N joH.o,/
(HCN)-»(NH^)nh,,no
1
I
1
l
i
i
1
0 I /¦"—\ N
1 (no) —
jr 4*
°oH
°5 h
K
S gn __
I NNj
JL«
i
i
i
i
! I
i
i
0 ' r~\
I (no)
vohT
;n °2 1
Hi
^—""oh h—
NO
NHj
©
Figure 2. N0X production and reduction for a fuel staged combustion with coal as
primary fuel and gas as reburning fuel
7B-55
-------
NH, (VolX)
-J
CD
cn
O)
6
cx
D,
n
c
o
•
t/5
t/5
6
<1>
X
O
£
O
£
0.80 0,85 0,90 0,95
Stoichiometry Reduction Zone \2
g
Pl,
p.
(D
e
o
N
¦d
a>
0$
e
o
4-i
cd
±J
c
-------
Figure 4, Furnace of the 160 MWei steam generator Feme 3
9068
Figure 5. Reduction gas nozzles
7B-57
-------
700
^ 675
6
a 850
625
c3 680
O
fcS 575
•O,
W
558
525
c
.2
W5
« 500
O
Z 475
450
I
£
Unstaged Burner Operation
1".
Staged Burner Operation
fc
30000
50000
60000
Flue Gas Recirculation (to the mills ) T m3/h 1
Figure 6. Results of air staging (burner) and flue gas
recirculation (to the mills)
600
r-i
£
550
Q.
a.
500
450
400
350
300
250
200
ry
o
150
•s*
100
in
o
50
z
e
LOAD
* 92%
Unslaked
92/
73%
n#8
P Staged Operaiior.
. ir. Furnace
Rebum Fuel Fraction
¦' 81%
&
10%'
¦ 86%
14%
' 91%
19%
1 1.05 1.1 1.15 1.2
Stoichiometry Primary Zone
Figure 7. Reduction by reburning -
influence of reburn fuel fraction
600
E 550
CL
a
" 500
450
400
350
300
~ 250
oj
o
x* 200
IT)
„ 150
O
z
100
Two Chamber Operation
©
. Non optimized Reburning
O ©
_©©© o
£ o
%
-------
without
200
ppm
no im 02>reburning
4300/- -
a
200 ..
ppm With
0NO reburning
3300 / -
2300 .
IB00.
1300,
4g0B
[mm]
Figure 9. Cross section measurement of NO in the reduction zone with and
without reburning gas (half cross section behind one chamber, gas injection
is located 12 meters below the depicted cross section on the left side)
ia.0/_ _ _. M im 4* J
/ / / /
/ m - Mm jL /
' / / / /
/ / / /
/ ¦/ t /
/ / /
80! 2 100 34 00 4?BB ' [mm]
Figure 10. Cross section measurement of unburnt gas in the reduction
zone (half cross section behind one chamber, gas injection is
located 12 meters below the depicted cross section on the left side)
7B-59
-------
Figure 11. Grid for fluid
flow computations
Figure 12. Calculated distribution
of velocities
a) Without Flue Gas Momentum b) With Flue Gas Momentum
Figure 13. Calculated distribution of stoichiometrics without and
with flue gas as mixing momentum
7B-60
-------
350
£ 325
a
Q.
1-J 300
275
250
225
200
175
o
150
in
o
100
Boiler Load 92%
Reburn Fuel
Fraction 19%
* Without Flue Gas Momentum
V With Flue Gas Momentum
125 •
B
.85 .9 .95
Stoichiometry Reduction Zone
Figure 14. Effect of flue gas momentum
on final NQX emissions
see
ppm Without Flue
NO Gas Momentum
0
[mm ]
/
/
A 300 ,
midd le of the furnace '
/
see
ppm With Flue 3000/-
140 • Gas Momentum
Figure 15. Effect of flue gas momentum on local
N0X emissions in the reduction zone
7B-61
-------
Boiler Load 78 - 86 %
V With Flue Gas Momertnmi
Cx With NH3 Addition to Flue
Gas
7 A Xj = 1,2
~ A X3 = 1,1
j
J
' ¦ . /
: /
(
/ /
¦ ¦ *4 / f
s* y/
25% 20% 15%
Rebum Fuel
Fraction
¦ 1 1 I 1 I i 1 1 1 » 1 ¦ l i > ¦ I . t » «
350 r
n I
E 3251
a
a
" 300[
2?5
2S01
. 225
200[
~ 175
OJ
o
.\* 1501
in
„ 1251
Boiler Lead 78 -86®
X Withou: Flue Gas Momentum
© Wish NH3 Addition to Burnout Air
.6 .85 .9 ,95
Sto i ch i ometry Redaction Zone
iee
25% 20%- 15* Rebura Fuel
Fraction
-.6 .05 .9 .95 I
Stoichiometry Reduction Zone
Figure 16. Effect of ammonia
addition to the flue gas
Figure 17. Effect of ammonia
addition to the burnout air
One Chamber Operation
NHj Addition to (he Flue Gas
X3 - 1,05 - 1,2
*
4fi.
£4
A
a m
.8 .9 1 1.1
Sto i'ch i ometry Reduction Zone
Figure 18. NOx emissions for one
chamber operation with ammonia
addition to the flue gas
9
r-i
x e
X
!A
cr
x
u
c .
c
0
JQ
L.
to
u
T5
&
c
k
3
S3
C
ID
One Chamber Operation
NH3 Addition lo the Flue Gas
X3 - 3,05 - 1,2
£ * A
A
A.
% A
A
7
B
5
4
3
2
I
®? re rs i 1.1
Sto-i ch 1 ometry Reduction Zone
Figure 19. Unburnt carbon in the
fly ash for one chamber operation
(Corresponding to Figure 18}
7B-62
-------
COMPUTER MODELING OF N20 PRODUCTION BY 'COMBUSTION SYSTEMS
Richard K. Lyon, Jerald A, Cole, John C. Kramlich,
and Wm, Lanier
Energy and Environmental Research Corporation
18 Mason
Irvine, CA 92718-2798
.7B-63
-------
-------
ABSTRACT
The observed rate of increase of Ne0 (0.10V! to 0.26% annually) is a
matter of environmental' concern. While it is generally agreed that this
increase is a result of human activity, there is no consensus as to the
relative importance of different sources. Several studies have suggested that
pulverized coal fired combustion systems might be responsible, but the high
levels of Na0 found in these studies were later found to be an artifact, the
results of chemical reactions which occur during sample aging. Measurements in
which precautions are taken against this problem show very low NH0 levels for
flue gas from pulverized coal firing but do show substantial Ne0 concentrations
for fluid bed combustion.
In this paper computer modeling calculations are done for two mechanisms
of NeQ production, the selective reduction of NO by HON and sample aging. The
farmer plausibly accounts for the production of Na0 in fluid bed combustion and
may also be responsible for the small but apparently real amounts of NE0 found
in flue gas from pulverized coal firing. Calculations for sample aging,
however, show that preventing this mechanism from producing small amounts of
NeO may be substantially more difficult than was initially believed. Thus
sample aging may also account for the small amounts of Ne0 presently found in
flue gas from pulverized coal firing.
There have been speculations in the literature that the flue gas from
pulverized coal firing may be an important indirect source of Ne0, i.e. it was
speculated that chemical reactions which occur during sample aging may also
occur in the flue gas after it is released to the atmosphere. Our calculations
indicated that this-does occur but only to a very minor extent.
7B-65
-------
INTRODUCTION-
The observed rate of-increase of Ne0 <0..18X to 0.26*/. annually) is a
matter of concern both because Ng.0 is a greenhouse gas and because it has a
major and unfavorable influence on .the ozone layer (1,2,35. While it is
generally agreed that.this increase is a result of human activity, there is no
consensus as to- the relative, importance of different sources. While McElroy's
calculations <3,4) suggest that der.it"ification of chemical fertilizers could
account for the observed increase, others have criticized his calculations as
an order of magnitude too high (5,6). Weiss and Craig
-------
COMPUTER. MODELING METHODS
Calculations were', done with the reaction rate model shown in Table I
using an Acuchem program (17). Additional calculations were also done with the'
model shown in Table S using the PC version of ChemKin developed by Albert
Chang of Stanford University <1B),
RESULTS AND DISCUSSION
Mechanism of Direct NSO Production during Pulverized Coal Firing
As discussed above in recent measurements of NE0 in flue gases of
pulverized coal fired systems precautions were taken against Ne0 formation
during sample aging. Since these measurements show greatly reduced but still
apparently real amounts of Ne0 one might conclude that some small production'of
Njs.0 does in fact occur during pulverized coal firing. Since it is well proven
that fluid bed combustion produces large amounts of NeQ one might plausible
concluded that whatever mechanism is involved there, is also operative to a
small degree during pulverized coal firing. Alternatively one might conclude
that the precautions taken against the production of Ne0 during, sample aging
were largely but not completely effective.
The production of NeQ' by sample aging shown in Figure 1 is
oversimplified in one important respect; in Figure 1 it was assumed that all
the NOx in the sample is initially present as NO, Figure E shows calculations '
for the removal of NQx from the vapor phase by reaction with sulfite ion
solution for two cases, a gas mixture containing 6Q0ppm NO and one containing
5^0 ppm NO plus 6Gppm N0e. While the' former shows a slow steady decay of the
NOx} in the latter case there is an initial drop which consumes much of the
NOe. Figure 3 shows the corresponding calculations for the production of HNOg.
in the liquid' phase. As one might expect, when NQj. is not initially present,
HNOr- is formed slowly and only after an induction period, while when NOe is
initially present, there is a burst of HNQ<= formation at the start of the
reaction. As shown, in Figure 4 when N0e is initially absent, fJE0 is produced
only after a significant induction, but when it is present, the formation of
NrtO begins immediately. Indeed when N0i? is initially present the sample need
only age for 12 seconds to produce Sppm Na0.
Thus for samples which initially contain NQa it is considerably more
difficult to avoid the production of Ne0 by sample aging. Consequently, if one
tests one's experimental procedures using synthetic gas mixtures which contain
NO but no N0a, these procedures may appear adequate to prevent NrO production
during the sampling process, but still fail for real flue gases'which do
contain NBQ.
In this regard, it is interesting to note, that in reference 15,
measured NPD/N0x ratios of 0.01 or less were found for 10 different coal fired
installations, but for a gas turbine a value of 0.S1 was found. If the Na0
found in these measurements is a result of inadequate precautions against
sample aging, one would expect the highest NP0/N0x ratio to be found for the
installation in which the NOx contained the largest fraction NC*. It is well
known that the NOx emitted by gas turbines can contain a much larger fraction
of N0e than found in other combustion systems.
7B-67
-------
Indirect N5Q Production during Pulverized Coal Firing1
As mentioned above there is a question of whether or not the NO* and S0a
in flue gas may not represent an indirect source of Na0. When flue gas exits
the top of a stack, it both mixes with the atmosphere and cools to a
temperature that allows some of the water vapor it contains to condense. Thus
two competing processes occur, i.e. formation of an aqueous phase allows the
processes which produced Ne0 in aging laboratory samples to occur in the flue
gas, but mixing with the ambient atmosphere will rapidly quench those
processes. Thus one can imagine two ways in which flue gas can act as an
indirect source of Ne0; some Ne0 production can occur immediately after release
to the' atmosphere and a much slower NeQ production might occur after, the mixing
with the atmosphere via NOx and S0e reacting in clouds.
The former is a complex process and would be difficult to model
accurately but from the calculations shown in Figure 4 it seems likely that it
is a real but minor source of Ne0, In order to do calculations for the
production of N4.0 by reaction of NOx and S0a once they have been diluted to
ambient atmospheric .concentrations a set of typical conditions was assumed.
Thus ambient concentrations of &ppb and SO ppb were assumed for N0S and S0fe
respectively, L,* the ratio of liquid phase to gaseous phase, was taken at .0
x 10 ^, a typical value for a cloud. It was also assumed that .the reaction
of N0e with SOeto form NP0 was in competition with other reactions and that
the most important of these was the reaction of NGe with OH to form HN03. The
ambient concentration of OH free radicals in the cloud was assumed to be 1.7 x
10"4" molecules/cc and a rate constant of 1.1 x 10,-ti was used for the reaction
NOa + DH = HNQa. (18)
Figure 5.shows the results of these calculations for a case in which the
aqueous phase was assumed to have an initial pH of 7, The NDe + 01-1 = HNQa
reaction is found to be faster than Ne0 formation by'a factor of more than
101*. Assuming an initial pH of less than 7 made Ne0 formation even less
important. Thus production of N>.Q from NQx and SQe after they have mixed in
the ambient atmosphere is trivial and combustion systems are indirect sources
of N;jO only to the minor extent that Ne0 forms prior to the mixing of the flue
gas with the atmosphere.
N50 Production during Fluid Bed Combustion
While the very small concentrations of Ne0 currently being found in the
flue gas of pulverized coal fired systems may or may not be real, the fact that
fluid bed combustion can produce large concentrations of Ns0 seems to be well
proven. Reference 19 reports an interesting set of experiments which may
provide an explanation for this high Ne0 production. In reference 19 it is
reported that substantial NO reductions can occur in the free board of a fluid
bed. combustion system and that these reductions can occur at temperatures as
low as 1050C.K and reaction times as short as 0.2 sec. Since these NO reductions
occurred in the presence of VA Oe, some form of selective noncatalytic
reduction is clearly involved, but the observed NO reduction does not appear to
be due to reaction with NHa. Thus the mechanism by which the NO was reduced is
unclear. •
7B-68
-------
Figure t>9 quoted from reference 20 shows the result of flame modeling
calculations done with a reaction mechanism very similar to that shown in Table
S. The model's prediction is that there exists a narrow range of temperatures
in which HCN selectively reduces NO, the product of this reduction being Ns0.
Reference 20 also reports experimental results which confirm this prediction.
Based on these results reference £0 suggested that Ne0 in the flue gases
from pulverized coal firing was produced by the following mechanism. Nitrogen
containing char is produced in the primary combustion. Some of this char
escapes the primary combustion zone and reacts to form HCN down stream at lower
temperature where the reduction of NO by HCN to form Ne0 is favorable. This
reaction only produces NeO in a narrow range of temperatures"because at
temperatures above this range N«0 decomposes and at temperatures below the
range the HCN/NO reaction does riot occur.
Looking at Figure 6 one might suppose that this mechanism for Ne0
production is not applicable to fluid bed combustion systems because they
operate below the temperature window. Figure 7, however, shows that the
temperature window for NeO5 production is a sensitive function of the reaction
time. Selective reduction of NO to Ne0 by HCN can occur in the free, board of a
fluid bed combustion system and thus may be the explanation of the NO removal
reported by,reference 19.
Practical Implications
Fluid bed combustion is generally regarded as a developing technology ¦
and hence the fact that- fluid bed combustors may emit Na0 might seem to be a
potential rather than an actual problem. There is, however, one application in
which fluid bed combustion is a major industrial process, fluid bed catalytic
cracking. Within the cat cracking process the catalyst used to "crack" higher
molecular weight hydrocarbons to smaller molecules becomes coated with coke
and' catalytic activity is restored by fluid bed combustion of the spent
catalyst. The temperature of this combustion is low to protect the catalyst
and consequently any Nc0 produced would survive. Further, the amount of
nitrogen in the coke which is available for Ne0 is substantial, since
chemically bound nitrogen in the hydrocarbon feed goes preferentially into the
coke. Thus, since a major fraction of the world's total oil production goes
through the fluid bed cat cracking process, it is quite possible that this
process contributes significantly to anthroprogenic NeQ emissions.
7B-69
-------
CONCLUSIONS
Recent measurements of the NK0 levels in flue gas from pulverized coal
firing typically show ccncentrat ions of a few pom. These NSQ levels may be
real and the result of the reduction of NO by traces of HCN, or they may be an
artifact, a result of the fact that it is more difficult to prevent Ne0
production by sample aging than was initially, believed.
While there has been speculation that the emissions of S0a and NOx by
pulveriHed coal firing may indirectly be a substantial source of NBQ, our
modeling calculations indicate that indirect NeO production is a minor process•
Fluid bed combustion, however, can produce substantial emissions of Ne0
and our modeling calculations suggest that these emissions can plausibly be
explained in terms of the reduction of NO by HCN. It is regrettable that no
data are presently available for the production of Ne0 by fluid bed catalytic
cracker regenerators, since these installations may be a substantial source 'of
NeO.
7B-70
-------
REFERENCES
1 Weiss., R.F., J. Geophy. Res,, 86,7105-7195 (1981).
2 Khalil, M.A, and Rasmussen, R.A., Tellus, 353, 161-169 (1983).
3 Marl and, G., and Rotty, R.M. J.A.P.C.A., 35, 1033-1038 <1985!.
4 McElroy, M.B., as reported by J. E. Bishop, The Wall Street Journal, p.9,
Nov. 13, 1975.
5 Crutzen, P.J., Geophys. Res. Lett., 3, 169-172 (1976).
6 Liu, S.C., Cicerone, R. J., Donahue, T.M., and Chameides, W.L., Geophys.
Res. Lett., 3, 157-160 (1976).
7 Weiss, R.F. and Craig, H., Geophys. Res. Lett., 3, 751-753, (1976).-
8 Pierotti, D.« and Rasmussen, R.A, , Geophys. Res. Lett., 3, 265-267 <1976).
9 Hao, W.M., Wofsy, S.C., McElroy, N.B., Beer, J.M., Toqan, M.A., J. Geophy.
Res., 92, 3098-3194 (1987).
10 Castaldini, C., Water land, L.R., and Lips, H.I., EPA-600-7-B6-003a, 1986.
11 Ryan, J, V., and R. K. Srivastava, EPA/IFP workshop on the emission of
nitrous oxide from fossil fuel combustion (Ruei1-Malmaison, France, June 1-2,
1988>, Rep, EPA-600/9-89-089, Environ, Prot, Agency, Research Triangle Park,
N.C., 1989. (Available as NTIS PB90-126038 from Natl. Technol. Inf. Serv., •
Springfield, Va.)
12 Muzio, L. J., and Kramlich, J. C., Geophysical Research Letters, 15, 1369-
1372, (1988)
13 Lyon, R. K., and Cole, J. A,, Combustion and Flame, 77, 139 (1989)
14 Muzio, L. J., Montgomery, T. A., Samuel sen, G. S., Kramlich, J. C,, Lyon,
R. K., and Kokkinos, A., 23rd Symposium (International) on Combustion, in
press,
15 Kokkinos, A, ECS UPDATE, Spring-Summer 1989, No 15 pp 8-10
16 Linak, W. P., et. al., Journal of Geophysical-Research, 95, 7533-7541
(1990)
17 Braun, W., Herron, J. T.. and Kahaner, D. K., Int. J. Chew. Kin. 20 51-62
(1938)
18 Baulch, D. L. , Drysdale, D. D., Home, D. G. and Llyod, A. C., Evaluated
Rate Constants, Butterworth, 1976
19 Walsh, P. M., Chaung, T. 2., Dutta, A., Beer, J. M., and Sarofin, A. F..
19th Symposium (International) on Combustion, 1281-12B9 (1982)
20 Kramlich,J. C., Cole, J. A., McCarthy, J. PI., Lanier, W. S., and McSorley,
J. A., Combustion and Flame, 77, 375-384, (1989)
7B-71
-------
-------
1000
-vl
CD
i
-sj
ro
800 -
600
>
E
CL
CL
400
200 -
150 200 250 300 350
TIME, MINUTES
Figure 1. Experimental and Kinetic Calculations of N^O Formation in Sampling Containers
-------
-------
ppm
L sec
[N02jo -- 0 [N02]o = cOppm
1000ppm 302 600ppnn MOx, 0 or 60pom M02.
4 0C, 6-52mo!e'* liquid waier
Figure 2. Modeling of N0X Removal from the Gas Phase by
Reaction with Sulfite ion
-------
ppm
50. 60
t, sec
+ [N02]o = 0
lOOOppm S02, 600ppm NOx, 0 or 60ppm N02,
4 0C, 6.52mole% liquid water
[N02]0 = 60ppm
Concentration of HN02 expressed
as ppm based on gas phase
Figure 3.
Modeling of HN02 formation with N02 initially
present and absent
-------
pprn
t, sec
""S- [N02]o = 0 [N02]o = 60ppm
lOOOppm S02, 600ppm NOx, 0 or 60ppm N02
40C, 6.52mole% liquid water
Figure 4. Modeling of N20 formation with N02 initially
present and absent
-------
TIME, seconds X1000
+-[N02]/[N02]i ' -0- [N20]/[N02]i X 10000 [HN03]/[N02]i
10ppb S02, 6ppb N02, Initial pH = 7 pH at 10E+5 sec - 3.33
4 0C, L = 4.8E-7 ccL/ccG
Figure 5.
Competition between N20 formation and HN03
formation after the flue gas mixes with the
atmosphere
-------
INJECTION TEMPERATURE, K
Figure 6. Quoted from Reference 20.
7B-77
-------
PPTI
T, K (Thousands)
t = 0.02 sec ~H— t = 0.04 sec ~t = 0.10 sec t = 0.20 sec
200ppm HCN. 600ppm NO, 10% 02. 5% H20,
balance inert
Figure 7. Calculation of the Effect of Reaction Time on N20
Formation
-------
TABLE.I
Q 12
Chemical mechanism, rate constants and equilibrium constant3 at 25 C.
(rate constants are in units of L/mol/s or L /mol /a)
Gas Phase Reaction Rate Constant
1.
NO + NO + Oj - NOj + HO2
6
.13
E+3
Liquid Phase Reactions
2 .
N02 + HSO3- = no2- + HSO3
3
00
E + 5
3.
HSOj + HSO-, (+ HjO) - H2S03 + .MjSO,
5
00
£+5
4 .
2N02 + HjO «» HN02 + HN03
1
O
O
E+7
5',
HNOj + HSOj- = N0SO3- + H-C
2
40
E + 0
6.
NOSO3- + H* (+' H20) •» HNO + HjSO^
5
00
E+l
1 .
HNO + HNO - N20 + HjO
3
00
E+4
8.
NOS03- + HSOj- + HNO (SO.,) 2"
. 8
50
E+l
9 ,
UNO {SO, i + H+ (+ H20) - HONIISO-,- +' II* + HS04-
1
90
E-2
10
HNO (S03) 2" + HjO = HONKSOj- + H* + HSO^-
1
50
E-6
Equilibrium Processes
ll.NOjfgasS = NOj(aq)
12.S0j(gas) = SO-i3q;
13.S02(aq> = H* t HSO,
14-, HNOj - h' ¦> NO--
15 . HS04- = H* SO,
Henry's Law Constants
H = 0.01 M/atm
H = 1,30 M/atm
Equilibrium Constants
K = 1.54 t-2 M
K « 5.10 £-4 M
K - 1.20 E-2 M
7B-79
-------
TABLE 2
ELEMENTARY REACTIONS USED IN MODELLING
REACTION
A
ri
Ea
{cc/mol,sec,)
cal/mol
NH3+M=NH2+H+M
4.80E16
0
. 0
93900.0
2
NH3'+H=NH2+H2
2.5CE13
0
, 0
17100.0
3
NH3+0=NH2+0H
1.5E12
0
. 0
604 0s. 0
4
NH3+0H=NH2+H20
- 3.3E12
0
, 0
,2120'. 0
5-
NH2+H=NH+H2
5,OE11
0
. 5
2000.0
6
NH2+0=NH+0H
1.7E13
0
. 0
1000.0
7
NH2+0H=NH+H20
5.5E10
0
, 68
¦ 1290.0
S
NH2+02=HN0+0H '
5,1E13
0
. 0
30000.0
9
NH2+N0=HNH+0H
6.1E19
-2
. 46
1866.0
10
NH2+N0=N2+H20
9.1E19
-2
. 46
1866.0
11
NH2+HNONH3+NQ
• 1.8E14
0
. 0
1000.0
12
NH2+NNH=N2+NH3
1.0E13
0
. 0
0.0
13
NNH+M=N2+H+M
2.00E14
0
.0
.3,0000. 0
14
NNH+N0=N2+HN0
9.10E11
0
. 0
0.0
15
NNH+OH=N2+H2G
3.00E13
0
,0
0 . 0
16
HNO+M=H+NO+M
1.9E16
0
.0
48 680,0
17
:HN0+0H=N0+H20
3.6E13 '
0
.0
0.0
18 ¦
NH+02=HN0+0
3.00E13
0
.0
3400.0
19
0H+H2-H20+H
2.2E13
0
. 0
5150,0
20
H+02=OH+0
2.3E14
0
. 0
16800.0
21
0+H2 =OH+H
1. 8E.10
1
. 0
88 90.0
22
20H=C+H20
6.3E12 •
0
. 0
1090.0
23
H+02+M=H02+M
H20/21./
1.5E15
0
.0
-995.0
24
H+H0-2 = 20H
2.5E14
0
.0
1900.0
25
0+H02=02+0H
4". 8E13
0
, 0
1000 . 0
26
0H+H02=H20+02
5.0E13
0
, 0
1000.0
2 7
H02+N0=N02+0H
3.40E12
0
. 0
-260.0
28
N02+H=N0+0H
3,5E14
0
.0
1500,0
29
N02-0=N0+02
1.0E13
0
. 0
600. 0
30
N02+M=N0+0+M-
1.1E16
0
. 0
66000.0
31
0+0+M=02+M
¦ 1.4E18
- 1
.0
340, 0
32
N20+H=N2+0H
7.59E13
o
o
151
32
N20+M=N2+0+M
1.6E14
0.0
,51600
33
N20+0=N2+02
1.0E14
0,0
28200
34
N20+0=NO+N0
1.0E14
0.0
28200
35
C0+0H=C02+H
1,50E7
1. 3
-770
36
C0+H02=CO2+0H
5.75E13
0.0
22930
37
C0+02=C02+0
2.51E12
0 . 0
47690
3 8.
C0+0+M=C02+M
5.8 8E15
0.0
4100
ag-
NCO+O-NO+CO
5.6E13
0.0
0'
io
NC0+N0=N20+C0
1.0E13 -
0.0
-390
41
NCO+H=NH+CO
5.0E13
0.0
0'.
42
NC0+NH2 = NH+HNCO
1.0E11
0.5
5000
43
0+H2=HNC0+H
8.6E12
0..0
9000
44
NCO+OH=NO+CO+H
1.OE13
0.0
00 .
45
HNC0+0H=KC0+H20
1,0E11
0.5
6290
4 6
HNC0+H=NH2+C0
2.0E13
0.0'
3000
47
HCN+OH=HKCO+H
4.80E+11
0.0
11000'.
48
HCN+0H=CN+H20
1.5GE+13
0'. 0
10929.
49
HCN+0=NC0+H
1.40E+04
2. 64
4980
50
CN+02=NC0+0
5,60E+12
0.0
0.
7B-80
-------
LOW NO- LEVELS ACHIEVED BY IMPROVED COMBUSTION
MODIFICATION ON TWO 480 MM GAS-FIRED BOILERS
Mark D. McDanneT _
Sheila M. Haythornthwaite
CARNOT
15991 Red Hill Avenue, Suite 110
Tustin, California 92680
Michael D. Escarcega/.
Barry L. Gil man"
Southern California Edison Company
2244 Walnut Grove Avenue
Rosemead, California 91770
ABSTRACT
While most applications to meet new and emerging N0X regulations have focused on
retrofit technologies (low-NOx burners, urea, SCR), there are still opportunities for
additional NQX reduction via improved combustion optimization.
Southern California Edison, as part of their compliance efforts for a new N0X rule,
which ultimately requires N0X limits of approximately 20 ppmc, retained Carnot to
assist them in designing and conducting a combustion optimization program on two 480
MW gas-fired boilers. As a result of detailed combustion optimization test programs
on the two boilers, N0X was reduced by 24 to 56% over the load range at an average
cost-effectiveness of $,59/1b N0X. Through increased windbox FGR, improved BOOS
patterns and overfire air, N0X levels at full load were reduced from 91 to 62 ppmc.
These reductions will help SCE meet current and near-term N0X limits, and will
substantially reduce construction, and operating costs of any future SCR systems.
8-1
-------
INTRODUCTION
While most applications to meet new and emerging N0X regulations have focused on
retrofit technologies (low-N0x burners, urea, SCR), there are still opportunities for
additional N0X reduction via improved combustion optimization.
Southern California Edison, as part of their overall compliance plan for South Coast
Air Quality Management District (SCAQMD) Rule 1135, retained Carnot to assist them in
designing and conducting combustion optimization programs on two 480 MM gas-fired
boilers (Alamitos 5 and Redondo 8). This paper presents the results of the two test
programs, which provided immediately implementable N0X reductions of 24 to 56% over
the unit load range at an average cost-effectiveness of $.59/1b NOx.
Included in the paper is a description of the technical and regulatory background on
N0X emissions from the two boilers, a description of the two boilers, a description
of the approach taken in designing and executing the program, the results of the
program, and, a discussion of the results.
BACKGROUND
All of SCE's boilers in the South Coast Air Basin are subject to SCAQMD Rule 1135,
which includes system-wide 24-hour average N0K limits that start at 1.10 lb N0x/MW-hr
(approximately 100 ppmc*) in 1990 and steps down to 0.25 lb N0x/MW-hr (approximately
23 ppmc) in 1999. Additionally, Alamitos 5 and Redondo 8 are subject to rule 475,
which was passed in 1970 and limits NQX on gas fuel to 125 ppmc (approximately
1,38 Ib/MW-hr) for a 15-minute averaging period. Figure 1 presents a summary of N0X
limits on these two boilers.
When the 125 ppmc limit was Imposed, SCE implemented off-stoichiometric combustion
(overfire air ports and/or burners out of service) on 24 boilers in the South Coast
Air Basin, and additionally implemented flue gas recirculation (FGR) to the windbox
on four of these boilers, including Alamitos 5 and Redondo 8. Implementation of these
* ppmc = parts per million by volume, corrected to 3% 02, on a dry basis
8-2
-------
techniques reduced N0X levels from approximately 900 ppmc to 100 ppmc on both
Alamitos 5 arid Redondo 8..
In SCE's overall Rule 1135 compliance plan, there are a number of N0X reduction
efforts either planned or already evaluated on these two units. On Alamitos 5, urea
injection and installation of one row of low-N0x burners have been tested, and the
installation of a Selective Catalytic Reduction (SCR) system is planned. On
Redondo 8, an SCR system consisting of blocks of (honeycomb) catalyst placed in the
duct between the economizer and air preheater is scheduled for 1991. It is within
this context that combustion optimization was evaluated and implemented.
UNIT DESCRIPTION
Alamitos 5 and Redondo 8 are two of four identical 480 MW Babcock & Wilcox opposed-
fired units operated by SCE 1n the South Coast Air basin. The units are capable of
firing either natural gas or fuel oil. This program addresses gas firing only, since
Rule 1135 has limited application to oil firing and since oil is rarely burned.
Relevant details on the boilers are listed below:
• . . Manufacturer: Babcock & Wilcox'
• Rated Capacity: 480 MW (net)
• Steam temperature: 1,000°F superheat and reheat
• Steam pressure: 3500 psig (supercritical)
• Burner arrangement (see Figure 2):
-- Opposed fired
-- 32 burners, 16 per wall
-- 4 rows of 4 burners each on each wall
-- furnace split by division wall
• N0X control:
-- third elevation of burners out of service
-- FGR to windbox
-- OFA ports
• Newly installed Rosemount digital control system
• 02 trim system in service
• CO trim system installed but not yet in service
PROGRAM DESCRIPTION
The objective of the program was to determine what level of N0X reductions could be
achieved by modifying and optimizing combustion and boiler operating conditions prior
8-3
-------
to the Installation of SCR or other back-end N0A reduction technologies. Specific
benefits expected were:
1. Help meet Rule 1135 limits immediately.
2. By reducing inlet NO^ levels, reduce the size and cost of
future SCR installations.
A comprehensive program involving five discrete phases was designed. The five phases
are listed below, followed by a brief description of each phase:
• Records search
• Interview operating staff "
• Physical inspection and repair
• Optimization test program
• Load following tests
Records Search
The first step of the program was to review available test and operating data on the
units to help plan the test program.
Interview Operating Staff
Interviews were held with station engineers, maintenance and instrumentation
supervisors, shift supervisors, and boiler operators to familiarize test personnel
with unit operation and to familiarize station personnel with the objectives of the
program. Unit operation was observed with at least two different shifts of operators.
Physical Inspection and Repair
Prior to performance of the combustion optimization test programs, thorough boiler
inspections were conducted during maintenance outages. The objectives of the
inspections and outages were,to:
1. Evaluate the condition of all fireside operating equipment
including fans, dampers, and burners.
2. Identify any equipment requiring repairs or adjustments, and
verify that repairs were made.
3. Allow the test crew to become familiar with boiler design
and equipment.
4. Wash boiler to provide a known cleanliness lever.
8-4
-------
Performance of the inspections and repairs ensured that equipment problems would not
adversely impact unit operation during the test program.
Optimization Test Program
The optimization test programs consisted of 105 tests on Alamitos 5 and 51 tests on
Redondo 8. The test matrices were designed to evaluate the impact on NOx emissions
from the following variables:
• Unit load
• Excess 02
• Flue gas recirculation (FGR) to the windbox
• Overfire air ports
• Alternate BOOS patterns
• Air register throttling to selected burners
• Superheat/reheat proportioning dampers {Alamitos 5 only)
• Fan balancing
Each test included collection of gaseous emission data at the economizer exit, a full
set of unit operating data from the control room, and external unit data as needed
(damper positions, air register settings, windbox 02, etc.)- For most tests,
North/South composite data was collected. This involved collecting average gaseous
data from the North side, average gaseous data from the South side, and a composite
sample. For selected tests, full 32-point gaseous traverses were performed. When
test conditions were established and unit data were collected, the impact of test
variables on unit heat rate was watched carefully. The need to isolate one test
variable at a time to determine its impact on combustion did result in some test
conditions where operation was not optimum; this was considered in evaluation of the
results.
toad Following Tests
The test programs on both units were concluded with two sets of load following tests.
These tests involved establishing recommended low-NOx operating conditions and
monitoring N0X, 02, and CO while ramping boiler load between 160 MW and 480 MW. The
purpose of these tests was to determine if the low-NOx operating modes could be
maintained, and expected N0X values seen, over the entire load range with no
operational problems.
8-5
-------
RESULTS
The results are presented separately for the two units, as follows. For the sake of
brevity, detailed impacts of individual test variables are presented only for
Redondo 8; similar results were obtained for A1amitos 5.
Redondo 8
The tests identified two modifications to baseline operation (as described under Unit
Operation) that resulted in significant NQX reductions over the full load range, and
two further modifications that resulted in small additional N0X reductions. The
modifications which reduced N0X significantly are:
• increasing flue gas recirculation to the windbox to
the maximum achievable; and
• minimizing excess 02 until CO formation is seen.
Modifications which produced smaller N0X reductions are:
• taking burner pair 6 out of service (while leaving the
air registers open); and
• opening of the OFA ports at 480 MM and during load
following tests.
The results of combining these techniques are summarized in Table 1, detailed in Table
2, and illustrated in Figure 3. Note that Figure 3 does not include the opening of
the OFA ports, which were only evaluated during the load following tests.
Increased Flue Gas Recirculation effects are shown in Figure 4. Test points on
Figure 4 are scattered somewhat due to the inclusion of all test variables. However,
the trend of N0X reduction with increased GR is clear. This is most notable at 480
MW. Higher GR was limited at this load because of a fan amp limit. If fan capacity
could be increased to enable 25% GR, the projected N0X would be approximately 40 ppm
@ 3% 02 (see dotted extension line on graph).
Minimizing excess 0-, was performed at 250, 360, and 480 MW. The 02 setpoint for
minimum 02 was determined by gradually lowering excessive air until 100 to 200 ppm of
GO was seen consistently at that condition.
Table 2 shows the percent reduction attributable to minimizing 02 at the various
loads. This reduction increases with lower load, and more reduction may be possible
at 160 MW, where significant CO formation had not begun.
8-6
-------
Taking burner pair 6 out of service reduces the N0X fairly' uniformly across the load
range, as shown in Table 2. Figure 5 shows graphically the impact on NQX of taking
6 00S. The reduction caused by this modification is small, but the improvement in
boiler operation is significant. Figure 6 shows the CO level both with and without
6 00S, At 480 MW extremely high CO was created with all burners in service*, this was
removed by taking 6 pair 00S.
Another impact of this modification was to improve the excess 02 balance between the
north and south sides of the boiler. A series of tests led to the conclusion that
Burner 6 south is starved for air. This results in lower 02 and higher CO levels on
the south side. Taking Burner 6 out of service improved both 02 and CO balance
between the two sides.
Opening the overfire air ports at 480 MW reduced N0X by 7 ppm, or 11%. This condition
was established while at full load. Opening the OFA ports was not evaluated at other
loads due to difficulty in determining the positions of the ports early in the test
program. Once the open position was established by observing NO^ reduction at 480 MW,
the ports were kept open for one set of load following tests. Figure 7 shows the
reduction achieved across the load range by opening the N0X ports. While this
reduction is small, the modification does not impact boiler operation, and could
easily be made a permanent operating condition.
Load following tests showed that optimum low-N0x conditions could be maintained over
the full unit load range, without any operating problems. The results of the load,
following tests are shown in Figure 7. N0X levels are shown with N0X ports both open
and closed. A slight reduction with N0X ports open is seen over the entire load
range.
Other variables that were investigated during the program were air register throttling
on inboard burners to provide increased air flow to starved outer burners, and
alternate BOOS patterns. These tests provided insight into unit operation, but
implementation caused undesirable effects such as increased N0X, difficult operation,
or a large 02 or CO imbalance between the north and south sides of the boiler.
Alamitos 5
The tests on Alamitos 5 identified three modifications to baseline operation (as
described under Unit Operation) that resulted in significant reductions in N0X
emissions over the full load range: increased flue gas recirculation to the windbox,
opening of the OFA ports, and taking burner pair 6 out of service (while leaving the
8-7
-------
air registers open). The results of combining these three techniques are summarized
in Table 3, and illustrated in Figure 8.
The results show that substantial NOx reductions were achieved across the load range,
with the percentage reductions decreasing as unit load increases (from 56% at minimum
load to 27% at maximum load).
Table 4 shows the incremental reductions achieved by each of the three techniques.
The reductions achieved by each technique were cumulative across the full load range.
The largest reductions (11 to 36%) were achieved by increasing FGR to the windbox.
Reductions of 10 to 18% were achieved by taking Burner Pair 6 00S, and reductions of
1 to 9% were achieved by opening the N0X ports.
Load following tests showed that these conditions could be maintained over the full
unit load range, without any operating problems. The results of the load following
tests are shown in Figure 9.
Other variables that were investigated during the program were excess 0Z level,
superheat/reheat proportioning damper position, air register throttling on lower
burners to provide increased combustion staging, air register throttling on selected
burners in an effort to overcome an air/fuel imbalance, FD and GR fan biasing and
balancing, and alternate BOOS patterns. Those tests provided insight into unit
operation, but did not provide substantial N0X reductions.
Reductions in excess 02 did provide some N0X reductions, but the existing boiler 02
curve is so low (1% 02 over most of the load range) the 02 levels could only be reduced
approximately 0.2% before the onset of CO. Placing the CO trim control system in
service will allow maintenance of minimum 02 levels over the load range, and should
result in additional N0X reductions of 2 to 5% (based on minimum 02 tests conducted
during this program).
The tests also identified a significant north/south 02 imbalance in the furnace. A
series of tests led to the conclusion that the imbalance is mostly due to burner 6
North (an upper, rear, corner burner) being starved for air. The problem was
partially alleviated by taking the burner pair out of service for- N0X control.
DISCUSSION
This section presents discussions on the potential impact of the three recommended
combustion modification techniques (increased windbox FGR, Burner 6 out of service,
minimum excess 02) on unit operation, including heat rate. This discussion applied
to both units.
8-8
-------
Heat Rate
Any change in operation should be evaluated in terms of its impact on unit heat rate.
Operating costs for a N0X technique can become significant if they have a significant
impact on boiler efficiency. Emissions data, unit operating data, heat rate factors
and fuel cost factors were combined to determine an operating cost in terms of 5/1b
NQX reduced for increased FGR, taking Burner 6 out of service, and opening the NGX
ports. The cost benefit of reduced excess oxygen levels was also considered.
Tables 5 and 6 summarize the heat rate penalties, and present the operating cost of
the techniques combined in dollars per pound of N0X reduced. On Alamitos 5, heat rate
penalties of $0.34 to $0,83/lb N0X were seen. On Redondo 8, the only load at which
a cost is seen is 250 MW. Here NQX costs SO,31/lb reduction. At all other loads, the
heat rate is improved by reducing excess 02.
The results showed two areas in which unit heat rate penalties were incurred, and one
in which heat rate was improved: increasing FGR to the windbox increased auxiliary
power consumption, and taking Burner 6 out of service increased average excess 02
levels as measured by the test van. Minimizing 02 reduced the NQX level and improved
heat rate by lowering the excess air used.
It should be noted that these cost-effectiveness values are so low in part because
these techniques involve an incremental extension of N0X reduction techniques already
implemented on the boilers. Costs for boilers which do not already have windbox FGR
or some form of off-stoichiometric firing would be higher.
Other Impacts'on Unit Operation
None of the four low N0X techniques used in this study had any deleterious effects on
unit operation that were detected during the test programs. When the techniques were
implemented unit load was stable, flame appearance and stability were acceptable, and
there were no significant changes in tube metal temperatures.
There are some areas in which the techniques might impact unit operation in the long
run. The most important may be a loss in load capacity safety margin while operating
with Burner 6 out of service. In the baseline condition there are 24 firing burners,
and with Burner 6 out of service there are 22 firing burners. If a burner pair trips
at full Toad, there would be either two or four fewer firing burners in service
(depending upon whether it is an upper or lower burner pair that trips). With Burner
6 out of service, it would be more likely that available unit load would be curtailed
if a burner pair tripped. Prior to implementing Burner 6 00S, the magnitude of the
possible curtailments would need to be determined and an evaluation made of the
relative value of reduced N0X emissions vs. the risk of increased load curtailments.
8-9
-------
Another area of concern with taking Burner 6 out of service is that the increased heat
release rate per firing burner (an increase of 9% would occur) might cause overheating
in the burner throat area. This would have to be evaluated prior to implementation.
Implementation of increased flue gas recirculation to the windbox should be
coordinated with appropriate safeguards, since the booster fans have a high enough
capacity that they can blow put the flames at lower loads. New digital controllers
have been installed on the booster fans, the hopper control dampers, the FGR fans, and
superheat/reheat proportioning dampers. With the new booster fan controllers, curves
of damper position vs. unit load can be programmed in. However, to protect against
injecting too much FGR there should be a windbox 02 monitoring system. Such a system
could be either used for operator information or tied into the control system to
provide an alarm and/or feedback signal.
Operating with the OFA ports open should not provide any operation problems. As noted
before, it is currently difficult to access the OFA ports to open or close them. The
ports should be welded open. The chains currently installed do not allow easy
operation.
An important aspect to consider in applying combustion optimization techniques is the
boiler control system. These two boilers have newly installed digital control systems
that allow effective and safe control of the fuel and air systems within close
tolerances. On boilers with older control systems it may not be possible to achieve
such tight control.
CONCLUSIONS
The major conclusions of the program are:
1. Improved combustion optimization can provide
significant N0X reductions (23 to 55%) beyond those
achieved to meet compliance with the first generation
of SCAQMD N0X rules.
2. The incremental operating cost of these N0X reductions
is negligible (average of $.59/1b N0„) compared to
retrofit technologies. In some cases , operating
savings are achieved due to excess 02 reductions.
3. These techniques can be implemented safely with no
adverse impact on unit operation.
8-10
-------
16D
165
-
JN
o
150
-
3*
135
-
n
A 1
®
120
E
105
[
o.
a
90
X
-
.tf
?5
i ¦
E
eo
-
3
45
-
o
30
15
0
- NOTE:
z
ASSUMES UNIT HEAT RATE =
" 9,000 Btu/kW-hr
I' . . 1 i
i 1 1
1970 1980 1990
YEAR
1.5 ¦?
5
. z
£
1
X
O
z
2000
Figure 1. Gas Fuel NOx Limits on Alamltos Unit 5 arid Redondo Unit B
WEST FIRING WALL iVIEW FROM INSIDB
0
D
~
D
©
0
' ©
©
?s -
Q
©8S
5N _
©
Qm
©
0
' ©
©
35 Q
o4S
1N©
-2N
©
4'4"
" 9-»- » »¦ » «¦ * «¦ * 9. * rirS
PLAN VIEW
FIRING WALL
FIRING WALL
DIVISION
WALL
FIRING WALL
FIRING WALL
I ¦>
>- - -•>— — J
33'
-48-
Figure 2. Burner and NOx Port Locations on Alamitos Unit 6
(Redondo 8 Is a Mirror Image)
8-11
-------
*
«
E
&
£
OSAIEUHE
°MT, BURNER I OQt
100 250 309 400
UNIT LOAD, MW NET
Figure 3. NOx versus Load lor Baseline and Best
Conditions for Redondo Unit 8
O
*
100
to
«0
re
8a
so
m
so
20
10
0
a teoiirw
O 2S0 HW
OWKW
O 480 MW
X \
°\ °v
\ ^
X"
WCREASIN^TOR
21 20
WINDBOX 02, %
Figure 4. NOx vs. Windbox for Redondo Unit 8
8-12
-------
UMIT LOAD, WW NET
Figure 5. Impact on NOx of Taking Burner 6
Out of Service for Redondo Unit 8
UNIT to AD, WW NET
Figure 6. Impact on CO of Taking Burner 6 Out of
Service for Redondo Unit 8
8-13
-------
UNIT LOAD, MW NET
Figure 7, NOx versus Load for Load Following Tests
at Best Conditions for Redondo Unit 8
a IOC 20D 300 400 WO
UNIT LOAD, MW NET
Figure 8. NOx versus Load for Baseline and Best
Conditions for Alamltos Unit 5
8-14
-------
UNIT LOAD, MW NET
Figure 9. NOx versus Load for Load Following Tests at Best
Conditions for Alamltos Unit 5
8-15
-------
TABLE 1
SUMMARY OF NOx REDUCTIONS
ACHIEVED IN REDONDO 8
COMBUSTION OPTIMIZATION PROGRAM
Unit Load
MW Net 160 250 360 480
Baseline,
ppm NO @ 3% 0, 26 39 63 88
lb/MW-hr1 0.38 0,55 0.84 1.19
lb/MW-hr2 0.32 0.47 0.72 1.02
Best Case,
ppm NO 0 3% Gz 20 24 30 55
lb/MW-hr1 0.22 0.28 0.39 0.73
lb/MW-hr2 0.22 0.29 0.35 0.64
% Reduction 23% 38% 52% 38%
(ppm @ 3% 02)
First lb/MW-hr number is calculated from plant CEM data divided by plant MW data
Second lb/MW-hr number is calculated from trailer N0X ppm and Roseimount heat rate
by I/O method
TABLE 2
PERCENT REDUCTION ACHIEVED BY THREE
N0X REDUCTION TECHNIQUES AT REDONDO UNIT 8
Unit Load
MW Net
160
250
360
480
Increased GR
to windbox
5%
26%
43%
37%
Minimize 02
*
CSI
.13% '
12%
7%
Take Burner
6 OOS
5%
7%
5%
6%
Combining all 3
techniques
. 23%
38%
52%
34%
* At 160 MW, 02 could be reduced further before significant CO formation
8-16
-------
TABLE 3
SUMMARY OF NGX REDUCTIONS
ACHIEVED IN ALAMITOS 5
COMBUSTION OPTIMIZATION PROGRAM
Unit Load
MW Net
150
250
360
480
Clean Furnace
Baseline NOx;
ppm 0 3% 02
Ib/MW-hr
Best Case N0X:
ppm @ 3% 02
Ib/MW-hr
% Reduction
Dirtv Furnace
Baseline N0X:
ppm i 3% 0?
Ib/MW-hr
Best Case NO :
ppm § 3% u,
Ib/MW-hr
% Reduction
32
0.42
14
0.19
51
0.61
29
0.35
43%
56%*
58
0.69
36
0.43
38%
59
0.67
35
0.40
41%
66
0.75
43
0.49
35%
94
1.06
69
0.78
27%
96
1.09
***
All three reduction techniques were not combined at 150 MW with a clean
furnace
Reduction from clean furnace baseline
No comparable baseline data available at full load with dirty furnace,
since maximum FGR was required to meet 125 ppm compliance limit
8-17
-------
TABLE 4
PERCENT REDUCTIONS ACHIEVED
BY THE THREE NOx REDUCTION TECHNIQUES
ON ALAM1TOS 5
Unit Load
MW Net
150
250
360
480
Increase GR to
windbox
36%
36%
29%
11%
Take Burner 6 00S
18%
10%
13%
13%
Open NOx ports
9%
. 9%
9%
1%
Combined techniques
56%.
43%
41%
27%
8-18
-------
TABLE 5
HEAT RATE PENALTIES ASSOCIATED WITH NOx REDUCTION TECHNIQUES
REDONDO UNIT 8
Load
160 MW
250 MW
360 MW
480 MW
Increase FGR
(higher aux. power)
0.06%
0.16%
0.33%
0.21%
Burner 6 00S
{higher 02)
0.12%
0.04%
-0,04%,
0.08%
Minimum 02
-0.48%
-0.04%
-0.32%
-0.32%
Net heat rate
penalty (gain)
(0.30%)
0.16%
(0.03%)
(0.03%)
Avg. heat rate,
Btu/kW-hr*
10,209
9,645
9,327
9,415
Base hourly fuel
cost, $/hr**
5,717
8,439
11,752
15,817
Efficiency penalty (gain),
S/hr
($17)
$14
($4)
($5)
lb/hr NQX Reduced
12
45
135
167
$/lb N0X Reduced
(1.42)
0.31
(0.03)
(0.03)
* Average of data collected during test program
** Assumes $3.50/MMBtu fuel cost
8-19
-------
TABLE 6
HEAT RATE PENALTIES ASSOCIATED WITH NOx REDUCTION TECHNIQUES
ON ALAMITOS 5
Load
150 MW
250 MW
360 MW
480 MW
Increase FGR
(higher aux, power)
0.33%
0.20%
0.21%
0.06%
Burner 6 OOS
(higher 02)
0,28%
0.12%
¦ 0.04%
0.12%
NO Ports Open
(higher 02)
-0.10%
0,14%
0.12%
0.12%
Net heat rate
penalty
0.51%
0.56%
0.37%
0.30%
Avg. heat rate,
Btu/kW-hr*
10,880
9,820
9,430
9,320
Base hourly fuel
cost, S/hr* ,
5,710
8,590
11,880
15,660
Efficiency penalty,
$/hr
$29
CO
¦
w
$44
$47
Ib/hr N0X Reduced
35
66
99
137
5/1b N0X Reduced
0.83
0.73
0.44
0.34
* Average of data collected during test program
** Assumes $3,5Q/MMBtu fuel cost
8-20
-------
NOx REDUCTION AND OPERATIONAL PERFORMANCE OF TWO FULL-SCALE
UTILITY GAS/OIL BURNER RETROFIT INSTALLATIONS
N. Bayard de Volo
L. Larsen
Energy Technology Consultants, Inc.
Irvine, California
L. Radak
R. Aichner
Southern California Edison Co.
Rosemead, California
A. Kokkinos
Electric Power Research Institute
Palo Alto, California
8-21
\
-------
-------
ABSTRACT
In 1989-90 Southern California Edison Company replaced the original burners
firing natural gas and residual oil fuels in two large, opposed-fired boilers of
different capacities and design. The replacement burners were manufactured by Todd
Combustion, Inc of Stamford, Connecticut. The principal objectives of the retrofit
were: 1} to improve flame shape and stability, and 2) to achieve NOx emission levels
with all burners in service at full load, in combination with Flue Gas Recirculation
(FGR), equal to or less than the levels previously achieved by Off-Stoichiometric
firing with FGR.
Tests were conducted on both boilers, firing gas and oil fuels separately, to
define the flame shape and stability and the NOx emissions over a wide range of
load, excess air and FGR rate for both pre- and post-retrofit configurations.
Further reduction in NOx emissions achievable with the new burners firing in an Off-
Stoichiometric mode, with FGR, was also determined over the same range of
operational variables.
This paper is an interim status report presenting preliminary results of the
pre- and post-retrofit testing program funded by SCE and EPR1.
8-23
-------
INTRODUCTION
In 1987 Southern California Edison Company (SCE) initiated projects to replace
existing gas/oil burners on two large boilers, Alamitos Generating Station, Unit 6,
and Ormond Beach , Generating Station, Unit 2. The principal motivation in each case
was to improve flame quality (stability, attachment, etc.) over the load range, but
especially at low firing rates. Additional motivations included improving boiler
efficiency and reducing NOx emissions.
In order to define the actual improvements achieved by each retrofit, SCE
instituted a program to perform comprehensive testing of both boilers before and
after the burner retrofits. EPRI provided additional funds to expand the parametric
testing and to promote the dissemination of the NOx technology results to the
electric utility industry. Energy Technology Consultants, Inc. (ETEC) was retained
to provide consulting services to plan and conduct the testing program, to analyze
the test results and to report on the program findings. This paper is written to
present some preliminary results comparing pre- and post-retrofit NOx emissions for
natural gas and oil fuels. The program is still in progress and a considerable
portion of the post-retrofit testing remains to be completed for both gas and oil
fuels. Nevertheless, because there is currently so 1ittle public information
available on full-scale, Low-NOx gas/oil burner performance, it was thought to be
useful to present these preliminary results at this time.
Considerable success has been achieved by utilities having to comply with
restrictive NOx regulations applying to existing gas/oil fired units by implementing
Off-Stoichiometric (O.S.) firing. In this mode of operation, selected burners are
taken out-of-service (BOOS) while fuel flow is compensatingly increased to the
remaining burners to maintain boiler load requirements. As a result, the active
burner combustion process is made fuel rich and consequently NOx formation is
reduced. Although NOx emissions can be significantly reduced in this manner for
both gas and oil fuels, operational performance can also be degraded somewhat as a
consequence of having to raise excess air levels to maintain' acceptable CO
concentrations on gas fuel and plume opacity/particulates on oil fuel. In addition,
a degradation in flame holding and stability can also result. SCE has employed O.S.
firing on all of its units for many years achieving significant reductions in NOx
emissions but has also experienced the deterioration of boiler performance and
combustion on selected units.
The basic concept of low NOx burners is to achieve fuel rich combustion, and
hence reduced NOx formation, by controlling local mixing of fuel and air. This
approach offers the promise of equaling or exceeding the NOx reduction capability of
O.S. firing while avoiding the possible performance and operational deficiencies
associated with the latter approach. The potential gains however must be balanced
against the capital cost of the burner retrofit in comparison to O.S. firing which
is implemented operationally without equipment expenditure.
This paper should be of interest to utilities who anticipate having a future
need to reduce NOx emissions from their gas/oil fired boilers. The subject program
represents one of the few instances in which data are to be developed for a low NOx
burner utility, boiler installation and for which a comparison of the relative NOx
reduction capabilities and overall performance of the two NOx control approaches can
be established. It is for this reason that SCE and EPRI have jointly funded the
program reported herein.
8-24
-------
PRE-RETROFIT OPERATION
Unit 6 at the Alamitos Generating Station (AGS-6) is a B&W, opposed-fired,
gas/oil fuel boiler/turbine/generator set rated to produce 480 MWe. The boiler was
designed with 16 two-burner cells, arranged in two rows of four cells on the front
and rear furnace walls. Ring type gas spud burners and constant-differential,
pressure-atomized, swirl-tip oil burners were provided. Dampered overfire air
ports, fed from the windbox, were provided above each top elevation burner cell.
Two gas recirculation fans were originally provided to extract flue gases from the
economizer exit and direct those gases to the furnace hopper area as an aid to
controlling steam temperatures at low firing rates (FGR).
The boiler began operation in 1966 and subsequently became subject to a Los
Angeles County APCD regulation limiting NOx emissions to 225 ppm (dry, 3% 0^) for
natural gas fuel and 325 ppm for fuel oil. The uncontrolled NOx emission with gas
fuel at full load was approximately 700 ppm. NOx emissions were reduced to within
the regulatory limit on both fuels by implementing O.S. firing. The optimum firing
configuration was determined to be with the bottom burners of the upper cells (i.e.
3rd elevation) out of service for both gas and oil fuels and with the OFA ports
closed.
Subsequently, the APCD NOx emission limit for natural gas was reduced to 125
ppm and 225 ppm for oil. Two booster fans were installed to extract flue gas from
the main gas recirculation fan outlets and to inject the flue gas into the
combustion air through orifices in the flow-metering air-foils within the air ducts
between the air preheaters and the windbox as depicted schematically in Figure 1.
The combination of windbox FGR (WFGR) and O.S. operation achieved compliance with
the reduced emissions limits for both fuels and the boiler has been operated in this
mode ever since.
Unit 2 at the Ormond Beach Generating Station (OBGS-2) is a Foster Wheeler,
opposed-fired, gas/oil fuel boiler/turbine/generator set capable of producing 800
gross MWe. The boiler was constructed with two sets of 2-burner cells at each of
four elevations on the front and rear furnace walls. Each two-burner cell is fed by
one gas and one oil supply pipe/valve, however, each individual burner had its own
air register control. Each burner had a constant-differential pressure-atomized,
swirl-tip oil gun and a cane-type gas burner with (8) eight canes fed from an
external ring manifold.
The boiler was originally designed to produce NOx emissions below 500 ppm
(dry, 3% 0?) for both gas and oil fuels. This was to be accomplished by including
overfire air (OFA) ports fed by the windbox. In 1969 it appeared that the Ventura
County APCD intended to establish a NOx emission limit of 250 ppm (dry, 3%02) for
both fuels. During construction of the OBGS units (1 & 2) WFGR was added to both
units. For each unit one dual-inlet fan extracted flue gas from the economizer
outlet ducts and injected the gas into the two combustion air ducts leading to the
windbox. The general configuration is depicted schematically in Figure 2. The WFGR
injection is accomplished through an array of perforated pipes located within each
air supply duct a few feet upstream of the rear windbox.
Upon commercial operation of 0BGS-2 in 1973, compliance with the 250 ppm NOx
limit was achieved with either fuel at full load by a combination of FGR, OFA and
limited O.S. firing. In 1975 the Ventura County APCD reduced the allowable NOx
emissions with gas fuel to 125 ppm (dry, 3% 0,). Because oil fuel was used
exclusively,for several years, compliance with the 125 ppm limit for gas fuel was
not demonstrated until 1977. Compliance was achieved by operation with 8 BOOS,
8-25
-------
maximum FGR (around 18%)) and load restriction to about 720 gross MWe. The use of
the OFA ports was discontinued.
Both units at OBGS have experienced severe boiler vibration under a variety of
"normal" operating conditions, possibly aggravated by the use of low-NOx firing
procedures. The optimum operating modes were determined on the basis of compliance
with NOx emission limits and acceptable vibration control, and consisted of maximum
FGR at full load (throttled back at reduced load) and with 8 out of 32 burners out
of service (3rd elevation-gas fuel, 2nd elevation-oil fuel).
Several substantial efforts were made to alleviate the incidence of boiler
vibrations, including installation of burner air register shrouds and readjustment
of boiler back-pass dampers. These efforts were partially successful in reducing
vibration.
As with the AGS units, operation at OBGS increasingly emphasized reduced load
operation at times of off-peak-demand. SCE determined that the flame conditions at
lower loads (ca 250 MWe) were not as secure as they desired. In addition, the OBGS-
2 steam system was modified in 1985 to permit continuous generation as low as 50
MWe. This increased the concern with flame stability (lift-off, etc.) at the
extremely low firing rates.
LOW NOx BURNER RETROFIT
In 1986, the Steam Generation Division at SCE, in conjunction with the System
Planning and Research Department, contracted with Todd Combustion (formerly a
Division of Fuel Tech, Inc.) to provide 32 gas/oil burners to replace the existing
burners at AGS-6, principally to improve low-firing-rate flame conditions but also
to provide reduced NOx emissions. Shortly thereafter, the Steam Generation Division
solicited competitive bids to provide 32 gas/oil burners for installation on OBGS-2.
The contract was also awarded to Todd Combustion. Again, the emphasis was on stable
combustion at all firing rates, with low-NOx and increased efficiency as additional
objectives.
Prior to installation of the Todd burners at AGS-6, SCE obtained a Permit to
Construct from the South Coast Air Quality Management District (SCAQMD), which
stipulated that the NOx emissions post-retrofit must not exceed 113 ppm on gas fuel
and 203 ppm on oil fuel. An additional requirement was that NOx emissions over the
load range must be at least 10% below comparable emissions pre-retrofit, and that CO
emissions could not increase.
The Todd Dynaswirl® burner relies upon control of the combustion air in
several component streams, as well as the controlled injection of fuel into the air
streams at selected points, for maintaining stable, attached flames with low NOx
generation. Figure 3 schematically illustrates the internal configuration of the
burner.
For gas firing, fuel is introduced through six pipes, or pokers, fed from an
external manifold. The pokers have skewed, flat tips, perforated with numerous
holes and directed inward toward the burner center!ine. Gas is also injected
through a central gas pipe with multiple orifices at the furnace end. A single oil
gun is located along the burner center!ine, inside the gas pipe.
Primary and secondary air streams flow from the surrounding windbox plenum
through a spun cone inlet to the burner. A shut off damper is provided at the
burner inlet. The primary air stream flows into the burner and down the center of
the venturi around the center fired gas gun where it mixes with the center gas
8-26
-------
forming a stable flame in front of the swirlers. The secondary air flows into the
burner flows near the outer walls of the venturi where it mixes with fuel from the
gas pokers and is ignited by the stable center flame. The testing air stream is
controlled by a separate slide damper and flows between the venturi evase and the
burner throat quarl. A piezometer ring is provided at the venturi vena contracta
for comparison to pressure at the burner inlet; the pressure signal of about 2.5
times the windbox to furnace pressure loss provides an accurate measurement of
combustion air flow rate.
The oil gun is a conventional constant-differential, pressure-atomized burner.
The original single orifice swirl tip was replaced with a multi-orifice proprietary
design to reduce boiler vibration, however the turndown ratio is still of some
concern, and efforts continue to improve the turndown while maintaining good flame
quality and low NOx emissions. A swirl impeller is attached to the oil gun support
pipe just at the end of the primary sleeve section.
In performance of the retrofit contract, Todd Combustion performed flow model
analyses of the windbox air flow distribution. Based upon those analyses, baffles
and turning vanes were installed at selected points in the windbox to improve the
uniformity of air flow to all burners.
Following selection of the Todd Dynaswirl burner for retrofit to OBGS-2, SCE
obtained a "Permit to Construct" from the Ventura County APCD. The permit
conditions specified that the new burners would produce no increase in the emissions
of NOx, CO, total particulate and Volatile Organic Compounds (VOC), over the
operating load range, as compared to pre-retrofit emissions. Windbox modifications
to improve air flow uniformity were also made on this unit.
TEST METHODOLOGY
Comprehensive measurements of gaseous emission species (NOx., CO, 02) were made
for the pre- and post-retrofit testing phases of both boiler retrofits. The scope
and conduct of both boiler test programs were essentially identical.
Gaseous emissions were measured by an extractive sampling/conditioning/
measurement system contained within a mobile van. Gaseous analyses included
chemiluminescent (NOx), non-dispersive infrared (CO, C02) and fuel cell (oxygen)
types. All measurements were made after drying the sample gases.
The sample flue gas was extracted through stainless steel probes located in a
matrix across the economizer exit ducts. Measurements could be made of any single
probe sample or a composite of any combination of probes. Composite samples ensured
an equal portion from each probe by passing each individual sample through a
valve/bubbler prior to mixing within a common manifold.
At AGS-6, a similar matrix of probes was located in the air supply ducts
between the air foils (FGR injection) and the windbox. At OBGS-2 the FGR/Air
mixture was measured by sampling from pressure-tap tubing located adjacent to each
burner air register.
The FGR rate was calculated as the volumetric percentage of the flue gas
extracted from the exit ducts and injected into the combustion air. The calculation
was made based upon the dilution of gas species caused by the mixing process, i.e.
the comparative concentrations of 02, C02 and NOx within the flue gas alone and the
flue gas/air mixture supplied to the burners.
8-27
-------
Since the QBGS-2 permit required a demonstration that particulate arid
hydrocarbon emissions did not increase following the burner retrofit, tests were
conducted to measure TSP (oil fuel only) and VOC (both fuels) pre- and post-
retrofit. TSP was measured using a modification of EPA Method 5, in which the back
end catch was analyzed in addition to the front end catch (filter plus probe
washing). VOC was measured by capturing flue gas in Tedlar bags and analyzing for
C2 to C8 by GC/MS. Triplicate measurements of Total Suspended Particulate (TSP) and
Volatile Organic Carbon (VOC) were made for each of four load.levels from 250 to 700
MWe. Analyses were made to determine the carbon content of the TSP filter catch and
the organic hydrocarbon content of the back-end catch.
Each test was conducted with operation as close to steady state as possible,
with the load blocked on manual control. The boiler fuel, air and steam controls
were generally on "automatic" except that excess air trim and PGR settings were
manually controlled. In general, each test lasted from 30 minutes to 2 hours,
depending upon the complexity of gas measurement desired. In addition to the
emissions measurements,, considerable data were recorded regarding operating
conditions (e.g., fuel and air flows, pressures and temperatures, control/damper
settings, steam conditions, motor amps, boiler excess 02 and stack opacity).
TEST RESULTS
This section of the paper presents a brief discussion of selected test results
acquired to date. As pointed out previously, although pre-retrofit testing has been
completed, only limited test data have been acquired for the post retrofit, low NOx
burner configuration for the two units. Due to the limited extent of this latter
data and some present uncertainty in calculated WFGR rates (discussed below), it is
premature to draw definitive conclusions as to the demonstrated NOx control
capabilities of the two Todd burner installations and comparison with the pre-
retrofit NOx control configurations. This paper should be viewed therefore as an
interim status report which will be superseded by a future publication documenting
the completed program.
The testing of both units was constrained by the necessity to continue to
comply with the regulatory NOx limits of 125 PPM and 225 PPM respectively on gas and
oil fuels. This constraint prevented testing to determine the NOx reduction
capability of the Todd burner by itself in the absence of the utilization of WFGR at
higher loads, since emissions compliance could not have been maintained. This same
constraint applied to the pre-retrofit testing relative to demonstrating the
individual control capabilities of WFGR and 0.S firing on the two units. Some
estimate of these individual influences for both NOx control configurations for
Alamitos #6 have been made using historical data and PGR effectiveness trends as
discussed later in the paper.
ALAMITOS UNIT #6
Figure 4 shows representative test results acquired for the Todd burner
installation on AGS-6 over the load range. The calculated WFGR rate and measured
average exhaust gas 0, concentration associated with each test data point is
indicated. In general, the data reflect the maximum NOx reduction capability of the
installation. The indicated 0, levels at the higher loads ( >260 MW) are the
minimum achievable within the 5CE constraint of maintaining exhaust gas CO
concentration below 300 PPM. The lower load minimum 02 levels are constrained by
the necessity to maintain a minimum level of air flow as dictated by safe operating
procedures. The indicated. WFGR rate at the highest loads is near the maximum
8-28
-------
capability of the WFGR system for the test conditions. At the lower loads, the
indicated maximum WFGR rate is constrained by flame stability concerns although no
flame degradation in this regard was noted for the indicated levels.
The upper data point shown in Figure 4 at 480 MW applies to the all- burner-
in-service (ABIS) mode of operation which was the intended employment by SCE for the
for Todd burner installation. The level of NOx emissions achieved represents a
reduction of 87% for the combination of burner and 16% WFGR from the uncontrolled
level of approximately 700 PPM (best estimate based on historical data, could
possibly be higher). At 19% WFGR, the maximum capability of the FGR system, NOx
emissions would have been in the range of 75 PPM (extrapolated from Figure 5 data)
representing an 89% reduction from uncontrolled baseline.
The curve in Figure 4 is for O.S. operation with 8 BOOS. Although the O.S.
mode of operation was not intended by SCE at the time for normal employment, SCE
wanted to demonstrate the maximum NOx reduction achievable since it now must comply
with a significantly reduced emission limit. As Figure 4 indicates, the O.S. mode
of operation combined with 19% WFGR resulted in a further full load NOx reduction of
35% (from 75 PPM to 49 PPM) which represents a 93% NOx reduction from the
uncontrolled baseline level. This NOx control mode has been implemented by SCE for
normal operation.
A comparison of pre and post retrofit test results for a range of WFGR rates
is shown in Figures 5-7. The measured average exhaust gas Q? concentration
associated with each data point is indicated. The single data point shown in Figure
5 for the Todd burner operating in an ABIS mode indicates that less NOx reduction
was achievable than for the pre-retrofit O.S. mode.
With respect to the O.S. mode of operation, most of the post retrofit data
acquired thus far have been for higher WFGR rates than for the pre-retrofit data and
the minimal overlap for the two sets of data prevent a direct comparison over a
range of WFGR rates. However, the data do seem to demonstrate consistent trends
indicating that the Todd burner is capable of achieving lower NOx levels in an O.S.
mode than was possible pre-retrofit. This result appears to be due primarily to the
burner's capability to operate at lower 0, levels (discussed later) since both sets
of data show a clear trend of decreasing NOx with decreasing excess 02. This may be
only a partial explanation and the Todd burner may in fact produce lower NOx
emissions than pre-retrofit operation at identical excess 02 and WFGR levels. A
regression analysis will be performed on the expanded future data base to more fully
assess this question.
The WFGR rates were determined according to the procedure previously outlined.
There is a degree of uncertainty associated with the indicated values, however,
since a comparison between the calculated rates determined by the different methods
(02 or NOx dilution) showed random differences in the range of 10-15%. Since FGR
exerts a strong influence on NOx level, this degree of uncertainty could result in
appreciable error In the data as plotted and misleading apparent trends. This
potential deficiency will be more fully assessed in the continuing program and it is
believed that the relative level of uncertainty in calculated WFGR rates can be
reduced.
Figure 8 shows a comparison between pre and post retrofit NOx control
performance capability for the various control configurations. The NOx levels for
uncontrolled baseline and BOOS configurations are estimated based on 20 year old
test data. The indicated NOx levels for the other configurations are either current
measurements or extrapolations from these measurements. The comparison is tentative
since it is based on current limited data but is presented to provide the reader
8-29
-------
with a present estimate of the Todd burner NOx control capability for the Alamitos
unit as well as a comparison with the pre-retrofit control capability. The
comparison indicates that for like configurations, there is little difference in pre
and post retrofit NOx control capability in absolute terms, the maximum being either
91% or 93%. However, in relative terms, the difference of 16 PPM is significant to
SCE's NOx emission reduction objectives.
The demonstrated percent reductions are measured from an uncontrolled NOx
baseline level of 700 PPM. Experience with implementing O.S. firing has shown that
the percent reduction achievable on a particular unit is dependent on the magnitude
of the initial, uncontrolled NOx emission rate and decreases as this rate is
reduced. Therefore, it is likely that lower NOx control capability could generally
be expected for Todd burner installations on boilers exhibiting lower uncontrolled
NOx emission rates.
Figure 9 compares pre and post retrofit C0/02 trends. As shown, the Todd
burner demonstrated significantly improved performance over that achievable for the
pre-retrofit NOx control configuration. This gain in minimum achievable excess 02
level is partly responsible for the lower NOx emission rate obtainable with the Todd
burner retrofit and also offers a benefit in terms of boiler thermal efficiency.
The improved C0/02 performance of the Todd, burner installation can be
attributed in part to improved air/fuel flow uniformity to the burner arrays on the
two firing walls. This was achieved by a combination of windbox modifications made
in conjunction with the burner installation and balancing of the burner fuel and air
flows during shakedown testing. Therefore, part of the NOx and heat rate gain can
be credited against the windbox modifications independently of the burner
installation and the remaining part,to the burner itself. The relative contribution
of these two factors has not yet been assessed but answering this question is useful
in terms of comparing the NOx control capability of O.S. firing (whose
implementation could be accomplished in conjunction with windbox modification) with
the installation of a Todd INB.
Figure 10 is a plot of recorded CEM data (note scale is in LB/HR) acquired
post retrofit during the month of August, 1990 for unit operation over the normal
load range in both AGC and operator control modes. The significant data scatter can
be attributed to the normal variability of key parameter settings such as excess 02
and FGR rate and instrumentation variability. A similar plot has been prepared for
the pre-retrofit NOx configuration for the same period in 1987. Figure 11 shows the
best curve fits for each of the mentioned data sets and also a replot of the lowest
obtainable post retrofit NOx emission demonstrated as shown previously in Figure 4
Call in LB/HR).
The plots illustrate that single point data acquired in controlled testing of
the maximum NOx control capability configuration can significantly underestimate
achievable operational emissions as monitored by a CEM for demonstration of
regulatory compliance purposes. 'A comparison of the upper two curves also confirms
that the Todd burner installation was successful in reducing NOx emissions during
normal AGC operation.
Figure 12 shows pre-retrofit NOx emissions at selected loads on oil firing for
the ABIS and BOOS modes of operation. Post retrofit oil firing data have not yet
been acquired and the data are shown for general interest.
In terms of operational performance, the Todd burner installation has
satisfied all of SCE's original objectives with the exception of turndown on oil
firing which has not yet been demonstrated. Flames are stable over the load range
8-30
-------
Including minimum load and do not exhibit any tendency to lift off under normal
operating conditions. In addition, operating excess 02 level has been significantly
reduced for gas firing thereby yielding a meaningful improvement in boiler thermal
efficiency.
ORMQND BEACH UNIT #2
Figure 13 shows representative pre and post retrofit test results over the
load range for OBGS-2 firing gas fuel. The data points apply to minimum excess 0,
levels and approximately to the same near maximum WFGR rate at each load level. The
data indicate that the Todd burner installation reduced NOx emissions to below
obtainable pre-retrofit levels for the ABIS mode of operation and a further
increment in NOx reduction was achievable for O.S. operation (third row BOOS).
Uncontrolled full load NOx emissions are believed to have been in the range of
1200-1500 PPM and therefore the controlled full load emissions for any of the
configurations (LNB or original burner with O.S. and with WFGR) represent a
reduction of at least 92%. This magnitude of percent NOx reduction is nearly
identical to that achieved on AGS-6. Unlike that unit however, post retrofit ABIS
NOx emissions at OBGS-2 are lower than the best obtainable pre-retrofit NOx
emissions by approximately 10% at full load. The test results in the O.S. mode
shows an incremental reduction of 20% from the pre-retrofit level at full load as
indicated in Figure 13.
The general range of pre and post retrofit CO concentrations measured verses
excess 02 is shown in Figure 14 for gas fuel at loads of 550 MW and above. The C0/02
trends are approximately the same for the pre-retrofit O.S. and post retrofit ABIS
modes of operation while post retrofit operation in an O.S. mode exhibited higher CO
concentrations at comparable 0? levels. These results are at variance with those
demonstrated for AGS-6 which snowed an improvement in the CO/O, post retrofit trend
for the O.S. operating mode in comparison to pre-retrofit results. CO
concentrations for this latter unit operating in an ABIS mode have not yet been
measured. The results are surprising since the windbox modifications made to
improve air flow uniformity were expected to result in an improvement in the C0/02
trend as compared to pre-retrofit conditions.
A comparison of pre- and post-retrofit NOx emissions for oil firing is shown
in Figure 15. The data indicate that the Todd burner achieved lower NOx emissions
at full load operating in an O.S. configuration than was obtainable for pre-
retrofit. Since the data are limited and there is some uncertainty in the Indicated
WFGR rates, further analysis is required to confirm this result.
For gas fuel there was no increase in measured V0C emissions for operating
conditions consistent with lowest-NOx emissions, {O.S. operation, low excess 02 and
high FGR rate). Similarly for oil fuel there was no measured increase in either
solid carbon or condensible hydrocarbons, again under lowest-NOx operating
conditions.
The post-retrofit condition of the flames was substantially better than pre-
retrofit under all operating conditions, even at 50 MWe with all air registers open,
high FGR rates (up to 40%) and high excess air (25% of rated flow). Under all
conditions the flames were closely attached to the burner tip/throat area and were
steady and symmetrical. Prior to retrofit the flames were frequently detached from
the burner throat by as much as three to four feet, pulsated irregularly and were
occasionally irregular in shape.
8-31
-------
Prior to the burner retrofit, severe boiler vibration (rumbles and furnace
wall pulsations) were experienced under certain "normal" conditions of load, excess
air, FGR rate and burner firing pattern. Although the severe vibration could
usually be avoided, or corrected by an experienced operator, the condition was of
concern to the operating and engineering staff. Following the burner retrofit, the
unit generally operates more smoothly and the most severe vibrations no longer
occur. It should be noted that simultaneously with the burner retrofit, the FD fans
were modified from constant-speed with inlet vane flow control to variable speed
with no inlet vanes. Although it is uncertain whether the fan modifications
contributed to the reduced vibration, the change has definitely reduced the
operating noise level and has significantly improved the control and steadiness of
the air flow.
CONCLUSIONS
It is premature in view of the limited post-retrofit test data acquired thus
far to draw definitive conclusions relative to the pre and post retrofit NOx
emission control performance comparison. It is possible, however, to make some
observations on the basis of the data that have been acquired which are expected to
be valid at the conclusion of the program.
1) Full load gas fired NOx emissions for both units with the Todd burner
installation combined with approximately 20% WFGR have been reduced by 93%
from the uncontrolled baseline NOx level. This reduction was achieved by
operating in an O.S. mode with 25% BOOS.
2) The pre-retrofit NOx control configuration of O.S.operation (25% BOOS)
combined with 20% WFGR demonstrated nearly the same NOx reduction as post-
retrofit from the uncontrolled baseline level for full load gas fired
operation. The difference in demonstrated relative NOx control capability
amounting to a further reduction of about 20% from the pre- retrofit level
could be meaningful for utilities facing very stringent NOx emission
control regulations such as SCE.
3) Achievable NOx emissions employing either control configuration during
normal AGC operation will be significantly higher than that demonstrated
in the controlled testing conducted in this program.
4) The C0/0? performance demonstrated by the Todd burner installations owed
conflicting trends in comparison to the pre-retrofit test results. VOC
emissions on gas fuel and particulates on oil fuel did not increase with
the installation.
5) The burner retrofit demonstrated significantly improved operational
performance relative to pre-retrofit in terms of flame holding, stability
and boiler vibration.
8-32
-------
8-33
-------
00
1
CO
Front
wlndbox
4
3
Burner
level 2
1
Air from FD fans (2)
Damper
FGR fan (2)
Figure 2: Alamitos - Unit 6 Air/FGR Configuration
-------
Injection Venturi (2)
Array (2)
Figure 3: Ormond Beach Unit 2 Air/FGR Configuration
-------
100
BO
O
*
CO
<§> 60
¦q
a
I2l
o
40
20
ALAMITOS #6 |
-
GAS FUEL
• 16%WFQR
• ABIS
0.8 % Excess 02
A 3rd Row BOOS!
Uncontrolled NOx Emission 700 ppm
* 19
A44
/ 1.0
T 10.2
^23 '
0.9
\
0.9
40 ~ ~ 40
3.2 1.1
i i I l.
l
0
Figure 4:
120
100
soo
600
200 300 400
GROSS GENERATION (MW)
Minimum Achievable Post Retrofit NOx Emission Over the Load Range
110
¦f 100
CO
2-
¥
Q.
Q.
x"
O
z
90
80
70
60
50
40
a2.o
A A 2.2
-i.2.3
2.2
A.
A 2.8 '
2,0 A A2-4
1.6 A 2.2
A1-0
0.8 % Excess 0.
2 .
ALAMITOS #6
GAS FUEL, 480 MW
a Pre Retrofit - BOOS
• Post Retrofit - ABIS
± Post Retrofit - BOOS
1.3
a'*®
~
1.45
^0,8
M.4
1.0
10
12 14
WFGR, %
16
18
20
Figure 5: Comparison of Pre and Post Retrofit NOx Emission at 480 MW
8-36
-------
120
100
m
@ 80
¦Q
E
a.
o.
X
O
60
40
20
¦ A'
A2.5
Au
6
A.2.6
Au
A2.3
A2.fi
2,11.9
45*
A. 85
ALAMITOS #6
GAS FUEL, 360 MW
A Pre Retrofit - BOOS
A Post Retrofit - BOOS
A 2-5% Excess 0.
A1.4
A-,4 #.
1.6
;i.O ±1.3
1-1 A,85
10
25
30
15 20
WFGR, %
Figure 6: Comparison of Pre and Post Retrofit NOx Emission at 360 MW
35
cf
CO
TJ
CL
a
*
O
z
90
80
70
60
50
40
30
20
10
' -'2.8
' A,8
ALAMITOS #6
GAS FUEL, 260 MW
A Pre Retrofit - BOOS
A Post Retrofit - BOOS
^2.3 % Excess 0,
A2.5
A:
1.7 A ^.2
1.8 ^
££0
2.3 A 'r2'7
Z$,2 Z^.O A2-1
¦ 1.9 A A^'8
0.7 A 1.3 A ^ ? ^.6
M'|l4
0,9 M
A2.3
10
l1.5
N.i
40
20 30
WFGR, %
Figure 7: Comparison of Pre and Post Retrofit NOx Emission at 260 MW
50
8-37
-------
1,000
Q.
Q.
X*
O
z
800
600
400
200
700 fi
ppm ¦
Pre Retrofit
Post Retrofit
Uncontrolled
Baseline
BOOS
Rsductiort from .
' 75%
Baseline i 174
91%
1
89%
65
1 5
m
i P
BOOS ABIS ABIS
+
+
19% FGR
19% FGR
93%
49
BOOS
19% FGR
Figure 8: Comparison of Alamitos #6 Pre arid Post Retrofit NOx Emission at Full
Load on Gas Fuel
700
600
500
ALAMITOS #6
GAS FUEL
250-480 MW
Q.
Q.
400
o
300
200
100
0.5
Post Retrofit
(O.S.)
EXCESS 02, PERCENT
Figure 9: Comparison of Pre and Post Retrofit CO Emission
8-38
-------
500
450
400
350
w
€
300
n
250
X
o
z
200
150
100
50
0
ALAMITOS #6
GAS FUEL
50 100 150 200 250 300 350 400 450 500
LOAD, MW
Figure 10: CEM Data for the Month of August, 1990
550
500
450
400
350
XT
300
£
><
250
o
z
200
150
100
50
0
ALAMITOS #6
GAS FUEL
Beat Fit of August, 1987.
CEM Data, (Pre-Rfitrofit)
Best Fit of August, 1990,
CEM Data, (Po»t-Ritroft)
Port Retrofit
Minimum Achievable
(Figure-4)
50
500
100 150 200 250 300 350 400 450
LOAD, MW
Figure 11: Comparison of Pre and Post Retrofit "Best Fit" Curve of CEM Data
and Post Retrofit Minimum Achievable NOx Emission
550
8-39
-------
170
160
-04
150
CO
£•
¦O
140
130
a
o.
O
z
120
110
100
ALAMITOS #6
OIL FUEL
o ABIS
a BOOS
Ol 2-6% FOB
3.5
©17.8
3.5
£14.2 % FGR
2.8 %<^
Q24.2
3.5
A207'
3.2
A 18,4
3.6
250
300 350 400 450
LOAD, MW
Figure 12: Pre Retrofit Oil Fuel NOx Emission
500
120
400 500 600
GROSS GENERATION, MWe
Figure 13: Comparison of Minimum Achievable Pre and Post Retrofit NOx Emission
over the Load Range for Gas Fuel
8-40
-------
1,200
1,000
33
Q.
Ql
800
600
o
° 400
200
Pre Retrofit (O S )
Post Retrofit
-------
-------
COMPARATIVE ASSESSMENT OF NOx REDUCTION TECHNIQUES
FOR GAS- AND OIL-FIRED UTILITY BOILERS
Gary L, Bisonett
Steam Generation Department
Pacific Gas and Electric Company
San Francisco, California 94106
Mike McElroy
Electric Power Technologies, Inc.
Berkeley, California 94705
8-43
-------
-------
ABSTRACT
Pacific Gas and Electric Company conducted a comparative assessment of commercially available
NO* control technologies that might be applicable to our gas- and oil-fired boilers. One phase
of the assessment, cofunded by EPRI, was a comparative cost and feasibility analysis of various
commercially available technologies, including combustion modifications, low NOx burners, and
selective catalytic reduction. The results of this study axe being incorporated into efforts
to identify a cost-effective systemwide NOx control strategy for our system. The comparative
assessment was conducted based on a typical boiler in our system to allow technology comparisons
to be made on a consistent basis; Once the information for each technology was developed, the
site specific factors that affected each technology were identified so that the results could be
generalized and modified for other boilers in our system. One aspect of the project was to
develop a computer program, also cofunded by EPRI, to help PG&E compare various NOx control
strategies for possible application in our system. The computer program provides a first-cut
analysis of NOx reduction costs given different projected NOx limits and compliance strategies.
8-45
-------
INTRODUCTION
Pacific Gas and Electric Company (PG&E) performed a multi-faceted engineering program to
identify and evaluate options for reducing NOx emissions from its gas- and oil-fired electric
generating units. The program, involving the 39 boilers in the PG&E system, had two primary
goals; (1) Evaluate and compare the technical and economic merits of commercially available
retrofit NOx control technologies and their applicability to PG&E's boilers; and (2) Develop a
computer model to assist PG&E in developing an optimum system-wide NOx control strategy.
The program was prompted by concerns for lower NOx emission requirements for California
utility boilers.
The program was performed with refunding and technical participation from the Electric Power
Research Institute (EPR1). The involvement of EPRI was in recognition that the PG&E program
would be a valuable case study for the utility industry, and the results could assist other
utility companies planning or engaged in similar NOx control assessments.
PG&E is one of the largest investor owned gas and electric utilities in the United States.
PG&E's fossil fuel fired electric generating capacity is centered in seven stations located
throughout the Company's service territory which encompasses much of northern and central
California, PG&E's gas- and/or oil-fired boilers total over 7,600 megawatts of electrical
capacity, and represent a wide cross-section of manufacturers, furnace designs, combustion
systems, equipment sixes, and vintages. PG&E's 345 MW opposed-fired boilers (manufactured
by Babcock and Wilcox) comprise one-third of the capacity, and were the focus of the program.
NOx control measures have been previously implemented on these and several other PG&E boilers,
including overfire air, flue gas recirculation, low excess air operation, and biased firing.
The California Clean Air Act which was passed in 1988 requires local air pollution control
districts to develop plans to attain ambient air quality standards in California. The
California ozone ambient air quality standard is 25 percent more stringent than the Federal
ozone standard. This requires a very aggressive program on the part of regulators to develop
plans to attain the California ozone standard. PG&E's goal is to work closely with regulators
to identify emission reduction plans that are both exist effective and responsive to the air
quality needs of the communities we serve.
Since the completion of this study, PG&E has continued to develop site-specific information to
identify cost effective strategies for reducing NOx emissions. This program is ongoing and will
continue as information from other installations, R&D, and die regulatory process becomes
available.
8-46
-------
PG&E NOx CONTROL ASSESSMENT PROGRAM
Program Scope
The PG&E program consisted of the following work elements:
1. Verify existing boiler NOx emissions as a function of load for each boiler,
using existing field test data, supplemented as necessary with NOx
emission predictions based on furnace heat release rate correlations.
2. Compile detailed listings of boiler-specific operating and physical data
that are related to NOx formation.
3. Evaluate the applicability and NOx reduction potential of operational
modifications (e.g., bumers-out-of-service and biased firing) for the
entire PG&E boiler population. This work was based upon previous
experience with such controls within PG&E and elsewhere in the utility
industry.
4. Assess the technical feasibility of retrofitting state-of-the-art low-NOx
combustion systems for three selected boilers, and develop NOx reduction
and cost factors for the technically feasible options.
5. Perform limited field tests on one unit (Contra Costa Unit 6) to validate
predictions of NOx reduction achievable by combustion modifications.
6. Conduct comprehensive technical and economic assessments for low-NOx
combustion and Selective Catalytic Reduction (SCR) for a selected boiler
(Contra Costa Unit 6).
7. Rank each potential NOx control option evaluated during the study by cost,
NOx reduction potential, and technical risk. Also, identify the site
specific factors that influenced the rankings.
8. Construct a NOx emission forecast model which utilizes the above results to
identify the NOx controls required to meet specified system-wide or
regional emission limits at minimum cost.
9. Develop hypothetical NOx compliance strategies for different levels of
system-wide NOx reduction utilizing the NOx emission forecast model.
Contra Costa Unit 6 was selected for the retrofit feasibility study (Item 4 above), and for
detailed engineering and cost evaluations (Item 6), because it is representative of a boiler
design that constitutes one-third of the PG&E fossil system capacity. Less detailed feasibility
studies where also performed on two other PG&E boiler designs which posed distinctly different
retrofit situations (Moss Landing Units 6 & 7, and Pittsburg Units 5 & 6). Each of the three
selected boilers were already operating in a reduced-NOx mode (with flue gas recirculation to
the windbox and combustion staging) which was the baseline condition for the feasibility and
engineering studies.
8-47
-------
For the three plant sites, operation with natural gas and residual oil was considered. Fuel oil
with a 0.5 percent sulfur content, based on the maximum allowed by regulatory requirements, was
assumed. Since fuel nitrogen content is not constant and variations would affect the NO*
reduction attainable by a given combustion NOx control, values of 0.3 percent and 0.6 percent
(by weight) nitrogen in the oil were considered for purposes of NOx predictions.
PesCTptipn pf gpjlerS
Contra Costa Unit 6 - The unit is a forced draft, opposed-fired, drum-type boiler manufactured
by Babeoek & Wilcox with a rated generating capacity of 345 MW (gross). The unit was built in
1964. The unit fires oil and natural gas through 24 circular register burners arranged in two
rows of six burners on each firing wall. The furnace contains two division walls separated from
the furnace end walls and each other by two columns of burners. An elevation drawing of the
boiler is provided in Figure 1. In 1973-1974, overfire air ports were installed to reduce NOx
emissions in order to meet new NOx emission limits. Overfire air ports were installed in the
windbox, one above each burner column, for a total of twelve ports. In addition, the existing
hopper gas recirculation system was upgraded to mix up to 18 percent flue gas into the secondary
air duct feeding the windbox.
Moss Landini Units 6 and 7 - These two identical units, rated at 750 MW (gross), began operation
in 1967-68. These units, manufactured by Babcock & Wilcox, are forced-draft, supercritical
boilers. The units are opposed wall fired and were originally equipped to fire oil or natural
gas with 3-nozzle cell burners arranged in a two-high by four-wide array on each Firing wall (a
total of 24 burner throats on each wall). In the early 1970's, the existing hopper gas
recirculation system was modified to permit operation with up to 18 percent flue gas
recirculation with provisions to direct recirculated flue gas to the windbox for NOx control.
Also, the top nozzles of the upper four cell burners on each wall were modified to pass air
only, acting as localized overfire air ports to provide an additional NOx reduction.
Pittsburg Units 5 and 6 - The two identical units, designed by Babcock and Wilcox, began
operation in 1960-61. The units are forced draft, natural circulation boilers, with a rated
generating capacity of 330 MW (gross when fired with either natural gas or oil fuel. The units
were designed for future coal firing with a conversion to balanced draft. The boilers are
opposed fired with 24 burners arranged in two-high by six-wide array on each wall. In the early
197Q's, the units were modified to reduce NOx emissions by adding flue gas recirculation to the
windbox and installation of overfire air ports above the top burner row.
Program Participants
A majority of the work was performed by outside contractors selected on a competitive basis.
The major participants and their areas of prime responsibility are as follows:
• EFRI - Cofunding and participation in project technical direction.
• Babcock & Wilcox Company - Retrofit evaluation of low-NOx combustion
equipment options and Selective Catalytic Reduction.
• Fossil Energy Research Corporation - Development of NOx Emission Forecast
Model
• KVB. Inc. - Compilation of current (baseline) boiler NOx emission factors,
and evaluation of NOx reduction via operational modification.
8-48
-------
* Electric Power Technologies, Inc. - Provide technical and administrative
support to PG&E, including assistance in program planning, selection of
subcontractors, and analysis of results.
* PG&E - Overall project management and NOx reduction field tests at Contra
Costa Unit 6.
NOx CONTROL TECHNOLOGIES EVALUATED
The NOx control technologies that were considered in the NOx control evaluation include;
1. Operational Modifications to Existing Equipment
2. Combustion Equipment Modifications
* Two Stage Combustion (TSC)
* Rebuming
3- Postcombustion NOx Control
* Selective Catalytic Reduction (SCR)
Operational Modifications. The operational modifications evaluated were: (1) low excess air;
(2) bumers-out-of-service (BOOS), including selected gas spuds out of service for natural gas
firing, (3) fuel biasing, (4) optimization of existing overfire air ports (where installed); and
(5) optimization of existing windbox flue gas recirculation (where installed). Other
modifications considered, but not found to be cost-effective, were reduced combustion air
preheat and water injection.
Combustion Equipment Modifications. The combustion equipment modifications were commercial
combustion systems, offered by B&W. Each involved retrofit of low-NOx PG-DRB burners,
installation of dual register overfire air ports, and installation of a compartmentalized
windbox. Conceptually, the systems differed primarily in the arrangement and number of burners
on the firing walls, location of overfire air ports, requirements for additional furnace height,
and the control and distribution of air and fuel among the overfire air ports and burner
elevations. Each system was evaluated for a range of flue gas recirculation rates, both within
the existing FGR capacity and under conditions of increased FGR capacity. The scope of
modifications and retrofit equipment associated with each combustion technology is summarized in
Table I.
Four versions of rebuming were evaluated:
(a) In-Fumace NOx Reduction (IFNR)
(b) Pseudo-In-Fumace NOx Reduction (Pseudo-IFNR)
(c) Derate In-Furnace NOx Reduction (Derate-IFNR)
(d) Dual-Mode In-Fumace NOx Reduction (DM-IFNR)
8-49
-------
Versions (b), (c) and (d) were essentially compromise designs which attempt to minimize boiler
modifications [e.g., minimize or eliminate need for additional furnace height] compared to a
non-compromise, full rebuming system [version (a)]. Pseudo-EFNR utilized minimum furnace
residence time criteria for rebuming reactions, Derate-EFNR involved a load reduction on the
unit to satisfy rebuming residence time requirements, and DM-IFNR involved operation in an IFNR
mode below a certain load and TSC operation at higher loads.
A limited evaluation of B&W's XCL burners was also performed, as this technology became
commercial during the course of the study.
Selective Catalytic Reduction. The postcombustion SCR technology was a commercial system
offered by B&W through a licensing agreement with Babcock-Hitachi in Japan. The scope of
modifications and retrofit equipment are summarized in Table 1,
RESULTS
Operational Modifications
Maximum NOx reductions achievable from implementation of operational modifications to existing
combustion equipment were predicted to range from approximately 10 percent to as high as 60
percent from boiler to boiler (at full load).
The range reflects the varying degrees of NOx control already in place, and the site-specific
factors that influence the applicability and performance of these controls. The NOx reductions
typically associated with each control technique are as follows:
In general, due to the low cost of implementing operational changes, these options should be
considered as a first NOx control alternative.
Combustion Equipment Modifications
State-of-the-art low-NOx combustion controls, aimed at achieving minimum NOx emissions via
modifications to combustion equipment - specifically, TSC and rebuming -- were not universally
applicable to all boilers in the PG&E system. Moreover, the predicted NOx reductions with these
technologies, where technically feasible, varies considerably from unit to unit. Predicted NOx
reductions range from 20 percent to as high as 70 percent from existing levels, reflecting the
impact of site-specific factors, associated compromises in NOx control system design, and
specific NOx control design and operating conditions. These NOx reductions were calculated from
existing "baseline" boiler operating conditions in which the current use of flue gas
recirculation and various degrees of conventional combustion staging already result in reduced
NOx emissions. Larger percentage NOx reductions would be expected if the study boilers had not
been previously equipped with these NOx control measures.
Operational Modification
HP* Rsijurtion
Low Excess Air
Bumers-Out-Of-Service
Fuel Biasing
Overfire Air Optimization
FGR Optimization
5-10 percent
15-60 percent
20-50 percent
10-15 percent
5-20 percent
8-50
-------
The boiler-specific results concerning technical feasibility are summarized in the following
paragraphs. The predicted NOx emissions are summarized in a separate subsection below.
Contra Costa Unit 6: TSC could be applied, with burner rearrangement and significant ductwork
and windbox modifications. The relatively tight furnace, originally designed with minimum
residence times, would not accommodate any version of rebuming without major extensions in
furnace height. The change in furnace height required for implementation of IFNR is illustrated
schematically in Figure 2.
, Moss Landing Units 6 & 7: Application of low-NOx combustion systems is difficult due to
the 3-nozzle cell burner design, and the physical interferences from steam headers and mixing
equipment located halfway up the furnace walls in the windbox. A TSC system could be installed
but would require major modifications to the firing walls, including complete rearrangements of
the burner amy and windbox to accommodate new burners and overfire air ports. Pseudo-IFNR is
the only reburning option determined to be feasible, but would require a substantial increase in
furnace height as well as firing wall modifications similar to TSC. For both control options,
use of XCL burners instead of PG-DRB burners could reduce the retrofit complexity and cost.
Pittsburg Units 5 & 6: The relatively high residence time in the furnace (originally designed
for future coal conversion) greatly enhances retrofit feasibility. TSC can be retrofitted with
only minor modifications to the overfire air ports (the PG-DRB burner would fit into existing
burner openings). IFNR can also be applied without major furnace modifications-an additional
row of burners and new overfire air ports would be required.
Selective Catalytic Reduction
It is feasible to retrofit Selective Catalytic Reduction (SCR) to the Contra Costa Unit 6 to
achieve postcombustion NOx removals of approximately 80 percent. The design conditions and
operating parameters were concluded to be similar to SCR units operating in Japan,
Two possible SCR arrangement were evaluated for Contra Costa Unit 6: (1) Base Case -single SCR
reactor located in the existing air heater location, requiring relocation of air healers and FD
fans towards the stack; and (2) Alternate Case - two SCR reactors located above the existing
air heater locations, with air heaters and fans undisturbed. Schematics of both configurations
are shown in Figures 3 and 4.
8-51
-------
NOx Reduction Summary fpr Control Options
The predicted NOx reductions for the combustion modification options are summarized in Table 2
for the three boilers evaluated.
Figure 5 compares the NOx reductions predicted for combustion modifications and SCR applied to
Contra Costa Unit 6,
Piims impacts
For SCR, and the advanced combustion systems that were technically feasible, there appear to be
no adverse impacts on power plant performance, operation, or reliability that would preclude
their implementation. However, potential impacts were identified and incorporated into the
overall evaluation of control options. The potential impacts considered include;
Combustion Modifications
- Increased auxiliary power for higher FGR rales, where required.
• Potential increase in furnace tube wastage due to reducing
conditions.
- Boiler control system complexity.
- Changes in furnace excess air and resulting effects on plant heat
rate.
- Boiler startup and shutdown procedures.
- Potential for flame impingement.
- Burner turndown.
- Restrictions on rate of lead change.
- Potential localized connective pass tube overheating.
SdeqimCatalytic Rgdvctipn
- Potential air heater plugging when burning oil fuel.
- Increased minimum load or economizer bypass to maintain minimum SCR
temperature.
- FD fan upgrading to overcome increased system pressure drop.
- Boiler startup and shutdown procedures.
- Increased maintenance for SCR catalyst replacement and air heater
cleaning.
- Air heater wash water treatment.
- Ammonia emissions.
8-52
-------
Cost of NOx Control
The cost of retrofitting combustion modifications and SCR (1989 dollars) were evaluated
according the standard EPR1 Economic Premises. Capital costs ($/kW) included all materials,
engineering, installation, contingencies, and home office fees for a turn-key retrofit project.
Levelized costs (mills/kWh) included all operating and maintenance labor and materials,
administrative costs, and carrying charges. Levelized costs reported herein are for a base case
30-year levelization period and 30 percent capacity factor (other assumptions were evaluated in
the study to examine cost sensitivity to these parameters),
Low-NOx combustion system costs estimated for Contra Cost Unit 6 ranged from approximately
$4Q/kW to $50/Kw, with total levelized costs ranging from approximately 3 to 4 mills/kWh. These
cost estimates are higher than generic cost estimates in the open literature.
The capital cost of SCR ranged from approximately S72/kW to $82/kW, and total levelized costs
range from approximately 3 to 8 mills/kWh, depending on specific design and operating
assumptions.
A comparison of the costs of technically feasible NOx control options (JSC and SCR) for Contra
Costa Unit 6 are compared in Table 3. Approximately 30 percent of the Engineering & Material
costs for TSC are for low-NOx burners, burner accessories, and overfire air ports. For SCR,
approximately 40 percent of the Materials & Engineering cost is for the SCR reactor vessel,
including the casing, framework, and initial catalyst charge.
General Observations. The results of the study reinforce the following considerations
regarding the evaluation of utility boiler retrofit NOx controls:
1. The selection of an optimum NOx control approach for a specific boiler is
rarely obvious, without first performing detailed engineering and cost
analysis of the available technology options.
2. To provide a meaningful comparison of NOx control options, it is imperative
that a systematic approach be used which analyzes each potential control
technology under the same technical and economic premises.
3. Relying on generic technical and cost data is not advisable for evaluating
retrofit feasibility, NOx control cost, and potential NOx reductions for a
specific boiler or a utility generating system. Such an approach could
easily lead to substantial errors relative to a systematic, detailed
engineering and cost analysis of the same boilers.
4. Depending on site-specific constraints and NOx reduction requirements, it
is likely that a combination of NOx reduction techniques will provide the
overall least cost means of achieving those requirements.
Applicability and Value to Industry
The PG&E retrofit analyses involved a single boiler manufacturer's NOx control technology
applied to a few specific boilers. Although the technologies are representative of generic
classes of NOx controls that are offered by other vendors, it is likely that conclusions
regarding technical feasibility and cost would differ if performed by another manufacture
applying its versions of these technologies.
8-53
-------
There are other boiler design types within the U.S. utility industry that are not represented by
the units selected for evaluation in this study. Such boilers, including tangentially-fired
units and cyclone-fired boilers burning gas and oil fuels, can be anticipated to pose
substantially different retrofit constraints. Thus, a comparable feasibility analysis performed
on these units could have different results than those in this study.
Although the technical and cost evaluations may be pertinent to some retrofit situations
encountered elsewhere in the industry, for the reasons enumerated above, feasibility and
engineering/cost analyses specific to each utility company are required. However, the
methodology used in this study is generally applicable across the industry, and can be applied
by other utility companies performing NOx assessments of their generating systems.
The value of this methodology will be further demonstrated as PG&E proceeds towards final
selection and application of NOx controls for their generating system.
PG&E NOx EMISSION FORECAST MODEL
The PG&E NOx Emission Forecast Model determines the NOx emission controls required to meet
specified emission limits and their related cost to PG&E. The costs are calculated both in
terms of capital costs and levelized costs. The model also determines changes in the system
heat rate due to the application of NOx controls. The model will allow PG&E to evaluate various
load and fuel use scenarios with different emission limits imposed. The model calculates annual
NOx emissions using boiler-specific information on operating hours and the loading, combined
with information on boiler specific NOx-versus-load and heal rate-versus-load curves. The model
has the capability to take PG&E's "adjusted load data* (a slightly modified version of the Total
Daily Production, or TDP, files) and produce seasonal, monthly, and annual load profiles and
capacity factors for each boiler. Therefore, although the model calculations are designed
around a system annual operating basis, year-to-year variations in. load demand and fuel use may
be accommodated.
A generic version of the model will be made available to EPR1 member utilities as part of a
software system now being assembled by EPRI.
CURRENT PG&E ACTIVITIES
PG&E is continuing to develop information on NOx control technologies that might be applicable
to our power plants. We are conducting studies to evaluate NOx control cost and feasibility for
more of the boilers in our system. This information will be used as input to the NOx emission
forecast model to help us develop a cost effective system-wide NOx reduction strategy. Our goal
is to identify a range of NOx reduction strategies that are both cost effective and responsive
to the needs of the communities we serve.
We are also planning to conduct a "proof of concept" test using urea injection on a 345 MW
boiler. Urea will be injected into one-third of the flue gas in the convective pass of the
boiler. The test boiler has two division walls that divide the furnace and flue gas paths into
three flow streams. The results of this test will be used to determine if urea injection has
the potential to provide cost effective NOx reductions on our 345 MW boilers.
8-54
-------
PAam OAl « KKTRC COMPANY
com* eeita rawti *want — uiMfti no. i am >
AJKTIOCH, CAUPOtNIA
•MMBBK9I * **«Ul MUtkiMt MMU4 MUH
CAfAfifTv. u niAM i«i neaua ummb wtunutii cxnxr ntiiuii wi i.ot
AUI3WMKJ WOtUM* PBUWIl • . |J« MWM PI.UI.Utl 1TIAM TfirfHHUiTUU, P . . ¦ IJM
UHlAt ItW tl^lMTUKl. * . , , !#!
8-55
-------
FIGURE 2
CHANGES IN FURNACE GEOMETRY FOR RETROFIT COMBUSTION SYSTEMS
CONTRA COSTA UNIT 6
m
i
m
©
t
m
»
h
u w
US •
« <£
ISl 10
¦
e z
¦ o
•" Z
*S
* x
e
ij)
*
«
B
¦
-------
-------
FIGURE 4
SCR ARRANGEMENT FOR CONTRA COSTA UNIT 6
ALTERNATE CASE
8-58
-------
FIGURE 5
PREDICTED NOX EMISSIONS FOR
CONTRA COSTA UNIT 6 - FULL LOAD
NOx. ppm @3% 02 (dry)
Original Existing TSC IFNR SCR
Design (FGR+OFA)
HH Fuel Oil (0.3 N2) Wi Natural Gas
-------
Table 1
MAJOR MODIFICATIONS AND EQUIPMENT ITEMS FOR
NOx CONTROL OPTIONS - CONTRA COSTA UNIT 6
Two Stage Combustion
In-Fumace NO* Reduction
SCR (Base Case)
Fan* and Ductwork:
- Replace FGR fan rotor.
» New PCI outlet ducts and
dampers.
- OFA ducts and dampen.
- PC ductwork/piping and dampers.
• Replace air heater outlet ducts.
• Generally, game items as for TSC.
(Detailed design not performed)
- New FD fans, drives, and
foundations.
- Increased stiffening on flues and
ducts.
- Structural supports, platework,
expansion joints, dampers, turning
vanes, etc for installation of SCR,
relocated air heater, and new FD
fans.
Ifeilsr NtaJlfiWisnr
- Partial replacement of sec.
superheater (SSH) tubes.
- Replace SSH attemperaior to
increase capacity.
- Compartmentalized windbox.
- Major extension of furnace height
(furnace bottom extended
downward) requires
modifications/replacement of
furnace wall panels, structural
supports, and water circuitry.
* Compartmentalized windbox.
- Reposition air heater toward stack
(install SCR reactor in existing air
heater location).
- Modify furnace convection pass
buckstay/support systems.
(jombustiop iq«ip!T>?nf;
- 24 PG-DRB burners with
accessories (installed in eristing
furnace openings}.
- 12 Dual Register OFA ports
(installed In existing furnace
openings).
- Modified fuel supply valving.
- Generally similar equipment
items as for TSC except for
add ItionsI re w of burners (i .e., 36
PG-DRB burners required).
- None
Other,
- Bailer control system modifications
(minimal)
• Boiler control system
modifications and
instrumentation expected to be
more extensive than for TSC.
• SCR reactor vessel, incl catalyst.
- Ammonia storage, vaporization, and
injection systems.
- SCR controls and instrumentation.
- Modified underground utilities (due
to interference^.
8-60
-------
Table 2
LOW-NOx COMBUSTION FEASIBILITY STUDY RESULTS
Contra Costa Unit 6
Moss Landing 6 & 7
Pittsburgh 5 & 6
TSC
IFNR
P-IFNR
D-IFNR
DM-IFNR
TSC
P-IFNR
TSC
IFNR
Test Case Description:
PG-DRB Burners
24
36
36
36
36
36
36
24
36
Burner Arrangement
2Hx6W
Opposed
3Hx6W
Opposed
3H x 6W
Opposed
3H x 6W
Opposed
3H x 6W
Opposed
3Hx6W
Opposed
3H x 6W
Opposed
2H x6W
Opposed
3Hx6W
Opposed
Overfire Air Ports
12
12
12
12
12
12
12
12
12
FGR Rate
20%
20%
20%
20%
20%
18%
18%
18%
18%
Predicted NOx Reduction
at Full Load:
Fuel Oil (0.3%N)
31%
52%
45%
58%
30%
21%
54%
40%
47%
Natural Gas
61%
73%
70%
75%
62%
50%
69%
58%
66%
Increased Furnace Height
No
Yes
Yes
No
Yes
No
Yes
No
No
Other Considerations
(1)
(2)
(2)
(3)
(3)
Preliminary Feasibility
Yes
No
No
No
No
Yes
No
Yes
Yes
(1) Load restricted to 55-60% of MCR.
(2) Existing 3-nozzle cell burners require extensive changes in burner arrangement and windbox to accommodate PG-DRB
retrofit. Physical interferences from steam piping and mixing devices along furnace wall complicate retrofit.
(3) Coal-design furnace provides sufficient residence time for combustion staging within existing furnace cavity.
-------
Table 3
COSTS OF TSC AND SCR FOR APPLICATION TO
CONTRA COSTA UNIT 6
Two Stage SCR SCR
Combustion (Base Case) (Alternate)
Capital Cost (S/kW)
Material & Engineering
17.5
30.8
33.7
Installation
12.7
15.6
19.3
Other (1)
m
25.9
TOTAL CAPITAL REQUIREMENT
45.7
723
82.5
Level ized Cost (mills/kWh)
Fixed and Variable O&M
0.8
1.1
1.2
Consumables <2)
0.0
1.3
1.3
Carrying Charges (Capital)
2.9
4.5
52
TOTAL LEVELIZED COST
3.7
6.9
7.7
Notes:
(1) Includes contingencies, general facilities, taxes, and pre-production costs.
(2) Includes replacement catalyst and ammonia for SCR.
8-62
-------
ANALYSIS OF MINIMUM COST CONTROL APPROACH
TO ACHIEVE VARYING LEVELS OF NOx EMISSION REDUCTION
FROM THE CONSOLIDATED EDISON CO. OF NY POWER GENERATION SYSTEM
D. Mormile
J. Pirkey
Consolidated Edison Co. of New York
New York, NY
N. Bayard de Volo
L, Larsen
B. Piper
M. Hooper
Energy Technology Consultants, Inc.
Irvine, CA
y
8-63
\
-------
-------
Analysis of Minimum Cost Control Approach
to Achieve Varying Levels of NOx Emission Reduction
from the Consolidated Edison Co. of NY Power Generation System
D, Mormile
J. Pirkey
Consolidated Edison Co. of New york
New York, NY
N. Bayard de Volo
L. Larsen
B. Piper
M. Hooper
Energy Technology Consultants, Inc.
Irvine, CA
ABSTRACT
Con Edison of New York operates a system of gas and oil fired boilers for
power generation and district heating which is located in New York City. Although
current NOx emissions from these boilers are in the range of NSPS limits, a further
reduction could be mandated as a consequence of a future NOx regulatory strategy to
achieve compliance with ambient ozone standards. In recognition of this
possibility, Con Edison initiated a program in 1989 to determine how NOx emissions
might be best controlled and at what cost.
Tests have been conducted on each unit type/fuel combination to determine
current NOx emission levels and the reduction potential achievable by employing
operationally implemented off-stoichiometric firing. A PC based model of the system
has been formulated which can predict system NOx emissions integrated over any
potential compliance period for the application of any unit specific combination of
NOx control technologies. The model considers capital and operating costs on a unit
specific, control concept design basis and calculates system cost levelized over a
specified period for each case considered.
This paper presents a review of the program status and a preliminary summary
of results obtained to date. The program is not yet completed.
8-65
-------
INTRODUCTION
In 1989, The Consolidated Edison Company of New York, Office of Environmental
Affairs, initiated a program to define cost-effective strategies to contend with
possible future NOx emission regulations. The purpose of the program was threefold:
1) To assess the cost and effectiveness of all viable NOx control
technologies as applied to the Con Edison fossil fuel boilers and to
define the optimum means of achieving any specified level of NOx
emissions.
2)' To provide information to assess the economic and emissions impacts of
proposed regulation levels and forms so that Con Edison might formulate
a corporate position relative to rulemaking activities of regulatory
agencies.
3) To identify areas to which Con Edison might best direct internal R&D
funding to nurture the development of NOx control technologies to serve
its future needs.
The program, still in progress, comprises four major tasks: 1) testing of
representative boilers to characterize both the baseline NOx emissions throughout
the Con Edison system and the emissions reductions possible with O.S. firing
techniques; 2) compilation and assessment of information on the control
effectiveness and application costs of all pertinent NOx control technologies;
3) formulation of a PC-based computer model of the Con Edison fossil fuel boiler
system to permit assessment of baseline NOx emissions and the cost and NOx emissions
resulting from application of selected control technologies; and 4) analysis of
optimum NOx control strategies to achieve compliance with a variety of potential
emission requirements, using the results from the previous three tasks.
The testing portion of the program consists of the measurement of NOx
emissions from a selected set of boilers representing the total Con Edison
population of boilers. Each boiler was tested with normal firing procedures over
its firing range (load) and for each fuel (natural gas or residual oil) commonly
burned. The baseline NOx emissions were characterized vs excess 0? level at each
load level tested. Measurement of 02, CO and NOx was made at multiple locations in
the boiler exit ducts using a mobile flue gas analysis laboratory. On some boilers
tests were also performed to define the potential NOx reduction achievable by firing
in an off-stoichiometric (O.S.) mode, consisting of shutting off fuel to selected
burners while leaving their air registers open, thus stratifying the air/fuel mix in
the combustion zone. In all, 21 boilers have been tested, out of a total population
of 31 electric generation and 33 steam sendout boilers.
The compilation and assessment of NOx control technology effectiveness and
costs was accomplished with a combination of public and proprietary NOx emissions
test data for a wide range of control technologies. To the extent possible, the
available data were adjusted to reflect the most likely control effectiveness and
cost of implementation which would occur upon application to specific Con Edison
boilers.
A PC-based, spreadsheet model was composed to calculate the NOx emissions,
electric and steam production, and fuel consumption of each Con Edison boiler for
any specified time period, load schedule, fuel mix and NOx control technology
implementation. A discussion of some features of the program is contained below.
8-66
-------
Some preliminary analyses of optimum NOx control strategies have been
completed using the computer model. The initial results are discussed in the paper.
The purpose of this paper is to present these preliminary results, which may be of
some interest to other utility and regulatory investigators. The authors emphasize
that the analysis is incomplete at present. Additional boiler testing is planned,
refinements are being incorporated into the computer model and the assessment of NOx
control technologies continues to be updated.
CURRENT OPERATION
Con Edison operates a system of 64 fossil-fuel-fired steam boilers located
within the city of New York, ranging in size from 150,000 lb/hr to over 8 million
Ib/hr steam capacity. Eleven large boilers generate only electricity (173 to 972
MWe each) with condensing turbines. An additional twenty boilers produce,
electricity and also send out live, extraction or exhaust steam for commercial
heating use. Thirty-three smaller boilers produce steam only for send-out. The 64
boilers are distributed among thirteen separate plants in the boroughs of Staten
Island, Brooklyn, Queens and Manhattan. Table 1 presents a summary description of
the boilers operated by Con Edison and included in the current analysis. Additional
electric generating plants, partially owned by Con Edison but operated by others,
are not included in this study. Similarly, combustion gas turbines are excluded at
present.
As shown in Table 1, some units burn either gas or oil fuel (or a combination
of both) while the remainder burn exclusively natural gas (60th St) or residual oil
(all of the rest). Boilers with dual-fuel capability are generally restricted to
oil fuel in the months of December through February due to curtailment of gas
supplies. When both fuels are available, current fuel prices generally favor gas
firing. In recent years the relative system-wide fuel mix has been from around 50
to 75% oil on an annual basis (BTU value).
The electric generating boilers represent a spectrum of tangential, face and
opposed fired boilers manufactured by CE, B&W and FW. Most of these were originally
designed for coal firing and therefore represent relatively large furnace volumes
(and consequently, low NOx emissions) for the unit firing capacity. This
characteristic is discussed further below.
The total capacity of Con Edison-operated fossil-fuel electric generation is
approximately 6,700 MWe of which about 5,100 is steam-electric located in New York
City. The remainder comprises gas turbines and shares of steam-electric units
located elsewhere. Figure 1 depicts representative monthly generation and fuel
usage projected for the early 1990's. From the figure it is clear that two annual
peak generation periods occur, one in December/January and the other in July/August.
In 1990 the peak generation days were on January 8 and July 5. As can be seen in
Figure 1 the total actual generation by fossil-fuel steam units is around 40% of the
maximum possible over,the year.
From Figure 1 the seasonal shift in fuel mix is clearly seen, with oil
predominating from October through April and gas fuel sharing the load throughout
the summer. This seasonal fuel-mix characteristic has significant implications on
NOx emissions and control strategies.
As mentioned above, the Con Edison boilers were, for the most part, designed
for coal firing and therefore exhibit low NOx emission characteristics. Table 2
shows a comparison between similar classes of boilers (size, design) at Con Edison
and at other utilities with typical gas/oil-design boilers. All data shown are from
test data acquired within the past several years. The Con, Edison baseline emissions
8-67
-------
measurements have not been completed. It is clear that the Con Edison boilers have
considerably lower baseline NOx emissions with gas fuel than comparable boilers
elsewhere. With oil fuel the difference is not as clear, although the Con Edison
emissions are among the lower emission levels. The principal implication of the low
initial (baseline) NOx emission levels at Con Edison is that the percentage
reduction in NOx emissions achievable with most NOx control technologies depends to
some degree on the initial NOx level prior to application of the technology.
The baseline NOx emissions shown in Table 2 and used for analysis of potential
NOx reduction capability are derived from short-term, carefully controlled
engineering tests performed with steady-state boiler operation. While these data
are useful for defining the effects of various controllable operating parameters on
NOx emissions, it should be understood that continuous, day-to-day operation of a
unit does not necessarily produce, on average, the same NOx emissions as a short-
term engineering test, even at nominally the same firing conditions. Thus, there is
a degree of uncertainty as to the actual NOx emission to be expected over a longer
time span.
Under Automatic Generation Control (AGC) the load on a unit (firing rate) is
controlled by a central dispatch computer and can cycle continuously over its normal
load range. In this transient mode of operation it is not always possible to
maintain the "optimum" specified firing conditions (excess 02, burner pattern, etc)
vs. load. In order to avoid unsafe conditions as the unit is automatically
controlled over the load range, operators will tend to set a safety margin of excess
02 above the ideal, steady state point at a given load level, and thus the NOx
emission will be increased somewhat. Also, over a longer period of time; boiler
furnace walls may become dirty between soot-blowing periods, burners may deteriorate
siightly and other uncontrollable factors may tend to increase NOx emissions over
the values defined in short-term testing. Figure 3 illustrates the considerable
variability of baseline NOx emissions with AGC control in comparison to the baseline
NOx emissions derived from short-term testing. Thus, in order to maintain NOx
emissions consistently below a specified regulatory limit, the operator would have
to either reduce the average NOx emission well below the limit (so that the peak NOx
emission was still below the limit) or reduce the variability of the NOx emissions
about the average value by maintaining tighter control of excess 02, boiler wall
cleanliness, etc,
NOx CONTROL TECHNOLOGIES
The technologies selected for inclusion in the study are those which have been
historically employed on an operational basis for NOx control on gas/oil fired
utility boilers and certain other developing technologies close to
commercialization. Descriptions of these technologies have been well documented in
the published literature and the discussion presented here is confined to pertinent
information relating to NOx control capabilities. Considerable uncertainty exists
as to the control capabilities of most of the candidate control options. The NOx
reduction algorithms employed in the preliminary analysis are current best
estimates. An effort is being conducted as part of the program to refine these
estimates for final analysis.
OfF-STOICHIOMETRIC FIRING fO.S.l
This control option has been effectively employed by a number of utilities to
achieve significant NOx reductions on gas/oil fired boilers. Figure 2 (abstracted
from Ref. 1) shows the results achieved by one utility (Southern California Edison
Co.) employing O.S. firing on a range of boilers firing natural gas fuel. These
results are representative of those demonstrated in other utility systems which
8-68
-------
generally indicate a NOx reduction dependency on initial, uncontrolled NOx level.
The shaded area in the figure depicts the range of NOx reductions demonstrated in
the current Con Edison test program and confirms the dependency of control
capability on initial NOx level. Similar trends have been demonstrated for oil fuel
firing. The Con Edison O.S. test data generally fall in the range of 30% NOx
reduction, which is substantially less than the control capability normally
associated with this technology but is explained by the low baseline NOx levels.
The steady state, short term data acquired in the test program for O.S. firing
have been used in the analysis for the performance of this control option. This
data may substantially overstate the magnitude of NOx reduction that could actually
be achieved during normal AGC operation. Figure 3 shows a comparison between steady
state and AGC test data for one of Con Edison's units in uncontrolled and O.S.
operating modes. The AGC data shows considerable scatter and does not reflect any
NOx reduction benefit for O.S. firing in comparison in the steady state data.
Similar data scatter has been observed for baseline operation. The data scatter is
due primarily to variations in operating excess air and to boiler cleanliness
effects resulting from switching back and forth between natural gas and fuel oil
firing. It may be possible to narrow the data scatter band by improving operating
procedures and air flow control, but differences between steady state and AGC NOx
emissions cannot be eliminated. The implication of these results is that both
baseline and O.S. operating mode NOx emissions should be predicted on the basis of
AGC operation, which is the intent for the final analysis.
LOW NOx BURNERS fLNBY
There are very few installations of LNB's on gas/oil fired utility boilers and
there is little published data reporting NOx control performance. Ref. 2 provides
preliminary data for installation of one such burner design on two gas/oil fired
utility boilers. The test results demonstrated an improvement over that which had
been achieved for O.S. firing in the range of 10-20%. On the basis of these
results, the analysis assumes an NOx control performance for the LNB control
technology of 10% greater NOx reduction than that achieved in the O.S. testing of
the Con Edison units.
UREA INJECTION (UREA)
UREA injection is a developing technology which is likely to have widespread
future application in utility systems for NOx control Versions of this technology
are currently being demonstrated on several boilers in the Southern California
Edison system. NOx reduction data acquired in these programs have been employed for
the present study to formulate a NOx control algorithm. The data have been
extrapolated to lower initial NOx levels than tested by kinetic analysis. The model
thus formulated was used in the analysis and is shown in Fig. 4.
The EXXON Thermal DeNOx technology which is similar to UREA injection except
that the reagent is ammonia, could be employed as an alternative to UREA injection.
For the purposes of this initial study, the UREA technology has been assumed to be
representative of this general category of NOx control approach.
WINDBOX FLUE GAS RECIRCULATION fWFGR)
WFGR has been employed on both new and existing gas/oil fired boilers for NOx
control. The technology has been demonstrated to be a very effective NOx control
option but little data exists in the published literature pertaining to it's control
performance. Reference 2 reports some data for two retrofit installations in the
Southern California Edison system. This data has been utilized to formulate a NOx
8-69
-------
control model for natural gas and fuel oil firing which is shown in Fig. 5. The
nitrogen content of the fuel oil applying to the test data is 0.3% which is
essentially the same as for the Con Edison fuel.
REBURNIN6
The Con Edison boilers are particularly suitable for the application of the
Reburning technology because of their uncharacteristically large furnaces for
gas/oil fired units. This technology was not considered in the analysis, however,
due to the lack of sufficient data to estimate NOx control performance, particularly
at low initial NOx levels.
SELECTIVE CATALYTIC REDUCTION fSCRI
SCR was assumed to have a NOx reduction capability of 80% for all initial NOx
levels.
SYSTEM NOx MODEL
A PC-based spreadsheet model was written to calculate the NOx emissions and
cost of control for any combination of control technologies for the Con Edison
system, and for each boiler unit individually. The model comprises three
functional areas: data input, calculations and summary.
In the data input area the user enters the conditions defining the specific
case to be evaluated. After the first run, only those data which change from case
to case need to be entered each run. The input data fall into three categories:
general, description of the case, NOx control selection, and unit loading schedules.
The general description data include case number and narrative description of the
case conditions. The NOx selection input consists of completing a matrix table of
NOx control technologies for each unit in the system. The final data input consists
of loading schedules for each unit for both short term (1 hour to many days) and
annual periods. The short-term period is intended to provide the total and average
NOx emissions from each unit over a specified duration (e.g. 8 hours, 1 day, 1 week,
etc). The annual period is used to calculated the NOx emissions, generation, fuel
consumption and variable control costs over a year's time. For each time period the
user inputs the hours of operation of each unit, at each of five (5) load levels and
for each fuel used. The specification of hours of operation at each load level is
important since NOx emissions are variable (usually non-linear) with load, and
therefore the load history must be known in order to calculate integrated NOx
emissions.
Also located in the data input area, but usually not changed by the user, are
tables of NOx reduction effectiveness and generic costs (capital and O&M) for each
control technology. Capital costs are specified in $/KW and variable O&M costs in
terms of S per unit of generation or of tons of NOx removed.
The calculation area of the model begins with tables of baseline NOx
emissions, (Ib/mmBtu) vs load for each unit and each fuel fired. Similar tables of
NOx emissions vs load are provided for O.S. firing conditions.
Controlled NOx emissions (in Ib/mmBtu) are calculated sequentially for each
technology specified in the data input area. Thus, each technology effectiveness
(and resulting NOx output) is dependent upon the output NOx level of the preceding
technology. For example, if both LNB's and FGR are selected for a unit, then the
FGR effectiveness at each load level of the unit will depend upon the LNB output NOx
level at the corresponding load. Of course, each technology not selected has no
effect on the NOx level.
8-70
-------
Following the last application of NOx technology to each unit, the final
outlet NOx level is determined at each load level for each fuel. Based upon the
hours of operation at each load level for each fuel specified in the input tables,
the total short-term and annual NOx emissions (1b/NOx) are integrated for each unit,
along with the total generation (kwh) and thermal input (Btu).
The cost of NOx control is calculated for each unit by summing each cost
element (capital, fixed O&M, variable O&M) for each technology used. The capital
cost for each selected technology is the generic cost (S/kw) times the unit rating
(kw) times a unit-specific multiplier which represents the degree of difficulty of
applying each technology to that unit. Similarly, the variable O&M cost of each
unit is calculated as the sum of each applied technology's variable O&M cost, which
is the product of the generic cost (S/kw or S/ton NOx) times the annual usage (kwh
or tons NOx) times a unit-specific cost multiplier. Fixed annual O&M costs are the
specified generic fixed O&M costs (S/yr) times a cost multiplier for each unit.
Finally, capital costs are levelized by multiplying the total capital cost for each
unit by a recovery factor representing a specified time period (e.g. 20 years) and
a rate of return (e.g. 10%). Similarly, the total annual O&M costs are levelized
according to standard procedures to account for rising O&M costs over the economic
life of the project, essentially in accordance with the EPR1 TAG procedures. The
capital and O&M levelizing factors are entered by the user.
The final function of the spreadsheet model is to compile the emission and
cost results for each unit into a total for the system (including appropriate system
averages, such as Ib/mmBTU NOx emission) and to present the results in a concise
tabular format.
By calculating the-unit specific emissions and costs .(and therefore the system
emissions and costs) for a successive series of varied NOx control applications, the
user can determine the lowest-total-cost combination of controls which will result
in total system emissions meeting any specified level for any specified time-
averaging period.
ANALYSIS RESULTS
The Con Edison System NOx model has been constructed and is fully operational,
but preparation of input information has only been partially completed. Selected
analyses have been performed, however, by utilizing that information which has been
developed and by otherwise employing prior information in ETEC's possession and best
estimates. The results of these analyses are reported herein and although they are
subject to some level of uncertainty in terms of magnitude, derived trends and
observations based on these trends are believed to be generally valid.
Figures 6 and 7 show calculated system NOx emissions for 24 hour periods
coinciding with peak generating days in July and December for baseline operation and
for various NOx control strategies. Each plotted data point corresponds to a
specific control strategy consisting of the application of various combinations of
NOx reduction technologies to each unit in the system. Solid symbols denote that
the indicated control combination has been uniformly applied to all units in the
system while open symbols indicate selective utilization. In this latter case, the
letters "Fg" indicate WFGR applications on only gas/oil fired boilers (excluding oil
only units) and a numeral denotes the limited number of unit applications of the
technology identified by the end letter in the sequence (ie OU{5) denotes O.S. on
all units and UREA on 5 units).
8-71
-------
The results apply to actual unit load duration curves for 1990 but the fuel
mix has been altered to reflect maximum gas burning in July and maximum oil burning
in January (ie, dual fuel units burn either all gas or all oil depending on the
month). This allocation of fuels burned approximates that shown in Fig, 1 which is
based on a PROMOD projection. The indicated NOx emissions for each strategy have
been determined by summing the respective integrations over each unit's load
duration curve of the emission rate applying to the fuel burned and the combination
of control technologies installed on the unit.
The baseline (uncontrolled) NOx emissions indicated in the figures have been
determined on the basis of the steady state test data acquired to date and estimates
for as yet untested units. The levels shown understate actual NOx emissions since
they do not reflect the effects of AGC operation, dual fuel firing and boiler
cleanliness in switching between fuels. Each of these factors would tend to
increase unit baseline, and hence system, NOx emissions. The reduced emission
levels shown to be achievable by the application of the various strategies are also
overstated in this regard since they are based on the baseline emissions. Aside
from this factor, the achievable reductions have been determined employing
potentially overly optimistic estimates of the NOx control capabilities of the
individual control options, as pointed out previously. As a consequence of the
above factors, the results as shown are probably too low and the rate of decline.in
achievable emissions with increasing control cost is too steep.
The analysis results shown in Figure 6 and 7 are primarily of interest to Con
Edison. It is possible, however, to draw certain observations based on the
indicated trends that may be of more general interest to other gas/oil utilities and
these are discussed below.
OPTIMUM NOx CONTROL STRATEGIES
The purpose of the analysis was to determine the minimum control cost to
achieve varying levels of NOx emission reduction. This cost would be represented by
a curve defining the locus of minimum control cost strategies for achieving
successively reduced levels of NOx emission. Defining such a curve by employing the
model is an iterative procedure in which various strategies are analyzed and the
calculated NOx emission levels and costs are compared. This procedure was followed
in the present case and the optimum strategies determined are those shown in Figures
6 and 7 as being the lowest points at any cost level.
The strategies that were analyzed only broadly define the optimum curve since
intermediate steps have not yet been evaluated. For instance, the locus of
strategies between 0.S. on all units and O.S. plus UREA on all units would be
defined by the intermediate steps of sequentially adding UREA combined with O.S. to
successive units. Two such intermediate steps are shown in Figures 6 and 7 for
0U(3) and 0U(5).
The analysis results indicate that the optimum strategy to achieve a specific
level of NOx emissions would consist of maximizing the system wide utilization of
the lowest cost technologies first before employing on any unit the next most costly
technology. For instance, it would always be more cost effective to employ UREA on
additional units compared to employing the next most costly technology, which in
this case would be WFGR, on any additional unit. This analysis result is summarized
below:
8-72
-------
Strategy for Increasing
Levels of NOx Control
Order of Control Option Application
All units Successive units
I
II
III
IV
O.S. + UREA
O.S. + UREA + WFGR
O.S.+ UREA + WFGR + SCR
O.S.
LN8 could be employed as a substitute technology for O.S., providing an added
10% increment in NOx reduction. However, the combination of O.S. plus UREA would
always be more cost effective than the utilization of LNB's. WFGR would be employed
in an optimum strategy only on gas/oil fired boilers since it's control capability
for reduced initial NOx levels is too low for cost effective utilization on oil-only
boilers.
The above ranking order for utilization of control technologies would apply
only to situations in which an emission regulation were expressed as a LB/day
emission limit averaged over a system. Alternative forms of emission limits would
likely result in a different ordering of technologies for optimum employment.
DIMINISHING RETURN
Figures 6 and 7 graphically illustrate the diminishing return of increasing
expenditure to reduce NOx emission from the Con Edison system. This observation is
quantified in the table below which applies to the optimum locus of strategies in
Figure 6.
The table values show for instance that an 80% emission reduction would
require a factor of three greater expenditure than a 70% reduction. This trend is
actually understated since the achievable emission reductions shown in the figures
are optimistic as explained previously.
SEASONAL INFLUENCE ON COST OF CONTROL
Figure 8 replots the optimum strategies defined in Figures 6 and 7 in terms of
daily emissions averaged on a LB/MMBTU basis. Peak day NOx emissions are shown to
be higher in January than in July. The reason for this is attributable to higher
baseline NOx emissions in January due to substantially increased oil firing, to
differences in unit loading schedules and to generally reduced NOx control
effectiveness for some of the technologies for oil firing.
The difference in emission rates for the two seasons is particularly
significant if a regulation were passed of a form limiting emissions on a LB/MMBTU
basis. The inset table in the figure shows that the cost of compliance in this
instance would be at least a factor of two greater for January in comparison to
July. The purpose of such a regulation, however, would be to reduce ambient ozone
concentrations, which tend to be most pronounced during the summer months because of
System NOx Emission
Reduction, %
Cost of control
Mill/KWH
50
70
75
80
.4
1.4
1.8
4.5
8-73
-------
meteorological conditions favoring their formation. Therefore, a regulation of this
form would result in an additional expenditure that would serve no environmental
purpose. In such a situation, the emission limit should be formulated to cost
effectively achieve it's intended purpose.
CONCLUSIONS
1) A system NOx emissions model of the type described can be a useful tool in
assessing the implications of a potential regulation in advance of it's
promulgation for preparing a utility for the regulatory process.
2) The Con Edison boilers have low uncontrolled baseline NOx emissions because of
their design and low capacity factors. In such instances, it is more difficult
to reduce NOx emissions because of the reduced effectiveness of NOx control
technologies for low initial NOx levels.
3) The process of establishing NOx emission regulations should recognize that
relatively small differences in control limits can have a dramatic effect on the
required cost of control.
4) The form of an emission regulation can inadvertently result in the expenditure
of unnecessary control costs if it does not specifically address it's intended
purpose.
REFERENCES
1) Bagwell, F.A., et.al., "Utility Boiler Operating Modes for Reduced Nitric
Oxide Emissions", JAPCA, November, 1971
2) Bayard de Volo, N., et.al., "NOx Reduction and Operational Performance of
Two Ful1-Scale Utility Gas/Oil Burner Retrofit Installations", 1991 Joint
Symposium of Stationary Combustion NOx Control, Washington, D.C.,
March 25-28, 1991
8-74
-------
TABLE I
CON EDISON GENERATING UNITS
Planl
Planl
Unit
Capacity
Mig
Firing
No of Burn.
Fuel
Function
MW
Config.
O-OII
G-Gas
POWER
ARTHUR KILL
20
345
B&W
Face
32
O
30
440
CE
Corner
40
O
ASTORIA
10
187
B&W
Face
22
G.O
20
173
B&W
Face
22
G,0
30
365
B&W
Face
32
G.O
40 '
375
CE
Corner
32
G,0
50
375
CE
Corner
32
G,0
RAVENSWOOD
10
95
CE
Corner
32
G,0
20
395
CE
Corner
32
G,0
30
900
CE
Corner
64
O
POWER PLUS
EAST RIVER
50
148
• B&W
Opposing
12
G,0
STEAM
60
148
B&W
Opposing
12
G.O
SENDOUT
(Pwr Only)
70
180
FW
Face
18
G,0
59TH St
110
72
B&W
Face
8
O
111
43
B&W
Face
5
O
112
43
B&W
Face
5
O
113
43
B&W
Face
5
O
114
79
CE
Corner
8
O
115
79
CE
Corner
8
O
WATERSIDE
41
71
CE
Corner
8
G,0
42
71
CE
Corner
8
G,0
51
71
CE
Corner
8
G.O
52
71
CE
Corner
8
G,0
61
97
CE
Corner
8
G,0
62
97
CE'
Corner
8
G,0
80
160
CE
Corner
16
6.0
90
160
CE
Corner
16
G.O
74TH ST.
120
64
CE
Corner
8
O
121
64
CE
Corner
8
O
122
64
CE
Corner
8
O
HUDSON AVE.
71,72
CE
Face
8
O
81,82
CE
Face
8
O
100
187
B&W
Face
16
O
STEAM
MLB/HR
SENDOUT
RAVENSWOOD
4 units
275 EA
B&W
Face
6
O
E.RIVER SO,
10 units
150 EA
FW
Package
2
O
59TH ST.
3 units
150 EA
FW
Package
2
O
74TH ST.
, 6 units
150 EA
FW
Package
2
O
GOTH ST.
6 units
150 EA
FW
Package
2
G
8-75
-------
TABLE II
CON EDISON BOILER CURRENT NOx EMISSIONS
AND COMPARISON WITH OTHER UTILITY BOILERS
FULL LOAD, UNCONTROLLED NOx EMISSIONS,
PPM
FIRING
SIZE
GAS
OIL
CONFIGURATION
MW
OTHER
CON ED
OTHER
CON ED
UTILITY
UTILITY
Single Face Fired
175
405
175
750
450
180
300
250
187
175
300
215
520
250
230
337
370
345
_
250
365
225
325
Opposed Fired
148
275
250
225
550
...
230
360
250
350
890
425
480
700
320
750
1200
750
T Fired
320
335
225
395
150
175
440
—
200
900/2
—
275
8-76
-------
2,500
00
1
g 2,000
CD
^ 1,500
tr
HI
LLl
o
^ 1,000
CO
CO
o
LLl
0)
500
Maximum Fossil Fuel
Generating Capability for one month
4880 GWH
50 million
40
30
20
10
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
MONTH
FIG. 1 Projected Con Edison Fossil Fuel Generation
and Fuel Consumption for the early 1990's
OQ
E
E
z
o
i—
Q_
Z)
CO
O
O
_i
LLl
=>
LL
-------
1,100
1,000
' 900
800
700
600
500
400
300
200
100
0
0
U
6.
G. 2
Normal
Operation
High
Excess
Air
Furnace
Operation
Fuel-Rich Burner Operation
SCE Units
175MW, Face Fired
~ 320 MW, Corner Fired
• 220 MW, Face Fired .
A 480 MW, Front & Back Fired
Con Edison Units
I; Typical Units
0.8
0.9
125
111
4,6
2.3
1,4 Burner Equivalence Ratio
71 Burner % Air
Furnace % Excess 0„
Off-stoichiometric Combustion for Natural Gas Firing
8-78
-------
400
300
200
100
AGC OPERATION
UNCONTROLLED
o os-
STEADY STATE TEST DATA
UNCONTROLLED
OS.
o
o
A
° ° Q.
A /\ A M. A A/vl
^ ^°g!F
"A " ~ Sr O
A
O
50
100
MW
150
FIG. 3 NOx EMISSIONS BAND DURING AGC OPERATION
ON GAS FUEL FOR ASTORIA UNIT # 10
200
-------
240
260
SO 120 160 200
Initial NOx, ppm
Figure 4. NOx Control Effectiveness of UREA Injection versus Initial NOx Level
100
90
80
70
c
o
60
"fj
3
T3
50
0
DC
40
X
O
z
30
20
10
0
Gas Fuel
Oil Fuel
240
280
0 40 80 120 160 200
Initial NOx, ppm
FIG. 5 NOx Control Effectiveness of 20% WFGR versus Initial NOx Level for
Gas and Oil Fuels
8-80
-------
200
00
I
oo
m
_i
o
o
o
CO
o
CO
CO
2
LU
^OFfl
AUJP) a**
u
UREA INJECTION
s
SELECTIVE CATALYTIC REDUCTION
alf
Aalu A aOOS<2>
A^.orj A105®
A . LFU
LFflU A
1 1 1 1
IFUS<5)
1
AS
I i i i
^LFUS
1
2 3 4 5 6 7 8 9
LEVELIZED COST OF CONTROL, Mill/kwh
10
11
12
FIG. 6 Optimum System NOx Control Strategy to Achieve Varying Levels
of Emission Reduction for Peak Generating Day in July
-------
250
JANUARY
LEGEND
CD
_l
A^
TECHNOLOGIES APPUED TO ALL APPLICABLE UNrTS
O
o
200
re
^LRIS(5)
TECHNOLOGIES APPUED TO SELECTED UNrTS
o
CO
B
LFU TO ALL APPLICABLE UNrTS, SCR TO (5) UNrTS
BASEUNE, NO CONTROL
o
CO
150
~
0
D>
2
O
L
OFF-STOICHIOMETRIC FIRING
LOW-NOx BURNERS
CO
. L F
A
F
g
FLUE GAS RECIRCULATION
OAS FIRED UNITS ONLY
LU
>
^OU<3) ^OF
A«JP)
u
UREA INJECTION
<
Q
100
4if
- ^ LU(5) A
s
SELECTIVE CATALYTIC REDUCTION
<
LU
Q_
DC
I
50
A°
A OUSG)
^ou A
^ ^LUSP)
A alfu
LFgU m
^LFUSp)
A
CM
LFUS
~
0 1 2 3 4 5 6 7 8 9 10 11 12
LEVELIZED COST OF CONTROL, Mill/kwh
FIG. 7
Optimum System NOx Control Strategy to Achieve Varying Levels
of Emission Reduction for Peak Generating Day in January
-------
00
1
CX>
CO
CD
E
E
m
_i
CO"
O
= 0.20
CO
CO
LU
Q
*
<
LU
Q_
CC
X
CM
0.35
POSSIBLE
% ADDmONAL COST OF CONTROL
0.30
-
EMISSION
LIMIT
LB/MMBTU
FOR JANUARY COMPLIANCE
IN COMPARISON TO JULY
A 0.13
100
0.25
B 0.08
x 120
JULY, 1990
JANUARY, 1990
2 4 6
LEVELIZED COST OF CONTROL, Mill/kwh
FIG. 8 Added Cost of NOx Control to Comply With LB/MMBTU
Emission Limit in January in comparison to July
-------
-------
REDUCED NOx, PARTICULATE, AND OPACITY ON THE
KAHE UNIT 6 LOW-NOx BURNER SYSTEM
Stephen E. Kerho
Dan V. Giovanni
ELECTRIC POWER TECHNOLOGIES, INC.
Menlo Park, California
J. L, B. Yee
HAWAIIAN ELECTRIC COMPANY, INC.
Honolulu, Hawaii
David Eskinazi
ELECTRIC POWER RESEARCH INSTITUTE
Palo Alto, California
8-85 .
-------
-------
ABSTRACT
Hawaiian Electric Company (HECO) completed major combustion system
modifications in mid-1988 on Kahe Unit 6, a Babcock & Wilcox (B&W) oil-fired unit
rated at 146 MW, The modifications were undertaken to reduce emissions of NOx and
particulate matter, and to restore operational flexibility that had been restricted with
burner-out-of-service operation previously used for NOx control. Modifications
included installation of the B&W PG-DRB burners, front and rear wall overfire air
(OFA) ports, extensive ductwork for the OFA and flue gas recirculation (FGR) flows,
and upgrading of the automatic burner control system. This installation represented
the first application of this type of low-NOx firing system to a utility boiler in the
United States.
As reported in 1989, the NOx reduction goal of emissions below 0.23 lb/MBtu was
achieved and particulate emissions were controlled to below 0.1 lb/MBtu. However
opacity levels increased from pre-retrofit levels of approximately 6% to between 15-
20%. In an attempt to reduce opacity levels and still comply with NOx emission
limits, HECO and the Electric Power Research Institute jointly sponsored a follow-on
Phase 2 performance improvement program conducted by Electric Power
Technologies, Inc to evaluate the potential of new atomizer designs to reduce NOx,
particulate, and opacity. The program demonstrated significantly jreduced opacity and
particulate levels while maintaining NOx emissions below 0.23 lb/MBtu even though
the levels of OFA and FGR were reduced.
8-87
-------
INTRODUCTION
In July 1987, the Hawaiian Electric Company (HECO) contracted with the Babcock &
Wilcox Company (B&W) to retrofit a low NOx combustion system on their 146 MW
(grossT oil-fired Kahe Unit 6. The unit is front wall-fired and bums oil with a
maximum sulfur content of 0.5%. Up to this time, the unit had been operating with
flue gas recirculation (FGR) to the combustion air and burners-out-of-service (BOOS)
in an attempt to satisfy the operating permit requirement for maximum NOx
emissions of 0.23 Ib/MBtu (180 ppm, dry, 3% O2). Typical emissions using these
controls were 0,28 Ib/MBtu NOx (219 ppm) and 0,06 - 0.08 lb/MBtu particulate matter
(PM). Normal opacity levels were in the 4-6% range, which is below the visible
threshold.
The principal objective of the retrofit was to reduce NOx emissions to below the
regulatory requirement while minimizing particulate matter (PM) emissions.
Additionally it was intended that the retrofit technology would allow a return to all-
bumers-in-service operation, thereby improving the operating flexibility of the unit
which had been impaired with BOOS operation. Specifically, a higher turndown was
expected from improved flame stability at low loads (the lowest load for dispatch was
95 MW with BOOS), and a higher reliability in achieving full load was expected with
the ability to accommodate burner maintenance outages without load reduction. The
project was the first installation in the United States of the integrated application of
low-NOx burners, FGR to both the combustion air and directly to the burners, and a
state-of-the-art front and rear wall overfire air (OFA) design to a heavy oil-fired utility
boiler. The combustion system, designated "PG-DRB", is licensed by B&W from
Babcock-Hitachi (BHK) who commercialized the technology in Japan.
The retrofit was successful in meeting the NOx requirement of the operating permit
and in providing the desired improved operating flexibility. However, operating
problems such as undesirable opacity levels led to a follow-on Phase 2 program of
combustion optimization work and equipment modifidation. This paper presents the
results of the follow-on program which was conducted in 1990.
OVERVIEW OF 1987 NOx SYSTEM RETROFIT
Kahe 6 is a radiant reheat type steam-electric unit manufactured by B&W. An
elevation view is presented in Figure 1. For NOx control, the boiler was originally
equipped with nine B&W dual register burners arranged in a 3 X 3 array on one wall,
and flue gas recirculation to the windbox which permitted up to 20% of the flue gas to
8-88
-------
be mixed with combustion air prior to the burners. The retrofit PG-DRB system
consisted of the following elements:
1. PG-DRB burners
2. Dual fluid (steam/oil) atomizers
3. Utilization of existing FGR to the windbox combustion air
4. Primary gas (PG) system which supplies FGR directly to the burners
unmixed with the combustion air
5. Overfire air system
6. Upgraded control system
The PG-DRB burner, shown in Figure 2, consists of an oil atomizer/impeller located
axially in the primary (core) air zone of the burner. The core air is introduced into the
center zone through slots located at the back of the burner. Core air flow is limited to a
maximum of approximately 10% of the total air flow. The flow to this region can be
controlled with a small sliding disk. The core zone is surrounded by the PG zone,
which is enclosed by the inner and outer air zones. Pure gas recirculation is fed
through a perforated plate located at the entrance to the PG zone annulus which helps
to distribute the flow around the periphery of the zone. A butterfly-type valve
provides controllability of the PG flow to individual burners. Air to the inner and
outer air zones is controlled by a single sliding disk. An impact-suction pitot tube grid
is installed prior to the inner and outer air zones to allow measurement of the airflow
in these zones. The pitot grid consists of a manifold which encompasses the burner
with six finger-like extensions into the total air flow zones. These measurements,
together with air slide position, provide the capability of controlling air flow to the
individual burners. The inner air zone contains gear driven spin vanes, while the
outer zone has fixed spin vanes followed by gear driven spin vanes. The manually
operated gear driven vanes provide the ability to vary swirl characteristics and thus
the resulting flame shape of the burner.
The OFA system was designed to divert up to 30% of the total combustion air to six
OFA ports located on the front and rear boiler walls (three ports per wall),
approximately 10 feet above the top burner elevation. Each OFA port is equipped with
damper assemblies and air spin vanes to allow independent control of air quantity,
velocity, and furnace penetration. A schematic showing the port design is provided in
Figure 3. Like the burners, the OFA ports were equipped with flow monitors, allowing
on-line measurement of separate flows through the spin annulus and central core of
each overfire air port. Flow modeling tests using a scale model of the windbox and
furnace were used by B&W to obtain air flow distribution information for the windbox
8-89
-------
and OFA system. The model results were used to establish placement and sizing of the
OFA ports for optimum mixing. The modeling results were the basis for the decision
to use six ports (instead of three) and the recommendation for a nominal 70:30 rear-to-
front wall distribution of overfire air.
Summary of Retrofit Low NOx System Performance Evaluation
The results of the program were presented in detail at the 1989 Symposium (Reference
1) and are summarized below. The retrofit realized its principal goal to reduce NOx
emissions to below 0.23 lb/MBtu with all burners in service. At 145 MWg, NOx and
PM emissions levels of 0.21 and 0.07 lb/MBtu respectively were achieved with a stack
opacity of 15%. The fuel nitrogen content was approximately 0.3% (wt.). The test was
performed using 10% FGR (defined as the amount of recirculated flue gas divided by
the sum of the total air and fuel flows) to the windbox and 27% of the total air to the
OFA system (split 70% to the rear ports and 30% to the front ports}. These acceptance
test results typified the best overall emissions performance achieved and required an
extensive test effort during the commissioning of the equipment to control PM
emissions and opacity. Although the opacity levels noted above are within the
regulatory requirement of <20% for a 6 minute average, they are considered
undesirable because a visible plume results. These results represented an over 75%
reduction in NOx from pre-retrofit levels with all burners in service (ABIS) and
without FGR.
During commissioning, a strong inverse relationship between NOx and PM/opadty
was encountered. Initially, when the combustion equipment was tuned to achieve
NOx levels below 0.23 lb/MBtu, the corresponding PM emissions were typically 0.13-
0.15 lb/MBtu and opacity exceeded 20%. The magnitude of this trade-off was
unexpected from previous experience reported by BHK in Japan, where over 10,000
MW of PG-DRB is operational. It appears that this trade-off is a fundamental feature
of the PG-DRB system when fired with heavy oils. Further assessment of the Japanese
experience in the light of these results led to the conclusion that a similar trade-off
exists at Japanese installations, however it is not an issue there because the boilers are
equipped with electrostatic precipitators for particulate and opacity control.
Initial Oil Atomizer Selection
In order to reduce PM emissions, a comprehensive program was implemented by
B&W during commissioning to optimize oil atomization with the PG-DRB burner
system. Improved atomization would result in smaller oil droplets which burn out
8-90
-------
more completely, resulting in reduced PM emissions. During the course of the
program, a number of B&W dual fluid (steam/oil) atomizer designs and atomizer
spray cone angles were evaluated. These included the Y-Jet, Racer, modified Racer
(Racer with increased steam rates), T-Jet, and a developmental I-Jet design. These
atomizer types are characterized by their geometry, steam-to-oil mass flow rates, and
the size of the oil droplets produced. The Racer, Y-Jet, and T-Jet designs were flow
characterized using water and air as the working fluids. Drop size distribution
information was obtained using an Aerometrics Phase Doppler Particle Analyzer. The
conversion of water/air data to oil/steam was done using viscosity, surface tension,
and mass ratio corrections which were obtained from the literature. For oil properties
and operating conditions at Kahe, the nominal Sauter Mean Diameter (SMD) of the oil
droplet size distributions were 400, 320, and 235 microns for the Racer, Y-Jet, and T-Jet
respectively. The T-Jet was judged to provide the best performance and was selected by
B&W for continuous operation. The importance of reducing drop size was
demonstrated by the reductions in PM and opacity achieved from the initial levels: PM
emissions were reduced from 0-13 - 0.15 lb/MBtu to 0.07 Ib/MBtu and opacity levels
from over 20% to 15%.
LONG-TERM OPERATING EXPERIENCE
Operation at Kahe 6 after approximately two years was characterized by a number of
combustion related problems. Although NOx levels were generally below the 0.23
lb/MBtu regulation, opacity levels had increased from the already undesirable levels
to near 20%, which left no operating margin to allow for variabilities in operation or
oil properties. For example, during one 90 day period, approximately 650 instances of
opacity readings (6 minute average) above 20% were recorded.
In an effort to resolve these problems and restore operating margin, HECO and EPRI
Jointly sponsored a follow-on Phase 2 performance improvement program conducted
by Electric Power Technologies, Inc. (EPT). The basis for this program was the belief
that additional potential remained to modify the oil spray and further reduce
emissions.
PERFORMANCE IMPROVEMENT PROGRAM
Focus on Oil Atomizer Design
Optimized atomizer designs which produce mean drop sizes below 175 microns have
been developed as part of the Heavy Oil Combustion (HOC) Program funded by the
8-91
-------
Consolidated Edison Company of New York, the Empire State Electric Energy Research
Corporation (ESEERCO), and EPRJ. Included in these designs are dual fluid atomizers
with steam consumptions of 10 -15% (steam/oil). Atomizers providing drop sizes in
this range have been shown (Reference 2) to result in substantial reduction in PM
emissions. The data also indicate that improving atomization and mixing of the fuel
and air through atomizer design results in increased NOx emissions. The magnitude
of the NOx increase can be mitigated however by lowered excess 02 levels and by the
use of increased OF A, both of which should be possible due to improved combustion
and a lowered smoke point. Additional potential for lowering NOx is available
through the control of atomizer design parameters which clan modify the distribution
of oil within the spray. It is important that details of the atomizer design also be
specifically tailored to match the aerodynamic flow patterns of the burner itself which
are controlled by impeller design and spin vane settings. Finally, added potential for
control of droplet size is provided by changing fuel viscosity (firing temperature) and
steam/oil mass ratio.
Performance Improvement Program Scope
The Kahe 6 performance improvement program consisted of the following elements;
Atomizer Design. Flow Characterization, and Fabrication. The aerodynamic flow field
of the PG-DRB burner was approximated using a computerized model which assumes
isothermal and in viscid flow. Using design criteria developed as part of the HOC
project, a number of atomizers were designed to match the burner flow field. Each of
the atomizers were flow tested using water and air as the working fluids. The data
were corrected to fuel oil and steam conditions and included a characterization of the
oil spray in terms of mean droplet size and droplet size distribution measurements
together with pressure/flow characterization over the atomizer flow range to assure
that flow variations between individual atomizers were minimized.
Unit Performance Characterization With Existing Equipment - A short performance
test was conducted to document existing Kahe 6 emissions and operational
performance. The data provided a quantitative measure of the performance
degradation which occurred over the two year period since commissioning.
Measurements of NOx, CO, and 02 were obtained using continuous instrumentation
to analyze flue gas obtained from a matrix of 24 sampling probes installed at the
economizer outlet.
8-92
-------
Unit Inspection - Following the performance test a scheduled unit outage provided an
opportunity to inspect the equipment to assure that burner and OFA spin vane or disk
positions had not changed and to repair any malfunctioning equipment.
Post-Outage Performance Test - Following the outage, the performance test was
repeated to document any changes in emissions performance and to establish a
reference condition for the subsequent atomizer changes, Following these tests, the
new atomizers were installed and performance evaluation tests were conducted. The
evaluation criteria included flame appearance, flow/pressure characteristics, NOx
emissions, and opacity. On the basis of these results, the atomizer providing the best
performance in terms of NO* and opacity was selected for full characterization.
The full characterization tests included NOx and opacity versus O2, OFA, Load, and
FGR. PM mass emissions were measured at full load for two fuels.
Fuel Nitrogen/NOx Correlation - Tests conducted prior to the PG-DRB retrofit
provided NOx data from fuel oils with nitrogen contents ranging from 0.24% to 0.45%.
The data were used to develop a correlation between NOx emissions and fuel nitrogen
level for both BOOS and ABIS operation. Hie post outage performance test provided
NOx data at one fuel nitrogen level. For this task, a fuel oil with a higher level of fuel
nitrogen was obtained and tests run to develop a similar NOx/fuel nitrogen
correlation for a low-NOx burner system. Test variables included excess air (at full
load, constant OFA), and burner theoretical air (obtained by varying the amount of
OFA).
PERFORMANCE IMPROVEMENT PROGRAM RESULTS
Unit Performance With Existing Equipment
Typical full load NOx emissions were found to vary from 0.18 to 0.23 Lb/Mltu with
opacity levels ranging from 12 - 20%. Operating conditions included 22-24% OFA, 14-
17% FGR to the windbox, and approximately 3% excess O2. Fuel nitrogen levels were
approximately 0.4% (by weight). Consistent coke buildup on the oil gun tips was also a
daily maintenance problem. Although most of the coke deposits could be removed by
retracting the gun, a heavy scale remained which built up to the point where it became
difficult to remove the gun from the boiler for cleaning. A point-by-point traverse of
the sampling probes in the economizer duct revealed an O2 imbalance of
approximately 1% between the north and south sides of the furnace. Additional tests
8-93
-------
which were conducted with the OFA level reduced to a minimum showed no change
in the O2 non-uniformity, indicating a fuel/air imbalance at the burner front.
A boiler/windbox inspection was conducted during the unit outage. All spin vanes
and impellers were found to be operable and in good condition. Nothing was noted
(i.e., jammed spin vanes or air slides) which would have resulted in either an air or
fuel maldistribution to individual burners.
Post-Outage Performance Tests
Reference tests were conducted to document unit operation,and performance
immediately after the outage. The results were similar to those observed prior to the
outage, with NOx emissions of 0.19 lb/MBtu and opacity of 18% with 26% OFA, 20%
FGR to the windbox, and an operating excess O2 level of 3.2%. A gaseous traverse
again indicated an O2 imbalance of nearly 1% between the averages of the north and
south sides of the furnace.
Because of the importance of more uniform combustion conditions in achieving low
NOx and PM emissions, the fuel and air flows to the burners were balanced before
beginning the performance evaluation tests. Individual burner airflows could be
controlled by adjusting air slides and burner spin vanes. Spin vane settings were
selected to optimize individual flame appearance and shape. Air slide positions were
set to equalize the flow to each burner using the individual burner airflow monitors
Hand valves and flow meters in the oil lines supplying each burner also allowed fuel
flows to be balanced. The tests were conducted with minimum OFA flow so that
economizer excess O2 measurements could be used to judge balance at the burner
front. The effort was successful in reducing the difference in average O2 between the
two economizer ducts from 1% to below 0.5%.
A similar effort was conducted to balance OFA flows using OFA port damper and spin
vane settings. As-found operation of the OFA system was consistent with B&W's
recommendation that the OFA flow be biased 70% to the rear wall and 30% to the front
wall. During the flow balancing tests, it was found that further biasing the flow to a
80:20 distribution improved opacity. The new OFA settings maintained the O2 balance
described above.
The NOx and opacity levels noted with balance operating conditions were 0.23
lb/MBtu and 14-16% respectively with operating levels of 13% FGR and 23% OFA.
8-94
-------
These values are not significantly different from the post-oulage reference test results
(unbalanced conditions) when corrections for differences in FGR rate and OFA level
between the two tests are made. Nevertheless, the effort to balance the boiler was ail
important element in the overall program since it assured that the potential of
atomizer design to control NOx and FM emissions could be fully evaluated. The
existence of burners receiving either too little air or too much fuel relative to other
burners would limit the effectiveness of any atomizer design in reducing opacity.
The atomizer evaluation which was conducted involved a total of eleven different
designs. The design variables investigated included the following:
• Mean droplet size (5MD)
• Spray cone angle
• Steam/Oil differential pressure
• Distribution of oil within the spray cone
• Non-uniform spray patterns
A range of each of these variables was explored in order to find the configuration
which best fit the aerodynamics unique to this specific burner. The best performance
as judged by the NOx, opacity, and excess O2 levels achieved was provided by a non-
uniform spray pattern, and this design was selected for full performance optimization
and characterization tests.
Full load performance with the selected design is summarized in Table 1 and
compared to the post-outage reference test performance. As shown by the data,
emissions performance was significantly improved with opacity reduced from 18% to
10% and the same NOx level of approximately 0.19 lb/MBtu (148-152 ppm) achieved
using substantially lower levels of OFA (20% vs 26%) and FGR (11% vs 20%). Boiler
excess 02 levels were increased slightly to 3.5% in order to maintain opacity at a 10%
level. Atomizer steam consumption was also reduced from over 0.20 lb steam/lb oil
to 0.14. Oil gun coking, which had been a daily occurrence was eliminated. Particulate
emissions of 0.05 lb/MBtu were measured at peak load conditions of 145 MWg as
compared to the original acceptance test results of 0.07 lb/MBtu obtained after
commissioning the PG-DRB system in 1988.
8-95
-------
TABLE 1
COMPARISON OF POST-OUTAGE REFERENCE
TEST WITH NEW ATOMIZER PERFORMANCE
LOAD 130-135 MWg
Reference
New Atomizer
NOx, lb/MBtu
0.189
0.194
NOx, ppm, dry, 3% Oa
148
152
Opacity, %
18
10
Excess O2, %
3.2
3.5
O2 Imbalance, % O2
1
<0,5
OF A, %
26
20
FGR To Windbox, %
20
11
Atomizing Steam/Fuel
>0.2
0.14
Steam/Oil Differential, psig
50
30
Fuel Nitrogen, Wt.%
0.29
0.27
Oil Gun Coking
Daily
None
Particulate*, lb/MBtu
0.07
0.05
* Particulate tests conducted at peak load of 145 MWg; As found results from 9/7/88
B&W Acceptance Test
Figure 4 summarizes the NOx/opacity emissions history of Kahe 6 by comparing pre-
retrofit operation (219 ppm NOx; 6% opacity), post-retrofit/new operation (NOx range
141-172 ppm; opacity range 12-17%), operation two years after retrofit (NOx range 141-
180 ppm; opacity 12-20%), and finally operation with a balanced boiler and new
atomizers (NOx 152 ppm, opacity 10%). Figure 5 is a similar historical summary of PM
emissions, which decreased from the pre- and post-retrofit level of 0.07 lb/MBtu to 0,05
lb/MBtu.
Unit operation to this point did not utilize the PG system to supply recirculated flue
gas directly to the burners. Although previous tests demonstrated that an additional
NOx reduction of nearly 10% was achievable with 4% FGR through the PG system,
opacity levels also increased. With opacity levels already high, use of the system was
restricted to minimum levels of PG flow. The reduced opacity levels achieved with
8-96
-------
the new atomizers however allowed the use of PG as an additional NOx contTol.
Increasing the PG flow to 4% FGR reduced NOx emissions to 0.18 lb/MBtu (141 ppm)
from the level shown in Table 1 with opacity increasing slightly to 11%. Based on
these results, use of the PG system was incorporated into day-to-day operation.
Unit Performance Characterization With New Atomizers And Balanced Boiler
The last phase of the testing involved a systematic variation of boiler and combustion
system operating parameters. The purpose of these tests was to characterize the effect
on NOx emissions of variables such as excess oxygen, flue gas recirculation, overfire
air, fuel nitrogen level, and unit load. The results axe presented in Figures 6 through
10 and are discussed below.
Excess Oxygen. The effect of excess O2 variation on NOx emissions is shown in Figure
6. The tests were conducted at 130 MWg, with 20% OFA, 14% FGR (windbox + PG) and
a fuel nitrogen content of 0.26%. The data indicate a strong sensitivity of
approximately 45 ppm for each 1% change in O2. This sensitivity is higher than the 30
ppm/17c O2 found for the original PG-DRB system.
Flue Gas Recirculation. The effect of variation in FGR rate (windbox + PG) is shown
in Figure 7 for two levels of excess O2. Increasing the FGR rate from 10 to 20% would
reduce NOx by approximately 17%. This sensitivity is similar to that noted for the
original PG-DRB system.
Overfire Air. As with the original PG-DRB system, OFA is a very effective means of
reducing NOx emissions. As shown in Figure 8, increasing the OFA level from 8% to
22% results in a decrease in NOx of nearly 40%. The burner air /fuel ratio at 22% OFA
was approximately 95% of stoichiometric. The results shown in Figure 8 also illustrate
the strong trade-off between NOx and opacity.
Fuel Nitrogen Level. The fuel nitrogen level for the oils burned during the tests was
in the range of 0.26% - 0.3%. In an attempt to establish the influence of this variable
on NOx emissions an effort was made to procure a fuel with a nitrogen content above
0.4%. The maximum level which could be obtained however, was 0.34%. Comparing
NOx data obtained with 20% OFA and 10% FGR on 0.26% and 0.34% nitrogen fuels
indicated that NOx emissions increased approximately 30 ppm for fivery 0.1% increase
in fuel nitrogen. Figure 9 provides a comparison of NOx/fuel nitrogen sensitivities
obtained with the original pre-retrofit of the PG-DRB system burners operating both
8-97
-------
with all burners-in-service and off-stoichiometric (removing 3 of 9 burners from
service) and the PG-DRB system with new atomizers and a balanced boiler. The
influence of burner stoichiometry on NOx/fuel M sensitivity can be seen in the
decrease from 55 ppm NOx/O.1% fuel N for the air-rich all burners-in-service case, to
35 ppm/0.1% for fuel rich operation of the original burners, to 30 ppm for the PG-DRB
system.
Load, NOx emissions over the load range from maximum, valves-wide-open
operation (145 MWg) to minimum (55 MWg) are shown in Figure 10. As shown by
the data, NOx emissions are relatively insensitive to load variation. The excess 02 at
the economizer varies from approximately 3% at 145 MWg to 5% at 55 MWg.
CONCLUSIONS
NOx and PM/opadty emissions generally exhibit a very strong inverse relationship in
that modifications made to equipment and/or operation which reduce NOx emissions
usually result in increased levels of PM and opacity. The results of the program
demonstrated the potential for atomizer design to simultaneously control both NOx
and PM/opacity. The degree of success achievable with this approach is dependent on
carefully matching design variables such as spray angle and fuel distribution to the air
flow patterns unique to a specific burner design.
The ability to monitor individual burner and OFA flows proved to be indispensable in
optimizing and operating the combustion system.
The use of a non-uniform spray pattern atomizer design resulted in reduced NOx
emissions which allowed a reduction in the amount of OFA and FGR required for
NOx control. This reduction in OFA and FGR levels, together with a balanced
distribution of fuel and air to the burners and OFA ports and improved atomization
quality resulted in reduced opacity levels. Although balancing flows did not in itself
significantly change NOx or opacity levels, the control of fuel and air flows did not
allow those burners which had been operating with a deficiency of air to become a
limiting factor in the improvement potential of the atomizers themselves.
The lowered opacity and the reduced levels of OFA and FGR now required for
operation significantly improved the available operating margin required to allow for
variabilities in operation and fuel properties. The performance characterization tests
demonstrated that OFA and FGR rates are dominant influences on NOx emission
8-98
-------
levels and quantified the additional NOx reduction potential available by increasing
their level of use.
Additional operational improvements which were demonstrated included the
elimination of an oil gun coking problem which made gun removal for cleaning and
maintenance difficult, reduced attemperation requirements due to reduced levels of
FGR, and reduced atomizer steam consumption.
ACKNOWLEDGEMENTS
The authors wish to acknowledge Combustion Components Associates, Inc. who
provided design support for the atomizer evaluation and conducted the atomizer
laboratory characterization, both Kilkelly Environmental Associates, Inc. and A. G.
Enterprises who operated the emissions monitoring system, and METCO
Environmental who conducted the particulate mass emissions tests. Additionally, the
cooperation and advice received from the Kahe Station operating and maintenance
personnel during the program is gratefully acknowledged.
REFERENCES
1. J. L. Yee, R. B. Freitas, D. V, Giovanni, S. 1. Kerho, M. W. McElroy. "Retrofit of
an Advanced Low-NOx Combustion System at Hawaiian Electric's Oil-Fired
Kahe Generating Station." 1989 Symposium on Stationary Combustion
Nitrogen Oxide Control, Volume 2, EPRIGS-6423, July 1989
2. D; V. Giovanni. "Predicting Carbon Emissions From a Utility Boiler Firing
Residual Fuel Oil." Fuel Utilization 1989 Workshop: Proceedings, EPRI GS-6459
8-99
-------
Figure 1. Kahe Unit 6 Elevation Drawing
8-100
-------
(mown event)
III WW! ll» _
Hln
*11 riOU MEKttOt
OtfUt MHULUt
Wll MWU.M
e*l UCltC M»ll Mn.1
me smium?
MIEt
-------
Load 130-135 MWg
Fuel N 0.3%
Indicates Bottom
Of Range Noted
1981-88
(BOOS, FGR)
PG-DRB
(NEW)
PG-DRB
(2 YEARS)
PG-DRB
(EPRI)
Figure 4. Historical Summary of Kahe 6 NOx and Opacity
1981-88
PRE-RETRO
PG-DRB
(NEW)
PG-DRB
(EPRI)
Figure 5. Historical Summary of Kahe 6 Particulate Emissions
8-102
-------
o
tn
v.*¦
«
Q
eu
c-
5
2
250
200-
150
100-
50-
0
¦
¦
¦
a
LOAD 130 MWg
FGR 14%
OFA 20%
FUEL N 0.26%
-
1 "I1 1 1
, ¦¦¦ 1 ,
2.5 3 3.5 4 4.5 5 5.5
CONTROL ROOM EXCESS 02, %
Figure 6. Effect of Excess 02 on NOx Emissions
fM
o
D
Cm
Ch
k"
O
Z
LOAD 130 MWg
OFA 22%
FUEL N 0.27%
PG 3-5%
' » I 1 1"
I I I
5 10 15 20 25 30
FLUE GAS RECIRCULATION (WINDBOX+PG), %
Figure 7. Effectiveness of FGR in Reducing NOx Emissions
8-103
-------
250
25
0
200
feS
f*l
150
fti
D
S
Ch
100
eu
K
O
2
50
LOAD 130 MWg
FGR 14%
FUEL N 0.28%
10 15 20
OVERFIRE AIR (% OF TOTAL AIR)
Figure 8, Effect of OFA on NOx and Opacity
o
m
a
Q
CU,
b
X"
I
500 ¦
450-
400'
350
300-
250-
200-
150-
100-
50-.
LOAD -130 MWg
FGR - 10%
02- 3%
| ABIS
1
BOOS
i
'
" PG-
nnn
20% C
)¥A~
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
FUEL NITROGEN, WT. %
Figure 9. NOx Dependence on Fuel Nitrogen Content
8-104
-------
200-
at
Q
Bn
CM
*
O
2
120-
100-
80-
60-
40-
20-
0-
40
60
T 1 T
r*11-
80 100 120
LOAD, GROSS MW
140
160
Figure 10. NOx Dependence on Unit Load
8-105
-------
-------
DEMONSTRATION OP ADVANCED L04f ~ NO* COMBUSTION TECHNIQUES
AT THE GAS/OIL-FIRED FLEVO POWKR STATION UNIT 1
J.G. Mitkarap
KEMA
P.O. Box 9035
6800 ET Arnhem
The Netherlands
J. van der Koolj
Sep, Dutch Electricity Generating Board
P.O. Box 515
6800 AN Arnhem
The Netherlands
G. Roster
Stork Boilers
P.O. Box 20
7550 GB Hengelo
The Netherlands
J.R. Sijbrlng
EPON
P.O. Box 10087
8000 GB Zwolle
The Netherlands
8-107
-------
-------
ABSTRACT
The Dutch Electricity Production Companies have developed a concerted N0X
Abatement Programme to reduce NO* emissions of power stations. One of the major
projects was the demonstration of advanced low-nox combustion techniques at the
Flevo power station unit I. The boiler is horizontally opposed fired with gas and
heavy Euel oil and the Eollowing techniques were demonstrated:
• low N0X burners.
• Flue gas recirculation, applied separately through the burners as
well as mixed with the combustion air.
• Two-stage combustion by deep staging.
• In Furnace Reduction by reburning technology (at 70% load).
In the framework oE the demonstration programme a measurement programme was
performed, producing a vast amount of information about nox emission and the
consequences of combustion modifications for boiler performance and boiler
maintenance. NO* concentrations down to 50 mg/mg3 for gas firing and 160
mg/mQ3 for oil firing could be obtained. In order to gain additional
Information about the Nc^ reduction mechanism and about a possible danger for
enhanced water tube corrosion measurements were also conducted inside the furnace
and at the furnace wall.
8-109
-------
INTRODUCTION
Since 1971 NO* emission has been under debate in licensing procedures for new
power stations and gradually NO* control technology has been applied to an
increasing extent in many power stations in the Netherlands. The growing concern
on acidification of the soil led to a nation wide regulation of the emissions for
nox and SO2 and in 1987 a General Administrative Order on emissions of large
combustion installations came into force. For new conventional boilers with
gas-firing the N0X concentration was limited to 200 mg/mg3 at 3% O2 and
for oil-fired boilers to 300 mg/mp3 at 3% 03. For gas firing these values
may be further reduced to 100 mg/mg3 at 3% O2 in 1992 and 60 mg/mo3 at
3% 02 in 1994 (draft proposal of December 1990). For oil firing the proposed
limits are 150 mg/rtiQ3 at 3% O2 in 1992 and 110 mg/mo3 at 3% 03 in
1994.
The electricity companies in the Netherlands have established that reduction of
*»°x emissions has a high priority. Therefore they formulated a Concerted nox
Abatement Programme, in which the emphasis is put on NOx reduction by combustion
modification. The different activities are coordinated by Sep (Dutch Electricity
Generating Board); most projects are new completed or almost completed. One of the
major projects was the demonstration of advanced low-NOx combustion techniques
at the Flevo power station unit 1. The main purpose of the demonstration project
was to establish the lowest possible NO^ emission by the most advanced
combustion techniques for gas and heavy fuel oil for front wall fired
installations known at the time of the retrofit. A second objective was to assess
how these techniques could be applied without loss of reliability of the boiler,
safety and dynamic behaviour and with a minimum loss of boiler efficiency. The
project was financed by Sep and NOVEM (the Dutch Association for Energy and
Environment).
The boiler was commissioned in 1968 and has a capacity of 185 MWe. The boiler was
originally equipped with two rows of four parallel flow burners both in the front
and the back wail of the furnace, in the spring of 1988 the boiler was converted
to low NOx firing and the following modifications were implemented:
• The burners were replaced by double register burners with the
possibility to apply primary gas (PG). A diagram of the burner is
presented in Figure 1. The primary gas is recirculated flue gas
which can be supplied through a separate annulus between the
primary air and the secondary air of the burner, in case there is
no PG cooling, air is supplied through this duct. The burners are
able to fire gas and oil. The gas spuds are placed between the
primary air duct and the PG duct.
8-110
-------
• In addition to PG, recirculated flue gas can be mixed with the
combustion air (GM = gas mixing).
• On top oE the two rows oE main burners one row oE planetary
burners was added on both sides of the boiler. The construction
oE these burners Is similar to the construction oE the main
burners. With these burners Euel can be burnt with a very low
stoichiometry so as to create hydrocarbon radicals with a
lifetime long enough to react with the NO of the main combustion
zone of the boiler (In Furnace NOx Reduction; IFNR).
• On top of the planetary burners one row of after air ports was
added on both sides of the boiler. An after alt' port consists of
a central pipe and around this central pipe a concentric duct,
which is connected with a register. The axial impulse and thus
the penetration of the air can be varied by the register setting
and by a damper on the central pipe.
ft diagram of the boiler is presented in Figure 2. The boiler Is an overpressure
boiler and In order to apply these measures and to avoid CO-leakage, the front and
back walls of the boiler were rebuilt and all boiler walls were made gastight.
Boiler control and the flame safeguarding was modified. The investment cost of the
retrofit project was about 21 million Dutch guilders (about 12.5 million us
dollars).
Boundary conditions
A requirement for the conversion was that conventional firing (all burners
operating with an excess of air) without flue gas recirculation would remain the
normal firing mode. This led to requirements for burner capacity and the heat
transfer surface of the convection banks. In practice the boiler capacity with
IFNR was therefore limited to 70 % in case of Elue gas recirculation through the
burners. Another reason to limit the load to 70 % in case of IFNR was the
residence time required for burnout.
The requirement that the boiler would remain optimised for conventional firing
implied that for IFNR-firing dampers had to be put In an extreme position, with
the consequence that the pressure loss in the ducts was high with IFNR and the
fans were operating close to the pumping limit. This In turn limited the amount of
flue gas that could be recirculated (particularly GM) and the lower limit that
could be reached with the stoichiometry of the planetary burners: a stoichiometry
of 0.4 - 0.45 was the lowest value achievable; for almost all experiments the
planetary burners were operated on a stoichiometry of 0.5.
EXPERIMENTAL PROGRAMME
The research programme was executed in the period July 1988 - June 1990. The major
part of the programme was dedicated to parameter research, which included:
• The firing mode:
-- Conventional firing.
8-111
-------
— Two stage combustion
-------
Gas Firing
It was found that a substantial reduction in N0X emission could be achieved. The
minimum NOx concentration was 45 - 50 mg/mg^ at 3% O2. which corresponds
to a reduction oE almost 95% in comparison with the original NO^ concentration
of "780 mg/rag3 at 3% 02- This concentration of 45 - 50 mg/mQ3 at 3%
O2 could be obtained by deep staging as well as by 1FNR, both in combination
with flue gas recirculation. Figure 3 provides an overview of the relative
effectiveness of the different techniques. The modification with low NOx burners
and the increased number of burners gave a reduction of NO* concentration to 285
rag/mo3 at 3% O2" a further reduction to about 100 mg/mp3 at 3% 02 is
possible both by PGR and by applying either deep staging or IFMR. The lowest value
can be obtained by a combination of the techniques. The data from Figure 3 were
obtained with 11% FOR, consisting of 10% GM and *?% PG; an amount of 17% FGR could
be applied for ail firing modes.
Experiments with two stage combustion with variation of the amount of after air
had indicated that the minimum NOx concentration could be reached with a
stoichlometry of about 0.8 and that the NOx concentrations tended to increase
again with a further reduction of the stoichlometry. This is shown in Figure 4,
where the NO* concentration is plotted as a function of the stoichlometry at the
burners, both for the situation with FOR and without FGR. It can be seen from the
figure that already with a moderate staging quite a substantial reduction In NOx
emission can be obtained. In the figure two measurement series are represented,
with a time difference of about half a year. For conventional firing the
difference in NC^ concentration is substantial. The difference may be attributed
to differences in burner settings, but no satisfying explanation was found; it was
concluded that this type of differences is probably inherent to experiments in a
large installation, In which not all parameters can be controlled as accurately as
In a test furnace.
The maximum amount of FGR that could be utilized for 1FNR firing was about 20%
(11% GM and 9% PG) and the minimum NOx concentration had still not been reached;
the reason for the restricted amount of flue gas recirculation was not only the
already mentioned fan capacity, but also the flame safeguarding which had to be
operational for all firing modes. With 20% FGR an NOx concentration of about 40
rag/mg3 at 3% O2 was measured with ifnk.
It appeared that GM was slightly more effective in reducing the NC^ emission
than PG or PG and GM combined, probably because of the more complete mixing with
the combustion air. in case of GM supply the flame temperature, being the most
important parameter for thermal NOx formation, Is probably better reduced than
by PG. However, because of the limited amount of GM that could be applied, the
support of PG was still helpful for a further reduction. Even for low initial
NOx concentrations FGR was very effective, as can be seen in Figure 5, where the
NOjj concentration Is plotted as a function of the amount of flue gas
recirculation (GM + PG). For conventional firing with 17% FGR the reduction was
63%; for IFNR the reduction was still almost 50%. For all firing modes the
recirculated flue gas was equally distributed over main burners and planetary
burners.
For deep staging and for IFNR it was necessary to increase the excess air from 4%
to almost 15% in order to keep CO emission below 50 ppm. For moderate staging the
amount of excess air was 8%. There were virtually no possibilities for improving
the mixing of the combustion products with the after air (and thus reducing the
excess air) for deep staging or IFNR. In order to provide a sufficient amount of
8-113
-------
after air the pressure drop across the after air ports had to be minimized by a
full opening of the dampers and minimum rotation.
The effect of excess atr and boiler load on the N0X concentration was small for
the lower NO* levels. For deep staging and for IFNR the effect of the boiler
load was about 10 mg/mQ^ over the load range studied of 35 - 10*- For deep
staging, and for IFNR the effect of excess air on NOx concentration was
negligible.
At the beginning of the project it was expected that IFNR would give a lower UOx
concentration than TSC. Compared with a moderate staging this was indeed the case,
but compared with deep staging there was hardly any difference. This, however,
demonstrated the need to gain a better understanding of the processes in the
boiler and several questions were raised:
• Do the CH-radicals indeed form at the planetary burners and is
their lifetime long enough? and if so, is there enough
penetration into the furnace?
• is it possible that the mx concentration at the end of the
main combustion zone is too low and that destruction of NO Is
compensated for by prompt nox formation at the planetary
burners?
• A stolchioraetry (SR) of 0.9 is often considered to be an optimum
in the reburnlng zone; is it possible that a stoichiometry of 0.7
is too low?
In order to have a higher initial concentration and to Increase the stoichiometry
of the reburning zone experiments were also carried out with a stoichiometry of
the main combustion zone (SRI) up to about 1.03 while keeping the stoichiometry at
the planetary burners at 0.5. The results are shown in Figure 6 as a function of
the amount of fuel for the planetary burners ("planetary fuel"). As reference
points the concentrations measured for two-stage combustion are taken with the
corresponding stoichiometrics. The results show that there is indeed a decrease in
NOx concentration, but it is unknown which part may be attributed to dilution,
if for instance with 33% planetary fuel a reduction of 33% by dilution is assumed,
it can be seen from the Figure that the reduction of NO by a chemical mechanism is
negligible.
Figure 6. also presents experiments with a reduced amount of planetary fuel. From
visual observation and later on also from tracer measurements it was assessed that
the penetration of planetary fuel was insufficient for values of 20% and lower.
There was, however, no way to increase the momentum except for increasing the
stoichiometry at the planetary burners, which was not considered practical.
Oil firing
For oil firing the same series of experiments were conducted as for gas firing. A
minimum NOx concentration was found of about 160 mg/mg3 at 3% 02* Before
retrofit an NC^ concentration of 500 mg/mg3 at 3% O2 was measured by
firing a heavy fuel oil with 0.18% nitrogen. This Implies a reduction of 65%, but
the difference in fuel nitrogen of the two measurements makes it difficult to
compare the values. By applying only deep staging an NOx concentration of 180
mg/mo^ at 3% O? could be obtained; a subsequent addition of FGR reduced the
8-114
-------
NO* concentration onLy by an extra 20 rag/mo3 down to 160 mg/mo3 at 3%
02- ft survey of the results can be found in Figure 1, with IFNR the lowest NO*
concentration that could be reached was 190 rag/rag-* at 3% 02; the addition
of PGR only caused an increase of the NO* concentration. The low1 NOx burners
had only little effect. In comparison with the situation before modification the
decrease in NOx emission for conventional firing is small: a reduction by 65
rag/rag3 to 435 mg/mg3 at 3% O?. By applying 11% FOR the NO^
concentration could subsequently be reduced to 360 mg/mo3 at 3% °2-
In general the effect of flue gas recirculation was small. This can be seen in
Figure 8, where for the different firing modes the NOx concentration is plotted
as a function of flue gas recirculation. The reason Is that for oil firing only a
part of the NO is generated by the thermal route. For conventional firing and for
a moderate staging GM was more effective than PG or GM + PG; for deep staging PG
was more effective than PG or GM + PG. For IFNR there Is an increase in NOx
concentration when applying FGR. This may be explained by a reduction of the
residence time in the reburnlng zone when Increasing the amount of FGR.
For TSC the minimum NOx concentration was found for a burner stolchiometry of
about 0.80; without FGR this figure was slightly lower as can be seen in Figure 9,
probably due to a longer residence time in the boiler for this situation.
Rs with gas firing, questions were raised about the effectiveness of IFNR.
Therefore, experiments were also carried out with an increased stolchiometry and
consequently an increased NO production of the main combustion zone. The
stolchiometry of the main combustion zone was increased up to about 1.05 while
keeping the stolchiometry of the planetary burners at 0.5. The results are
presented in Figure 10, in which a comparison is made with TSC. The interpretation
is more complicated than for gas firing, because for oil there is an additional
contribution by the conversion of the fuel-bound nitrogen of the planetary fuel.
From experiments with gas on the main burners and oil at the planetary burners a
conversion of the fuel nitrogen at the planetary burners of about 30% was
estimated. This was higher than the conversion for deep staging, which was 17% in
the case of an NOx concentration of 160 mg/mo3 at 3^ °2- l-ow levels
the NO formation of the planetary fuel compensated for the reduction by reburning
and was probably also responsible for the fact that the NOx concentration
obtained by deep staging could not be reached by IFNR. For higher NOx
concentrations the planetary fuel appeared to be effective in reducing NOx
emission.
Experiments with 20% planetary fuel showed a reduced effectiveness in decreasing
NOx emission.
Rs for gas firing, the excess air had to be Increased to about 15% for deep
staging and IFNR to keep the CO concentration below 50 ppm. For IFNR-oil firing a
visible plume could not be avoided, although the measured concentrations of solids
were only slightly higher than for conventional firing and in the same order of
magnitude as for TSC: * 40 mg/mo3. Scanning electron microscope pictures,
however, revealed a different nature of the solids. For conventional firing and
TSC the solids consisted mostly of cenospheres, whereas for IFNR the concentration
of cenospheres was small, but the filters showed quite a thick layer of black
dust. This could be ascribed to the formation of hydrocarbon chains by
devolatllized oil components.
8-11 5
-------
oil on the main burners and gas at the planetary burners
One of the objectives of the project was also to get a first Impression oC the
possibilities o£ applying IFNR to coal firing with gas or oil as a possible
planetary fuel. This was one of the reasons why experiments with a reduced amount
of planetary fuel were incorporated in the programme, because replacing one third
of the fuel is regarded as too much for coal firing; 10% and maybe 20% is more
within the acceptable range. It was felt that the best approach of coal-IFNl was a
situation with oil on the main burners and gas on the planetary burners.
The minimum NOx concentration that could be obtained was 110 - 120 rag/m03
at 3% O2 with 33% gas. This was achieved with a stolchiometry of 0.80 - 0.85 on
the main burners and 0.5 on the planetary burners. The results of the experiments
are given in Figure 11; the stolchiometry of the main burners was varied between
1.05 and 0.80. The reduction obtained in relation to 100% oil firing (0% planetary
fuel) again depended on the initial NC^ concentration. The maximum decrease
found was 53%, corresponding with a reduction from 470 rag/mQ3 at 3% O2 to
220 mg/mg3 at 3% O2. In order to estimate the chemical reduction of NO,
NOx values have been corrected for the dilution by multiplying the measured
NOx concentration by (100/(100 - % planetary fuel). The results of this exercise
are presented in Figure 12, where the corrected values are compared with two-stage
oil firing. For conventional firing (stolchiometry = 1.05) and a moderate
staging it was found that there is about 25 - 30% reduction that cannot be
explained by dilution only.
IN-FURNACE MEASUREMENTS
Visualization of the reburning zone
Information on the penetration of the flame of the planetary burners could be
obtained by the atomization of Naci- solution in the combustion air in one or more
planetary burners; injection In one burner gives an impression of a single flame.
The flame was visualized by recording the emission line of sodium of 590 run by
means of a spectrograph. With a measurement of the 430 nm emission line the
presence of hydrocarbons could also be detected. The spectrograph was mounted as
close as possible to a peeping hole at the planetary burner level, but
nevertheless the angle of vision was small and the measurement was difficult. By
using several peeping holes and adjusting the optics to different angles it was
possible to evaluate the flow pattern of the flames. A diagram for 33% planetary
fuel is presented in Figure 13, indicating a sufficient degree of penetration;
with 20% planetary fuel penetration was found to be Insufficient. The zones
Indicated in the figure are the visualized areas. Measurement of the CH-radicals
in some situations showed a concentration in the order of 1-10 ppm (being the
lower detection limit with this method) up to 2.5 metre's distance from the
planetary burners; this corresponds with about one quarter of the furnace width.
Combustion gas composition measurements
Combustion gas composition measurements inside the Eurnace were performed for IFNR
and deep staging. The composition was measured by inserting a suction probe
through one of the side walls of the boiler, in Figures 14 and 15 the gas
composition at three representative locations Is presented. The lines represent
measurement, locations just below the planetary burners (Tl), between the planetary
8-116
-------
burners arid the after air ports (T2) and above the after air potts (T3). Although
the number of measurements at locations Tl and T2 was restricted due to the
limited length of the probe (required for room to manoeuvre). It can be seen that
there is a difference between tsc and 1FNK. Especially at point T2 the NO
concentration was much lower for IFNR. indicating a reaction of NO with the
planetary fuel. The NO concentration at T3 and at the outlet of the boiler,
however, was the same in both situations. This implies that most probably a
reduction in NO in the IFNR case is compensated for by NO formation with the after
air. This could be prompt NO formation from the planetary fuel or a reaction to NO
again of the Intermediates, which were formed at the planetary burners (by the
reaction of NO with hydrocarbons). Also, measurements were carried out at a
position corresponding to area 3 of Figure 13. An O2 concentration up to 4% was
measured, indicating an admixture of the after air already at this level.
Combination of the results of the parameter investigation and the in-furnace
measurements showed the following;
• The presence of a reduction zone is likely.
• There is a reduction of NO as long as the initial NO
concentration is high enough (or as long as the stolchiometry at
the reducing zone is not too low: these effects could not be
separated).
• The reduction zone did not extend to the middle of the furnace.
Combustion gas measurements at the furnace wall
With IFNR-oil firing (with a reburning zone stolchiometry SR2 of 0.7) the
concentrations of o2. CO, co2. H2S, S02 and h20 were measured at 23
locations along the furnace wall. High H2S concentrations were found in the
entire area below the after air ports with a maximum of 1100 ppm (just below the
planetary burners). The 02 concentration in this area was 0% (+ 0.02%) and the
CO concentration was 3 - 5%. Above the after air ports the conditions were
oxidizing. Below the burner zone, in the hopper area, the environment was
oxidizing, calculations with 950 • 1100 ppm H2S. with the assumption of a
thermodynamic equilibrium at the furnace wall and a temperature of 425 "c at the
surface of the evaporation tubes, gave a partial sulphur pressure of about 1
-------
BOILER OPERATION
Burner damage
From February 1989 to November 1989 the performance of the programme was
Interrupted. The boiler had to be taken out of operation due to extreme coke
formation on one of the planetary burners. The parameter research had just been
finished and the boiler had operated for 1% weeks on IFNR with oil. The purpose
was to study the behaviour of several test materials under continuous IFNR
operation with oil. During a follow-up inspection it was established that the
burner in question was severely damaged by combustion Inside the burner, but a
more important outcome was that several other burners were damaged as well. It was
found among other things that the most important damage was deformation and
corrosion of a number of PG ducts, mainly of the planetary burners but also of the
burner row below. The damage had happened at sichroraal (xlOerftllB) outlet rings of
the PG ducts. It was not until the end of September, however, that an opportunity
was found to repair the burners. In the period February 1989 to October 1989 the
boiler was operated in load-following manner with conventional gas firing using 23
burners. No experiments were conducted in this period. In October 1989 the damaged
burners were repaired.
In order to minimize the chance of repeated damage the following actions were
taken:
• Metallurgical investigation of the damaged burner parts and the
application of test strips to the PG rings of several burners.
• Registration of the temperature of the critical burner parts in
order to define a range of firing modes for safe operation.
The highest temperatures - above 1000 °c - were measured at the PG rings at full
load and at 10% load for conventional gas firing without PG. For partial load,
IFNR or the application of 5% PG the temperatures of the PG rings were reduced to
about 400 °C; for deep staging without PG the temperatures were a little higher
(up to 600 °c). For oil firing the temperatures of the PG rings were below 600 °c
under ail circumstances. The metallurgical investigation showed thick layers of
oxides on some PG rings and recrystalllsatlon of the materials, indicating
temperatures of 950 °c and higher. Also, some sulphldation was found. The general
conclusion was that without PG supply the cooling air to the burners had been
insufficient and the high temperatures in combination with a reducing atmosphere
had prevented the formation of a protective oxide skin.
Protective measures were taken, but they probably came too late and in February
1990 it was established that some burners (PG rings) were damaged again. For
technical and economical reasons the PG rings were removed at the first possible
opportunity, which was during a boiler stop in June 1990. For the time left it was
decided to restrict the test programme to those experiments which were considered
most essential. These were the in furnace experiments already mentioned. The rest
of the scheduled programme - dynamic tests, extensive efficiency measurements and
the corrosion test programme (which was interrupted by the burner damage) - was
cancelled. For the future it is envisaged to apply a moderate TSC as the normal
firing mode.
8-118
-------
Operational experience
• Switching from one firing mode to an other went smoothly.
• The starting procedure and the starting speed of the boiler had
remained unchanged.
• The dynamic behaviour of the boiler is acceptable.
• There have been no problems with furnace vibrations.
• Despite the combination of an overpressure boiler and reducing
conditions at the furnace wall, safety was sufficient at ail
times.
Boiler efficiency
Although no extensive measurements were carried out to determine boiler efficiency
for the different firing modes, calculations were made according to DIN 1942 as a
next best assessment. These calculations showed a maximum difference for gas
firing of 0.4%: from 95.5% (conventional firing) to; 95.1% (IFNR with FOR). For oil
firing there was an increase in flue gas temperature at the boiler outlet,
resulting in a maximum decrease in efficiency from 93.7% (conventional firing) to
92.0% for IFNR with FGK. The effect on unit efficiency, however, was greater due
to an increased amount in superheat spray water for IFNR and TSC. With respect to
new installations the construction of the boiler will be adapted to an advanced
firing mode and it is not expected that for new units there will be a significant
loss in efficiency, provided that the mixing processes in the boiler are
controlled. f
EVALUATION
The results of this demonstration project are considered to be satisfactory. This
is especially true for the results obtained for gas firing, whereby the goals for
low NOx values were fully achieved. For oil firing the results are also
satisfactory, considering the nitrogen content of the oil. The results may be
applied to other installations for the calculation of N0X emissions, assuming
that the furnace load, residence time in the reducing zone and the time needed for
burnout can be accounted for. In terms of percentages, the reduction is probably
strongly related to the present installation and it cannot be expected that for
other installations the same percentage in NOx reduction can be obtained.
The problems encountered with the damaged PG rings emphasize the Imperative to pay
special attention to the integrity of the burners when applying new combustion
techniques. It was felt that the problems encountered were primarily a consequence
of the demonstration character of the programme, which included two fuels and a
lot of different flow and flame conditions at the burners. This can be avoided by
reducing the amount of different fuels and parameter settings and by a proper
estimation of the gas composition and temperatures for the resulting settings.
On the basis of the analysis of the reasons for the burner damage it was concluded
that the damage could have been avoided if better cooling had been applied, either
by cooling air or by PG. The solution finally chosen was a removal of the pg
rings. Even in this case with only GM the NOx concentration is still low: with
10 - 12% GM an NOx concentration of 60 - 70 mg/mg^ at 3% O2 can be
8-119
-------
obtained. This indicates that, although the problems encountered had a quite large
Impact on the performance of the scheduled programme, the final Judgment on the
demonstration project may not be affected by these problems.
The demonstration project has shown that a balanced combination oE advanced
low-N0X combustion techniques is a powerful tool in the abatement of N0X
emission. It has also provided information on the limits of low-NOx control
technology for front-wall-fired boilers with gas- and oil-firing.
8-120
-------
t
ruEl OIL
Figure 1. Low-NO* Dual Fuel Burner with FG
Addition
Figure 2. In Furnace NOx Reduction System
Cor oil and Natural Gas Firing
8-121
-------
800
780
0
£
m
sp 600
$
a o
E
D!
o
z
400
200
BEFOFttE
' (no PGR)
CONVENTIONAL
FGR=0%
CONVENT IONAL
FGR=1?%
TSC SR=0.95
FGR=0%
E3 TSC SR=0.99
FG=t= 17%
I:. 11 TSC SR=O.80 '
EZ3 tsc SR=o.eo
FGR=i?%
IFNR
FGR=0%
! I IFNR
FGFt=17%
Figure 3. Summary Gas Firing; Effect Firing Mode at 10% Load
300
0
n
w
fi o
E
x
a
o
2
200
100
f I
/
A
M,
¦
-J
—'
3-c=3Z5S
0
0.60
6
O
FGR=0%
series 1
FGR=0%
series 2
FGR= 1.7%
series 1
FGR=17%
series 2
0.70
0.80
0.90
1.00
1.10
Figure 4.
stoichiometry burners
Effect of Staging; Gas Firing 70% Load
8-122
-------
0
2?
ro
+-»
m
fl o
£
D>
0
Z
400
300
200
100
CONVENTIONAL
PG
CONVENTIONAL
GM
CONVENTIONAL
PG+GM
TSC SR=0,97
PG
TSC SR=0.9?
GM
TSC SR=0.97
PG+GM
IVNR
GM
IVNR
PG+GM
flue gas recirculation !% of air)
Figure 5, Effect Flue Gas Recirculation; Gas Firing 70% Load
o
0s
n
to
i ' 1 ' ' ' ' 1 ' ¦
•s.
a _
--- *
— 9
~ ~ A
¦ *— *-
a SR1 = 1.02
FGR=0%
a SR 1=1.04
FGR=17%
o SR 1 =0.94
FGR=0%
* SRI =0.94
FGR=17%
^ SR 1 =0.89
FGR=0%
~ SR 1=0.90
FGR=17%
~ SR 1 =0.79
FGR=17%
10 15 20 25 30 35
planetary fuel (%)
Figure 6. Variation Amount Planetary Fuel; Gas Firing 70% Load
8-123
-------
600
0 500
m
«
01
400
300
200
100
500
180 1QC*
i6cr
215
y
N%—
0.18
N%=0.34
EH BEFORE
(no PG/GM)
CONVENTIONAL
FGR=0%
conventional
FGR=17%
TSC SR=0.95
FQ=i=0%
E3'TSC SR=0.95
FGR=17%
CHI] TSC SR=C.B0
FGR=0%
I ..".I TSC SR=0.B0
FGR= 17%
t I IFNR
FGR=C>%
I i IPNR
FGR=17% .
Figure 7. Summary Oil Firing; Effect Firing Mode at 70% Load
o
m
co
ci o
cn
.0
500
4O0 -
300
200
100
conventional
PG
conventional
GM
conventional
PG+GM
TSC SR=0.95
PG
TSC SR=0.95
GM
TSC SR=0.95
PG+GM
TSC SR=0.80
PG+GM
IFNR ¦
PG+GM
0,5 10 15 20 25
flue gas recirculation (% of air)
Figure 8. Effect Flue Gas Recirculation; Oil Firing 70% Load
8-124
-------
500
150
0.60
0.70
0.80
0.90
1.00
stoichiometry burners
Figure 9. Effect of Staging; Oil Firing 70% Load
FGR=0% •
FGR=17%
1.10
o
sP
os
n
(0
o
D*
E
O
Z
500
400
300
200 -
100
0
10
20
30
planetary fuel (%)
4 SR 1 = 1.05
V SR 1=0.95
o SR 1=0.85
o SR 1=0.79
40
Figure 10. Variation Amount Planetary Fuel; FGR = 0%; Oil
Firing 10% Load
8-125
-------
0
£
ro
ra
m ©
e
o
e
X
O
500
400
300
200
100
0
10
20
30
planetary fuel (%)
40
A SR 1-1.05
' SR 1 =0.95
~ SR 1 =0.85'
o SR 1=0.80
Figure 11. Variation Amount Planetary Fuel; FOR
10% Load
0%; Oil-Gas
o
£
CO
*•>
©
ft O
e
D)
o
z
500
« 450
150
0.70 0.80 0.90 1.00
stoichiometry bumers/SR 1
o
planetary
fuel 20%
planetary
fuel 35%
1.10
Figure 12. Effect of Planetary Fuel; Oil-Gas; Comparison with
TSOOil
8-126
-------
-- After air ports
- Planetary burners
»Main burners
Figure 13, Estimated Flame Shape at the
Planetary Burners with IFNR Gas Firing
a
a
c
o
'¦&
to
+-*
c
-------
E
a
a
c
0)
o
c
o
a
2
25
20
15
10
•
i -O T1
' \T3
-6— T 1
-A— T2
-*-'T3
100- 200 300- 400
distance to funace wall (on)
500
Figure 15. NO Concentration inside the Furnace for TSC with
SR = 0.80
8-128
-------
Appendix A
1991 Joint Symposium on Stationary Combustion NOx Control
03/25/91-03/28/91
The Capital Hilton
Washington DC
List of Attendees
Jan van der Kooij
Environmental Affairs Dept.
Sep/Dutch Elec.Generating Board
Utrechtseweg 310
6812 AR Arnhem
THE NETHERLANDS
+31/85 721473
Pierre van Grinsven
Senior Development Engineer.
KSLA - Kon Shell Lab Amsterdam
Badhuisweg 3
1031 CM Amsterdam
T1IE NETHERLANDS
020/303818
Harold Abbasi
Mgr., Applied Combustion Research
Institute of Gas Technology
4201 West 36th Street
Chicago, 1L 60632
312/890-6431
Andris Abele
Program Supervisor
So.Coast Air Quality Hgmt.District
9150 Flair Drive
El Monte, CA 91731
818/572-6491
Alberto Abreu
Sr Air Pollution Ctrl Engr
San Diego Air Pollution Ctrl Dist
9150 Chesapeake Dr
San Diego, CA 92123
619-694-3310
Jerry Ackerman
Mgr., Contract Research Marketing
Babcock & Wilcox
1562 Beespn Street
Alliance, OH 44601
216/829-7403
Rau Acosta
Asst. Ops. Bupt.
Florida Power & Light
P. 0. Box 13IIB
Ft. Lauderdale, FL 32316
305/527-3543
Michael Acme
Senior Proj. Engineer
Kilkelly Environmental
P. 0. Box 31265
Raleigh, NC 27622
919/781-3150
Ken Adams
Senior Scientist
Ontario Hydro
700 University Avenue
Toronto, Ontario
MSG 1X6 CANADA
416/592-4333
Rui Afonso
Senior Engineer
New F.nglnirl Power
Research & Development
25 Research ISrive
Westhorongli, HA 01582
N/A
Bhuban Agatw.il
Gen. Mgr., EA Division
Foster Whealer Energy Corp.
8 Peach Tree Hill Road
Livingston NJ 07039
201/535-2372
Annette Ahart
Section Leader
EG&G WASC, Inc.
P. 0. Box 880
Horgantown, WV 26507-0880
304/291-4463
A-1
-------
Raymond Aichner
Supv.Plant Engineering
Southern California Edison
6635 S. Edison Drive
Oxnard, CA 93033
805/986-7244
Jeffrey Allen
Special Combustion Projects Manager
NEI-International Combustion Ltd.
Sinfin Lane
Derby DE2 99J
ENGLAND
332 271111
Maurice Alphandary
N/A
AEA Technology - ETSU
B156 Harwell Laboratory
Oxfordshire 0X11 QRA 44
UNITED KINGDOM
N/A
Leonard Angello
Technical Manager
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2673
Patrick Aubourg
Manager, R&D
Owens Corning Fiberglass
2790 Columbus Road, Rte.16
Granville, OH 43023
614/587-7604
Robert Badder
Power Production Manager
City of Independence Power & Light
21500 E. Truman Road
Independence, HO 64056
816/796-4400
P, Baimbridge
First Engineer
PowerGen Pic.
Moat Lane, Solihull
West Midlands
ENGLAND
(ENG.)021-701-3873
Aldo Baldace.i
Manager
ENEL-CRTN
Via A. Plsfliio, 12
Pisa 56100
ITALY
0039/50-535744
Lothar Balling
Manager, DeNOx
Siemens KWU/T123
Frauenauracherstr. 85
Erlangen, 8520 GERMANY
9131/186151
Maureen Barbemi
Conference Coordinator
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2127
Joe Barkloy
Chemical Engineer
Tennessee Valley Authority
P, 0. Box 150
West Padnr.ali, KY 42086
502/444-4657
William Dartok
Senior Vice President
Energy & Environmental Research
P. 0. Box 139
Whltehonse, NJ 08888
908/534-5033
R. J. flatyko
Mgr., Environmental Projects
Babcock & Wilcox
20 S Vnn Buren Ave.
Barberton, OH 44203
216/860-1654
Frank Bauer
Corporate Consultant
Stone & Webster
Three Executive Campus
Cherry HJ13, NJ 08034
609/482-3284
A-2
-------
J*
Nick Bayard da Volo
President
ETEC'
One Technology, Suite 1-809
Irvine, CA 92718
714/753-9125
Peter Beal
Manager, Business Development
NEI-International Combustion Ltd,
Sinfew Lane
Derby
ENGLAND
332/27 11 11
Frank Beale
Mgr., Boiler Burner Systems
John Zink Company
4401 South Peoria Ave
Tulsa, OK 74170
918/748-5180
Robert Becker
President
Environex, Inc.
P. 0. Box 159
Wayne, PA 19087
215/975-9790
Janos Beer
Scientific Director
Massachusetts Instit. of Technology
MIT Combustion Research Facilities
Cambridge, MA 02139
617/253-6661
Edward Behrens
Product Manager, DeNOx
Joy Environmental Equipment Co.
404 E. Huntington Drive
Monrovia, CA 91016-3633
818/301-1215
F. Bennett
Sr. Systems Engineer
Potomac Electric Power
1900 Pennsylvania Ave., N.W.
Washington, DC 20068
202/872-2442
Mogens Berg
N/A
ELKRAFT Power Company Ltd,
Lautruphoj 5
DK-2750 Ballerup
DENMARK
+45 42 65 61 04
Elliot Berman .
President
Project Sunrise, Inc.
6377 San Como Lane
Camarillo, CA 93012
805/388-0208
Leif Bernergard
Technical Officer
Swedish Environm.Protection Agency
S-171 85 Solna
SWEDEN
+468 799 II 19
Naum Bers
N/A
Consultant
2111 Jefferson Davis Highway
Apt. 1219 North
Crystal City, VA 22202
N/A
Kamal Bhattacharyya
Head, Emissions Evaluation
Ministry of Environment
Air Management Branch
810 Blanshard Street
Victoria, BC VBV 1X5 CANADA
604/387-9946
Ramon Biarnns
Managing Director
Land Combus 11 on
2525-B Pearl Buck. Rd
Bristol, PA 19007
215-781-0810
Richard Biljetlna
Assistant Vice President
Institute of Gas Technology
3424 S. State
Chicago, IL 60616
312/890-6418
A-3
-------
Gary Bisonett
Senior Steam Gen.Engineer
Pacific Gas & Electric Co.
245 Market Street, 434A
San Francisco, CA 94106
415/973-6950
John Bitler
President
Environmental Catalyst Consultants
P. 0. Box 247
Spring House, PA 19477
215/628-4447
Verle Bland
Emissions Control Supervisor
Stone & Webster
P. 0. Box 5406
Denver, CO 80217-5406
303/741-7684
Richard Boardman
Senior Engineer
Westinghouse Idaho Nuclear Co,
P. 0, Box 4000 MS 5218
Idaho Falls, ID 83402
208/526-3732
Danny Bolerjack
Maintenance Foreman
Alabama Power Co.
Miller Steam Plant
4250 Porter Road
Quinton, AL 35130
205/674-1207
Richard Borio
Executive Consulting Engineer
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-2229
Steven Bortz
Manager, Western Lab
Research-Cottrell Envir.Serv/Tech.
9351 B Jeronimo
Irvine, CA 92718
714/830-2255
Ernest Boufford
Senior Air Pollution Control Engr.
State of Connecticut
165 Capitol Ave., loom 136
Hartford, CT 06106
203/566-8230
Richard Boyd
Program Manager
Radian Corporation
2455 Horsepen Road
Herndon, VA 22011
703/834-1500
Bernard Breen
President
Energy Systems Associates
1840 Gateway Three
Pittsburgh, PA 15222
412/392-2380
Fiorenzo Bregani
Senior Researcher
ENEI.-CRTN
Milan, ITALY
N/A
John Brewster
Ass't. Plant Manager
Cajun Electric
112 Telly Street
New Roads, LA 70760
504/638-3773 ,
Frank Brideri
Chemist
U.S.Envlronmental Protection Agency
Air & Energy Eng'g Research Lab.
Research Triangle Park, NC 27711
919/541-7808
Daslav Brkic
Manager/Chemical & Envir. Catalysts
U0P
25 East Algonquin Road
Des Plains, II, 60017-51017
708/391-2677
A-4
-------
R. G. Broderick
Consultant
RJM Corporation
10 Roberts Lane
Ridgefield, CT 06877
203/438-6198
Bert Brown
Vice President, Technology
Joy Environmental Equipment Co.
404 E. Huntington Drive
Monrovia, CA 91016-3633
818/301-1172
William Browne
Environmental Engineer
U.S.Environmental Protection Agency
841 Chestnut Bldg.
Philadelphia, PA 19107
215/597-6551
C. P. Brundrett
Manager, Emission Control
V. R. Grace & Co. - Conn.
10 East Baltimore St.
Baltimore, HD 21202
301/659-9125
Hans Buening
Sen. Staff Engineer
Radian Corporation
7 Corporate Park
Irvine, CA 92714
714/261-8611
Galen Bullock
Project Engineer
Carolina Power & Light
P. 0. Box 1551
Raleigh, NC 27602
919/546-2768
Daniel Butler
Deputy Group Leader
Los Alamos National Laboratory
Group T-3, MS B216
Los Alamos, NM 87545
505/667-9099
Gary Camody '
Product Mgr., Burner Systems
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-5039
E. J. Campobenedetto
Mgr.,N0x Control Systems
Babcock & Wilcox
P. 0. Box 351
Barberton, OH 44203
216/860-6762
Gene Capriotti
Vice President, Sales
Nalco Fuel Tech
2001 West Ma 1n St., Ste. 295
Stamford, CT 06902
203/323-8401
Ben Carmine
Supervising Engineer
Houston Lighting & Power Co.
P. 0, Box 1700
Houston, TX 77251
713/922-2191
Steven Carpenter
Market Analyst
Diamond Power
P. 0. Box 415
Lancaster, OH 43130
614/687-4363
Doug Carter
General Engineer
U.S. Department of Energy
1000 Independence Ave., S.W.(FE-4)
Washington, DC 20585
202/586-1]fifi
Carlo CastaJdini
Project Hnnager
Acurex Corporation
555 Clyde Avenue
P. 0. Box 7044
Mountain View, CA 94039
415/961-5700, X3219
A-5
-------
P, Cavelock
Sr. Project Engineer
Potomac Electric Power
1900 Pennsylvania Ave,, N.W.
Washington, DC 20068
202/672-2447
Charles Chang
Mechanical Engrg.
h.h. Department of Water & Power
P. 0, Box 111
Los Angeles, CA 90051-0100
213/481-3235
Kwok-Ping Ching
Environmental Protection Officer
Environ.Protect.Dept.,Hong Kong Gov
28thfloor, Southern Centre
130 Hennessy Road
Wan Chai, HONG KONG
852-8351074
Roger Christman
Program Manager
Radian Corporation
2455 Horsepen Road
Herndon, VA 22071
703/834-1500
Landy Chung
President
Phoenix Combustion, Inc.
P, 0. Box 2257
Ashtabula, OH 44004
216/964-6396
Ed Cichanowicz
Project Manager
Electric Power Research Institute
1019 Nineteenth St, N.W.
Suite 1000
Washington, DC 20036
202/293-7515
David Clay
Manager
Kraftanlagen Heidelberg
c/o AUS, 1140 East Chestnut Ave.
Santa Ana, CA 92701
714/953-9922
John Cochran
Ass't.Group Leader,Air Qual.Control
Black & Veatch
8400 Ward Parkway
P.O. Box 8405
Kansas City, HO 64114
913/339-2190
Thomas Coerver
Engineer Supervisor
Louisiana Dept.of Environ.Quality
P. 0. Box 44096
Baton Rouge, T,A 70804
504/342-8912
Mitch Cohen
Consultant
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-2482
William Coler
Senior Market.lug Specialist
Babcock & Wilcox
1562 Beeson
Alliance, Oil 44601
216/829-7317
Robert Collette
Project Mgr.; Low NOx Projects
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-5687
Richard Collins
Mechanical Engineer
Tennessee Volley Authority
1101 Market Street (MR 3B)
Chattanooga, IN 37402-2801
615/751-7935
Robert Combs
Corporate Research Specialist
Virginia Power
Innsbrook Technical Center
5000 Dominion Blvd.
Glen Allen, VA 23060
804/273-2975
A-6
-------
Joseph Comparato
Mgr> , Process Development
Nalco Fuel Tech
P.O. Box 3031
1001 Frontenac Road
Napervilie, IL 60566-7031
708/983-3247
Raymond Connor
Technical Director
Industrial Gas Gleaning Institute
1707 L Street, N.W. , Ste, 570
Washington, DC 20036
202/457-0911
Thomas Cosgrove
Manager, Testing Services
Research-Cottrell Envir.Serv/Tech.
P. 0. Box 1500
Somervi1le, NJ 08876
908/685-4619
David Cowdrick
Senior Engineer
Tampa Electric Co.
P. 0. Box 111
Tampa, Ft 33601
813/228-4111,X46269
H.Tom Creasy
Engineer
Virginia Dept.Air Pollution Control
P. 0. Box 10089
Richmond, VA 23240
804/786-0178
David Crow
Manager, Faber Div.
Tampella Keeler
2600 Reach Road
Williamsport, PA 17756
717/326-3361
D. P. Cummings
Associate Engineer
Ontario Hydro
700 University Avenue
Toronto, Ontario MSG 1X6
CANADA
416/592-4505
Donna Currie
Engineering-Generating
Potomac Electric Power
1900 Pennsylvania Ave., N.W.
Washington, DC 20068
202/331-6280
G. D'Anna
Ansaldo Component 1 Representative
Ansaldo Component!, B&W Interntl.
c/o Babcock & Wilcox International
20 South Van Buren Ave,
Barberton, 011 44203
216/860-6029
Manny Dahl
PEPS, Project Manager
Babcock fit Wilcox
20 South Van Buren
Barberton, Oil 44203
216/860-6634
Donna Dant
Environmental Engineer
Louisville Gas & Electric
P. 0. Box 32010
Louisville, KY 40332
502/627-2343
R. M, Davies
Manager, Engineering Science
British Gas Pic
Midlands Research Station
Solihull, West Midlands B91 2JV
ENGLAND
0/21-705-7581
Charles Davis
Sr. Staff Engineer
Virginia Power
5000 Dominion Blvd.
Glen Allen, VA 23060
804/273-2619
Michael Deland
Chairman
President1s Council/Environ.Quality
Executive Office of the President
The White House
Washington, DC
N/A
A-7
-------
Mukesh Desai
Supervisor, Env. Technology
Bechtel
9801 Washington Blvd.
Gaithersburg, HD 20878
301/417-3158
Arun Deshpande
Abatement Engineer
Ministry of Environment
A.P.I.O.S. Office
135 St. Clair Ave.,W, StclOOO
Toronto, Ontario,M4V IPS CANADA
416/323-5055
Larry Devillier
Eng.Supervisor, Air Permits
Louisiana Dept.of Environ..Quality
P. 0. Box 44096
Baton Rouge, LA 70804
504/342-8926
J.G. DeAngelo
N/A.
New York State Electric & Gas
4500 Vestal Parkway, E.
Binghamton, NY 13902
607/729-2551
N. N. Dharmarajan
Principal Engineer
Central & South West Services
1616 Woodall Rodgers Freeway
Dallas, TX 75202
214/754-1373
Richard Diehl
Dir.,Coal Tech.,Energy Tech Office
Textron Defense Systems
2385 Revere Beach Parkway
Everett, MA 02149
617/381-4282
Daniel Diep
Senior Research Engineer
Nalco Fuel Tech
One Nalco Center
Naperville, 1L 60563-1198
708/305-2047
Joseph Diggins
Mgr. Pittsburgh District Sales
Foster Wheeler Energy Corp,
300 Corporate Center Dr. Ste.130
Coraopolis, PA 15108
412/264-0611
Roger Dodds
Air Quality Engineer
Wisconsin Electric Power
333 W. Everett St.
Milwaukee, WI 53201
414/221-2169
Patrick Doherty
Senior Engineer
Coastal Power Production Co.
310 First Street, 5th floor
Salem, VA 24153
703/983-4365
Stephen Doll
District Manager
Riley Stoker Corporation
4108 Park Road, #315
Charlotte, NO 28209
704/527-8877
Brandon Donahue
Client Manager
ABD Combustion Engineering
1200 Ashwood Parkway, NE
Suite 510
Atlanta, GA 30338
404/394-2616
Les Donaldson
Mgr., Emissions Control Research
Gas Research Institute
8600 W. Bryn Mawr Avenue
Chicago, Hi 60631
312/399-8295
Dirk Doucet
N/A
Gulf States Utilities
P. 0. Box 2951
Beaumont, TX 77704
409/838-6631
A-8
-------
Barry Downer
Boiler Engineer
National Power PLC
Whitehil Way Swindon
Wilts, ENGLAND
(SWINDON) 892263
Brian Doyle
Principal
Brian Doyle Engineering
Six Sunset load
Putnam Valley, NY 10579
914/528-0139
John Doyle
Sales Engineer
Babcock & Wilcox
7401 W. Mansfield Ave, Ste.410
Lakewood, CO 80235
303/988-8203
Dennis Drehmel
Dpty.DirPollution Control Div.
U.S.Environmental Protection Agency
AEERL (MD-54)
Research Triangle Park, NG 27711
919/541-7505
H.C.W. Drop
N/A
Rodenhuis & Verloop
Oosterengweg 8
1221 JV Hilversum
THE NETHERLANDS
+31 35 88 1211
Richard Dube
Consultant
Stone & Webster
245 Summer Street
Boston, HA 02107
617/589-7831
J. D.H. Dumou1in
N/A
EPON
Dr. Stolteweg 92
8025 AZ Zwolle
HOLLAND
038/ 27 29 00
David Duncan
Air Permits Coordinator
Texas Utilities Electric
400 N. Olive St., LB 81
Dallas, TX 75201
214/812-8479
llao Duong
Engineer
Dayton Power St Light
P. 0, Box 468
Aberdeen, OH 45101
513/549-2641,X5B32
Michael Durham
Vice President, R&D
ADA Technologies, Inc.
304 Inverness Way South
Suite 110
Englewood, CO 80112
303/792-5615
Michael Durilla
Sr. Tech. Service Engineer
Engelhard Corporation
101 Wood Avenue
Iselin, NJ 08830-0770
908/205-6644
Hans-Jurgen Durselen
Engineer
RWE Energie AG
Lannerstr. 30
405D Monchenglodbach 4
Essen, GERMANY
02166/58943
George Dusatko
Vice President & Gen. Mgr.
Energy Systems Associates
1840 Gateway Three
Pittsburgh, PA 15222
412/392-2372
Richard Dye
General Engineer
U.S. Department of Energy
FE-4
Washington, DC 20585
202/586-6499
A-9
-------
Owen Dykema
President
Dykema Engineering, Inc.
23429 Welby Way
Canoga Park, CA 91307
818/348-3751
Ed Ecock
Steam Gen.Engineer
Consolidated Edison of N.Y.
Four Irving Place
New York, NY 10003
212/460-4830
Raj Edwards
President
EnviroTech International
335 Park St, NE
Vienna, VA 22180
703/938-5138
D. R. Eisenmann
V.P.,SCR Systems Div,
Peerless Mfg. Co.
2819 Walnut Hill Lane
Dallas, TX 75229
214/357-6181
John Eldridge
Prof.of Chemical Engineering
University of Massachusetts
39 Kendrick Place
Amherst, MA 01002
413/253-5991
William Ellison
Director
Ellison Consultants
4966 Tall Oaks Drive
Monrovia, NO 21770
301/865-5302
Thomas Emma1
Senior Staff Engineer
Radian Corporation
3200 East Chapel Hill Road
Research Triangle Park, NC 27709
919/541-9100
Michael Escarcega
Sr. Environmental Engineer
Southern California Edison
P. 0. Box 800
2244 Walnut Grove Ave.
Rosemead, CA 9 J 770
818/302-4032
Art Escobar
Environmental Engineer
Virginia Dept.Air Pollution Control
9th Street Office Bldg.
Richmond, VA 23219
804/786-5783
David Eskinozi
Project Manager
Electric Power Research Institute
3412 Hillvicw Avenue
Palo Alto, CA 94304
415/855-2918
Lee Ewing
Engineer
U.S. Department of Energy
9141 Vendomn Drive
Bethesda, MIJ 20817
301/353-5442
Nancy ExconHe
Proposal Manager
Babcock & Wilcox
74 E. Robinson Avenue
Barberton, OH 44203
216/860-2328
Edward Farkas
Senior Engineer
Canadian Gns Research Institute
55 Scarsdalr Rr>ad
Don Mills, Ontario
M3B 2R3, CANADA
416/447-6465
Hamid Farzan
Sr. Research Engineer
Babcock & Wilcox
1562 Beeson St.
Alliance, OH 44601
216/829-7385
A-10
-------
Michael Fatigatl
Liaison Engineer
Babcock & Wilcox
4332 Cerritos Ave. Ste.204
Los Alamitos, CA 90720
714/236-0432
George Feagins
Environmental Engineer Senior
Virginia Dept.Air Pollution Control
121 Russell Road
P. 0. Box 1190
Abingdon, VA 24210
703/676-5582
Paul Feldraan
Director R&D
Research-Cottrel1 Envir.Serv/Tech.
P. 0. Box 1500
Somerville, NJ 08876
908/685-4880
W. K. Felts
Air quality Regulatory Analyst
Potomac Electric Power
1900 Pennsylvania Ave., N.W.
Washington, DC 20068
202/331-6179
James Ferrigan
N/A
WahLeo, Inc.
4707 College Blvd.
Leawood, KS 66211
913/491-9292
Abe Finkelstein
Chief, Clean Air Technologies
Environment Canada
Unit 100 Asticou Center
Hull, Quebec
CANADA
819/953-0226
Tom Fletcher
Combustion Research Facility
Sandia National Laboratories
P. 0. Box 969
Livermore, CA 95376-0969
415/294-2584
John Foote
Senior Engineer
University of Tennessee
Space Institute
B.H. Goethert Parkway
Tullahoma, TN 37388
615-455-0631
John Galtsk.U 1
Engineer
U.S.Environmental Protection Agency
230 South Dearborn
Chicago, 1L 60604-1504
312/886-6705
Ivo Gal Hubert i
Professor
University of Padova
Via Gradenlgo 6A
35131 Padovn
ITALY
33/49-828-754]
Michael Gambutg
V.P., Western States Op.
Todd Combustion, Inc.
61 Taylor Reed Place
Stamford, CT 06906
203/359-1320
Wayne Gensler
Combustion Engineer
Selas Corporation America
P. 0. Box 200
Dresher, PA 19025
215/283-8338
Robert Giammar
Mgr.,Process Engineering Dept.
Battelle Memorial Institute
505 King Avenue
Columbus, 0!! 43201
614/424-7701
A. F. Gi Jlespie
Engineering Manager
Foster Wheeler Ltd.
P. 0. Box 3007
St. Catharines, Ontario
CANADA
416/688-4434
A-11
-------
Dan Giovanni
President
Electric Power Technologies, Inc.
P. 0. Box 5560
Berkeley, CA 94705
415/653-6422
Philip Goldberg
Coal Utilization Division
Pittsburgh Energy Tech. Center
P. 0. Box 10940, MS 92211
Pittsburgh, PA 15326
412/892-5306
Toby Gouker
Mgr., Stationary Emission Control
W. R. Grace & Co. - Conn.
7379 Rt. 32
Columbia, MD 21044
301/531-4131
Loic Gourichon
Engineer
CERCHAR
Rue Aime Dubost
62670 Hazingarbe
FRANCE
33/21 72 81 88
Mary A. Gozewski
Editor
Coal & Synfuels Technology
1401 Wilson Blvd., Suite 900
Arlington, VA 22209
703/528-1244
Martin Grant
Senior Engineer
AUS Combustion Systems, Inc.
1140 East Chestnut Ave.
Santa Ana, CA 92701
714/953-9922
Michael Grimsberg
Tekn. Lie.
University of Lund
Dept. of Chem. Eng.II,Box 124
5-221 00 Lund
SWEDEN
+46/46-108276
John Grusha
Mgr.,Firing Systems Engrg.
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-3497
Manoj Guha
Mgr., Technical Assessment
American Electric Power
One Riverside Plaza
Columbus, OH 43220
614/223-1285
James Guthrie
Assoc.Air Resources Engineer
State Air Resources Board
P. 0. Box 2815
Sacramento, CA 95812
916/327-1508
Steven Guzinski
Mechanical Engineer
Naval Energy & Envir.Support Activ.
NEESA-11A
Port Huennmc, CA 93043-5014
805/982-5388
Greg Haas
Mechanical Engineer
Exxon Research and Engineering
2BOO Decker Hrive
Haytown, TX 77522
713/425-7892
Donald Hfigsr
President
Damper Design, Inc.
1150 Hauch Chunk Rd.
Bethlehem, PA 18018
215/861-0111
Leo JIakka
Project Development Mgr.
CANSOLV
Union C.irblde Canada Ltd,
Box 700, Pofnte-Aux-Trembles
Quegec HIB 5KR CANADA
514/4993-2617
A-12
-------
Robert Hall
Branch Chief
U.S.Environmental Protection Agency
Combustion Research Branch (MD-65)
Research Triangle Park, NC 27711
919/541-2477
M. Halpern
Proj.Licensing Coor.-Gen.Engrg,
Potomac Electric Power
1900 Pennsylvania Ave,, N.W.
Washington, DC 20068
202/331-6489
David Ham
President
EnviroChem, Inc.
54 Bridge Street
Lexington, MA 02173
617/863-1334 .
Doug Hammontree
Project Manager
Burns & McDonnell
4800 East 63rd Street
Kansas City, M0 64141-6173
816/822-3115
Frank Harbison
Senior Analyst
Louisiana Power & Light
P. 0. Box 60340, Unit N-31
New Orleans, LA 70160
504/595-2308
Stan Harding
Vice President
IE I
317 Marion Drive
McMurray, PA 15317
412/941-9202
Robert Hardman
Research Engineer
Southern Company Services
P. 0. Box 2625
Birmingham, AL 35202
205/877-7772
Doug Hart
Prin.Engr., Firing Systems
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-2439
Gary Hausman
Project Engineer
Pennsylvania Power & Light
Two North Ninth Street
Allentown, PA 18101
215/774-6562
Robert Hnyes
Operations Specialist
Illinois Power Co.
500 S. 27th Street
Decatur, II. 62525
217/424-8101
John Healy
Mgr..Generating Schedule/Cost
Potomac Electric Power
1900 Pennsylvania Ave., N.W.
Washington, DC 20068
202/872-3596
Dennis HelfH l:ch
Mgr., Technology Assessment
Research-Cottrell Envir.Serv/Tech.
P. 0. Box 1500
Somervilie, N.I 08876
908/6B5-4147
Todd He 1.I ewe 11
Engineering Support Manager
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-4919
S, Hashemi
Sr. Project F.ngineer
Potomac Electric Power
1900 Pennsylvania Ave., N.W.
Washington, DC 20068
202/331-6495
A-13
-------
R, Henry
Sr. Project Engineer
Potomac Electric Power
1900 Pennsylvania Ave,, N.W.
Washington, DC 20036
202/872-2441
Hark Hereth
Senior Chemical Engineer
Radian Corporation
2455 Horsepen Road
Herndon, VA 22071
703/834-1500
Andrew Hetz
Environmental Engineer, Senior
Virginia Dept.Air Pollution Control
7701 Timberlake Road
Lynchburg, VA 24502
804/947-6641
Steven iliggins
Engineer, R&D
Pennsylvania Electric Co.
1001 Broad Street
Johnstown, PA 15907
814/533-8883
Duane Hill
Mrg., Performance Admin.
Dairyland Power Coop
3200 East Ave. S
LaCrosse, W1 54602
608/787-1424
Richard Mimes
Project Engineer
Fossil Energy Research Corp,
23342 C South Pointe
Lagnna Hills, CA 92653
714/859-4466
Anna Hinderson
Process Engineer
ABB Carbon AB
612 82 Finspong
SWEDEN
4-46-122 81000
John Hofmann
Vice President, Engineering
Nalco Fuel Tech
1001 Frontenac Road,
Naperville, II. 60563
708/983-3252
Gerald Hoi linden
Sr. Program Manager
Radian Corporation
633 Chestnut Street
Chattanooga, TN 31450
615/755-0811
Kevin Hopkins
Senior Engineer
Carnot
15991 Red HilJ Road
Suite 110
Tustin, CA 92680-7388
714/259-9520
Richard Hotchklss
N/A
National Power
Kelvin Ave., Leatherhead
Surrey KT22 7SE
ENGLAND
703-374488
Reagan Houston
President
Houston Consul i'.ing, Inc.
252 Foxhunt Lane
Hendersonvi iIe, NC 28739
704/642-3722
Vincent Huang
Program Manager
A. 0. Smith Corp.Techno logy
12100 V. Park Place
Milwaukee, WT 53224
414/359-4255
Alex lluhmomt
Mgr.,Air Pollution Control Sys.
Public Service Electric & Gas
80 Park Plaza
P. 0. Box 570, MC-19E
Newark, NJ 07101
201/430-6997
A-14
-------
Terry Hunt
Professional Engineer
Public Service Company of Colorado
5900 East 39th Avenue
Denver, CO 8020?
303/329-1113
Peter Imm
Principal Engineer
01in Corporation
P, 0. Box 2896
Lake Charles, LA 70602
318/491-3481
Ivan Insua
Senior Engineer
Salt River Project
P. 0. Box 52025
Phoenix, AZ 85072-2025
602/236-5240
Robin Irons
Team Leader, NOx Control Tech.
PowerGen
Ratcliffe Technology Centre
Ratcliffe-on-Soar
Nottinghamshire, NG11 0EE, ENGLAND
602/830591, X2437
Bruce Irwin
Engineering Manager
Hauck Manufacturing Co.
P. 0. Box 90
Lebanon, PA 17042
717/272-3051
Reda Iskandar
V.P., Sales & Marketing
Cormetech, Inc.
8 E. Denison Parkway
Corning, NY 14831
607/974-4313
Keijo Jaanu
Technology Development Mgr.
KEMIRA, Inc.
P. 0. Box 368
Savannah, GA 31402
912/236-6171,X149
Rudolf Jaerschky
Director, Power Plant Department
Isar-Amperwerke AG (IAW)
Brienner Strasse 40
Munchen 2, GERMANY 8000
089/5208-2621
James Jarvis
Senior Staff Engineer
Radian Corporation
8501 Mo-Pac Blvd.
Austin, TX 78720-1088
512/454-4797
Jeff Jensen
Civil/Mechanical Design Supervisor
Wisconsin Public Service Corp.
600 North Adams
Green Bay, Wl 54307
414/433-1864
Ken Johnson
Environmental Affairs Manager
Duke Energy Corporation
400 S. Tryon St,
Wachovia Center
Charlotte, NC 28202
704/373-5089
Larry Johnson
Project Manager
Southern California Edison
2131 Walnut Orove Avenue
Rosemead, OA 91770
818/302-8542
Robert Johnson
Regional SalManager
Wahlco. Inc.
4707 College Blvd., Suite 201
Leawood, KS 66211
913/491-9292
Steve Johnson
Vice President
PSI Technology Co.
20 New England Business Center
Andover, MA (HALO
508/689-0003
A-15
-------
Dale Jones
N/A
NoelI, Inc.
1401 East Willow Street
P.O. Box 92318
Long Beach, CA 90800-2318
213/595-0405
Anda Kalvins
Environ.Studies Specialist
Ontario Hydro
700 University Avenue
Toronto, Ontario MSG 1X6
CANADA
416/592-3193
Bent Karll
Senior Manager
Nordic Gas Technology'Centre
Dr, Neergaards Vej 5A
DK-29 70 Horsholm
DENMARK
45/45 76 69 95
Anders Karlsson
Reporter
Technical Outlook
Swedish Off.of Science fit Tech.
600 New Hampshire Ave., N.W.
Washington, DC 20037
202/337-5170
Hans Karlsson
Professor
University of Lund
Dept. of Chem.Eng. II, Box 124
S-221 00 Lund
SWEDEN
+46/46-108244
Wally Karrat
Advisory Engineer
IBM - T.J.W. Research
Route 134
P. 0. Box 218
Yorktown Heights, NY 10598
914/945-35166
Borchert Kassebohm
Director
Stadtwerke Dusseldorf AG
Am Wiedenhof 7
D 4000 Dusseldorf, GERMANY
0211/821-2459
Sandy Kaupang
Air Pollution Control Engineer
Burns & McDonnell
4800 East 63rd Street
Kansas City, M0 64141-6173
816/333-4375
Bruce Kautsky
Boiler Specialist
United Engineers & Constructors
460 E. Swedpsford Rd
Wayne, PA 19087
215-254-5155
Donald Knweck.i
Section Mnnngor
Foster Wheeler Energy Corp.
Perryvi11e Corporate Park
Clinton, NJ 08809
908/730-5466
Jim Kennedy
Service Rep
Foster Wheeler Energy Corp.
2001 ButterId Road
Downers Grove. IL 60515-1050
708/241-2850
Stephen Kerho
Consulting Engineer
Electric Powr Technologies, Inc.
24672 Venablo Lane
Mission Viejn, CA
714/380-7316
Tanveer Khan
R&D Engineer
Ahlstrom Pyropower, Inc.
8970 Crestmar Point
San Diego, CA 92121
619/552-2323
A-16
-------
Hark Khinkis
Asso.Dir,.Applied Combustion Res.
Institute of Gas Technology
4201 West 35th Street
Chicago, IL 60632
312/890-6452
J.Leslie King
Combustion Engineering Manager
Babcock Energy Ltd.
PorterfieId Road
Renfrew, PA4 8DJ
SCOTLAND
41/886-4141
Allan Kissam
Senior Salesman
Joy Environmental Equipment Co.
404 E. Huntington Drive
Monrovia, CA 91016-3633
818/301-1100
Edward Kitchen
Senior Engineer
Fichtner USA, Inc.
Overlook 1, Suite 360
2849 Paces Ferry Rd., NW
Atlanta, GA 30339
404/432-6983
John Kitto
Program Manager
Babcock & Wilcox
1562 Beeson St.
Alliance, Oil 44720
216/829-7710
Peter Knapik
Manager, R&D
UOP
25 E. Algonquin Rd.
Des Plaines, IL 60017-5017
708/391-2554
Bernard Koch
Director, Project Development
Consolidation Coal Company
4000 Brownsville Road
Library, PA 15129
412/854-6612
Angelos Kokklnos
Project Manager
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2494
Zofia Kosim
Environmental Engineer
U.S.Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
202/475-9400
Gerrit Roster
Process Service Engineer
Stork Boilers
Postbus 20
7550 GB Hengelo
THE NETHERLANDS
074/401416
Vaclav Kovac
Design Engineer Specialist
Ontario Hydro
700 University Ave.
Toronto, Ontario
CANADA
416/592-5243
Toshio Koyanagi
Senior Engineer
Mitsubishi Hnnvy Industries
2 Houston Center, Suite 3800
Houston, TX 77010
713/654-415!
Ed Kramer
Sr. Production Engineer
PS1 Energy
P. 0. Box 40fl
Owensville, IN 47665
812/386-4212
Henry Krigmont
DirTechnics I Dept.
Wahlco, Inc.
3600 W. Scgerr,trom Ave,
Santa Ana, CA 92704
714/979-7300
A-17
-------
K.S. Kumar
Manager, Applications
Research-Cottrel1 Envir.Serv/Tech,
P, 0. Box 1500
Somerville, NJ 08876
908/685-4876
Naveen Kumar
Project Engineer
Sargent & lundy
55 E. Monroe
Chicago, IL 60603
312/269-6706
Yul Kwan
Consulting Engineer
Applied Utility Systems, Inc.
1140 East Chestnut Avenue
Santa Ana,CA 92701
714/953-9922
H, K. Kwee
N/A
Stork Boilers B.V.
P. 0, Box 20
7550 6B llengelo
THE NETHERLANDS
31/74 40 18 57
Richard La Flesh
Principal Consulting Engineer
ADD Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/205-2583
Yan t.achowicz
Product Mgr., Burner Systems
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-2581
Don Langley
Regional Service Manager
Babeock & Wilcox
7401 W. Mansfield Ave #410
Lakewood, CO 80235
303-988-8203
El len 1*3 num
Mgr.Conferences & Exhibits
Electric Power Research Institute
3412 Hi11view Avenue
Palo Alto, CA 94304
415/855-2193
Leonard Lspatnick
Environmental Research Engineer
Public Service Electric & Gas
80 Park Plaza, T16H
Newark, NJ 07101
201-430-8129
Dennis Laudnl
Research Engineer
University of North Dakota
Energy & Environ.Research Center
P. 0. Box 8213, University Station
Grand Forks, ND 58202
701/777-5138
Tom Laursen
Development Engineer
Babeock & Wilcox
20 S. Van Buiren Ave.
Barberton, OH 44203
216/860-6142
A1 LaRue
Advisory Engr/Combustion Systems
Babeock & Wilcox
20 S. Van Bur^n Avenue
Barberton, OH 44203
216/860-1.493
Steve Legedzn
Mgr., Industrial Process Tech.
Consumers G.i
-------
Joel S, Levine
Senior Research Scientist
NASA Langley Research Center
Atmospheric Sciences Division
Hampton, VA 23665
804/864-5692
Julian Levy
Dir., Atmospheric Science Div.
Versar, Inc.
9200 Ramsey Road
Columbia, HD 21045
301/964-9200
Robert Lewis
Consulting Engineer
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-5968
John Lewnard
Principal Process Engineer
Air Products and Chemicals, Inc.
7201 Hamilton Blvd.
Allentown, PA 18195-1501
215/481-6932
Sergio Ligasacchi
Thermal/Nuclear Research
ENEL-CRTN
Via A. Pisano, 120
Pisa 56100
ITALY
050/535622
William Linak
Project Officer
U.S.Environmental Protection Agency
AEERL (MD-65)
Research Triangle Park, NC 27711
919/541-5792
Robert Lisauskas
Director, R&D
Riley Stoker Corporation
45 McKeon Road
Worcester, HA 01610
508/792-4801
Mike Little
Chemical Engineer
Tennessee Valley Authority
P. 0, Box 150
West Paducah, KY 42086
502/444-4654
Jim Locher
Engineer, Production
Pennsylvania Electric Co.
1001 Broad Street
Johnstown, PA 15907
814/533-8547
Judith Lomox
N/A
Maryland Power Plant Research Prog.
301/974-2261
Robert Lott
Project Mannger
Gas Research Institute
8600 West Bryn Mawr Ave.
Chicago, IL 60631
312/399-8228
Phillip Lowe
President
INTECH, Inc.
11316 Rouen Drive
Potomac, HD 20854-3126
301/983-9367
Tien-Lin Lu
Senior Mechanical Engineer
Arizona Public Service Company
P. 0. Box 53999
Phoenix, AZ 85072-3999
602/250-4731
Richard Lyon
Senior Scientist
Energy & Environmental Research
18 Mason
Irvine, CA 92718-2789
201/534-5833
A-19
-------
Denis Maftei
Sr.Process Engineer
Ministry of Environment
880 Bay Street, 4th floor
Toronto, Ontario H5S 1Z8
CANADA
416/326-1649
Herwig Maier
Dept. Mgr., Steam Gen & Envir.Tech.
Energie-Versorgung Schwaben AG(EVS)
Hauptverwaltung
Krlegsbergstrabe 32
7000 Stuttgart 1, GERMANY
0711/128-2849
Jason Makansi
Executive Editor
Power Magazine
11 West 19th St., 2nd floor
New York, NY 10011
212/337-4074
Rene Manga1
Engineer
Ontario Hydro
Research Division
800 Kipling Avenue
Toronto, Ont., M8Z 5S4 CANADA
416/231-4111,X6162
Mansour Hansour
President
Applied Utility Systems, Inc.
1040 East Chestnut Avenue
Santa Ana, CA 92701
714/953-9922
John Marion
Mgr.,Fuel Systems Development
ABB Combustion Engineering
Kreisinger Development Laboratory
1000 Prospect Hill Road
Windsor, CT 06095-0500
203/285-4539
B. L. Marker
N/A
New York State Electric & Gas
P. 0. Box 3607
Binghamton, NY 13902
607/729-2551
Eugene Marshall
Principal Engineer
Pacific Corp Electric Operations
14007 West North Temple
Salt Lake City, UT 84140
801/220-2235
Greg Marshall
District Sales Manager
Foster Wheeler Energy Corp.
2001 Butterfield Road, Ste. 206
Downers Grove, 1L 60515-1050
708/241-2050
John Marshnl1
Manager
Riley Stoker Corporation
45 McKeon Road
Worcester, MA 01613
508/792-4826
G, B. Martin
Deputy Director
U.S.Environmental Protection Agency
Air & Energy Engrg.Research Lab
MD-60
Research Triangle Park, NC 27711
919/541-2821
Sadahira Maruea
Mgr., Business Development
Nippon Shokubai America, Inc.
101 East 52ml Street
New York, NY I 0022
212/759-7890
Doug MaxweJI
Principal Research Engineer
Southern Company Services
P. 0. Box 2fi25
Birmingham, AI. 35202
205/877-7614
A-20
-------
Michael Maxwell
Chief, Gas Clean.Tech.Branch
U.S.Environmental Protection Agency
AEERL (MD-04)
Research Triangle Park, NC 27711
919/541-3091
Phil May
N/A
Radian Corporation
P.O. Box 1300
Research Triangle Park, NC 27709
N/A
T. J. Hay
Planning Project Manager
Illinois Power Co.
500 S. 27th St.
Decatur, IL 67S25
217/424-6706
Michael McCartney
Dir.,Fuel Systems Controls Engrg.
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095-0500
203/285-4677
John McCoy
Senior Consultant
Electricity Supply Board Internet'1
Stephen Court
18/21 St. Stephen's Green
Dublin 2, IRELAND
353/01 785-155
Mark HcDanne1
Vice President & General Manager
Carnot
15991 Red Hill Road
Suite 110
Tustin, CA 92680-7388
714/259-9520
Barry McDonald
President
Carnot
15991 Red Hill Road
Suite 110
Tustin, CA 92680-7388
714/259-9520
Michael McElroy
Project Manager
Electric Power Technologies, Inc.
695 Oak Grove Ave.
Menlo Park, CA
415/322-0843
Marilyn Mcllvaine
Managing Editor
Mcllvaine Company
2970 Maria Ave.
Northbrook, IL 60062
708/272-0010
Robert Mcllvaine
President
Mcllvaine Company
2970 Maria Ave.
Northbrook, IL 60062
708/272-0010
John McKie
Environment.!] Engineer, Senior
Virginia Dept.Air Pollution Control
6225 Brandon Ave., Suite 310
Sprinfield, VA 22150
703/644-0311
William McKJnuey
Vice Pres.,New Business Develop.
United Catalysts, Inc.
P. 0. Box 32370
Louisville, KY 40232
502/634-7218
Robert McMurry
Design Engineer
Duke Power Company
500 S. Church Street
Charlotte, NC 28262
704/373-6346
Tom McNay
N/A
Cincinnati Gas & Electric
P. 0. Box 960
Cincinnati, 01! 45201
513/632-2676
A-21
-------
Guntcr Mechtersheimer
Mgr., Environmental Technologies
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-2853
David Meier
Sales Manager, Utilities ,
Beltran Associates, Inc.
1133 East 35th Street
Brooklyn, NY 11210
718/338-3311
James Meyers
Chemical Equipment Engineer
Detroit Edison Company
2000 Second Ave., H-128A WSC
Detroit, MI 48226
313/897-0806
Paolo Michelotti
Engineer
F.T.C. Legnano
Via Monumento, 12 - Legnano
Legnano 20025
ITALY
0331/522 111
Charles A, Miller
Mechanical Engineer
U.S.Environmental Protection Agency
Air & Energy Engineering Res.Lab
MD-65
Research Triangle Park, KC 27711
919/541-2920
Katherine Miller
Environmental Engineer
Virginia Dept.Air Pollution Control
801 Ninth & Grace Streets
Richmond, VA 23219
804/786-1433
John Hincy
Market Development Manager
Nalco Fuel Tech
P. 0. Box 3031
Naperville, 1L 60566-7031
708/983-3258
Tadahisa Miyasaka
Chief Representative
Electric Powrr Development Co.
1825 K St., N.W.,Suite 1205
Washington, Of* 20006
202/429-0670
Cal Mock
General Sales Manager
Babcock & Wilcox
3333 Vaca Valley #300
Vacaville, CA 95688
707/451-1100
Larry Monroe
Head, Combustor Research Group
Southern Research Institute
P. 0. Box 55305
Birmingham, Al. 35255-5305
205/581-2879
Ed Moore
R&D Manager
Hauck Manufacturing Co,
P. 0. Box 90
Lebanon, PA 17042
717/272-3051
Terry Moore
Environmental Engineer, Senior
Virginia Dept.Air Pollution Control
7701 Timber Jflke Road
Lynchburg, VA 24502
804/947-6641
Bruce Morgan
Environments I Staff Engineer
Rust International
100 Corporate. Parkway
Birmingham, AT, 35243
205/995-7112
Mark Morgan
Mgr., Engrg. & Services
PSI Technology Co.
20 New Eng)nnrl Business Center
Andover, MA 01810
508/689-0003
A-22
-------
Dominick Mormile
Manager, Air Quality Control
Consolidated Edison of N.Y.
4 Irving Place
New York, NY 10003-3586
212/46Q-6275
Per Horsing
Hgr.DeNQx Technology
Haldor Topsoe A/S
Nymollevej 55
DK-2800 Lyngby
DENMARK
+45/45 27 2000
Herman Mueller-Odenwald
N fk
Kraftanlagen Heidelberg
c/o AUS 1140 East Chestnut Ave
Santa Ana, CA 92701
714/953-9922
Paul Musser
Program Manager
U.S. Department of Energy
Fossil Energy, FE-232 GIN
Washington, DC 20585
301/353-4346
Lawrence Muzio
Vice President
Fossil Energy Research Corp.
23342-C South Pointe
Laguna Hills, CA 92653
714/859-4466
Ram Nayak
Principal Mechanical Engineer
Stone & Webster
Three Executive Campus
P. 0. Box 5200
Cherry Hill, NJ 08003
609/482-3582
Lewis Neal
President
N0XSO Corporation
P. 0. Box 469
Library, FA 15129
412/854-1200
Mike Nelson
Senior Engxnesr
Southern Company Services
P. 0. Box 2625
Birmingham, AL 35202
205/870-6518
Sumitra Ness
Research Engineer
University of North Dakota
Energy St Environ.Research Center
15 North 23rd Street
Grand Forks, ND 58202
701/777-5213
Richard Newby
Principal Engineer
Westinghouse STC
1310 Beulah Road
Pittsburgh, PA 15235
412/256-2210
Julie Nicholson
Principal Engineer
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-3745
Satashi Nonako
Manager
Mitsubishi Heavy Industries America
1000 Prospect Hill Road
Windsor, CT 0^095
203/2B5-249!
Dave Nott
Special Projects Supervisor
Central Illinois Light Co.
300 Liberty Street
Peoria, It,' 6 I6D2
309/697 -14]2
Jim Nylandr.r
Senior Engineer
San Diego 0ns &. Electric
4600 Carlsbad Blvd.
Carlsbad, CA ^2008
619/931-7294
A-23
-------
James O'Brien
Project Engineer
Pennsylvania Power & Light
Two North Ninth Street (N-5)
A1lentown, PA 18101
215/774-4352
John O'Leary
N/A
Nalco Fuel Tech
2001 W. Main St., Suite 295
Stamford, CT 06902
203/323-8401
Raymond 0*Sullivan
Manager, Power Engineering
Orange & Rockland Utilities, Inc.
One Blue Hill Plaza
Pearl River, NY 10965
914/577-2630
George Offen
Program Manager
Electric Power Research Institute
3412 Hlllview Avenue
Palo Alto, CA 94304
4156/855-8942
Earl Oliver
President
Oliver Associates, Inc.
2049 Kent Drive
Los Altos, CA 94024
415/964-4838
Paul Orban
Engineer, Boilers
Pennsylvania Electric Co.
1001 Broad Street
Johnstown, PA 15907
814/533-8537
Robert Orchowski
Sr. Compliance Assurance Engr.
Duquesne Light Co.
One Oxford Centre
301 Grant Street
Pittsburgh, PA 15279
412/393-6099
Case Overduin
Supervising Engineer
Southern California Edison
2131 Walnut Grove Avenue
Rosemead, CA 91770
818/302/8323
Louis Paley
Compliance Monitoring Coordinator
U.S.Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460
703/308-8723
Y.S. Pan
Project Manager
U.S. D0E/PETC
P. 0. Box 10940
Pittsburgh, PA 15236
412/892-5727
Paul Paris!
Development Engineer
Union Carbide
P. 0, Box 700
Polnte-aux-Trembles
Quebec HID 5A8 CANADA
514/640-7400,X1296
Reginald Parker
Environment*1 Engineer
NYSDEC
50 Wolf Rood
Albany, NY 12233
518/457-2044
Raraesh PoteJ
Principal Engineer
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor. CT 0^095
203/285-2027
Roy Payne
Senior Vice President
Energy & Environmental Research
18 Mason
Irvine, CA 92718
714/859-8851
A-24
-------
David Pearsall
Product Manager
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT
203/285-5127
Jarl Pedersen
Manager
Burmeister & Wain Energy
23, Teknikerbyen
DK-2830 Virum DENMARK
+45/4285 7100
Thomas Penn
Mgr., Generating Projects
Potomac Electric Power
1900 Pennsylvania Ave., N.W.
Washington, DC 20068
202/872-2446
Henry Pennline
Chemical Engineer
U.S. Department of Energy
PETG
P. O. Box 10940
Pittsburgh, PA 15236
412/892-6013
Michael Perlsweig
Program Manager
U.S. Department of Energy
Fossil Energy, FE-232 GTN
Washington, DC 20585
301/353-4348
Mildred Perry
Group Leader, Flue Gas Chem
U.S. DOE/PETC
Box 10940
Pittsburgh, PA 15236
412-892-6015
Karin Persson
Chemical Engineer
Swedish Energy Development Corp.
Biblioteksgaten 11
S-11146 Stockholm
SWEDEN
+468 679 8610
Henry Phillips
N/A
Consultant
22 Beacon Hill Drive
Metuchen, NJ 08440
201/549-0332
Richard Phil Hps
Engineer
Union Electric Co.
1901 Chouteau Ave,
St. Louis, HO 63103
314/554-3485
Robert Philp
Research Coordinator
Energy, Mines & Resources Canada
555 Booth Street
Ottawa, Ontario K1A 0G1
CANADA
613/996-2175
Matthew Pler.hocki
Contract Manager
Babcock & Wi lcox.
20 S. Van Bnren Ave
Barberton, Oil 44203-0351
216/860-1704
Bill Pi erce
District Sales Manager
Babcock & Wilcox
3333 Vaca Vfillny Parkway
Suite 300
Vacaville, CA 95688
707/451-1100
Larry Piersoti
Project Man,igor
Babcock & Wilcox
20 S. Van Bnrwi Ave.
Barberton, OH 'i4203
216/860-J 103
Jack Pirkey
Principal Research Engineer
Consolidated Ediaon of N.¥,
4 Irving Plnn^
New York, NY 10003
212/460-250'.
A-25
-------
William Pitman
Environmental Engineer
Tennessee Valley Authority
400 W. Summit Hill Drive '
Knoxville, TN 37902-1499
615/632-6699
E, L. Plyler
N/A
U. S.Environmental Protection Agency
AEERL (MD-54)
Research Triangle Park, NC 27711
N/A
John Fohl
Senior Scientist, Energy
W. J. Schafer
8001 Irvine Center Drive
Suite 1110
Irvine, CA 92718
714/753-1391
Terry Poles
Director, Market Development
Engelhard Corporation
101 Wood Ave
Iselin, NJ 08830
908-205-6633
Robert Porter
Ass't,Project Manager
TransCanada PipeLines
55 Yonge Street, Bthfloor
Toronto, Ontario M5E 1J4
CANADA
416/B69-2161
John Protapas
Senior Project Manager
Gas Research Institute
8600 W. Bryn Mawr Avenue
Chicago, IL 60631
312/399-8301
Edward Preast
Project Manager
Florida Power & Light
P. 0, Box 14000
Juno Beach, FL 33408-0420
407/694-3112
Shaik Qader
Project Manager
Ebasco Services, Inc.
3000 West MacArthur Blvd.
Santa Ana, CA 92704
714/662-4093
Greg Quartucy
Engineer
Fossil Energy Research Corp.
23342 C South Points
Luguna Hills, CA 92653
714/859-4466
Brian Quil
Mechanical Engineer
Naval Energy & Envir.Support Activ
NEESA-11A
Port Huenemn, CA 93043-5014
805/982-3512
Les Radak
Senior Research Engineer
Southern California Edison
2244 Walnut Orove Avenue
Rosemead, CA 91770
818/302-9746
G.P. Rajendran
Research Chemist
E. 1. Du Pont de Nemours
P. 0. Box 80302
Wilmington, 11R 19880-0302
302/695-27R4
Jay Ratnfii-nrown
Dir..Environmental Projects
SAIC
1710 GoodrJdge Dr.
Box 1303
McLean, VA 22102
703/448-6343
William Reamy
EnvironmentsI Engineer
Baltimore Gas & Electric Co.
1000 Brandon Shores Road
Baltimore, HIV 21226
301/787-5378
A-26
-------
James Reese
Manager
Applied Utility Systems, Inc.
1140 East Chestnut Avenue
Santa Ana, CA 92701
714/953-9922
Christopher Rellly
St. Engineer, R&D
New York State Electric & Gas
4500 Vestal Parkway, East
Binghamton, NY 13902-3607
607/729-2551,X4105
Anthony Renk
Supervising Engineer
Florida Power & Light
P. 0. Box 078768
West Palm Beach, FL 33410
407/640-2289
Diane Revay Madden
Project Manager
U,S. D0E/PETC
P. 0. Box 10940
Pittsburgh, PA 15236-0940
412/892-5931
Cathy Rhodes
Public Health Engineer
Colorado Dept. of Health
4210 East 11th Ave.
Denver, CO 80220
303/331-8570
Michael Rini
Sr. Consulting Engineer
ABB Combustion Engineering
Kreisinger Development Lab
1000 Prospect Hill Road
Windsor, CT 06095-0500
203/285 -2081
J. R. Rizza
President
Todd Combustion., Inc.
61 Taylor Reed Place
Stamford, CT 06906
203/359-1320
Rodney Robertson
Project Manager
Burns & McDonnell
P. 0. Box 419173
Kansas City, M0 64141
816/822-3062
Chris Roble
Consulting Engineer
United Engineers & Constructors
P. 0. Box 5888
Denver, CO 80217
303/843-2803
Farzan Roshdieh
Senior Engineer
Applied Utility Systems, Inc.
1140 East Chestnut Avenue
Santa Ana, CA 92701
714/953-9922
Geoff Ross
Senior Program Engineer
Environment Canada
Industrial Programs Branch
Ottowa, Ontario KlA 0H3
CANADA
819/997-1222
Edward Rubin
Professor
Carnegie Mellon University
Schenley Park
Pittsburgh, PA 15213
412/268-5897
Dave Rundstrom
Research Scientist
Southern Ca1i fornia Edison
2244 Walnut Grove Ave.
Rosentead, CA 91770
818/302-956!
Pia Rydh
Chemical Engineer
Vattenfall Knergisystem AB
Box 528
16215 Vallingby
SWEDEN
+46/8 739 55 68
A-27
-------
Joseph Saliga
Systems Engineer
Fluor Daniel, Inc.
200 W. Monroe St,
Chicago, IL 60606
312/368-3862
Pia Salokoskl
Engineer
Imatran Voima OY
Rajatorpan tie 8 P. 0. Box 112
SF-01601 Vantaa
FINLAND
358/0 508 4837
N. C, Samish
Staff Research Engineer
Shell Development Co.
P. 0. Box 1380
Houston, TX 77251
713/493-7944
Howard Sandler
Principal
Sandler & Associates
111 Paciflca, Ste. 250
Irvine, CA 92718
714/727-2676
Emelins Sandoval
Engineer
Pacific Gas & Electric Co.
One California St., F-836D
San Francisco, CA 94106
415/973-5422
Edmund Schindler
Project Manager
Todd Combustion, Inc.
61 Taylor Reed Place
Stamford, CT 06906
203/359-1320
Richard Schlager
Div.Head, Environmental Sciences
ADA Technologies, Inc.
304 Inverness Way South, Suite 110
Engiewood, CO 80112
303/792-5615
Henry Schreiber
Project Manager
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2505
David Schulz
Regional Power Expert,
U.S.Environmental Protection Agency
Region 5
230 S. Dearborn - 5AC-26
Chicago, IL 60604
312/886-6790
Herbert Schuster
N/A
Deutsche Babcoek Energie
Duisburgerstr. 375
D-4200 Oberhaussen
FEDERAL REPUBLIC OF GERMANY
N/A
Blair Seckington
Supervising Engineer
Ontario Hydro
700 University Avenue
Toronto, Ontario MSG 1X6
CANADA
416/592-519!
Charles Sedmnti
Chemical Engineer
U. S . F.nvl ronmcni.al Protection Agency
AEERL (flD-04)
Research Trinngle Park, NC 27711
919/54,1-7700
James Seebold
Staff Engineer
Chevron Corporation
100 Chevron Wny
Richmond, CA 14802-0627
415/620-33)3
Tim Seelaiis
Mgr., Business Development
Pure Air
Two Windsor PJoza
Allentown, PA 18195
215/481-5373
A-28
-------
Daniel Seery
Sr. Program Manager
United Technologies Research Center
Silver Lane
East Hartford, CT 06108
203/727-7150
Dave Shilton
Senior Environmental Engineer
Pacific Power & Light
920 SW 6th Ave., Suite 1000
Portland, OR 97204
503/464-6479
Gary Shiomoto
Engineer
Fossil Energy Research Corp.
23342 C South Pointe
Laguna Hills, CA 92653
714/859-4466
Dale Shore
Program Manager
Radian Corporation
7 Corporate Park, Ste, 240
Irvine, CA 992714
714/261-8611
J. M. Shoults
Manager, Permitting
Texas Municipal Power Agency
Environmental Affairs
P. 0. Box 7000
Bryan, TX 77805
409/873-2013
William Siegfriedt
Director, Process Engineering
Fluor Daniel, Inc.
200 W. Monroe Street
Chicago, IL 60606
312/368-3828
Ralf Sigling
Engineer
Siemens/KWU
llammerbacher Str, 12 + 14
Erlangen 8520 GERMANY
01149/9131-18-6169
Paul Singh
Sr. Vice President
Procedair Industries
625 President Kennedy
Montreal, Quebec H3A 1K2
CANADA
514/284-0341
Bill Smith
Combustion Specialist
Burns & McDonnell
P. 0. Box 419173
Kansas City, M0 64141
816/822-3074
Chris Smith
Proposal Mgr., Burner Systems
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-5573
David Smith
Senior Chemist, Environment
Saskpower Corporation
2025 Victoria Avenue
Regina, Sask. S4P 0S1
CANADA
306/566-2290
J. W.R. Smith
Gen, Mgr., Rales & Marketing
Babcock Energy Ltd.
11 The Doul«.vnrd
Crawley, W. Sussex RH10 1UX
UNITED KINGDOM
0293/528755
Ken Smith
Engineer
Southern California Edison
2700 Edison Wny
Laughlin, NV
702/298-1197
Lowell Smith
Vice President
ETEC
One Technology, Suite 1-809
Irvine, CA 92718
714/753-9126
A-29
-------
Hart in Smith
Coal Research Establishment
British Coal Corporation
Stoke Orchard
Cheltenham, Glos
ENGLAND
0242 673361
Todd Sommer
Vice President, Engineering
EER Corp.
1645 N. Main St.
Grrville, OH 44667
216/682-4007
Robert Sommerlad
Mgr., Combustion Tech.
Research-Cottrell Envir.Serv/Tech.
P. 0. Box 1500
Somerville, NJ 08876
908/685-4776
John Sorge
Research Engineer
Southern Company Services
P. 0. Box 2625
Birmingham, AL 35202
205/877-7426
Arend Spaans
Engineer
Stork Boilers
Postbus 20
7550 GB Hengelo
THE NETHERLANDS
31/74-401328
David Speirs
Principal Engineer
ABB Combustion Engineering
99 Bank Street
Ottawa, Ontario KIP 6C5
CANADA
613/560-4458
Barry Speronello
Principal Development Scientist
Engelhard Corporation
Henlo Park CN40
Edison, NJ 08818
908/205-5155
Cindy Spittler
Marketing Manager
Radian Corporation
50 Century Blvd.
Nashville, TN 37214
615/885-4281
Hartmut Spliethoff
Scientific Assistant
University of Stuttgart
IVD Institute
Pfaffenwaldring 34
7000 Stuttgart 80 GERMANY
49/711-685-3396
Christopher Stala
Project Mgr..Advanced Materials
Gas Research Institute
8600 W. Bryn Mawr Avenue
Chicago, IL 60631
312/399-8233
Susan Stamey-Hall
Staff Scientist
Radian Corporation
3200 E. Chapel Hill Rd
P. 0. Box 13000
Research Triangle Park,NC
919/541-9100
James Staudt
Mgr., NOx Control
PS I Technology Co.
20 New Englnnd Business Center
Andover, flA 01810
508/689-0003
Richard Storm
V.P., Technical Services
Flame Refractories, Inc.
Highway 742
P.O. Box 649
Oakboro, NC 28129
704-485-3371
Richard P. Storm
Senior Service Engineer
Flame Refractories, Inc.
P. 0. Box 649
Oakboro, NC 28129
704/4B5-3371
A-30
-------
Joseph Strakey
Assoc.Dir..Clean Coal :
Pittsburgh Energy Tech. Center
P. 0. Box 10940
Pittsburgh, PA 15236
412/892-6124
Peter Straiigway
R&D Consultant
Niagara Mohawk Power Corp.
300 Erie Blvd., West, A-2
Syracuse, NY 13202
315/428-6532
Sabine Streng
N/A
Lentjes AG
Hansa-Allee 305
4000 Dusseldorf
GERMANY
N/A
Lamar Sumerlin
Principal Engineer
Southern Company Services
P.O. Box 2625
Birmingham, AL 35202
205-870-6519
Kohei Suyama
Project Manager
Mitsubishi Heavy Industries
2 Houston Center, Suite 3800
Houston, TX 77010
713/654-4151
Timothy Sweeney
Supervisor
Foster Wheeler Energy Corp.
Perryville Corporate Park
Clinton, NJ 08809
908/730-5436
Thomas Szymanski
Mgr., Product Research
Norton Company
P. 0. Box 350
Akron, OH 44309
216/673-5860
Masaki Takahashl
Visiting Researcher
MIT/EPDC
One Amherst Street
Cambridge, MA 02139
617/253-7828
Harry Tang
Sr. Research Engineer
Shell Development Co.
P. 0. Box 1380
Houston, TX 77251-1380
713/493-8424
Tai Tang
Associate Engineer
KIN Engineering & Applied Sciences
1034 NW 57th Street
Gainesville, FL 32605
904/331-9000
Roberto Tarll
Manager
ENEL
Production & Transmission Dept.
Via A. Pisano, 120
56100 Pisa, ITALY
0039/50-535754
Robert Teetz
Mgr..Chem.Div.,Env.Engrg.Dept.
Long Island Lighting Co.
P. 0. Box 426
Glenwood Landing, NY 11547
516/671-6744
Donald Teixeira
Tech. Mgr., Fossil R&D
Pacific Gas & Electric Co,
3401 Crow Cnnyon Road
San Ramon, CA 94583
415/866-5531
Preston Tempnro
Plant Manager
KPL Gas Service
Mile Post #30
P.O. Box 249
Lawrence, KS 66044
913-843-8118
A-31
-------
Angela Testa
Visiting Researcher
Eniricerche (Italy)
c/o HIT - Chemical Engineering
60 Vassar St., Bldg. 31-261
Cambridge, MA 02139
617/253-1721
Paul Thompson
President
Tenerx Corporation
P. 0. Box 1444
303 Laurel
Friendswood, TX 77546
713/482-5801
Richard Thompson
President
Fossil Energy Research Corp.
23342 C South Pointe
Laguna Hills, CA 92653
714-859-4466
David Thornock
R&D Engineer
ABB Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-2931
Richard Tischer
Project Manger
U.S. Department of Energy
P. 0. Box 10940
Pittsburgh, PA 15102
412/892-4891
Majed Toqan
Prog.Mgr., Prin.Research Engineer
Massachusetts Instit. of Technology
Dept. of Chemical Engineering
60 Vassar St., Bldg, 31-261
Cambridge, HA 02139
617/253-1721
Ian Torrens
Department Director
Electric Power Research Institute
3412 Hlllview Avenue
Palo Alto, CA 94304
415/855-2422
A-:
H.ILJ, Tossaint
Mgr., Combustion Engrg.
Stork Boilers
P. 0. Box 20
7550 GB Hengelo (0)
THE NETHERLANDS
31/74 40 1015
Donald Toun
Advisory Engineer
Babcock & Wilcox
20 S. Van Vuren
Barberton, OH 44203
216/860-1986
Shiaw Tseng
Project Engineer
Acurex Corporation
P. 0. Box 13109
Research Triangle Park, NC 27709
919/541-3981
Lance Turcotte
Assoc. Consulting Engineer
Ebasco Services, Inc.
759 South Federal Highway
Stuart, FL 34994-2936
407/225-9476
Henry Turner
Utility Plant Manager
IBM
P. 0. Box 218
Yorktown Ht, NY 10598
914/945-3,720
Minoru Uchid/t
Mgr., Nuclear Project Dept.
Chiyoda Corporation
12-1 Tsurumichuo 2-Chome, Tsurumi
Yokohama, JAPAN
045/506-7062
Toshio Uemnrrt
Senior Engineer/Combustion Systems
Babcock-Hitnchi K.K.
No. 6-9 Takara-machi
Kure-city, I!Iroshima-prefectur
JAPAN
0823/21-1163
-------
Andy Uenosono
Senior Project Coordinator
Hitachi America, Ltd.
2000 Serra Pt, Parkway"
Brisbane, CA 94005-1835
415/244-7602
K. Ueshima
Ass't. Mgr..Environ,Plant Engrg.
KHI/Joy Environmental Equipment
1-1, Higashi Kawasaki-cho 3-chome
Chuo-ku, Kobe
JAPAN
078/682-5230
David Underwood
Vice President, Sales
Aptec'
RD 1, Box 583
Honey Brook, PA 19344
215/942-3651
James Vader
Project Manager
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2316
Mohammad Vakill
Project Manager
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2541
James Valentine
President
Energy & Environmental Partners
480 Hemlock Road
Fairfield, CT 06430
203/254-7166
Bauke Van Kalsbeek
Vice President
Sierra Environmental Engineering
3505 Cadillac Avenue, K-l
Costa Mesa, CA 92626
714/432-0330
Bill Van Nlfitiwenhuizen
N/A
Babcock & WilnoK
581 Coronation Blvd.
Cambridge, Ontario N1R 5V3
CANADA
519/621-2130
Michel Vandycke
Head, Chemical Engineering
Stein Industrie
19-21, Av. Morana Saulnier
78141 Velizy-Villacoublay
Cedex, FRANCE
34-65-46-02
Joel Vatsky
Dir., Combustion & Environ.Systems
Foster Wheeler Energy Corp.
Perryville Corporate Park
Clinton, NJ 08B09-4000
9087/730-5450
Dahlgren Vaughan
Environmental L Engineer
Virginia Dept.Air Pollution Control
300 Central Rd,,Suite B
Fredericksburg, VA 22401
703/899-4600
Gary Veerkamp
Sr. MechanicsJ Engineer
Pacific Gas & Electric Co.
One California, Room F827
San Francisco, CA 94106
415/973-1576
Denise Viola
Commercial Manager
Engelhard Corporation
101 Wood Avenue
Iselin, NJ 08830
908/205-5039
Gary VonBargen
Project Engineer
Wisconsin F. 1 or.trie Power
P.O. Box 204ft
Milwaukee, Wt 53201
414-221-2310
A-33
-------
Peter Waanders
Contract Manager
Babcock & Wilcox
20 5. Van Buren Ave.
Barberton, OH 44203
216/860-1967
Frederick Wachtler
Project Manager
Foster Wheeler.Energy Corp.
Perryville Corporate Park
Clinton, NJ 08809
908/730-5438
Paul Wagner
Project Engineer
Delmarva Power
195 & Route 273
P.O. Box 9239
Newark, DE 19714
302-454-4844
Peter Warne
Senior Instrumentation Engineer
Monenco Consultants Ltd.
Power Division
400 Monenco Place, 801 6 Ave., S.W.
Calgary, Alberta T2P 3W3 CANADA
403/298-4678
Kevin Washington
Power Resources Staff Specialist
Florida Power S> Light
6001 Village Blvd.
West Palm Beach, FL 33407
407/640-2412
Richard Waterbury
Principal Engineer
Florida Power & Light
16423 79th Terrace, N.
Palm Beach Gardens, FL
407/747-7643
Robert Weimer
Chief Engineer
Air Products and Chemicals, Inc.
7201 Hamilton Blvd.
Allentown, PA 18195
215/481-7626
Steven Weiner
Program Manager
Air Products and Chemicals, Inc.
7201 Hamilton Blvd.
Allentown, PA 18195
215/481-4372
M. Weiss
MgrGenerating Systems Engr.
Potomac Electric Power
1900 Pennsylvania Ave., N.W.
Washington, DC 20068
202/872-2431
Tom White
Project Engineer
Sargent & Lundy
55 E. Monroe
Chicago, 1L 60603
312/269-6137
Kenneth Wildmnn
Development Engineer
Eastman Kodak Co.
Kodak Park Bldg 31
Rochester, NY 14652-3709
716-477-0666
Donald Wilhelm
Sr. Chemical Engineer
SFA Pacific, Inc.
444 Castro St., Suite 920
Mountain View, CA 94041
415/969-8876
Ronald Wilknlss
N/A
Mobil Oil Corporation
3700 W. 190th Street
Torrance, CA 90509
213/212-4587
Steve Wilson
Principal Research Engineer
Southern Company Services
P.O. Box 2625
Birmingham, At. 35202
205/877-7835
A-34
-------
Phil Winegar
Senior Engineer
New York Power Authority
Research & Development
1633 Broadway
New York, NY 10019
N/A
H. B, Wylie
Senior Engineer
Baltimore Gas & Electric Co,
1000 Brandon Shores Road
Baltimore, HO 21226
301/787-5245
Anthony Yagiela
Cyclone Reburn Project Manager
Babcock & Wilcox
20 S. Van Buren Avenue
P. 0. Box 351
Barberton, OH 44203-0351
N/A
Larry Winger
Mgr., New Ventures
Engelhard Corporation
101 Wood Avenue
Iaelin, NJ 08830
908/205-5266
Johan G. Witkamp
Project Manager
KEMA
Utrechtseweg 310
6900 ET Arnhem
THE NETHERLANDS
085/56 3625
James Wittmer
Supervisor, Project Mgmt.
Central Illinois Light Co,
300 Liberty Street
Peoria, IL 61602
309/693-4840
James Wolf
Senior Engineer
Virginia Power
5000 Dominion Blvd.
Glen Allen, VA 23060
804/273-2617
Brian Wolfe
District Manager
Babcock & Wilcox
7401 West Mansfield, #410
Lakewood, CO 80235
303/988-8203
Gregg Worley
Environmental Engineer
U.S.Environmental Protection Agency
345 Court land St., N.E.
Atlanta, GA 30365
404/347-2904
Misao Yamamura
Mgr., NO.2 Land Boiler.
Mitsubishi Heavy Industries
1-1 Akunoura-Machi
Nagasaki 850-91
JAPAN
81/958-28-6400
Ralph T. Yang
Chair, Dept. of Chem. Engineering
State University of N.Y. at Buffalo
Buffalo, NY 14260
716/636-2909
Shyh-Ching Yang
Mgr.,Energy Resources Laboratories
Industrial Toch. Research Institute
Bldg.64,195 Roc.4, Chung Hsing Rd.
Chutung Hslnchu, Taiwan
REPUBLIC OP ClUNA 31015
886/35-916439 /
James Yeh
Chemical Engineer
U.S. Department of Energy
P.O. Box 10940
Pittsburgh, PA 15236
412-892-5737
Cher if Youss«f
Research Project Engineer
Southern California Gas Co
Box 3249 Terminal Annex
ML 73ID
Los Angeles, CA 90051
818-307 -2695
A-35
-------
Kenneth Zak
Development Associate
W. R. Grace & Co. - Conn.
7379 Route 32
Columbia, MD 21044
301-531-4383
Kent Zamrait
Project Manager
L.A. Department of Water & Power
111 N. Hope St.,Room 931
Los Angeles, CA 90012-2694
213/481-5019
Aldo Zennsro
Combustion Engrg.Manager
Ansaldo Component!
Via Sarca 336
Milan 20126
ITALY
010392/6445 2204
Jim Zhou
N/A
Babcock & Wilcox
581 Coronation Blvd.
Cambridge, Ontario N1R 5V3
CANADA
519/621-2130
Qian Zhou
Research Engineer
N0XS0 Corporation
P. 0. Box 469
Library, PA 15129
412/854-1200
A-36
-------
1, REPORT NO. 2.
EPA-600/R-92-093 c
3. RECIPIENT'S ACCESSION'NO,
4. TITLE AND SUBTITLE
Proceedings; 1991 Joint Symposium on Stationary
Combustion NOx Control, Washington, D. C. , March
25-28, 1991, Volume 3. Sessions 6-8
5J3EPOBT DATE
i July 1892 ]
6. PERFORMING ORGANIZATION CODE
7, AUTHOR(S)
Carolee DeWJtt, Compiler
S, PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
William Nesbit and Associates
1221 Farmers Lane
Santa Rosa, California 95405
10. PROGRAM ELEMENT NO.
11. CONTRACT /GRANT NO.
NA (EPRI Funded)
12. SPONSORING AGENCY NAME, AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, North, Carolina 27711
13, TYPE OF REPORT AND PERIOD COVERED
Proceedings; 3/89 - 3/91
14. SPONSORING AGENCY CODE
EPA/600/13
TECHNICAL REPORT DATA
15.supplementary notes AEERL project officer is Robert E. Hall, Mail Drop 65, 919/541-
2477. Volume 1 includes Sessions 1-3, and Session 2 includes Sessions 4 and 5.
16, abstract proceedings document the 1991 Joint Symposium on Stationary Combus
tion NOx Control, held in Washington, DC, March 25-28, 1991. Jointly sponsored by
EPRI and EPA, the symposium was the sixth in a biennial series devoted to the
international exchange of information on recent technological and regulatory develop-
ments for stationary combustion nitrogen oxides (NOx) control. Topics covered inclu
ded the significant increase in active full-scale retrofit demonstrations of low-NOx
combustion systems in the U. S. and abroad over the past 2 years; full-scale oper-
ating experience in Europe with selective catalytic reduction (SCR); pilot- and bench-
scale SCR investigations in the U.S. ; increased attention to selective noncatalytic
reduction (SNCR) in the U, S. ; and NCx controls for oil- and gas-fired boilers. The
proceedings are published in three volumes.
7,
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDEO TERMS
c. COSATI Field/Group
Pollution
Nitrogen Oxides
Combustion
Catalysis
Fossil Fuels
Pollution Control
Stationary Sources
Catalytic Reduction
Noncatalytic Reduction
13 B
07B
2 IB
07D
21D
8, DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21, NO, OF PAGES
432
20. SECURITY CLASS {This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
A-37
-------
------- |